Theory of Mind in Preadolescence: Connections Among Social Understanding, Self-Concept, and Social Relations

THEORY OF MIND IN PREADOLESCENCE: CONNECTIONS AMONG SOCIAL UNDERSTANDING, SELF-CONCEPT, AND SOCIAL RELATIONS

Sandra Leanne Bosacki

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Education Ontario Institute for Studies in Education of the University of Toronto

O Copyright by Sandra Leanne Bosacki 1998 National Library Bibliothwue nationale l*P of Canada du Canada Acquisitions and Acquisitions et Bibliographic Services sewices bibliographiques 395 Wellington Street 395, rue Wellington Ottawa ON KIA ON4 OttawaON KIA ON4 Canada Canada

The author has granted a non- L'auteur a accorde me licence non exclusive licence allowing the exclusive permettant a la National Library of Canada to Bibliotheque nationale du Canada de reproduce, loan, distniute or sell reproduire, preter, distriiuer ou copies of this thesis in microfom, vendre des copies de cette these sous paper or electronic formats. la fonne de microfiche/filrn, de reproduction sur papier ou sur format electronique.

The author retains ownership of the L'auteur conserve la propriete du copyright in this thesis. Neither the bitd7auteur qui protege cette these. thesis nor substantial extracts fkom it Ni la t]h&seni des extraits substantiels may be printed or otherwise de celle-ci ne doivent Etre imprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation. Abstract Theory of mind in preadolescence: Connections among social understanding, self-concept, and social relations Doctor of Philosophy, 1998 Sandra Leanne Bosacki Graduate Department of Education Ontario Institute for Studies in Education of the University of Toronto

The investigation of the ability to attribute mental states to others and to see them as the basis for people's actions has been referred to as "theory of mind (TOM) research. That is, TOM research attempts to explain how children come to understand social action in both themselves and others. Despite the rising interest in this approach to social understanding, empirical evidence gleaned from the early adolescent years remains sparse. Accordingly, the purpose of this study is to assess social understanding in preadolescents from a TOMperspective, and to determine if such understanding relates to concepts of self and social relations differently in girls and boys. It was hypothesized that preadolescents' ability to understand thoughts and emotions in others would be associated with their sense of self and their social-relational ability.

One hundred and twenty-eight preadolescents @ = 11;9 years; 64 girls, 64 boys) completed a self-concept questionnaire, vocabulary tasks, peer social competence ratings and participated in an interview that involved two ambiguous social narratives followed by questions to assess social and self-understanding. Teacher ratings of the participants' classroom behaviour were also collected. Based on composite social story scores, results indicated positive associations between children's social understanding and a) perceptions of their behavioural conduct, b) ability to understand their own mental states and feelings and c) social competence ratings. Separate gender analysis revealed different correlational patterns among the variables. More specifically, multiple regression and multivariate analyses showed that among girls, global self-worth (i.e., reported personal happiness) acted as a moderator between social understanding and social competence ratings whereas this was not found for boys. Results are discussed in relation to (I) individual differences in social understanding, self- concept and social relations and (2) effects of socio-cultural context. Educational implications include theoretical foundation for developments in holistic curriculum for preadolescents. In conclusion, the present findings may increase educators' awareness of the inner world of girls and boys in the sixth grade which in turn may further the discourse on social cognition and holistic education. ACKNOWLEDGEMENTS

I am grateful to my advisor, Janet Wilde Astington for her constant guidance and support throughout this endeavour. I should also like to thank the other members of my thesis committee, David Booth and Mary Lou Arnold for providing me with valuable insights and advice. I also wish to express my thanks to the additional members of my examination committee, Anne McKeough, Jack Miller, and David Olson for their comments on the manuscript I wish to thank Susan Elgie for her personal time and statistical expertise and Deepthi Kamawar, Theresa McCarthy, and Janette PeUetier for their support on both academic and personal levels. Finally, a special thanks goes out to Mr. Dennis Tschirhart and to all the principals, teachers, and students who participated in this study. In particular, I am grateful to the Institute of Child Study Laboratory School in Toronto, Ontario for their participation in the pilot study and to the Brant Haldimand-Norfolk Catholic District School Board for their participation in the main study. Without their cooperation, the completion of this thesis would not have been possible. Above all, I wish to thank my parents and sister Leslie, for their constant emotional support over the past three years. In times of uncertainty, they provided me with the love and strength I needed to continue my work TABLE OF CONTENTS .. Abstract II Acknowledgements iv Table of Contents v List of Tables X List of Figures xii ... List of Appendices XUl I. Introduction 1.1. General Problem 1 -2. Present Study 1-3. Sipficance of This Study 1.4. Outline of Subsequent Chapters 11. Literature Review 2.1 . Introduction 2.2. Theory of Mind: A New Approach to Social-Cognitive Development 2.3. Theories of Theories of Mind 2 -4. Overview of Theories of TOM 2.5. Theory of Mind in Childhood 2 -6. Preadolescent Theory of Mind: A Case for Complexity 2.6.1. Conceptual Role-Taking 2.6.2. Empathetic Sensitivity 2.6.3. Person Perception 2.6.4. Self-conception and Self-Understanding 2.6-5. Language, Communication and Social Relations 2.7. Social Understanding, Self-concept and Social Relations in Preadolescence 2.8. Gender Effects 2.9. Summary m. Research Questions and Hypotheses 3.1. Introduction 3.2. Research Questions 3 -3. Hypotheses 3 -4. Rationale of the Research Design N. Method 4.1 . Introduction 4 -2. Instnunentation 4.2.1. Development of New Social Understanding Measures 52 4.2.2. Assessment of an Understanding of Higher-Order Mental States 53 4.2.3. Codhg of Social Understanding Narratives (Social and Self- Understanding Interview - Part A - Interpersonal Understanding) 6 1 Scoring of Conceptual Role-taking Questions (Section A) Interval level data 64 Nominal or categorical level data 66 Scoring of Empathetic Sensitivity Questions (Section B) Interval level data 66 Nominal or categorical level data 68 Scoring of Person Perception Questions (Section C) Interval level data 68 Nominal or categorical level data 69 Scoring of Alternative Explanation (Section D) Interval level data 70 Nominal or categorical level data 70 4.2.4. Total Social Understanding Story Score 7 1 4.2.5. Between-S tory Correlations and Differences 73 4.2.6. Self-concept 74 Self-perception Profile for Children (SPPC) 75 Social and Self-understanding Interview (part B) - Intra- personal Understanding 76 Coding of the self-understanding interview 77 Assessment of understanding self-esteem as a concept 79 Response coding for self-esteem question 79 4.2.7. Social Relations 80 Teacher's Rating Scale of Child's Actual Behaviour (TRF) 80 Peer-rated Social Competence 8 1 Peer-rated Likability (Per Acceptance) 82 4.2.8. Vocabulary Measures 82 Mental State Verbs Task (MVT) 83 Scoring of MVT 84 Gates-MacGinitie Vocabulary Test 84 4.2.9. Summary of Instnunents 85 4.3. Participants 85 4.4. Procedure 87 4.4.1. Establishing Contact and Consent 4.4.2. Administration of the SPPC, SCNF, Peer Likability Scale, Gates-MacGinitie Vocabulary Test and the MVT 4.4.3. Administration of the Social and Self-understanding Interview (Inter-Intrapersonal Understanding) Debriefing 4.4.4. Summary of Data Analyses Results: Preliminary Analyses 5.1. Overview 5.2. Descriptive Statistics 5.2.1. Distribution of Scores 5.2.2. Assessment of Social Understanding 5.2.3. Order Effects of Tasks 5 -2-4. School Effects 5.2 -5. Gender Main Effects 5 -2-6. Teacher-Gender X Student-Gender Interaction Effects 5.3. Correlational and Regression Analyses 5 -3.1. Social Understanding and Self-Concept 5 -3-2. Mental State Verb Understanding, Self-Concept and Social Relations 5 -3-3. Social Understanding and Self-understanding 5 -3-4. Social Understanding and Social Relations Social understanding and peer-ratings of social competence Social understanding and teacher-ratings of children's behaviour 5 -3-5. Self-Concept and Social Relations (Peer and Teacher Ratings) 5.3.6. Sibling Effects 5.4. Summary Results: Hypothesis Testing of Mediation and Moderation Models 6.1. Overview 6.2. Testing Self-Concept as a Mediator 6.3. Testing Self-Concept as a Moderator 6.3.1. Perceived Global Self-worth as a Moderator Between Social Understanding and Social Relations Characteristics of higMow groups based on social understanding ability and perceived global self-worth of

vii girls and boys 6.4. Summary W. Results: Further Analyses of Social Understanding Stories and Self-Understanding Interview 7.1. Overview 7.2. Social Understanding Stories 7.2.1. Emotional Valence of Social Understanding Story Responses 7.3. Self-understanding Interview 7.3.1. Analyses of Perceived Global Self-worth Justifications 7.3.2. Analyses of Self-Esteem Definitions 7.4. Summary VPII. Discussion Introduction Support for the Hypotheses 8.2.1. Social Understanding, Self-Concept and Social Relations Intervening role of self-concept between social understanding and social relations 8.2.2. Gender Differences 8.2.3. Additional Important Findings 8.2.4. Additional Si,@icant Findings (Categorical Level Data) Meaning of Findings (Interval Level Data) 8.3.1. Relations among Social Understanding, Self-Concept and Social Relations 8.3.2. Social Understanding and Mental State Verb Understanding 8.3.3. Social and Self-Understanding 8.3.4. Social Understanding and Social Relations 8.3.5. Social Understanding and Teacher Ratings 8.3.6. Self-Concept as Mediator or Moderator 8.3.7. General Gender Differences 8.3.8. Siblings Meaning of Findings (Categorial Level Data) 8.4.1. Responses to Social Understanding Stories 8 A.2. Responses to Self-Understanding Interview Role of Language summary IX. Educational Implications 9.1. Overview 9.2. Introduction 9 -3. Whom Does This Include? 9.4. What Should Be Taught 9.5. How? Teaching and Learning Activities 9.6. Assessment Tools 9.6.1. Teacher Rating Scale of "Mindreading" Classroom Behaviour 9.6.2. Metacognitive Task 9.8. Summary X . Conclusions 10.1. Introduction 10.2. Methodological Issues 10.2.1. Limitations 10.3. Limitations of the Measures 10.3.1. Self-concept 10.3.2. Social and Self-Understanding 10.3.3. Social Relations (Peer and Teacher Ratings) 10.4. Future Directions/Challenges 10.5. Concluding Statement XI. References XII. Appendices A. Coding Guides and Examples of Transcripts B. Instruments C. School Letters and Consent Forms List of Tables -Title 1. Interpersonal Understanding Interview: Social Understanding Story 1 2. Interpersonal Understanding Interview: Social Understanding Story 2 3. Intercomlations Among Social Understanding Story Subscales 4. Means, Standard Deviations, and Gender Effects for Social Understanding Story SubscaIes Means, Standard Deviations, and Gender Effects for SPPC Self-Concept Subscales and Self-Unders'mdhg Interview Means, Standard Deviations, and Gender Effects for SPPC Teacher's Rating Scale of Child's Actual Behavior Means, Standard Deviations and Gender Effects for Peer Ratings of Social Relations and Vocabulary Measures Pearson Correlations Between Gates-MacGinitie Vocabulary Test (1994) and Social Understanding Stories, Self-Concept and Self-understanding Pearson Correlations Between Gates-MacGinitie Vocabulary Test (1994) and Social Relations (Peer and Teacher Ratings) and Mental State Verb Understanding Means and Standard Deviations for Social Understanding Story Subscales and Univariate Analyses of Teacher-gender X Student-gender Pearson Correlations Between Social Understanding Story Total Score, Self-Concept and Self-Understanding Pearson Correlations Between Social Understanding Story Subscales and Self-Concept Hierarchical Multiple Regression Analysis Predicting Social Understanding Using Gates-MacGinitie Vocabulary Test (1994) and Self-Concept Pearson Correlations Between Social Understanding Story Total Score and Social Relations (Peer and Teacher ratings) and Vocabulary Measures Pearson Correlations Between Social Understanding Story Subscales and Social Relations (Peer and Teacher Ratings) Hierarchical Multiple Regression Analysis Predicting Social Understanding Using Gates-MacGinitie Vocabulary Test (1994) and Teacher's Rating Scale of Child's Actual Behavior Hierarchical Multiple Regression Analyses Testing Perceived Behaviour Conduct as a Mediator Hierarchical Multiple Regression Analyses Testing Perceived Global Self-worth as a Moderator with Social Relations as the Outcome Variable Mean Social Relations Scores for Girls and Boys as a Function of Social Understanding and Perceived Global Self-worth Frequency Percentages of Positive Responses to Question 3b of Social Understanding Stories Frequency Percentages of Total Self-Understanding Categorical Responses for High and Low Social Understanding Groups of Girls and Boys Within-Gender Frequency Percentages of Total Self-Understanding Categorical Responses for High and Low Social Understanding Groups Means, Standard Deviations, and Effects of Self-Esteem Definitions for Peer Social Ratings, Social and Self-Understanding and Perceived Global Self-worth The Co~ectingCurriculum: Curriculum for the Psychological and Social Lives of the Preadolescent List of Figures -Title Page 1. Theoretical Models of Relations Among Social Understanding, Self-concept and Social Relations 45 List of Appendices -Title Al. Coding Guide and Examples of Transcripts for Social and Self-Understanding Interview (Part a) - Interpersonal or Social Understanding Social and Self-Understanding Interview (Part b) - Intrapersonal or Self-Understanding Self-Perception Profile for Children (SPPC; Harter, 1985) Social and Self-Understanding Interview Student Directions Social aad Self-Understanding Interview (Part b) - Intrapersonal or Self-Understanding Teacher's Rating Scale of Child's Actual Behavior (TRF,Harter, 1985) Social Competence Nomination Form (adapted fiom Ford, 1982) and Peer

Likability Scale (adapted from Matthews & Keating, 1995) Mental State Verbs Task (MVT, adapted from Astington & Olson, 1990; Booth & Hall, 1994) Gates-MacGinitie Vocabulary Test, D 516 (Form 4) (MacGinitie & MacGinitie, 1994) Student Directions

Gates-MacGinitie Vocabulary Test, Level D 516 (Form 4) (MacGinitie & MacGinitie, 1994) Cover Letter to School Principal Cover Letter to Classroom Teacher Declaration of Informed Consent for Teacher Class Introduction to Research Study Cover Letter to Parents Declaration of Informed Consent by ParentIGuardian Debriefing Instructions Thank-You Letter to Parents CHAPTER I Introduction 1.1. General Problem Over the past decade, social-cognitive research has increasingly come to envision the child as an interpretative psychologist (Astington & Olson, 1995; Bennett, 1993; Erwin, 1993). That is, such research views the child as an intersubjective theorist pruner, 1996); one who depends on a mentalistic construal of reality to make sense of the social world. Based on the collective works of various social-constructivists (e.g., Berger & Luckmann, 1967; Bruner, 1990; Case, Okamoto, Griff~n,McKeough, Bleiker, Henderson, & Stephenson, 1996; Gergen, 1985; Vygotsky, 1965) and symbolic-interactionists (e.g., Baldwin, 19 13; Cooley, 1912; Mead, 1934), such an approach proposes that children come to understand or make meaning from their experiences ,~dedby the tenets of relativism, constructivism, narrative and self agency. Moreover, the Vygotskian notion that cognitive growth sterns from social interaction is congruent with holistic or humanistic and psycho-cultural approaches to development (Bruner, 1996). This approach draws on various theories that assume children rely on both emotional and co,onitive structures to make sense of the world (e.g., Hobson, 1991; Piaget, 1963). Such an integrative approach may assist researchers to answer the increasingly common question of how children come to make meaning from their social experiences and eventually become "soc-emotionally literate" or socially intelligent (e.g., Gardner, 1991; Goleman, 1995; Levinson, 1995). Thus, the larger question becomes what conceptual framework can provide a unifying theory that emphasizes the interactions among thought, emotion and action? In search of such a theory, social scientists have recently begun to investigate the social- cognitive underpinnings of the ability to understand the social world. The main goal of such research is to find a theory that will assist in their exploration of how children acquire the knowledge that others are thinking and feeling beings. Accordingly, over the past decade, many researchers have approached the area of social cognition from what is referred to as a "Theory of Mind (TOM)perspective. This unique way of viewing social understanding has also been referred to as folk psychology, commonsense psychology or belief-desire reasoning (see Moore, 1996). A TOMperspective on social understanding is unique in that it is founded on the premise that all humans are folk or commonsense psychologists. That is, humans understand social information by means of ascribing mental states to others and thizlking that overt behaviour is governed by these states. This ability to "read" others' minds and to predict how people will act in social situations, focuses on the understanding of mental states such as beliefs, desires and intentions (Moore, 1996). More specifically, to understand social behaviour, children must first understand mental representation. That is, they must understand that there is a difference between thoughts in the mind and things in the world (Astington, 1993). By inferring mental states from people's actions, children learn to understand that minds are active and contain mental states that can bring about events in the world. Thus, the same world can be experienced in different ways by different people. Each person may have a distinctive belief about reality. This relatively new approach to social cognition claims that a largely implicit conceptual himework containing intentional elements allows children to understand, explain, and predict their own and other people's behaviour and mental states (Von Eckardt, 1994). In line with this view is the widely held assumption that this metallizing ability allows children to make sense of social behaviour by ascribing desires and intentions to others' actions for the specific purpose of regulating their interactions with others (Astington & Gopnik, 1991). Moreover, it is believed that the ability to recognize, represent and understand others' thoughts and emotions in early childhood provides the socialcognitive foundation for the later development of social and emotional competency (Lalonde & Chandler, 1995; Slomkowski & Dunn, 1996). Interestingly, although the interest in the development of folk psychology has been paralleled by an interest in the social-cognitive processes of the preadolescent (10- to 13-year-old) (e-g., Alsaker, 1995; Brooks-Gum, 1989; Harter, 1993; Matthews & Keating, 1995), the two research areas have failed to connect. Perhaps one of the greatest impediments that has prevented researchers iiom aciap~ga Toki approach to preadoiescent sociai co,oniuon has 'been the iac~of conceptual agreement among TOMtheorists. Examples of some of the ongoing conceptual debates include the argument of how exactly a "theory of mind" develops beyond preschool, and what exactly the processes that develop (Moses & Chandler, 1992; Flavell & Miller, 1997). Although TOM research could enrich investigations of preadolescent social cognition, particularly in the areas of self-concept (Gopnik, 1993; Wellman, 1990), perspective-taking (both affective and cognitive) (Baron-Cohen, 1995; Chandler, 1987; Perner, 1988) and person perception (Yd, 1992% 1992b, 1993), the two research areas have continued to work independently without building on each other's findings. The school is a complex, social institution, where children's ability to make sense out of social information plays a crucial role in self-development and subsequent social relations (Bruner, 1996). Given such a data-rich environment, one would assume that the context of the elementary school classroom could provide the ideal test-frame for a TOM approach to social understanding. However, to this date, there have been no attempts to delineate the links between social reasoning and social relations in preadolescents, based on the folk psychological tenets of belief, desire and action. Despite the claim that early adolescence, is a pivotal time in many areas of social-cognitive development including cognitive reflexivity (e.g., Piaget, 1963), self-concept formation (e-g.,

Damon & Hart, 1988; Erikson, 1968; Harter, 1985) and interpersonal relations (e.g., Rosenberg, 1989; Selman, 1980), a TOM approach to exploring the links between such social-cognitive areas remains to be taken. Such an inquiry may promote a better understanding of the two main tasks of preadolescence which are: 1) the intrapersonal task of constructing a coherent psychosocial identity (Damon & Hart, 1988; Erikson, 1968; Rosenberg, 1989), and 2) the interpersonal task of understanding multiple and contradictory intentions of others, allow&g judgments to be made in an uncertain world (Chandler, 1987). Thus, drawing on various social-cognitive (Selman, 1980, Weiner, 1985) and epistemological theories and research (Gilligan, 1983; Perry, 1970), a folk psychological approach to social cognition may help to illustrate the linkages among the uncierstanding of meniai states in others, seif-concept and sociai reiarions. 4

1.2. Present Studv Given this communication problem between two seemingly similar areas of research (i.e., similar in their investigation of attribution of mental states to oneself and others), the purpose of this study is twofold. First, based on folk psychological principles, this study develops a task to assess preadolescents' ability to attribute mental states to others. Second, using this task, this study explores and describes the relations among the ability to attribute mental states to others, concepts of self and social relations. For the purposes of the present study, these three core constructs are operationally defined below. First, in the present study, social understanding is defined as the ability to attribute mental states (i.e., beliefs, intentions, emotions) to others (Astington, 1993). More specifically, this mentalizing ability refers to understanding multiple perspectives, emotional states and the concept of a "person." (Chandler, 1987; Weban, 1990). In this study, social understanding is assessed by a story-telling interview describing socially ambiguous events. Specifically, children were asked to reason and make inferences concerning the intentional and emotional states of the protagonists. Some clarification of the use of the terms "social understanding" and "theory of mind is required here. As noted by previous researchers (e.g., Astington, 1993; Moore, 1996), the term "theory of mind can refer to both the mentalizing ability to attribute mental states to others and as a label for a particular theoretical approach that studies social understanding. Such an approach is concerned with the development of a form of knowledge involving folk psychological or intentional elements (as elaborated in chapter 2). Also, the terms "theory of mind and "social understanding" are often equated and can also be defined as attributing mental states to both self and other. Given that the present study views the understanding of self and other as separate constructs, the present study will refer to social understanding only as the ability to attribute mental states to others. Second, self-concept is defined as the ability to judge one's competence in various a,.--:-, ,.. -.,.ll ..., 4, A,#.Ul., ,,A :..,&&. ,,,I, ,,-,,:-.,A --.,dl. ,, ,-,-,, 'PI-:.. :, >,C,,A ,, UVU-3, LW WbU LW LV UbDbLLLJb QllU JUOULJ VllG 3 PLbGLVbU WVlLU 43 4 PLSVll. 11113 13 UGLlllGU 43 5 the affective aspect of the self and is assessed by Harter's (1985) Self Perception Profile for Children (SPPC). This affective aspect of the self-concept, or how much one likes oneself as a person, has also been referred to as "self-esteem" (e.g., Harter, 1996) and thus the present study sometimes uses the two terms interchangeably. The cognitive aspect of the self is defined as the extent to which children understand their self-judgements (self-understanding) (Harter, 1985), and is assessed by a semi-structured interview, with questions based on the children's responses to the SPPC. Third, the term social relations is defined as the general ability to interact socially with others. To avoid the conceptual and methodological confusion surrounding the term b'social competence" (see Oppenheimer, 1989; Schaffer, 1995, for reviews), the present study operationally defines the term social relations as a combination of one's popularity or peer acceptance and one's abiliiy to act in a hypothetical social situation. To assess this global social- relational ability, the study employed both peer and teacher ratings of classroom social behaviour. Peer ratings are assessed by an adapted version of the Social Competence Nomination Form

(SCNF) (Ford, 1982, Matthews & Keating, 1995) including a peer likability scale that assessed children's popularity or likability. Teacher ratings of children's behaviour are assessed by Harter's (1985) Teacher's Rating Scale of Child's Actual Behaviour . The present study will use the term "social competence" only in reference .to the scores obtained from the peer ratings on the adapted version of the SCNF (see section 3.2). In addition to these three main constructs, this study tests for general vocabulary ability using the Gates-MacGinitie Vocabulary Test (MacGinitie & MacGinitie, 1994) and for specialized vocabulary ability using the Mental State Verbs Task (adapted from Astington & Olson, 1990; Booth & Hall, 1994; Siddiqui, 1995). These tasks are included to test for a theoretical link between language development and the study's main variables of social understanding, self- concept and social relations. The central hypothesis of this study is that significant associations exist among social understanding, self-concept and social relations, with self-concept either acting -- .. -- -na--+n- w u LuwuaLwL WL L~WU~A~LWL. 1.3. Simificance of This Studv Based on the combination of a psycho-cultural (Bruner, 1996) or ethnopsychological approach to TOMand the gender intensification hypothesis (Hill & Lynch, 1983), the findings from this research offers a detailed look at the dynamics among social understanding, self-concept and social relations in preadolescence. That is, few studies on preadolescent cognition take into account the importance of the self-concept as an intervening variable between social understanding and social relations in both girls and boys. Furthermore, although there exists a plethora of studies on the general constructs of social understanding, self-concept and social relations, this investigation is unique in that it specifically tests how self-concept intervenes between the two constructs of social understanding and social relations. Finally, with regard to TOMresearch, this study adds to what little research examines the development of an "interpretive" or "constructivist" TOMin older children and adolescents. Through the use of a newly designed narrative story task, the current study shows that such a task can be seen as both 1) a source for data for understanding preadolescents' awareness of mental states in others and 2) a possible vehicle for the development of such an awareness. This study also furthers the discourse on TOMby focusing on gender issues, which are often ignored or inadequately assessed in the few TOMstudies that do involve older participants (Fabricius, Schick, Prost & Schwanenflugel, 1997; Fox, 1991; Lalonde & Chandler, 1997; Nelson, Plesa & Henseler, 1998). Thus, through the integration of multiple theoretical perspectives and related empirical evidence of both social-co,onitive research and theory of mind research, this study aims to bridge the gap between the two domains by discussing ways of linking theory to guide future research on social understanding in older children and adolescents. The educational sigmficance of this study is that it suggests the application of psychological research to practical issues of pedagogy and curricdum development. This study may help to increase educators' awareness and understanding of the inner lives of preadolescent girls and boys. That is, the study may also help educators to become more cognizant of the different ways in

--WIILL;u .*-. - *- @i & boys-m&s~mG & &2T3i'ei swi.. &,wie&e. Ad&iJon;liiY, &,hgs2cjm he study also highlight the role language and metacognition play in preadolescents' abilities to reason about mentd states in self and others. Such findings provide the groundwork for future curriculum development that employs narrative and metacognitive strategies to promote both interpersonal and intrapersonal understanding in preadolescent girls and boys. This study, then,

a may further the discourse on theory of mind research and gender issues in social-cognitive development. In line with other psychologists interested in both the affective and cognitive aspects of

education (e.g., Bruner, 1996; Coles, 1990,1997; Donaldson, 1992; Donelan-McCall & Dunn, 1996; Gardner, 1991), the present study provides support for a balanced curriculum, one that values both the intellect and the emotions. Moreover, the present findings emphasize the value of understanding the link between mind, emotion and language in the school milieu. More specifically, the current study's use of a socially ambiguous narrative task to assess social understanding suggests that such a task may also shed some light on how preadolescents create their own subjunctive reality (Astington, 1990; Bruner, 1986). That is, ambiguous text invites readers to create their own meanings partly from their imagination, including their emotions and feelings of self-worth and partly from the text. Such findings may encourage teachers to engage students in activities that promote interpretive understanding and intentional creativity (Schank & Cleary, 1995). For example, activities that require awareness of and reflections on one's own thoughts and feelings, such as self-narration, bibliotherapy, and psychodrama, allows a kind of self-experimentation that may promote understanding of one's own mind and how it relates to others. In conclusion, this study borrows from the areas of both co,onitive science and social development in the hope that such a multidisciplinary approach to social co,onition may provide some fresh insight into the connections between the mental and social life of the preadolescent. Thus, the study highlights the need for a collaborative effort on behalf of both educators and psychologists to continue to investigate the subtleties and complexities of the preadolescent mind and experience. 8

1.4. Outline of Subseauent Cha~ters Following this introduction, Chapter 2 reviews the research that has been done in the areas of social understanding, self-concept and social relations. Due to the vastness of each of the research areas, the chapter focuses on research conducted during the preadolescent years and describes the areas of overlap between the two research fields of Theory of Mind (TOM) and social cognition. Chapter 3 outlines the specific research questions and hypotheses of the present study. Chapter 4 describes the methods employed in the study. It describes the participants, the task and the procedures adopted. This chapter also describes the limits of the common methods used to assess social understanding and provides the rationale for the development of a new social understanding measure. A large section is devoted to the development and coding of the social and self understanding measures because they were specifically designed for this study. Finally, the methods of statistical analyses are outlined. Chapter 5,6 and 7 describe the results of the present study. Chapter 5 describes the preliminary analyses performed on the measures including testing for school and task order effects. Chapter 5 also reports the results obtained from correlational and multiple regression analyses. Chapter 6 describes the multivariate analyses performed to test whether self-concept moderates or mediates the relation between social understanding and social relations. Finally, Chapter 7 reports additional analyses to further describe results found in Chapter 6. Chapter 8 contains the discussion based on the results presented in Chapter 4-7. The results are first presented with regard to the main hypotheses and then the findings are interpreted. Chapter 9 discusses the educational implications of these findings, outlining both pedagogical and curricular issues. Chapter 10, the final chapter, outlines methodological issues and the limitations of the skdy. The chapter ends with suggestions for future research and final conclusions. CHAPTER 11 Literature Review 2.1. Introduction Despite the fact that early adolescence is a pivotal time in both the mental and social world of girls and boys, the relations between these two worlds remain to be thoroughly explored. In particular, the ways in which girls and boys learn to understand social behaviour may influence their thoughts and feelings about themselves, and how they interact with others. Thus, the ability to attribute mental states to others may have different implications for the self-concept and social relations of preadolescent gir!s and boys. Briefly defined, social cognition is the understanding and knowledge of people and their doings (Flavell & Miller, 1997). More specifically, it is concerned with how children conceptualize others and how they come to understand the emotions, intentions and perspectives of others (Erwin, 1993). Given that numerous approaches to the study of social cognition exist (e-g., social-information processing: Dodge, 1983; neo-Piagetian: Case et al., 1996, etc.), it is beyond the scope of this study to present a detailed, theoretical discussion of social-cognitive development. The purpose of this chapter is to view the area of social co,@tion from a TOMperspective by sketching conceptual and empirical areas of interest relevant to the present study. The absence of conclusive evidence and contradictory findings presented in this section shows that further research is needed. This review will outline fruitful points of contact between recent work or; children's understanding of mind and related topics in social co,onition. Such an overview will illustrate how theory of mind research can provide some further insight on understanding the preadolescent mind.

2.2. Theorv of Mind: A New Approach to Social-Coaitive Develo~ment As the cmnt Zeitgeist continues to emphasize the importance of the social context in the study of child development (Astington & Olson, 1995; Tomasello, Kruger, & Ratner, 1993), theories of folk or commonsense psychology are being re-written and findings are being re- interpreted from an interpretive rather than a causal view, emphasizing the social and emotional implications of thought. The theories originally proposed by Cattaneo (1864, cited in Mugny & Carugati, 1989) and later advanced by Baldwin (1913) and Mead (1934) which claim that the development of higher cognitive processes is a consequence of social interactions have been actively investigated over the past seventy years - mainly beginning with Piaget's (1929, 1963, 1965) work on children's understanding of mental states. Until about 1983, this research on children's inner, mental worlds had been defined as social-cognitive research and classified as either 1) attribution-theoretical, metacopitive or 2) in the general Piagetian tradition of perspective or role-taking (Flavell & Miller, 1997). However, following Premack and Woodruff's (1978) article on "theory of mind" in chimpanzees, there appeared a surge of studies on children's abilities to attribute mental states to themselves and others, beginning with Wirnmer and Pemer's (1983) classic study of children's representation of false belief. To date, theory of mind (TOM) research is one of the most active research areas in developmental psychology (Moore, 1996), and currently holds the greatest preeminence in the field of social-cognitive development (Flavell & Miller, 1997). Although TOMresearch currently suffers from conceptual disputes regarding structural and developmental issues (Moses & Chandler, 1992; Karmiloff-Smith & Russell, 1994), many writers claim that it provides a fruitful way to investigate the complex field of social cognition. In particular, Flavell and Miller (1997) have explicitly outlined the many ways in which a theory of mind approach to social cognition distinguishes it from its predecessors. However, it must be noted that a complete picture of the comparisons between the two research areas is beyond the scope of this literature review. What follows are a brief summary of the main distinguishing features of TOMresearch that are relevant to the current study. First, TOMresearch offers great potential to unify investigation of self and other understanding within the framework of social relations by providing a dynamic developmental theory that emphasizes thought, emotion and action (Leekam, 1993; Moore, 1996). Flavell and Iviiiier note that whiie the broader fieid or" sociai cognition is grounded in Baiciwin's sociaiiy oriented theory that claims social and self understanding is an interpersonal process carried out by a social mind through collaborative activities (1902; cited in Damon & Hart, 1988), TOMresearch is more specifically focused on how children come to understand mental states and how such knowledge is applied. Second, the underlying theoretical and philosophical foundation of TOMresearch is in itself a feature that distinguishes the field from the broader area of social cognition. Theory of mind research stems from the theoretical basis of folk psychology based on Dennett's (1987) intentional stance that views people as rational agents who act to fulfill their desires, given their beliefs (Astington, 1993). More specifically, TOMresearch has tended to focus on the child's understanding of specific mental states such as beliefs and desires. Accordingly, TOMresearchers claim that this understanding of the representational nature of mind is central to socialcognitive development. Alternatively, social-cognitive research has focused on children's understanding of a subset of mental states such as intentions and emotions in slight of other major mental entities such as beliefs and desires (Flavell & Miller, 1997). The TOMapproach to social cognition is also distinctive in that it advocates the importance of the relation between thought and language (Astington, 1998). Specifically, the importance of the link between speech acts, mental states and how they interrelate with social communication. As explained in section 2.65, on one view, TOMis social understanding made explicit in metacognitive language. That is, we come to understand and predict social interaction by understanding mental states which are expressed in people's speech acts (Astington, 1993). By focusing on various aspects of pragmatics, TOMresearch offers a unique contribution to the area of social cognition. A TOM approach can thus address the way in which the structure of cognition, language and social relations serves to frame children's reality in specific ways, offering ways of creating experience. Thus, one could claim that the field of TOMresearch has in effect returned to Mead's (1934) problem of attempting to describe and unpack the nexus between mind, self and society. A final contribution of TOMresearch to understanding mental states is the particular focus on the fact that the world is not always as it seems. This focus on the ability to interpret and understand ambiguous sources of information has implications for various areas of social cognition, particularly areas that deal with self and other deception, and the understanding of complex emotions. For example, studies on how children interpret ambiguous events have shown that highly aggressive boys interpret ambiguous events as intentional rather than as possibly accidental, whereas normal boys do not (Dodge & Frame, 1982; Erdley & Asher, 1996). As school aged children develop an understanding of the mind as an interpretive device and become aware of informational ambiguity, they realize that events per se are sometimes ambiguous sources of information. Within the context of the present study, preadolescents' interpretations of socially ambiguous narratives may help to illustrate how preadolescents understand socially ambiguous situations and may add to the related socialcognitive studies of the understanding of social ambiguous actions and intentional reasoning (e.g., Dodge & Frame, 1982; McKeough, Yates & Marini, 1994).

2.3. Theories of Theories of Mind The TOMapproach to social understanding posits that in acquiring a folk psychology, children come to understand the mind as a representational entity that mediates between the individual and the world (e.g., Davies & Stone, 1995). Although considerable debate has surrounded how such a representational theory of mind is made possible and just exactly how it develops, it has been recently argued that any full theory of the development of TOM must entail three components of social understanding including the understanding of: other agents and their activities, self, and the interactions between self and other (e-g., Moore, 1996). Currently, the four main approaches to TOMdevelopment include the theory theory, the simulation theory, modularity or nativist theory and the most recent socioconstructivist or intersubjectivity theory (Astington, 1996; Moore, 1996). Due to the complexity of this area, a full treatment of these

.L--2-- 2- L -----2 .L- -C.L- .I---:.. TT--"- a2-c ------2---CAI-- .L--2-----ZVI I-- UIGVUC~13 u~yuuuUS abupc UI ULC UICDID. ncubc, uucr ~uiillc~UL UIC UICUIICS WUI UG presented, placing emphasis on the socioconstructivist approach believed to be the most relevant to a preadolescent theory of mind.

2.4. Overview of Theories of TOM All TOMtheories must deal with both the issues of TOMacquisition and its end-state, in other words how a mature individual "explains and predicts the actions of self and other" (Carmthers & Smith, 1996, p. 4). Although the aforementioned authors argue that these two issues are separate, I will address the various theories as approaches to TOMdevelopment, thus encompassing both acquisition and end-states and later, briefly comment on the ~plicationseach theory may have for development of self and other understanding in preadolescence. Firstly, modularity or nativist theorists propose that children's understanding of the mind is innate by claiming that children are born with modular, neurological mechanisms dedicated to mental state computations (e.g., Baron-Cohen, 1995; Carey, 1985; Fodor, 1992; Leslie, 1988). More specifically, modularity theorists agree that a TOMinvolves an innate system such as that proposed by Leslie (1994). Baron-Cohen's (1995) modules include the Intentionality Detector Theory of Mind mechanism (ToMM), the Eye-Direction Detector (EDD), and the Shared Attention

Mechanism (SAM) which code psychological relations between agents and objects (Barresi & Moore, 1995). With the exception of Leslie, the majority of modularity theorists would claim that the concept of self as potentially equivalent to others is built into the structure of the representational system, thus one could predict that self and other understanding remain equivalent throughout one's lifetime. In contrast, both theory-theorists (e.g., Gopnik, 1993; Gopnik & Wellman, 1995; Perner, 1991; Wellman, 1990) and simulation-theorists (e-g., Hanis, 1989; Johnson, 1988) place more emphasis on the role social experience plays in the development of self and other understanding, asserting that children's development of understanding human minds is not innate but acquired through the process of theorizing or leaming (although whether or not this process is cognitive,

: L --L --i xxn----- LL _ __ - x--- -L. -1- _ - awuu ul uvuleulaiu .-ru vc ~cau~vcq.vv llcrciw LIIG uuuwi1~yu~culy GGJGCSihr: roie oi experience in children's development of social understanding, both theory-theorists and simulation theorists support the general view of constructivist cognitive psychology which focuses on the adaptable and malleable nature of theory construction. However, the two views differ with regards to how this theory of mind is acquired with theory-theorists arO.uingthat children arrive at social and self knowledge via a process analogous to scientific theory construction consisting of a collection of intuitive and concrete lawddes that predicts and explains their own and others' behaviour. Thus, to understand action, the role of experience would remain indirect, informed by the postulated entity. Thus, according to the theory-theory view, human behaviours are explained by the application of postulated mental states to self and other. In opposition to the theory-theory view that claims humans understand the psychological relations of self and other in the same terms, simulation theorists (e.g., Gordon, 1986; Harris, 1989; Johnson, 1988) tend to focus on the child's own introspective access to her or his mental states. Such theorists argue that phenomenological awareness of such states provides a rich source of information about the mind that does not rely on postulated entities. It is argued that this "privileged access" sets knowledge about the mind apart from other kinds of knowledge that must be obtained by more indirect inferential process (Moses & Chandler, 1992). Thus, through the process of mental simulation where one imagines viewing the world from another person's perspective (Harris, 1989), it is the combination of social experience and mentalizing ability that guides a theory of mind as opposed to a theory of mind that ,&des one's social experience (theory- theory view). The modularity theory, theory-theory and simulation theory are similar in their emphasis on individual development and all share the view that children are creating or implementing a conceptual structure known as a theory of mind (Astington, 1996). Although these three different views continue to remain the most prominent in theory of mind research, there has recently been a growing interest in an alternative approach which proposes that children do not really acquire any theory of mind of their own but participate in social construction of the self and other ('aver &

---.. - --a ieaCoeaier, i995 j. Siemming &om various sociai theories (e.g., Iiincie, r Y IY; I revarthen, 1~IY; Vygotsky, 1965, 1978) that propose a social starting point for the mind, authors such as Bruner (1990) and Tomasello et al., (1993) claim that children learn to give meaning to what people do and internalize cultural narratives or norms about themselves and others by involving themselves within a particular culture. This sociocultural or social constructivist approach (Gergen, 1985) supports the recent self-organizational or dynamic systems approach to human development often mentioned in post-modem (e.g., Collin, 1996; Oppenheimer, 1989) and feminist literature (e.g., Belenky, Clinchy, Goldberger, & Tarule, 1986; Code, 1993), that claims that development is a complex system of interactions or processes between the thoughts and emotions of self and other (e-g., Barton, 1994). Proponents of such views of social understanding include those theorists (e.g., Dunn, 1988, 1995) who agree that a theory of mind is actually a theory of "mind-in-relationships" (Flavell & Miller, 1997) where the major cognitive task in social development is to develop mutual understanding or intersubjectivity (i.e., knowing other minds) (Levinson, 1995; Raver & Leadbeater, 1993). Such an approach views social understanding or knowing others' inner states not just as a conceptual process but as an affect-laden, interdirected action that occurs within a dynamic, social context (Hobson, 1991; Raver & Leadbeater, 1993). In particular, a constructivist, holistic approach to development which promotes a cognitiveemotional unity is further supported by Hobson's (199 1) claim that social understanding consists of both intersubjectivity and a conception of mental and physical attributes. A social-constructionist approach to a theory of mind supports the view that to survive in an increasingly complex and social world, children behave as constructivist narrators, continually interpreting and coconstructing meaning from social experience (e.g., Astington, 1990; Bruner, 1990; Fox, 1991). More specifically, the cognitive ability to understand human action and intentions has been referred to as the narrative mode of thought (Bruner, 1986; 1990; 1996). In contrast to logico-scientific or paradigmatic thought (thought that is concerned with physical reality and deals with issues of logic, analysis and truth), narrative thought is concerned with

..IcI.-~cIIcIA-~~s.rao+n Lrmn- ow-Aoo-nc +hn+Ann1 -4+h -orr+nl r+n+nr nrrA n-r.drrr.r XAn-nlvor YOJ UUUAU-UCU U V UUW) IIUAAAU UApAAULAUUU UAUL UW WILU AAAWUCLY ObULWU LUIU WAAAUUVW. ATAVA W V WA, narrative thought encompasses the ability to simultaneously recount events that take place in the physical world (i.e., landscape of action) and those that take place in the mental life of the characters (i.e., the landscape of consciousness) (Bruner, 1986). Furthermore, in line with the Vygotskian claim that thought is derived from social action, a TOMapproach to social understanding suggests that the understanding of narrative, mental states and social relations may all be linked and influence one another (Astington, 1990; McKeough, Templeton, & Marini, 1995; Nelson et al., 1998).

2.5. Theorv of Mind in Childhood For the past fifeen years, theory and research have accumulated on the development of a child's theory of mind. This work claims that between the ages of 3 and 5, children acquire a mentalistic model or theory to interpret and explain social interaction (e.g., Astington, 1993, Wellman, 1990). As noted earlier, this theory consists of their understanding of and ability to attribute mental states such as beliefs, desires and intentions to themselves and others. This commonsense ability to understand the representational nature of mind is based on the principles of folk or belief-desire psychology and has been found to develop first from an understanding of desire, to an understanding of belief, and fmdy to an understanding of a beliefdesire psychology (Bartsch & Wellman, 1995) or what Perner refers to as first-order representation (e.g., 1988, 1991; 1996). According to Perner this cognitive skill first enables one to represent concrete objects and events, followed by the ability to attribute first-order representations of mental states to oneself and others which later develops into the ability to attribute second-order representations of mental states (e-g., he thinks that she thinks the cookie has been moved) which occurs between the ages of 6 and 9 (Homer, 1995; Perner & Wimmer, 1985). Although this capacity to understand second-order mental representations plays a key role in social interactions (Homer & Astington, 1995; Sullivan, Zaitchik & Tager-Flusberg, 1994), some writers assert that this knowledge of mental representation (TOM) develops beyond second-

~4-t+- Lnlvvdn +ha nh:l&. +- .--dnrr+nrrd nLL.-+r +L:-A --Ar- ---.+-l -+-+I- :Il---~+-rlI-. &I- VAUVA LV UVAUUII LUII WULJ LV L.IIUUAJLLIUU uuu ULULUUL~ LILUUVIUC.L UGUW ~ULGS uwuaL- vy UG capacity to represent further recursions about beliefs about another person's belief (Bowler, 1992; Bowler, 1994; Happg, 1994; Happ6, Winner, & Brownwell, 1998; Perner, 1988). This more sophisticated level of recursive reasoning (Pemer & Wimmer, 1985) may also enable children to become progressively better at understanding social situations which include complex social concepts such as commitment (Mant & Pemer, 1988) and intentional deception (Leekam, 1993). Accordingly, Wellman (1990) suggests that these conceptual advances represent the increasing complexity and differentiation of a child's theory of mind as children become able to create specific theories within their general framework theory of mind including theories of self, intelligence and memory.

2.6. Preadolescent Theory of Mind: A Case for Com~lexitv Both social co,onitive (e.g., Bennett, 1993; Chandler, 1987, 1988; Gardner, 1985; Harter, 1993; Selman, 1980) and TOMtheorists (e.g., Perner, 1988; Wellman, 1990) agree that the ability to understand self and others within the context of social relations develops in complexity throughout one's lifetime. However, how this growth comes about and what it consists of remains debatable (e.g., Bennett, 1993; Brooks-Gum, 1989; Harter, 1993). Recently, various authors have emphasized that in order to acquire a better understanding of the concept of a general, overarching theory of mind, one must map out or chart what the adult or mature theory of mind consists of (Feldmm, 1992; Flavell & Miller, 1997; Leekam, 1993). Although a few attempts to apply the TOM approach to social understanding have been expanded to the preadolescent years (e.g., Chandler, 1987, 1988; Fabricius et al., 1997; Feldman, 1992; Mansfield & Clinchy, 1997; Nelson et al., 1998), the majority of literature written on preadolescent social cognition remains within the field of social or social-cognitive psychology. Furthermore, both Dunn (1988) and Levinson (1995) suggest that effective communication is dependent upon both the attribution of mental states to ourselves and others and the maintenance of a positive self-concept. Thus by focusing on the understanding of, and the coordination with, the perspective of others, a TOM qjpi3& ic sxi&GG@Gve &veiGpmefii &hep~~~&.c~fi~ my i0 zukdie Ge Ei&,ns between social understanding (theory of other minds and emotions), intrapersonal understanding (theory of one's own mind and emotions) and social relations (Gardner, 1985, Goleman, 1995). Social cognitive (Bennett, 1993; Chandler, 1987,1988; Gardner, 1985; Harter, 1993; Keating, 1990) and TOMtheorists (Pemer, 1988; Wellman, 1990) have reco,onized the significant role that higher-order mental states play in social interactions. Both groups of researchers contend that complex, reflexive reasoning skills are a prerequisite for the ability to understand self and others within the context of social relations. However, the issues surrounding how this reasoning ability develops, or what it consists of, lack consensus (Brooks-Gum, 1989; Harter, 1993). Despite the potential to shed light on the complex workings of the preadolescent mind, the two research paradigms of TOM and social cognition have failed to connect beyond the early school-age years. The majority of socialcognitive research still remains independent of TOM, including studies on attribution and perspective or role-taking (both of self and other) (e.g., Selman, 1980, 1989; Damon & Hart, 1988), person perception (e-g., Yuill, 1992); and empathetic sensitivity (e.g., Ferguson, Stegge, & Damhuis, 1991). Similarly, the realm of social-cognitive research has failed to integrate the mainly cognitive studies of higher-order mental processes among preadolescents (e.g., Boyes & Chandler, 1992; Miller, Kessel, & Flavell, 1970). For example, studies have shown that the emergence of relativist thought, or the process of becoming a reflective knower (Chandler, 1987), co-occurs with the ability to understand the meaning of promising (Astington, 1988), social commitment (Mant & Pemer, 1988), sarcasm and irony (Hap*, 1994; Keenan, 1995; Winner & Leekam, 1991), self-conscious emotions such as shame and guilt (Tangney, 1991) and metacognitive and metalin,.uiStic verbs (verbs that represent mental states) (Astington & Olson, 1990; Schwanenflugel, Fabricius, & Noyes, 1996). Moreover, despite the increasing interest in the development of a constructivist theory of mind beyond middle childhood (Astington, 1996; Fabricius et al., 1997; Gopnik, 1993; Lalonde &

Chandler, 1997; Mansfield & Clinchy, 1997; Schwanenflugel, Fabricius et al., 1996), at the time of this writing, there have been no studies that have attempted to conceptualize or systematically

siuciy iht: workings oiihe preuioiesceni mind as a dynamic, muiGaceted network OFcognitive and affective components that may serve as a template for self and other understanding. Furthermore, the present study's attempt to define social understanding in preadolescence within the framework of TOMresearch may shed some light on the wealth of findings from psycho-social studies that show a si,pificant drop in self-worth (e-g., Edwards, 1993; Gilligan,

1993; Harter, 1985) and an increase in self-consciousness (Buss, 1980; Simmons, Rosenberg, & Rosenberg, 1973) around the age of eleven. Although Chandler (1987,1988) attributes the development of relativistic thought as the main cause of generic self-doubt (implying a decrease in self-worth), more research is needed on preadolescent thought and social-emotional development ' to investigate both the drawbacks and benefits of developing an advanced ability to understand the minds or interior lives of others (Belenky et al., 1986). Moreover, despite the recent popularity of cultural psychology (Bruner, 1996), a large gap still exists in the TOMliterature concerning sociocultural issues such as gender (Astington, 1996; Lillard, in press). A social-constructivist TOMapproach to social understanding may contribute to recent f~ndingsthat have shown an increase in preadolescent stereotypic gender-role expectations and behaviours (e.g., Hill & Lynch, 1983; Unger & Crawford, 1992) and in the development of a negative self-concept among preadolescent girls (Brown & Gilligan, 1992; Callahan, Cunningham, & Plucker, 1994; Edwards,

1993; Matthews & Smyth, 1997; Taylor, Gilligan, & Sullivan, 1995; Unger & Crawford, 1992). Thus, such a lack of communication between disciplines advocates the need for gender differences

to be explored within the context of TOM(Mansfield & Clinchy, 1997). The need for an integrative, multilateral theory to explain an advanced theory of mind is supported by the assertion that the age of preadolescence (10-13) is the second period of individuation (Blos, 1967), when the fusion of interpersonal and intrapersonal understanding allows the preadolescent to form a sense of identity (Erikson, 1968; Gardner, 1985). The majority of literature on preadolescent social cognition has assumed that both self and person perceptions develop in parallel; that is, self- and other concepts arise simultaneously from social interactions, both develop in the same fashion and both share the same features (e-g., Baldwin, 1902; Mahler,

--- % D ----- n-ra A I*---L-.-I-. :- ..-----& -.h~-6~~- ---:-I ---:.:---.L--L-.- /n---- ?LG oc u~lpmu,17 I a~.nu~~uauv~~y, u apwu(;.ur wlru UUGL DVCI~QL~U~IUVGUIGUIWW \UQUUII 20

& Hart, 1988; Hatcher, Hatcher, Berlin, Okla, & Richards, 1990; Ruble & Dweck, 1995), a TOM perspective may provide an avenue for one to investigate the dynamic relations between these two processes. Until recently, little effort has been dedicated to the identification and explanation of factors associated with individual differences in the acquisition of a mentalistic understanding of one's own and other minds. Drawing on various theories of socialcognitive processes, particularly that of attribution theory (e-g., Weiner, 1985; Werner, 1948), conceptual role-taking

(Chandler, 1987; Piaget, 1929; Piaget & Inhelder, 1956; Selman, 1980), empathy (e.g., Harris,

1989; Hoffman, 1986; Tangney, 1991) and person perception (e-g., Barenboim, 1981; Livesley & Bromley, 1973) TOMresearch provides a framework from which to investigate how the understanding of self and other relates to social interaction. Given the philosophical foundations of folk psychology and Chandler's (1987) conceptual formulation of social understanding in preadolescence, researchers have yet to describe how this understanding influences self-concept and social relations in preadolescence. Such a framework will provide the opportunity to study the influences of a developing TOMon preadolescents' construals of other people, the self, and the reasons behind social behaviour. Similarly, the notion of TOM as an ability to co-construct or narrate one's social reality may also provide a framework in which to investigate the consequences of the process of becoming "perspectival", that is, becoming a constructivist epistemologist, may have on one's social and emotional development (Boyes & Chandler, 1992; Chandler, 1987). For example, although cognitive developmentalists have suggested that as children enter preadolescence they move from a dichotomous, "truelfalse" view of knowledge and mind to a more constructivist or "degrees of certainty" view (Fabricius, Schwaneneflugel, Kyllonen, Barclay, & Denton, 1989; Kitchener &

King, 1981; Schwanenflugel, Fabricius, & Alexander, 1994; Schwanenflugel et al., 1996), such a notion has failed to include aspects of emotional and social development such as self understanding or social relations. It is therefore important to investigate traditionally researched areas of preadolescent social understanding from the new perspective offered by TOMtheorists.

na.....Jle.. /Ioom h..- --.."-A +LA+.. -ell-Le-L-.- ,.----.. L 6 --.--A,. ---A--*--A:-- ..L- ~~~UALULW\r /o I 1 uw pupa- uur u buuawLauVG aypuabu ruwcuw uu~1~rauugLUG 2 1

preadolescent's mind may help to illustrate the social-cognitive and emotional processes that occur during preadolescence, when a shift from a realistic to a constructivistic epistemology occurs (e.g., Chandler, 1987; Erwin, 1993; Lalonde & Chandler, 1997). More specifically, Chandler suggests that the investigation of conceptual role-taking, empathetic sensitivity and person perception may provide a clearer picture of how preadolescents infer mental states in others. That is, the examination of these three constructs may help to explain the social-co,onitive processes underlying social understanding during preadolescence. Accordingly, the following section provides a brief overview of research findings in each of the three social-cognitive constructs. The section will focus on how a TOMapproach to social cognition can further the discourse on the preadolescent experience.

2.6.1. Conceptual Role-Taking As already noted by various TOMand social-cognitive researchers (Astington, 1993; Bennett, 1993; Flavell & Miller, 1997; Harris, 1989, Perner, 1991), research on children's understanding of mind is reminiscent of the notion of social role-taking or peispective taking in the

1960's and 1970's (eg, Flavell, Botkin, Fry, Wright, & Jarvis, 1968; Miller, Kessel, & Flavell, 1970). However, the majority of studies performed during the sixties and seventies involved preschool and early grade school children. Although an attempt was made by cognitive developmentalists such as Michael Chandler, John Flavell, Robert Selman, Lawrence Kohlberg and Jane Loevinger to investigate perspective-taking and ego-development in older children, the majority of the studies involved older adolescents and did not specifically focus on the socialcultural factors (see Kroger, 1996; Light, 1993; Schaffer, 1996). The most detailed account of the differentiation of self from other has been provided by Selman (1980). Selman created a model of interpersonal understanding that explains the way in which children become capable of coordinating their own point of view with those of other people and thus develop appropriate role playing skills. According to Selman's five stage theory, children

w-4&&y PiOgc~~from egxeficiC Sm&ge, ti) io aPpmiair: &*i oihers &, have points of view and that these may be different from their own. Selman's model contends that preadolescents (i-e., 10- to 13-year-olds) are capable of understanding multiple perspectives simultaneously. Known as the third person or mutual perspectives stage, the preadolescent is able to abstract the self from an interactive situation and view the perspectives of each person involved in the interaction. Their own viewpoint can be reflected upon from that of another person. Furthermore, Selman's (1980) model has also been criticized for an overemphasis on the structure of various stages too closely tied to Piagetian stages of cognitive development (see SchaRer, 1996). Theory of mind research also re-addresses children's and adolescents' problems with egocentrism (Chandler, 1987; Ellcind, 1967; Piaget & Inhelder, 1956). Adolescents' failure to differentiate between their own thoughts and the thoughts of others is referred to as the imaginary audience syndrome, whereas overdifferentiation between their own and the thoughts of others is known as the personal fable. From a TOMperspective, adolescents who suffer from egocentrism are not able to understand what someone else would think or feel. They are unable to imagine being in the same situation. That is, since they cannot take the roles or perspectives of another person, they are unable to put themselves in 'another person's shoes'. Research has shown that interpersonal understanding is a better predictor of adolescent egocentrism, whereas formal operations did not predict any forms of adolescent egocentricism (Jahnke & Blanchard-Fields, 1993). These results suggest that the links between self and other reasoning and processing capacities are complex and suggest the need for further investigation. Drawing on previous conceptual role-taking research, a TOMapproach to social understanding attempts to integrate the multiple cognitive abilities that one utilizes when one tries to make sense of human behaviour. A TOM approach also enables one to examine the complexity of social understanding by investigating the different aspects of perspective taking such as understanding others' thoughts and emotions. In contrast to past perspective-taking research that assumed thought causes emotion and behaviour, TOMresearch can help to illustrate the

A,- A,- .----A!-- -1 ,-v-~ :--,1- -L-.- -.. - - - .. . m Ir. ucusaawuiu laauulw uciwccu r;ug~&uiii3iid eiii0Guii. 'ii zdiii'uvii, ii I OM approach ewabies researchers to explore and delineate the links between cognitive abilities and social action and thus provide a bridge between cognitive developmental research and social-psychological research.

2.6.2. Empathetic Sensitivity Given that young children's emotional understanding is considered to provide the foundation for the later development of empathy (ability to recognize emotions in others) and prosocial behaviour (e-g., Eisenberg & Mussen, 1989), a TOMapproach to preadolescent social understanding assumes that in order to achieve effective social relations, preadolescents must interpret and understand both the thoughts and feelings of self and other. According to Hoffman (1986) and Davis (1980), empathy contains both affective and cognitive components and is related to one's ability to interact with others in role relationships. Although a growing number of TOM theorists are showing an interest in the role that understanding emotions or empathetic sensitivity plays in the understanding of minds (e-g., Astington, 1993; Baron-Cohen, 1995; Chandler, 1987; Dunn, 1995; Hobson, 1991), TOMresearch does not reflect this interest in emotionaVcognitive unity as seen in the lack of investigations of empathetic sensitivity in preadolescent social understanding. Research has shown that the majority of preadolescents are able to understand and mention multiple internal states and relate them to each other in a coherent fashion (Harris, 1989; Bruchkowsky, 1992). Studies have also shown that once children reach preadolescence, they develop the ability to understand that a person can have conflicting emotions andlor hide emotions from others (e.g., Harris, 1989; Harter & Buddin, 1987). Related research on the more complex, or self-conscious emotions such as shame and guilt show that children are not capable of understanding these complex emotions, or the fact that one can have conflicting emotions until preadolescence (e.g., Griffin, 1995; Tangney 1991). Furthermore, the importance of empathy in one's social understanding has been supported by research from two independent, although related areas, including TOM research on preschoolers

...h:rh hn,. &.,...- +Ln+:, ..Aa:L,, +, ,,,:4&.- ..-a,,*,,rl:-, ,--.-Lr--l .--a,,c,,J:,, -1 ",., "- W-WU uua JUW wu UIUL AU uu~lluwuLW CIVSUUVCI LUIUCILSLCLU-5,GUUUUIIQI UI~UG~~WUAU~pay 3 4~ 24 independently significant role in social interactions (Astington & Jenkins, 1995; Bartsch & Wellman, 1995; Dunn, 1995; Lalonde & Chandler, 1995). Similarly, socialcognitive studies of empathy in preadolescence have generally found that empathetic responding is positively related to popularity or peer acceptance (e.g., Erwin, 1993) and to peer competence (e-g., Ford, 1982; Zahn- Waxler, Cole, & Barrett, 1991). Thus, to achieve a fuller understanding of a preadolescent's theory of mind, the two research areas need to connect.

2.6.3. Person Perception Although research on person perception stems from diverse theoretical and methodological perspectives (Shantz, 1983), the basic assumption is that social interactions are strongly influenced by one's conceptualization of others. That is, how one perceives others is affected by one's social experience (e.g., Mead, 1934). In general, research has shown that children's understanding of the behaviour and personality of others progresses along a developmental continuum (Schaffer, 1996). This continuum reflects shift from viewing others in tenns of concrete, observable characteristics (e.g., "She is pretty"), to an increased understanding of others in terms of abstract, psychological characteristics (e.g., "He is more shy than his brother"). During preadolescence, research has shown that there is an increase in the use of psychological comparisons and categories reflecting consistent traits, interests and abilities, and beliefs (Barenboim, 1981; Flapan, 1968;

Livesley & Bromley, 1973; Peevers & Secord, 1973). Although studies of person perception and related studies of gender stereotyping have occurred independently of TOMresearch, the ability to attribute or ascribe gender-role stereotypes relates to the general rubric of trait attribution and thus suggests indirect implications for TOM research (Edwards & Potter, 1995; Flavell& Miller, 1997; Yuill, 1993). This ascription of gender-role stereotypes can be viewed as a heuristic device that enables girls and boys to understand their own and others' intentions and beliefs (Erwin, 1993). Consequently, the representations of these social roles may help shape evaluative perceptions about the self and other,

LLu& ?&jII cc-&G Ld ii& ;c gi&Sacid h-Lici"&Oris .IyyJ,.An.-\ For henotion that the selfconcept plays a role in the understanding of social-role ascription is supported by a study with preschoolers that found both emotional self-evaluation and cognitive development may influence the acquisition of gender-role concepts (Kuhn, Nash, & Brucken, 1978). With regard to preadolescence, Harter, Waters, Whitesell, & Kastelic (1997) recently studied the links between self-concept, levels of "voice" and gender-role orientation. Results from Harter et al.,'s study showed that levels of voice, or the ability to express one's opinions was highly correlated with self-worth. Furthermore, girls who endorsed a traditionally "feminine" gender-role orientation reported lower levels of self-worth. Taken together, results suggest that the ability to represent and understand others as psychological beings is related to self-perceptions and feelings, which in turn may affect social relations. The explanatory or predictive use of trait terms and gender-roles shares some of the concepts that are associated with TOM research by illustrating how people create implicit personality theories in order to predict or explain others' behaviour. Although person perception and trait attribution research continues to grow, the area continues to be neglected by TOM researchers (Heyman, in press; Yuill, 1992a,b). In general, research gleaned from socialcognitive psychology has shown that preadolescent girls and boys interpret each other's behaviours based on gender-role stereotypic attributes such as greater sociality and emotionality to girls (Powlishta, 1995) and greater instrumentality and autonomy to boys (Honess, 1979,1981; Urberg, 1982). Such findings provide support for various feminist epistemological theories (e.g., Chodorow, 1978; Gilligan, 1982) that claim females' conceptions of self and others are more psychologically oriented or rooted in a sense of connection and relatedness to others whereas males define themselves and others in terms of behaviours or accomplishments. Thus, the TOM approach could help to further the discourse on gender-role ascription and development by investigating the specific processes which enable preadolescents to create gender-typed implicit personality theories and why this occurs. 26

2.6.4. Self-Conce~tionand Self-understanding In addition to the aforementioned higher order social-cognitive processes, an inquiry into an advanced theory of mind must consider the role of the self-concept in the development and application of these processes (Wellman, 1990). As asserted by Shantz (1983), the developmental processess involved in understanding oneself and in understanding others appears to be intricately interwoven and support the view that both self and other perception derives from social experience (Berger & Luckmann, 1967; Mead, 1934). Similar to TOMresearch, the study of the self has had to deal with the problem of internal theoretical conflict since the field's conception (Baldwin, 1902; Cooley, 1912; James, 1892; Mead, 1934), perhaps causing its continued popularity within various social science disciplines especially in psychology (e.g., Bruner, 1990; Kegan, 1994; Ruble & Dweck, 1995) and philosophy (e-g., Bakhtin, 1981; Morrison, 1994; Nagel, 1986; Williams, 1973). Conceptually, the "theory" theory of the self shares the greatest affinity to TOMresearch given that it emphasizes the individual's construction of the self and deals with one's feelings and beliefs directed towards oneself (Damon & Hart, 1992; Harter, 1993). In addition, self-concept research has recently moved fiom the monolithic conception of self (e-g., Coopersmith, 1967) to one that is multidimensional (Harter,

1985) and hierarchical in nature (Marsh & Shavelson, 1989; Shavelson, Hubner, & Stanton, 1976). This notion of the dialogical or "omnibus" self, suggests that the self is a dynamic, multi- layered entity that reconceptualizes itself as a combination of cognitive abilities, emotional and social experiences (Bruner & Kalmar, 1997; Cedarblom, 1989; Day & Tappan, 1996; Markus & Wurf, 1987). From a developmental perspective, the self-concept becomes increasingly differentiated with age and gradually shifts from the physical and active self in early childhood to the

psychological and social self in early adolescence (see Damon & Hart, 1988; Harter, 1993). During preadolescence, the concept of the self increases in complexity and abstractness; the self is now conceptualized primarily in mental-rather than physical-terms. Many researchers have noticed the simiiariues in chiiciren's uncierstancihg of seif and oiher, that ihey boih progress aiong a developmental continuum from concrete, physicalistic, and situation-specific views of the self to abstract, psychological, and trait-like self-definition (Schaffer, 1996). Accordingly, although an interactionist or social-constructivist approach to self-concept research is in its infancy (Byrne & Shavelson, 1994), a growing number of social-co,.aitive researchers have begun to investigate how children and preadolescents come to understand the complex experience of their own individuality and how this differs from their understanding of others (e.g., Damon & Hart, 1992; Hart & Fegley, 1995; Harter & Monsour, 1992; Selman, 1980). For instance, studies that have investigated the concepts of self and other have found some developmental differences between the two concepts, suggesting that it may be possible to have a greater understanding of oneself than other, or vice versa (Damon & Hart, 1988; Hatcher et al., 1990). An ongoing debate among self-concept researchers involves the question of how to integrate the cognitive and affective aspects of the structure of self. While the majority of researchers in the past (e-g., Coopersmith, 1967; Rosenberg, 1989) have defined one's cognitive representation of self in mainly affective terms claiming that "self-esteem" assesses how negatively or positively one feels about one's selfhood, more recent research has be,- to focus on the more cognitive aspects of the self by investigating the justifications of these sekvaluations. As Damon and Hart claim, asking "how much understanding a person has does not approach the crucial question of how the self is understood @. 14,1988). Subsequently, researchers have begun to incorporate both qualitative (e.g., interview) and quantitative (e.g., self-report questionnaire) methods to obtain a clearer picture of the conceptual system of self-conception and its cognitive underpinning of self-understanding which encompasses the thoughts, attitudes and beliefs about oneself that distinguish self from others (Damon & Hart, 1988). Moreover, with multidimensional self-concept research in its infancy, many researchers are advocating further research in this area, especially questioning whether or not the structure of the self-concept is equivalent across gender (Byrne & Shavelson, 1987,1996; Harter, 1993; Harter et al., 1997). Within the research domain of TOM,the concept of the self has recently been increasingly 28 both the self and other (e-g., Barresi & Moore, 1995; Gopnik, 1993; Gopnik & Astington, 1988; Moore, 1996). Unfortunately, there exists a large gap between the research fields of TOMand social cognition, with the latter offering a wealth of literature on the relation between self-concept and socialdevelopment in preadolescence whereas the former continues to focus on infants and preschool children. Thus, by approaching self-concept development from a TOMperspective, the complex relations between a theory of mind and theory of self may become a little clearer. A fuller, more complex model of social understanding, then, may shed some light on the various research questions concerning self-concept during preadolescence. For example, a more complete model of social and self understanding could help to explain findings of recent studies that have shown self-consciousness (Rubin & Asendorpf, 1993), social anxiety (Rosenberg, 1989), and loneliness (Parkhunt & Asher, 1992) increase in tandem with cognitive development. In addition, this socialcognitive model could also be used to test the various theories of TOMdevelopment by comparing self and other interpretations. For instance, theory-theorists would expect self- and other-understanding to develop in tandem (e-g., Barresi & Moore, 1996; Gopnik, 1993), whereas the simulation theorist would expect self-understanding to develop before other-understanding (e.g., Harris, 1989). Alternatively, given that a sense of self is derived from social interactions (e.g., Cooley, 1912; Mead, 1934), a social-constructivist TOMtheorist would perhaps predict that other-understanding would develop before self-understanding. The investigation of self-development within the framework of TOMalso provides an opportunity to explore the role language plays in self-conception. Although the notion that language influences the individuation process is not new (e.g., Bakhtin, 1981; Vygtosky, 1965), Bakhtin's claim that the "selfing process" takes place in the context of relationship, such as the ongoing dialogue between self and others, has gained popularity within psycho-linguistic and feminist epistemological circles. In particular, recent literature has focused on the concept of "voice", that is, the changing social nature of the selfing process is expressed in "voice", or the articulation of one's subjective experience (e.g., Harter et al., 1997). Furthermore, many

-...en-I.a- -1.4- +I...+ +I... A--...l--,.-+ -C+L:.. -...:-A ..-A/-.. ..-lCL-..---.. ..XCC---L-+-> -----a:-- +- 4b-bUbAJ -1- LllUL Ulb Ub V CIAUYUlCILLL ULULL3 VVlbG QLlW Vl 3GL.l UG&VLlLGD UlLIGlGLLULILCLl aLLAJ1UUIL; LU gender during early adolescence (e.g., Brown & Gilligan, 1992; Debold, Tolman, & Brown, 1996). The fact that the self is developed within and by complex social interaction and experience of language suggests then, that the conception of self entails one's ability to speak (and listen) to oneself and to others. The majority of current TOMresearch on self and other have found significant correlations between the understanding of mental states in self and other (e.g., Homer & Astington, 1995; Pratt & Bryant, 1990; Wirnmer & Hartl, 1991). Such findings support the theory-theory hypothesis that claims children use the same conceptual system to reason about their own and other people's second-order mental states. Alternatively, there are a few social cognitive (Damon & Hart, 1988;

Hatcher et al., 1990), crosscultural (Ochs & Schieffelin, 1984) and TOMpreschool studies that provide support for the MeadianNygotskian claim that one needs to understand the concept of "other" in order to understand the self. Due to these confiicting findings, a collaborative effort between social-cognitive researchers and TOMtheorists is needed to investigate preadolescent self- and other-understanding from a TOMperspective. That is, using methods gleaned from 1) the more general research area of social cognition and 2) the more focused research area of children's understanding of mental states, a TOM approach to preadolescent social co,Onition may become an exciting avenue for future research.

2.6.5. Language. Communication and Social Relations As mentioned by past theorists and researchers (e-g., Blos, 1967; Coopersmith, 1967; Harter, 1985), preadolescence is considered the stage of re-individuation, when the child reorganizes and differentiates the self-concept. Given that the sense of self is derived from our social experience (e.g., Berger and Luckmanu, 1967), preadolescents' social experience with their peers play a si,onifant role in the development of their self-concept (Rosenberg, 1989; Selman, 1980; Sullivan, 1953). Accordingly, a social-constructivist TOMapproach to social understanding in preadolescence may assist in the investigation of the links between cognitive representations of scX and ohrand iheir iin'Ks to sociai communicauon. Tnat is, by focusing on the intenuonai elements of social understanding, both TOMand socialcognitive theorists have suggested that higher-order mental states are necessary for effective social interaction (e.g., FlavelI & Miller, 1997; HappC, 1994; Perner, 1988; Selman, 1980). According to folk psychology, social interaction is actually an interaction of minds which consists of mental states that are expressed through speech acts. That is, social interaction does not proceed directly via the interaction of these mental states, it proceeds indirectly by way of language (Astington, 1993). This notion is expressed in various speech act theorists' claims that communication relies on both speaker and hearer taking account of each other's knowledge and intentions (Austin, 1962; Grice, 1968; Searle, 1969) and the corresponding definition of communicative competence as intention attribution (Levinson, 1995; Sperber & Wilson, 1986). In relation to preadolescent development of cognition, as previously noted, by the age of 11, children are capable of abstract, recursive thought (Piaget, 1963). Case et al. (1996), have shown that by age 11, children's increased processing abilities (i-e., central conceptual structures, Case et al., 1996) allow them to integrate multiple perspectives and relate them to each other in a coherent fashion. Within the context of narrative thought, studies on narrative complexity and cognitive competence have shown that the structural complexity of children's narratives increases with age and that this increase may be attributed to information processing capacity (McKeough, 1992). Research on children's narrative shows that 10- to 11-year-olds are capable of understanding and creating interpretive narratives (e.g., Fox, 1991; McKeough et al., 1995). For example, in a study to explore the developing awareness of mind in children's narrative writing, Fox (1991) found that the majority of 11-year-olds were capable of spontaneously writing stories that focused on Bruner's landscape of consciousness as opposed to the landscape of action. That is, most of the 11-year-olds' narratives involved a self-reflective protagonist, or several characters capable of thinking and feeling. Such narrative tasks, then, may be a useful vehicle to investigate children's cognitive ability simultaneously to represent and reflect upon the relation between two

8+-eii.d. G-i-efi .hgi s-& a cOg&dve if&psicy. YriiiGeI've s&&*iii6~"n.d szs pruner, 1986; Case et al., 1996), a growing number of researchers have suggested that narrative tasks may provide some insight into the study of the conception and understanding of self and other in children (e.g., Astington, 1990; Charman & Shrnueli-Goetz, in press; Hatcher et al.,

1990; Tager-Flusberg, Sullivan, Barker, Harris, & Boshart, 1997). With regard to mental states, language and social interaction, various researchers have agreed that the understanding of mental states in others is reflected by the use of both metacognitive (e.g., believe, think, know) and metalinguistic verbs (e.g., assert, say, concede) (representing' mental states and speech acts respectively) (e.g., Astington & Pelletier, 1996; Fabricius et al., 1997; Slomkowski & Dunn, 1996). The use of such mental state verbs has implications for higher order interpersonal relations in that it represents one's own stance or position towards another person's mental states (Laing, 1961; Leech, 1980). For example, the statement "she claims that she is hungry" implies that she believes it is true that she is hungry, but the speaker may not share her belief. Thus, the expression of mental states in speech acts facilitates interpersonal communication and provides the basis for the ability to "read each other". Although preadolescents' ability to relate to their peers may partly be influenced by their ability to understand mental states in others as expressed through their speech acts, peer relations are often excluded from studies that investigate psycho-lin,pistic and pragmatic issues. For example, although links between communicative ability and theory of mind have been studied with preadolescents in terms of speech acts such as the understanding of irony and promising

(Astington, 1988; Hap*, 1993; Keenan, 1995; Winner & Leekam, 1991), within this age group, related areas of social cognition including social competence have yet to be approached fiom a TOM perspective. An integration of research on children's understanding of mental states and their acquisition of mental state verbs may assist in the investigation of how preadolescents come to understand intentional relations in the peer milieux including the understanding of feelings, intentions and beliefs of others (e.g., de Villiers, 1995, cited in Astington & Pelletier, 1996; Moore & Furrow,

pm; ~c~~/ze~l.2ge~,fht is, m94). 'p.L, g>/=L= tA.t +A=ccz~zc.~c~ +& cEL..=.= &y; cc +Ae intentional element of social understanding in the course of social interaction is supported by 1) TOMresearch on relations between preschoolers' theory of mind and social behaviours and, 2) intention attribution studies in social cognitive research, the exploration of understanding intentional relations in preadolescence seems warranted (Astington & Jenkins, 1995; Dodge &

Frame, 1982; Dodge, Murphy & Buchsbaum, 1984; Ford, 1982; Gnepp & Klayman, 1992; Howard & Men, 1989; Lalonde & Chandler, 1995; Slomkowski & Dunn, 1996). For the purposes of the present study, the next section will briefly outline the research showing links between social understanding, self-concept and social relations to justify the need for studying such constructs within the preadolescent classroom. Such a research agenda, then, may provide a rich data base for the future integration of social co,onitive, psycholinguistic and TOMresearch.

2.7. Social Understanding. Self-Conce~tand Social Relations in Preadolescence Despite the fact that a large amount of research exists on preadolescent social understanding, self-concept and social relations, few studies actually investigate the links among all three constructs. As previous researchers have noted, this paucity of research could be due to the fact that the links between social thought, self-cognitions and social behaviour become increasingly complex as children reach early adolescence (Dodge & Feldman, 1990). The classroom setting is a powerful social context in which the psychological adjustment of preadolescents can be affected. As various writers have noted, children often evaluate their own performances and sense of selves based on the feedback received from the school environment (e.g., Bruner, 1996; Harter et al., 1997). That is, a preadolescent's sense of self is partly determined by peer and teacher interactions within the classroom (Harter, 1996). Accordingly, peer relations help shape an individual's personality to the extent that children who avoid peer interactions or withdraw from the social milieu may be at risk of developing problems in the social- cognitive and social-behavioural domains (Sullivan, 1953). To study the role of peer relations in self and social cognitions, a large amount of research has been conducted on the link between self-concept and social behaviour. In brief, the majority of studies shows that positive feelings of self-worth are related to positive social experiences and prosocial behaviour (e.g., Boivin & Hymel, 1997; Patterson, Kupersmidt, & Griesler, 1990; Schaffer, 1996). More specifically, such studies have shown a positive relation between self- concept and social competence (e-g., Cross & Markus, 1994; Harter, 1985; Pellegrini, 1995), self- concept and peer acceptance (e.g., Jones & Gerig, 1994; Richards, Crowe, Larson, & Swarr, in press; Rudolph, Hammen, & Burge, 1995) and self-concept and attachment (e-g., Jacobsen,

Edelstein, & Hofmann, 1995). Although peer relations play a large role in the preadolescent's social-cognitive and affective development (e.g., Abrams, 1989; Sullivan, 1953; Wentzel & Asher, 1995), teachers also continue to affect the preadolescent's inner world. To support this view, Pekrun (1990) found that the influence of family, peers and teachers had a cumulative effect on 10- to 13-year- old$ general and academic self-concept, with teacher support correlating the highest. In addition to teachers and peers, socio-cultural factors such as family structure (i.e., birth order, number of siblings) (e.g., Peterson & Leigh, 1990; Sulloway, 1996), cultural background (e.g., Phinney, 1990), and media exposure (e-g., Wilgosh, 1994) have also been suggested to influence social- cognitive thought and behaviour. Research findings on the link between social-cognitive abilities and social relations in preadolescents have been contradictory and inconclusive. Socialcognitive processes such as conceptual role-taking, empathetic sensitivity and person perception have all been found to be related to both teacher and peer ratings of positive social behaviour ratings and peer acceptance (e.g., Ford, 1982; Pellegrini, 1985). In contrast, some studies have failed to find a relation between various socialcognitive abilities and sociometric status. In particular, studies have shown that aspects of social reasoning such as referential communication and means-ends problem- solving were not related to peer popularity among grade six students (Matthews & Keating, 1995; b.-L:, la-+\ a\U"u, 17,A,. ASEoGge ZGG Fc&ii= (l%%jsiiggest, the i& of ~oli~isi~niIriaciings concerning socialcognitive skills and peer ratings highlights the need for further study. Similarly, with regard to mental-state understanding, findings have shown relatively sophisticated cognitive abilities such as the ability to understand recursive mental states in others (i.e., "She thinks that he thinks...") may be linked to feelings of negative self-worth (Farber, 1989; Veith, 1980). For instance, Veith (1980) found that preadolescents' ability to understand recursive mental states in others was related to a lowering of the self-concept. Similary, feelings of low self-worth and depressed mood have also been found to be linked to sophisticated mental-state reasoning capacities, particularly among highly academically competent preadolescent girls (e.g., Gjerde, 1995; Ken, 1994). Taken together with other studies on adokscents that have shown links between cognitive abilities and self-concept (Bhatnager & Rastogi, 1986; Bosacki, 1995), these results support Farber1s(1989) contention that high psychological mindedness may have deleterious consequences such as a negative self-concept. Thus, such findings suggest that social and emotional correlates of interpersonal understanding or "psychological mindedness" need to be further examined in preadolescents (Park & Park, 1997). Given the previously mentioned link between social understanding and communicative competencies, preadolescents' classroom experience may partly be determined by their ability to understand mental states in others. Although the link between understanding mental states in others and school adjustment has not been studied among preadolescents, past studies on preschoolers and young school-aged children have shown positive links between the understanding of false-beliefs and: 1) positive teacher-ratings of socialemotional skills (Lalonde & Chandler, 1995), 2) skilled aspects of pretend play, (Astingtoc & Jenkins, 1995), 3) peer ratings of likability or popularity (Docket& 1997). In addition to false-belief understanding, studies have also shown a positive relation between positive peer relations and emotional understandinglaffective perspective-taking (Donelan-McCall & Dunn, 1996; Werner & Cassidy, 1997). However, within the realm of TOMresearch, the possible links between preadolescents'

.--1^-^&^^-1:^^ -C---&-l ^..L...." --a ---I&- -C&1--:- "-1c ----I-* ...-a --;-I -l-d~~...---...:- UIIUGI~UUI~~~~UA IIIGUUU ~UILGS auu QS~WVL UGU ~EU~UUCIG~QL~U DCICIACU lG1auuua LC~UQU~ undiscovered. The few studies that have investigated links between social cognition, self-concept and social relations in preadolescents have not worked within a folk psychological framework. That is, such studies have not incorporated the principles of folk psychology or beliefdesire reasoning into their research design or into the interpretation of their findings (e.g., Bosacki, 1995; Rubin, LeMare & Lollis, 1990; Schultz & Selman, 1989). Thus, TOMresearch on preadolescents needs to draw on recent studies that have examined mental state understanding and social relations with younger children. For example, TOMresearch with young school-aged children has begun to investigate how children's personal and emotional experiences within the classroom are linked to their mental-state understanding. Most recently, researchers have be,% to examine the personal school experience of young school-aged children and how this relates to their understanding of thoughts and emotions in others (Astington & Pelletier, 1997; Donelan-McCall & Dm, 1996). In particular, Donelan-McCall and Dunn's study of 6-year-olds' school experience found positive links between emotional understanding and positive peer relations. Thus, such findings promote the need for further investigations of social understanding and social relations in preadolescents. Furthermore, studies that have investigated the complex web of preadolescent social cognition and social behaviour rarely examine how the self-concept intervenes in the relation between social reasoning and social behaviour. The majority of research that investigates the various connections among three or more social-cognitive variables often assume that the construct of the self-concept acts as mediating variable. As Baron and Kenny (1986) note, research needs to investigate the possibility that the self-concept may act as either a mediator or a moderator. To illustrate, Boivin and Hymel(1997) recently studied peer experiences and self-perceptions in preadolescents. Although Boivin and Hymel found that negative peer status and peer victimization mediated the relation between social behaviour and self-perceptions, they did not test to see if these mediating variables also acted as a moderator. That is, by using only a mediational model, Boivin and Hymel assumed that social behaviour causes social preference or popularity. Alternatively, rrnrr..lnG+.. -n.rlA hn-.a nntaa ne n mrrrte-tar &+ha- en--..ren+Ln nr Ah;t.X.rn +ha :..rCi..err-e I~C pYbUCUALJ WVIIIU UUVUUULVU CW U ILIVUUAUCVA, UAWUA UVUpUWLUI6 VL AAAUAUAlALAb ULU UUAUUUW VA social behaviour on self-perceptions. As Baron and Kenny (1986) suggest, as opposed to using the terms mediator and moderator interchangeably, when studying the complex interrelations among three or more social-psychological variables, both mediational and moderation models need to be tested. Within the context of the present study, the influences of self-concept may have either a mediating or moderating effect on the relation between social understanding and social relations. An exception to Baron and Kenny's (1986) claim, a recent study on preadolescents tested for the mediating and moderating effects of self-esteem, attributional style and ethnicity between the relation of shyness and academic performance (Haines & Bartels, 1997). Results showed that self-esteem and a negative attributional style mediated the relation between shyness and academic performance. Given the lack of empirical evidence for established links between self-concept and mental state reasoning, and findings from past studies that show self-concept may have a negative influence on social behaviour, irrespective of social-cognitive ability (Schultz & Selman, 1989), it could be suggested that self-concept may have an inhibitory influence on the relation between interpersonal understanding and social relations. That is, irrespective of one's ability to understand mental states in others, if one holds a negative view of the self, one would be perceived by peers and teachers as less socially competent.

2.8. Gender Effects As previously noted, most studies on social cognition and on TOMresearch have tended to gloss over gender issues, or at least, have not specifically aimed to investigate gender-related differences andfor patterns. In particular, a sociocultural or "ethnopsychological" approach to TOM(Astington, 1996; Astington & Olson, 1995; Lillard, in press; Yuill, 1993) provides a valuable form of inquiry to investigate the effects of gender on social-cognitive development. This approach to TOMasserts the notion that individuals' ability to understand mental states in others is lqciy i,&~vis& & sK~&y+-,~~-cc~c~. s-qp313& &isvie-w, cf"ss+&iw& ,..s.h &di has shown that the development and content of a TOMmay be dependent upon one's culture or social experience (McCormick, 1994). A social-constructivist approach to TOMsuggests that individual differences in the ability to understand mental states in others may indicate that this ability is acquired in different ways for women and men. Put differently, the process of learning to understand self and other within a social context may be contingent not on whether a child is female or male, but on the way in which a child's gender interacts with herhis environment. Thus, if a child is seen as a "cultural invention" (Kessen, cited in Cahan, 1997, p. 2), gender helps to create a separate culture for that child (Edwards, 1993; Unger & Crawford, 1992). This conception of gender as a social category or particular culture (Maccoby, 1988, 1990, 1998; Weber & Mitchell, 1995), suggests that achowledging the contribution of cultural milieu to a TOMmay prove to be a fruitful avenue for future research on development of gender-role conception and behaviour (Zahn-Waxler, 1996). A socialconstructivist inqujr into the workings of the preadolescent mind and its implications for social interactions would enable researchers to view gender incontext (Deaux & Major, 1987; Riger, 1992; Shields, 1991), as it operates within a social milieu such as the classroom. In opposition to the more traditional sex-differences model (i-e., gender differences not mentioned apriori, only post-hoc statistical tests to indicate differences), such a gendered approach to research is in line with Hill and Lynch's (1983) gender intensification hypothesis. This hypothesis claims that during preadolescence, gender differences increase among girls and boys due to the increased pressure to conform to traditional gender-role stereotypes. For example, research has shown that during preadolescence, traditional gender-role behaviour and ascription (i.e., femininity = sociality, submissiveness; masculinity = autonomy, aggression) become intensified (Hill & Lynch, 1983; Tavris, 1992; Wilgosh, 1994). In accordance with research that shows a link between traditional female stereotypic behaviour and depression among women (Gove & Tudor, 1973; McGrath, Keita, Strickland, & Russo, 1990), intensification of gender identity among preadolescent girls may strengthen behavioural tendencies hypothesized to hold

ie~cvGcsfGi i-&3ci&zry- +piessififi s-ucii &erpso&dscns~~,,~L,,, edgemess io please and an increased concern for others (e.g., Zahn-Waxler et al., 1991). Thus, preadolescent girls who appear to be relatively competent in the ability to understand and be sensitive to the needs of others may be at risk for developing future self-concept disorders such as depression (Gjerde,

1995; Peterson, Sarigiani, & Kennedy, 1991; Silverstein & Perlick, 1995). In support of the gender-intensification hypothesis, many gender-related differences gleaned from research on social behaviour and social cognition in preadolescence are usually in he with traditional gender-role stereotypes. Concerning social behaviour, the majority of research findings show that compared to boys, girls are rated as more socially competent and popular by both their peers and teachers (e.g., Dodge & Feldman, 1990; see Unger & Crawford, 1992). Similarly, research on teacher perceptions shows that girls are perceived by their teachers to be more compliant and prosocial than boys (e.g., Harter, 1985; 1996). Furthermore, studies on rejected and clinically depressed children have shown that compared to boys, girls tend to display more internalizing behaviours such as unhappy, withdrawn behavioural patterns. In contrast, boys have been found to display more externalizing behaviours such as aggressive, antisocial patterns

(e.g., Hatzichristou & Hopf, 1996; La Greca, 1981). Regarding social-co,@tive abilities, research shows that boys score higher than girls on nonsocial spatial perspective-taking tasks (Coie & Dorval, 1973), whereas girls score higher than boys on social perspective-taking and empathy tasks (Feshbach, 1978; Jahnke & Blanchard- Fields, 1993; King, Akiyama & Elling, 1996; Zahn-Waxler, Radke-Yarrow, & Brady-Smith, 1977). Likewise, Offer, Ostrov, Howard and Atkinson (1988) reported that across 10 countries, girls were more likely than boys to express a desire to help a friend when possible and thus gave evidence of a "more sociable and empathic stance" than did boys @. 70). Gender-related differences have also been found in the socialcognitive area of person perception. In general, girls have been shown to emphasize such categories as interpersonal skills, psychological traits and social relationships (e.g., Honess, 1981, Peevers & Secord, 1973;

Sprague, Beauregard, & VoeUcer, 1987). In contrast, boys have been found to emphasize such c&tsgGr;cs &ggcssiGfi, $Lcics~6, zd &c&~c&~~j- & s-"r& iGysj. 39

Surprisingly, although TOMresearchers investigate social-cognitive abilities like the ones mentioned above, the exploration of gender-related differences has just begun. Of the few TOM studies that actually do test for gender effects, most report nonsignificant results (e.g., Jenkins & Astington, 1996). In contrast, two recent TOMstudies have found evidence to support gender- related findings gleaned from the more general social-cognitive research area. More specifically, results obtained from both young preschool children and adults show that girls and women score significantly higher on TOM-typetasks as compared to boys and men respectively. Regarding preschool children, Hughes and Dunn (in press), investigated theory of mind and emotion understanding in preschoolers. Hughes and Dunn found that compared to boys, girls referred to mental state verbs more frequently and their choice of verbs was more sophisticated or developmentally advanced. Regarding adults, Baron-Cohen, Jolliffe, Mortimore and Robertson (1997) found sigmficant gender effects on scores derived fiom a newly developed adult TOM-type task (the Eyes Task) designed to measure Baron-Cohen's (1995) notion of "the language of the eyes". In brief, 50 adults were asked to infer mental states from various photographs of humans by choosing a particular mental state verb. Although Baron-Cohen et al. did not control for general language ability, results showed that females scored significantly higher on the Eyes Task than males. That is, females were able to infer a greater number of mental states from photographs of eyes. Taken together, these two gender-related fhdings suggest the need for further investigation into gender effects within the context of TOMresearch. Although the research base in this area is thin, there is some correlational evidence that shows gender differences in the development of psychological understanding of self and others during adolescence. In particular, Hatcher et al., (1990) found that among 13-year-old girls, abstract reasoning was related to the understanding of others but not oneself, whereas the reverse was found for boys, abstract reasoning was positively related to self-understanding but not to the understanding of others. The authors claim that this finding suggests that girls are more likely to

..-An-+q-A nthn- +ha.. +n v.-rla.-~+--A +hnmcn1.mr Lnrnd tLn +nrrde-.--. -C -~d--.-. XXT--L-- -.-I&--- YLIVVALIUYIY VUIV~U LULY~LV LuAUYAOCLLUU LIIYLIWYA V W WIWCN VU UClLIIUUClUk J WI LLLUUGLLL 1V GDLGLLl bUlLUlG to label interpersonal understanding and empathetic sensitivity as "natural" personality traits for females. Likewise, implicitly supportive evidence derived fiom investigations which suggest that for girls, a heightened awareness and understanding of the mental states and feelings of others may have negative consequences for later psychoemotional functioning. For example, longitudinal studies investigating gender differences in developmental models of depression have found that as young girls, depressed adolescent females were more concerned with maintaining interpersonal relationships, more able to recognize the feelings of others and more likely to include moral issues in their play patterns than boys (Block, Gjerde, & Block, 1991; Gjerde & Block, 1991). Further supportive evidence derives from the literature on the socialization of empathy and guilt which suggests that high levels of empathy and guilt may serve as precursors for later depression in women (Bybee, 1998; Zahn-Waxler et al., 1991; Zahn-Waxler & Robinson, 1995). Similarly, in a study of 115 preadolescent girls and boys (8- to 12-years-old), Fraser and Strayer (1997) found that the most robust correlation between sensitivity and shame existed among the older girls (1 1- to 12-year olds). Hence, such studies support the view that interpersonally sensitive individuals may be preoccupied with the psychological needs of others to the detriment to their own, which may eventually lead to a lack of self-definition ( Park & Park, 1997; Silverstein & Perlick, 1995). Accordingly, connections between works on feminist epistemologies and TOM research would provide a valuable starting point for investigating the processes of understanding self and other in preadolescence (e.g., Belenky et al., 1986; Brown & Gilligan, 1992; Debold, Tolman, & Brown, 1996; Smith, 1990). The attempt to define and assess the processes that enable preadolescents to understand social situations may offer some insight on gender-related research findings. Such a research agenda would help to shed some light on the wealth of findings from social-emotional studies that show around the age of 1 1, some girls (in comparison to boys) experience the following: a si&~cant drop in feelings and thoughts of self-worth (e.g., Edwards, 1993; Harter, 1985; Silverstein & Perlick, 1995), an increase in self-consciousness (e-g.,

n-r)\ Shimom, iicisenberg, & Zvse~i~l~g,17. IJ,, aid iin increiingiy ncgaiivt: sense or" seir' WOM despite high academic achievement, particularly among "gifted girls" (e-g., Callahan,

Cunningham, & Plucker, 1994; Loeb & Jay, 1987; Matthews & Smyth, 1997; Reis, 1987). Thus, research on the ability to understand mental states in others and its links to sekoncept and social relations may help to illustrate the related phenomenon of why girls may be more at risk than boys to lose their sense of self or "inner voice" during preadolescence (Brown & Gilligan, 1992; Harter, 1996).

2.9. Summarv Drawing on the broad array of social-cognitive literature on preadolescents outlined in the above sections, empirical findings are mixed concerning the link between social understanding, self-concept and social relations. Although the investigation of sekoncept and social understanding in preadolescents began forty years ago (Bruce, 1958; Flapan, 1968), researchers and educators remain puzzled and intrigued as to how preadolescents learn to make meaning from their social experiences. As already noted, although there exist a large number of studies on social cognition in preadolescents, they either examine specific components of social cognition and social relations, or the interrelations between two constructs. Combined with Chandler's (1987) assertion that the development of relativistic thought is linked to the emergence of generic self-doubt (implying a decrease in self-worth), one could hypothesize from the aforementioned results that the development of higher order cognitive abilities such as recursive thought may have the potential to have a negative impact on a preadolescent's sense of self (Dabrowski, 1964). Furthermore, given that girls develop their sense of self through their co~ectionswith others and are more likely to internalize their social problems (e.g., Chernin, 1985; Chodorow, 1978; Goetz & Dweck, 1980), this realization that all knowledge is relative to perspective (Bruner, 1990) may have a more devastating effect on their self-concept as compared to boys (Silverstein & Perlick, 1995). For example, a more sophisticated understanding of others' mental states may also be related to a greater psychoemotional sensitivity that may lead one to feel psychoiogicaiiy isoiateci and vuinerabie ~@abrowsici,i9Sj. Such feeiings of isoiarion and vulnerablity, then, may lead to a decrease in one's sense of self-worth (Farber & Golden, 1997; Park & Park, 1997). Thus, more research is needed on preadolescent thought and social- emotional development in order to investigate the possibility of maladaptive socialemotional responses as a consequence of developing an advanced understanding of others' minds. In summary, this literature review presents a vast array of research on social cognition and social relations in preadolescence. However, the aforementioned studies do not specifically speak to the possibility that 1) a TOMtheoretical framework may be used to explain individual differences found between language, social-cognitive abilities and social relations, 2) the self-concept may play an intervening role between the social understanding and social relations, or 3) gender-related linkages and differences may occur between social understanding, self-concept and social relations. The current study will address these three main concerns. Gleaned fiom the literature, one could assume that a preadolescent TOMinvolves understanding higher order mental states, empathetic sensitivity, person perception and may be related to the ability to understand mental or metaco,onitivelmetalingui~ticverbs, feelings of self- worth and social-relational ability. Thus, given these criteria and based on past literature and pilot work, the present study aims to investigate the links between social understanding, self-concept and social relations in preadolescent girls and boys. Research Questions and Hypotheses 3.1. Introduction Grounded in the theoretical and empirical background discussed in Chapter 2, this chapter will outline the research questions of the present study. Following the research questions, the specific hypotheses of the present study will be presented.

3.2. Research Ouestions Based on the lack of a folk-psychological unifying theory and empirical evidence for an existing TOMapproach to social understanding in preadolescence, the main purpose of this study was twofold. First, the study intended to conceptualize and operationalize social understanding from a TOMperspective by focusing on the understanding of mental states and emotions. Second, this variable of mental state understanding was to be used to illustrate the linkages between individual differences among social understanding, self-concept and social relations. More specifically, the three variables were also to be used to test Baron and &MY'S (1986) statistical models of mediation and moderation. Finally, to show evidence of differential gender patterns, all of the main variables were to be examined for correlational and individual differences. To investigate how preadolescents make meaning from their social world, this study used a holistic, social-constructivist research approach that focuses on the connections and interactions between self and other, and emotion and cognition (Baldwin, 1913; Gergen, 1985; Mead, 1934; Vygotsky, 1978). That is, to examine whether preadolescents' understanding of mental states and emotions in others is associated with their sense of self and their ability to relate socially to others, this study borrowed from socialconstruction theory that contends a sense of self is derived from social experience (Berger & Luckmann, 1967). Such theories of the self contend that the self is not only constructed, but that its mode of construction is interpretive and dynamic. That is, events that people experience are automatically interpreted or filtered through their self- concept. Thus, social experience and self-concept are interdependent, one cannot exist without the 44 other (Bruner & Kalmar, 1997). In addition, the present study draws on reciprocal-constru~tivistmodels of socialcognitive development including Hinde's (1979) social complexity model, Dunn's (1988) relationship model of social understanding, Oppenheimer's (1989) activity levels model (dynamic systems) of social competence and Bruner's (1986, 1990, 1996) model of narrative thinking. Such models focus on the dynamic connections between the self and other, and reinforce the constructivist notion that self-perceptions and experience are intimately related In the context of the present study, these constructivist models would predict that psychological changes and differences are the result of complex interactions among interpersonal, cognitive and social processes. Based on the assumptions of the aforementioned models and the empirical evidence outlined in Chapter 2, this study predicted that the three constructs of social understanding, self-concept and social relations would be interrelated. In addition, based on the aforementioned models and empirical evidence, self-concept was predicted to act as an intervening variable between social understanding and social relations. That is, self-concept could act as either a moderator or mediator (see Fiewe 1). If self-concept acts as a mediator, the relation between social understanding and self-concept is significant. The ability to understand mental states and emotions in others acts through self-concept to affect one's social- relational ability. For example, if self-concept was to act as a mediator, high levels of social understanding would also be related to high levels of self-concept and high levels of social- relational ability. Thus, according to the mediation model, a child who displays high levels of social understanding will also have high levels of both self-concept and social-relational ability. Alternatively, Figure 1 also shows the relations among social understanding, self-concept and social relations if self-concept acts as a moderator. According to the moderation model, the influence of social understanding is not directly related to self-concept, but rather, self-concept partly influences or determines how social understanding would affect social relations. For example, in the context of moderation, high levels of social understanding would not necessarily

L r&;& ifi EgG levels se~Coiicqji & s~..-f&d~on... aDlllLy. .'a -... yne &ierminrnt one's Socid Social Relations rnUnderstanding

B . Moderation Model r-lSocial

Social Relations Self-Concept

Figure 1. Theoretical Models of Relations Among Social Understanding, Self-concept and Social Relations social-relational ability would be one's level of self-concept, irrespective of one's ability to understand social situations. To illustrate, according to the moderation model, a child could display high levels of social understanding but at the same time experience low levels of self- concept which in turn would lead to low levels of social-relational ability. Thus, self-concept as a moderator can either hinder or facilitate the influence of social understanding on social-relational ability. Finally, gender-related differences were to be examined based on the gender-intensification hypothesis that claim such differences are due to the increase in gender-role socialization during preadolescence (Galambos, Almeida, & Petersen, 1990; Hill & Lynch, 1983; Maccoby, 1990). Such theories contend that preadolescents' increased exposure to culturally shared notions of gender-appropriate behaviour gradually affects their self-perceptions, which in turn, leads to stereotypic gender-role behaviours. Based on a reciprocal-constructivist mode of inquiry, this study aimed to explore the links between the three main variables including the following: 1) Social understanding or the interpersonal aspect of social reasoning. That is, the ability to attribute mental states to others including the ability to understand multiple perspectives (conceptual role-taking), the ability to recognize and understand emotional states (empathetic sensitivity), and the ability to understand the concept of a person as a psychological being with stable personality characteristics (person perception) (Chandler, 1987; Schwanenflugel et al., 1994). This variable was assessed by means of a story-telling interview that will be described in detail in the following chapter. 2) Both the affective and cognitive aspects of self-concept, that is the ability to perceive, feel and judge differently about multiple aspects of the self including the areas of social acceptance, behavioural conduct, athletic competence, academic competence, physical appearance and global self-worth (Damon & Hart, 1988; Harter, 1985). To assess the affective aspect of self-concept, Harter's Self-Perception Profile for Children (SPPC, 1985) was used. To assess the cognitive

----+ -C --1C ------& A..C--A :- +L- -A-A------I& --->--&--A:-- -L:lA------I-- 2 A- mp+&Wl JGll-bWUbGpL, UGllUGU Ul LUG plGJGUL JLUUY PJ 3Gll'UllUGlJLcllIUlll& bllLlUlGll WGlG aSJSGU W justify some of their responses to the SPPC. 3) Social relations or the ability to interact socially with peers was measured by peer and teacher ratings of how well the child gets along and is popular with others. To assess teacher ratings, Harter's Teacher's Rating of Actual Behaviour rating scale (TRF, 1985) was administered to the classroom teachers. To assess peer-rated social competence or "the attainment of relevant social goals in specified social contexts, using appropriate means and resulting in positive developmental outcomes" (Ford, 1982, p. 323-324), children were administered an adapted version of Ford's Social Competence Nomination Form (SCNF, 1982) (Cauce, 1987; Matthews & Keating, 1995). This measure focuses on peer ratings of the more co,@tive aspects of social competence, that is, children are asked for judgements of their peers' competence in handling hypothetical social situations. To assess peer popularity or peer likability, that is, the extent to which peers are liked by others within classroom (e.g., Cauce, 1987), the peer likability sociometric scale was used (see Matthews & Keating, 1995). In addition, for the purposes of the present study, based on the suggestions of various authors (e-g., Ford, 1982; Harter, 1996, Schaffer, 1996), a composite score of social-relational ability was computed by summing the scores of peer-rated social- competence, popularity, and teacher ratings of social acceptance. In addition to these three main variables, to check for the influence of language on social- cognitive ability (Astington & Olson, 1990; Goody, 1995; Schwanenflugel, Fabricius, Noyes, Bigler, & Alexander, 1994), two aspects of language ability were explored. Based on past research that aims to explore socialcognitive processes without the confounding effect of general language ability (e-g., Charman & Shmueli-Goetz, in press; Hap@, 1994; Hughes & Dunn, in press; Siddiqui, 1995), to control for the confounding effect of general vocabulary ability, general receptive language ability (language comprehension) was assessed through the use of a standardized paper and pencil vocabulary task (general words). To explore the possibility that the understanding of mental states in others may be related to the understanding of metaco,onitive language, the understanding of mental state verbs was assessed by a paper and pencil mental state

-.--L- 6111- VGLU3 LQ3h. Based on the aforementioned theoretical models and empirical evidence, the following research questions were investigated separately for girls and boys to determine if distinct patterns exist for each gender: 1) Do relations exist among social understanding, self-concept and social relations? 2) What intervening role does self-concept play in the relation between social understanding and social relations (mediator or moderator) and how does this role differ among the various self-concept dimensions? 3) How do individual scores and correlational patterns differ?

3.3. Hvuotheses Based on the theoretical assumption that folk or beliefdesire psychology provides the infrastructure for both self-conception and social interaction (Pemer, 1988; Wellman, 1990), it was predicted that individual differences in preadolescents' ability to understand the thoughts and emotions of others would be related to their self-concept and social relations with their peers. Moreover, based on 1) past studies that have shown an association between social understanding and social relations (e.g., Astington & Jenkins, 1995; Dockett, 1997; Slomokowski & Dunn, 1996), and 2) both theoretical and empirical work that have investigated the role self-perceptions play between interpersonal thought and social action (e.g., Bruner & Kalmar, 1997; Keating, 1990; Rubin, LeMare & Lollis, 1990; Schultz & Selrnan, 1989), it was predicted that self-concept would somehow intervene (either mediate or moderate) in the link between TOMand social relations (see Figure 1). Due to the conceptual complexity and exploratory nature of the present study, one model (e.g., mediation) was not hypothesized to prevail over another. Following the suggestions of Baron and Kenny (1986), it was predicted that either 1) social understanding and self-concept would be directly related and thus, social understanding would act through self-concept to influence social relations (mediation model), or 2) social understanding and self-concept would not

L a:-.41-. -l-*-A *I---.. ------*> -Cr----L- 2: :---->I-- --.- G uu~~~ysiarcu ad-- uua, ~GU-LUULG~L------.wuuu arc-~~UIG U~UVII ---. UWU~ bill~~@~f ~C i~ii~~fi between social understanding and social relations (moderation model). In addition, based on pilot work and the aforementioned research findings that support the enculturationJsocial-constructivistapproach to social understanding (e.g., Bosacki, 1995, 1997; Hatcher, et al., 1990; Matthews & Keating, 1995; Zahn-Waxler et al., 1991), and the gender intensification hypothesis that claims gender differences increase during preadolescence due to culturally shared notions of gender-appropriate behaviour (Hill & Lynch, 1983), it was predicted that gendered linkages would occur among the three variables. That is, the patterns of the links among the three variables were expected to differ for girls and boys. For example, girls are socialized to understand and care for others more than themselves and are encouraged to base their sense of self-worth on their relationships (Belenky et al., 1986). Unlike boys, they are not encouraged by societal messages to accept themselves or to become autonomous and self-reliant (e-g., Edwards, 1993). Consequently, during preadolesence, when gender-typed behaviours are emphasized, it could be expected that girls will exhibit greater social understanding than self- competence as compared to boys and greater social insight may actually have a negative impact on their sense of self and social interactions by increasing their feelings of insecurity and self- consciousness. Accordingly, it was hypothesized that as girls became more sophisticated in their social understanding, their self-concept and social relations would decrease. This effect was not predicted for boys. Thus, by combining a psycho-cultural approach to TOM (e-g., Bruner, 1996; Lillard, in press) and the gender intensification approach to gender differences (Hill & Lynch, 1983), this study offers a more detailed look at social understanding and social relations in preadolescence. The richness of this approach, then, is that it builds on past socialcognitive research by taking gender into account and by explicitly testing the intervening role of self-concept between the mental and social life of the preadolescent. Finally, given the recent studies on the influence of siblings on general psychological development and social-cognitive abilities such as understanding mental states (e.g., Sulloway, innL. A-L-A-- 0- r--i~-- inna .L ...A-.a---i--: ------.-a -- ..--I--.--.: --.-- ~77~~TWUU~.~CZ. JG-a, 177~j,LUG ------~GBGUL DLUUY QLDV ubviplaLGu GA~W~LVLJ~VGDU~OUVU .:-,..:-- of sibling effects.

3.4. Rationale of the Research Desim Although longitudinal studies permit the most robust causal claims, time and financial constraints led the present study to adapt a more timeefficient correlational research design. The research design of the present study was not cross-sectional, that is, it did not test age effects of the various constructs. Unlike the majority of developmental social-cognitive and TOMresearch that investigates age effects, the present study focused on individual differences within one age level (preadolescents). As previously mentioned, the main purpose of the present study was to describe the connections among the variables of social understanding, self-concept and social relations. Given the claim that the most fruitful approach to studying social understanding is a multimethodological approach (e.g., Astington & Olson, 1995; Feldman, 1992), the present study aimed to combine a variety of methods. That is, the present study combined both experimental and semi-naturalistic methods by including a combination of both quantitative measures (self-report questionnaires, peer and teacher ratings) and interpretive measures (semi-structured interviews including a mixture of forced choice and open-answered questions). Within the context of social- cognitive research, arguments can be found to support both an experimental approach and more interpretive methods. For example, it has been suggested that within the research domain of TOM, the use of a strictly experimental approach precludes the problem of subjectivity and thus may be more accurate in its representation of the participants' perspectives than more interpretive methods such as an interview, (e.g., Perner, 1988). Alternatively, recent approaches to studying mental state understanding in children are moving toward more interpretive, hermeneutic tasks such as open-ended narrative tasks as opposed to structured, forced-choice tasks such as the false-belief task (e.g., Charman & Shmueli-Goetz, in press). Given this lack of consensus as to what constitutes the optimal design to assess social-cognitive constructs in preadolescents, the present study attempted to find a balance between experimental and hermeneutiditerpretive methods with

+I., csv-a-+nd~- An+ r..-L n Aanrinr rr.rr.rlA -rr..

4.2. Instrumentation 4.2.1. Development of New Social Understanding Measures Due to the paucity of research on preadolescent social-cognitive understanding from a TOM perspective, pilot work was conducted in order to test measures of social and self understanding. These tasks were developed in order to investigate the connections among social understanding, self-concept and social relations for a preadolescent sample. Since 1983, social understanding from a TOMperspective has been studied by first- (e.g.,

Wimmer & Perner, 1983) and second-order (e.g., Homer & Astington, 1995; Perner & Wimmer, 1985) false-belief tasks consisting of dramatized stories followed by a series of questions which are used to test the child's ability to ascribe false belief to others. Such stories are sometimes accompanied by an appearance-reality task where the children have to distinguish between what something looks like and what it really is (e.g., a piece of sponge painted to look like a rock) (Flavell, Flavell, & Green, 1983). Although the majority of such TOMresearch has focused on preschoolers, there has recently been an attempt to study an advanced theory of mind or the understanding of third-order mental states through the use of elaborated false belief stories among autistic individuals (e-g., Bowler, 1992; HappC, 1994). Although this method has the advantage of obtaining a quantifiable score to represent a "theory of mind in a relatively short period of time, given the complex and dynamic mental life of the preadolescent, such a measure might not prove to be either internally or externally valid and perhaps not comprehensive enough.

4.2.2. Assessment of an Understanding:of Higher-Order Mental States In general, there exists a lack of agreement among TOMtheorists on such issues as whether or not a "theory of mind" or the interpersonal understanding aspect of social reasoning actually continues to develop. Moreover, if one assumes that this mentalizing ability continues to develop once it is acquired, further debate surrounds the issue of what exactly develops (e.g., see Astington, in press). This theoretical confusion makes it a challen,@ng task to develop an instrument to measure particular aspects of social understanding in preadolescence. Given the assumption that social understanding increases in complexity and becomes multidimensional, consisting of both co,onitive and emotional representations (Wellman, 1990), then both affective and cognitive aspects of self- and other- understanding must be considered in an attempt to assess the interpersonal understanding aspect of social reasoning in preadolescence. More specifically, Chandler's (1987) recommendation that an advanced "TOMability" consists of conceptual role- taking or perspective-taking, empathetic sensitivity, and person perception stresses the importance of gaining insight into the contents and processes of others' minds. In addition, Wellman's (1990) proposal, that developments in children's belief-desire psychology should be reflected in self- and personconception, suggests the need for research that would describe and compare social reasoning in areas of both self- and other- understanding. Thus, taken together, both theories highlight the importance of exploring how people attribute mental states to both self and others, which in turn may provide a complete model of a "theory of mind in preadolescence. As previously mentioned in Chapter 2, the wealth of socialcognitive studies that exist on preadolescent inter- and intrapersonal understanding could assist in the investigation of a TOM TOMhave failed to cross. This lack of communication has been mainly due to the fact that most of traditional social-cognitive research has focused on elementary school-aged children and adults, whereas TOMresearch has focused on preschool children. Among the studies that have examined preadolescent social cognition, the TOM approach is rarely, if at all mentioned. Few researchers have taken advantage of investigating different dimensions of social cognition within the flexible framework of TOM (for an exception see Mansfield & Clinchy, 1997). For example, areas that have previously been studied in the field of social cognition where the focus has been on the understanding of social relationships could be studied from a TOMperspective, such as conceptual role-taking (e-g., Chandler & Helm, 1984; Gnepp & Klayman, 1992), person perceptionltrait attribution (e-g., Yuill, 1992a; Howard & Allen, 1989), and empathetic sensitivity (eg., Hughes, Tingle & Swain, 1981; Tangney, 1991). A TOMperspective would focus on the understanding of the mental states of individuals and the understanding of a network of interpersonal mental connections. Such a perspective would help to unpack the process of understanding recursive thinking which involves the sophisticated capacity of realizing that several people can simultaneously hold belief states about one another (e.g., "She's thinking that he's thinking that I'm thinking..." (Miller, Kessel, & Flavell, 1970). Building on TOM studies with younger children (e.g., Astington & Jenkins, 1995; Dockett,

1997; Dunn, Brown, Slomkowski, Tesla, & Youngblade, 1991) and various claims that higher- order mental states are related to social interaction (e.g., Carpendale, 1998; Perner, 1988), an inquiry into preadolescent social understanding should consist of social stories or vignettes followed by questions which would probe whether or not a child could interpret social inferences or meaning fiom the story. Recently, many socialcognitive researchers have proposed that vignettes consisting of ambiguous social situations may be the most effective way to elicit spontaneous attributional thinking fiom children (e-g., Leekam, 1993; Levinson, 1995; Mansfield

& Clinchy, 1997). Such stories could be followed by questions used to probe the child's representational understanding of mind by including questions on social role-taking or causal

*tt&hm&nnrlo rr U7h-t Anxrnm +h;rrlr th-t --fin ;r nn;nn tn rlrr - -.rh-73\ -on-- -or-a-Lrr- lo rr C.CYIYYUVA.U \Y-e-).. YYL YV J YY a&. UaUC yWLIYaa ILI bVLU6 LV UV VV UJ ;/) pL3VUpAbbYUVU \U.5., Choose a character and describe what kind of person they are - why?) and empathetic sensitivity (e-g., How does this person in the story feel - why?). Past studies have measured various aspects of interpersonal understanding among children through the use of either ambiguous (e.g., Dodge & Frame, 1982; Gnepp & Klayman, 1992) or straightforward social vignettes (Selman, 1980) that include either implicit social intentions or explicit social intentions respectively. The narratives are usually followed by written questions as in a semi-structured interview (e.g., Selman, 1980). However, most studies usually focus on only one aspect of interpersonal understanding and have thus failed to recognize the complex and multidimensional nature of preadolescent social understanding. Furthermore, studies that have investigated various aspects of social understanding have either 1) failed to incorporate the concept of ambiguity or uncertainty into the stories (e-g., Selman, 1989), 2) failed to ask children to justify or to provide reasons for their answers (e.g., Hatcher et al., 1990), or 3) have used ambiguous provocation scenarios as opposed to more ambiguous situations that are of common occurrence in the preadolescent classroom (e-g., Dodge & Frame, 1982; Erdley & Asher, 1996). Thus, due to the sparse research on recursive mental-state understanding under uncertainty in preadolescent girls and boys, the present study used ambiguous social vignettes combined with a short interview on self-evaluation and understanding. It was expected that such a task would elicit spontaneous social judgments made under uncertainty (Kahneman, Slovic, & Tversky, 1982) which in turn would shed some light on the relation between interpretive understanding or "verstehen" of the self and other (Oatley, 1992; Kaufinan, 1994). Accordingly, from a constructivist perspective that views social understanding as a vehicle used to co-construct or narrate one's social reality (Astington, 1990; Bruner, 1990, 1996), and based on past research that has investigated higher-order mental-state understanding through the use of narratives (Bowler, 1992; Charman & Shmueli-Goetz, in press; Fox, 1991; Happ6, 1994), and pilot study findings (Bosacki, 1997), two brief vignettes were constructed, consisting of ambiguous social situations that are relevant to the preadolescent's school experience (Social

TT-J--.--X-- m.--- nr----- m#---:- m--!-i TT.. -%- UUUGI~UULU~CWL~ 1: luau~yl~v~iug~c,DW~ UIIUCIS~~ Siory 2: Kenny~%iaric,see -1 ames .. i 56

and 2). In particular, findings from the pilot study showed a significant positive link between the summed score of the two social stories and a story adapted from HappC's (1994) advanced test of theory of mind ability (Strange Story Task). In line with recent findings from Baron-Cohen et al. (1997), such a link suggests that the social stories used in the present study contain a mental state attribution component. A further distinguishing feature of the ambiguous social story tasks is that no words are spoken among the characters. All communication is by means of eye-contact and non-verbal communication (i.e., smiling, nudging). Thus, all mental states must be inferred from the characters' ambiguous actions as opposed to language. Story 1 describes the scenario of two girls approaching another girl who is new to the school and is playing by herself in the playground. Story 2 describes a school sports scenario where two boys are choosing another boy to join their team. Each vignette is socially ambiguous in that each describes nonverbal behaviours that could have multiple interpretations. Social ambiguous vignettes were created based on the notion that interpretations of ambiguous social situations are an effective method of eliciting spontaneous attributional thinking (Weiner, 1985a). Moreover, tasks that involve ambiguous social situations have also been claimed to trigger or promote higher-level social information processing or social reasoning ability as compared to more explicit, obvious social situations (Waldman, 1996). That is, ambiguous stories may encourage children to draw on their representational and interpretive understanding of mind (e-g., Carpendale & Chandler, 1996; Leekam, 1993; Levinson, 1995). Also, this study used ambiguous narratives because past research has revealed robust individual differences in how children and preadolescents interpret and respond to ambiguous scenarios (e.g., Dodge & Frame, 1982; Erdley & Asher, 1996; Slaby & Guerra, 1988; Waldman, 1996). Furthermore, to counterbalance possible gender effects of the stories (see Goldberger, Tarule, Clinchy & Belenky, 1995), in each story, the characters were of the same gender and both stories were presented to the entire sample. To assess the social-cognitive processes involved in interpreting social inferences or meaning from the story, each vignette was followed by three sets of questions aimed to measure

..L ^-^+:E- CLf-L- --3-- &--I --I- .-I-?..- USap111r; r;u~ccyw------.-61 ~uguci-ulucl c;uuc;cp~uiu rulc-~suullg, eiiipaik3k semiiiviiy iirld ps0a Table 1 Interpersonal Understandine Interview: Social Understanding Stow 1 Nancy and Margie are watching the children in the playground. Without saying a word, Nancy nudges Margie and looks across the playground at the new girl swinging on the swingset. Then Nancy looks back at Margie and smiles. Margie nods, and the two of them start off toward the girl at the swingset. The new girl sees the strange girls walk towards her. She'd seen them nudging and smiling at each other. Although they are in her class, she has never spoken to them before. The new girl wonders what they could want.

1. Does the r?ew $1 see Nancy and Margie cudeaing and smiling at each other? Yes/No 2. Has the new girl ever spoken to Nancy and Margie before? YedNo If response to either comprehension question is incorrect, re-read the story to S. S must respond conectly to both Comprehension questions before continuing with the interview.

General prompt guide (applies to all questions): if response is "I don't know," or the child does not answer, repeat the question. If S responds with a second "I don't know," or still refuses to answer, go on to the next question. A. Third Order Conce~tualRole-Taking 1. Why did Nancy smile at Margie? 2. Why did Margie nod? 3. a) Why did Nancy and Margie move off together in the direction of the new girl? b) Why do you think this1 How do you know this? 4. a) Does the new girl have any idea of why Nancy and Margie are walking towards her? Yes No

b) How do you know that the new girl has or [doesn't have] any idea of why Nancy and Margie are walking towards her? The phrasing of 4b is contingent upon the response to 4a Specifically, if response to 4a is "yes," ask "How do you know the new girl has an idea of why Nancy and Margie are walking towards her?" If response to 4a is "no" ask "How do you know the new girl doesn't have any idea of why Nancy and Margie are walking towards her?" response to 4a is "I don't know" repeat question. If S responds to the second asking of question 4a with an "I don't know," go on to next section. B. Empathetic Sensitivity 5. a) How do you think the new girl feels b) Why? Does she feel anything else? Why? Keep repeating until no answer. Table 1 (Cont.) If response includes an ambiguous or bizarre feeling such as "different" or "wondering", ask for further clarification - "What do you mean by feeling different?" ; "Is wondering a feeling?"; Continue to prompt for valence if response is still ambiguous - "Is it a goodhad or in between feeling?" C. Person Perceution 6. Choose a character in the story and describe her. What kind of things can you think of to describe her? What kind of person do you think she is? If S describes something the character does or some aspect of the story without mentioning any kind of trait, prompt with "What kind of a person do you think she is?" Also, if response includes an ambiguous term with no explanation (e.g., she's nice) prompt with "What do you mean by nice?"; "What makes her nice?" D. Alternative Exulanations 7. Is there another way that you can think about this story? Yes No If so, how? If response is: "No," "I don't know," or if S asks "What do you mean?" continue with "Could there be another reason why Nancy and Margie are nudging and smiling at each other?" If response is still no, ask, "So is this the only way to think about the story?" Note. For the purposes of the present study, this measure was designed by the author (Bosacki). Table 2 Interoersonal Understanding: Interview: Social Understanding: Stow 2 Kenny and Mark are co-captains of the soccer team. They have one person left to choose for the team. Without saying anything, Mark winks at Kenny and looks at Tom who is one of the remaining children left to be chosen for the team. Mark looks back at Kenny and smiles. Kenny nods and chooses Tom to be on their team. Tom sees Mark and Kenny winking and smiling at each other. Tom, who is usually one of the last to be picked for team sports, wonders why Kenny wants him to be on his team.

1. Does Tom see Mark and Kenny winking and smiling at each other? Yes/No 2. Is Tom usually the first person to be picked for team sports? Yes/No [Follow up as for Story 11 A. Third Order Conce~tualRole-Taking 1. Why did Mark smde at Kemy? 2. Why did Kemy nod? 3. a) Why did Kenny choose Tom to be on the team? b) Why do you think this/ How do you know this? 4. a) Do you think that Tom has any idea of why Kenny chose him to be on the team? Yes No b) How do you know that Tom either has or [doesn't have] any idea of why Kenny chose him? Follow up as for Story 11

5. a) How do you think Tom feels? b) Why? Does he feel anything else? Why? @?oUow up as for Story 11 C. Person Perce~tion 6. Choose a character in the story and describe him. What kind of things can you think of to describe him? What kind of person do you think he is? Follow up as for Story 11 Table 2 (Cont.) D. Alternative Ex~lanations 7. Is there another way that you can think about this story? Yes No If so, how? Follow up as for Story 11 && For the purposes of the present study, this measure was designed by the author (Bosacki). perception (see Tables 1 and 2). The tasks were designed to tap the socialcognitive competencies that enable humans to "read people and social situations by means of understanding their mental states (i.e., intentions, desires, beliefs). Thus, this task was developed in the attempt to strike a balance between very projective, open-ended narrative tasks (e.g., Fox, 1991; Selman, 1980) and more forced-choice, experimental tasks (e-g., Baron-Cohen et al., 1997). For example, in the Nancy/Ma.gie story, Table 1 shows that to assess children's understanding of conceptual perspective-taking they are asked such questions as, "Why did Nancy and Margie move off together in the direction of the new *I?, followed, "Why do you think this/How do you know this?" To assess empathetic sensitivity or emotional understanding and person perception, children were asked, "How do you think the new girl feels and why?" and asked to choose a character in the story and describe her. In a final assessment of the ability to understand multiple perspectives of a situation, children are asked, "Is there another way that you can think about this story?" First, to control for effects of memory or comprehension, two forced choice control questions were asked once the story was read (before the set of TOMquestions), e.g., "Does Tom see Mark and Kenny winking and smiling at each other?" The story is re-read to the child if either response to the control questions are incorrect.

4.2.3. Coding- of Social Understanding- Narratives (Social and Self-Understanding- Interview - Part A - Interpersonal Understandind The children's answers, and the researcher's questions and probes, were audiotaped and transcribed verbatim. The children's answers were scored from the transcripts (see Appendix Al). The responses were coded according to 1) level of structural complexity and 2) thematic content including emotional valence. First, a coding scheme was developed to obtain a single, quantifiable score for TOMor understanding of mental states and feelings in others. This score was used to determine whether or not differences existed in cognitive complexity. Regarding this quantitative scoring, the questions for each story were grouped into four

nrrrl:-.- --+nrroGnn +hnt :n annL nn+anrr.-.. -rrrora-+nd thn nrrr.c*....-tr rrF- \r-5 \raL\r5VAAU3, UUL13, UUUU -LU6VA J AUYAWUULVU UAU UVUPUUWW VL. A. Understanding conceptual perspective taking or third order mental state understanding (5 items). B. Empathetic sensitivity (2 items). C. Person perception (1 item). D. Alternative explanation (1 item). In the pilot study, the wording of the fourth question of Section A was "Does the new girl know why Nancy and Margie are walking towards her/Does Tom lcnow why Kenny chose him to be on the team? The question was changed to "Does the new girl have any idea of why Nancy and Margie are walking towards her?"/"Does Tom have any idea of why Kenny chose him to be on the team?". This change was based on the pilot study finding that showed 18/20 students chose the "no" response to the question. For example, in Story 2, the students claimed that the new girl did not know that the girls were walking towards her because "the story said so - she wondered why" or because the new girl had not "spoken" to the girls yet. Thus in the pilot study this question was not included in the total score. The responses to the items were coded based on various coding schemes gleaned from pilot work and from both social-cognitive (e.g., Case et al., 1996; Damon & Hart, 1988; Flapan, 1968; Harter, 1985; Hatcher et al., 1990; Selman, 1980; Yuill, Perner, Pearson, Peerbhoy, & van den

Ende, 1996; Murgatroyd & Robinson, 1997) and TOMliterature (e.g., Happb, 1994; Homer, 1995; Nelson, et al., 1997; Yuill, 1992b). Overall, the coding scheme represented the level of interpersonal understanding based on the increasing cognitive complexity of the responses. Thus, the coding of the responses reflected the child's ability to understand mental states in others and recursive intentional relations, moving fiom simple, obvious ("surface") characteristics to the interaction of several different abstract psychological concepts and the integration of multiple and paradoxical perspectives (see Appendix A1 and A2 for scoring details for both parts of social and self-understanding interview). Drawing on scoring procedures used in previous studies (e.g., Case et al., 1996; Happb,

100A\ ;.. *he --on-4. ,.t..rl.. ~~1..A,, ,,mu -.,r -r- --,-re 2-2,- --Lrl---*- --- 2:. a?-- I //7j, AAA UAU pCIJCIuL JLUUJ, VlUJ VILC* JC*ULCI W a3 ijl VbUPl LG;.3pVll>G, 51 VUlgYQI ULlpllW LlGUlL LVI their "best" answer. That is, if a participant gave an "I don't know" response followed by an appropriate answer, the higher score was given. Similarly, if a participant gave a response that appealed to both physical/behavioural and mental states, the justification would be scored as a mental state. Thus, the total score represented the participant's greatest capacity to understand social situations. If a participant did not first respond to the question, or replied with an "I don't know", the researcher repeated the question. If the participant responded with an "I don't know", the response was scored as a zero. Although this coding scheme contrasts with some social-cognitive research (e.g., Selman, 1980), the decision was based on past research that has coded "don't know" responses as zero and not as missing data (e.g., Hap@, Winner, & Brownell, 1998; Hatcher et al., 1990; Johnson, 1997). In line with past research, a major aim of the present study was to quantify preadolescents' responses to social situations in order to achieve a total score that would represent their ability to understand mental states in others. If a child responded with a high number of "don't know" responses, herhis total sum score would represent a less co,onitively complex response (based on the structural complexity of the responses). However, in the present study, few participants responded with "don't know" after the first prompt. To address this issue of response complexity, data analyses were computed on total scores that included the zero scores of the "don't know" responses and on total scores that did not include scores of zero (i.e., "don't know" responses scored as missing data). T-test analyses showed that the two scores did not significantly differ @ < .05). Thus, the small incidence of prompted "don't know" responses would not likely have significantly affected the results. In general, within psychological-educational research, when a child responds with a "don't know", it could mean that either the child genuinely does not know the answer or that dhe is being evasive and saying "don't know" merely to avoid an answer. In the present study, it appeared to me that the children who did respond to the second prompt with a "don't know" genuinely appeared not to know the answer. I aimed to provide a comfortable and warm interviewing environment where the children wouiri leai ssura enough io provide respnses iiiiii were genuine. rmInus, I assume &ti ihc children's answers were a true reflection of their knowledge regarding the particular interview question. Although it was beyond the scope of the present study, as suggested by various researchers (e.g., Gilligan & Brown, 1992; Harter et al., 1997), future studies on preadolescent social cognition need to investigate the different reasons why some children respond with "I don't know" within research. The answers to the stories were coded by a second, independent coder. Inter-rater reliability was assessed on 38% of the transcripts and the average Cohen's kappa (measurement of agreement that takes chance into account) will be reported for each subscale of the interview in the following sections together with the specific coding details. After inter-rater reliability was assessed, all discrepancies were resolved by consultation between the coders. Second, the responses were coded on a categorical level to determine whether individual differences existed in the various content and emotional valence of the responses. The reason for this coding was mainly exploratory, stemming from various findings from social-cognitive studies that suggest children differ in their emotional valence of responses (i.e., negative/positive/neutral) to social stories, interviews, etc. For example, studies have shown children's responses are more positive when responding to stories of the same gender (e.g., Powlishta, 1995). Other studies on peer relations and social cognition have shown that aggressive boys are more likely to respond in a negative or hostile manner to ambiguous stories (e.g., Dodge & Frame, 1982; Erdley & Asher, 1996; Waldman, 1996). Furthermore, children have also been found to differ according to their valence of responses when describing school experiences (Donelan-McCall & Dunn, 1996). Based on these past studies, the present study coded the social story responses as either positive, negative or neutral. Such a coding scheme was used to explore the possibility that children who differ in their mental state understanding and social relational ability may also differ in the emotional tone of their responses. Scoring: of Conce~tualRole-Takine Ouestions (Section A) Interval level data (to be included in total social understanding: score). Responses to yuesiions i, 2 and 3a wen: scored according io wheiht;r or noi ihey were correci, as eiiher a one (1) or a zero (0), depending on their reference to mental or internal states (thoughts, feelings, judgements, motives). A score of 1 (correct) was given if the response included a reference to a

mental state (eg, "They are going to talk to her because they feel sorry for the new girl") andlor

communicative intent (e.g., "Mark smiled at &MY to tell him he wanted to pick Tom for the team"). Children's responses were coded as incorrect and received zero points if their responses did not refer to any inner states such as a mere repetition or paraphrase of the story (e.g., "They're friends"). If children responded with "I don't know" or refused to respond, the question was repeated. A second "I don't know", was coded as a zero. However, in the present study, all children were able to answer the first two questions of the stories (see conceptual role-taking questions, A1 and A2 in Tables 1 and 2). In other words, no responses included two repetitive "I don't knows". The justification questions (see conceptual role-taking questions A3b and A4b in Tables 1 and 2) were scored according to the complexity of the responses. In contrast to coding the response as either incorrect/correct, the justifications were coded based on the following criteria: tangential responses (e.g., "I don't know"; "I can't explain") received a score of zero; responses that included a behavioural or situational explanation focusing on physical as opposed to mental action received a score of 1 (e-g., "Because they're smiling"); responses that included a mental state or acts of communication and perception received a score of 2 (e-g., "Because they wanted to be her friend."); responses that included an integration of two or more mental states and related them to each other in a coherent manner such as moral judgements (e.g., "Because I've seen kids on the playground who are always the last ones to be picked and you know they're sad but you don't want them to feel bad, you don't want to put them on the spot because it's not the right thing to do") or recursive psychological statements (e.g., "Because Tom thought that they were acting weird and stuff because they were winking and smiling and stuff and he doesn't know if they really want him on their team or not"). A total score was created by summizlg the questions both with and without the response to NancyMargie without and with question 4b was -28 and -33 respectively. Similarly, Cronbach's alpha for the KennyMark without 4b and with 4b was .33, .44 respectively. This increase in the alpha coefficient when question 4b was left in shows that this particular question increased the internal consistency of the conceptual role-taking scale. Further reliability analysis showed that the average kappa for the NancytMargie was 1.00 (without 4b) and -98 (range from -90 - 1-00with 4b) and the KennyMark story was .99 (range -95 - 1-00with and without 4b). Categorical or nominal level data. The emotional valence of question 3a was also scored, with a score of 1 given if the response described a positive scenario, (e.g., "They are walking towards her so that they can make friends"), a score of 3 was given if the response described a ne~tralscenario (e.g., "They are going over to talk to her"), and a score of 2 was given for a response that included a negative scenario (e.g., "Kenny picked Tom because he knows he's a bad player and he wants to make fun of him"). Cohen's kappa for the NancyiMargie story and KennyMark story was .90 and 1.00 respectively. In addition to emotional valence, responses to question 3a were also scored according to their references to the story, the self, or to convention. Responses received one point in the appropriate category if they referred to: a) the story (e.g., She's new to the school"); b) the self or personal experience (e.g., "This has happened to me before") and c) convention or social norms (e.g., "That's what usually happens"). An "I don't know" response was scored as a 7. Cohen's kappa for the story, self and convention categories for the NancyMargie and KennyMark stories were .95, .92, .80 and .93, .85 and 1.OO respectively. Responses to 4a ("Does the new girl have any idea of why Nancy and Margie are walking towards her?" - Story 1; "Does Tom have any idea of why Kenny chose him to be on the team?" - Story 2) were coded as either a 1 (e-g., yes, sort of, some idea, probably), a 2 (e.g., No) or a 3 (e.g., maybelmight have). Cohen's kappa was 1.00 for both stories. Scoring of Empathetic Sensitivity Ouestions (Section B) Interval level data. The complexity of the responses to the question of "How do you think

4L, L-lrn,, f,,l" ,,a ...l.,.? " ,,,,a :, c,., ,,dm ,\ LL- :-:LIl L:-l :--1--a-a UG ,,,,.IIGW ~UUL UIU 1-3 LLUU WILY. ,.,, WGLG D~UICIU UI LWU piw aj UIG uucu aua WGL wluclu WGIUUGU descriptions of emotions, b) and the justification as to why the character felt the particular emotions. Based on works of empathy and emotion understanding (e-g., Case et al., 1996; Donelan-McCall & Dunn, 1997; Hart & Fegley, 1995; Hoffman, 1986; Harris, 1989; Harter & Buddin, 1987; Lewis, 1993, Oatley & Jenkins, 1996), the responses in section a) were coded according to the complexity of the emotion states described. Children received a score of zero if they responded with "I don't know". A score of 1 was given if the response included simple emotions and their derivatives such as angry, sad, happy, scared, upset, etc. A score of 2 was given if the response contained a reference to a psychological~cognitiveor identity-related emotion such as shy, uncertain, nervous, confused, excited, surprised, relieved, etc. Such cognitive emotions are more complex than the basic emotions but do not imply the self-reflection and inner conflict described in the next level of self-conscious emotions (Lewis, 1993). A score of 2 was also given if the response included a combination of two different emotions placed in temporal order, or a two-part behavioral event involving two distinct emotions, "She felt happy at fmt then sad; "He will feel sad if they tease him but happy if they are nice to him." A score of 3 was given if the response referred to complex/moral andlor self-conscious emotions such as guilty, embarrassed, vulnerable, rebellious, self-conscious, etc. Children also received a score of 3 if their responses included recursive identity-related emotions such as phrases and emotions that imply self-reflection (e.g., "She felt self-confident, she felt good about herselr'). A score of 3 was also given if the response included a simultaneous occurrence of two contradicting emotions, that is, emotions with contradicting valences (e.g., "She felt happy but also nervous"; "Half of him was sad and half of him was happy"). Finally, a response that included a coherent combination of three or more psychological emotions was also scored as a 3 (e.g., "She felt nervous, confused and shy.") The average kappa for this particular response for the NancyMargie and KennyNark story was .93 and .97 respectively. Justification responses to "Why does she feel that way?" were coded using the same conceptual framework as the conceptual role-taking responses (i.e., tangential response = 0,

m~~va&n-nlAaLn.G~..-l - 1 -e.*nLnle&nnl - 9 --A L+a-+aA --.nLnl-&nnl - 2\ UALUUUVAACIYWUU V AVLUUL - L ,YJJ UAAVAVeULY - A CUAU AAALUeULVUYJJ UAAVAV5I-UA - J/. The average kappa for the empathetic sensitivity justifications for the NancyMargie and KennyMark story were 1.00 and -97 respectively. A total empathetic sensitivity score was created by summing the two empathy items resulting in a maximum score of 6. Further reliability analysis revealed that the two items within each story obtained high item-total correlations, Nancy/Margie story (B1 - Empathy total, ~(126)= .87, B2 - Empathy total, ~(126)= 37) and for the KennyMark story (B 1 - Empathy total, ~(126)= .90, B2 - Empathy total, 1(126) = -89). Internal consistency was also assessed for the two items by means of Cronoach's alpha that revealed a coefficient of -61 and -52 for the NancyMargie and KennyMark story respectively (alphas of .50 or above are often considered adequate for 2- or 4-item scales, especially if the hypothesized relations are significant; Nunnally, 1978). NorninaVCatworical level data. Based on the previously mentioned emotional development literature, valences of the reported emotions were also coded. Raw counts of emotion terms or phrases were calculated for three categories including positive, negative and neutral emotions. Positive emotions included happy, wanted, needed, special, pleased, lucky, confident etc., negative emotions included depressed, sad, lonely, upset, anxioudworried, disappointed, confused, etc. Neutral emotions included puzzled, surprised, mysterious, strange, etc. Cohen's kappa for the three emotional valence categories (positive, negative and neutral) for the NancyMargie story was 1.00 for all three categories and for the KennyMark story was .97,1.00 and .97. Scoring of Person Perception Ouestions (Section C) Interval level data. The complexity of the responses to the question used to assess how children conceptualize a person, (e.g., "Choose a character in the story and describe herhim - what kinds of things can you think of to describe (character's name?") were coded according to their level of description. Based on person perception literature (e.g., Barenboim, 1981; Davis,

Conklin, Smith, & Luce, 1996; Hart & Fegley, 1995; Yuill, 1992a,b, 1993), and following the general coding scheme of the present study, children's responses were scored in complexity

-...L-- L-- L-L..-L,...... l A-..L-L--- *.. ----I--. -----L-l-L-..l ----..L----OL:l>-- --..-:---a IQU~~~LLVU wuavivua u~auq~uvua LV ~~~JLGA ~~~UVIVE~LUU ~~I~QLL~VUS. LIUIUGII ICCCIVCU & 69 score of zero if they responded with "I don't know". A score of 1 was given if the response included a reference to a) a behavioral or physical description such as "She was blonde" or "He was good at soccer" orb) a stereotypical trait phrase or clich6s such as "She was nice", or "He was bad." A score of 2 was given if the description included a reference to a) a psychological trait such as "She was thoughtful" or "He was shy", b) intentional states such as "She wants to be nice", c) combination of stereotypical personality clich6s and trait characteristics such as "nice and popular", or d) simple unelaborated contradictions such as, "He might be nice but might not be." A score of 3 was given if the response referred to a) self-reflexive or recursive mental states such as "She thinks she's conceited" or "He knows that he's mean" orb) complex psychological andlor elaborated contradictory traits such as "He was both nice and mean because a part of him wanted to be friends with Tom but a part of him wanted to make fun of him." or "She was one of those people who expects and assume that things will happen" or c) psychological comparisons such as "She was much more sensitive than the average person". The average kappa for this particular response for the NancyMargie and KemyMark story was .96 and .92 respectively. Thus, for each story, the maximum person perception score was a total of 3. Nominal or categorical level data. Based on person perception literature that has focused on the emotional valences of traits (e.g., Chaplin, John, & Goldberg, 1988; Gnepp & Chilarnkurit, Norman, 1963; Peabody, 1967) the responses to the person perception question were also coded according to their emotional valence. Raw counts of psychological traits and phrases were calculated for three categories including positive, negative and neutral traits. Examples of positive traits included nice, good, kind, loving, gentle, etc., negative traits included mean, shy, rude, bully, etc., and neutral traits included curious, quiet, independent, tough, etc. Emotions were counted as a trait only if the child did not mention it in section B and described the character as "She is always happy" or "He is a sad person." Cohen's kappa for the three emotional valence categories (positive, negative and neutral) for the Nancyl'argie and KennyMark story was 1.00, .92, 1.00 and .97, ,96 and .93 respectively. Scoring of Alternative Ex~lanation(Section D) Interval level data. Similarly to the previous sections of the storytelling interview, responses to the question that assessed children's ability to envision multiple perspectives, "Is there another way that you can think about this story?", was coded according to their levels of complexity. Based on perspective-takindconceptual role-taking and moral development studies with older children (e.g., see Brown & Gilligan, 1992; Johnston, 1988; Kitchener & King, 1981), children's responses were scored in complexity ranging from behavioural descriptions to complex psychological comparisons. Children received a score of zero if their responses were "No", "I don't know" or a simple "Yes" with no explanation (e.g., "Yes but I can't explain it"), or if the "Yes" response was followed by a repeat of the response to question 3a from the conceptual role- taking section. In other words, the child would say yes, there is another way to think about the story but would only describe their original hterpretation. A score of 1 was given if the response was "Yes" and was followed with a brief explanation with reference only to physicaVbehaviora1 descriptions (e-g., "Maybe they have sleepy eyes"). A score of 2 was given if the response was "Yes" and was followed by a brief explanation with reference to an intentional action or a mental state, (e.g., "Maybe they wanted to make friends with the new girl"; "They are going to make fun of her or tease her"). Finally, a score of 3 was given if the response was "Yes" and was followed by an alternative explanation referring to the concept of intention and one or more references to mental states (e.g., "Maybe they made a different plan to go over and just laugh and ignore her because they might have thought that she wasn't a good friend). The average kappa for this particular response for both stories was .92. The maximum person perception score was a total of 3. Categorical or nominal level data Similarly to Question 3a in Section A, the emotional valence of Section D (Alternative Explanation) was also scored, with a score of 1 given if the response described a positive scenario, (e.g., "They are waking towards her so that they can make friends"), and a score of 2 was given for a response that included a negative scenario (e.g.,

Ill7----. -:-I--> w+-- L ------L 1 ------L-*- AGUJ plbhtu I uu -OUDG UC MU w3 UG 3 ti btid &~i &iid w~Sifi iiii~ fdii 0t hhi"j, md zi score of 3 was given if the response described a neutral scenario (e.g., "They are going over to talk to her"). Cohen's kappa for the NancyMargie story and KennyMark story was .94 and .97 respectively.

4.2.4. Total Social Understanding Story Score The items of each story were summed resulting in the maximum total score for each story (without question 4b and with question 4b) of 18 and 21 respectively, with a high score representing a more sophisticated understanding of mental states and feelings (i.e., a more complex TOM). Analysis of item 4b "Does she have any idea...", revealed that, in contrast to the pilot study, 41% of the girls and 3 1%of the boys responded with either a yes or a maybe. Although only 36% of the participants claimed that the characters had "some idea" of why NancyMargie were walking toward the new girl, or why Kenny picked Tom, almost all of the participants (85%) attempted to justify their answers (question 4b). Based on the notion that children's justification of a particular response assesses their understanding of a particular concept, relizbility analyses were performed on both of the total story scores (with and without 4b). Specifically, internal consistency of the items were measured by calculating Cronbach's alpha coefficients for the NancyMargie story (-64without 4b, -67 with 4b) and KennyMark story (.62 without 4b and .69 with 4b). Thus, the increase in alpha coefficients shows that the addition of question 4b increased the internal consistency of each story. In addition, these alpha coefficients are somewhat higher than the ones obtained in the pilot study (without question 4b), alphas for NancyMargie and KennyMark stories were -59, .48, respectively. Given that an alpha of -60or greater is recommended for basic research (Nunnally, 1978), for a construct that has been newly created and not undergone much use or revision, such reliability scores are adequate. To obtain further internal consistency data, intercorrelations were calculated among the four subscales within each of the stories, (the conceptual role-taking subscale was analyzed with and without item 4b). Table 3 shows significant @ c .05) positive correlations existed among aU of the

~c~sc~cs~pi&kc 2: zcy, t-M&-gicstq-, hitCi cGze;z~cfis i&igcd fiGE -19 icGfic5p ic;c-~&ig Table 3 Interconelations Arnonp Social Understanding: Story Subscalesa

Story 1 (NancyMargie)

Subscale 1 2 3 4 5

1. TOTROR -- a - -- -- 2. TOTROR4B ------3. TOTEMP -38 .58 - - - 4. TOTPP .3 1 .40 .36 - - 5. TOTALT .19* .45 .25 .35 -

Story 2 (KennyMark) 1. TOTROR ------2. TOTROR4B ------

3. TOTEMP -34 .72 - - - 4. TOTPP .18* .58 -38 - - 5. TOTALT .23 .66 .27 .32 -

Note. Data obtained from social understanding story scores. TOTROR = Total third order conceptual role-taking score, without question 4b. TOTROR4B = Total third order conceptual role-taking score, with question 4b. TOTEMP = Total empathetic sensitivity score. TOTPP = Total person perception score. TOTALT = Total alternative explanation score. aN = 128.

*g < .05. All other correlations Q < .001. (without 4b) - alternative explanation)] to .58 [empathetic-sensitivity - conceptual role-taking (with 4b)l. For the Kenny/Mark story, intercorrelations ranged from -18 [conceptual role-taking (without 4b) - person perception) to .72 [empathetic sensitivity - conceptual role-taking (with 4b)l. Table 3 shows that for both stories, the greatest intercorrelation was between empathetic sensitivity and conceptual role-taking. This result suggests that the ability to understand others' perspectives or point-of-view is related to the ability to understand others' emotions. Additional reliability analysis showed significant subscale-total (withiwithout 4b) correlations for both stories. Results showed that the correlations of the subscales-total (with 4b) were higher than the correlations of the subscales-total (without 4b). Regarding the subscale-total (with 4b) correlations, for the NancyMargie story, correlations ranged from .55 to .77. For the KennyMark story, correlations ranged from .58 to -80. Thus, for both conceptual and empirical reasons, the results reported in the present study (Chapters 5 and 6),include the justification response to question 4a of the social stories (question 4b), "How do you know that the new girl/Tom haddoesn't have any idea.." Conceptually, the inclusion of the justification score ensures that the total social understanding score represents the highest level of social reasoning. That is, even if the child did not think that the protagonist had "any idea" of what was occurring in the story, it was the child's ability to just@ this response that adds to the complexity of the total social understanding score. Thus, a child would not receive a lower score if she claimed that the protagonist had "no idea" Empirically, the reliability analysis shows that the internal consistency of the tasks is greater when question 4b is included in the total social understanding score.

4.2.5. Between-Story Correlations and Differences Based on the conceptual framework of the social understanding variable, and psychometric evidence from both pilot work and the present study, two composite scores were computed for subsequent data analysis by averaging the sum of the two stories (including item 4b; study, it was also expected that the three subscales of the two stories would be related. Correlational analyses showed the four subscales to be significantly related ranging from ~(126)= .20, g c .05 (conceptual role-taking subscales) to ~(126)= -42, g < -001 (empathetic sensitivity subscales). The correlation between the two total scores (sum of sections A,B,C,D) is 1(126) =

-60 @ < .001) which is in agreement with results from pilot work that also obtained a significant positive correlation between the two story totals k(18) = .89, Q c .001). To assess for story order bias, paired 1- tests were conducted and found no order effects. Also, to assess for story gender bias (i.e., girls and boys performing better on the stories that used characters of the same gender), paired T-tests were conducted within each gender group to investigate if @ls performed better on the NancyMargie story and boys on the KennyMark story. For story totals including item 4b, results showed that although there was no same-gender bias, the scores on the NancyMargie story (Story 2) were significantly higher than the KennyMark story (Story 1) @ c .001) for both girls (Story 2 M = 15.90, SD = 2.23, Story 1 M = 13.69, SD = 2.09, L(63) = -4.51) and boys (Story 2 M = 13.93, SD = 3.43, Story 1 M = 12.06, SD = 3.52, -t(63) = -4.79). Such within-gender, between-story differences were also found in the pilot study, although this difference was found only among the girls who scored significantly higher (f(9) = - 6.71, g < -001) on the NancyMargie story (M = 11.1) than the Kenny~Markstory (M = 9.6). Interestingly, the pilot study revealed anecdotally that 60% of the children preferred the NancyMargie story (6 girls, 6 boys). Taken together, these results show that a gender-related response bias existed irrespective of the participants' gender, suggesting that perhaps the NancyMargie story was better able to tap into children's understanding of ambiguous social situations than the KennyMark story.

4.2.6. Self-Conceut Given the various conceptual and methodological issues surrounding the construct of self- concept (Hart & Fegley, 1995), for the purposes of the present study, the "theory" theory view of assumes that the self-concept is a multidimensional cognitive representation or theory that is based on an understanding of one's own beliefs, desires and intentions. Drawing on theories that claim the self-concept consists of both a cognitive component (self-understanding) and an affective component (self-esteem) (Damon & Hart, 1988; Harter, 1985), pilot work assessed both aspects of the selfconcept and found that self-esteem and self-understanding were not related [as assessed by Harter's Self-perception Profile for Children (SPPC; 1985) and self-understanding interview based on the SPPC respectively]. Thus, to reflect the complexity of the self-system, the present study assessed both the affective and cognitive aspects of the self. Self-perce~tion~rofile for children (SPPC). To measure the affective aspect of the self, the Self-perception Profile for Children was used (SPPC; Harter, 1985, see Appendix B 1). This is a 36 item self-report questionnaire designed to measure children's perceptions of themselves across various domains. Accordingly, the measure is divided into six subscales including: 1. Scholastic competence - children's perception of their scholastic ability. 2. Social acceptance - children's perception of the degree to which they feel accepted by their peers or feel popular. 3. Athletic competence - children's perception of their ability in participate in physical activity such as sports and outdoor games. 4. Physical appearance - children's perception of the degree to which they feel happy with their physical attributes, e.g., hair, face, body, etc. 5. Behavioural conduct - children's perception of the degree to which they like the way they behave, or do the right things and avoid getting into trouble. 6. Global self-worth - children's perception of the degree to which they like, or are happy with themselves as a person; their perceived worth as human beings.

The SPPC is frequently used in socialcognitive research among preadolescents (e-g., Boivin & Hymel, 1997; Evans, Brody, & Noarn, 1995), with internal consistency reliabilities for all subscales ran,oing between -71 and -86 (Cronbach's Alpha) (Harter, 1985) and test-retest reliability (-59 tc -87, .;L.lli'kg-,.?A S'r;t;scds>. Social and self-understanding:interview se art b) (intraoersonal understandina). In addition to self-evaluation, based on the suggestions of contemporary social-cognitive researchers (Damon & Hart, 1988; Harter, 1985; 1993) and on pilot work, the children were administered a brief, six question semi-structured interview (one question per domain) that assessed the children's particular theories or understanding of their level of competence in a given domain (see Appendix B2 and B3). That is, children were asked to provide a justification for six of their SPPC self-statements. The six items were chosen by Harter (1985) to each best represent one of the self-concept domains from the SPPC including; 1. Scholastic competence - "Some kids do very well at their classwork BUT other kids don't do very well at their classwork (SPPC, item 25). 2. Social acceptance - "Some kids have a lot of friends BUT other kids don't have very many friends" (SPPC, item 8). 3. Athletic competence - "Some kids do very well at all kinds of sports BUT other kids don't feel that they are very good when it comes to sports" (SPPC, item 3) 4. Physical appearance - "Some kids think that they are good looking BUT other kids think that they are not very good looking" (SPPC, item 34). 5. Behavioral conduct - "Some kids usually act the way they know they are supposed to BUT other kids often don't act they way they are supposed to" (SPPC, item 17) 6. Global self-worth - "Some kids are very happy being the way they are BUT other kids wish they were different (SPPC, item 30). Thus, for each of the six SPPC items, the children were asked the following key question, "What is the main reason for why (it's really truelsort of true) that -(item content)", e.g., "What is the main reason for why it is really true that you are good at your school work?'). To answer these questions, it was assumed that the children must reflect on their self-judgements. Thus, such responses may provide a more accurate picture of how children understand their own theories about their social selves, academic selves, physical selves, etc. (Harter, 1985). Coding of the Tntra~ersonalor Self-Understandin9 Interview The children's answers, and the researcher's questions and probes, were audiotaped and transcribed verbatim and the children's answers were scored fiom the transcripts. To obtain a quantifiable measure of the complexity of the child's understanding of herhis self-theory (Harter, 1985), each of the six responses were scored on a scale from 0 to 3 based on pilot work and criteria drawn from the aforementioned social-cognitive literature (e-g., Damon & Hart, 1988; Flapan, 1968; Hatcher et al., 1990; Selman, 1980). A score of zero was given if the child was unable to answer or replied "I don't know." As in the social story interview (see section 4.2.3.) children who initially responded with "I don't know" were asked the question again. If the response was still "I don't know", the researcher went on to ask the next question. More specifically, for item 34 of the self-understanding interview ("What is the main reason why you think that it is really true that you are good looking?") if participants did not respond after a few seconds or appeared uncomfortable, the researcher prompted with, "Is it what you think or what other people think?" The self-understanding section of the interview was completed after all six questions were asked. A score of one (1) was given if the child offered a self-description such as "I have lots of friends" or "I do what the teacher says" but did not give any explanation or interpretation of the description. There were no references to intention, traits or self-reflection. A score of two (2) was given if the child offered a self-description followed by an explanation such as "I have lots of friends because I'm easy to get along with" or "I do well at school because I study hard." Finally, a score of three (3) was given if the child's response was more sophisticated than a level 2 score by including a combination of thoughts and feelings and integrating psychological aspects of both self and other. The self-judgement was followed by an explanation based on a combination of knowledge and affect implying greater self-reflection. For example, "I'm smart because I've never gotten under a 'C'". I try my hardest that's the most important, that I try my hardest to make my parents proud." Responses also included social comparisons at a psychological level as opposed

A- L-L,-.: -----1 --,--2 ---- :- 1 --.-1 L-.- I- - l1XXT-11 :A -L ---.-1 T 1- -1- -A ---.--lC ICT ---.13 1- - LV VFiuavlvula ~vupa1~vuaUI IGVC~ LWV, \G.L;., vv cu IL 311uw s uuw I luun ai IUy3tll. II I GUUU ut someone else I would be someone who people think are really pretty and everyone likes them for the way they look even though I know that's not how you're supposed to judge a person but I still wish I could be them."). For additional examples of responses see coding guide in Appendix A2. A total self-understanding score was obtained by summing the six scores, resulting in a possible total score of 18 that reflects a greater self-understanding (understanding of herhis self- judgement) as compared to a lower score. Inter-rater reliability assessment revealed a kappa of -96

(range -91-1.00 for the six individual items.). In agreement with pilot work (N = 20) which found an acceptable rate of internal consistency (I = -70, Cronbach's Alpha), for the present study, an alpha of .72 was obtained. In addition to coding the response according to level of complexity, the responses were also coded categorically, based on Harter's suggestions to explore the basis of our self-concept. That is, whether our self-concept is derived fiom our own beliefs or derived from statements of what others say about us - referred to as Cooley's (19 12) notion of the 'looking-glass self. Accordingly, the self-understanding responses were given a one (1) if they referred to personal agency, implying that a sense of self is derived from self-judgements, (e-g., "I'm good at school because I do my work."). A score of two (2) was given if the responses referred to the influence of significant others on their sense of self implying that their self-worth is derived from other people's jud,omnts (e.g., "I'm pretty because that's what my mother always tells me"). A score of three (3) was given if responses included references to both self and other (e.g., I'm smart in school because I try hard and my teacher says I can do anything I want"). Finally, "I don't know" responses were scored as a seven (7). Inter-rater reliability analysis for this scoring method resulted in an average kappa of .95 (range -90-1.00). An additional categorical variable was added by scoring the child's response to the question, "What is the main reason for why it's really/sort of true that you are very happy being the way you are?" or "What is the main reason for why it's really/sort of true that you wish you were different?" (question is dependent upon the child's response to item 30 on the SPPC). This q-;ies.dGc;.. isr;sed iG zscss cz&-cii's -GGcGt.L&-,g PiOi-~&S .Acili -G?& ove&, global sense of self-worth. Based on content, the responses were coded into the seven categories including socialhiiends (e-g., "Because I'm popular"), school (e.g., "Because I'm good in school"), family (e.g., "Because I have a caring family"), self (e-g., "Because I'm pretty"), multiple reasons (e.g., "Because I have lots of friends and do well at school"), no ms?in reason (e-g., "I just like being me"), and "I don't know" responses. Inter-rater reliability analysis (Cohen's kappa) for this scoring method demonstrated perfect agreement (kappa = 1.00) between the coded responses of two independent coders. Assessment of understanding self-esteem as a concept. Given the lack of research on children's understanding of self-esteem or the self as a concept (Boivin & Hymel, 1997), children were asked to define self-esteem in their own words. Specifically, they were asked, "In your own words, what does the term "self-esteem" mean to you?" This question was added for exploratory purposes, mainly to investigate individual differences among children's commonsense understandings of the term "self-esteem". Res~onsecodinsr for self-esteem question. The children's answers, and the researcher's questions and probes, were transcribed verbatim and the children's answers were scored from the transcripts. Based on suggestions from Harter et al., (1997), the self-esteem def~tionswere coded according to their reference to self. More specifically, if the response included an objective reference to the self, (e-g., "Self-esteem is how you think about yourself') it was scored as a 1. A score of 2 was given if the response included a reference to both self and other (e.g., "Self-esteem is how you feel about yourself and what other people think about you"). A score of 3 was given if the response included a subjective reference to the self, that is, if the response included a personal example of self-esteem, (e-g., "I have highnow self-esteem because..."). Finally, "I don't know" responses were scored as a 7 (kappa = .96). Such a rating scale was based on past research that suggests differences in children's definitions of social knowledge terms may reflect differences in epistemological assumptions (Harter et al., 1997; Mansfield & Clinchy, 1997). That is, the way in which children define social concepts such as "self-esteem" may be linked to how they view 80

4.2.7. Social Relations Based on pilot work and past social-cognitive literature mentioned in Chapter 2, to assess social interactions with peers the following measures were used: the Harter's (1985) Teacher's Rating Scale of Child's Actual Behavior (see Appendix B4), a modified version of the Social Competence Nomination Form (SCNF, Cauce, 1987; Ford, 1982) and a rating scale sociogram or a Peer Likability scale to assess peer acceptance (Matthews & Keating, 1990) (see Appendix B5). Each measure will be described in the section below. Teacher's Rating: Scale of Child's Actual Behavior fI'Rl3. Based on evidence from both socialcognitive literature on preadolescents and TOMresearch on preschoolers, to explore the relations between teacher ratings of children's classroom behaviour and their self-perceptions, Harter's (1985) Teacher's Rating Scale of Child's Actual Behavior was used. In particular, studies in peer relations have found teacher ratings of social behaviour and social competence to be positively associated with social-co,onitive skills and peer status (e-g., Cantrell & Prinz, 1985;

Edelbrock & Achenbach, 1984). Similarly, evidence fiom TOM studies shows that teacher ratings of social-emotional maturity and prosocial behaviour is associated with preschool children's performance on false-belief and emotional understanding tasks respectively Galonde & Chandler, 1995). Constructed by Harter (1985) to parallel the SPPC, for each of the five specific domains, teachers rate the child's actual behaviour in each area (not how they think the child would answer). That is, the teachers are asked to judge the child's adequacy in each of the following five domains with three items each: Scholastic Competence - Item 1,6, 11; Social Acceptance - Item 2,7,12; Athletic Competence - Item 3,8,13; Physical Appearance, Item 4,9,14 and Behavioural Conduct, Item 5, 10,15). In the present study, the Social Acceptance subscale was used to measure teachermtings of peer acceptance. For example, an item reflecting teachers' perceptions of a child's social acceptance would include: "This child finds it hard to make friends OR for this child it's pretty easy." (Item 2). Teachers must decide what side of the sentence best describes the

t&h they -m9the ~t~k~~ztejtAer rp&;l pec_r SCfi cf 11122- 8 1

The Global Self-worth items are not included in the teachers' scale since they do not translate into attributes which an objective observer can rate. Harter found that only three items per subscale were needed to obtain highly reliable judgements. For the present study, the reliability (Cronbach's Alpha) for each TRF subscale was the following: school .92; social .93; athletic competence, .96; physical appearance -95; behavioral conduct -95. Furthermore, for the present study, the alpha for the total TRF score (15 items) was -92. Peer-rated social commence. The peer group becomes increasingly salient in the life of a preadolescent (e.g., Harter, 1985; Sullivan, 1953). In line with this view, past research has shown that peer ratings offer a more accurate assessment of students' actual social competence as compared to those of teachers or adults (e-g., Harter, 1996; Matthews & Keating, 1995). Therefore, in order to obtain a peer-rated measure of social competence, the present study used a slightly altered version of Matthews and Keating's (1995) adapted version of Ford's (1982) Social Competence Nomination Form (SCNF). This paper and pencil measure consists of three brief vignettes consisting of social situations on which students rate their peers (on a 5-point scale) on the basis of their relative competence in the social problem tasks presented in the stories (see Appendix B5). The vignettes included A) organizing a student play, B) representing the class in expressing sympathy to an ill school staff member and C) helping peers with personal problems. The SCNF has been extensively used in preadolescent samples and is considered to be a highly reliable and valid measure of social competence with high internal consistency reliabilities (Cronbach's alpha, low to middle 90s) and it is significantly related to various social intelligence measures such as Hogan's Empathy Scale (Ford & Tisak, 1983, Hogan, 1969). In the present study, the SCNF was found to have high internal consistency as demonstrated by Cronbach's Alpha (for 3 items, .97), high intercorrelations, ranging from .9 1 to 1.00 and high item-total intercorrelations, (.97, .97 and .96 for items A, B and C respectively). Given this high internal consistency, to allow between-class comparisons of the scores and to obtain a total peer-related score of social competence, the peer ratings were transformed to standardized &scores and a cc~pcsitescc~ Y;ZS c=t& (SLZ cf A, I! zdC). 82

Peer-rated likability (peer acceptance). Given researchers' claims that a rating sociogram is the least problematic and most appropriate method of assessing peer status (e.g., Iverson, Barton & Iverson, 1997; Matthews & Keating, 1995) ,a more global measure of the social relations construct was messed by a rating scale sociogram (see Appendix B5). Each participant was given a class list and asked to rate each other student in the class on a 5-point scale as to how much she or he would enjoy spending time with that person. This measure was included in the present study based on theoretical claims and empirical evidence gleaned from peer relation literature which has found that peer popularity or sociometric status is related to general measures of self-concept (Asher & Dodge, 1986), to various domains of competence including social and academic competence (Cauce, 1987; Wentzel & Asher, 1995) and to social-cognitive skills (see Dodge & Feldman, 1990 for a review). A further rationale for including this measure in this study was based on literature that claims sociometric status and social cognition has been relatively unexplored with regard to gender issues (Daniels-Beimess, 1989;

Dodge & Feldman, 1990). As noted in section 3.2., for the purposes of the present study, to assess general social- relational ability, a composite score was computed, consisting of summed &scores of peer-rated social-competence, popularity, and teacher ratings of social acceptance.

4.2.8. Vocabulary Measures In addition to the assessment of the three main variables, language or vocabulary ability was also assessed. The main rationale for including these language measures was based on strong theoretical claims and empirical reports that social-cognitive development is inseparable from language (e-g., Astington, 1990; Bruner, 1996; Nelson, 1996; Olson & Homer, 1998; Vygotsky, 1965). More specifically, researchers have noted that metacognitive language ability plays a significant role in TOMdevelopment (e.g., Astington, in press; Astington & Pelletier, 1997; Fabricius, et al., 1989; Olson & Bruner, 1996). Also, in regard to methodology, studies have 83

Happ6,1994; Jenkins & Astington, 1995; Matthews & Keating, 1995). For these reasons, the Mental State Verbs Task (adapted from Astington & Olson, 1990; Booth & Hall, 1994; Siddiqui, 1995), was used in the present study as a more specialized vocabulary test, to assess vocabulary specifically related to the understanding of others' perspective (see Appendix B6). In addition, the

Gates-MacGinitie Vocabulary Test, Level D 516 (MacGinitie & MacGinitie, 1994) was used to assess general language competency (receptive) (see Appendix B7 and B8). Each of these two language measures will be described below. Mental State Verbs Task (MVT). To assess social understanding in preadolescents, based on pilot work and recent TOM and psycholinguistic literature that suggests the understanding of metarepresentational or mental verbs may reflect the development of the understanding of mental states in middle childhood and early adolescence (Astington & Olson, 1990; Fabricius, Schwanenflugel, Kyllonen, Barclay, & Denton, 1989; Schwanenflugel et al., 1996), a mental state verbs task was included in the present study. Astington and Olson (1990) claim that children's ability to apply metacognitive and metalinapistic verbs to others may be related to their understanding of others' perspectives (i-e., the attitude expressed in their mental states and speech acts). Astington and Olson (1990) suggest that the use of mental state verbs to characterize others' mental states is more complicated than applying them to the self alone because it includes an implicit reference to one's own state. For instance, the use of the metacognitive verbs 'know' and 'think' to characterise one's own mental state verbs refers only to a degree of certainty. In contrast, the application of mental state verbs to others simultaneously marks one's own stance towards a proposition and the other person's mental state or speech act. For example, if she thinks what I believe to be true, I say "She knows", but if I believe it not to be true I say "She thinks", although from her perspective, the proposition and her attitude may be identical. In the present study, sensitivity to mental state distinctions were assessed by a series of 15 short texts followed by multiple choice responses. Both the order of the vignettes and the mental state verbs were counterbalanced across participants.

l?-11--.:-- c:>>:-.-: /infie\ 1c :.------ll-+-> L-- --"La--- rUllUWlllg DlUUlrjU {IPPJJ, 1J LlCllW WClG C;UllCblCU UUlll VQLlUUD metalin@stic/metacobonitive verb tasks (Astington & Olson, 1990; Booth & Hall, 1994) representing the speech act verb categories of "say", (assert, concede, predict, interpret, confirm), and the mental state verb categories of "think (remember, forget, doubt, infer, hypothesize) and "know" (observe, recognize, conclude, understand, reflect). An example of a "know" item where the correct answer is "observed (from Booth & Hall, 1994) is as follows: The other day, Jane was exploring in the woods. While walking, a large bird flew right in front of her face. She kept her eyes open because she wanted to see this large bird up close. When talking to her fiiend Sally later, Jane told Sally about the bird. Sally denied that Jane had seen a bird. Jane said. " T know a large bird flew in front of mv face." A. Jane observed the large bird. B. Jane recognized the large bird. C. Jane understood that it was a large bird. D. Jane realized that it was a large bird. Scoring: of MVT. Following Astington and Olson (1990), the p cipan scores were corrected for guessing, by subtracting .33 from each wrong answer, negative scores count as zero. Thus, a relatively high score would reflect a more sophisticated understanding of mental state verbs. Reliability analysis revealed a split-half reliability of -53 (Spearman-Brown corrected) which is typical of new instruments that have not undergone a considerable amount of investigation and revision. Siddiqui (1995) and Booth and Hall (1994) reported similar low reliabilities for the comparable items in their studies (.53 and .56 respectively). Gates-MacGinitie Vocabulary Test. This 45-item paper and pencil, multiple choice subtest taken from the Gates-MacGinitie Reading Test series is frequently used by Canadian schools to assess language competency. Indicated by normative data obtained from Grade 6, this test has been found to be a valid and reliable measure of language proficiency with a Kuder-Richardson Formula 20 (KR-20) reliability coefficient of -91 and strong correlations with other reading tests such as the Iowa Tests of Basic Skills (ITBS) @ = -80) and the Comprehensive Tests of Basic

c:c -- 1 A~::P~:: inn^\ u-s \L x uuj - .I J j \LvLabUIILlUG UL. LVLObUUUG, L 7-1. 4.2.9. Sumrnarv of Instruments To summarize, the instruments that were administered in the present study are listed below: Social Understanding 1. Social and Self-understanding Interview (Part a) - Interpersonal or Social Understanding (two vignettes and corresponding questions) Self-Conce~t 2. Self-perception Profile for Children (SPPC; Harter, 1985) 3. Social and Self-Understanding Interview (Part b) - Intrapersonal or Self-Understanding (Six self-understanding questions based on responses gleaned from SPPC) Social Relations 4. Teacher's Rating Scale of Child's Actual Behavior (TRF,Harter, 1985) 5. Social Competence Nomination Form (adapted from Ford, 1982)

6. Peer Likability Scale (adapted ffom Matthews & Keating, 1995) Vocabularv

7. Mental State Verbs Task (MVT, adapted ffom Astington & Olson, 1990;

Booth & Hall, 1994) 8. Gates-MacGinitie Vocabulary Test, Level D 516 (Form 4) (MacGinitie & MacGinitie, 1994) Demographic information regarding the child's age, main language spoken at home and number of siblings was also collected.

4.3. Partici~ants One hundred and seventy-two 6th-grade children drawn from eight different classrooms in six Roman Catholic elementary schools were contacted for participation. The six schools belonged to a semi-rural school district and were located in middle-class neighbourhoods. Three of the participating classrooms contained only grade six students, four classrooms contained a grade 86 varied between classrooms, ranging from 41% (7117) to 96% (23124), with an overall participation rate of 74% (1281172) (mean number of participants per school = 21). This rate of participation meets the minimum rate of 70% suggested by Crick and Ladd for sociometric studies (1989). The final sample &? = 128), consisted of those children whose parents returned a written consent form in response to my cover letter sent home by the classroom teacher. The mean age of the participants was 11;9 (1 1 years 9 months, range 10;9 to 13;2). There were 64 girls (mean age 11 ;9, range 1O;9 to 12;s) and 64 boys (mean age 11 ;9 range 11 ;3 to l3;2). The ethnic composition of the sample was 98% white Euro-Canadian, with 97% of the participating sample reporting English as their first language. Of the four participants who reported English as their second language, three reported Portuguese as their fmt language and one reported

Polish; they were all fluent speakers of English as a second language, as judged by the classroom teacher. The study also included the children's current sixth-grade teachers consisting of five women and three men. In addition to the reasons previously mentioned, early adolescence was selected as the developmental level of interest due to the emergence during this period of reflectivelabstract thought (e.g., Piaget, 1965, Chandler, 1987), an increase in gender-role expectations and behaviours and the increase of self-conflict (e.g., Blos, 1967, Harter, 1985; Hill & Lynch, 1983). Furthermore, the lack of research at this developmental level on social understanding and gender suggests that such an investigation would be fruitful. As previously mentioned, a pilot study was conducted to ensure the feasibility of the present study and to test out newly designed tasks (Bosacki, 1997). Twenty 11- to 12-year-olds, from a university laboratory school, were tested on some of the tasks, or earlier versions of them. None of this work is reported in detail, but reference will be made to various results obtained during pilot testing, including some results from tasks which were not used in the final study. 4.4. Procedure 4.4.1. Establishing Contacts and Consent Upon written permission received from the participating school board, the researcher contacted the principal of each of the participating schools and arranged a meeting with the principals and the teachers who were to be involved in the study. Prior to the meetings, cover letters and informed consent letters were distributed to both the principals and teachers (see Appendix C1 - C3 for letters). After receiving written permission from the principals, and signed consent forms fiom the teachers, the researcher entered each classroom and described the study to all the children while they attended their primary classroom (see Appendix C4 for introduction speech). After a discussion of the voluntary and confidential nature of the study, children were instructed to take consent forms home to their parents and discuss participation (see Appendix C5 and C6). The researcher returned to the participating schools periodically during the week following the distribution of the consent fokto collect the signed ones. Children who did not return a consent form were not allowed to participate in the study.

4.4.2. Administration of the SPPC. SCNF. Peer Liability Scale. Gates-MacGinitie Vocabulary Test and the MVT Testing in each classroom began the week following the initial introductory speech and distribution of consent forms. In each Grade 6 classroom, The SPPC, SCNF, sociogram (Peer Likability Scale), Gates-MacGinitie and the Mental State Verbs task were group-administered according to the corresponding instructions during re,dar class time in a period of approximately 45-50 minutes. Before administering the measures, the responsibility for keeping all responses confidential both on the part of the students and the researcher was discussed. All participating children were given the following general instructions (the first three paragraphs of the Class Introduction to Research Study, see Appendix C4). To ensure confidentiality, subject numbers were assigned to each participant and children were instructed not to place any identifying

:-.frr-..Lrr- -- *LA----..LA--..:-- 'hT----L-:---&- --.-----.-.a -:&Lee ..- .L- Leal- -c *L- ..I...-...---- Irrrvuuauvu vu UG yu~auvuuau~.IWWU~~U~I~~W WGIG IIIVV~~ GIUGI LU UG uabh UI UG ~~Q~~IWUI or to a different room to complete class work that had been previously assigned by their teacher. In addition, desks were separated when necessary, and children were allowed to prop up binders or books to create a visual barrier. In addition, to minimize any possible test anxiety, children were reminded that these questionnaires were not related to their school work and would remain confidential. For four of the classes, the pencil and paper tasks were administered inside the regular classroom. The remaining four classes had the tasks administered within an empty room away from the regular classroom (e.g., the library or an unused classroom). Once the questionnaire booklets were distributed (the order of the questionnaires was as follows: SPPC, SCNF, sociogram, with Gates-MacGinitie Vocabulary Test and the MVT counterbalanced, the researcher followed the directions found in the administration and instructions section of the manual for the SPPC and read aloud the "Instructions to the Child section (Harter, 1985; see Appendix Bl). Following the instructions, the researcher read aloud the 36 statements, allowing the students to complete each question. Children were asked to raise their hands if they did not understand any of the sentences and the researcher attempted to define the statement in similar terms. Upon completion of the SPPC, children were then instructed on how to complete the SCNF and the sociogram, with specific attention given to how to apply the rating scale of 1 to 5 (see Appendix B5). Following the instructions used by past researchers (e.g., Cauce, 1987; Ford, 1982; Matthews & Keating, 1995), each of the three stories was read aloud to the participants, followed by a time period to enable the children to rate each of their classmates accordingly. Once the SCNF was completed, children were asked to complete the sociogram or the Peer Likability Scale. Children did not rate themselves and only rated their classmates who participated in the study. For the final in-class, group administered tasks, children were instructed on how to complete both the MVT and the Gates-MacGinitie Vocabulary task (see Appendix B6 - B8). Children were then instructed to begin and to remember to work on their own and to raise their hands if they had any questions. Following the instructions for both of the language tasks, enough

Amn ..roc 01lrrtxrnA fnv 011 mOrh'A..rOntc tn mnmnlatn thn tmrtr m+ thn;r nw.m nmmn /nnnrnv;mn+alr. 113 1 C LLLUI v. w -Y ..IY AVI YYI r-uIy-W cv IV-~AYCI -Y LWIW UL LUYY vvvu YLWY\U~~AVILLILIULYIJ ZV-z J 89 minutes). When a child was finished, the questionnaire booklet was collected for subsequent data analyses, and she was instructed to sit quietly and given the option of reading a book or working on an uncompleted class assi,wnent.

4.4.3. Administration of the Social and Self-Understandin~Interview (Inter-Intra~ersonal Understanding) Following the group-adminstration of the questionnaires and reading tasks, the researcher thanked the children for participating and gave the following general instructions to the second part of the study: We are now finished the first part of the study. For the second part, I will be taking each of you separately to the library/resource room where I will read you two short stories followed by some questions. I will then ask you questions about some of your answers on the "What I Am Like" questionnaire. The entire interview should take about 20 to 25 minutes. Please remember that you have the right to refuse to answer any of the questions and to stop the interview whenever you wish. Also, your responses will remain anonymous. Are there any questions before I begin to take people to the library? Children were individually withdrawn from the classroom and in a face-to-face manner in an empty classroom or library, participated in a two part interview consisting of 1) the understanding of mental states and intentional relations in others and 2) the understanding of self- judgements. The interview was always conducted in this order (aside from the counterbalancing of the social stories) to develop a researcher-participant relationship necessary for the more personal questions of the self-understanding interview. Following the general interview instructions (see Appendix B2), during the first section of the interview, the researcher read to each participant two brief vignettes describing a socially ambiguous situation (see Tables 1 and 2). Following the reading of each vignette (the order of the vignettes was counterbalanced), the children were asked

L-..c a -La:-- /----LA\-----I ----A :--.. -.I. .I------A :---L-.I bwu lulb-~uuIb~\r~~l~~, ~UUUUIqUC3UVU3. 3~~1IC~~UUWSS ic UICDG GUIIUW YUG~UUU~uzzu .LC be correct in order for the researcher to carry on with the remaining questions. If the child provided an incorrect answer, the researcher re-read the story and repeated the process. If the child answered both control questions correctly, the question sequence described above was given. Gleaned from past research that has used similar narrative-interview procedures (e.g. ,Case et al., 1996; Donelan-McCall & Dunn, 1997; Martin, 1996; Saarni, 1988; Szagun & Schauble, 1997; Yuill et al., 1996;) limited prompts were used if second responses to the open-ended questions were not given spontaneously. For example, when asked open-ended questions such as, eg., "Why do you think this or how do you know this?", "How do you think the new girl feels and why?", and "What kind of things can you think of to describe herhim?", depending on their responses, children were asked "Does she feel anything else?" or "Is there anything else?" until the child responded with a "No, that's all", or "No, there's nothing else I can think of'. The researcher then advanced to the next question. Additional prompts included ones for the Empathetic Sensitivity questions (Section B); if the initial response included global/stereotypical terms like "nice", "bad",good or ambiguous terms like "left-out", children were asked "What makes her nice?" or "And how would that make her/him feel?" Likewise, if the child's response included an ambiguous or bizarre emotion such as "different" or "wondering", prompts such as, "What do you mean by feeling different?" were asked for further clarification. If the child's responses remained ambiguous, the child was presented with the following forced choice question, "Is it a goodhad or inbetween feeling?" For section D (Alternative Explanation); if the children did not respond to "Is there another way that you can think about the story" or if they said "what do you mean?" they were asked, "Is there another reason why KennyMark (NancyMargie) are winking/smiling (nudging/smiling) at each other?" If the child's response remained "No", the researcher asked, "So is this the only way to think about the story?" Administration ended once the child said "yes" to the final question in Section D. Upon completion of the first part of the interview, the researcher read the instructions to the szGnd, "h';"QI;ziSo;;d"p-5 of +U1;2 kCrVjc7(set ApF-- 33). FAs of -he&..iJie-ii; assessed the participants' understanding of their previously completed answers on their SPPC questionnaire. Thus, based on their answers on the SPPC, the participants were asked six questions, each question pertaining to a specific domain of the SPPC (see Appendix B3). In particular, the participants were asked to justify or to provide reasons for their choice of answers on the SPPC questionnaire. As previously mentioned, although not part of the ori,@nal research design, after the first 16 children were tested, the term "self-esteem" appeared frequently in their responses to the inter- and intrapersonal interview questions. To assess participants' understanding of the phrase "self- esteem", at the end of the interview, the researcher asked the children, "In your own words, what does the term "self-esteem" mean to you?" "I don't know" or no response was prompted with "What is the first thing that comes to your mind when you hear the term 'self-esteem'?" The interview was terminated if the child responded with "I don't know." This question was given to the remaining 88% of the children (57 girls, 55 boys). The entire interview was recorded on audiotape for subsequent transcriptions and analyses. Debriefing. Once the interview was f~shed,the researcher debriefed each of the participants before going back to the classroom (see Appendix C7 for debriefing instructions). The procedure lasted about 15-20 minutes per child and total testing time in each of the schools ranged from two to three days. Once testing was completed in the school, the researcher returned to the participating classroom to distribute thank you letters to the parents of children who participated (see Appendix C8). In addition, the researcher briefly explained to the entire class the main purpose of the study and provided an opportunity for the children to ask questions or mention any concerns. Prior to testing, the Teacher's Rating Scale of Child's Actual Behavior (Harter, 1985, see Appendix B4) was given to each of the classroom teachers. The teachers were asked to read the rating instructions and then to complete the questionnaire for each child participating in the study. The teachers were also informed that these forms should be completed independently and that they

.cxrnm tn enmrrlntn thn Cnc.ww-n+A;nn tn thn;r n.rm +;ma rehnAnln n-mo +nc+Lr* hnA ba- nrrm-ln+nA VVYAI CUYVAAApIYCV LIIY AUAIWUYwYACYIIb CV WAY- V I. I. LYAAY -IIYVY.IY. VAIYUCW-5 llUU VVVll CIVILIyIULVU at their school, the teacher rating scales were collected for subsequent data analyses.

4.4.4. Summary of Data Analvses Following the collection and scoring of the data, the analyses were organized into three sections, described in the following three chapters. First, Chapter 5 describes preliminary analyses that examined the overall distribution of scores and descriptive statistics. To check for various effects such as the order of the tasks, school and gender, multivariate analyses of variances were performed. To test if relations existed among social understanding, self-concept and social relations (one of the main goals of the study), Pearson correlations were conducted on the entire sample, followed by separate analyses for &Is and boys on all of the variables. Fisher's Z transformations were used to check for any significant differences among the correlations. Second, based on Baron and Kenny's (1986) models of moderators and mediators, Chapter 6 describes the present study's use of various multivariate analyses to test the hypothesized role of self-concept in the relation between social understanding and social relations. To test for the Self-Concept-=-Mediator Model, the following regression analyses were conducted for each of the SPPC subscale totals: 1) the mediator or Self-concept variable (SC) was regressed on the independent variable of Social Understanding (TOMTOT4B), 2) the dependent variable or Social Relations (SOCTOT) variable was regressed on Social Understanding (TOMTOT4B) and 3) SOCTOT was regressed on both TOMTOT4B and SC. Thus, the three regression equations provided the tests of the connections between the three main variables, with perfect mediation occurring ifTOMTOT4B has no effect on SOCTOT when SC is controlled. Alternatively, the SPPC subscale totals were also tested for the Self-Concept-as Moderator- Model by means of hierarchical multiple regressions, with SOCTOT as the outcome variable for all equations. Specifically, SOCTOT was regressed on 1) TOMTOT4B, 2) the particular SPPC subscale total (SC) and 3) the interaction term between social understanding and self-concept (TOMTOT4B X SC). A significant interaction effect (TOMTOT4B X SC) would show self- 93

test of moderation was also performed by means of 2 X 2 ANOVAs. Following the dichotomization of the TOMTOT4B and SPPC variables into High and Low scores (based on a Median split), the Self-Concept-as-Moderator Model was tested by means of a 2 (TOMTOT4B) X 2 (SC) ANOVA, with SOmOT as the outcome variable. Thus, a significant interaction effect (TOMTOT4B X SC) would indicate that self-concept acts as a moderator in the link between social

understanding and social relations. As previously noted, although analyses were performed on all of the SPPC subscales, Chapter 6 reports only the SPPC subscales that obtained significant results. Third, Chapter 7 contains further analyses that were performed on the categorical data derived from the social stories and the self-understanding interview. These more "interpretive" analyses were performed to provide 1) further support for the quantitative analyses and 2) a fuller, or more descriptive picture of the private and social lives of the preadolescent. Finally, exploratory analyses of the effect of gender were performed on both levels of the data (interval and categorical) and appear throughout the three chapters. CHAPTER V Results: Preliminary Analyses 5.1. Overview This chapter describes the preliminary analyses performed on the data. To examine the distribution of scores, descriptive statistics of the main variables were conducted. To check for school and gender effects, multivariate analyses were performed on the various measures utilising a 6 (School) X 2 (Gender) ANOVAs and MANOVAS (depending on the nature of the dependent variable). To investigate the ielations among social understanding, self-coilcept md social relations, bivariate Pearson correlations were performed among all main variables for the entire sample a= 128) followed by correlations performed on the independent subsamples of girls @ = 64) and boys

(g = 64). To test for differences between the correlation coefficients between the two gender groups, Fisher Z transformations were performed and tested for significance @ c .05). Due to the correlations between the Gates-MacGinitie Vocabulary Test scores and the majority of the main variables, correlational analyses were repeated with the effect of vocabulary partialled out. Results from these partial correlations showed that the majority of the relations among the variables remained significant independent of vocabulary ability, thus, only the bivariate Pearson correlations are reported.

5.2. Descri~tiveStatistics 5.2.1. Distribution of Scores Once scoring and reliability analyses were completed, normal distribution of the scores was assessed by examination of the frequency of the scores as displayed by histograms and stem and leaf diagrams. Means and standard deviations of the scores can be found in Tables 4-7. The majority of the scores approximated normal distribution with the exception of the teachers' ratings and the self-concept responses which were found to be both slightly negatively skewed (a positive bias among the scores). In particular, the subscale that included teachers' ratings of the students' Table 4 Means. Standard Deviations. and Gender Effects for Social Understanding Storv Subscalep

F Group

- ---- Variable Total Girls Boys

(N = 128) @= 64) (n = 64)

TOTROR 8.55 (1.84) 9.16 (1.72) 7.94 (1.77) 16.54** TOTROR4B 12.38 (2.69) 13.37 (2.07) 11.42 (2.93) 18.18** TOTEMP 8-15 (2.16) 8.86 (2.00) 7.43 (2.13) 19.70** TOTPP 4.18 (1.02) 4.42 (0.83) 3.94 (1.14) 8.81* TOTALT 3.32 (1.61) 3.64 (1.60) 3.20 (1.63) 3.MM TOMTOT 24.20 (5.00) 26.08 (4.21) 22.52 (5.03) 23.50** TOMTOT4B 28.13 (5.86) 30.25 (4.64) 26.00 (6.21) 23.58**

-Note. F represents results from MANCOVA, partialling out the effects of Gates-MacGinitie Vocabulary Test (1994) total score. Mean scores shown with standard deviations in parentheses. Variable Definition Possible range TOTROR = Total third order conceptual role-taking score, 0- 12 without question 4b. TOTROR4B = Total third order conceptual role-taking score, 0- 18 with question 4b. TOTEMP = Total empathetic sensitivity score. 0- 12 TOTPP = Total person perception score. 0-6 TOTALT = Total alternative explanation score. 0-6 TOMTOT = Total social understanding score without question 4b. 0 - 36 TOMTOT4B = Total social understanding score with question 4b. 0 - 42 a Scores derived from social understanding stories. The higher the scores, the greater the social- understanding.

-0 .. Q < .Ul. *E< .U5. lVAE< -10. Table 5 Means. Standard Deviations. and Gender Effects for SPPC Self-Conce~tSubscales and Self-

Understandin_oInterview

F Group

Variable Total Girls Boys (N=128) (n=64) (a = 64)

Self-concept (SPPC)a BEHAVE 18.17 (4.00) 18.98 (1.84) 17.36 (4.00) 5.49' SCHOOLS 17.84 (4.11) 17.89 (4.46) 17.78 (3.75) -01 SOCIAL 18.07 (4.76) 18.63 (4.13) 17.52 (5.30) 1.89 SPORTS 18.44 (4.16) 18.13 (4.20) 18.75 (4.13) -79 LOOKS 17.94 (4.35) 17.39 (4.55) 18.51 (4.08) 4.24* GLOBAL 19.61 (3.55) 19.81 (3.72) 19.41 (3.39) -08 Self-Understandingb 12.37 (2.50) 13.00 (2.51) 11.77 (2.36) 7.008*

Note. SPPC = Self-perception Profile for Children, Harter (1985). For the SPPC subscales and self- understanding score, F represents results from MANCOVA and ANCOVA respectfully, partialling out the effects of Gates-MacGinitie Vocabulary Test (1994) total score. Mean scores shown with standard deviations in parentheses. BEHAVE - Total perceived behavioural conduct score. SCHOOLS = Total perceived academic competence score. SPORTS - Total perceived athletic competence score. LOOKS - Total perceived physical appearance score. GLOBAL - Total perceived global self-worth score. aPossible scores ranged from 6 - 24. The higher the score, the greater the self-concept. bTotal self-understanding score from self-understanding interview. Possible scores ranged from 0 - 18. The higher the score, the greater the self-understanding. -p<.Oi.-g<.Q5.A i.ie<.i~. Table 6 Means. Standard Deviations. and Gender Effects for SPPC Teacher's Rating Scale of Child's

Actual Behaviors

F Group

Variable Total Girls Boys

(N=128) b=64) (JI. = 64)

TRFBEHAVE 9.95 (2.41) 10.84 (1.91) 9.05 (2.52) 22.94**

TRFSCHOOLS 9.10 (2.36) 9.34 (2.42) 8.86 (2.81) 2.81M TRFSOCIAL 8.89(2.72) 9.30(2.56) 8.48(2.82) 3.11*

TRFLOOKS 9.22 (1.80) 9.61 (1.84) 8.84 (1.74) 6.04* TRFSPORTS 8.64 (2.46) 8.77 (2.36) 8.52 (2.55) .31

Note. SPPC = Self-Perception Profde for Children, Teacher's Rating Scale of Child's Actual Behaviour, Harter (1985) (TRF). With the exception of global self-worth, teacher rating subscales are parallel to that of the SPPC. F represents results from MANCOVA, partialling out the effects of Gates-MacGinitie Vocabulary Test (1994) total score. Means scores shown with standard deviations in parentheses. TRFBEHAVE = Total teacher rating behaviourd conduct. TRFSCHOOLS - Total teacher rating of academic competence. TRFSOCIAL = Total teacher rating of social acceptance. LOOKS = Total teacher rating of physical appearance. TRFSPORTS = Total teacher rating of athletic competence. aPossible scores ranged from 3 - 12. The higher the score, the greater the teacher rating of student's competence.

**p< .01. *Q< -05. Mp < .lo. Table 7 Means. Standard Deviations. and Gender Effects for Peer Ratings of Social Relations and Vocabularv Measures

Group

Variable Total Girls Boys = 128) @=64) @=a)

Peer Ratings of Social Relations

SCNF .OO (1.00) -36 (-97) -.36 (.go) 29.06**

Vocabulary Measures

GMAT 29.95 (7.36) 29.52 (7.10) 30.38 (7.67) .66

F represents results from ANCOVAs, partialling out the effects of Gates-MacGinitie Vocabulary Test (1994) total score. Mean scores shown with standard deviations in parentheses. For all scores, the higher the scores, the greater the understanding and competence. Variable Definition Possible range SCNF = Total score computed from an adapted version of Social Competence -2.23 - 1.80 Nomination Form (SCNF). PRA = Peer acceptance or peer likability score. -2.64 - 1.84 GMAT = Total score from Gates-MacGinitie Vocabulary Test Score. 0 -45 MVT = Total score from an adapted version of Mental State Verb Task. 0 -15 a For both of the peer-rating measures, scores were standardized within each classroom. MVT scores were corrected for guessing. ** Q < .01. *Q < .os. Q < .lo. physical looks lacked variability with 62% of the scores resulting in a 9/12 subscale total. This shows that the majority of teachers chose to rate their students as "sort of' attractive, thus, they did not want to rate their students as either "unattractive" or "very attractive". Such a positive bias is commonly found among research that asks teachers to rate their students and similzr!y, studies that ask students to rate themselves on their own competencies; hence, standardized measures including the SPPC have been designed to help to minimize social desirability effects (e.g., Harter, 1985; Ladd & Profilet, 1996). Regarding the sociometric scores, following procedures used in related research (e.g., George & Hartmann, 1996; Matthew & Keating, 1995), to allow for inter-class comparison of the sociometric scores, the scores of the SCNF and the sociogram were transformed to standardized scores (Z scores). Furthermore, the means and standard deviations obtained from the present study's remaining measures (e-g., the MVT, the Gates-MacGinitie and the SPPC), were in agreement with findings from past studies (e-g., Astington & Olson, 1990; Gates-MacGinitie, 1992; Harter, 1985).

5.2.2. Assessment of Social Understanding Although it was originally planned that the combined score of the two socially ambiguous stories and the mental verbs task would provide an adequate measure of the understanding of mental states and feelings in others, with a higher score representing a more sophisticated level of social understanding, such a composite TOMscore was not used in the present study due to the lack of correlations between the MVT and the social stories (Pearson correlations revealed no significant relations @ < -05) between the MVT total score and each of the social story total scores, with and without question 4b). Although the MVT total score and the sum score of the two stories together (with question 4b) revealed a si,@ficant correlation k(126) = -18, g c .@I),this relatively small correlation suggests that the two tasks are measuring different kinds of understanding. Thus, in the present study, TOM or social understanding was reflected by the composite score of 100

(4.2.3), although parallel analyses were performed on the composite story score without question 4b, results revealed similar but not as robust effects as with the story total with question 4b. Thus, results are reported only for the social story score inclusive of question 4b.

5.2.3. Order Effects of Tasks A one-way analysis of variance (ANOVA) on the Mental State Verb task &ta with the four task lists serving as the independent variable revealed that there was no effect of the order of the questions (list effects) on the scores @(3,124) = -78, g > -05). Paired $-tests with the order of task presentation (Gates-MacGinitie or MVT first) as the independent variable revealed no significant order effect on either the vocabulary scores e(126) = .30, g > .05) or the mental state verb task scores e(126) = -.06, g > .05). To test for the order effect of the presentation of the social stories, paired 1-tests were performed on both of the story totals, with results showing no significant differences among the NancyMargie story e(126) = -1.4, g > .05) or the Ke~yMark story a (126) = -1.7,~> -05).

5.2.4. School Effects Due to the small number of students in two of the classes of two different schools (5 and 7 respectively), to ensure adequate statistical power, each of the smaller classes was added to the respective larger class in each school, resulting in a total of six schools. To check for differences between girls and boys both between and within the different schools, 6 (school) X 2 (gender) ANOVAs were performed on each of the dependent measures that were not comprised of intercorrelated subscales (e-g., Gates-MacGinitie). Likewise, 6 (School) X 2 (Gender) MANOVAS were performed on the dependent variables that included significantly intercorrelated sub-scales

(e.g., SPPC). Results indicated a main effect for school @ (5, 122) = 3.89, E < .01). Tukey tests showed one school in particular performed si,@ficantly higher (M = 13.15) than two other schools

(M = 10.75 and M = 11.26, respectively) on the total KennyMark story score. Due to the relative fiO~Gg2zc~~J. of +kcc~h&z & sKi-GEG,,z kcke-V~rili& fro= ;hc= siiiIiB:c-*- EG 101 significant between-school differences had been expected. Thus, given that the majority of the dependent variables did not differ with respect to school, and results indicated no significant interaction effects (School X Gender), school as a variable was subsequently eliminated from further analyses.

5.2.5. Gender Main Effects Overall, the results from the ANOVAs and the MANOVAS revealed a number of significant gender main effects for each of the main groups of dependent measures (social understanding, self- concept and social relations). Results obtained for each section will be reported. With respect to the social understanding stories, the MANOVA revealed a sigdicant multivariate main effect for gender, WWs lambda = .8 1, F (1, 127) = 7.04, (g < .001) followed by significant univariate gender main effects for all of the subscale totals except for the Alternative Explanation response which was marginally significant @ c .lo). Examination of the means shows that compared to boys, girls' scored higher on the social story subscale totals (see Table 4). Given the strong association between the social story responses and the vocabulary measure (see Table 8), such gender effects could be due to differences in general vocabulary ability rather than differences in social understanding. To control for the possible confounding variable of language ability, an analysis of covariance (MANCOVA) was performed in which the mean scores of girls and boys on the social stories were examined with their Gates-MacGinitie Vocabulary score acting as the covariate. Table 4 shows that even when the effect of vocabulary ability is controlled for, the effects of gender remain significant @ < .01) with girls scoring significantly higher than boys on three out of the four subscales (marginally higher on the fourth). It appears, then, that girls scored higher than boys on the social understanding measure, independent of vocabulary ability. Results from the multivariate analysis of variance (MANOVA) performed on the SPPC subscales showed significant multivariate main effects effects for gender, Wilk's lambda = -88, F (1, 127) = 2.8, @ c .001). Follow-up univariate analysis of variance for each of the self-concept eavLeeq1ae -.raelaA &-;

Understanding: Stories. Self-concept and Self-understanding

Variable Total Girls Boys

(N = 128) = 64) (IL= 64)

Social Understanding Storv Subscalesa GMAT TOTROR4B -09 TOTEMP .45** TOTPP .36** TOTALT .27* TOMTOT .46** TOMTOT4B .53** Self-Conce~t(SPPC) BEHAVE -07 SCHOOLS -13 SOCIAL -02 SPORTS .21M LOOKS -04 GLOBAL -04

.04

Note. GMAT - Gates-MacGinitie Vocabulary Test TOTROR4B = Total third order conceptual role-taking score, with question 4b. TOTEMP - Total empathetic sensitivity score. TOTPP - Total person perception score. TOTALT - Total alternative explanation score; TOMTOT - Total social understanding score without question 4b. TOMTOT4B = Total social understanding score with question 4b. SPPC - Self-Perception Profile for Children, Harter (1985). BEHAVE - Total perceived behavioural conduct score. SCHOOLS - Total perceived academic competence score. SPORTS - Total perceived athletic competence score. LOOKS - Total perceived physical appearance score. - GLOBAL Total perceived global self-worth.- score. a Scores derived fiom sociai undestanding stories. ~l'oraiseii-undemanding score derived from seif understanding interview. **p<.01. *p< .05. Mp<.lO. to boys, girls scored significantly higher on the behavioral conduct that is, girls perceived themselves as more well-behaved. Such results are congruent with Harter's (1985) original set of findings although Harter also found that boys scored significantly higher than girls on perceived athletic competence and perceived global self-worth while in the present study, these differences were not significant. To test for gender effects with vocabulary ability controlled, a MANCOVA was performed on the SPPC subscales with scores on the Gates-MacGinitie as the covariate (see Table 5). Results showed that compared to boys, girls still perceived themselves as significantly more well-behaved even while controlling for the effects of vocabulary ability. Also, once the effects of vocabulary were controlled for, the gender difference concerning physical appearance became significant with boys scoring higher than girls. This result suggests that language ability may act as a confounding variable in gender differences among girls and boys self-perceptions of their physical appearance. Regarding the self-understanding measure, results showed a significant main effect for gender E(1,127) = 7.00, g c .001, with girls (NJ = 13.00) scoring higher than boys (M = 11.77). To control for the possible influence of vocabulary ability, an ANCOVA was performed with the Gates-MacGinitie Vocabulary scores as a covariate. Results showed that the gender difference remained significant even while controlling for vocabulary ability, demonstrating that girls scored higher on the self-understandhg measure than boys independently of their vocabulary ability.

, MANOVA results from the peer social competence rating showed a multivariate main effect for gender, Wills lambda = -61, F (1, 127) = 5.1, @ c .001). Univariate ANOVAS showed main effects for gender for both measures, with girls scoring significantly higher than boys on the SCNF and the sociometric or peer acceptance rating (see Table 7). As with the social understanding and self-concept measures, a MANCOVA was performed with the Gates- MacGinitie Vocabulary Test scores as a covariate and results showed that independent of vocabulary ability, girls' social competence ratings remained significantly higher than boys whereas girls' peer acceptance ratings were only marPoinally higher than boys.

Tm -rr*rA +A tnorhn-' m&.rna nCA41Awr.'a Lh.,.<,...- rpn..l+e Lm+he Kfi A hTnVA rhr.-.-A LU AYbYYU bU LYWAAUALI AUC11160 UA UAAAAUIWAA J VCIIIUVAUCUO, AWLUW UVIII UlU AT-.(U T A DUVWbU 104 significant multivariate main effects for gender, Wilk's lambda = .84, F (1, 127) = 4.63 1, (E < .001). Univariate ANOVAs revealed significant main gender effects for the behavioural conduct and physical appearance and marginal gender effects for social acceptance, with girls receiving higher ratings in all areas (see Table 6). To assess the independence of these ratings from the children's vocabulary ability, a MANCOVA was performed with the vocabulary scores as the covariate. Results showed that gender differences among the teachers ratings remained significant, even after controlling the effects of vocabulary ability. Interestingly, results from the MANOVA showed that girls and boys did not differ in their scores on the Gates-MacGinite Vocabulary Test or on the Mental State Verbs task (see Table 7). These results demonstrate that in the present study, girls and boys did not differ on general vocabulary ability or in their understanding of mental state verbs.

5.2.6. Teacher-Gender X Student-Gender Interaction Effects Due to the overrepresentation of female teachers, teacher ratings were not tested for gender effects. However, once the sample was divided into four groups (girls with female teacher, girls with male teacher, boys with female teacher, boys with male teacher), the means of the variables were tested for interaction effects (gender of teacher X gender of student, see Table 10). Controlling for general vocabulary ability, ANCOVAs and MANCOVAs were performed on the appropriate variables. Significant interaction effects were found for only the subscale scores on the social understanding stories. Results of a MANCOVA on the subscale scores of the social understanding stories yielded significant multivariate interaction effects (teacher-gender X student- gender), Wilk's lambda = -91, E(5, 119) = 2.48,~< .05. The univariate analyses of covariance illustrated in Table 10 shows that only the total empathetic sensitivity score and the alternative explanation score reached marginal significance. Examination of the means shows that for all of the social understanding subscales, girls with male teachers scored the highest whereas boys with female teachers scored the lowest. This result shows that independent of vocabulary ability, girls

+t& de *Ecztr"?+& t&= ~~t.3.~~ p.iZ l~~e~iz~g (.-.& &ttd& b. sxid sS;q Table 9 Pearson Correlations Between Gates-MacGinities Vocabularv Test (1994) and Social Relations peer and Teacher Ratinos) and Mental State Verb Understanding

Group

Variable Total Girls Boys

(N=128) (11=64) @= 64)

Social Relations GMAT Peer-Ratings SCNF PRA .22* .16 .30* SPPC Teacher Rating TRFBEHAVE .19 -16 .28* TRFSCHOOLS .47** .44** .52** TRFSOCIAL .08 .16 .28* TRFLOOKS .06 -01 .15 TRFSPORTS -.03 -14 -.I6 Mental State Verbs Task (MVT) .47** .53** .42** Note. GMAT - Gates-MacGinitie Vocabulary Test total score. SCNF - Total score computed from an adapted version of Social Competence Nomination Form. PRA - Peer acceptance or peer likability score. SPPC - Self-perception Profile for Children, Teacher's Rating Scale of Child's Actual Behavior, Harter (1985) (TRF). With the exception of global self- worth, teacher rating subscales are parallel to that of the SPPC. TRFBEHAVE = Total teacher rating behavioural conduct. TRFSCHOOLS = Total teacher rating of academic competence. TRFSOCIAL = Total teacher rating of social acceptance. TRFLOOKS = Total teacher rating of physical appearance. TRFSPORTS = Total teacher rating of athletic competence. MVT - Total score !corrected) from an adapted version of Mental State Verbs Task. Table 10 Means and Standard Deviations for Social Understandine Storv Subscales and Univariate Analvses of Teacher-gender X Student-gender

TeachedStudent Group

Social Understanding Male/Gi.rl MaleBoy Female/Girl FemaleBoy

Story Subscalesa @ = 25) h=22) @ = 39) (n = 42)

TOTROR 9.24 (1.36) 8.32 (1.52) 9.10 (1.93) 7.74 (1.87) -33 TOTROR4B 13.44 (1.85) 12.23 (2.30) 13.26 (2.19) 11-00 (3.15) .92

TOTEMP 9.48 (1.53) 7.48 (2.04) 8.46 (2.13) 7.36 (2.20) 3.59M TOTPP 4.44(0.87) 4.27(0.94) 4.41(0.82) 3.77(1.21) 1.33 TCYTALT 3.77(1.61) 3.40(1.36) 3.44(1.63) 2.93(1.72) 3.19M

Note. E represents results fiom univariate analyses that followed a MANCOVA, partialling out the effects of the

Gates-MacGinitie Vocabulary Test (1994) total scores. Results showed si,pificant multivariate interaction (teacher- gender X student-gender) effects (see text). Mean scores shown with standard deviations in parentheses.

Variable Definition Possible ranoe

TOTROR - Total third order conceptual role-taking score, without question 4b. 0 - 12

TOTROR4B = Total third order conceptual role-taking score, with question 4b. 0- 18

TOTEMP - Total empathetic sensitivity score. 0- 12

TOTPP - Total person perception score. 0-6

' TOTALT - Total alternative explanation score. 0-6 a Scores derived from social understanding stories. The higher the scores, the greater the social understanding. " g < -10. scores), whereas boys with female teachers achieved the lowest social understanding scores.

5.3. Correlational and Remession Analvses 5.3.1. Social Understanding: and Self-Conce~t To test the relation between social understanding and self-concept, Pearson bivariate correlations were performed between the total social stories score (TOh4TOT4B) and the six subscales of the SPPC (see Table 1 1). Results obtained for the entire sample revealed only one significant correlation; behavioral conduct was positively correlated with TOMTOT4B (r(126) = -27, g < .01). This result shows that children who perceived themselves as more well-behaved, scored higher on the social stories. Repeat analyses for the two gender groups revealed that this relation was simcant for boys (r(62) = .30,g c .05) but not significant for girls (r(62) = .09, ns). To further examine the relations between social understanding and self-concept, Pearson correlations were performed between the subscale totals of the social stories and the subscales of the SPPC (see Table 12). Correlations obtained fiom the entire sample revealed that behaviouralconduct was the only self-concept subscale significantly related to the social story scores @ c .05), with positive correlations between conceptual role-taking (. 18), empathetic sensitivity (.27) and person perception (-21). Examination of the correlations among the girls and boys revealed that sipficant correlations between social understanding and self-competence were found only among the boys, and only with the empathetic sensitivity subscale. Table 12 indicates that for the boys, empathetic sensitivity total scores were significantly positively related to perceived behavioral conduct and perceived academic competence and marginally related toperceived physical appearance and to perceived global self-worth. In contrast, for girls the relations between empathetic sensitivity and self-concept appeared to be in the opposite direction, with all of the correlations negative (although not significant, see Table 12). Fisher Z transformations showed significant gender differences @ c .05) in the correlation

czff;,cic2s+wp.!zn eI?jF&P,GC Stzit;,yit;l=n 1) FTP2y& k&x,AyJccE&2Ct lG = 2-83>,2) Table 11 Pearson Correlations Between Social Understanding Storv Total Score. Self-Conce~tand Self-understanding

Group

Variable Total Girls Boys

(N=128) @=64) (9 = 64)

Social understanding Story Total Score (TOMTTOT4B)

BEHAVE SCHOOLS SOCIAL SPORTS LOOKS GLOBAL Self-Understanding

-Note. The total social understanding story score is the sum of all of the social understanding story subscales. TOMTOT4B = Total social understanding story score. SPPC = Self-Perception Profile for Children, Harter (1985). BEHAVE = Total perceived behavioural conduct score. SCHOOLS = Total perceived academic competence score. SPORTS = Total perceived athletic competence score. LOOKS = Total perceived physical appearance score. GLOBAL = Total perceived global self-worth score. =Totalself-understanding score from self-understanding interview. ***p< -001. **p< .01. Mpc .lo. perceived physical appearance (Z = 2.13) and 3) perceived global self-worth (Z = 2.19). These results provide weak support for the predicted direction of the relation between social understanding and self-concept, showing a nonsignificant negative relation for girls and a significantly positive relation for boys. That is, the correlational analyses shows that girls with higher empathetic sensitivity scores reported lower self-concept scores, whereas the opposite was found for boys; boys with higher empathetic sensitivity scores reported higher self-concept scores. These results also show that self-perception regarding one's conventional behaviour may be an aspect of self-concept that is particularly related to social understanding, especially empathetic sensitivity. To assess the amount of variance in social understanding that self-concept as a whole accounted for independent of language ability, hierarchical multiple regression analyses were conducted. After controlling for language, all of the SPPC subscales were entered into the equation as one block. In general, results showed that as an undifferentiated construct, self- concept accounted for a small but significant amount (8%)of the variance in social understanding (F = 3.55, p < .05). However, parallel analyses for both gender groups showed that among boys, once language was controlled for, self-concept accounted for 10% of the social understanding variance, whereas for girls it accounted for 6%. Furthermore, to assess the relations between the multiple dimensions of the selfconcept and social understanding (independent of language ability), a combination of hierarchical and stepwise multiple regression analyses were conducted (see Table 13), in which general language ability was entered first (GMAT) followed by a stepwise regression analyis on the SPPC subscale scores that were entered as predictors of social understanding (total social story score). Analyses were conducted for the total sample and the gender groups. Table 13 shows that after controlling for general vocabulary ability (entering Gates-MacGinitie Vocabulary scores into the equation first), perceived behaviour conduct was the only SPPC subscale to enter the regression equation, accounting for 5% of the variance in the social story total scores. This result shows that children's

--on&nrrr nFhn.tr ..mil Lho.vort +ha.. n- n;-:&m-+l-. -A-.+AL--+~+rr cL- --rLr-r- :- &I.-:- ---:-I YYIYIYUVIICI YA IIWVV WYII-WIIUVYU UAUJ UAU J16LCIlIUUIUJ QUUUIWUk LU U1G VQI1UC.G Ill UlFill DmlQl Table 12 Pearson Correlations Between Social Understanding Story Subscales and Self-Conce~t

Self-Conce~t(SPPC) variable BEHAVE SCHOOLS SOCIAL SPORTS LOOKS GLOBAL

SociaI Understanding? Tot.&J = 128)

1. TOTROR4B .18* .05 .10 .10 -07 .ll

2. TOTEMP .27** .ll .04 -.01 -.01 .05

3. TOTPP .21* .13 -.02 -.01 -.05 .05

4. TOTALT -08 .OO .O 1 .06 -.04 .OO

Girls = 64)

1. TOTROR4B -06 -.01 -.01 -13 -08 .09

2. TOTEMP -.@Ia -.02 -.02 -.03 -.15, -.16b

3. TOTPP .20 .12 .02 -.I9 .00 .02

4. TOTALT .ll .02 .O 1 -.04 -11 .02

Boys = 64)

1. TOTROR4B .18 .ll .13 -13 .17 .10

2. TOTEMP .44,** .25* .02 -05 .23, M .23b M

3. TOTPP .16 .14 -.09 -16 -.06 .05

4. TOTALT -00 -.05 .OO .18 -.17 -.04

Note. Correlations with matching subscripts significantly differ at e < .05, Fisher's Z Test. See Table 8 for variable definitions. t Scores derived from social understanding stories. **e<-01. *e<-05. Mp < .lo. understanding, even after controlling for the effects of general vocabulary ability. Table 13 also indicates that for boys, perceived athletic competence was the only SPPC subscale to enter into the equation, accounting for 6% of the variance in the social story total scores. In contrast, Table 13 shows that for girls, once general vocabulary ability was entered in the equation, the self-concept scales did not si,onificantly contribute to variance in the social story scores. Thus, results from multiple regression analyses performed on each gender suggest that language ability and self- concept play different roIes in preadolescent girls' and boysf social understanding.

5.3.2. Mental State Verb Understanding. Self-concept. and Social Relations Given the lack of a robust relation between the social story scores and the Mental State Verb (MYT) scores (r(126) = .l8,2 c .05)., correlations between mental state verb understanding (as assessed by the MVT) and the SPPC subscales were calculated separately from the social story totals. In contrast to the relations between social understanding (as assesssed by social story totals) and self-concept, significant negative correlations @ c .05) were found between perceived athletic competence and mental state verb understanding for the entire sample (r(126) = -.19) and for boys (r(62) = -.28) although these relations became nonsignificant once the effect of vocabulary ability was partialled out. Among girls, there were no si,onificant relations between mental state verb understanding and self-concept. Furthermore, for all three samples (total, girls, and boys), there were no significant relations found between mental state verb understanding md self-understanding. To assess the relation between mental state verb understanding and teacher ratings, Pearson bivariate correlations were performed between the Mental State Verb 0scores and the various subscales of the Teacher's Rating of Child's Actual Behaviour (TRF). A sigmficant positive correlation @ c .05) was found between teacher ratings of academic competence and the total MVT score (r(126) = -19) among the girls (r(62) = .30), whereas this relation was not significant among the boys (r(62) = .08, ns). Significant negative correlations were found between teacher ratings

~f yweivec! ~tfiek~~~~t,P~= & me~td~taa i+~-ct~Acg fcr the bjg(452) = --3c, Table 13 Hierarchical Multiple Regression Analvsis Predicting Social Understanding Using: Gates-MacGinitie

Vocabulary Test (1 994) and Self-Conce~t

Variable -B Mult. R -Rz -R2 Change -F

Total = 128) Step 1 GMAT .24*** .29 .10 -10 13.14***

Step 2 BEHAVE .33*** .38 .15 -05 10.42***

Girls b = 64) Step 1 GMAT .29*** -45 -20 .20 15.33***

Boys b = 64) Step 1 GMAT .24* -29 .09 -09 5.54* Step 2 SPORTS .41* -39 .15 .06 5.12*

Note. Social understanding represented by total social understanding story score. Self-concept represented by Self-Perception Profile for Children (SPPC, Harter, 1985). B = unstandardized Beta coefficients. GMAT - Gates-MacGinitie Vocabulary Test (1 994) total score. BEHAVE - Total perceived perceived behavioural conduct score. SPORTS - Total perceived athletic competence score. NO subscales of the SPPC significantly entered the equation. ***p< .001. *p< .05. 113 g = .02 whereas the relation was not significant for the girls (r(62) = .05, ns; z = 2.0, g c -05). However, once the effects of general vocabulary ability were partialled out (GMAT), all correlations failed to reach significance. To assess the relation between the understanding of mental state verbs and peer-related social competence, repeat analyses as described above were performed on the total score obtained from the Mental State Verbs Task and the peer-rated social competence and likability variables. ResuIts showed that the MVT totd was significantly positively related to peer-rated social competence scores for the total sample (r(126) = .22, g c -05) and among the girls (r(62) = -30, p c -10) whereas this relation did not exist for the boys (r(62) = -15, ns). Given the strong association between the MVT and language ability found both in past research (e.g., Booth & Hall, 1995) and in the present study (r(126) = .47,g < .001), partial correlations were conducted by partialling out the Gates-MacGinitie Vocabulary scores to control for the effect of vocabulary ability. Partial correlations revealed that the relations between mental state verb understanding and peer-rated social competence became nonsignificant once the effect of vocabulary ability was controlled for. Likewise, Pearson bivariate correlations between the peer acceptance scores and the MVT showed no significant correlations. Results showed no correlations between MVT and peer- rated likability.

In support of the prediction that the self-concept consists of both cognitive and affective components, correlations between the total score on the self-understanding interview (representing the cognitive component) with the subscale totals of the SPPC (representing the affective component) showed that the two groups of scores were weakly related. Among the three correlational matrices (total sample, girls, and boys), the only correlation to reach significance at the .05 level was between self-understanding and perceived behavioural conduct in boys (r(62) = .37). These findings support previous findings from pilot work that also found no relation ~~~~~nseE-*E&-ct2ctg zdthe SPPC s2bdes. Given that the self-understanding interview measured different aspects of the selfconcept than the SPPC, to assess the relation between the understanding of self-perceptions and social understanding, correlations were performed between the total score obtained fkom the self- understanding interview and the total score obtained from the social stories. Table 11 shows robust positive correlations @ c .001) between the self-understanding score and the social understanding score for the total sample, the girls, and the boys (-57, -46 and -61 respectively). FoIIow-up partial correlational anaiyses with the Gates-MacGinitie Vocabulary scores partialled out, revealed that this relation remained significant independent of vocabulary ability. Thus, children with higher self-understanding scores also achieved higher social story scores, representing a greater social understanding.

5.3.4. Social Understanding and Social Relations Social understand in^ and mer-ratings of social commtence. To examine the relation between social understanding and social competence, Pearson correlations were performed between the summed Z-scores of the SCNF, the sociogram or peer likeability scale and the total obtained from the social stories (see Table 14). Results showed that children's total peer- ratedsocial competence scale was significantly related to their total score on the social stories, for the total sample (r(126) = .39, Q c .001) and also among the gLls and the boys ~(62)= .30; ~(62)

= .30, Q < .05). However, once the effects of vocabulary ability were partialled out, only the partial correlation among the total sample remained significant (g(126) = .29, g < -001). Pearson correlations were also conducted between the sociogram or peer acceptance ratings and the social story total score. Results showed that compared to the SCNF scores, the relation between the peer acceptance ratings and the social story score was not as robust with a significant correlation @.< .05) for the entire sample (r(126) = .23) but the correlations for the girls and boys did not reach ~i~cance(. 11 and -17 respectively). Also, partial correlational analysis revealed that the relation between the peer acceptance ratings and the social story total became marginally c;an;firant nnre the effertc nf vnrahmrlanr tvere t-nntrnllerl fnr 1G1(135!, = -16, --=---- "------A&--- -- ."----, ..-a- -"------A < -1c)- TLlez Table 14

Pearson Correlations Between Social Understanding Story Total Score. Social Relations (Peer and

Teacher Ratings) and Vocabulaq Measures

Group

Variable Total Girls Boys (N=128) @=64) @ = 64)

Social Relations Social Understandha Story Total Score rTOMTOT4B)

Peer-Ratings SCNF PRA SPPC Teacher-Rating TRFBEH4VE TRFSCHOOLS TRFS OCIAL TRFLOOKS TRFSPORTS Vocabulary Measures GMAT

Note. SCNF = Total score computed from an adapted version of Social Competence Nomination Form. PRA = Peer acceptance or peer likability score. SPPC = Self-Perception Profile for Children, Teacher's Rating Scale of Child's Actual Behavior (Harter, 1985) (TRF). TRFBEHAVE = Total teacher rating behaviod conduct. TRFSCHOOLS = Total teacher rating of academic competence. TRFSOCIAL = Total teacher rating of social acceptance. TRFLOOKS = Total teacher rating of physical appearance. TRFSPORTS = Total teacher rating of athletic competence. GMAT = Total score from Gates-MacGiitie Vocabulary Test (1994) score. MVT = Total score (corrected) from an adapted version of Mental State Verbs Task ***p< .001. **pc .018Q< .05. M Q c .lo. 116 together, these results show that social competence ratings related to social understanhg, whereas peer popularity did not. Given the positive correlation between the SCNF score and the peer acceptance score k(126) = .79, p < .001), the two Z-score totals were summed and Pearson correlations were conducted with this total peer social competence rating and the social story total. Results showed that for the entire sample, the total peer-rated self-competence scores were significantly related to the social story total b(126) = .24, g < -01). Partial correlations controlling for the effects of vocabulary ability showed that this relation remained marginally significant once the vocabulary scores were partialled out (pr(126) = .l7, p < .lo). Separate correlation matrixes for the two gender groups failed to reveal any significant correlations among the total peer-rated social competence score and the total social story score. Given the present study's prediction that the relation between social understanding and social competence ratings wouid differ for girls and boys (see Table 15), the peer-rated social competence scores (SCNF) were correlated with the subscale totals of the social stories. Examination of the three separate Pearson correlation matrices revealed that for the total sample, the peer-related social competence scores were significantly positively related @ c .05) to the social story subscale totals with correlation coefficients ranaging from .22 (SCNF-Alternative Explanation) to -33 (SCNF-Empathetic Sensitivity). For girls, the correlations between peer-rated social competence and the social story subscale totals ranged fiom .20 @ < .lo, Alternative Explanation) to -38 @ < .01, Person Perception). For boys however, peer-rated social competence was significantly related only to conceptual role-taking (.36,g < .01) and empathetic sensitivity (-24, p < .05). To control for the effects of vocabulary ability, partial correlations were performed, with the Gates-MacGinitie score as the covariate. Results showed that for the entire sample, peer-rated social competence remained significantly related to conceptual role-taking and empathetic sensitivity. However, the correlational matrices for each gender revealed that once the effect of

V~&$p hil lit^ WIG p&se4 2% Fr--m&G S~iz~~~nd~n~a .woe e;rm;&t.m-+lw. -1nfaA +nnrrlr. J ---J --- y--U- .V W eI6AILUILLLlUJ AWAUCVU CVVLUJ Table 15 Pearson Correlations Between Social Understanding Storv Subscales and Social Relations (Peer and Teacher Ratings)

SPPC Teacher Ratings Peer Ratings

Variable BEHAVE LOOKS SCHOOLS SOCIAL SPORTS SCNF PRA

Social Understanding? Total a= 128)

1. TOTROR4B .06 .18* .15M .09 .19* .28** .19*

2. TOTEMP .17M -.02 .12 -.05 -.02 .33** .20*

3. TOTPP .13 .ll .15M .OO -08 .30** -15 M

4. TOTALT -03 .13 -08 .04 .08 .22** -04

- -

Girls @ = 64)

1. TOTROR4B -.02 -.Ma .06 -.16b .13 .25* .02

2. TOTEMP -.01 -.I2 .ll -.I3 .06, 25' -00

3. TOTPP .16 .04 -23 M -07 .07 .38** -17

4. TOTALT .02 .05 .15 -.04 .12 .20 -15

Boys @ = 64)

1. TOTROR4B -.03 .28,* .19 .22bM .23 .36** -21 M

2. TOTEMP -.I2 -.08 .08 -.09 -.I2 .24M .13

3. TOTPP -.02 .08 .05 -.I2 -08 .14 .05

4. TOTALT -.05 .14 .O 1 .07 -04 .16 .12

-Note. Correlations with matching subscripts significantly differ at Q c -05, Fisher's Z Test. See Tables 8 and 9 for variable definitions. Scores derived from social understanding stories. ** gc .Ol. *e<.05. .lo. person perception and conceptual role-taking among girls and boys respectively. Repeat analyseswere performed for peer acceptance ratings and the total peer-rated social competence scores. Results showed no significant correlations among social understanding scores and the peer acceptance ratings but the total peer-rated social competence score was significantly related to conceptual role-taking among the entire sample (Dr(126) = -23 12 < -01) and among boys (jg(62) = -29, 12 < .Ol), even with the effects of vocabulary ability partialled out. No sigmficant correlations were found among the girls. Social understanding and teacher-ratings of children's behaviour. To investigate the connections between social understanding and teacher ratings of children's classroom behaviour, Pearson bivariate correlations were performed between the subscale and overall totals of the social stories and the subscale totals of Teacher's Ratings of Actual Behavior (see Tables 14, 15). Table 14 shows no significant correlations between the teacher ratings and the overall total of the social stories (TOMTOT4B), with correlations ranging from 1(126) = -.11 (Teacher rating of girls7social acceptance - TOMTOT4B) to ~(62)= -22 (Teacher rating of girls' academic competence - TOMTOT4B). To further examine the relations between teacher ratings, correlations were performed between the subscale teacher rating totals and the subscale totals of the social stories. Table 15 shows that although no significant correlations were found for the entire sample, separate correlational matrices for girls and boys show that among boys, positive correlations exist between conceptual role-taking and teacher ratings of 1) physical appearance b(62) = .28, g < .05), 2) athletic competence k(62) = .23, g < -10) and 3) social acceptance k(62) = .22, g < -10) with no such relations found among the &ls (-.04, .13, -.I6 respectively). These results show that boys who scored higher on the conceptual role-taking subscale were rated as better looking and more athletically and popular by their teachers. Fisher Z transformations showed that between-gender correlation coefficientsfor the relation between conceptual role-taking and teacher ratings of physical appearance and social acceptance were significantly different (Z = 1.8 1, Q < .10 and z = 2.13, g < -03 respectively).

'T-Lla 1 < alr~ch~mrr a marAnallxr o;-;C;,-ant ,-n..lrrlrlt;nn Lhrr-n +-a,-hnr -tArrr AC ArlS7 A UVAY Ad -0- 0-V I. 0 U AU-6YII.Y J OAb-IUU- IVAAIAYUVIIVIL V. WAA AULWbCI VI PA 119 academic competence and girls' person perception ( ~(62)= .23,g < .lo). This results shows that as girls' scores on the person perception subscale increased, their teachers rated them as more academically competent. However, given results from previous research (e.g., Verschueren, Marcoen, & Schoefs, 1996) and correlations obtained from the present study that show children's vocabulary ability may influence teacher ratings of children's behaviour (see Tables 8 and 9), . partial correlations were conducted between the teacher ratings and the social story totals with the Gates-MacGinitie Vocabulary scores as the covariate. Results showed that once the effect of children's vocabulary ability was controlled for, no significant correlations existed between teacher ratings and social understanding scores. Hierarchical multiple regression analyses showed that after controlling for general vocabulary ability, teacher ratings (entered as one block), accounted for only 5% of the total social understanding story variance (see Table 16). Table 16 also shows that teacher ratings of social acceptance shares the strongest relation to social understanding (B = -.50, a.Similar analyses were performed for girls and boys. Results showed that once the effects of general vocabulary ability are contolled for, teacher ratings entered as a block account for 10% of the variance in social understanding in girls and for 4% of the variance in social understanding for boys. Table 16 also shows that for girls, teacher ratings of social acceptance are negatively related to social understanding (I3 = -.71, Q < .05), whereas for boys, the strongest relation is found between physical appearance and social understanding (B = -46, ns).

5.3.5. Self-concept and Social Relations Peer and Teacher Ratings) Although the main purpose of this study was to examine the relations between social understanding and both self-concept and social relations, correlations were also performed on the SPPC subscales and the teacher and peer ratings. Significant correlations were found between the two sociometric measures (social competence ratings and popularity) and all of the teacher rating subscales. Teacher ratings of social acceptance obtained the highest partial correlations

(kA2F+2r?t ef geI!ttmr \I~&~iL3'&zt;').?!it& the ~r ,r?.t&gS cf +L\ Szi~=c~Fttz-~*+Am bs ! Table 16

Pierarchical Multiple Remession Analvsis hedictinp Social Understanding Using Gates-MacGinitie Vocabulm

Test (1994)and Teacher's Rating Scale of Child's Actual Behavior

Step Variable B Mult. R -R2 -RL Change -F

Total = 128)

1. GMAT .25* -31 -09 .09 13.14***

2. Total Teacher Ratings - - .56 .14 .05 03.21*** BEHAVE .29 LOOKS .37 SCHOOLS -.06 SOCIAL -SO SPORTS -.47

Girls @ = 64)

1. GMAT .27* .45 .20 .20 15.33***

2. Total Teacher Ratings - - .58 -30 .10 04.07* BEHAVE .09 LOOKS -.07 SCHOOLS .20 SOCIAL -.71 SPORTS .2 1

Boys @ = 64)

1. GMAT .32* -29 .08 .08 5.71*

2. Total Teacher Ratings - - .35 .12 .04 1.34 BEHAVE -26 LOOKS .46 SCHOOLS -.34 SOCIAL -.23 SPORTS -35

& Social understanding represented by total social understanding story score. Teacher ratings based on the Teacher's Rating Scale of Child's Actual Behavior 0(Harter, 1985) . B = unstandardized Beta coefficients. GMAT - Gates-MacGinitie Vocabulary Test (1994) total score. BEHAVE - Total teacher rating of behavioural conduct. LOOKS - Total teacher rating of physical appearance. SCHOOLS - Total teacher rating of academic competence. SOCIAL - Total teacher rating of social acceptance. SPORTS - Total teacher ratings of athletic competence. ***ac .001. *pc .05. 121 and popularity b(126) = .59, .58, g < -001, respectively). In contrast, the peer ratings of social competence and popularity obtained the lowest partial correlations with the teacher ratings of athletic competence (Dr(126) = .4 1, .3 1,g < -001, respectively). These correlations between teacher and peer ratings of social competence are in line with a recent study by Harter (1996), although she did not partial out the effects of general vocabulary ability. Given the vast amount of empirical evidence supporting a strong link between positive self- concept and social-relational ability (e.g., Dodge & Feldman, 1990; Harter, 1985; Matthews & Keating, 1995), this relation was also expected to be found in the present study. Regarding peer- rated social competence, for the total sample, significant positive correlations @ < -01) were found between the SCNF total and the SPPC subscales of behavior conduct (r(126) = .30), perceived academic competence (r(126) = .28), perceived social acceptance (r(126) = -37) and perceived global self-worth k(126) = .25). Significant positive correlations @ < .01) were also found between peer-rated likability and the SPPC subscales of perceived behavioural conduct (r(126) = .26), perceived social acceptance k(126) = SO) and perceived global self-worth (r(126) = .37) but not with perceived academic competence (r(126) = -13, ns). Separate gender analyses showed similar results for boys. However among girls, peer-rated social competence and likability were si@cantly related to both perceived social acceptance and perceived academic competence, but perceived global self-worth was only related to peer acceptance (r(62) = .30, p < .05), and not to peer-rated social competence (r(62) = .16, ns). That is, girls who were rated as relatively more socially competent by their peers viewed themselves as more popular and academically competent but did not necessarily report more positive feelings of self-worth. To control for the effects of general vocabulary ability, partial correlations were also computed and showed similar results. In general, these results demonstrate that children who received higher peer-rated social competence ratings also viewed themselves as more well- behaved, competent in school and popular. However, students who were rated as popular did not necessarily view themselves as academically competent. Furthermore, results show that for girls,

A&, ea-ea AC ,l~L,l c~lC-.v.nr~hcnc n --a- ;C -n&+&.nl.. -1-+nd +a-.----I-~C- L--*--* L- ---:-I -I- WOWWA 6AWULI. -AA .. VALU UO U PI-U 13 YV31UVUIJ IUICbL\N bu~~~a&~Ly UUL Lu3UbM.I competence ratings. In contrast to self-perceptions, no significant correlations were found between self- understanding scores and the peer-related social relations score. Pearson bivariate correlations between the SPPC subscales and the teacher rating subcales (TRF)showed that each of the corresponding subscales were significantly related to each other, ranging from ~(62)= 25, g < .05 (Self-perceptions and teacher ratings of behavioural conduct in boys) to ~(62)= -61, E < .OOl (Self-perceptions and teacher ratings of social acceptance). These results suggest that teachers ratings of children's behaviour is similar to children's self-ratings of the same behaviour. To control for the effect of vocabulary ability on the relations between self- competence and teacher and peer ratings, parallel analyses were conducted with the effects of vocabulary ability partialled out, resulting in similar fhdings. Correlations between the SPPC scbscales and the teacher rating scales also showed that teacher-rated behavioural conduct was related to all of the SPPC subscales except for athletic competence. Regarding the perceived global self-worth SPPC subscale, for girls, the link was the strongest between perceived global self-worth and teacher ratings of social acceptance (r(62) = .32, g < .OS), whereas for boys, perceived global self-worth was most strongly related to teacher ratings of academic and athletic competence (r(62) = .3 1, .29 respectively, p < .05). In general, these results show that links between teacher ratings and preadolescent's self-perceptions may be linked in different ways according to gender.

5.3.6. Sibling Effects Correlational analyses of the main variables with sibling data revealed few significant relations, with the effect of vocabulary ability partialled out, the only relation found was a positive correlation (r(126) = -18, g < .05) between the total number of siblings and the peer acceptance scale. This suggests that the greater number of siblings one has, the greater is one's peer acceptance or popularity. However, Pearson correlations calculated separately for each gender

A lthniiuh amnna a'--orniin r7 --r-x~ald ---- -many +&ic~t %wit& t&z 1rrl;n y&.;,&hleS & $&u0 Aata-- 'MU"' 0- --- a the girls, the Gates-MacGinitie score was not related to any sibling variables, among the boys, a marginally sigmficant correlation @ < .01) was found between the Gates-MacGinitie vocabulary score and the number of older sisters, whereas for girls, the same correlation was -18m. To test the significance of the difference between girls' and boys' correlation coefficients, Fisher's Z transformations revealed that the relation between vocabulary ability and the number of older sisters a child has si,onificantly differs according to the gender of that child, z = 2.18, E < -05. Thus, for boys, as their number of older sisters increases, their vocabulary scores decrease. To further examine siblings effects on social and self-understanding, partial Pearson correlations were performed, controlling for the effects of the Gates-MacGinitie score. Regarding the connections between sibling data and social relations, among girls, a significant partial correlation @ < .05) was found between the total number of sisters and peer acceptance, ~(62)= -27. In contrast, among the boys, there were no significant correlations between sibling data and social competence ratings. Interesting gender differences emerged among the sibling data and the self-concept scales. Among girls, a significant relation @ < .01) was found between the total number of younger brothers and perceived behavioral self-concept, ~(62)= -39 whereas for boys such a relation did not exist, p (62) = .05, z = 1.9, p = .06. Among boys, perceived behavioral self-concept was si&~cantly negatively correlated @ < .01) with the total number of older sisters, ~(62)= -.3 1 whereas this relation was not significant for girls, ~(62),-.04, z = 2.0,~=.046. Taken together, these results suggest girls with a higher number of young brothers tend to see themselves as more well-behaved, whereas boys with a higher number of older sisters tend to see themselves as less well-behaved. Another self-concept variable that showed signifkant correlational gender differences was that of school or perceived academic self-concept. Among the boys, perceived academic competence was significantly negatively correlated with the total number of older brothers, ~(62)= -.36, p < .01, whereas this relation was not significant for girls, pr(62)

= .05, z = 2.04, Q < -05. This result demonstrates that boys tend to see themselves as less academically competent as the number of their older brothers increases. differences (although the differences between the conelations were not significant). Among girls, the total number of sisters was marginally @ c .lo) negatively related to their scores on the social perspective-taking or conceptual role-taking scores on the social stories, ~(62)= -.24, whereas this relation was not sigmficant for boys, ~(62)= .lo. This result shows that girls' scores on the conceptual role-taking scale decreased as their number of sisters increased. Among boys, the total number of older siblings was significantly negatively related to their empathetic sensitivity scores on the social stories, ~(62)= -.22, whereas this relation was not significant for girls, pr(62) = - .02. Thus, the more older siblings boys have, the less ability they have to score on the empathetic sensitivity scale. Taken together these results demonstrate that an increase in siblings (sisters for ,girls and older siblings for boys) is negatively related to social understanding ability. Also, in the light of recent research on the effects of family size and siblings on children's social understanding that has found a significant positive conelation between social understanding and a larger family size (e-g., Jenkins & Astington, 1996; Perner, Ruffman & Leekam, 1994), results from the present study call attention to the need for closer inspection of the gender of children's siblings.

5.4. Summarv Overall, results showed that independent of vocabulary ability, girls scored higher than boys on both the social and self understanding measure. With regard to behavioural conduct, girls rated themselves and were perceived by their teachers as more well-behaved than boys. Girls also received higher peer and teacher ratings of social competence and popularity than boys. No gender differences were found among the general vocabulary and the metacognitive language scores. Furthermore, marginal interaction effects between the gender of teacher and the gender of students indicated that girls with male teachers scored the highest on the social understanding score, especially on the empathetic sensitivity section. In contrast, the social understanding scores showed that boys with female teachers scored the lowest. Furthermore, results regarding the links between self-concept and social-relational ability provide support the multidimensionality of the 125

With regard to the links between self-concept and teacher and peer ratings of social- relational ability, results show that particular dimensions of preadolescent's self-concepts are positively associated with both teacher's ratings and peer rating of social behaviour, however such relations may differ according to gender. More generally, such results provide support for the notion that social constructs such as selfconcept and social-relational ability are multidimensional and reciprocal in nature (e.g., Cauce, 1987; Harter, 1996). With regard to siblings, gender differences were found in that the number of sisters was negatively correlated to the conceptual role-taking aspect of social understanding among girls. In contrast, among boys, the number of older siblings was negatively related to the empathy aspect of social understanding. Taken together, these sibling effects suggest that an increase in siblings may have a negative influence on children's ability to understand mental states in others. Related fmdings showed that for boys, the number of older sisters was negatively related to their general vocabulary ability. This fmding suggests that older sisters may have a negative influence on their younger brothers' language ability. In general, the results from the correlational analyses concerning the relation between social understanding and peer-rated social competence and popularity suggest that peer-rated social competence may have stronger links to social understanding as opposed to peer acceptance or "likability." Also, results suggest that links between social understanding and social competence among peers may differ for girls and boys, with particular aspects of social understanding having a differential influence on children's peer interactions. More specifically, the ability to perceive the psychological characteristics of others and the ability to understand another person's perspective may influence peer social competence in girls and boys respectively. Such findings are in line with mixed findings gleaned from pilot work and current research on social understanding and social interactions where some studies have found social understanding to be significantly related to peer popularity (e.g., Dockett, 1997). CHAPTERVI Results: Hypothesis Testing of Mediation and Moderation Models 6.1. Overview One of the main objectives of the present study was to test whether self-concept acted as either a mediator or as a moderator in the relation between social understanding and social relations. Put differently, this study aimed to test whether or not social understanding acted through selfconcept in affecting social relations (mediation model), or whether self-concept affected the direction andlor stre~!~.thof the relztio~between social understanding and social relations (moderation model). Such an investigation may help to delineate the complex links between social understanding and social relations. Accordingly, the use of such models warn against the tendency of some researchers to assume that a simple linear relation exists between social cognitive and social-relational ability. That is, the examination of the intervening role of self-concept may help to explain why some preadolescents score high on social-co,pitive tasks, but do not like themselves andlor experience difficulty in the social-relational domain. As previously mentioned, to test these models in the present study, Baron and &MY'S (1986) analytic procedures to make the moderator-mediator variable distinction were explicity followed. Although analyses were computed for both aspects of the self-concept (SPPC and the self- understanding interview), significant results were found only with the SPPC. Thus, only the tests of the SPPC subscales as either a moderators or mediators will be reported.

6.2. Testing Self-Conce~tas a Mediator According to Baron and K~MY'S(1986) guidelines, to demonstrate mediation, the predictor or independent variable must be si,onificantly related to a) the mediating variable and b) the dependent variable. To test whether or not selfconcept acted as a mediator between social understanding and social relations, a composite score was created to represent a total social relations score by adding the following Z-score totals together: teachers' ratings of social acceptance, rating of peer acceptancelpeer likability and peer ratings of social competence. The creation of this composite score is in line with the notion that the social ratings of teachers and peers may have a cumulative effect on children's self-concept (e.g., Harter, 1996; Pekrun, 1990). To fmt satisfy the initial conditions of mediation, Pearson bivariate correlations were computed between the independent variable or the total social story score (TOMTOT4B) with a) each potential meditor (SPPC self-concept subscale) and b) the dependent variable defined or the total social relations score (SOCTOT). Examination of the correlations revealed that only the perceived behavioral-conduct subscale of the SPPC correlated significantly with TOMTOT4B (j(126) = .27, g < -01) and SOCTOT (r(126) = -20, g < -05). Thus, the test for mediation was performed with only the perceived behavioural-conduct aspect of the self-concept acting as the potential mediator. Following Baron and Ke~y(1986), a series of three hierarchical multiple re,oression equations were estimated (see Table 17). First, the social understanding story total (independent variable) was used to predict perceived behaviour-conduct (mediator) as the outcome variable, second, the social story total (independent variable) was used to predict the total social relations score (dependent variable) as the outcome variable, and third, both the social story total score (independent variable) and perceived behaviour-conduct (mediator) were used to predict the total social relations score (dependent variable) as the outcome variable. To establish mediation, in the first equation the social story total must affect perceived behaviour-conduct; in the second equation the social story total must affect the total social relations score; and in the third equation, perceived behaviour-conduct must affect the total social relations score. Table 17 shows the three regression equations for the total sample and for the separate samples of girls and boys. For the total sample, only the first and third equations show a significant effect. Thus, perceived behaviour-conduct does not meet the conditions of the meditional model and consequently, it does not mediate the relation between social understanding and social relations. To control for the effects of vocabulary ability, repeat analyses were conducted but the Gates-MacGinitie scores were entered into the regression equation before the ether v~+&les*y-es~ek -%Med~e~~ificzct- -AJ$e, ic tst f~:gzza&: &ffe=zZs L+= Table 17 Hierarchical Multi~leRemession Analvses Testing. Perceived Behaviour Conduct as a Mediator

Variables in the Equation B Mult. R E2 &- I?

Total = 128) Outcome: BEHAVE TOMTOT4B .18** .27 .07 .07 9.74** Outcome: SOCTOT TOMTOT4B .12 -16 .03 .03 3.46 Outcome: SOCTOT TOMTOT4B .08 BEHAVE .21* .25 .06 -06 4.27*

Girls (Jl=64) Outcome: BEHAVE TOMTOT4B .06 -09 .O 1 -01 0.55 Outcome: SOCTOT TOMTOT4B .02 -03 .OO .OO 0.05 Outcome: SOCTOT TOMTOT4B .O1 BEHAVE .23 .23 .05 .05 1 .70

Boys (n = 64) Outcome: BEHAVE TOMTOT4B .19* -30 .09 -09 6.13~ Outcome: SOCTOT TOMTOT4B -08 -12 .Ol -01 0.88 Outcome: SOCTOT TOMTOT4B .05 BEHAVE .13 .17 .03 .03 0.91

&& B = unstandardized Beta coefficients. Perceived Behaviour Conduct from Harter's (1985) Self-perception Prome for Children (SPPC). BEHAVE - SPPC subscale total representing perceived behaviour conduct. TOMTOT4B = Total score of social understanding stories. SOCTOT - Total social relations score (smedz-scores of total peer and teacher social ratings). **Qc .01. *Q < .05. Table 18 Hierarchical Multi~leRemession Analvses Testing: Perceived Global Self-worth as a Moderator with Social Relations as the Outcome Variable

Step Variables B hd!dLB B2 PZ Change -F

Total a= 128) 1. TOMTOT4B -.I9 .02 .OO .OO 0.06 2. GLOBAL -.01 .30 .09 .09 6.18** 3. GENDER .~7~ .33 .ll .02 5.09** 4. TOMTOT4B -01 .34 .ll .OO 3.89** X GLOBAL

Girls = 64) 1. TOMTOT4B - 1.22** .ll -01 .O 1 0.77 2. GLOBAL -1.52* .34 -11 .10 3.90* 3. TOMTOT4B .06** .48 .23 .13 5.82** X GLOBAL

Boys @ = 64)

1. TOMTOT4B -50 -02 -00 .OO 0.02 2. GLOBAL -92 .27 .07 .07 2.40 3. TOMTOT4B -.03 .3 1 -10 .03 2.13 X GLOBAL

Note. Social relations represented by the computed total social relations score. B = unstandardized Beta coefficients. Perceived global self-worth from Harter's (1985) Self-Perception Profile for Children (SPPC). GLOBAL = SPPC subscale total representing perceived global self-worth. TOMTOT4B = Total score of social understanding stories. ** Q < .01. *Q < .05. M~ < .lo. mediational models, parallel analyses were conducted for girls and boys, both showing nonsignificant results.

6.3. Testing Self-Conce~tas a Moderator In general, moderation implies that the relation between two variables changes as a function of the moderator variable. Thus, the statistical analysis must measure and test the differential effect of the independent variable on the dependent variable as a function of the moderator. In this study, to test the moderation model that predicts self-concept would act as a moderator by affecting the direction andlor strength of the relation between social understanding and social relations, correlations among the variables were inspected to search for the needed nonsi,onificant relation between the independent variable (social story total) and the potential moderator (self-concept). Given that only one of the six SPPC subscales significantly correlated with the social story total, all the subscales except for the behaviour conduct subscale were tested for moderation. Given that both the independent variable and the moderator are continuous variables, following Baron and K~MY(1986), a series of hierarchical multiple regressions was performed on five of the SPPC subscales. For each of the five subscales, the total social understanding story score (TOMTOT4B), gender, the particular SPPC subscale and the interaction term (TOMTOT4B X GLOBAL) acted as independent variables and were used to predict the total social relations score (SOCTOT) as the outcome variable. A significant interaction effect between the moderator and the independent variable of self-concept (TOMTOT4B X SPPC subscale) would show that self- concept moderates the link between social understanding and social relations. Results from the five sets of the three multiple regression equations showed that a significant moderation effect was found for only one of the self-concept subscales (GLOBAL - perceived global self-worth). Results of the regression analyses (see Table 18) indicate that perceived global self-worth (GLOBAL) was a significant predictor of the total social relations score, accounting for 9% of the variance. Gender was entered as the third step and was found to

+" 2 Egh*,r si*5cz$ F+ctcz cf LI~$#,+A y.A$ +,ac. ~#,*~,?m.1m.+kg f#,: zyz cf variance. Finally, the fourth step involving the interaction term (TOMTOT4B X GLOBAL) was entered into the equation and did not account for a significant amount of variance. This lack of significance of the interaction term illustrates that for the total sample, perceived global self-worth did not moderate the relation between the social understanding stories scores and the total social relations scores. The sigdlcant effect of gender suggested that the moderation effect of perceived global self-worth might differ for girls and boys, thus, the three-step multiple regression model was repeated separately for the girls and the boys. Table 18 indicates that perceived global self- worth acts as a moderator only among the girls. That is, the interaction between perceived global self-worth and understanding of social stories signif~cantlypredicts girls' social relations scores, accounting for 13% of the variance. Furthermore, to check for language effects, parallel analyses were performed with the Gates-MacGinitie scores entered into the equations first. These analyses showed similar results. Thus, perceived global self-worth significantly moderated the relation between social understanding and social relations independent of general vocabulary ability.

6.3.1. Perceived Global Self-worth as a Moderator between Social Understanding and Social Relations As Baron and Kenny (1986) suggest, if the moderator and the independent variable are both continuous, another method of testing for a moderation effect is to dichotomize the moderator and independent variable where the analysis is a 2 X 2 ANOVA and moderation is indicated by a si@cant interaction effect. To further examine the moderating effect of perceived global self- worth on the relation between social understanding and self-concept, based on median splits for each variable, two levels of the independent and the moderator variable were created. Based on the median split calculated fiom the social story totals, the sample was divided into two groups representing high and low social understanding, with scores equal to or greater than 29 or less than 29 respectively. Similarly, two groups were created based on a median split with the perceived

=nrr-l ----alnhal "--celf-wn&h .. "-- --"--"crnrec -"rpnrpc~nting I------k;& & 12~?/F,'c~+& g12bs seLf-x:Jc,rf&, +t& rU- to or greater than 20 or less than 20 respectively. McNemar's test was used to check for si,pificant differences among the distributions of participants in the four cells (see Table 19). Regarding the distribution of girls and boys among the cells, results showed that there was a significantly greater number @ < -05) of girls (g = 28) than boys (g = 12) in the High ToM/High Self-worth group. Counter to prediction, no significant gender distribution difference was found for the High ToMfLow Self-worth group (15 girls, 10 boys). To check for the moderating effect of perceived global self-worth on the relation between social understanding and social relations, a 2 X 2 X 2 (Gender X Social Understanding X Perceived Global Self-Worth) ANOVA was performed first, with the total social relations score as the dependent variable, to check for overall main and interaction effects. Although no sigmficant effects were found, to further examine the gender patterns of total social relations score among the four groups, 2 X 2 (Social Understanding X Perceived Global Self-Worth) ANOVAs were performed separately for girls and boys with the total social relations score as the dependent variable. Results of the ANOVAs showed that for both girls and boys, the interaction effect (Social Understanding X Perceived Global Self-Worth) was nonsignificant F(1,63) = 1.97 and -F(1,63) = .20 respectively. Although the lack of significant interaction terms indicate that perceived global self-worth did not si,onificantly moderate the influence of social understanding on the total ratings of social relations, inspection of the means among the four groups demonstrated interesting gender patterns. Characteristics of higMow mou~sbased on social understandiw ability and ~erceived global self-worth of girls and bovs. Table 19 illustrates the means of the social relations scores as rated by the peers, teachers and combination of peers and teachers. Although a marginally significant three-way interaction @ < .lo) was found only for the total peer social relations score, F (1, 127) = 3.57, interesting trends were found. Table 19 shows that across all sets of social relations scores (peers, teacher, peers and teacher), ,&Is in the high social understandingflow

nlA.nl ..,=,lC ..~~r*h -rn:.r,A +Lo -i---.-.-a-- +L1 2-1- 1- -,-L-L t--..-. YYIYYATVU 6IVWUl 0-U W VIUI 6IVUp ACIlrbA VCIU Ub IWWbDLIUU~D LUllWll5 ULb pL3. l,LL ClULLUCWl., Wy3 Table 19 Mean Social Relations Scores for Girls and Boys as a Function of Social Understanding and Perceived Global Self-worth

Ratings

Group -n Peep TeacheP Peer and Teachelc

High TOMTOT4Bl High GLOBAL

Girls 28 3.46d (4.36) 9.68 (2.38) .82 (2.46)

Boys 12 .30d (3.98) 8.50 (1.73) -.40 (.65)

High TOMTOT4BI Low GLOBAL

Girls 15 -.23 (7.38) 8.20 (3.14) -.72 (3.26)

BOYS 10 -1.39 (5.1 1) 8.40 (2.73) -.51 (.78)

Low TOMTOT4Bl High GLOBAL

Girls 13 -.20 (5.02) 9.38 (2.10) .51 (2.18)

BOYS 24 -.99 (5.63) 8.83 (2.73) -.06 (36)

Low TOMTOT4Bl Low GLOBAL

Girls 8 2-44, (4.67) 9.88 (2.53) 1.02 (2.62)

BOYS 18 -4.25, (6.14) 8.06 (3.62) -36 (.93)

For all means, the higher the scores, the greater the social-relational ability. Means with matching subscripts differ at < -05, Tukey HSD Test. Mean scores shown with standard deviations in parentheses. TOMTOT4B group = Total social understanding score, scores derived from social understanding stories. High and low social understanding groups based on total social understanding story score. High = total score 129. Low = total score < 29. GLOBAL group = Total perceived global self-worth score from Self-perception hofile for Children (SPPC, Harter, 1985). High = total perceived global self-worth score 120. Low = total perceived global self-worth score < 20. aPeer total score obtained from summed pscores of Social Competence Nomination Form and Peer Likability. bTeacher total ratings of social acceptance, from SPPC. cTotal social relations score obtained from summed Z-scores of total peer and teacher social ratings. who belonged to the group of low social understandingflow perceived global self-worth were rated as the least social. These findings suggest that perceived global self-worth may play a larger role in the relation between social understanding and social relations for girls than for boys. That is, despite the fact that girls had a sophisticated social understanding, if they had low perceived global self-worth, they were perceived by their peers and teacher as the least social. Such an effect was not found for boys. Thus, these results provide marginal support for a moderating role of perceived global self-worth between social understanding and social relations among girls. In addition, although there was no sigmficant main effect for gender or an interaction effect, inspection of the means of the teacher ratings of behaviour-conduct showed that girls in the low social understandinflow perceived global self-worth group received the highest teacher behavior-conduct ratings = 11.13), whereas the same group among the boys received the lowest ratings (M = 8.33). These results show that teachers rated girls with low social understanding and low perceived global self-worth as the most well behaved, whereas the boys placed in the same group were rated by their teachers as the least well behaved. Furthermore, girls within the low social understandingmigh perceived global self-worth group were rated as the least well-behaved (M = 10.46) whereas boys within this same group were rated as the most well- behaved. Overall, the results from the three-way ANOVAs suggest that teachers' rating of behaviour- conduct may be influenced by gender, children's perceived global self worth, and their social understanding ability. As in the analyses used to test the mediation model, to control for the possible effect of vocabulary ability, parallel analyses (three-way ANCOVAs) were performed on all of the dependent variables with the Gates-MacGinitie Vocabulary score as the covariate. Results fiom the ANCOVAs showed that once the effect of vocabulary was partialled out, the marginal three-way interaction effects became nonsignificant, however, the pattern of means among the scores of the four groups remained the same. 6.4. Summary In general, the results mentioned in the above analyses provide some support for the prediction that complex interactions would exist among social understanding, self-concept and social relations, especially when gender is taken into account. Although the results from the regression analyses did not provide any evidence for a mediating role of perceived behaviour- conduct, there was some evidence to show that perceived global self-worth may have a moderating effect on the relation between social understanding and peerlteacher ratings of social relations among girls. Such results support recent efforts to distinguish between the role of mediation and moderation among relations that involve self-concept (e-g., Boivin & Hymel, 1997; Haines & Bartels, 1997). Furthermore, the reverse gender patterns surrounding teachers' and peers' perceptions of children with high or low social understanding and/or perceived global self-worth, emphasize the need to take gender into account when studying self-concept's role as either a mediator or a moderator between social understanding and peerlteacher ratings of social-relational ability. CHAPTERVII Results: Further Analyses of Social Stories and Self-Understanding Interview 7.1. Overview In addition to the quantitative analyses described in the past two chapters, analyses were performed on the categorical data In particular, the responses to the social stories and the self- understanding interview were examined by means of Chi-square analyses and descriptive, content analyses. Such interpretive analyses were performed for two main reasons: 1) to provide support for the quantitative analyses and 2) to provide a richer or "thicker" description of the preadolescent mind.

7.2. Social Understanding: Stories In further support of the ANOVA results, Chi-square analyses were performed on each of the questions for the social stories among. Results showed significant or marginal gender effects among three out of the four subscales (conceptual role-taking, empathetic sensitivity and person perception). The scores of girls signif~cantlyor marginally remained independent of the boys' scores on all of the subscales. Girls' responses were more likely to be categorized as psychological or integrated psychological whereas the boys' responses were more likely to be categorized as behaviourallsituational. The justification responses to the question of "Why do you think that NancyMargie are moving in the direction of the new girl?/Why do you think Kenny chose Tom to be on the team?" were analyzed as to whether or not they referred to a) oneself or personal experience, b) directly to the story, or c) convention (e.g., "That usually happens at school"). Frequency analyses showed that for both stories, the majority of the children responded by referring directly to the story (77% and 68% for the NancyIMargie and KennyMark stories respectively). Responses that contained references to convention were mentioned by 20% of the children for the NancyMargie story and 17%of the children for the KennyMark story. Finally, only 9% of the children referred to personal experience when answering this question for the KennyMark story and 13% of children for the NancyMargie story. Although Chi-square analyses did not reveal any significant gender effects among the responses, examination of the response frequencies (see Table 20) revealed that for both stories, as compared to boys, a greater number of girls referred to convention and personal experience whereas this was not found for direct references to the text.

7.2.1. Emotional Valence of Social Stow Res~onses To examine the emotional valence of the responses to the social story questions, the means of the proportions of positive, negative and neutral responses for the individual stories and their total were explored. To assess for gender differences in the proportional distribution of positive, negative and neutral reponses, 1-tests were performed, showing no significant gender effects. However, within each gender group, paired &-testsrevealed that girls' responses were significantly more positive @(1,63)= 2.03, g < -05) for the KennyMark story = -62, SD = .29) as compared to the NancyMargie story (M = .54, SD = -27). Moreover, girls' responses were si,onificantly more negative @(1,63)= -2.44,~< .05) for the NancyMargie story (M = .37, SD =

.26) as compared to the KennyMark story (M = .27, SD = -28). In contrast, although boys' proportions of negative and positive responses did not differ between two stories, the proportion of neutral responses was signficantly higher @(1,63)= -2.10,~c .05) for the NancyMargie story

(M = -15, SD = -20) than for the KennyMark story (M = -08, SD = -12). These results suggest that ,+ls' responses were more negative for the story containing characters of the same gender. Such findings are in conflict with Powlishta's (1995) study that showed a gender bias in girls and boys' responses to social narratives with girls and boys responding more positively toward stories with same-gender protagonists. In addition, Pearson bivariate correlations were performed between the total proportion of positive, negative and neutral responses and the main variables. Although no significant correlations were found among the entire sample, si,pScant conelations were found between the piGFpiGG Gf Gcg& pc"-6 Fsidyc sxi& s.6ics z6 &he sccies ct~&escci& Table 20 Freauencv Percentages of Positive Responses to Ouestion 3b of Social Understanding: Stories

Response Category Story Self Convention

Story 1 (NancyLMargie) Girls 63 19 271, BOYS 72 13 14b

Story 2 (KennyLMark) Girls 77 11 23a Boys 77 6 11,

Note. Question 3b refers to the justification of the response to Question 3a. For example, in the Nancy/Margie story, Question 3b would follow: "Why do you think that/How do you know that (Nancy and Margie moved off together in the direction of the new girl)?" Data obtained from social understanding stories. For both girls and boys, n=64. Percentages with matching subscripts differ at < .lo. McNemar's Test. understanding measures among the separate correlation matrices for girls and boys. Results indicated that, for girls, a significant positive correlation was found between the proportion of negative responses to the social stories and the total Mental State Verbs Task score (r(62) = .3 1, g < .05); among boys this correlation was nonsignificant (r(62) = -.12, ns) with Fischer Z transformations showing that the two correlation coefficients differed significantly (g = 2.4, g c .05). Correlations between the proportional total scores of emotional valence of the social story responses and the totd scores of the social stories (TOMTOT4B) showed that for boys, a significant positive correlation was found between the proportion of negative responses and the total social story score (L(62) = .29, g c .05) whereas this relation was not significant among the girls (r(62) = -04, ns). Repeat analyses with the Gates-MacGinitie scores as a covariate showed that the correlations remained significant even with the effects of vocabulary ability partialled out. Taken together, these results demonstrate that as &Is and boys scored higher on the mental state verbs task and the social stories, they reported more negative reponses to the social stories.

7.3. Self-understanding Interview The justifications of the self-statements obtained from the SPPC were categorized as 1) self-appraisal (e-g, "I do well in school because I try hard"), 2) other appraisals (e.g., "I do well in school because my parents tell me I do"), 3) self and other appraisals such as social comparisons (e.g., I do well in school because I'm smarter than everyone else.") and 4) tangential (e.g., " I don't know."). Frequency analyses showed that overall, self-appraisals were the most common, accounting for 73% of the total number of responses. The second most common response were the other-appraisals, occurring 20% of the time with only 5% of the responses containing appraisals that referred to both self and other. Only 2% of the total number of responses were tangential. Cross-tabulations were performed on the total number of self-understanding responses by gender and showed that overall, a larger percent of the self-other appraisals were made by girls (72%) as compared to boys (28%). Also, a larger percentage of tangential responses were zbt&ed fie= e;s (7 1 T?) CcEpzred tz 29%firEa the gk!s!s. To assess the differences in responses between children who reported overall high or low perceived global self-worth, the two groups (high and low self-worth) were analyzed according to their responses and results showed that high self-worth group accounted for 63% of the total self- appraisal responses (girls and boys combined). Girls in the low self-worth group had the lowest number of self-appraisals (16%), whereas boys in the high self worth group had the greatest number of total self appraisals (3 1%). In contrast, the highest number of "other appraisals" occurred within the high self-worth girls (3 1% of total "other" responses) whereas the lowest number of other appraisals (21%) occurred among the high self-worth boys. To examine the breakdown of the self-understanding responses more closely, Chi-square analyses were performed on the responses to each of the selfconcept categories. Results showed that for the behaviour-conduct question, "What's the main reason for why you actldon't act the way you know you are supposed to?", significantly more girls &I = 22) than boys (Q = 12) responded with an "other" or "seWother" appraisal x2(2, N = 34) = 6.26, p c .05. It was found that compared to low self-worth girls (11 = 14), a significantly greater number of high self-worth girls &I = 29) responded with self appraisals to the question, "What's the main reason for you to thinkldon't think that you are good-looking? (physical appearance), x2(2, N = 43) = 11.00, p c

.01. Likewise, results showed that compared to low self-worth boys (Q = 13), a significantly greater number of high self-worth boys (n = 28) responded with self appraisals to the question, "What's the main reason for you to think that you have many friends?" (social acceptance), x2(2, N = 41) = 7.24, p < .05.

To investigate the differences between the high and low social understanding groups' self- understanding categorical responses, the frequency of the responses were analyzed for the two social understanding groups as well as gender. Frequency distribution of the responses did not differ for the low and high social understanding groups. Multiple response analyses showed that overall, self-appraisals accounted for the highest percentage of responses (73%), followed by other responses (20%), self-other responses (5%) and lastly, tangential responses (2%). However, once the sample was divided according to gender, Table 21 shows that the response distribution between the high and low social understanding groups differed between the girls and boys. The highest number of self-appraisals were found among low social understanding boys (34% of total self-appraisals), whereas the lowest number was found among the low social understanding girls (16% of total self-appraisals). Table 22 shows that within each gender group, the distribution of girls' self-appraisals between the low and high social understanding groups was 34% and 66% respectively, compared to the reverse distribution found among the boys (66% and 34%) (the individual cells between the gender were si,onificantly different from one another - McNemar's Test). Table 21 illustrates a simiIar reversed gender pattem for the "other" appraisals, where the highest number of other-appraisals occurs among the high social understanding girls (37% of total other-appraisals). In contrast, the lowest number of other-appraisals occurs among the high social understanding boys (18%). Within each gender, the high/low distribution of "other" responses for the girls is 71% and 29% respectively, compared to 38% and 62% for the boys (see Table 22). This reversed pattem of self and other appraisals found among high and low social understanding groups of girls and boys suggests that girls who were placed in the high social understanding group were more likely to respond with other-appraisals whereas boys placed in the low social understanding were more likely to respond with self-appraisals. Also, between-gender comparisons showed that high social understanding boys had significantly @ < .001, McNemar's

Test) fewer "other" appraisals (38% of total boys' "other" appraisals ) whereas high social understanding girls accounted for 71% of the total number of girls' "other" appraisals. Overall, results from the response analyses of the self-understanding question demonstrate that children within the high self-worth group are more likely to refer to their own sense of self when asked to justify their own self-statements. Results also show that girls are more likely than boys to refer to others when asked to justify self-statements. Table 2 1 Frequencv Percentages of Total Self-Understandin? Categorical Res~onsesfor High and Low Social Understanding Grouus

Response Category -n Self Other Selflother Tangential

Girls @ = 64) High Social Understanding 41 32 37 53 14 Low Social Understanding 23 16 16 19 14

Boys @ = 64) High Social Understanding 36 18 18 6 22 Low Social Understanding 28 34 29 22 50

Note. High and low social understanding groups based on total social understanding story score. High = total score 2 29. Low = total score c 29. Total number of responses for self = 565, other = 153, selflother = 36, tangential = 14. Table 22 Within-Gender Frequencv Percentages of Total Self-understanding Categorical Res~onsesfor High and Low Social Understanding Grou~s

Response Category -n Self Other Selmther Tangential

Girls (Ir = 64)t High Social Understanding 41 66 7Ia 761, 50

Low Social Understanding 23 34 29 24b 50

Boys @ = 64)$ High Social Understanding 36 34 38, 2% 30

Low Social Understanding 28 66 62 80, 70

------Note. High and low social understanding groups based on total social understanding story score. High = total score 2 29. Low = total score c 29. Frequencies with matching subscripts significantly differ at 2 < .001, McNemar7sTest. t Total number of responses for self = 272, other = 82, selflother = 26, tangential = 4.

Total number of responses for self = 293, other =71, selflother = 10, tangential = 10. 144

7.3.1. Analyses of Perceived Global Self-worth Justifications Responses to the self-understanding question that asked children to provide reasons for why they reported thst they were happy being the way that they were or that they wished they were different were analyzed according to the frequency of categories. The most common response (51% of the children, 34 boys, 3 1 girls) was that the child did not have a main reason as to why they were happylwished they were different (eg, "I don't know why I am happy - I just am"; "I just wish that I was someone else"). The second most common response (25% of the children 20 girls, 12 boys) referred to a social aspect of their lives (e.g., "Because I have lots of friends", "I wish that I had more friends"). Only 5% of the children (3 girls, 3 boys) mentioned that they were happylunhappy because of their family with the remaining 12% of the children mentioning multiple reasons (e.g., "Because I have lots of friends, good at school, do well in sports, etc."). Chi- square analysis showed that girls and boys did not differ in their responses x2(2,N = 112) =

5.13(5), p = .40).

7.3.2. Analvses of Self-Esteem Definitions Frequency analyses of the responses to the question of "What does self-esteem mean to you?" showed that a greater number of responses (54%, 30 girls, 30 boys) referred to an aspect of the self from an objective or a secondhird person perspective (e-g., "It's how you feel about yourself."; "It's what people think they can do."). Twenty percent of the responses (13 girls, 9 boys) mentioned aspects of self and other (e.g., "It's how you feel about yourself and what other people think about you."), whereas only 10% of the responses (8 girls, 3 boys) referred to personal experience. That is, only 10% of the responses included a subjective or first-person definition of self-esteem (e.g., "I know that I have self-esteem because I like the way I look"). Finally, 17% of the responses (6 girls, 13 boys) were tangential (e.g., "I don't know"; "I've never heard of that before."). Also, a marginally significant gender effect was found x2(2,N = 112) =

7.32, g < -07) showing that girls and boys responses to the self-esteem questions were marginally independent of one another. Specifically, more girls than boys included personal experiences in their definitions whereas more boys than girls responded with "I don't know." To assess if children differed in their social and self-understanding according to how they responded to the self-esteem question, one-way ANOVAs were conducted with the four categories of self-esteem responses acting as four independent groups with each main variable as the dependent measure. Results showed that the four groups (objective self, self and other, privatelsubjective self and tangential) did not sigmficantly differ with respect to their scores on the Gates-MacGinitie Vocabulary Test, on the Mental States Verbs Task or on the teacher ratings or the SPPC subscales. As illustrated in Table 23, significant differences were found among the groups with respect to peer-ratings of social-competence and popularity. Post-hoc Tukey-HSD tests found that for social competence ratings, the self-other group scored significantly higher a= .3 1) than the tangential group @ = -.61). With regard to peer likability, the objective @ = 3.02) and subjective-self (AJ = 3.17) group were rated as significantly more popular than the tangential group

(M = 2.46). Also, the tangential group received the lowest peer ratings in both social competence and popularity. Table 23 also shows siaMcant between-group differences between the social story and self understanding total scores. Mean scores show that children who included references to personal experience in their self-esteem definition scored the highest in both social stories (M =

32-18) and self-understanding (M = 13.9 1). Post-hoc tests (Tukey HSD and Scheffc's), showed that these children scored signif~cantlyhigher on the social stories and self-understanding interview than children in the tangential groups (24.1 1 and 11.47 respectively). Overall, results from these ANOVAs show that children who differ in their perceptions of self-esteem may also differ in how they understand ambiguous social situations and their own self-statements. Also, the fact that the different groups of self-esteem definers differed in similar patterns among both the social stories and the self-understanding interview supports the strong association found between the total social story score with the total self-understanding interview. Table 23 Means. Standard Deviations. and Effects of Self-Esteem Definitions for Peer Social RatingaSocial and Self-Understanding and Perceived Global Self-worth

Group

Variable Objective Self SeWOther Subjective Self Tangential

@ = 60) @ = 22) (Il=11) (g =19)

SCNF -02 -31, -22 -.61a 3.33* (1.06) (m (-61) (1 -08) PRA -l%d -00 -3led -.66, 3.58* (.96) (1-w (.46) (1.04) TOMTOT4B 28-37,, 31-86,, 32. ISgh 24.1 1, 3.78**

(454) (4.92) (5.02) (7-00) SUTOT 11.47 13.27 13.91, 11.47, 9.59*** (2.38) (2-0s) (2.0) (2.78) GLOBAL 19.22 20.55 20.18 19.47 -91 (3-23) (4.18) (2.93) (3.52)

&& Groups based on categorical codes of self-esteem definitions. F represents results from ANCOVA, partialling out the effects of Gates-MacGinitie Vocabulary Test (1994) total score. Mean scores shown with standard deviations in parentheses. Means with matching subscripts significantly differ at Q < .05, Tukey HSD. The higher the scores, the greater the understanding and competence. Variable Definition Possible range SCNF - Total score from adapted version of Social Competence -2.23 - 1.80 Nomination Form. PRA - Peer acceptance or peer likability score. -2.64 - 1.84 TOMTOT4B = Total social understanding story score. 0-42 SUTOT - Total self-understanding score from self-understanding interview. 0- 18 GLOBAL - Total perceived global self-worth score from the Self-perception Profile for Children, (SPPC, Harter 1985). 0 - 24 a For both of the peer-rating measures, scores were standardized within each classroom. *** p < .0001. **gc .01. * Q < -05. Smq With regard to the emotional valence of the social story responses, within each gender group, girls' responses were significantly more negative for the story containing characters of the same gender. Also, the proportion of negative responses to the social stories was positively related to mental state verb task score in girls and the total social understanding score in boys. Concerning content of the responses, children were more likely to refer to the text than to convention or personal experience. Girls more often than boys referred to convention and personal experience. Findings from the self-understanding interview showed that the majority of children atttributed their self-perceptions according to their own sense of self or agency. However, girls were more likely than boys to attribute their self-perceptions to significant others (parents, friends, siblings). Further analyses showed that girl with relatively sophisticated social understanding were more likely to use other-appraisals, whereas boys with relatively low social understanding were more likely to use self-appraisals. Finally, results from the self-esteem question found that the majority of children gave an objective definition of self-esteem. However, compared to boys, girls were found to offer more subjective definitions of self-esteem (i-e., reference to personal experience), and fewer "I don't know" responses. Furthermore, analyses of variance showed that children who defined self-esteem based on their personal experience obtained the highest social and self-understanding scores. This finding suggests a link may exist between the ability to reflect on one's own mental life and the ability to understand the mental life of others. That is, the more children reflect upon their own sense of self and private experiences, the more they may come to understanding the private worlds of others. CHAPTERW Discussion 8.1. Introduction The main purpose of this study was to explore the relations among social understanding, self-concept and social relations in preadolescents. More specifically, the present study addressed the following questions: 1) Do relations exist among social understanding, self-concept and social competence ratings (peer and teacher)?; 2) What intervening role does self-concept play in the relation between social understanding and social competence (mediator or moderator) and how does this role differ among the various self-concept dimensions? and 3) Zow do individual scores and correlational patterns differ according to gender? To check for differences in the entire sample and within each gender group, data analysis was performed on the entire sample as one undifferentiated group and also between and within each gender group. To answer the research questions, the discussion is presented in three sections: 1) it is discussed whether or not the hypotheses were supported by the findings, 2) based on a reciprocal-constructivist model of inquiry, explanations are offered for both the expected and unexpected findings and 3), the pedagogical and curricular implications of the study are addressed.

8.2. Support for the Hvuotheses 8.2.1. Social Understanding. Self-Conce~tand Social Relations Based on social-constructivist theory and the assumption that folk or belief-desire psychology provides the infixstructure for both social understanding, self-conception and social interaction (Perner, 1988; Wellman, 1990), the general hypothesis for the entire sample = 128) was that individual differences in preadolescents' ability to understand the thoughts and emotions of others would be related to their self-concept (both self-perceptions and self-understanding), and social relations with their peers. This hypothesis was only partially supported, with correlational analyses indicating, in general, that social understanding was weakly associated to self-concept but more strongly related to social relations. More specifically, perceived behavioural conduct was the only aspect of self-concept si,onificantly related to social understanding. However, a robust positive relation was found between social understanding and self understanding, or the cognitive aspect of self-concept. This finding supports past research that has found positive connections between self and other understanding (Hatcher et al., 1990; Homer & Astington, 1995; Pratt & Bryant, 1988; Wimmer & Hartl, 1991). In contrast to the dimensions of self-concept, strong associations were found between social understanding and peer ratings of social competence, and significant but somewhat weaker relations between social understanding and peer likability. This finding supports previous work that has suggested a link exists between psychological understanding and social behaviour (e-g., Astington & Jenkins, 1995; Dockett, 1997; Dunn, 1995; Werner & Cassidy, 1997). Counter to prediction, no sigdicant relations were found between social understanding and teacher ratings. Multiple regression analyses revealed that teacher ratings accounted for only 5% of the total variance in the social understanding scores, with social understanding marginally positively related to teacher ratings of academic competence and behavioural conduct. These findings are in contrast to recent studies that have found strong positive relations to exist between teacher ratings of social competence and young children's ability to understand others' mind and emotions (e.g., Astington & Pelletier, 1997; Lalonde & Chandler, 1995). Intervening role of self-conce~tbetween social understanding and social relations. Based on the theoretical claims and empirical work presented in Chapter 2, it was predicted that self- concept would somehow intervene (either mediate or moderate) in the link between TOMand social relations. Although this prediction is more specific than the general correlational hypothesis, due to the conceptual complexity and exploratory nature of the concepts assessed in the present study, it was not hypothesized that one model (e-g., mediation) would prevail over another. Overall, results fiom the multiple regression analyses and ANOVAs did not provide any evidence for a mediating role of self-concept, but there was some evidence to show that perceived global self worth may have a moderating effect on the relation between social understanding and peer and iaa-har -&one r\C em;-1 r..lo&nnol oh;lhr Covwm-ato nenclnr 9n~lxrencchnxrrocl +hot omnnn &rlc tho -.A". -.A". AULYI6CI "A ~WA...-*YAYUYYIY YYLY., . YIp..Y... b"YYI* UIYJYIY YYV .. YY LYUC UIII"Y(j bYALI, UIY 150 interaction term between perceived global self-worth and social understanding accounted for a significant amount of the variance (13%) found in the peer and teacher social-relations ratings. Such results are in line with social-constructivist theories of development that suggest the self- concept plays a crucial role in how humans come to make meaning about social situations and in how humans interact with one another (e.g., Bruner, 1990; Dunn, 1988; Hinde, 1978). Thus, the present findings provide support for recent studies that have attempted to determine the role of self- concept among social reasoning and relations (e.g., Boivin & Hymel, 1997; Haines & Bartels, 1997).

8.2.2. Gender Differences Based on the suggestions that separate data analyses for girls and boys may reveal results that are "hidden" within a larger sample of mixed gender (e.g., Tavris, 1992; Unger & Crawford, 1992; Tavris, 1992), parallel analyses were conducted on the two samples of girls and boys. In general, this approach appeared to be fruitful in that it produced some interesting signilkant results that did not appear in the initial data analysis of the undifferentiated group. Although it was not predicted, results indicated that independent of vocabulary ability, girls performed significantly higher on the social and self understanding measures (social stories and self-understanding interview, respectively) than the boys. These results were fairly robust, remaining significant even after the effect of vocabulary ability was controlled for. Such results support Hatcher et al.'s (1990) findings that girls scored higher than boys on both social and self understanding tasks across grades four through twelve. Further gender analyses revealed that girls scored significantly higher than boys on both self-perceived and teacher ratings of behavioural conduct. Con,gruent with previous research on self-perceptions in preadolescents (Harter, 1985), results from the present study suggest that girls perceive themselves and are perceived by their teachers as more well-behaved or more likely to conform to conventional behaviour than boys. Interestingly, the self-concept results obtained in concept that shows girls perceive themselves more negatively than boys and are more likely to suffer from low self-esteem and related self-concept disorders than boys (American Association of

University of Women, AAUW, 1990; Brown & Gilligan, 1992; Canadian Teachers' Federation, CTF, 1991; Pipher, 1994; see Silverstein & Perlick, 1995). The only self-concept subscale that provided some support for low self-concept in girls is the physical appearance subscale. Results indicated that boys scored significantly higher than girls on physical appearance, suggesting that boys perceived themselves as more physically attractive and are happier with their physical appearance than girls. This result provides evidence for a growing area of research on body image and self-concept in preadolescence that shows girls are more concerned, negative and sensitive about their body weight and physical appearance than boys (e.g., Edwards, 1993; Furnharn & Radley, 1989; Hesse-Biber, 1996). Furthermore, as compared to boys, girls were rated as more socially competent and popular or likable by their peers. This result supports related studies that find girls generally receive higher social competence ratings than boys (e-g., Matthews & Keating, 1995). Interestingly, no gender- related differences were found on the metalinDouistic measure and the vocabulary task. These results are consistent with recent research which has found no gender-related differences to exist among metacognitive language abilities (Astington & Pelletier, 1997; Matthews & Keating, 1995). However, the finding that girls did not score significantly higher than boys on the general vocabulary task contradicts past research that has shown girls to score significantly higher than boys on vocabulary and language ability (e-g., Halpern, 1992). However, the possibility of gender-related differences in verbal and language ability has been an ongoing debate among educators and psychologists, and many would argue that the real difference lies in the methodological approach and sociocultural influences (e.g., see Unger & Crawford, 1992). Thus, researchers need to remain cautious when interpreting gender-related differences and implications of the present gender-related findings will be discussed further on in the chapter. The hypothesis that correlational patterns among the three variables would differ for girls scores would be associated with lower levels of self-concept and social-relational ability received partial support from the finding that perceived global self-worth significantly moderated the relation between social understanding and social relations only among the sample of girls. Moreover, perceived global self-worth remained a sigmficant moderator even with general vocabulary ability partialled out. Following analyses for testing self-concept as a moderator, examination of the means of the four groups of children (high and low levels of perceived global self-worth and social understanding) revealed that it was the girls in the high social understanding but low self-worth group who received the lowest social relations ratings fiom both peers and teachers. In contrast, among boys, results were more in line with the mediation model. Athough the mediation model was not supported statistically, correlational analyses showed a positive relation between social understanding and self-concept (perceived behavioural conduct), and between social understanding and peer social competence ratings. Thus, based on these findings, it appears that sophisticated social understanding ability is not necessarily related to high social relations ratings among girls and that this association is somewhat more dependent on perceived global self-worth than for boys. That is, although girls may possess a sophisticated social ability to understand mental states in others, if they also have a low or negative sense of self-worth, they may be perceived as less socially competent and popular by their teachers and peers. 8.2.3. Additional Irn~ortantFindines In addition to investigating the connections among social understanding, self-concept and social relations, the present study aimed to explore the various aspects of social understanding and to create means to assess this mentalizing ability in preadolescents. Given that the ability to score on the social stories and the mental state verb task derives from the common understanding that one social event may be interpreted in multiple ways, a strong relation was expected to exist between the two tasks. Counter to prediction, a weak positive association was found between total scores obtained from the mental state verb understanding and the social stories. This weak relation does

-A*, ..,A ,, ,,,, & ,,,,,, L .LA&,L ,,,., ,,,,:&L., 1:,1- La------):-I --->---*--a:-- ^- a ----A- 3 UUL DUY~LLLWUL LGJCQLGU uaLDUVWJ a ~DIUVG lum UCLW~~Uauucu uuuc~a~illlugmu ucuw 153 state language use (e-g., Astington & Pelletier, 1997; Schwanenflugel et al., 1996; Siddiqui, 1995, Tager-Flusberg et al., 1997). In contrast with Flavell's (1992) findings that found perspective-taking related to mental state and cognitive verb knowledge, results from the present study showed that empathetic sensitivity was the only social story subscale that related to mental state verb understanding. Thus, results from the present study support the view that social understanding is a multidimensional phenomenon where only specific dimensions such as emotional understanding may be related to metacognitive language ability. Alternatively, the present study's finding of lack of a correlation between the social story scores and the mental state verb understanding is in line with a recent study that found no relation between false belief tasks and use of mental state terms in a narrative (Charman & Shmueli-Goetz, in press). Thus, as noted by Fox (1991) and Charman and Shmueli-Goetz, narrative tasks such as the social story tasks used in the present study could be tapping into other areas of metalinguistic knowledge other than the understanding of mental state verbs. Interestingly, separate correlational analyses for gender showed that the relation between social understanding and mental state verb understanding was significant among the girls, whereas this relation was not significant among the boys. Such results suggest that the social story and mental state verb task were assessing different concepts of mental state knowledge. Finally, results from the present study affirmed the link between vocabulary ability and social understanding and mental state verb knowledge. Consistent with previous research (e.g.,

Astington & Olson, 1995; Booth & Hall, 1995), both mental state verb and social understanding tasks were associated with general vocabulary ability. Although no specific prediction was made, another aim of this study was to investigate the different aspects of self-concept including the assessment of both the affective (self-perceived competence) and cognitive aspect (self-understanding) of the self. Consistent with the view that self-concept is multidimensional and complex (e.g., Bruner & Kalrnar, 1997; Harter, 1993; Markus & Nurius, 1986), scores on the self-concept scales were not related to scores obtained

An- tha anlf-..nAntrtanA;nm ktnrr4av.r A 1thn.r.rh th;e In-b rrC-la&-- nr..vlA La nv-l&-eA :- +,-r ZIV- UIY OWY -UWILILLUIVIII~ -LWA VIYVV. SALIIVUSAACLL.~ Auun VA AWAUUVAA wvcuu a,u UA~AUUAUU u LULIIW of methodology (i.e., self-report vs. interview), the strong psychometric properties of Harter's (1985) SPPC, and the strong reliability and internal consistency results obtained for the present study's self-understanding interview suggest an alternative interpretation. Thus, perhaps the lack of relation between the two self-concept measures suggests a conceptual difference; that is, the two tasks assess different aspects of the self. Another related finding wsthat vocabulary ability was strongly related to the social stories and mental state verb task but was only weakly related to the self-measures. In particular, the only significant findings between vocabulary ability and self measures were found among boys. Results showed that boys who scored higher on the general vocabulary task perceived themselves as less athletically competent but more well-behaved and they reported a greater self- understanding. In general, these results suggest that irrespective of gender, language plays a significant role in the ability to understand mental states in others, whereas the relation between language and self may be partly dependent upon gender. That is, general language ability may play a greater role in boys' self-perceptions and self-understanding as compared to girls. Regarding sibling effects, in general, results showed that children's social understanding and general vocabulary ability was related to the number and gender of their siblings. More specifically, results demonstrated that among girls, their ability to understand the perspectives of others was negatively related to their number of sisters. Among boys, empathetic sensitivity was negatively linked to their number of older siblings. A further fmding for boys showed that a negative relation existed between their general vocabulary ability and their number of older sisters.

8.2.4. Significant Additional Findings (Categorical Level Data) To provide a richer or "thicker" description of how grade six children interpret social situations and view themselves, some of the children's more open-ended answers and justifications were explored. Analyses of the social story responses revealed that although girls' and boys' responses did not differ in emotional valence, within each gender group, girls' responses were signifcantly more negative for the story containing characters of the same gender. Such findings zv pzfiic: +t& pAy+~sy=rCh +: hsshcwwz 2 sGe-gezder bLs c~&yz's responses to short narratives (Powlishta, 1995). Also, the proportion of negative responses to the social stories was positively related to mental state verb task score in girls and the total social understanding score in boys. Such results suggest that children who had a sophisticated ability to understand others also responded more negatively, and hint at the possibility that sophisticated social understanding may not necessarily be solely related to positive attributes. Results from the analysis of the justification to the perspective-taking question (e.g., "Why do you think that NancyMargie are moving in direction of the new girl) showed that the majority of children (over 70%) included direct references to the text, whereas only about 115th of the children included references to convention and only about 1110th of the children referred to personal experience. Gender analysis showed that girls were more likely than boys to refer to convention and to personal experience. These results provide some support for the claims that as @ls approach preadolescence, compared to boys, they are more likely to draw from subjective or personal experience in response to social scenarios (e-g., Honess, 1981; Mansfield & Clinchy, 1997). Response analyses of the self-understanding interview revealed that the majority of responses (73%) included self-appraisals, followed by other-appraisals (20%) with few children referring to both self and other (5%). In general, the findings showed that girls were more likely than boys to refer to other-appraisals in their responses. It was also found that compared to children who reported relatively low global self-worth, children who reported relatively high self- worth were more likely to refer to their sense of self when asked to just@ their self-statements. Such results are in agreement with past self-concept theorists such as Cooley (1912) and recent arguments made by Harter (1997) who propose that self-definition in terms of "other" may not be considered as psychologically healthy as defining oneself in terms of one's own sense of agency. Moreover, when divided according to their social understanding ability, results showed that girls with high social understanding are more likely to use other-appraisals. That is, &Is with a relatively high social understanding ability defined themselves in terms of what other people d.-..-L* e- -..:..I ..L-.-& 6L-- 1 uuupr aaru avvur ul~u.Iii ZGIitizt., kij~G?& :G~VSX~& iiiiG~i~~~dkg ~~1y.-. io use self-appraisals. Put differently, boys with relatively low social understanding ability defined themselves in terms of their own sense of selves, their self-judgements were not based on the perceptions of others. Such results suggest that children who have a relatively sophisticated ability to understand mental states in others differ in how they define their sense of self, which in turn may be dependent upon their gender, with girls more likely to define themselves in terms of "other" whereas boys are more likely to define themselves in terms of "self." Response analyses of children's answers to the question of "What does self-esteem mean to you?" illustrated that over 50% of children referred to an aspect of the self from an objective perspective (e.g., "it's how people feel about themselves."), whereas only twenty percent of the responses included a personal experience or a more subjective definition (e-g., "I know that I have self-esteem because I like myself.") Gender analysis showed that compared to boys, girls were more likely to offer subjective defhtions and less likely to respond with "I don't know." Once grouped according to their self-esteem definition, results from further analysis of the children's mean scores across the various measures showed that overall, children who responded with a personal or subjective definition of self-esteem scored the highest on both social and self understanding measures. In regards to teacher and peer ratings of social relations, children who included reference to both self and other received the highest ratings, whereas children who responded with an "I don't know" received the lowest ratings. In general, such results support the finding that social understanding and self understanding are associated. 8.3. Meaning: of Findings (Interval level data) 8.3.1. Relations among: Social Understanding. Self-Conce~tand Social Relations The finding that perceived behavioural conduct was the only aspect of self-concept that was significantly related to social understanding shows that children who have a relatively sophisticated ability to understand mental states in others also perceive themselves as more well behaved. That is, as children's ability to see others as psychological beings and interpret their behaviour from an intentional perspective increased, they were more likely to see themselves as following societal

....I..- nc -..--.--d-- fl---:-+a-+ ..AL La+& -..-+ -a..:..1--n-1 el....- la--#.-& ..-el -n- -,.-+ LWb3 VLbUIIVUllUVll. bVll313LbLlL WLUl WUI PLWL JUbLLU lllUlLU UbVbLUPIIIUllL IILbLULLUU UUUVlb L-bllL 157

TOM literature (Astington, 1996; Nucci & Nucci, 1982; Smetana, 1993; Turiel, 1983), the present findings suggest that as children become more advanced in social understanding and gain greater social knowledge, they become more aware of morality and convention. Thus, perhaps a more sophisticated understanding of social-moral rules and convention enable children to judge their own behavioural conduct and to answer questions such as "Is it really true or sort of true that you usually do the right thing." Further investigation of the different dimensions of social understanding as assessed by the social story subscales revealed that the strongest relation was with empathetic sensitivity. As recently argued by TOMresearchers (e-g., Astington, 1996; Chandler, 1988; Dun, 1995, 1996) this finding supports the notion that social understanding is a multidimensional phenomenon, comprised of unique but related sociako,gitive abilities. The strong positive relation between the empathy score and children's perceptions of their behavioural conduct suggests that the children who have a sophisticated ability to understand emotions in others may also be more likely to perceive themselves as well-behaved and as children who follow the rules. This finding indirectly supports empathy and emotional sensitivity literature that claims the ability to understand emotions in others is a prerequisite for more sophisticated social-moral understanding (e-g., Coles, 1997). Similarly, perceived behavioural conduct was the only subscale significantly associated with self-understanding. In general, the lack of relation between the self-concept (SPPC) and the self-understanding measure supports the view that the concept of self is not only multidimensional in regards to content (e.g., Harter, 1993), but also has qualitative differences includkg affective and cognitive components (Damon & Hart, 1988). Thu, the more cognitive aspect of the self concept (assessed in the present study by the self-understanding interview) may not necessarily be related to feelings of self-worth (assessed in the present study by the SPPC). For example, in the present study, a child could have scored high on the self-understanding interview (i.e., possess a sophisticated or complex self-theory) but may have reported negative feelings about herself in particular areas (i-e., scored low on the SPPC). Hence, like the multidimensional notion of social

....aa.r+n-rf;r..r +ha II-~~C~I,--. ----:-+ -- :-drr+------I- -e L-. UIIUUIUULII-6,LUU WVIIC~CI~L VI 3eu uay 413~~VWLDL WL cau UU1b4LG UGLWVLA VI 3GpQlklLG UUL connected representations (e.g., Markus & Nurius, 1986). In addition, perceived behavioural conduct was found to be positively related to peer ratings of social competence. More specifically, perceived behavioural conduct shared the second highest correlation (following perceived social acceptance) with peer social competence ratings, suggesting that children who see themselves as well-behaved were also rated as socially competent by their peers. Based on the reciprocalconstructivist model of socialcognitive development, the positive relation can be explained in two ways. First, it could be that children who perceive themselves as well-behaved actually are well-behaved and thus seen as responsible, socially competent by their peers. Second, it could be that children earn a "social reputation" within the classroom and this helps shape their self-perceptions (Zarbatany, Van Brunschot, Meadows, & Pepper, 1996). Thus, children may not perceive themselves to be well-behaved until they begin to interact socially with others. Separate gender analyses indicated that the relations between perceived behavioural conduct and social and self understanding and peer ratings of social competence remained robust among the boys, whereas for the girls the relations were nonsignificant. Given that vocabulary ability related positively to perceptions of behavioural conduct for boys but not for girls, language may play a partial role in determining this gender-related difference. That is, boys who perceived themselves as well-behaved scored highly on the self and social understanding measures and were rated as socially competent as a result of their verbal sophistication. However, once the effects of language were partialled out, the relations remained sigmficant, suggesting that factors other than language may help to explain the relations. The finding that perceived behavioural conduct was related to both self and social understanding and peer rated social competence among boys but not girls, could be explained in terms of stereotypical social-role expectations. Past research has shown that traditional gender-role stereotypes perpetuate the view that girls are more likely to follow convention and passively obey authority, whereas boys have been viewed as unconventional, and encouraged to question or rebel

ono&ct thn ot~fircn.3- /Cmot--a 1002- T 1.r-a- A fl-...F,-rl 1 onm A --..a:-.-c------I- ----:~1 U~YY"~C UIY UUILUCI ynw \UIYYWIU, A// J, UU~A ~lrQLUW AWIU,112~1. AC.~VLUUI~ LV pay ~ILU~UCIIQI theories, children are subjected to societal voice and ear training that teaches them how to interact socially and to behave according to social convention (e.g., Gilligan, 1993). Gradually, children are said to internalize these traditional social-role expectations which are said to increase during

preadolescence (Hill & Lynch, 1983). Consequently, young adolescents actually begin to perceive themselves as either conventional or unconventional based on their gender stereotype (see Tavris, 1992). Findings from the present study support this view, with girls perceiving themselves as more well-behaved than boys. Thus, the fact that correlations were found between the perceived behavioural conduct and other measures for boys but not for girls could have been due to the lack of variability in the girls perceived behavioural conduct scores. That is, due to traditional gender- role expectations, the majority of girls perceived themselves as more well-behaved than boys, and this lack of variability in the scores prevented any si,onificant relations occurring.

8.3.2. Social Understanding and Mental State Verb Understanding- As already noted, based on past theoretical argument and empirical evidence that suggests the ability to understand mental states in others is related to the ability to understand actual mental state or cognitive verbs ,a strong relation between the social story measures and the mental state verb task measure was expected. That is, if children are able to reflect a sophisticated social understanding by attributing mental states to characters in socially ambiguous narratives, their ability to recognize and understand mental state verbs should also be sophisticated. However, correlational analyses revealed that such an expected relation was barely significant among the entire sample, and among the boys, the total scores of two measures remained independent of one another. To the contrary, among the girls, the two total scores were significantly related to one another but once the effects of general vocabulary ability were controlled for, such relations ceased to exist. Thus, findings from the present study do not support past literature which would expect to find a significant relation between social understanding and the metalinguistic ability of understanding mental state verbs (Fabricius et al, 1989; Siddiqui, 1995). Given the strong correlations between mental state verb understanding and general language ability, the present

r--$G srrggcsi tvleii-dgGic xVr& T& 2ci& g "=&Gary reiai& to ihe understanding of mental states and speech acts. Conceptually, the reason for the lack of relation between the social understanding story scores and the mental state verbs task could be that the social understanding stories assessed social reasoning whereas the MVT assessed linguistic abilities. As Matthews and Keating (1995) suggest, the MVT items require that the children understand the story and can distinguish between a set of related terms by choosing appropriate words for different contexts. The MVT also requires that the children must master the vocabulary needed to select the appropriate answer. In agreement with Matthews and Keating's (1995) study that showed no relations between the MVT (original form, Astington & Olson, 1990) and various social measures, results from the present study showed that preadolescents' ability to understand mental state language and choose the correct MVT response was not necessarily related to the ability to reason about mental states in others given a hypothetical ambiguous social situation. This suggests that some preadolescents may have understood the linguistic terms of mental states, but were unable to express these thoughts during the social understanding interview. Given the suggestion that higher order mental processes may involve complex interactions between emotions and thought (e-g., Bruner, 1996; Elliot & Feldman, 1990; Goleman, 1995; Keating, 1990), perhaps the socially ambiguous nature of the stories may have evoked emotional responses from the participants whereas the MVT did not, thus affecting the results. With regard to instrumentation, the MVT and the social understanding measures differed in both their construction and their administration. That is, the MVT was a pencil-and-paper multiple- choice task that the participants completed independently within the classroom. The social understanding task was a story-telling interview administered individually within a quiet room, outside of the regular classroom. Such instrumentation effects may also have influenced the results (see Chapter 10.2 for further discussion). Although it was not explicitly anticipated, a robust positive correlation was found between the total scores obtained from the social stories and the self understanding interview. This finding remained significant even with the effects of vocabulary ability partialled out, thus it suggests that a strong relation exists between the ability to understand mental states and emotions in others and the ability to understanding mental states and emotions in oneself. Put differently, the present finding that the social story responses were positively associated with the responses elicited fiom the self understanding interview suggests that the ability to articulate one's "theory of mind of other people is related to one's ability to articulate one's "theory of self'. Within the framework of theory of mind literature, this finding is consistent with the "theory-theory" explanation of social understanding that contends one's organizing principles of how people think, feel and behave stem from social-cognitive reasoning or "theory-making" capacities (e-g., Gopnik & Wellman, 1995; Wellman, 1990). According to this view, the ability to mentally represent and understand multiple perspectives operates similarly for both self and other in that it helps to create theories for both self and other. Concerning the present findings, children's theories of other people and of themselves were linked because both theories were derived from an underlying main cognitive conceptual framework. This adds to the body of evidence that supports the "theory-theory" view of psychological understanding (e-g., Gopnik & Astington, 1988; Pratt & Bryant, 1988; Wimmer & Hartl, 1991). Alternatively, the relation between the social understanding measure and self-understanding measure could also be explained in terms of the more recent socio-cultural narrative theory of mind (e.g., Astington, 1990,1996a; Bruner, 1996). According to this view, the concept of "other" and "self' are actually one and the same, both phenomena are considered processes where "self" is created through interactions with "other" (Kalmar & Bruner, 1997). Thus, our sense of self is created from our social experience, including our perceptions and understandings of others as psychological beings. Given the neverending feedback loop between self, other and society, the fact that social and other understanding were related in the present study suggests that p~eG~lesce~s'irnmrrtmAc ~f selr & +_r z% hn:czk!;r hnAe+n+ c-s'ctz+i bAhgcz each other's experience. As discussed in earlier chapters, a similar area of research known as socialcognition has also studied the relation between social and self-understanding, but more from the social- psychological view as opposed to TOMresearch that focuses on epistemological development and related metacognitive abilities (see Flavell& Miller, 1997). Similar to TOMresearch, studies in the area of socialcognitive development are contradictory when it comes to determining whether or not the understanding of self and other develop in tandem and are related. The present study's frndings provide support for other studies that have found relations between social and self understanding (e.g., Hart & Fegley, 1995; Hatcher et al., 1990). However, the present findings do not support the contention that the capacity for self-understanding and the ability to understand others is independent of each other (e.g., Damon & Hart, 1988; Hatcher & Hatcher, 1997). Such studies claim that based on modem Westem cultural belief that sociability and understanding others is a more "natural" personality trait for females, the ability to understand others may actually be easier to attain than self-understanding. Although the present study did not set out to measure whether or not self or other understanding differed in complexity, future studies need to design parallel tasks that would enable researchers easily to compare social and self understanding. In contrast to the strong positive relation found between social understanding and self- understanding, no such relation was found to exist between self-understanding and mental state verb understanding. This lack of correlation suggests that the ability to recognize and understand mental state verbs in others (as assessed by the Mental State Verbs Task) is not related to the ability to articulate one's self-theory. As previously mentioned, this lack of correlations could have been due to the methodological differences of the two tasks. Thus, a more parallel measure of understanding mental state verbs in self could have produced significant associations. 8.3.4. Social Understanding and Social Relations The lack of an association between the ability to understand ambiguous social situations and the ratings obtained from the sociometric and teacher ratings shows that peer and teachers' ratinoc nf cwial-relatinnal ahilitv mav he Aeterrninerl hv fartnrc nth~rthan nnm'c ~h;l;+.r tc ------, - --WJ " " -"..I" Ul-C J understand mental states and emotions. For example, general language ability such as vocabulary skills may play a role in one's social status and competence and the present study found such a relation. Specifically, although positive relations were found between the total social story score and the peer social competence ratings, once the effects of general vocabulary ability were controlled for, this relation remained significant among the entire sample but partial correlations performed for each gender failed to find a similar relation. Correlational analyses also showed that general vocabulary ability was positively related to peer ratings of social competence but not to girls' teacher ratings of social acceptance or to their peer ratings of likability. The ability to understand mental states verbs was also not found to be related to any of the social relation measures. In general, results from the present study show that social understanding was more strongly related to social competence ratings as compared to popularity or social status. Thus it appears that language and social-cognitive abilities are related to one's ability to interact socially with peers. In contrast, peer popularity may be less influenced by social-cognitive factors, and more dependent upon an interaction of factors such as social behaviour and physical appearance (e-g., Dodge, 1983). The finding that general vocabulary ability was positively related to boys' social status but not to the social status of girls suggests that language may play a greater role in determining boys' popularity as compared to girls. Given modem Western society's emphasis and value of physical attractiveness among women (see Unger & Crawford, 1992), it could be expected that a girl's physical appearance may play a larger role in determining her social status than her general vocabulary ability (Coie, 1990). The present results add to the ongoing debate concerning the link between social cognition and social relations, or in more general terms, between social reasoning and social action. Findings gleaned from theory of mind and socialcognitive literature can be found to support various arguments, first, that socialco,onitive ability is related to social-relational ability (e.g., Dockett, 1997; Ford, 1982; Matthews & Keating, 1995). A second contention is that social- Rubin, 1972). Finally, a third and more complex argument is that socialcognitive ability is not directly related to social-relational ability, but may be indirectly related, through other intervening factors. Although the correlations between the social understanding measure and the sociometric and teacher ratings of social relations support the second argument, a major goal of this study was to test this third possibility and the relevant findings will be discussed in the forthcoming sections. 8.3.5. Social Understanding and Teacher Ratings The lack of relation between TOMand teacher ratings suggests that children's ability to understand thoughts and emotions in others is not associated with how their teachers perceive their classroom behaviour. Although this lack of relation has been also found in similar studies that have attempted to link children's social understanding to their teachers' ratings of social competence (Astington & Pelletier, 1997; Werner & Cassidy, 1997), in such studies, this lack of correlation has been due to the lack of variability of teachers' ratings. However, in the present study, aside fiom ratings of physical appearance, all the teacher subscde scores approximated normal distribution. Methodologically, the lack of association could be explained in relation to problems of teacher ratings scales in general which will be discussed in Chapter 10. Briefly, children could have exhibited behaviours that when rated by another person or according to a different scale could have produced significant correlations. Methodological problems aside, assuming that the teacher rating scales adequately assessed children's behaviour and that teachers and peers agreed upon their ratings Wgh codations), the fact that peer but not teacher ratings were related to social understanding suggests that the connections among social understanding, self-concept and social relations are extremely complex, especially after taking into account effects of gender and language ability. In line with past research that has found teacher ratings to be associated with children's language ability (Verschueren et al., 1996), the present study found that children's vocabulary ability was significantly related to teacher ratings of both behavioural conduct, and academic competence. Subsequent correlations with vocabulary ability partialled out revealed that teacher ratings and

+ilhilrez'$ $&Z lyv4ent2a&cr $CC,t-$ x:p,- yt However, even with the effects of vocabulary ability controlled for, teacher ratings of children's behaviour were positively associated with the peer ratings of children's social competence and likability. In particular, correlational analyses of the entire sample showed that teachers' ratings of social acceptance were the most strongly related to the children's peer ratings, with teachers' ratings of academic and athletic competence rating the lowest. Separate correlations performed for each gender showed similar patterns for ratings of peer likability, but for rating of peer social competence, the lowest correlations existed with teacher ratings of behavioural conduct. Such results suggest that the link between teachers' perceptions of how well-behaved their students are is the least likely to be related to their students' peer ranking of social competence. Thus, this lack of correlation provides support for the moderation model, suggesting that the relation between social understanding and teachers' perceptions of behaviour are moderated by levels of self- esteem. Further correlational analyses comparing teacher ratings to the self-perception measures showed that the majority of teacher ratings were found to be positively related to the various self- perception subscales. This suggests that teachers and their students may agree upon their view of behaviour within the classroom, and provides evidence for teacher expectancy research where teachers' expectations of their students have been shown to have an effect on student behaviour and student self-concept (e.g., Brophy, 1983; Jones & Gerig, 1994). According to the "self- fulfilling prophesy" literature (Brophy & Good, 1970; Jussim & Eccles, 1995; Rosenthal & Jacobson, 1969), this effect is explained by claiming that teachers' expectations of students' behaviour help to shape the students' self-perceptions and thus, eventually molds the students' behaviour to fit the original expectations. In terms of the present study, teachers' perceptions of students as popular, academically competent, well-behaved, etc., may have influenced the students to perceive themselves in a similar manner. Interestingly, the teacher ratings of children's behavioural conduct were positively related were the least likely to be rated as well-behaved. Examination of the relation of children's perceived global sense of self-worth showed that for girls the link was the strongest between their overall sense of personal happiness and teacher ratings of social acceptance, whereas for boys, the strongest links occurred between perceived global self-worth and teacher ratings of academic and athletic competence. Such results support past research on teachers' stereotypic gender-role expectations that has shown teachers tend to emphasize social competence in girls and academic and athletic competence in boys (AAUW, 1990; Dusek & Joseph, 1983). To the contrary of the self-perception subscales, no relation was found to exist between teacher ratings and the children's total self-understanding score. Conceptually, this lack of relation might reflect the difference between the two constructs of competence and understanding. That is, the teacher ratings reflect children's competence within the classroom and the self-understanding score reflects the children's own understanding of self-judgements. This lack of relation could simply be a reflection of assessment tool differences (teacher ratings versus semi-structured interview). On the other hand, it could be that semi-structured interviews, as opposed to self- report measures, may provide a richer, less socially desirable description of children's self- theories, which may not be related to how their teachers perceive them as individuals. Given the methodological problems of studying self-understanding, much work remains to be done in the area of teacher ratings of classroom behaviour and their students' self-understanding. The interaction effects from the analyses of covariance of the social understanding scores showed that girls with male teachers scored the highest on the social understanding and were perceived as the most well-behaved. In contrast, boys with female teachers scored the lowest on the social understanding scores and were perceived as the least well-behaved. Perhaps the children's social understanding scores and behavioural conduct ratings were implicitly influenced by their teachers' traditional gender-role expectations. More specifically, girls with male teachers might have achieved high social understanding scores and behavioural conduct ratings because this is how their male teachers expected them to behave (based on traditional stereotypes of femininity). 167 understanding scores and behavioural conduct ratings is that female teachers expected male students to behave according to the traditional masculine stereotype. That is, female teachers may not have expected male students to be very adept at understanding the thoughts and emotions of others and to behave non-conventionally in the classroom. In general, the present study showed that the relation between teacher ratings and children's social and self understanding is much less clear than the relation between teacher ratings of students' behaviour and students' self-perceptions and peer sociometric ratings. Perhaps children's implicit theories or understanding of others and themselves are less likely to be affected by teachers' expectations as their self-perceptions and social behaviour are. Another possibility could be that children's self-perceptions and their sociometric ratings also play roles in the relation between social understanding and teacher perceptions. Thus, to help explain such a complex web of relations, a major goal of the present study was to test two specific models that were aimed to tease out specific connections between social understanding, self-concept and social relations.

8.3.6. Self-concept as Mediator or Moderator The second main question that this research set out to answer involved the investigation of whether or not self-concept played an intervening role in the link between social understanding and social relations. That is, did particular aspects of children's self-concepts either mediate or moderate the influence of their social understanding on their social relations. Regressional analyses showed that no aspects of self-concept mediated the relation between social understanding and social-relational ability. In other words, self-concept did not represent the generative mechanism through which preadolescents' ability to understand mental states in others was able to influence their social relations. In contrast, results fiom regression analyses performed on the entire sample showed that perceived global self-worth acted as moderator between social understanding and social relations. That is, out of all the self-perception subscales, perceived global self-worth or children's overall r& Gf pry.c y.Ac& Fc;-LAllklg -*=CIC G-ly 6f.&e seg*oii cepi to ,ucnce &e between social understanding and social relations. This suggests that among girls, although perceived global self-worth was not directly related to social understanding, it influenced the connection between social understanding and social relations. Interestingly, this effect was not found among the sample of boys. To examine further if gender played a role on how perceived global self-worth moderated the influence of social understanding on teacher and peer ratings, the sample was divided into four separate groups (high perceived global self-worth, high social understanding; high perceived self- worth, low social understanding; low perceived global self-worth, high social understanding; low perceived global self-worth, low social understanding). Inspection of the teacher and peer ratings showed that the within each gender, the groups of children who received the lowest social competence ratings were the girls in the high social understanding, low perceived global self-worth group and the boys in the low social understanding and low perceived global self-worth group. Such findings suggest that girls who may be at the greatest risk of being negatively perceived by their peers are those who experience negative feelings of self-worth and have a sophisticated ability to understand mental states in others whereas boys who also experience feelings of negative self-worth and have difficulty in understanding mental states in others are most at risk. Thus, it appears that a sophisticated ability to understand mental states in others, coupled with an overall negative sense of self-worth may have a negative influence on peer and teacher ratings of social relations in girls but not among boys. As a moderator then, low levels of perceived global self-worth may have hindered the positive influence social understanding may have had on social relations. That is, given the assumption that high levels of social understanding may have a positive idluence on social relations, a low sense of self-worth could have inhibited such positive effects. Such an inhibitive effect may explain why some girls who had high levels of social understanding received relatively low social ratings. Alternatively, given the assumption that high social understanding may have a negative effect on social relations, a high level of global self-worth may have acted as a buffer against the sophisticated social understanding ability, coupled with a high level of self-worth received relatively positive peer and teacher social ratings. The fact that a sophisticated ability to "read social situations" coupled with negative feelings of self-worth may have a detrimental influence on preadolescent girls' social lives, provides support for studies that have investigated TOMand its relation to socialemotional development (Dunn, 1995), and social co,pitive studies on gifted girls (e.g., Eccles, Bauman & Rotenber, 1989, see Kerr, 1994; Silverman, 1995). Both sets of studies have suggested that although commonly deemed as a "social skill", a sophisticated emotional and social understanding may not necessarily guarantee an easier social experience in school. In fact, some studies have shown that children who excel in social understanding tasks may be more sensitive or vulnerable to how others are thinking and judging them at school and consequently may experience emotional and social problems later on in life such as social rejectionhsolation, depression, etc. (e.g., Petersen et al., 1991; Silverstein & Perlick, 1995). Thus, given that a superior mentaking ability enables one to read the mind of others, it could be that children who possess a sophisticated social understanding are more likely to imagine what others are thinking and saying about them, which could lead to new, or add to already existing negative feelings of self-worth. The finding that perceived global self-worth played a larger role in the relation between social understanding and social relations among girls could be explained in terms of psycho- cultural theories of children's self-development (Belenky et al., 1986; Chodorow, 1978, Gilligan, 1982). Based on these theories, a girl's sense of self is derived from her relationships with others and feelings of psychological connectedness, whereas for a boy, his sense of self is derived from the ability to remain separate, and is derived more from his actions. Thus, given the social nature of the concepts investigated in the present study, it is not surprising that for girls, their global sense of self-worth or psychological well-being played a larger role. Similarly, a psycho-cultural perspective of self-development can also help to explain why a negative sense of self-worth in girls played a moderating role by compensating for the possible

=EfftS sOPfis;-C~t~Szis ~+e~t~cg hzyP,hs ~2 Sziza+2dcz- pu2t is, 2 170 girls experienced low global self-worth, they were perceived as relatively socially incompetent, independent of their social understanding. Given social-cultural norms that convey stereotypes of women who are more caring and able to understand others, the ability to do this task very well may actually hinder the self-development of some girls and lead to feelings of low self-worth or self- dislike. On the other hand, given society's value of traditionally masculine traits such as autonomy, exploration, egotism (Jack, 1991; Tavris, 1992), girls in the present study who scored relatively highly on the social understanding measure may have perceived this ability to understand emotions and thoughts in others as "feminine." Given society's negative view of such traditionally feminine traits, these girls may have felt devalued as members of society, which in turn may have led to feelings of low self-worth. The present results can be tentatively viewed, then, as indicating that societal expectations of traditional gender-roles creates a sense of low self-worth in girls which in turn may lead to social behaviours that are perceived as incompetent by their teachers and peers. Such findings support the psychocultural view that contends the transactional relation between culture and mind allows children to create their sense of selves through a process of personal meaning-making (Bruner, 1996; Erikson, 1953,1968; Levitt & Selman, 1996).

8.3.7. General Gender Differences The third main objective of the present study was to explore individual differences among the scores of the measures and the categorical data. As previously mentioned, individual scores were examined for any gender-related differences, while ensuring that the effects of vocabulary were partialled out. In sum, results from analyses of covariance showed that girls scored significantly higher than boys on social and self understanding, perceived behavioural conduct, teacher ratings of behavioural conduct and peer ratings of social competence and likability. Interestingly, the only score in which the boys scored higher than the girls was that of perceived physical appearance. That is, boys perceived themselves as more physically attractive than girls.

Fsy,zg~r&g the +g2ue ~en*~~s, zc &%-%f~s feAqd &-.;;-o= +A=gczcd 171 vocabulary test scores and the total score obtained from the mental state verb understanding score. Each result will be discussed in turn. The finding that girls scored signdicantly higher on the social and self understanding tasks suggests that in the present sample, girls possessed a more sophisticated ability to understand mental states in others and to articulate their self-theory than boys. Despite the rising interest in gender analyses in TOM research, the majority of studies have failed to fmd significant gender- related differences (e.g., Astington & Pelletier, 1997, Jenkins & Astington, 1995). Aside from Hughes and Dunn's (in press) study, the lack of gender-related results in TOMresearch could be explained in tenof the ages of their participants. Past TOMresearch has mainly focused on children under the age of seven or eight, whereas gender-related differences in social cognition do not usually manifest themselves until around preadolescence (Halpern, 1992). With regard to Baron-Cohen et a1.k (1997) study, the gender effect could have merely been due to superior language ability among the sample of women. Perhaps as TOMresearch continues to investigate older participants (and test for gender effects), gender-related differences will continue to be discovered. The finding that girls were more adept at social understanding than boys is consistent with findings from research gleaned from the areas of both social cognition and TOM. As previously mentioned in Chapter 2, although not labelled "TOM"research, this study's finding that girls scored higher than boys on the social understanding scores supports related studies in socialcognition that have showed girls score higher than boys in tasks that assess various aspects of psychological understanding such as conceptual perspective-taking, empathetic sensitivity, and person perception (e-g., Hatcher et al., 1990, Hoffinan, 1986; Honess, 1981; Roberts & Strayer, 1996). Regarding TOMresearch, the present study adds to the growing number of studies that reflect a superiority among females in the area of mental state attribution and mental state language (Baron-Cohen et al., 1997; Hughs & Dunn, in press). Furthermore, the present study's finding that the magnitiude of the gender differences was

+Ln.rra..+,."+ +I..- r--r-.r~LILr -^^-:L-~c. -err------L- -:-11 C a:--- a------A .,----. -U WUULUJL LLUlVll5 UlC, bU~LLUlGUb3GU3lUVlLJ 3CIVlC3 3UpPLl.3 3lllJ.Ll&i 1~~~1lg3LlULll I-&& that has shown the greatest gender differences occur among empathy tasks (e.g., Roberts & Strayer, 1996). Thus, this study supports past research that has shown girls are relatively more sophisticated than boys in recognizing and understanding emotions in others (e.g., King, Akiyama & Elling, 1996). The findings from the present study related to genderdifferences surrounding issues of self-perceptions and teacher and peer ratings of behaviour are also consistent with past research. In particular, that finding that girls perceived themselves as more well-behaved and were rated as more well-behaved by teachers and socially competent by their peers supports the majority of studies that claim similar findings (e.g., Harter, 1985; Matthews & Keating, 1995). Furthermore, the finding that boys perceived themselves as more physically attractive than girls is consistent with past studies that have found boys to be more accepting of their body image and physical features than girls (e-g., Silverstein & Perlick, 1995). Taken together, the gender-related differences found in the present study illustrate the complexity involved when one takes into account the effects of gender in social-cognitive processes. As various researchers and writers have noted (e.g., Edwards, 1993; Hoff-Summers, 1994; Maccoby, 1998), much caution needs to be taken when interpreting gender-related findings. For example, the gender-related differences found in the present study provide both evidence for and against the gender-intensification hypothesis that explains gender effects in terms of increased differential socialcultural experience during preadolescence (Galambos, Almeida, & Petersen, 1990; Hill & Lynch, 1983). In particular, the finding that girls scored higher in social understanding, perceived and teacher ratings of behavioural conduct and sociometric ratings supports the gender-intensification hypothesis claim that traditional gender-role expectations help to shape gender-related behaviours and cognitive abilities in girls and boys. From such a perspective, the present findings reflect the contention that gender differences often reflect gender-role stereotypes (e.g., Matthews & Keating, 1995). That is, due to society's value of traditional "feminine" traits such as understanding and caring about others, sociability and conventionality, rGch LG....+i c..&-a &? rsGxdLy gL-.Ls yrL ..& &..AY ...c $Y AcL- ie..chcz, p&-cz*G hi- Fcis. Hence, as children approach early adolescence, gender differences in social behaviour and thought become prevalent (see Halpem, 1992). The finding that boys perceived themselves as physically more attractive than girls also supports the gender-intensification hypothesis in that boys are socialized to focus their sense of self-worth on their achievements, whereas girls are socialized to place most of their self-value on their physical appearance and body size (e.g., Hesse-Biber, 1996). Thus, society's obsession with physical appearance in women may lead preadolescent girls to become more sensitive and perhaps self-critical and negative when it comes to judging their own appearance. However, the finding that girls and boys did not differ in their general vocabulary ability and their mental state understanding does not support the gender intensification hypothesis in that they do not support the stereotype of the "talkative girl" (Eagley & Mladinic, 1989; Hyde & Linn, 1988). According to this hypothesis, due to exposure to gender-role expectations and differential behaviour based on gender-roles of their teachers, parents and peers that promote traditional feminine traits such as girls being more talkative and understanding of others, girls would be expected to develop greater vocabulary skills and a greater command of mental state verb understanding. As an alternative explanation to the gender-in&sification hypothesis, the gender-related findings from the present study could be interpreted from a more psychocultural perspective. Such a perspective would argue that in addition to socialcultural effects, individual psychological traits and cultural factors also could play a role in the relations among social understanding, self- concept and social relations (Bruner, 1996; Harter et al., 1997; Tannen, 1994). For example, the gender-related differences may not have been due to the preadolescents' actual biological sex, but rather due to whether or not the children perceived themselves as traditionally "feminine" or "masculine" incorporating such stereotypical traits of communion and agency (Eagley & Mladinic, 1989). Thus, the actual gender-related differences found in the present study could have been due to gender-role perceptions as opposed to whether or not the children were female or male. Further underlying psychological and cultural differences such as nonverbal intelligence or spatial-visual &Z~,V& s+~c-~czcF~c siz+s Zlc ccdd bz.rc &?-it& +Ah gczdcr-xlctd FGkgS. Fucpzsc~t 174

results can be viewed, then, as indicating that controlling for general vocabulary ability may not be sufficient to tease out the gender-related differences among such complex concepts as social and self understanding, self concept and social relations.

8.3.8. Siblinns Although the present study failed to find any significant differences with regards to social understanding and the number of siblings in the entire sample, separate correlational analyses of the gender groups illustrates the need to take gender into account when investigating the influence of family structure on socialcognitive abilities. An increase in number of siblings (sisters for girls, older siblings for boys) was negatively related to social understanding ability. This finding suggests that the greater number of sisters girls had, their scores on the social understanding measure decreased, whereas for boys, as the number of their older siblings increased, their scores on the social understanding measure decreased, irrespective of their older siblings' gender. Such results contradict past TOMstudies that have found the number of siblings has a positive effect on children's social understanding (e.g., Jenkins & Astington, 1996; Dunn et al., 1991; Perner et al., 1994). This finding can be explained in terms of exposure to mental state language and the connection between language and thought. Given the related finding that boys' general vocabulary scores also decreased as their number of older sisters increased, an explanation for this particular sample could be that as the number of older siblings increased in boys' families, they were less exposed to general conversation. Thus, this lack of exposure may lead to a lack of use of mental state talk which in turn may lead to the lack of understanding mental states in others. Thus, although the social interactions between the boys and their older siblings may not actually decrease, the quality and content of their conversations with family members may perhaps decrease in psychological and emotional terms. Furthermore, the present finding concerning older sisters and their younger brother's

he+t& 911cx~?.12y'S (1996) c!& t& 5 !zrm- farn;l;oe tho xrnnnm-v ~~~qkid~~ LY" Y c.zI&"* ~hze% aw--- WY, J children are the most shy and more reticent than their siblings. In relation to the present study, perhaps having a large number of older sisters somehow overwhelms or intimidates younger boys and thus influences their language ability. Given that a thorough examination of sibling effects is beyond the scope of the present study, future research needs to continue to investigate the influence of siblings on socialco,anitive and social-hagistic development (Sulloway, 1996). Self-perceptions could also play a role in that the present study found that boys perceived themselves as less well-behaved as their number of older sisters increased. Also, boys' perceived academic competence was found to decrease as the number of their older brothers increased. Taken together, these findings suggest that older siblings may have a negative effect on social understanding in boys due to some overall lack of confidence in academic ability and independent behaviour. That is, younger boys may feel less competent in academic areas, which in turn may lead to poor lan-wge and social understanding abilities. Younger siblings may also be somewhat more dependent upon the older siblings and perhaps rely on their older siblings for assistance in social situations @unn & Mum, 1986), verbal communication (Astington & Jenkins, 1995), and school work (Glass, Neulinger, & Brim, 1974; Klagsbrun, 1992). Thus, the more older siblings a boy has, the less likely he may be to develop strategies needed to develop his social understanding, language and general academic skills.

8.4. Meaning of Findings (Categorical level data) 8.4.1. Responses to Social Understanding Stories In addition to quantitative analyses performed on summed scores, content analyses of the responses to some of the open-ended questions of the social and self understanding measures were performed. In regards to the questions pertaining to social understanding within the social stories, the majority of the children included responses that directly referred to the text as opposed to references to convention or personal experience. This supports past studies that have shown children are more likely to refer to the written text as opposed to their own personal experience of hGk-ie&of Sxi& L:GIii-a~~~iiS (ze&, 1990; Gison, 1887j. i>rZI-wiiig 00 sOCiwdiur&ihcOry ihai emphasizes the influence on children's thought procesess of viewing written text as the objective truth via media, newspapers, and schooling (e.g., Bruner, 1996; Olson, 1994; Olson & Bruner, 1996; Postman, 1995), as children in the present study attempted to imagine and construct responses to the questions pertaining to ambiguous social events, they were most likely to refer to the text as a source of authority. Interestingly, as compared to boys, girls' responses were more likely to include references to personal experience. This gender-related finding is consistent with literature on epistemological development in preadolescents that claims children are influenced by society's value of societal gender norms that perceive women to be adept in social knowledge (e.g., Belenlcy et al., 1986). Thus, girls' responses to social tasks are more likely than boys to refer more to subjective and personal knowledge (Mansfield & Clinchy, 1997; Martin, 1996; Jack, 1991). Accordingly, the present results suggest that girls were more likely than boys to refer to their own personal theories and experience and thus, more likely to view themselves as the ultimate authorities on an ambiguous social situation. Further analyses of the social story responses found a sigmficant positive relation between the proportion of negative social story responses in social stories with the mental state verb understanding total score for girls and the total social understanding score for boys. Put differently, as the responses to the social story questions became increasingly negative, girls' and boys' total scores on the mental state verb understanding task and the social stories increased, respectively. This finding provides possible evidence for the contention that a sophisticated ability to understand mental states in others may include some kind of Machiavellian aspect or "theory of nasty minds" (Byme & Whiten, 1997; Frith, HappC & Siddons, 1994; Whiten & Byme, 1988). For example, although children may become expert at understanding social situations and reading the minds of others, they may not necessarily use this skill in a morally acceptable or positive manner (i.e., they may lie, steal, tease, etc.). The present study's findings thus provides support for past research that has studied the links between social understanding, deception and who possess sophisticated social reasoning skills are also adept at deceptive strategies, particularly when in conflict experiencing negative emotions such as anger or distress @unn, 1996). Findings from the present study, then, point to the need to study not only a child's competence to be able to understand mental states and emotions in others, but to study how children will apply such valuable knowledge or skill to the real world. That is, future studies need to investigate not only how much social understanding preadolescents have, but they need to determine what exactly children do with such a skill. The finding that girls' responses were significantly more negative in reference to the story that contained girl protagonists suggests that girls were more likely to report negative social scenarios, negative emotions and negative personality traits when responding to the questions pertaining to the story with three girls @ancy/Margie story). This finding contradicts Powlishta's (1995) findings that showed children's responses to narratives reflected a same-gender bias, with children responding more positively to stories that included same-gender characters. Building on Crook's (1988) argument, that all humans exist within a socially constructed cage, perhaps it is the result of social comparison that lead to such negativity in the girls' responses. Also, given Silverstein and Perlick's (1995) contention that modem society's obsession with achievement, competition and perfection in all aspects of our lives is amplified among preadolescent girls, it becomes not surprising that girls were more negative when responding to the NancyMargie story. Alternatively, the negative responses could have been due to the difference in story content, with the Nancy/Margie story describing a new student scenario, whereas the KennyMark story (all boy characters) describing a team sport scenario. Thus, the fact that the NancyMargie story described a scenario about a new girl meeting some unfamiliar children could have triggered negative personal memories, which in turn, may have coloured the girls' responses in a negative light.

8.4.2. Res~onsesto Self-understandin Interview Further findings such as response analyses of the self-understanding questions showed that thp: ~jc~i cf rhililre~rcfe~cd tc t+k nwn cp~ccf sLf xVvhznresp2A&g tc the +est;YczS 3f "What's the main reason for why you...?" This is in line with related research that shows children are more likely to refer to their own sense of agency or volition when asked to justify or explain self-statements (Harter, 1997). Such results suggest that children are more likely to refer to their own personal or subjective knowledge when discussing content involving the self. Further analyses showed that children who possessed a high global sense of self worth, were more likely to refer to their own sense of self when asked to justify their own self-statements. That is, children who were more likely to report an overall sense of well-being and perceive themselves to be generally happy were also more likely to refer to refer to a sense of personal agency when asked to justrfy self-judgments. Moreover, when further divided according to their social understanding ability, self-worth and their gender, results showed that overall, girls were more likely than boys to refer to others when asked to justify their self-statements. In particular, girls with high social understanding and low perceived global self-worth were more likely to use other-appraisals, whereas boys with low social understanding and low perceived global self-worth were more likely to use self-appraisals. Such results support the psycho-cultural view espousing that our modem culture encourages girls to define themselves in terms of others and to attempt to please others with their

achievements (Chodorow, 1978, Brown & Gilligan, 1992). Also, the fact that it was the children with the low self-worth who were more likely to report other-appraisals is also in line with both theoretical and empirical work that claims a private or personal sense of self is more psychologically healthy than a "looking glass self' or social self (Harter, 1997). Put differently, it is more psychologically healthy to define yourself in terms of personal standards as opposed to detining yourself in terms of others' standards or ideals. The finding that girls with high social understanding ability and low self-worth were more likely to report other-appraisals suggests that although some girls may have the ability to gain insight into the contents of other people's minds, they may use this ability to behave in the way in which they think that others want them to act, or say what they think others want to hear, thus presenting a "false self' (Harter et al., 1997, Jack, 1991). The concept of a false or pretend self stems from the notion of a "divided self", where the self-system may be divided into different parts (Laing, 1959). More contemporary, feminist writers have elaborated on this notion of the divided self by describing two aspects of the self or "self-voices" (Harter, 1997; Jack, 1991; Taylor et al., 1995). First, there is the voice of personal experience and observation; known as the "authentic self' or the true "I". The second voice is that of the objective, moralistic self or the "cultural eye" that includes a cultural consensus about feminine goodness, truth and value. Psychological problems arise when this judgmental "eye" begins to override the authentic self s viewpoint (Jack, 1991; Taylor et al., 1995). With regard to the gender intensification hypothesis, during preadolescence the "cultural eye" becomes stren,atened and may become more powerful than the authentic self, particularly among ,&ls and boys who ascribe to societal expectations of "feminine goodness." Within the context of this study, perhaps girls with a sophisticated social understanding ability have developed their "cultural eye" to the detriment of the authentic "I". According to this explanation, then, results from this study illustrate how a sophisticated mentalizing ability may have deleterious effects on one's emotional and psychological state. In regards to boys, the fact that boys in the low social understandingAow perceived global self-worth group also referred to other-attributions more than self-attributions suggests that boys in this particular group would rather behave in terms of pleasing others. That is, similar to the high social understandinflow perceived global self-worth group, these boys would fulfill the expectations of others, rather than fulfill their own. This result also suggests that such boys may attribute negative behaviours and failures (external attribution) to others rather than take responsibility for their own actions (internal attribution). Taken together, such results are in line with attribution and locus-of-control research that has found a link between aggression and behavioural conduct problems and external locus-of-control or external attributions in boys (Dodge & Frame, 1982; Weiner, 198%). 8.5. Role of lanrmage Although this study did not explicitly set out to examine the relation between social understanding and language, social cognition cannot be studied effectively if language is not taken into consideration (Tannen, 1994). Given findings from past research that have illustrated a strong link between language ability and social understanding (e.g., Astington & Pelletier, 1997, HappC, 1994), and language ability and social relations (e.g., Dodge & Feldman, 1990; Levinson, 1995), it was expected that language would be related to some of the measures, explaining why vocabulary ability was controlled for throughout &ta analysis. Overall, the main findings that related to language and social understanding suggest that both general vocabulary ability and the understanding of mental state verbs play a role in preadolescent's social and self understanding, their self-concept and social relations. More specifically, the finding that vocabulary ability was positively related to the social understanding and mental state verb task, but to only some of the self measures (and only among the boys) suggests that language is more related to gaining insight into other people's minds than one's own. The present results can be tentatively viewed, then, as indicating that perhaps preadolescents are in need of a specialized language or vocabulary that is specifically designed for talk about the self. Perhaps the "language of the mind is too focused on other people's minds and emotions instead of one's own thoughts and feelings (Astington & Pelletier, 1996). However, another possibility could be that the language of mind is the same for both self and other, and that the difference lies in the fact that the preadolescents in the present study did not feel comfortable or secure enough to talk about themselves on a personal level. Related to further analyses of the self, this possible lack of "language of the self' also can assist in the explanation of the absence of articulate self-esteem definitions. Perhaps these preadolescents do not have the language or vocabulary to have adequately defined the term self- esteem. Such a limited "self vocabulary" may also limit the possibilities for the notion of inner or private s3p~~ll(Bdchtin; 19111; Vyg~tslly~1965!, nr the zbili!y fn sawfly tallr t~ nceself (i-e., the 181 voice in the mind). The findings fiom the ANOVAs among the different scores obtained from the four groups of self-esteem definitions showed that the preadolescents who did not know what self- esteem meant scored the lowest on the self-understanding interview. In contrast, preadolescents who included a reference to personal experience (in defining self-esteem) scored the highest on both self-understanding and social story scores. However, the groups of children did not differ with respect to general vocabulary ability. Such results suggest, then, that children who did not know the definition of selfesteem may have had the cognitive tools to understand the question but did not have the linguistic content to talk to oneself about oneself. Thus, if people think with a dialogue template or follow a narrative mode (Bruner, 1990; Goody, 1997), perhaps these particular children did not have a dialogue template or a narrative mode for the self. On the other hand, the reason for the poor definitions could have been due to lack of exposure to self-esteem discussions in the classroom. Similar to Astington and Pelletier's (1996) araoument that teachers' increasing rnetaco,onitive talk within the classroom may increase children's metacognitive language use, perhaps the vague self-esteem definitions reflected the fact that few classroom teachers in the present study discussed the term self-esteem or self-concept. The finding that general vocabulary ability was positively related to perceptions of behavioural conduct (both self and teacher) and negatively related to athletic competence only among boys suggests that boys who were verbally sophisticated perceived themselves, and were perceived by their teachers as well-behaved. Teachers were also found to perceive boys who scored relatively high on the mental state verbs task as less athletic. Also, although general vocabulary ability was sigmficantly related to peer social competence ratings in both girls and boys, only among the boys was it ~i,~cantlyrelated to peer likability and self-understanding. Taken together, the stronger correlational results found among the boys implies that language ability may play a more significant or larger role in boys' self-concept and peerkeacher ratings than girls. The fact that language appeared to play a greater role among the boys is interesting given that it contradicts the majority of literature that claims language usually plays a larger role in the social- rncmitive rlevelnnmentr---- nf Grlc ac rnmnarerl---=--- tn kys(sklc Tchz, & Src-.yn, 1996; Tzzsz, --0--.- --.------.. 1994). 8.6. Summary In sum, results fiom this study suggest that the ability to understand mental states in others may have different social and emotional implications for preadolescent girls and boys. More specifically, the results shows that for girls, the combination of a sophisticated mentalizing ability with a negative self-view is linked to low teacher and peer ratings of social-relational ability. For boys, results show that the combination of a relatively unsophisticated understanding of mental states in others with a negative self-view is linked to low teacher and peer ratings of social- relational ability. Furthermore, it was also found for boys that a relatively sophisticated general vocabulary ability is linked to a negative self-view concerning athletic competence. Taken together, these results suggest that language may play a less beneficial role in the social world of boys as compared to girls. Furthermore, the results support the notion that self-concept plays a signif~cantreciprocal role in the relation between social cognition and social relations (see Bruner, 1996; Keating, 1990). That is, one's self-perceptions can a) help shape social understanding and social relations and b) also be shaped or influenced by social understanding and social relations. Overall, results fiom the present study are often ones that have already been expected by educators and researchers of preadolescents. The present finding that how a young person is perceived by others will, in part, help to determine her or his sense of self has long been implied in past educational research. However, this study may help to highlight the complexity of this idea and suggest that much more needs to be discovered about the role of the self-concept in the life of a preadolescent. A distinguishing feature of this study is that it describes the dynamics between the constructs of social understanding, self-concept and social relations among preadolescent girls and boys. Furthermore, the finding that global self-worth continues to play a somewhat greater role in the social and mental life of girls rather than boys is another finding that may not seem so surprising to those who spend time with preadolescents. Such a finding may merely support many aA*fiq+n-' n~4nncthqt cta-nh-:-ql manAn.--vnIa av--te4~-e -411 -4-t L +ha -la-0-n~- YUUIULWAU IIWUWIIU UYL LICYAYVLJ~IYLU ~YIIVWA AWAY wn~~uuwwouWNUC UA UIY YIWULWLIL However, the fact that the picture of stereotypes and self-mgprophesies in the classroom continues to present itself is sigmficant. Although socialcognitive research has come far over the past several decades, as the new millennium fast approaches, educators have still failed to change the societal outlook for girls and boys. One hopes that results fiom this study, and the framework and issues discussed above, will move educators closer to integrating academic and social interventions to promote personal and social competence in preadolescents. CHAPTER IX

Educational Implications "An idea is a sound from the heart." (Japanese proverb, cited in Barbeiri, 1995).

9.1. Overview The findings from the present study have many educational implications. In general, this study may assist educators in devising new teaching methodologies and curriculum that are specifically focused on the preadolescent. The following chapter will outline both the pedagogical and curricular implications of the findings including the development of new assessment tools.

9.2. Introduction Due to the lack of theoretical and empirical work relating theory of mind to schooling, it remains to be seen if the development of a theory of mind has any direct educational implications. However, a growing number of claims and studies propose that some connections may exist between TOMand schooling (e.g., Astington & Pelletier, 1996; Donelan-McCall & Dunn, 1996; Gardner, 1995; Olson & Bruner, 1996; Slomkowski & Dunn, 1996; Tomasello, Kruger, & Ratner, 1993). Thus, the present study contributes to the increasing amount of evidence of links between TOMdevelopment and classroom behaviour. Findings from this research suggest that social cognitions and language play an integral role in the social lives of preadolescents. More specifically, the ability to understand mental states in others and to understand metacognitive language was found to be linked to the way preadolescents thought and felt about themselves and others. Such findings suggest that the social and emotional life of the preadolescent remain a si,onificant part of the grade six experience. The interco~ections and complex interactions found among social understanding, self-concept and social relations demonstrate the need for educators to be aware of underlying factors that may affect the learning process in the preadolescent. Additionally, the specific gender patterns found among the three primary variables suggest that gender-stereotypic nonns may be associated with social cognitive and affective development among preadolescents. Hence, the present results suggest that educators need to adapt a holistic educational approach that strives to involve all aspects of child development in the teachinfleaming process. Results from the present study suggest that social cognitions play integral roles in the lives of preadolescents and have implications for all three of Schwab's (1983) curriculum commonplaces: 1) teacherAearner, 2) curricular models and content and 3) teaching1 learning or applications. Within the context of holistic education, the implications of the present findings will be discussed in relation to these three curricular issues.

9.3. Whom Does This Include? With regard to the teacher and the learner, the present findings that teacher ratings of children's behaviours are linked to how children actually perceive themselves supports the holistic vision that both child and teacher hold implicit theories and are equally capable of envisioning themselves as "intersubjective theorists" (Bruner, 1996; Olson and Bruner, 1996). The finding that teachers' and students' perceptions may be related suggests that the teacher and students share the role of learner and co-constructor of knowledge. This study might also encourage both teachers and students to think of themselves as collaborators, constantly learning from each other. Once students adopt such a perspective and realize that they have an equal opportunity to learn, to teach and to create knowledge, potential teacherlstudent power struggles that are common to early adolescent classrooms may be reduced (Giroux & Penna, 1979; Smith, 1990). For example, encouraging students to participate in curriculum development may provide students with a sense of personal power. That is, by contributing to such objectified knowledge as the class curriculum, students may begin to envision themselves as connected knowers (Belenky et al., 1986), ones who are active participants in the construction of objectified knowledge. Furthermore, given the elimination of traditional teacherlstudent roles, preadolescents may develop a greater sense of agency based on a greater belief in their ability which in turn may lead to greater competence in all areas of school (Bruner, 1996; Dweck, 1986). relations suggest that differential patterns of treatment by educators may reinforce biased assumptions about gender-appropriate roles for girls and boys. Although stereotypes may be useful to educators as cognitive frameworks for person-related information (Unger & Crawford, 1992), they have the potential to be detrimental to a student's sense of self worth. For example, an educator who stereotypes a child as "quiet and shy" is ignoring that child's distinctive personal qualities and herhis positive contributions to others and to society. Also, in line with the teacher expectancy research mentioned earlier (e.g., Brophy, 1983), teachers' stereotyped expectations such as the "talkative, social girl" and the "non-talkative, athletic boy" may actually help shape that targeted child's self-perceptions. As Bmner (1996) contends, "School judges the child's performance, and the child responds by evaluating himself or herself in turn." @. 37). Hence, educators need to be aware of the effects of gender-role socialization on preadolescent girls and boys and attempt to minimiirf: their damage through the avoidance of gendered metaphors. Moreover, educators of preadolescents need to challenge and redefine their definitions of femininity and masculinity and attempt to understand how gender-role stereotypes may affect preadolescents' self-perceptions and consequently shape their behaviour. Although not explicitly examined in the present study, the sibling findings have implications for the role that parents and family may play in school life of the preadolescent (Blos, 1967; Brown & Gilligan, 1992; Coles, 1990; Doyle & Markiewicz, 1996; Erikson, 1968). The finding that older siblings may have a negative effect on children's social understanding and language ability, particuiary in boys, suggests that perhaps parents may also convey stereotypical gender-role expectations to their children. For example, parents may expect the older siblings, especially the girls, to be more socially responsible and aware than their younger siblings. As with the teachers, the same self-fulfilling prophesy may occur in the home as well, with the expectations of the parents shaping their children's self-perceptions. Thus, the present study suggests that parents as well as teachers need to become aware of and to try to understand how gender-role stereotypes affect their treatment of their children. Accordingly, collaborative work involving

*.,--+c ta9,.ha- -..a ..-.,a~i~-,.~-+~--- &..I- +.. ,.~.~ii,.---+L- ---A:-- :-n ------c ---.. A:------>-- yMurrw,buwu~~ wu ~AUUUWA-uw wu AAUA~w UXUUU~CI LUG WU~VIU~-~GUS ur ur;puvc gtuucl- role stereotypes, both within and outside of the classroom.

9.4. What Should Be Taught? In addition to having pedagogical implications concerning both teacher and student, the proposed study is also sigrzlficant to holistic education in that it may generate possible curriculum

development (Clandinin & Connelly, 1992). The present findings that connections exist among social understanding, self-concept and social relations is consistent with the theoretical bedrock of holistic education that includes the three main principles of balance, inclusiveness, and interconnectedness (Miller, 1993). The study's finding of the robust relation between social and self-understanding suggests the need for a curriculum model that emphasizes both inter and intrapersonal aspects of understanding (Gardner, 1985). Such findings suggest that both students and teachers need to view each other as intentional agents, that is, human beings who share the same understanding of the mind and of learninglteaching as a collaborative, reciprocal process between teacher md student (Astington & Pelletier, 1996; Olson & Bruner, 1996; Tomasello et al., 1993). Consistent with the post-progressive internalist and personal perspective on learning (Bruner, 1996; Doll, 1993; Eisner, 1994; Gardner, 1991; Postman, 1995), and the social perspective on howledge construction (Cole, 1995; Resnick, 1991; Rogoff, 1990; Vygotsky, 1978; Wertsch, 1989), this study could also serve as a vehicle to promote TOMas an interdisciplinary form of inquiry. In line with already existing curriculum models that promote recursive, intuitive thinking, multidimensionality of the mind and learning as a personal process or cyclic journey (Bruner, 1996; Doll, 1993; Drake, 1992; Gardner, 1986; 1992), TOMcould help make the connections between conceptions of mind and curriculum more explicit (Olson & Bruner, 1996; Tomasello et al., 1993). The development of such a curriculum model would thus increase the awareness among educators and curriculum creators of the need to focus on the human, intentional aspect of education that considers the mind as a whole, as a construction of knowledge Furthermore, the complex relations illustrated in the present study support Gough's (1993) proposal to apply narrative and complexity theory to curriculum studies. To comprehend the magnitude, ambiguity and complexity of the social relations that occur within the school milieu, more complex discourses such as TOM are needed to assist children to make "human sense" out of their social interactions (Davis & Sumara, 1997; Donaldson, 1992). A TOMcurricular perspective, that emphasizes the implicit theories and mental states of both the teacher and the student may also help to uncover the issues surrounding the hiddenlimplicit (Eisner, 1994) or underground curriculum (Bruner, 1996) that views schools as "agents of ideological control which function to reproduce and to maintain dominant beliefs, values, and norms" (Giroux & Pe~a,1979, p. 214). Thus, this investigation of the understanding of mental states in self and other in preadolescents may encourage educators to consider the importance of the hidden social curricdum within the classroom and subsequently to aim to address such issues through the use of narrative and dialogical relations with their students. The finding that the selfconcept plays a major role in the relation between social understanding and s&ial relations, especially one's perceived self-worth, implies that a holistic, curriculum for preadolescents must have the child as the focus, or as Dewey (190211966) stated, "that the child and curriculum are two simple limits which define a single process." @. 11). Similar to various holistic curriculum models that emphasize self-inquiry (e.g., Bruner, 1965; Drake, 1992; Miller, 1993), the individual person is at the source of the leaming and participates in a cyclic journey of self-growth. The recursive property of the curriculum provides opportunity both to self-reflect and to participate in the cocreation of the curriculum. The present study provides evidence to support the development of a curriculum for the inner self that may be viewed as extremely valuable in holistic education for preadolescents. Given that early adolescence is a time of both self-definition and self-integration (e-g., Blos, 1967; Gilligan, 1993), holistic curriculum models that highlight the importance of one's own voice and personal experience seem particularly relevant. Such models that focus on the creation of a

-..~nmgl fi.1~4fi.1l.l- r\-.;An q.r nnnn~r4hrCnr eolC-nondn- n-A nrn1rvc:c .r.h;nh ;;rr h.- -err L-1- yuaeuu- YIYLIUCIICUY puVAYY YU UY~UALLUYLJAVA UYAA AYAA~UUAA wu UUIILJUAU,WILLUU AU LUU myuuy preadolescents to construct self-knowledge and to alter or challenge existing cognitive constructions of self and other. The finding the various aspects of the self-concept differed in importance for girls and boys provides further support for an inner curriculum. The present study reiterates the notion that the self-system is complex and may play different roles in the mental and social life of the

preadolescent (e.g., Byrne & Shavelson, 1996; Harter, 1997). Educators must then realize that students do not possess a unidimensional sense of self and that particular aspects of the self may be more salient than others. A curriculum that focuses on the self-concept of the child may thus provide some insight into the self-system of the preadolescent, and enable educators to develop programs that motivate the child from within. As noted by various holistic educators, curriculum developers need to keep in mind the ultimate goal of promoting a child's sense of well-being (Kane, 1990; Miller, 1998). Furthermore, a constructivist TOMapproach to social understanding in a preadolescent classroom may help to demonstrate that the dynamics of social understanding differ for girls and boys and may increase educators' awareness of the different ways girls and boys interpret and come to understand both self and social knowledge. In relation to giftedness (Reis, 1987; Wilgosh, 1994; Winner, 1996), assisting educators in achieving an explicit understanding of the preadolescent mind may also help to illustrate that advanced socialcognitive ability may have different consequences for girls and boys and may thus encourage educators to recognize both the social-emotional and cognitive needs of their students. Due to the lack of empirical work relating TOM to schooling, it cannot yet be said whether or not a more developed theory of mind has any educational implications. However, a growing number of studies and claims (Astington & Pelletier, 1996; Slomkowski & Dunn, 1996) suggest that there is some connection between TOM and schooling including the relation between TOMdevelopment and social behaviour which may subsequently have implications for the classroom. 9.5. How? teach in^ and Learning: Activities Encouraging educators to view curriculum fiom a TOMperspective has the potential to create a psychological-social curriculum that deals with both self-reflective, metacognitive skills or the ability to "go meta" (Bruner, 1996, p. 88), and socialcommunicativeskills such as empathetic sensitivity (e.g., Tomasello et al., 1993). Educational activities that focus on both self-reflectica and critical discussion can help promote both critical consciousness and social awareness among preadolescents (Giroux & Penna, 1979). Although reflexive thought or the "best kind of thinking" (Dewey, 1933/1966, p. 3) has been mentioned by various educators/psychologists in the past (Piaget, 1929, 1963; Vygotsky, 1978; Bruner, 1965), the majority of critical reasoning curriculum remains within the domain of gifted education (Lavecky, 1995; Swanson, 1992). In addition, the present findings support recent claims that language may play a critical role in TOMand self development (Astington & Pelletier, 1996,1997; Harter et al., 1997; Nelson et al, 1997). Various writers and curriculum designers have noted the importance of narrative and metacognition for the curriculum (e-g., Astington, 1990; Bruner, 1996, Keating, 1990; Kuhn, 1989; Olson, 1997; The Ontario Ministry of Education and Training, 1997). According to narrative theory (Bruner, 1986; 1990), children learn to make sense of their social world as they acquire the ability to tell stories about it. The use of writing and reading stories can help preadolescents to develop their ability to understand and to integrate multiple perspectives. That is, preadolescents can use narratives to integrate what they and others think, feel and do. Accordingly, educators of preadolescents need to encourage silent reading, book sharing and "story-time" with books that focus on the private, subjective experience of the preadolescent. For example, Margaret Atwood's (1988) Cat's Eve and Judy Blume's (1982) Then Again Mavbe I won't are excellent examples of books that explore the private world of preadolescent girl and boys respectively. Thus, in line with the 1997 Ontario Curriculum for Language, by examining the thoughts and emotions of characters in a novel, preadolescents can learn to clam and develop their perspective. The present study also shows the value of language activities that encourage dialogical

-.,e..-:..,. -.,e..-:..,. ..-.a ..-x,....l -&lad...... -.a -a*. a-a-.....-.-- *--..LA- 4- ..---- -4.-a--*.. :- -A-2.2-- &L-. A-u-5 wu QALUWU LQLL-uvu uuu LUUJ CIUQVLILUE~QL-U~L~ LV GUE~QE~G DLUUGUW AU ~UVIUG~uaL promote interpretive understanding and intentional creativity (Schank & Cleary, 1995). For example, in line with various educators' claims that metacognitive activities need to be integrated into the classroom (e.g., Keating, 1990; Kuhn, 1989; Olson, 1997; Paley, 1984), activities that utilize metacognitive abilities such as journal writing, bibliotherapy, and psychodrama allow a kind of experimentation within the self that may promote a greater understanding of one's own mind and how it relates to others. Thus, the use of metacognitive activities which are frequently referred to in the gifted literature (e.g., Lovecky, 1995; Swanson, 1992; Winner, 1996), should be made available to all children and need to be rigorously used within the classroom of the preadolescent. Related findings from the present study suggest metacognitive understanding may be linked to aspects of social and self understanding. Thus, based on past research that has shown exposure to metacognitive and metalinguistic verbs may increase the use of such verbs among children, which in turn may lead to a greater understanding of mental states (e.g., Astington, 1990; Astington & Pelletier, 1996; Schwanenflugel et al., 1996), both teachers and curriculum developers should also be encouraged to increase their use of metacognitive or mental language in both classroom talk and curriculum materials). Within a holistic educational framework, educators can increase their usage of metacognitive and psychological language by talking about their own thought processes and emotions. Thus, by making their own implicit, psychological theories more explicit, educators may foster the growth of metaco,@tive understanding in their students (Astington & Pelletier, 1996). The use of such language, then, may help to promote both individual growth and a united classroom consciousness. Educational programs for the @ed can also provide some answers for students who may be at-risk or particularly vulnerable to feelings of low self-worth and social incompetence. The present study suggests that girls with a relatively sophisticated TOMand boys with a relatively less complicated TOMmay suffer from such socialemotional difficulties. These findings support the view that the core educational issue remains the child's perceived worth as a person, and programs programs that aim to foster self-development (e.g., Davis & Phillips, 1994), various activities can be used to increase the self-confidence of such children. In particular, an arts-based proam that focuses on self-expression and understanding can be used to foster the self-connections that are necessary to successfully connect with others. In line with recent research on personal and emotional intelligence (Gardner, 1985; Goleman, 1995), the present study lays the groundwork for a holistic educational program that focuses on the metacognitive abilities of both social and self-understanding. Drawing on holistic and transformative educational principles and goals (see Miller, 1993), and borrowing from va$ous affective educational programs and curriculum documents (Ontario Ministry of Education and Training, 1993, 1994; 1997; Richardson & Evans, 1997; Vernon, 1997), the present study provides a framework for a metacognitive, emotional education curriculum for preadolescent (see Table 24). In general, the main goal of the program is to foster both personal and social growth through arts-based activities designed to improve preadolescents7understanding of their own feelings and behaviour and those of others. The activities listed in Table 24 aim to provide preadolescents with an understanding of the social world; to help them recognize the differing beliefs, desires and feelings that different people can bring to the same situations. Based on the constructs of social and self-understanding, Table 24 lists the corresponding metacognitive skills followed by the educational goals and suggested activities. Although space does not permit the discussion of all of the suggested activities, two are worth brief mention, as they are particularly relevant to self-development in preadolescence. First, Rudolf Steiner's (1976) concept/method of eurhythmy or "music and speech expressed in bodily movement" (Holland 1981: quoted in Reinsmith, 1989, p. 87) is an educational concept/method that especially speaks to the issue of the body-mind connection. An integral part of Waldorf Education, this form of movement education focuses on learning through rhythmic experience by moving to music (Richards, 1980). Such an activity can provide preadolescents with a personal sense of mastery by enabling them to create rhythmic movement to music. Consequently, this temp. ~f r,~~petp.~ceCE prQEte the. pregd~le~ent'cghae !Q bve and gcccpt QE~I'S mind and Table 24 193

The Connecting Cuniculum: Cuniculum for the Psvcholo~caland Social Lives of the Preadolescenf

Qverall aims: 1. to foster the understanding of self and other as psychologicai beings: 2. to further personal and social growth, 3. to foster the metacognitive abilities that underly the abilty to understand one's private and social worlds (i.e., foster "emotional literacy" skills).

Metacomitive Skill Educational Goal Educational Activitv

L Social Understanding - interpersonal understanding - develop ability to - read and reff ect on novels that focus on the attribute mental states psycho-emotional aspect of interpersonal to others relationships (peers, family, teacherlstudent)

1. Conceptual Role-Taking - understand multiple - dramatic role-playing (psychodrama) perspectivesfpoints of view - painting and writing from multiple - understand the beliefs perspectives and intentions of others - classroom debates - cosperative ,oames, peer teaching

2. Empathetic Sensitivity - understand the emotional - dramatic role-playing states of others - creative writing - conflict mediation

3. Person Perception - to view others as - creative writing/film, tv making psychological - bibliotherapy beings (i-e, understand - painting others' portraits, people sculpting personality traits) - create comics, chacteurs

4. Alternative Thinking - create alternate solutions - critical media analysis (tv, radio, (divergent thinking skills) magazines, film. etc)

IL Self-Understandino, - intrapersonal understanding - develop ability to -read and reflect on novels that focus on the attribute mental states to psycho-emotional aspect of intrapersonal self relationships (self-discovery, self-joumies)

1. Affective - develop ability to trust, - ,~dedimagery focused on bodyimind case and accept oneself connection - teacher-led meditation/mindfulness - eurythmy - self-portrait of facebody (using various mediums such as clay-self-sculture, paints, paper-mache, filmlvideo, music) - masking (creation of masks)

2. Cognitive - understand workings of one - bibliotherapy own's mind - journalling, dream work (both narrative and visual) - gain knowledge of how one - self-pomait of the mind (using various thinks mediums) - self-narration. self-autobiom~hicalwritin9 Notes. Activities within each domain (social and self respectively) are interchangeable. Activities can be applied to a variety of psycho-cultural themes (gender, ethnicity, social status/popularity). Social understanding activities can serve as both individual and cooperative group assignments. body. Such an activity, then, may help to immunize both girls and boys against selfconcept disorders where mind and body become psychologically separated. The second activity mentioned is that of guided imagery and grouplclassroom meditation that promotes the concept of mindfulness or the ability to be aware in the present (Goleman, 1995; Miller, 1993). In such activities, the teacher asks the children to imagine a particular scenario; one that is comforting and has personal meaning for those involved. The goal of visualization or guided imagery is to develop preadolescents' awareness of their thoughts, physical sensations, and emotions simultaneously. Activities that focus on mindfulness then, promote connections within all realms of the self including body-mind, thought-emotion, and logic-intuition. Furthermore, such activities could be also be viewed as relaxation techniques used to teach children co,$tive strategies that can later be used as a coping mechanism to deal with real-life emotional stress such as academic and social anxiety. Although the activities presented in Table 24 may be adapted to include any theme, results from the present study provide evidence to suggest that a holistic curriculum for preadolescents would need to include the theme of gender. To encourage preadolescents to challenge and redefine gender-role stereotypes and expectations, valuable activities could include critical analysis and discussion of the portrayal of such stereotypes in the media (see Daley et al., 1994, for an excellent curriculum guide on gender-related issues). Furthermore, such activities may encourage students to replace dualistic thinking with a more healthy, global, and holistic perspective by encouraging them to view things on a continuum (Greene, 1995; Noddings, 1992). For example, critical group discussions of dichotomous tern such as masculinelfeminine, fatlthin, good/bad could prove beneficial by inspiring preadolescents to both challenge and escape the tyranny of societal expectations that glorify linear, dichotomous thinking. In sum, through the integration of ideas from both cognitive psychology (metacognitive activities) and holistic curriculum models (art-based activities) such a program may provide a valuable contribution to preadolescent education. Such an educational program could be used to enhsnre the prohinil C1z?;,C1d2~t=A~r& h tk2t it xzm$6 tc bszz e,'.:+Ae &e;br;eGI; Gf affective and cognitive activities. However, it should be noted that due to the self-exploratory nature of some of the activities, the program is not intended to provide psycho-therapy for serious emotional problems. Educators wouId thus need to be aware of such children and to provide them with the proper psychological resource if need be. Self-connections can be further strengthened by adapting the ecological and spiritual perspective toward the self and education. Although the present findings do not speak directly to the issue of spirituality and emotionality in the classroom, this study can act as a catalyst by encouraging educators to develop curricula that attend to such crucial issues. In agreement with progressive educators who are advocates for the development of a spiritual and ecological understanding (e-g., Coles, 1990; Donaldson, 1992; Gardner, 1995; Goleman, 1995; Greene, 1995; Rozak, 1992; Sinetar, 1991), the integration of such competencies into mainstream education certainly deserves discussion and study in educational and psychological circles.

9.6. Assessment Tools 9.6.1. Teacher Rating: Scale of bbMindreading:"Classroom Behaviour A final curricular implication of the present findings concerns the development of metacognitive assessment tools for the educator. The present study provides some evidence to show that future work needs to be conducted on the creation of practical measures that teachers can use in the classroom to accurately assess socialcognitive skills such as metalinguistic and social understanding. With regard to the assessment of social understanding in preadolescents, a brief, but accurate teacher-rating scale of "mindreading behaviour" needs to be developed. Such a teacher rating scale measure would need to include a balance of both positive and negative items to reflect the subtleties of social insight. More specifically, such a scale would need to include items that could tap into the more covert, mentalistic aspects of aggression such as relational or psychological aggression including items such as teasing, sarcasm, ability to manipulate, etc. (Crick, 1996; Hughes, Soares-Boucaud, Hochmann, & Frith, 1997). 9.6.2. Metacomitive Task With regard to metacognitive understanding, the strong correlations found between mental state verb understanding and general language ability suggest that the Mental State Verb Task acted as a specialized vocabulary task related to the understanding of mental states and speech acts. Such findings suggest the need for a more specific task that actually assesses children's ability to talk about thought and emotions as opposed to merely recognizing cognitive verbs. The creation of such a task wouId be usefd as an educational assessment tool of metacognitive understanding, and as research instrument to continue to investigate the link between social understanding and the understanding of mental state verbs (Astington & Olson, 1990). However, before a reliable and valid measure of metacognitive verbs is created, the theoretical constructs of metacognition and metalinguistics need to be more fully teased apart.

9.7. Surnrnarv As many holistic educators have already noted (e.g., Kaufman, 1994; Miller, 1993; 1998; Noddings, 1992), holistic education cannot be reduced to a list of specific techniques. Holistic education is the art of teaching and learning from the heart, one that involves a dialogical relationship between the teacher and student within a community of learners. Such education flows from personal experience and our understanding of others and ourselves as psychological beings. By focusing on such issues, the present study may encourage not only preadolescents, but educators as well, to look within themselves and into the hearts and minds of others as they continue to further their self and social knowledge which in turn may assist their communication with others. The present fmdings thus support the notion that holistic eduction requires the development of both social and self understanding through the act of learning to listen to and reflect on the voice of others and oneself. In general, results from the present study suggest that 1) the ability to understand mental states in others may have different social and emotional implications for preadolescent girls and

vr&~iicAimcnci~~ ~f snA4 ~~crnitinn The bys, ~n2) Iqmg~xy $by li&ffwtfit_plcs e------findings imply that girls who have a sophisticated mentalizing ability combined with a negative self-view, may be viewed by their teachers and peers as less socially competent. Overall, these fmdings may help to increase educators' awareness and understanding of the inner lives of preadolescent girls and boys, and may encourage educators to realize that both the inter- and intrapersonal aspects of social reasoning may play an important role in learning. Moreover, findings from the present study support the view that both the cognitive and affective components of education are both significant and may foster the deveIopment of a more connected and balanced curriculum. CHAPTER X Conclusions 10.1. Introduction In this final chapter, the methodological issues concerning the newly developed instruments will be discussed first. Second, the significant limitations of the study will be outlined. Third, suggestions for future studies or possible research avenues to explore will briefly be mentioned followed by a final, concluding statement.

10.2. Methodolostical Issues The results of the study confirmed that the social stories and the self-understanding interview are effective measures of social understanding and self-understanding respectively. Reliability analyses showed that both measures obtained a high level of inter-rater agreement and adequate levels of internal consistency. Interestingly, both the social understanding story task and the self-understanding measure failed to correlate with the other two measures that were allegedly assigned to measure the same concepts (Mental State Verbs Task and the SPPC respectively). As previously mentioned, the fact that the social stories and the self-understanding interview failed to correlate with the MVT and the SPPC respectively, could have been due to either a methodological difference or a more conceptual one. Regarding the methodological difference, this seems unlikely given that both sets of measures (social stories, MVT; self-understanding interview, SPPC) related in the same manner to the Gates MacGinitie Vocabulary Test. The present results can be tentatively viewed, then, as indicating that the constructs of social understanding and self-concept are complex and multidimensional, with individual components that may either pertain separately to affect and cognition, or to both. Regarding the social stories, the strong interconelations among the four subscales further illustrates the multidimensional nature of social understanding. Such findings support both the theoretical claims (Astington, 1996; Chandler, 1987), and the corresponding evidence (Astington & Pelletier, 1997; Dunn, 1995) that suggest social understanding is comprised of both cognitive and affective components. Moreover, the present findings also reflect and support both the theoretical (e.g., Astington, 1993; Chandler, 1988; Keating, 1990; Perner, 1988; Wellman, 1990) and empirical claims (e.g., Case et al., 1996; Fox, 1991, Hatcher et al., 1990; Mansfield & Clinchy, 1997; Nelson et al, 1997) that propose the ability to gain social insight into contents of other people's minds is a very complex process during preadolescence. The finding that the empathetic sensitivity scale was the only subscale to correlate positively with the Mental State Verbs Task reveals that emotional understanding is related to the more linguistic task of understanding mental state verbs. This fmding provides indirect support for a recent study by Hughes and Dunn (in press) that showed a positive link between children's understanding of emotions and their understanding of belief. Taken together, these findings suggest that understanding emotions in others may play a large role in understanding mental state verbs, that is, the understanding of mental state verbs and the understanding of emotions may be intricately related and dependent on one another. In regard to the validity of the social stories, the present study did not address this issue due to the conceptual complexity of social understanding. Based on Chandler's (1987) definition of a more complex or interpretive theory of mind, the social stories task in the present study was designed to measure the components of social understanding including conceptual role-taking, empathetic sensitivity and person perception. Thus, to ensure adequate validity, future research needs to administer the social stories together with standardized tasks that have been commonly used to measure similar socialcognitive constructs such as the Interpersonal Reactivity Index 0 (Davis, 1980), the Epistemic Doubt Interview (EDI) (Boyes & Chandler, 1992) and HapHs Strange Stories task (HappC, 1994). Significant correlations among the tasks would provide evidence to show that the social stories are a valid measure of social understanding. Thus, such measures could prove to be useful in future validation studies. The Mental State Verbs Task employed in this investigation was also comparatively new.

Although all of the items were gleaned fiom past research (Astington & Olson, 1990; Booth &

U~11 1004. c;ri&nrr; 100<\ .&& Gnrn ..a ern011 Alnt ehrthr AT - 'XI\ rrn rfi*A*r h-A ..coA -11 nF A*-) -44 ., YIYYyU, &44"/, -I- ...... a .. YYIIYI YY". OCYYJ - -v,, .A" ULYV, A.L... -CL.I "I these items together as one task. It is problematic that the task's split-half reliability (Spearman- Brown corrected) was only -53. However, this value is similar to the values obtained by previous

researchers with similar items (Booth & Hall, 1994; Siddiqui, 1995). Both Booth and Hall and Siddiqui do not offer any explanation as to why the reliability of this task is so low aside from mentioning that the task is a fairly new instrument that has not undergone much investigation. A possible explanation may have to do with the fact that various sections of the task dealt with both metacognitive verbs (e.g., think, know) and metalinguistic verbs (e-g., say). As Astington and Olson (1990) note, the link between metacognition and metalinguistics is not fully understood and thus perhaps different sections of the task measured conceptually different items. Accordingly, more work needs to be done in the area of understanding metacognitive and metalin,@tic verbs and their link to understanding mental states.

10.2.1. Limitations Several limitations of the present study may have precluded the results from demonstrating a stronger relation among social understanding, self-concept and social relations. First, in relation to design, the present sample was one of convenience which was not truly representative of the general population. Based on past studies which found a positive correlation to exist between feelings of self-worth and social class (e-g., Smilansky, 1968), the sample may have been biased because the majority of the students were from Euro-Canadian, middle-class backgrounds. To eliminate such bias, future research should thus employ a random sample of participants from an population of grade 6 students, diverse in both ethnic background and social class. In relation to instrumentation, there were several general limitations that may have influenced the results. First, the issue of social desirability and self-disclosure may have affected the self-report measure of the SPPC and the self-understanding interview. Although Harter (1985) claims to have controlled for this issue, the fact remains that some children may have simply reported what they think that the researcher expected, not what they really felt or thought (Kiess,

1989)?\.p-egEA&g t& iss~5cf y--&sc!091A=, stt2&ts h/tfcaz2 t& elSzTd yvcvez z= ;=;+.= likely than boys and men to self-disclose about personal issues in both paper and pencil items and interviews (Gjerde, 1995; Unger & Crawford, 1992). In regards to the present study, the finding that girls scored higher than boys in self-understanding could be explained in terms of this gender bias. In other words, the fact that girls may be more likely than boys to talk about themselves and about personal issues could have caused the girls' responses to be more elaborate than the boys, which in turn may have effected their overall scores. Second, the possibility of a gender-bias in terms of experimenter effects could have also iduenced the results. Researchers have noted that the gender of the researcher or experimenter may affect how the participant responds in the study, with same-gender researcher-participant pairs performing better than mixed-gender (Barbieri, 1995). The fact that the researcher in the present study was female could have biased the results in favour of the girls which could have inflated their scores. Ideally, future research needs to control for such experimenter effects by employing an equal number of both females and males to administer the tasks. Third, time constraints imposed by the school schedule was another limitation. The lack of time could have directly affected the inter- and intrapersonal interview responses. The short time limit may have caused children in the present study to feel somewhat rushed and not psychologically comfortable enough to provide answers that truly reflected their social and self- understanding. Additionally, time constraints may have had an effect on the Teacher's Rating Scale of Child's Actual Behaviour (Harter, 1985). Although efforts were made to allow ample time for the teachers to fill out the questionnaire, perhaps some teachers felt that this time period was insufficient. Thus, the scores may have failed to accurately represent the students' actual behaviour within the classroom. A final general limitation concerns the issue of confounding cognitive factors such as memory, verbal and nonverbal intelligence. To address this complex issue, the present study attempted to control for memory and verbal intelligence by administering control memory questions and a general receptive vocabulary task respectively. However, due to financial and time 202 verbal ability were not employed. Thus, some of the differences and relations found in the present study could have been influenced by individual differences in the children's spatial-visual abilities andlor expressive verbal ability. More specific limitations of the individual measures employed in the present study will be addressed below.

10.3. Limitations of the Measures 10.3.1. Self-concept A major conceptual limitation regarding self-concept concerns the possibility that the relation between self-concept and mental health is not linear (Kohn, 1994). An awareness of the mistakes and limitations of the self is crucial for normal adjustment (Damon & Hart, 1992), and consequently, an extremely high self-concept score may reflect maladaptive rather than adaptive behaviour. Thus, the assumption that a high self-concept score reflects a psychologically well- adjusted person may not always be true. For example, in a review of peer relations, Hartup (1983) found that children with extremely high and extremely low self-concept scores tended to have friendship difficulties. Within the context of the present study, it is possible that some children with high self-concept could have received extremely low social ratings and thus could have affected the results by lowering the correlations. Thus, future studies need to take into account the fact that high scores on self-concept measures do not necessarily reflect optimal psychological well-being. Similarly, due to the complex nature of the self-system, the attainment of a true measurement of all self-constituents has yet to be achieved (Hoff-Summers, 1993). Due to both financial and time constraints, the present study investigated only certain aspects of the self- system. Similar to the multi-methodological method employed by Hart and Fegley (1995), future research needs to investigate more thoroughly all aspects of the self-system, exploring both affective and cognitive components. Also, in relation to gender, research needs to take into account the possibility that self-construction is an ongoing, active process. That is, the self can be

-&.-.,.a #... #. L,l:,f c,,: ,,,,,,,, ,L'llL ,,,, :,,I1 -,.I---.I- * -)--t--A- --A C.- -*--c7 vIur*- cw -ubL-IuI-g p1-~33 VI VF~CIV-L~ awl U~Ga WUUCLC ~CLi)l ~tlltlbi)r ii pmduci 203

of "being" (Bruner & Kalmar, 1997; Cederblom, 1989, Tannen, 1994). Accordingly, research on social cognition in preadolescents needs to borrow from contemporary literature that views the self as an active process, such as the recent work on self-narration and the autobiographical self (Bruner, 1990; Kehily, 1995; Martin, 1996). A further complexity of the self-system includes that of the relational self; how we think and feel about ourselves may differ according to our social environment (Harter, 1993). In the present study, although the SPPC supposedly assessed the children's overall self-perceptions in different areas of competence, these perceptions may be contingent upon the social context. For example, the answers to the SPPC global self-worth subscale may have been significantly different from the ones obtained in the present study if the children were asked to think of themselves in terms of their family. Although this issue has recently been addressed in older adolescents with Harter et al., (1997) having recently created an adapted version of the global self-worth scale to tap into relational self-perceptions, relational self-worth at the early adolescent level remains to be investigated. A further methodological problem with Harter's (1985) SPPC scale is that it fails to provide an accurate measure of how a child feels about her or his physical sense of self. Despite research that shows body image is a prevailing factor in how preadolescents' view their physical appearance (e-g., Edwards, 1993; Heaven, 1996), only one item in the SPPC addresses this issue (i.e., "Some kids wish their body was different BUT Other kids like their body the way it is"). Accordingly, future self-perception questionnaires or interview schedules need to include more detailed questions concerning body image and issues around the relation between body weight and attractiveness. For example, drawing from psychological clinics that treat selfconcept problems in children, questions from measures such as Davis' (1993) Concerns with Weight and Shape Scale (i.e., "I think a lot about my weight or shape." and "I feel proud about my weight or shape.") could add tremendous content validity to self-concept scales. 204

10.3.2. Social and Self Understanding Although the social story task and the self-understanding measure were found be effective measures of social and self understanding respectively, some methodological limitations may have influenced the present findings. Structurally the two tasks were not similar, the social story task focused the questions on a socially ambiguous narrative, whereas the questions in the self- understanding interview were based on previously obtained SPPC responses. Perhaps two more structuraliy similar tasks would have provided different results. Following examples of TOM studies that have assessed the understanding of false belief in self and other (e.g., Astington & Homer, 1995), in the social stories used in the present study, after each question the children could be asked how they would feel and think, and what they would do. Another methodological problem concerning the social stories involved the gender of the protagonists. Due to financial and time constraints, only two stories were created, each story describing similar social situations with three characters of the same gender. Drawing on recent research on feminist epistemology and epistemological development in preadolescents (Mansfied & Clinchy, 1997; Potter, 1993), ideally, a larger number of stories would have been created including same and both gender characters. However, in light of the environment in which this present study was conducted, such lengthy narrative tasks would have increased out-of-class-time, which would have created a new set of problems. Thus, the present study relied on the two stories as an assessment of social understanding. Finally, the technical issue of the tape recorder could have affected the children's responses. Unfortunately, this possible confounding variable was unavoidable and thus, the researcher attempted to create a psychologically comfortable atmosphere for the child as much as possible. Despite such attempts, some of the children remained visibly nervous and shy throughout the interview and subsequently, this "test anxiety" may have affected their responses. To help eliminate some of the effects the tape recorder may have had on the results, future studies should perhaps hide the tape recorder so it is not visible. Although participants must agree to be tz~~pcmrl_e& zrl_this knw % &~rpt+t they +lJ& ~ t& p~pzzcp_cf 2 kp A~cr&r,+Ae fact that they do not see the machine may automatically relieve some anxiety that the children may be feeling.

10.3.3. Social Relations (Peer and Teacher Ratings) Although the sociometxic and teacher ratings employed in the present study have been used in past research (Harter, 1985; Matthews & Keating, 1995), there were some limitations to these measures that may have affected the results. Concerning the peer ratings of social competence measure, within the research area of peer relations, there is an ongoing debate around whether or not social competence ratings really measure social competence or whether they measure social conformity (Oppenheimer, 1989). Thus, children rated as highly socially competent by their peers could have been those who conform to the classroom rules and who are not particularly socially competent. Also, researchers argue that constructs of social competence and popularity are defined differently among girls and boys (Daniels-Biemess, 1989). Thus, in the present study, although both girls and boys were asked to rate everyone in their classroom, the children may have rated their peers against very different standards. Regarding teacher ratings, many of the teachers in the present study commented on how unsure they felt about their ratings of the children. Consistent with past literature that has noted the positive bias of teacher ratings (e.g., Ladd & Profilet, 1996), many of the teachers in the present investigation also expressed their unwillingness to rate children negatively. Although physical appearance was the only subscale that was sigDlficantly skewed (i.e., most teachers said that it was "sort of true" that the student was attractive), the lack of variability may help to explain the lack of significant relations that was found between teacher ratings and social understanding. A final issue regarding teacher ratings deals with the issue of gender. Although the gender of the children may have affected the teachers' ratings (see Chapter 8), studies have shown that the gender of the teacher may have an effect on their responses (e.g., Dusek & Joseph, 1983).

Although the small sample of teachers in the present study (g = 8,s female, 3 male) precluded any interaction between the gender of the teacher and gender of the student affected the children's social understanding scores. Similar to the contention used in teacher expectation research (e.g., Brophy, 1983), differential expectations of female and male teachers could have shaped their perceptions of how they viewed their students. Thus, to control for such teacher gender-bias, future studies need to employ an equal number of female and male teachers and subsequently, analyze the data for gender-related differences. In general, although economical in terms of both finances and time, the use of teacher and peer ratings to assess actual behaviour of children within the classroom may not provide the most "objective" measure. To obtain a more "true" measure of the children's behaviour, the present study could have included methods such as having two or three researchers perform behavioural observations of the children then perform reliability on the data. Another technique that could have been used is videorecording of the children's real behaviour, although this may also have problems of subjective coding. In sum, the limitations mentioned above are ongoing debates and for the time being, remain research problems that are relatively unsolved.

10.4. Future DirectionsKhallen~es The results of this study both support and contradict the outcomes of some of the previous studies in the field of Theory of Mind and social cognition (e.g., Astington & Jenkins, 1995, Dunn, 1995; Hart & Fegley, 1995; Hatcher et al., 1995; Matthews & Keating, 1995). We still are, then, in need of more evidence and knowledge to create a fuller model of the social and emotional correlates of social understanding. Nevertheless, the present investigation provides further validation of a multimethodological approach to investigating social understanding in preadolescence. By employing techniques of both self-report pencil-and-paper measures and semi- structured story-interviews, this study increases confidence in the use of these measures to investigate the link between individual differences in social understanding, self-concept and social relations. By suggesting that sociocultural factors may underlie gender-related differences found among social understanding, self-concept and social relations, future research should be conducted that goes beyond the scope of the present investigation. For example, more research needs to investigate the influence of gender-stereotypic norms on the social cognitions of preadolescents. As contemporary researchers are beginning to suggest (Debold at al., 1996; Harter et al., 1997;

Jussim & Eccles, 1995), the impact of gender-role stereotypes may be examined by studying gender-role self-perceptions among preadolescents and their link to other socialcognitive and emotional variables. Perhaps researchers need to remain cognizant of the fact that simple gender effects explains neither how the difference got there not what maintains it. A more useful strategy may be to theorize the context in which gender effects operate and examine factors that drive when and how differences are likely to occur. Furthermore, as suggested by Harter et al. (1997), future research needs to investigate the links between the multidimensional self-concept and its relation to children's ability to articulate their subjective experience or their level of "voice" (Gilligan, 1982). That is, in addition to studying the influence of verbal and non-verbal abiliq, future studies need to create ways to assess whether children feel that their emotional voices or true selves are being "psychologically" heard by adults. Thus, the investigation of children's emotional need to be heard may assist researchers in answering the question of how and why children develop social insight and how this mentalizing ability relates to their ability to get along with others. The findings from the present study that suggest girls with high social understanding scores and low self-worth and boys with low social understanding scores and low self-worth may be at risk for social-relational problems, suggests the need for future TOMresearch to focus on clinical subsamples of children. For example, studies need to explore how social understanding and self relations develop in atypical populations, particularly psycho-social-behavioural problems such as aggressive, neglected, and depressedlinhibited children. Thus, to assist in the investigation of such a problem, contemporary TOM researchers need to integrate past work on

-~.v,-h~~-a+hnln-v en-;-1 wlai4nrr~.,..A enAa1 -n-Xnrr Atle n.rA Lnrve /n rr Pn:n 100n- n-d-, yUJYUUyUCIlVAV6 J y UUIILU A-IUUVUU LUAU U\IYI.U YU6UUVU AkA 6UAU UAAU WJU \U-6.y bVAW) L//V) YVU5W 208

& Frame, 1983; Nolen-Hoeksema, Seligman, & Girgus, 1991; Schultz & Selman, 1989). Results of this study warn against the investigation of social understanding in an undifferentiated group of children. Future studies need to take into account individual differences, (particulary gender), not just age. The use of independent sample 1-tests between gender and regression analyses with gender as an independent variable may not be sufficiently sensitive to pick up the sometimes subtle but potentially significant gender effects found within the data. Although the sample used in the present study was not adequate to speak to the issue of cultural and ethnic issues, these are additional sociocultural factors that need to be addressed in future investigations. Although large scale, cross-sectional studies that occur within one population are the most economically feasible, such studies provide culturally-biased information and do not allow researchers to examine the development of attitudes over time. Given the strong socio-cultural influences on preadolescent self-identity, following the method of Barbara Kerr (1994), it would be interesting to follow a group of children (girls and boys) and interview them at pivotal times in their development about emotional and psychological issues. Ideally, a crosscultural, longitudinal study on the links between social understanding and social relations would provide an excellent sketch of how preadolescents in various cultures view and understand both themselves and others. To investigate properly the complexities of the preadolescent mental and social life, the ideal method is to employ both "thick and "thin" methodologies (Geertz, 1973). Future research needs to use techniques from both the etic (outside-in) or more objective perspective drawn from cognitive research (e.g., standardized tests, controlling for verbahonverbal IQ) and the more emic (inside-out) or subjective techniques gleaned from social and anthropological research (e.g., visual-drawing tasks, narrative techniques, action research, autobiographical work). The use of such a multi-tasked approach, may provide data that is richer in content and more valuable in the end to both educators and researchers. 10.5. Concluding Statement The findings from this study support the reciprocal-constructivist model of social cognitive development by illustrating that self-concept plays an intervening role in the link between social understanding and social relations. The finding that the ability to understand thoughts and feelings in self and others is linked to one's ability to interact socially highlights the importance of framing our psychological and educational questions in terms of both cognitive and emotional considerations. This study is also of theoretical and practical significance. Theoretically, the present findings support the view that an advanced ability to gain insight into the minds of others may negatively effect how children perceive themselves and socially relate to others (Chandler, 1987; Donelan-McCall & Dunn, 1996; Dunn, 1995). In practical terms, this study may encourage educators to adapt a more humanistic and balanced educational approach by creating curricular activities that emphasize the importance of both personal and social understanding. The present findings also reiterate the importance of focusing on individual differences when approaching social understanding from a TOMperspective. The finding that a greater number of sigdicant differences was found as the present sample was further divided into subgroups illustrates the significance of shifting attention from similarities among members of an age cohort to individual differences. For example, upon initial inspection of the present sample, few differences were found, but a more detailed analysis provided significant differences in relation to gender, global self-worth and social understanding. Similarly, results obtained from the correlational analyses of three main constructs, social understanding, self-concept and social relations, provide further evidence for focusing on the multidimensional nature of social constructs. The fact that correlations among the subscales of the various constructs were often found to be more robust than correlations among the total combined

scores supports recent criticisms of single-factor psychological theories (e-g., Damon & Hart, 1992; Harter, 1997; Kohn, 1994). Hence, such findings stress the need for researchers to examine the various components of complex psychological constructs such as social

.,-rl-,*-,.J:-, --1C ,,,,,, c -,A ,,,:el ,,,,, c ,,,, UllUGlDUUlCLLII& DGU-CIUUCIGPL QLlU JUb1CU CIUIII~LCULC. Through the integration of both cognitive and psychosocial feminist theory, this study was intended as a beginning step toward a broader description and understanding of preadolescents' mental and social life. The present finding that correlational pattems among social understanding, selfconcept and social relations differed according to gender challenges previous research that has failed to find any gender-related differences in such areas (e.g., Fox, 1991; Hart & Fegley, 1995). The illustration of particular gender pattems supports recent studies that suggest a sophisticated mindreading ability may have a detrimental effect on social and emotional adjustment in some children (Dunn, 1995), while at the same time, it provides new evidence suggesting that such an effect may be especially si,onificant for girls. The present study, then, highlights the necessity for future social co,onitive research to include gender as a contributing factor. The present study also highlights the crucial role that language plays in the mental and social life of the preadolescent. The finding that for girls, general language and metalinapistic ability was linked to social understanding but not to self-understanding, suggests that language may help girls to gain insight into the minds of others but not their own. That is, although language may provide a vehicle through which the theories of others' mind and feelings are created (e-g., Astington, 1990; Bruner, 1990), findings suggest that among preadolescent girls, language may be of greater use in the creation of "theories-of-others" than "theories-of-self". Consequently, researchers need draw from psycho-linguistic literature (e.g., Goody, 1995; Lakoff, 1975; Potter, 1993; Tannen, 1994), and begin to integrate language-related tasks (both receptive and expressive) into studies of social cognitive development among preadolescents. In agreement with Pekrun (1990), this study emphasizes the critical role both educators and peers play in the socialcognitive development of the preadolescent. The fact that self-concept was found to have an important influence on the preadolescent's school experience may help to remind educators that preadolescence is a pivotal time in self-development (e.g., Dusek & Flaherty, 1981; Gardner, 1985; Keating, 1990). Educators need to be cognizant that early adolescence is a time when girls and boys experience increased social consciousness and social pressures to conform to rronltor-ctnmntrm> nnrmc T ;to&co tho rronltor-mlotoA finA;nne AomnnchtoA ;n th;r ehw?xr mrlrr aYYYIA UCIIYYLJ YAW IAVAIIl). YU1I .. A*) LUY bY.IYIA AIAILIY *--aV --VYUYYCIY CIYU UCYYJ -YJ help educators to realize that their students' behaviour cannot be separated from the larger social context and may thus encourage educators to become aware that they have differential gender-role expectations for students. Such findings may also help educators to understand the formation and perpetuation of these beliefs within the classroom and strive to minimize the extent to which these expectations may influence preadolescent self-perceptions. With regard to peers, the present findings provide yet another example of how important peers are to the social, mental and emotional life of the preadolescent in the classroom (e-g., Dodge

& Feldman, 1993; Kroger, 1996; Levitt & Selman, 1996). The fact that peer ratings of social competence shared more connections to social understanding and language but popularity was linked to perceived social acceptance suggests that popularity may have a greater effect on one's self-concept. The present results can be tentatively viewed, then, as indicating that it is popularity and not social competence that has the greatest impact on how worthy one sees oneself as a person. Furthermore, the differential gender patterns found in the peer ratings supports the socio-cultural view that gender, social competence and popularity are culturallydefined constructs (e-g., Maccoby, 1988, 1990, 1998). In sum, the results of this study provide important new evidence for study of preadolescent's understanding of self and other. Specifically, the present study illustrates the signif~cantrole self-concept plays in the relation between preadolescents' ability to make meaning from their social world and their ability to socially relate to others. The demonstration of such complex relations suggests that the link between the mental and social life of the preadolescent is not as clear as previously thought. Finally, this study provides another example of how crucial it is to continue to investigate the preadolescent's "theory of mind fiom a psycho-cultural perspective. Substantial research, then, is needed both to replicate and extend findings on how preadolescents make meaning from their social world and how this affects their sense of self and their ability to relate socially to others. Findings fkom the present study, then, point to the need to study not only a child's en--tanen tn nnAarrtqnA mantql ctatac qnA omn&nnc ;n nthatr hit qlrn tn efirrfwr hnx.. AGlrf-n .-A1 Yw-yYWYYI LW CLUUIILIUL.. Y .A.Y*.CUI LICCICILI ....U YIIIWUW..., A.. WLUY*LI, VYL -LIw LW LILY.., YV.. w-"awu ..- 212

apply such valuable knowledge or skill to the real world. That is, future studies not only need to investigate how much social understanding preadolescents have, but also determine what exactly children do with such a skill. In general, the present study suggests that the various domains of social understanding should be differentiated, not only with respect to each other, but also in terms of their links with affect and social-relational ability. In conclusion, this study demonstrates that the dynamics among social understanding, self- concept and social relations differ for preadolescent girls and boys. This study's focus on gender and the emotional and social correlates of TOMdevelopment further contributes to the "second wave" of TOMresearch (Astington, 1997). This new line of interest explores TOM from a sociocultural perspective. That is, it takes into account factors that have been neglected in past TOM research such as individual differences, social-emotional development and schooling effects. Drawing from areas of cognitive science and social development, then, this study's multilateral approach to exploring social cognition in grade six students, may encourage both educators and psychologists to engage in a collaborative effort to continue to investigate the link between the preadolescent's mental and social lives. Through the implementation of a balanced curriculum that includes both affective and co,@tive components within a "psychologically safe" classroom climate, educators can create a learning environment that will maximize the growth of both intra- and interpersonal functioning. Hence, the present study provides support for the reciprocalconstructivist model of social understanding and may help to bridge the gap between theory of mind and social cognitive research. In the end, this study deepens educators' awareness of the different ways in which girls and boys interpret and understand social and self knowledge, thus furthering the discourse in socialcognitive psychology and holistic education. REFERENCES Abrams, C. (1989). Differential association: Social developments in gender identity and intergroup relations during adolescence. In S. Skevington & D. Baker (Eds.), The social identitv of women @p. 59-83). London, UK: Sage Publications. Adler, P., & Adler, P. (1998). Peer Dower: Preadolescent culture and identity. New Brunswick, NJ: Rutgers University Press. Alsaker, F. (1995). Is puberty a critical period for socialization? Journal of Adolescence,

d18 427-444. American Association of University Women (1990). ShortchanPring girls, shortchangine America Washington, DC: The Analysis Group Incorporated. Armstrong, T. (1994). Multi~leintelligences in the classroom. Alexandria, VA: Association for Supervision and Curriculum Development. Asher, S., & Dodge, K. (1986). Identifying children who are rejected by their peers. Develo~mentalPsvcholow. 22,444-449. Astington, J. (1988). Children's understanding of the speech act of promising. Journal of Child Languaae. 15,157-173. Astington, J. (1990). Narrative and the child's theory of mind. In B. Britton & A. Pelle,gini, (Eds.), Narrative thought and narrative language @p. 151-171), Hillsdale, NJ: Lawrence Erlbaum. Astington, J. (1993). The child's discovery of the mind. Cambridge, MA: Harvard University Press. Astington, J. (1996). What is theoretical about the child's theory of mind? A Vygotskian view of its development. In P. Carruthers & P. Smith (Eds.), Theories of theorv of mind @p. 184- 199). Cambridge, England: Cambridge University Press. Astington, J. (1996, April). The roots of metacomition and morality. Paper presented at the annual meeting of the American Educational Research Association, New York Astington, J. (1997, April). Discussion of the social orieins and im~licationsof a theorv of mind. Discussant's comments presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Astington, J. (1998). Theory of mind, humpty dumpty, and the icebox. Human Develo~ment, 41,30-39. Astington, J., & Gopnik, A. (1991). Theoretical explanations of children's understanding of the mind. British Journal of Develo~mentalPsvcholow. 9,7-3 1. Astington, J., & Jenkins, J. (1995). Theory of mind development and social understanding. Cornition and Emotion. 9,151-165. Astington, J., & Olson, D. (1990). Metacognitive and metalinguistic language: Learning to talk about thought. Applied Psycholom: An International Review. 39,77437. Astington, J., & Olson, D. (1995). The cognitive revolution in children's understanding of mind. Human Development. 38,179-1 89. Astington, J., & Pelletier, J. (1996). The language of mind: Its role in teaching and learning. In D. Olson & N. Torrance (Eds.), Handbook of education and human develo~ment:New models of learning. teaching. and schooling (pp. 593-619). Oxford, UK: Blackwell. Astington, J., & Pelletier, J. (1997, April). Young children's theory of mind and its relation to their success in school. Paper presented at the biennial meeting of the Society for Research in Chiid Development, Washington, DC. Atwood, M. (1988). Cat's eye. New York: McClelland-Bantam, Inc. Austin, J. (1962). How to do things with words. Cambridge, MA: Harvard University Press. Bakhtin, M. (198 1). The dialoyic imagnation. (C. Emerson & M. Holquist, Trans.). Austin, TX: University of Texas Press. Baldwin, J. (1902). Social and ethical interpretations in mental life. New York: Macmillan. Baldwin, J. (1913). History of psychologv. London, UK: Watts. Baldwin, M. (1992). Relational schemas and the processing of social information. Psychological Bulletin. 112,461-484. Barbieri, M. (1995). Sounds from the heart: Learninn to listen to girls. Portsmouth, NH: Hernemann. Barenboim, C. (1981). The development of person perception in childhood and adolescence: From behavioral comparisons to psychological constructs to psychological comparisons. Child Development. 52,129-144. Baron-Cohen, S. (1995). Mindblindness. Cambridge, MA. BradfordMIT. Baron-Cohen, S., Jolliffe, T., Mortimore, C., & Robertson, M. (1997). Another advanced test theory of mind: Evidence from very high functioning adults with autism or Asperger syndrome. Journal of Child Psvcholoq and Psvchiatry. 38,813-822. Baron, R., & Kenny, D. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personalitv and Social psycho lo^. 5 1,1173-1 182. Barresi, J., & Moore, C. (1995). Intentional relations and social understanding. Behavioral and Brain Sciences. 18,256-279. Barton, S. (1994). Chaos, self-organization and psychology. American Psvcholo~st.49,5-14. Bartsch, K, & Wellman, H. (1995). Children talk about the mind. New York: Oxford University Press. Bed, C. (1990). Development of knowledge about the role of inference in text comprehension. Child Development. 61,101 1-1023. Belenky, M., Clinchy, B., Goldberger, N., & Tarule, J. (1986). Women's wavs of knowing;. New York: Basic Books. Bennett, M. (Ed.) (1993). The child as psvcholo@st: An introduction to the develo~mentof social co-mition. New York: Harvester Wheatsheaf. Berger, P., & Luckrnann, T. (1967). The sociai construction of reaiitv: A treatment in the sociologv of knowledee. Garden City, NY: Anchor. Bhatnager, P., & Rastogi, M. (1986). Cognitive style and basis ideal disparity in males and females. Indian Journal of Current ?svcholog;ical Research. 1.36-40. Block, J., Gjerde, P., & Block, J. H. (1991). Personality antecedents of depressive tendencies in 18-year-olds: A prospective study. Journal of Personalitv and Social Psvcholo~.60, 726-738. Blos, P. (1967). The second individuation process of adolescence. The Psvchoanalytic Studv of the Child. 22,162-1 86. Blume, J. (1982). Then again, mavbe I won't. New York: Basic Books. Boivin, M., & Hymel, S. (1997). Peer experiences and social self-perceptions: A sequential model. Developmental Psvcholow. 33,135-145. Booth, J., & Hall, W. (1994). Role of the cognitive internal state lexicon in reading comprehension. Journal of Educational PsvchoIo~~.86,413-422. Booth, J., & Hall, W. (1995). Development of the understanding of the polysemous meanings of the mental-state verb know. Cognitive Development. 10,529-549. Bosacki, S. (1995). Field independence-dependence. self-concept and plavfulness in preadolescents. Unpublished Masters thesis, University of Windsor, Windsor, Ontario, Canada. Bosacki, S. (1997, June). Theorv of mind. self-conce~tand social com~etencein preadolescence. Paper presented at the 58th Annual Convention of the Canadian Psychological Association, Toronto, Ont. Bowler, D. (1992). "Theory of mind" in Asperger's syndrome. Journal of Child Psvcholoq and Psvchiatrv. 33,877-893. - Boyes, M., & Chandler, M. (1992). Cognitive development, epistemic doubt, and identity formation in adolescence. Journal of Youth and Adolescence. 21,277-384. Brooks-Gum, J. (1989). Pubertal processes and the early adolescent tradition. In W. Damon (Ed.), Child development today and tomorrow @p. 155-176). San Franscico, CA: Jossey- Bass. Brophy, J. (1983). Research on the self-fulfilling prophecy and teacher expectations. Journal of Educational Psvchologv. 75,63 1-661. Brophy, J., & Good, T. (1970). Teachers' communication of differential expectations for children's classroom performance: Some behavioural data. Journal of Educational Psvcholo~~.61,365-374. Broughton, J. (1981). The divided self in adolescence. Human Development. 24,13-32. Brown, L., & Gilligan, C. (1992). Meeting at the crossroads. New York Ballantine. Bruce, P., (1958). Relationship of self-acceptance to other variables with sixth grade children oriented in self-understanding. Journal of Educational Psvcholow. 49,229-237. Bruchkowsky, M. (1992). The development of empathic cognition in middle and early childhood. In R. Case (Ed.), The mind's staircase: Exploring: the conce~tualundeminnings of children's thought and knowled~e. up. 153-170). New York: Erlbaum. Bruner, J. (1965). The growth of mind. American Psvchologist. 20,1007-1017. Bruner, J. (1986). Actual minds. possible worlds. Cambridge, MA: Hanrard University Press. Bruner, J. (1990). Acts of meanin?. Cambridge, MA: Harvard University Press. Bruner, J. (1996). The culture of education. Cambridge, MA: Harvard University Press. Bruner, J., & Kalmar, D. (1997). Narrative and metanarrative in the construction of self. In M. Ferrari & R. Sternberg (Eds.), Self-awareness: Its nature and development, @p. 1- 52). New York: Guilford. Buss, A. (1980). Self-consciousness and social anxiety. San Francisco, CA: Freeman & Company. Bybee, J. (1998). The emergence of gender differences in guilt during adolescence. In J. Bybee (Ed.), Guilt and children (pp. 114-122). San Diego, CA: Academic Press. Byme, B., & Shavelson, R. (1987). Adolescent self-concept: Testing the assumption of equivalent structure across gender. American Educational Research Journal. 24,356-385. Byme, B., & Shavelson, R. (1996). On the structure of social self-concept for pre-, early, and late adolescents: A test of the Shavelson, Hubner, and Stanton (1976) Model. Journal of Personalitv and Social Psvcholo9v. 70,599-6 13. Byme, R., & Whiten, A. (1997). Machiavellian intelligence. In A. Whiten & R. Byme (Eds.), Machiavellian intelligence 11: Extensions and evaluations (pp. 1-23). Cambridge, UK: Cambridge University Press. Cahan, E. (1997, Fall). On the uses of history for developmental psychologists or on the social necessity of history. SRCD Newsletter. 40, (3), p. 2. Callahan, C., Cunningham, C., & Plucker, J. (1994). Foundations for the future: The socio- emotional development of gifted, adolescent women. Roeper Review. 17,99-105. Canadian Teachers' Federation (1991). A capella: A report on the realities. concerns. ex~ectations and barxiers experienced bv adolescent women in Canada. Ottawa, ON: CTF. Cantrell, V., & Prinz, R. (1985). Multiple perspectives of rejected, neglected, and accepted children: Relation between sociometric status and behavioral characteristics. Journal of Abnormal Child Psvcholosw. 12,189-1 98. Carpendale, J. (1998, March). Developing an interpretive theory of mind: Children's construction of social understanding within interaction. Paper presented at the University of Toronto, Toronto, Ontario. Carpendale, J., & Chandler, M. (1996). On the distinction betwen false belief understanding and subscribing to an interpretive theory of mind. Child Develo~ment.67,1686-1706. Carey, S. (1985). Conce~tualchange in childhood. Cambridge, MA: Bradford/MIT. Carruthers, P., & Smith, P. (1996). Introduction. In P. Carruthers, & P. Smith (Eds.), Theories of theories of mind, (pp. 1-10). Cambridge, UK: Cambridge University Press. Case, R., Okamoto, Y., Griffin, S., McKeough, A., Bleiker, C., Henderson, B., & Stephenson, K (1996). The role of central conceptual structures in the development of children's thought. Monoma~hsof the Societv for Research in Child Develo~ment.61 (1-2, Serial No. 246). Cauce, A. (1987). School and peer competence in early adolescence: A test of domain-specific self-perceived competence. Develo~mentalPsvchologv. 23,287-291. Cederblom, J. (1989). Willingness to reason and the identification of the self. In E. Maimon, B. Nodine, & F. O'Connor (Eds.), Thinking. reasoning.. and writing (pp. 147-159). New York: Longman Inc. Chandler, M. (1987). The othello effect: Essay on the emergence and eclipse of sceptical doubt. Human Develo~ment.30,137-159. Chandler, M. (1988). Doubt and developing theories of mind. In J. Astington, P. Harris, & D. Olson (Eds.), Develo~ingtheories of mind (pp. 387-413). Cambridge, MA: Cambridge University Press. Chandler, M., & Helm, D. (1984). Developmental changes in the contribution of shared experience to social role-taking competence. International Journal of Behavioral Deveiopment. 7,145-156. Chaplin, W., John, O., & Goldberg, L. (1988). Conceptions of states and traits: Dimensional attributes with ideas as prototypes. Journal of Personality and Social Psvchologv. 54, 541-557. Chaman, T., & Shmueli-Goetz, Y. (in press). The relationship between theory of mind, language and narrative discourse: An experimental study. Current Psvcholow of Cognition. Chernin, K. (1985). The hunmv self: Women. eatin and identity. New York: Harper & Row. Chodorow, N. (1978). The reproduction of mothering. Berkeley, CA:University of California Press. Clandinin, D., & Connelly, F. (1992). Forms of curriculum inauiry. Draft prepared for the international encyclopedia of education. Oxford: Pergamon Press. Code, L. (1993). Taking subjectivity into account. In L. Alcoff, & E. Potter (Eds.), Feminist epistemolo~ies(pp. 15-48). London, UK: Routledge. Coie, J. (1990). Toward a theory of peer rejection. In S. Asher & J. Coie (Eds.), Peer reiection in childhood @p. 365-401). New York: Cambridge University Press. Coie, J., & Dorval, B. (1973). Sex differences in the intellectual structure of social interaction skills. Develo~mentalPsvcholoq. 8,261-267. Cole, M. (1995). Sociocultural-historicalpsychology: Some general remarks and a proposal for a new kind of cultural-genetic methodology. In J. Wertsch, P. Del Rio, A. Alvarez (Eds.), Sociocultural studies of mind (pp. 187-214). New York: Cambridge University Press. Coles, R. (1990). The spiritual life of children. Boston, MA: Houghton Mifflin. Coles, R. (1997). The moral intelligence of children: How to raise a moral child. New York: Plume. Collin, A. (1996). Re-thinking the relationship between theory and practice: practioners as map -readers, map-makers - or jazz players? British Journal of Guidance and Counsellin~.24, 67-8 1. Cooley, C. (1912). Human nature and social order. New York: Scribner. Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco, CA: Freeman & Company. Crick, N. (1996). The role of overt aggression, relational aggression, and prosocial behaviour in the prediction of children's future social adjustment. Child Develo~ment.67,2317-2327. Crick, N., & Ladd, G. (1989). Nominator attribution: Does it affect the accuracy of children's sociometric classifications? Merrill-Palmer Ouarterlv. 35,197-207. Crook, J. (1988). The experiential context of intellect. In R. Byrne & A. Whiten (Eds.), Machiavellian intelli~ence:Social ex~ertiseand the evolutionof intellect in monkevs. ams, and humans @p. 347-362). Oxford, UK: Blackwell. Cross, S., & Markus, H. (1994). Self-schemas, possible selves, and competence performance. Journal of Educational Psvchologv. 86,423-438. Dabrowski, K. (1964). Positive disinteeration. London, UK: Little Brown. Daley, M., Diakiw, J., McLeod, P., Pitcairn, S., Scrimger, N., & Sumner, D. (1994). The girl child: An investment in the future. Toronto, ON: UNICEF. Damon, W., & Hart, D. (1988). Self-understanding in childhood and adolescence. New York: Cambridge University Press. Damon, W., & Hart, D. (1992). Social understanding, seif-understanding, and morality. In M. Bornstein & M. Lamb (Eds.), Developmental psvcholoq: An advanced textbook (3rd ed., pp. 42 1-464). Hillsdale, NJ: Erlbaum. Daniels-Bierness, T. (1989). Measuring peer status in boys and girls: A problem of apples and oranges? In B. Schneider, G. Attili, J. Nadel & R. Weissberg (Eds.) Social comwtence in developmental perswtive (pp. 107-120). London, UK: Kluwer Academic Publisher. Davies, M., & Stone, T. (1995). Introduction. In M. Davies & T. Stone (Eds.), Folk Psvcholo~ (pp. 1-44). Cambridge, MA: Blackwell. Davis, B., & Sumara, D. (1997). Cognition, complexity, and teacher education. Harvard Educational Review. 67,105- 124. Davis, M. (1980). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psvcholom. 44,113-126. Davis, R. (1993). Concerns with Weight and Shape Scale. Unpublished manuscript. Davis, M., Conklin, L., Smith, A., & Luce, C. (1996). Effect of perspective tallcing on the cognitive representation of persons: A mer-&g of self and other. Journal of Personalitv and Social psycho lo^, 70,7 13-726. Davis, R., & Phillips, W. (1994). Turning: points: A ~svchoeducationalprom for overcoming an eating disorder. Unpublished manuscript. Day, J., & Tappan, M. (1996). The narrative approach to moral development: From the epistemic subject to dialogical selves. Human Development. 39,6742. Deaux, K, & Major, B. (1987). Putting gender into context: An interactive model of gender -related behavior. Ps~chologicalReview. 94,369-389. Debold, E., Tolrnan, D., & Brown, L. (1996). Embodying knowledge, knowing desire: Authority and split subjectives in girls' epistemological development. In N. Goldberger, J. Tarule, B. Clinchy & M. Belenky (Eds.), Knowledge. difference and power @p. 85-125). New York: Basic Books. Dennett, D. (1987). The intentional stance. Cambridge, MA: MIT Press. Dewey, J. (190211966). The child and the curriculum. Chicago, IIL: University of Chicago. Dewey, J. (193311966). How we think: A restatement of the relation of reflective thinking to the educative process. Boston, MA: D.C. Heath and Company. Dockett, S. (1997, April). Young children's peer ~o~ularitvand theories of mind. Poster presented at the Biennial Meeting of the Society for Research in Child Development, Washington, DC. Dodge, K (1983). Behavioral antecedents of peer social status. Child Develo~ment.54, 1386-1399. Dodge, K, & Feldman, E. (1990). Issues in social cognition and sociometric status. In S. Asher & J. Coie (Eds.), Peer rejection in childhood (pp. 119-155). New York: Cambridge University Press. Dodge, K, & Frame, C. (1982). Social cognitive biases and deficits in aggressive boys. Child Development. 53,620-635. Dodge, K, Murphy, R., Buchsbaum, K. (1984). The assessment of intention-cue detection skills in children: Implications for developmental psychopathology. Child Develo~ment.55, 163-173. Doll, W. (1993). A post-modern perspective on curriculum. New York: Teachers College Press. Donaldson, IM.(1992). Human minds: An exploration. New York: Penguin Press. Donelan-McCall, N., & Dunn, J. (1996). School work, teachers, and peers: The world of first grade. International Journal of Behavioral Development. 2 1,155-178. Doyle, A, & Markiewicz, D. (1996). Parents' interpersonal relationships and children's friendships. In W. Brukowski, A. Newcomb & W. Hartup (Eds.), The companv they kee~:Friendship in childhood and adolescence @p. 115-136). Cambridge, UK: Cambridge University Press. Drake, S. (1992). Developing an intemted curriculum using the story model. Toronto, ON: OISE Press. Dm,J. (1988). The begnning: of social understanding, Oxford, UK: Blackwell. Dm,J. (1995). Children as psychologists: The later correlates of individual differences in understanding of emotions and other minds. Emotion and Cognition, 9,187-201. Dunn, J. (1996). The Emanuel Miller memorial lecture 1995, children's relationships: Bridging the divide between cognitive and social development. Journal of Child Psvchology and Ps~chiatry.37,507-5 18. Dm,J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L., (1991). Young children's understanding of other people's feelings and beliefs: Individual differences and their antecedents. Child Develo~ment.62,1352-1366. Dm,J., & Mum, P. (1986). Siblings and the development of prosocial behaviour. International Journal of Behavioural Development. 9,265-284. Dusek, J., & Flaherty, J. (1981). The development of self-concept during adolescent years. Monoma~hsof the Societv for Research in Child Development. 46, (4, Serial No. 191). Dusek, J., & Joseph, G. (1983). The bases of teacher expectancies: A meta-analysis. Journal of Educational Psvcholow. 75,327-346. Dweck, C. (1986). Motivational processes affecting learning. American Psvchologist. 41, 1040-1048. Eagley, A., & Mladinic, A. (1989). Gender stereotypes and attitudes toward women and men. Personality and Social Psvcholo~~Bulletin. 15,543-558. Eccles, A., Bauman, E., Rotenberg, K. (1989). Peer acceptance and self-esteem in gifted children. Journal of Social Behavior and Personality. 4,401-409. Edelbrock, C., & Achenbach, T. (1984). The teacher version of the child behavior profile: I. Boys aged 6-1 1. Journal of Consulting and Clinical Psvcho1og;y. 52,207-217. Edwards, P. (1993). The A cappella papers: Self-esteem and adolescent women. Ottawa, ON: Canadian Teachers' Federation. Edwards, D., & Potter, J. (1995). Attribution. In R. Ham5 & P. Steams (Eds.), Discursive psvcholow in practice @p. 87-1 19). London, England: Sage Publications. Eisenberg, A. (1992). Conflicts between mothers and their young children. Merrill-Palmer Ouarterlv, 38,2 1-43 Eisenberg, N., & Mussen, P. (1989). The roots of prosocial behaviour in children. New York: Cambridge University Press. Eisner, E. (1994). Cornition and curriculum reconsidered (2nd Ed.). New York: Teachers College Press. Elkind, D. (1967). Egocentrism in adolescence. Child Development. 38,1025-1034. Elliott, G., & Feldman, S. (1990). Capturing the adolescent experience. In S. Feldman & G. Elliott (Eds.), At the threshold: The developing adolescent @p. 1-13). Cambridge, MA: Harvard University Press. Erdley, C., & Asher, S. (1996). Children's social goals and self-efficacy perceptions as influences on their responses to ambiguous provocation. Child Development. 67,1329- 1344. Erikson, E. (1953). Childhood and society. New York: Norton Erikson, E. (1968). Identity: Youth and Crisis. New York: Norton. Erwin, P. (1993). friends hi^ and peer relations in children. West Sussex, UK: John Wiley & Sons. Evans, D., Brody, L., & Noam, G. (1995). Self-perceptions of adolescents with and without mood disorder: Content and struture. Journal of Child Psvcholorrv and Psvchiatrv. 8, 1337- 1354. Fabricius, W., Schick, K., Prost, J., & Schwanenflugel, P. (1997, April). We don't see eve-to- eye: Develo~mentof a constructivist Theorv of Mind in middle childhood. Poster presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Fabricius, W., Schwanenflugel, P., Kyllonen, P., Barclay, C., & Denton, M. (1989). Developing theories of mind: Children's and adults' concepts of mental activities. Child Devel 60.1278-1290. Farber, B. (1989). Psychological mindedness: Can there be too much of a good thing? Ps~chotheraphv.26,2 10-2 17. Farber, B., & Golden, V. (1997). Psychological mindedness in psychotherapists. In M. McCallum & W. Piper (Eds.), Psychological mindedness: A contemporarv understanding (pp. 21 1-235). Mahwah, NJ: Erlbaum. Feldman, C. (1992). The new theory of theory of mind. Human Development. 35,107-1 17. Ferguson, T., Stegge, H., & Damhuis, I. (1991). Children's understanding of guilt and shame. Child Development. 62,827-839. Feshbach, N. (1978). Studies of empathetic behavior in children. In B. Maher (Ed.), Promess in experiinental mrsonalitv research (pp. 147). New York: Academic hss. Flapan, D. (1968). Children's understandine of social interaction. New York: Teachers College Press. Flavell, J. (1992). Perspectives on perspective-taking. In H. Beilin & P. Pufall (Eds.), Piaeet's theory: Pros~ectsand possibilities (pp. 107-139). Hillsdale, NJ: Erlbaum. Flavell, J., Botkin, P., Fry, C., Wright, J., & Jarvis, D. (1968). The development of role-taking and communication skills in children. New York: Wiley. Flavell, J., Flavell, E., & Green, F. (1983). Development of the appearance-reality distinction. Cognitive Psvcholo-q. 15,95-120. Flavell, J., & Miller, P. (1997). Social cognition. In W. Damon (Series Ed.) & D. Kuhn & R. Siegler (Vol. Eds.), Handbook of child ~sycholom:Vol2. Cognition. perception and language (5th ed.) (pp. 851-898). New York: Wiley. Fodor, J. (1992). A theory of the child's theory of mind. Cognition. 44,283-296. Ford, M. (1982). Social cognition and social competence in adolescence. Developmental Psvcholog. 18,323-340. Ford, M., & Tisk, M. (1983). A further search for social intelligence. Journal of Educational Psvcholog. 75,196-206. Fox, R. (1991). Developing awareness of mind reflected in children's narrative writing. British Journal of Developmental Psychology. 9,281-298. Fraser, S., & Strayer, J. (1997, April). Guilt and shame in middle childhood: Relationships with empathetic responsiveness. Poster presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Frith, U., Happi, F., & Siddons, F. (1994). Autism and theory of mind in everyday life. Social Develo~ment.3,108-124. Furnham, A., & Radley, S. (1989). Sex differences in the perception of male and female body shapes. Personality and Individual Differences. 10,653-662. Galambos, N., Almeida, D., & Petersen, A. (1990). Masculinity, femininity, and sex role attitudes in early adolescence: Exploring gender intensification. Child Develo~ment,

-961 1905-1914. Gardner, H. (1985). Frames of mind. New York: Basic Books. Gardner, H. (1991). The unschooled mind. New York: Basic Books. Gardner, H. (1995, November). Reflections on multiple intelligences: Myths and messages. &i Delta Ka~pan,200-209. Geertz, C. (1973). The internretation of cultures. New York: Basic Books. George, T., & Hartmann, D. (1996). Friendship networks of unpopular, average, and popular children. Child Developent. 67,230 1-23 16. Gergen, K. (1985). Social constructionist inquiry: Context and implications. In K. Gergen, & K. Davis (Eds.), The social construction of the person @p. 3-18). New York: Springer- Verlag . Gilligan, C. (1982). In a different voice: Psvcholo~icaltheory and women's develoument. Cambridge, MA: Harvard University Press. Gilligan, C. (1993). Adolescent development rediscovered. In A. Garrod (Ed.), Approaches to moral development (pp. 103-132). New York: Teachers' College Press. Giroux, H., & Pema, A. (1979). Social education in the classroom: The dynamics of the hidden curriculum. Theory and Research in Social Education. 7,21-42. Gjerde, P. (1995). Alternative pathways to chronic depressive symtoms in young adults: Gender differences in developmental trajectories. Child Development. 66,1277-1300. Gjerde. P., & Block, J. (1991). Preadolescent antecedents of depressive symptomatology in late adolescence: A prospective study. Journal of Youth and Adolescence. 20,2 15-230. Glass, D., Neulinger, J., & Brim, 0. (1974). Birth order, verbal intelligence, and educational aspiration. Child Development. 45,807-8 1 1. Gnepp, J., & Chilamkurti, C. (1988). Children's use of personality attributions to predict other people's emotional and behavioral reaction. Child Development. 59,743-754. Gnepp, J., & Klayman, B. (1992). Recognition of uncertainty in emotional inferences: Reasoning about emotionally equivocal situations. Develo~mentalPsvcholow. 28,145-158. Goetz, T., & Dweck, C. (1980). Learned helplessness in social situations. Journal of Personality and Social Psvcholow. 39,246-255. Goldberger, N., Tarule, J., Clinchy, B., Belenky, M. (Eds.). (1995). Knowledge. difference and power. New York: Basic Books. Goleman, D. (1995). Emotional intelligence. New York: Bantam Books. Goody, E. (1995). Introduction: Some implications of a social origins of intelligence. In E. Goody (Ed.), Social intelligence and interaction @p. 1-33). New York: Cambridge University Press. Goody, E. (1997). Social intelligence and language: Another rubicon? In A. Wbiten & R. Byrne (Eds.), Machiavellian intelligence IT Extensions and evaluations @p. 365-396). Cambridge, UK: Cambridge University Press. Gopnik, A. (1993). How we know our minds: The illusion of first-person knowledge of intentionality. Behavioral and Brain Sciences. 16,l- 14. Gopnik, A., & Astington, J. (1988). Children's understanding of representational change and its relation to the understanding of false belief and the appearance-reality distinction. Child Develo~ment.58,26-37. Gopnik, A., & Wellman, H. (1995). Why the child's theory of mind really & a theory. In M. Davies & T. Stone (Eds.), Folk Psvcholo~(pp. 232-258). Cambridge, MA: Blackwell. Gordon, R. (1986). Folk psychology as simulation. Mind and Language. 1,158- 171. Gough, N. (1993). Environmental education, narrative complexity and postmodern sciencelfiction. International Journal of Science Education. 15,607-625. Gove, W., & Tudor, J. (1973). Adult sex roles and mental illness. American Journal of Sociologv. 78,s 12-835. Greene, M. (1995, January). Art and imagination: Reclaiming the sense of possibility. Phi Delta Ka~pan,378-382. Grice, H. (1968). Utterer's meaning, sentence-meaning and word-meaning. Foundations of Langua~e.4,l-18. -11, S. (1995). A cognitive-developmental analysis of pride, shame, and embarrassment in middle childhood. In J. Tangney & K. Fischer (Eds.), Self-conscious emotions: The psvcholoev- - of shame. milt.- embarrassment and ride @p. 219-252). New York: Guilford Press. Haines, B., & Bartels, F. (1997, April). Shvness and academic ~erformance:Mediating or moderating- roles of self-esteem. attributional style. and ethnicity. Poster session presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Halpem, D. (1992). Sex differences in comitive abilities (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum. HappC, F. (1993). Communicative competence and theory of mind in autism: A test of relevance theory. Cognition. 48,101-1 19. Happt, F. (1994). An advanced test of theory of mind: Understanding of story characters' thoughts and feelings by able autistic, mentally handicapped, and normal children and adults. Journal of Autism and Developmental Disorders. 24,129-155. Happt, F., Winner, E., & Brownell, H. (1998). The getting of wisdom: Theory of mind in old age. Developmental Psycholow. 34,358-362. Harris, P. (1989). Children and emotion. Cambridge, MA: Basil Blackwell. Hart, D., & Fegley, S. (1995). Prosocial behavior and caring in adolescence: Relations to self- understanding and social judgement. Child Development. 66,1346-1359 Harter, S. (1985). Manual for the self-perception profile for children (SPPC). University of Denver. Harter, S. (1993). Developmental processes in the construction of the self. In H. McGurk (Ed.), Childhood social develo~ment:Contemmrarv ~ewectives@p. 45-78). East Essex, UK: Lawrence Erlbaum. Harter, S. (1996). Teacher and classmate influences on scholastic motivation, self-esteem, and level of voice in adolescents. In J. Juvonen & K Wentzel (Eds.), Social motivation: Understanding children's school adjustment @p. 11-42). New York: Cambridge University Press. Harter, S. (1997, April). The appraisals of self and other: Which comes first and does it matter? Paper presented at the biennial meeting of the Society for Research in Child Development, Washington, D.C. Harter, S., & Buddin, B. (1987). Children's understanding of the simultaneity of two emotions: A five-stage developmental acquisition sequence. Developmentd Psycholow. 23,388-399. Harter, S., & Monsour, A. (1992). Developmental analysis of conflict caused by opposing attributes in the adolescent self-portrait. Developmental Psvcholo~.28,25 1-260. Harter, S., Waters, P., Whitesell, N., & Kastelic, D. (1997, April). Predictors of level of voice amon? high school females and males: Relational context. support. and gender orientation. Paper presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Hartup, W. (1983). The peer system. In E. Hetherington (Ed.), Handbook of child psvcholow: Socialization. personality and social development @p. 103-196). New York: Wiley. Hatcher, R., & Hatcher, S. (1997). Assessing the psychological mindedness of children and adolescents. In M. McCallurn & W. Piper (Eds.), Psychological mindedness: A contem~oraryunderstanding @p. 59-75). Mahwah, NJ: Erlbaum. Hatcher, R., Hatcher, S., Berlin, M., Okla, K., & Richards, J. (1990). Psychological mindedness and abstract reasoning in late childhood and adolescence: An exploration using new instruments. Journal of Youth and Adolescence. 19,307-325. Hatzicbristou, C., & Hopf, D. (1996). A multiperspective comparison of peer sociometric status groups in childhood and adolescence. Child Development, 67. 1085-1102. Heaven, P. (1996). Adolescent health: The role of individual differences. London, UK: Routledge. Hesse-Biber, S. (1996). Am I thin enough vet? The cult of thinness and the commercialization of identity. New York: Oxford University Press. Heyman, G., & Dweck, C. (in press). Children's thinking about traits: Implications for judgements of self and others. Child Development. Hill, J., & Lynch, M. (1983). The intensification of gender-related role expectations during early adolescence. In J. Brooks-Gum, & A. Peterson (Eds.), Girls at ~ubt3I-t~:Biolo&al and psvchosocial perspectives (pp. 201-228). New York: Plenum. Hinde, R. (1979). Towards understanding relationships. London, UK. Academic Press. Hobson, P. (1991). Against the theory of Theory of Mind'. British Journal of Developmental Psvcholow. 9,33-5 1. Hoffman, M. (1986). Affect, cognition and motivation. In E. Higgins, & R. Sorrentino (Eds.), Handbook of social behaviour (pp. 244-280). New York: Guildford. Hoff-Sommers, C. (1994). Who stole feminism? New York: Simon & Schuster. Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and Clinical Psvchologv. 33,307-3 16. Homer, B. (1995). Children's attribution of second-order mental states and truth value iud~ement Unpublished masters thesis. University of Toronto: Toronto, ON. Homer, B., & Astington, J. (1995, March). Children's understanding:of second-order beliefs in self and other. Paper presented at the biennial meeting of the Society for Research in Child Development, Indianapolis, IN. Honess, T. (1979). Children's implicit theories of their peers: A developmental analysis. British Journal of Psvchologv. 72,445-45 1. Honess, T. (1981). Girls' and boys' perceptions of their peers: Peripheral versus central and objective versus interpretive aspects of free descriptions. British Journal of Psvcholow, 7Q, 485-497. Howard, J., & Allen, C. (1989). Making meaning: Revealing attributions through analyses of readers' responses. Social Psychologv Ouarterlv. 52,280-298. Hughes, C., & Dunn, J. (in press). Theory of mind and emotion understanding: Longitudinal associations with mental-state talk between young friends. Developmental Psvcholo~~. Hughes, C., Soares-Boucaud, Hochmann, J., & Frith, U. (1997). Social behaviour in pervasive developmental disorders: Effects of informant, group and "theory of mind". European Child and Adolescent Psychiatry. 6,19 1-198. Hughes, R., Tingle, B., & Swain, D. (1981). Development of empathetic understanding in children. Child Development. 52,122- 128. Hyde, J., & Linn, M. (1988). Gender differences in verbal ability: A meta-analysis. Psycholog5cal Bulletin. 104,53-69. Iverson, A., Barton, E., & Iverson, G. (1997). Analysis of risk to children participating in a sociometric task. DeveIopmentaI PsychoIow. 33,104-1 12. Jacobsen, T., Edelstein, W., & Hofmann, V. (1995). A longitudinal study of the relation between representations of attachment in childhood and cognitive functioning in childhood and adolescence. Developmental Psycholo-p. 30,112- 124. Jack, D. (1991). Silencing the self: Women and d ression. New York: Harper Collins. Jahnke, H., & Blanchard-Fields, F. (1993). A test of two models of adolescent egocentrism. Journal of Youth and Adolescence. 22,313-327. James, W. (1892). The principles of psvchologv. New York: Fawcett Publications. Jenkins, J., & Astington, J. (1996). Co,gnitive factors and family structure associated with theory of mind development in young children. Developmental Psycholoqv.32,70-78. Johnson, C. (1988). Theory of mind and the structure of conscious experience. In J. Astington, P. Harris, & D. Olson (Eds.), Developing:theories of mind. (pp. 47-63). New York: Cambridge University Press. Johnson, E. (1997). Children's understanding of epistemic conduct in self-deception and other false belief stories. Child Development. 68,1117-1 132. Johnston, D. (1988). Adolescents' solutions to dilemmas in fables: Two moral orientations-two problem solving strategies. In C. Gilligan, J. Ward, & J. McClean Taylor (Eds.), Mapping. the moral domain: A contribution of women's thinking. to svchological theory and education @p. 49-71). Cambridge, MA: Harvard University Press. Jones, M., & Gerig, T. (1994). Silent sixth-grade students: Characteristics, achievement and teacher expectations. The Elementary School Journal. 2,169-182. Jussim, L., & Eccles, J. (1995). Naturally occuring interpersonal expectancies. In N. Eisenberg (Ed.), Social development (pp. 74-108). Thousand Oaks, CA: Sage. Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgement under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge University Press. Kane, J. (1998). The new language of reform. Encounter. 11,2-3. Karrniloff-Smith, A., & Russell, J. (1994). The mind itself, compared to children's theories of it. In S. Guttenplan (Ed.), A com~anionto the ~hilos hv of mind, (pp. 257-259). Cambridge, MA: Blackwell. Kaufman, J. (1994). The wedge between emotion and co,onition: Feminism, knowledge and power. Holistic Education Review. 7,43-49. Keating, D. (1990). Adolescent thinking. In S. Feldrnan & G. Elliott (Eds.), At the threshold: The developing: adolescent (pp. 56-89). Cambridge, MA: Harvard University Press. Kegan, R. (1994). In over our heads: The mental demands of modem life. Cambridge, MA: Harvard University Press. Keenan, T. (1995). The role of echoic information in young: children's com~rehensionof sarcasm. Unpublished doctoral dissertation. University of Toronto, Toronto, ON. Kehily, M. (1995). Self-narration, autobiography and identity construction. Gender and Education. 7,23-32. Kerr, G. (1994). Smart girls two: A new ~svcholowof girls. women and gftedness. Dayton, OH: Ohio Psychology Press. Kiess, H. (1989). Statistical concepts for the behavioral sciences. Boston, MA: Myn & Bacon. King, C., Akiyama, M., & Elling, K. (1996). Self-perceived competencies and depression among middle school students in Japan and the United States. Journal of Earlp Adolescence. 16,192-2 10. Kitchener, K-,& King, K. (198 1). Reflective judgement: Concepts of justification and their relationship to age and education. Journal of Applied Psvcho1og;y. 2,89-116. Klagsbrun, F. (1992). Mixed feelings: Love. hate. rivalry. and reconcilation among: brothers and sisters. New York: Harper & Row. Kohn, A. (1994, December). The truth about self-esteem. Phi Delta Kappan, pp. 272-283. Kroger, J. (1996). Identity in adolescence: The balance between self and other. London, UK: Routledge. Kuhn, D. (1989). Making cognitive development research relevant to education. In W. Darnon (Ed.), Child development today and tomorrow, (pp. 261-287). San Francisco, CA: Jossey-Bass. Kuhn, D., Nash, D., & Brucken, L. (1978). Sex-role concepts of two- and three-year-olds. Child Develo~ment.49,445-45 1. Ladd, G., & Profilet, S. (1996). The child behavior scale; A teacher-report measure of young children's aggressive, withdrawn ,and prosocial behaviors. Devel mental Psvcholoy, 32.1008-1024. La Greca, A. (198 1). Peer acceptance: The correspondence between children's sociometric scores and teachers' ratings of peer interation. Journal of Abnormal Child Psycholop. 9, 167-178. Laing, R. (1959). The divided self. London: Penguin Books. Laing, R. (1961). Self and others. London: Pen,* Books. Lakoff, R. (1975). Language and woman's place. New York: Harper and Row. Lalonde, C., & Chandler, M. (1995). False belief understanding goes to school: On the social- emotional consequences of coming early or late to a hrst theory of mind. Cormition and Emotion. 9,167-185. Lalonde, C., & Chandler, J. (1997, April). The development of an interpretive theorv of mind. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Washington, DC. Leech, G. (1980). Exvlorations in semantics and ~ramatics.Amsterdam: Johns Benjamins B. V. Leekam, S. (1993). Children's understanding of mind. In M. Bennett (ed.), The child as ~svchologiist:An introduction to the develo~mentof social codtion @p. 26-61). New York: Harvester Wheatsheaf. Leslie, A. (1988). Some implications of pretense for mechanisms underlying the child's theory of mind. In J. Astington, P. Harris, & D. Olson (Eds.), Developing theories of mind (pp. 19- 46). New York: Cambridge University Press. Levinson, S. (1995). Interactional biases in human thinking. In E. Goody (Ed.), Social intelligence and interaction (pp. 221-260). New York: Cambridge University Press. Levitt, M., & Selman, R. (1996). The personal meaning of risk behaviour: A developmental perspective on friendship and fighting in early adolescence. In G. Noam & K. Fischer (Eds.), Develo~mentand vulnerability in close relationships @p. 201-233). Mahwah, NJ: Lawrence Erlbaum. Lewis, M. (1993). Self-conscious emotions: Embarrassment, pride, shame, and guilt. In M. Lewis & J. Haviland (Eds.), Handbook of Emotions @p. 563-573). New York: Guilford. Light, P. (1993). Developing psychologies. In M. Bennett (Ed.), The child as psvcholog;ist: An introduction to the development of social cornition @p. 191-198). New York: Harvester Wheatsheaf. Lillard, A. (in press). Ethnopsychologies and the origins of a theory of mind. Child Develo~ment. Liard, A. (1998). Ethnopsychologies: Cultural variations in theory of mind. Psvcholoeical Bulletin. 123,39-43. Livesley, W., & Bromley, D. (1973). Person perception in childhood and adolescence. London: Wiley. Loeb, R., & Jay, G. (1987). Self-concept in gifted children: Differential impact in boys and @ls. Gifted Child Ouarterly. 3 1,9- 14. Loevinger, J. (1976). Ego develo~ment:Conce~tions and theories. San Franscisco, CA: Jossey- Bass. Lovecky, D. (1995). Exceptionally Nedchildren: Different minds. Roe~erReview. 17,116-120. Maccoby, E. (1988). Gender as a social category. Developmental Psvcholo~v.26,755-765. Maccoby, E. (1990). Gender and relationships: A developmental account. American Psvchologist,

245 513-520. Maccoby, E. (1998). The two sexes: Growing up apart. coming together. Cambridge, MA: Harvard University Press. MacGinitie, W., & MacGinitie, R. (1994). Gates-MacGinitie reading tests. Toronto, ON: Riverside Publishing Company. Mahler, M., Pine, F., & Bergman, A. (1975). The psychological birth of the human infant. New York: Basic Books. Mansfield, A., & Clinchy, B. (1997, April). Toward the internation of ectivity and subiectivitv: A longitudinal studv of e~istemoloecaldeveio~ment between the ages of 9 and 13. Poster presented at the Biennial Meeting of the Society for Research in Child Development, Washington, DC. Mant, C., & Perner, J. (1988). The child's understanding of commitment. Developmental Psvcholow. 24,343-351. Martin, K. (1996). Pubertv. sexuality. and the self: Girls and bovs at adolescence. New York: Routledge. Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist. 41,954969. Markus, H., & Wurf, E. (1987). The dynamic self-concept: A social psychological perspective. Annual Review of Psvcholow. 38,299- 337. Marsh, H., & Shavelson, R. (1985). Self-concept: Its multifacted, hierarchical structure. Educational Psvchologist. 20,107-125. Martin, K. (1996). Pubertv. sexualitv. and the self. New York: Routledge. Matthews, D., & Keating, D. (1995). Domain specificity and habits of mind: An investigation of patterns of high-level development. Journal of Earlv Adolescence. 15,3 19-343. Matthews, D., & Smyth, E. (1997). Encouraging bright girls to keep sh;ning: Countering school and social pressures against gifted girls. Orbit. 28,3436. McCormick, P. (1994). Children's understanding of mind: A case for cultural diversity. Unpublished doctoral dissertation, University of Toronto,Toronto, Ont. McKeough, -4. (1992). A neo-structural analysis of children's narrative and its development. In R. Case (Ed.), The mind's staircase: Explorin~the conce~tualundeminnines of children's thought and knowledpe,(pp. 171-188). New York: Erlbaum. McKeough, A., Templeton, L., Marini, A. (1995). Conceptual change in narrative knowledge: Psychological understandings of low-literacy and literate adults. Journal of Narrative and Life History. 5,21-49. McKeough, A., Yates, T., & Marini, A. (1994). Intentional reasoning: A developmental study of behaviorally aggressive and normal boys. Develo~mentand Psvchopatholow. 6,285-304. McGrath, E., Keita, G., Strickland, B., & Russo, N. (1990). Women and depression. Washington, DC: American Psychological Association. Mead, G. (1934). Mind. self and society. Chicago, IL: Chicago University Press. Miller, J. (1993). The holistic curriculum. Toronto, ON: OISE Press. Miller, J. (1998). Review of David Marshak's The common vision: Parenting and educating for wholeness. Encounter. 1 1,556. Miller, P., Kessel, F., & Flavell, J. (1970). Thinking about people thinking about people thinking about... : A study of social cognitive development. Child Development. 41,613-623. Moore, C. (1996). Theories of mind in infancy. British Journal of Develo~mentalPsvcholom. 14, 19-40. Moore, C., Pure., K, & Furrow, D. (1990). Children's understanding of the modal expression of speaker certainity and uncertainty and its relation to the development of a representational theory of mind. Child Development. 61,722-730. Morrison, K. (1994). The self. In S. Guttenplan (Ed.), A com~anionto the philosophy of mind @p. 550-558). Cambridge, MA: Blackwell. Moshman, D. (1995). Reasoning as self-constrained thinking. Human Development. 38,53-64. Moses, L., & Chandler, M. (1992). Traveler's guide to children's theories of mind. Psvcholo~cal Inquiry. 3,286-301. Mugny, G., & Carugati, F. (1989). Social representations of intelligence. New York: Cambridge University Press. Murgatroyd, S., & Robinson, E. (1997). Children's and adults' attributions of emotions to a wrongdoer: The influence of the onlooker's reaction. Comition and Emotion. 11,83-101. Nagel, T. (1986). The view from nowhere. New York: Oxford University Press. Nelson, K. (1996). Lanrmage in co-mitive development: Emergence of the mediated mind. New York: Cambridge University Press. Nelson, K., Plesa, D., & Henseler, S. (1998). Children's theory of mind: An experiential interpretation. Human Development. 41,7-29. Newton, P. (1994). Preschool prevarication: An investisation of the cognitive ~rereauisitesfor deception. Unpublished doctoral dissertation. University of Portsmouth, Portsmouth, UK. Noddings, N. (1984). Caring: A feminine a~~roachto ethnic and moral education. Berkeley, CA: University of California. Noddings, N. (1992). Gender and the curriculum. In P. Jackson (Ed.), Handbook of research on cuniculum: A project of the American Educational Research Association (pp. 659-684). New York: Macmillan. Nolen-Hoeksema, S., Seligman, M., & Girgus, J. (1991). Sex differences in depression and explanatory style in c'nildren. Journal of Youth and Adolescence. 20,20-30. Norman, W. (1963). Toward an adequate taxonomy of personality attributes: Replicated factor structure in peer nomination personality ratings. Journal of Abnormal and Social Ps~c~o~o~.66,574-583. Nucci, L., & Nucci, M. (1982). Children's social interactions in the context of moral and conventional transgressions. Child Development. 53,403-412. Nunnally, J. (1978). Psvchometric theory. New York: McGraw-Hill. Oatley, K. (1992). Best laid schemes: The ~svcholozvof emotions. New York: Cambridge University Press. Oatley, K, & Jenkins, J. (1996). Understanding emotions. Cambridge, MA: Blackwell. Ochs, E., & Schieffelin, B. (1984). Language acquisition and socialization: Three developmental stories. In R. Shweder & R. LeVine (Eds.), Culture theory: Mind. self and emotion @p. 276-320). Cambridge, UK: Cambridge University Press. Offer, D., Ostrov, E., Howard, K., & Atkinson, R. (1988). The teenaye world: Adolescents' self-image in ten countries. New York: Plenum. Olson, D. (1994). The world on paper. Cambridge, UK: Cambridge University Press. Olson, D. (1997). Critical thinking: Learning to talk about talk and text. In G. Phye (Ed.), Handbook of academic learning;,(pp. 493-510). New York: Academic Press. Olson, D., & Bruner, J. (1996). Folk psychology and folk pedagogy. In D. Olson, & N. Torrance (Eds.), Handbook of education and human develo~ment:New models of learning. teaching and schooling @p. 9-27). Oxford, UK: Blackwell. Olson, D., & Homer, B. (1998). How children create knowledge. Cognitive Develo~ment. 13.249-255. Ontario Ministry of Education and Training (1993). Common curriculum. Toronto, ON: Ministry of Education. Ontario Ministry of Education and Training (1994). Ensenderins equitv: Transforming;- curriculum. Unpublished draft. Ontario Ministry of Education and Training (1997). The Ontario Curriculum. Grades 1-8, Language. Toronto, ON: Ministry of Education. Oppenheimer, L. (1989). The nature of social action: Social competence versus social conformism. In B. Schneider, G. Attili, J. Nadel, & P. Weissberg (Eds.), Social competence in deveio~mentalwrspective (pp. 41-69). London: Kluwer Academic Publishers. Paley, V. (1984). Bovs and Qrls: Superheroes in the doll comer. Chicago, IL: Chicago University Press. Park, L., & Park, T. (1997). Personal intelligence. In M. McCallum & W. Piper (Eds.), cal rnindedness: A contemporary understanding (pp. 133-168). Mahwah, NJ: Erlbaum. Parkhurst, J., & Asher, S. (1992). Peer rejection in middle school: Subgroup differences in behavior, loneliness, and interpersonal concerns. Developmental Psvcholoev. 28, 231-241. Patterson, C., Kupersmidt, J., & Griesler, P. (1990). Children's perceptions of self and of relationships with others as a function of sociometric status. Child

Peabody, D. (1967). Trait inference: Evaluative and descriptive aspects. Journal of Pesonality and Social Psvcholo_gy Monomaph. 7 (4, Serial No. 644). Peeves, F., & Sword, P. (1973). Developmental changes in attribution of descriptive concepts to persons. Journal of Personality and Social Psvchology. 27,120-128. Pekrun, R. (1990). Social support, achievement evaluations, and self-concepts in adolescence. In L. Oppenheimer (Ed.), The self-conce~t:Euro~ean ~ers~ectives on its develo~ment, aspects and applications (pp.107-119). Berlin, Germany: Springer. Pellea-, A. (1995). Social recess and plavground behaviour. New York: Cambridge University Press. Pellegrini, D. (1985). Social cognition and competence in middle childhood. Child Development, 56.253-264. Perner, J. (1988). Higher-order beliefs and intentions in children's understanding of social interaction. In J. Astington, P. Harris, & D. Olson (Eds.), Develo~ingtheories of mind @p. 27 1-294). Cambridge, MA: Cambridge University Press. Perner, J. (1991). Understanding the representational mind. Cambridge, MA: BradfordIMIT. Perner, J. (1996). Simulation as explicitation of predication-implicit knowledge about the mind: Arguments for a simulation-theory mix. In P. Carruthers and P. Smith (Eds.), Theories of theories of mind (pp. 90-104). Cambridge, UK: Cambridge University Press. Perner, J., & Wimmer, H. (1985). "John thinks that Mary thinks that...": Attribution of second- order beliefs by 5- to 10-year-old children. Journal of Exmimental Child Psvcholog7r, a 437-471. Perner, J., Ruffman, T., & Leekam, S. (1994). Theory of mind is contagious: You catch it from your sibs. Child Develo~ment.65,1228-1238. Perry, W. (1970). Forms of intellectual and ethical development in the college vears. New York: Hoh, Rinehart & Winston. Peterson, G., & Leigh, G. (1990). The family and social competence in adolescence. In T. GulIotta, G. Adams, & R. Montemayor (Eds.), Developing social competencv in adolesceace (pp. 97- 138). Newbury Park, CA: Sage Publications. Petersen, A., Sanigiani, P., & Kennedy, R. (1991). Adolescent depression: Why more girls? Journal of Youth and Adolescence, 20,247-27 1. Phinney, J. (1990). Ethnic identity in adolescents and adults: A review of the research. Psvcholog;ical Bulletin. 108,499-514. Pipher, M. (1994). Reviving Ophelia: Savine the selves of adolescent girls. New York: Ballantine Books. Piaget, J. (1929). The child's conception of the world. New York: Harcourt, Brace & Janovich. Piaget, J. (1963). The ori~insof intellicence in children. New York: Norton. Piaget, J. (1965). The moral judgement of the child. New York: Free Press. Piaget, J., & Inhelder, B. (1956). The child's conception of mace. London, UK: Routledge & Kegan Paul. Postman, N. (1995). The end of education: Redefining the value of school. New York: Vintage Books. Potter, E. (1993). Gender and epistemic negotiation. In L. Alcoff & E. Potter (Eds.), Feminist epistemologies @p. 161-186). New York: Routledge. Powlishta, K. (1995). Gender bias in children's perceptions of personality traits. Sex Roles. 32. 17-28. Pratt, C., & Bryant, P. (1990). Young children understand that looking leads to knowing (So long as they are looking into a single barrel). Child DeveloDIIIent. 61,973-982. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences. 1,5 15-526. Raver, C., & Leadbeater, B. (1993). The problem of the other in research on theory of mind and social development. Human Development. 36,350-362. Raver, C., & Leadbeater, B. (1995). Commentary. Human Develo~ment.38,190-193. Reis, S. (1987). We can't change what we can't reco,@ze: Understanding the special needs of fled femaies. Gifted Chiid Ouarteriv. 5 1.83-88. Reinsmith, W. (1989). The whole in every part: Steiner and Waldorf schooling. The Education Forum. 54,79-9 1. Resnick, L. (1991). Shared co,snition: Thinking as social practice. In L. Resnick, J., Levine, & S. Teasley (Eds.), Perspectives on sociallv shared cornition, @p. 1-20). Washington, DC: American Psychological Association. Richards, M. (1980). Toward wholeness: Rudolf Steiner education in America. Middleton, CT: Wesfeyan University Press. Richards, M., Crowe, P., Larson, R., & Swan; A. (in press). Developmental patterns and gender differences in the experience of peer companionship during adolescence. Child Develoiment. Richardson, R., & Evans, E. (1997). Connecting with others: Lessons for teaching: social and emotional competence. Waterloo, ON: Colwell Systems/Research Press. Riger, S. (1992). Epistemological debates, feminist voices: Science, social values, and the study of women. American Psvcholog:ist. 47,730-740. Roberts, W., & Strayer, J. (1996). Empathy, emotional expressiveness, and prosocial behaviour. Child Develo~ment.67,4491170. Rogoff, B. (1990). renticeship in thinking Cornitive development in social context. New York: Oxford University Press. Rosenberg, M. (1989). Society and the adolescent self-image (Rev. ed.), Middletown, CT: Wesleyan University Press. Rosenthal, R., & Jacobsen, L. (1969). Pymalion in the classroom: Self-fulfilling: prophecies and teacher expectations. New York: Holt, Rinehart & Winston. Rozak, T. (1992). The voices of the earth. New York: Simon & Schuster. Rubin, K. (1972). Relationship between egocentric communication and popularity among peers. Develo~mentdPsycholow. 7,364. Rubin, K, & Asendorpf, J. (1993). Social withdrawal, inhibition, and shyness in childhood: Conceptual and definitional issues. In K. Rubin, & J. Asendorpf (Eds.), Social withdrawal. inhibition and shyness in childhood @p. 3-17). Hillsdale, NJ: Lawrence Erlbaum. Rubin, K, LeMare, L., & Lollis, S. (1990). Social withdrawal in childhood: Developmental pathways to peer rejection. In S. Asher & J. Coie (Eds.), Peer rejection in childhood (pp. 217-249). Cambridge, UK: Cambridge University Press. Ruble, D., & Dweck, C. (1995). Self-conceptions, person-conceptions, and their development. In N. Eisenberg (Ed.), Social development @p. 109-139). Thousand Oaks, CA: Sage. Rudolph, K., Hammen, C., & Burge, D. (1995). Cognitive representations of self, family and peers in school-age children: Links with social competence and sociometric status. Child Development. 66,1385-1402. Saarni, C. (1988). Children's understanding of the interpersonal consequences of dissemblance of nonverbal emotional-expressive behavior. Journal of Nonverbal Behavior. 12,275-294. Schaffer, R. (1996). Social development. Oxford, UK: Blackwell. Schank, R., & Cleary, C. (1995). Making machines creative. In S. Smith, T. Ward, & R. Finke (Eds.), The creative co~itiveapproach (pp.229-247). Cambridge, MA: MIT Press. Schultz, L., & Selman, R. (1989). Bridging the gap between interpersonal thought and action in early adoiescence: The role of psychodynamic processes. Development and Psvchopathologv. 1,133- 152. Schwab, J. (1983). The practical 4: Something for curriculum professors to do. Curriculum Inquiry. 13,239-265. Schwanenflugel, P., Fabricius, W., & Noyes, C. (1996). Developing organization of mental verbs: Evidence for the development of a constructivist theory of mind in middle childhood. Comitive Development. 11,265-294. Schwanenflugel, P., Fabricius, W., Noyes, C., Bigler, K, & Alexander, J. (1994). The organization of mental verbs and folk theories of knowing. Journal of Memorv and Language. 33,376-395. Schwanenflugel, P., Fabricius, W., Alexander, J. (1994). Developing theories of mind: Understanding concepts and relations between mental activities. Child Develo~ment, 65. 1546-1563. Searle, J. (1969). Smch acts: An essav in the philosophv of lanrmage. Cambridge, UK: Cambridge University Press. Selman, R. (1980). The mowth of interpersonal understanding. New York: Academic Press. Selman, R. (1989). Fostering intimacy and autonomy. In W. Damon (Ed.), Child development today and tomorrow (pp. 201-233). Mahwah, NJ: Lawrence Erlbaum. Shantz, C. (1983). Social cognition. In P. Mussen (Ed.), Handbook of child psychology: Vol. 3. (pp. 495-555). New York: Wiley. Shavelson, R., Hubner, J., & Stanton, G. (1976). Self-concept: Validation of construct interpretations. Review of Educational Research. 46,407-441. Shields, S. (1991). Gender in the psychology of emotion: A selective research review. In K Stron,oman (Ed.), International review of studies on emotion (Vol I), (pp. 227-245). West Sussex, UK: John Wiley & Sons. Siddiqui, S. (1995). Cognitive consequences of exposure to print: A studv of decontextualization and knowled_geof mental state verbs. Unpublished doctoral dissertation. University of Toronto, Toronto, ON. Silverman, L. (1995). The moral sensitivity of gifted children and the evolution of society. Roeper Review. 17, 110-1 15. Silverstein, B., & Perlick, D. (1995). The cost of competence: Whv ineaualitv causes de~ression, eatin9 disorders. and illness in women. New York: Oxford University Press. Simmons, R., & Rosenberg, F., & Rosenberg, H. (1973). Disturbance in the self-image at adolescence. American Sociological Review. 38,533-568. Sinetar, M. (199i). Develo~inea 21st century mind. New York: Vuard. Slaby, R., & Guerra, N. (1988). Cognitive mediators of aggression in adolescent offenders: 1. Assessment. Develo~mentdPsvcholoqr. 24,580-588. Slomkowski, C., & Dun, J. (1996). Young children's understanding of other people's beliefs and feelings and their co~ectedcommunication with fiiends. Developmental Psvcholop, 2442-447. Smetana, J. (1993). Understanding of social rules. In M. Bennett (Ed), The child as psvchologist: An introduction to the development of social comition @p. 111-141). New York: Harvester Wheatsheaf. Smilansky, S. (1968). The effects of sociodramatic ~lavon disadvantaged re school children. New York: Wiley. Smith, D. (1990). The conce~tual~ractices of ~ower:A feminist sociologv of knowledge. Toronto, ON: University of Toronto Press. Sperber, D., & Wilson, D. (1986). Relevance: Communication and comition. Cambridge, MA: Harvard University Press. Sprague, D., Beauregard, D., & Voelker, S. (1987, April). Person perce~tionin children: The effects of age. sex. and popularity. Paper presented at the biennial meeting for the Society or Research in Child Development, Baltimore, MD. Steiner, R. (1976). Education and the child in the light of anthroposophy, G. Adam & M. Adams (Trans.). London, UK: Rudolf Steiner Press. Sullivan, H. (1953). The interpersonal theory of psvchiatry. New York: Norton. Sullivan, K., Zaitchik, D., & Tager-Flusberg, H. (1994). Preschoolers can attribute second-order beliefs. Developmental Psvcholop. 30,395-402. Sulloway, F. (1996). Born to rebel: Birth order. familv dvnamics and creative lives. New York: Vintage Books. Swanson, H. (1992). The relationship between metacognition and problem solving in gifted children. Roe~erReview. 15,4348. Szagun, G., & Schauble, M. (1997). Children's and adults' understanding of the feeling experience of courage. Cognition and emotion. 11,291-306. Tager-Flusberg, H., Sullivan, K., Barker, J., Harris, A., & Boshart, J. (1997, April). Theorv of mind and language acquisition: The devel ment of cognition verbs. Poster presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Tangney, J. (1991). Moral affect: The good, the bad and the ugly. Journal of Personality and Social Psvcholow. 61,598-607. Tannen, D. (1994). Gender and discourse. New York: Oxford University Press. Tavris, C. (1992). The mismeasure of women. New York: Simon & Schuster. Taylor, J.; Gilligan, C., & Sullivan, A. (1995). Between voice and silence: Women and girls, race and relationship. Cambridge, MA: Harvard University Press. Tomasello, M., Kruger, A., & Ratner, H. (1993). Cultural learning. Behavioural and Brain Sciences. 16,495-552. Trevarthen, C. (1979). Instincts for human understanding and for cultural comprehension: Their development in infancy. In M. von Cranach, K. Foppa, W. Lepenies, & D. Ploog (Eds.), Human etholom: Claims and limits of a new disci~iine(pp. 530-571). Cambridge, MA: Cambridge University Press. Turiel, E. (1983). The develo~mentof social knowledse: Morality and convention. Cambridge, UK: Cambridge University Press. Unger, R., & Crawford, M. (1992). Women and gender: A feminist psvcholo~~.New York: McGraw-Hill. Urberg, K. (1982). The development of the concepts of masculinity and femininity in young children. Sex Roles. 8,659-668. Veith, D. (1980). Recursive thinking and the self-concepts of preadolescents. The Journal of Genetic Psvcholom. 137,233-246. Vernon, A. (1997). Thinking. feeling and behaving: An emotional education curriculum for adolescents. Waterloo, ON: Colwell SystemIResearch Press. Verschueren, K., Marcoen, A., & Schoefs, V. (1996). The internal working model of the self, attachment, and competence in five-year-olds. Child Development. 67,2493-25 11. Von Eckardt, B. (1994). Folk psychology (1). In S. Guttenplan (Ed.), A companion to the philosophv of mind @p. 300-307). Cambridge, MA: Blackwell. Vygotsky, L. (1965). Thought and lanmage. Cambridge, MA: MIT Press. Vygotsky, L. (1978). Mind in sociee. Cambridge, MA: Harvard University Press. Waldman, I. (1996). Aggressive boys' hostile perceputal and response biases: The role of attention and impulsivity. Child Development. 67,1015-1033. Weber, S., & Mitchell, C. (1995). 'That's funnv. vou don't look like a teacher.' Interro~ating- images and identity in popular culture. London, UK: Falmer Press. Weiner, B. (1985a). "Spontaneous" causal thinking. Psycholo~calBulletin. 97,7444. Weiner, B. (198%). An attributional theory of motivation and emotion. Psvcholotzical Review, 92,548-573. Wellman, H. (1990). The child's theory of mind. Cambridge, MA: MIT Press. Wentzel, K., & Asher, S. (1995). The academic fives of neglected, and rejected, popular, and controversial children. Child Development. 66,754-763. Werner, H. (1948). Comparative psvchoIow of mental development. New York: The Free Press. Werner, R., & Cassidy, K (1997, April). Children's psvchological understandine and its relationship to social information processinp and social competence. Poster presented at the biennial meeting of the Society for Research in Child Development, Washington, DC. Wertsch, J. (1989). A sociocultural approach to mind. In W. Damon (Ed.), Child development today and tomorrow (pp. 14-33). San Francisco, CA: Jossey Bass. Whiten, A., & Byrne, R (1988). The Machiavellian intelligence hypotheses: Editorid. In R. Byrne & A. Whiten (Eds.), Machiavellian intelligence: Social expertise and the evolution of intellect in monkevs. apes. and humans (pp. 1-9). Oxford, UK. Blackwell. Wilgosh, L. (1994). The underachievement of girls: A societal rather than a gender issue. Education in Canada Sprin, 18-23. Williams, B. (1973). Problems of the self: Philoso~hicalpapers 1956-1972. Cambridge, MA: Cambridge University Press. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cornition. 13,103-128. Wimmer, H., & Hartl, M. (1991). Against the Cartesian view on mind: Young children's difficulty with own false beliefs. British Journal of Developmental Psvcholow, 9L, 125-138. Winner, E. (1988). The ~ointof words: Children's understanding:of metaphor and ironv. Cambridge, MA: Harvard University Press. Winner, E. (1996). Gifted children: Mvths and realities. New York: Basic Books. Winner, E., & Leekam, S. (1991). Distin,pishing irony from deception: Understanding the speaker's second-order intention. British Journal of Developmental Psvcholow, 9.257-270. Yuill, N. (1992a). Children's conception of personality traits. Human Development. 35,265-279. Yuill, N. (1992b). Children's production and comprehension of trait terms. British Journal of Develo~mentalPsvcholow. lO,l3 1-142. Yuill, N. (1993). Understanding of personality and dispositions. In M. Be~ett(Ed.), The child as psvcholo~st:An introduction to the develo~mentof social cognition- @p. 87-1 10). New York: Harvester Wheatsheaf Yuill, N., Perner, J., Pearson, A., Peerbhoy, D., & van den Ende. (1996). Children's changing understanding of wicked desires: From objective to subjective to moral. British Journal of Develo~mentdPsvchologv. 14,457-475. Zahn-Waxler, C. (1996). Environment, biology, and culture: Implications for adolescent development. Develo mental Psvcholoq. 32,571-573. Zahn-Waxler, C., & Robinson, J. (1995). Empathy and guilt: Early origins of feelings of responsibility. In J. Tangney & K. Fischer (Eds.), Self-conscious emotions: The psvcholoq of shame. milt. embanassment. and pride (pp. 143-173). New York: Guilford. Zahn-Waxler, C., Cole, P., & Barrett, K. (1991). Guilt and empathy: Sex differences and implications for the development of depression. In J. Garber & K. Dodge (Eds.), The develo~mentof emotion redations and dvsreeulation @p. 243-272). Cambridge, UK: Cambridge University Press. Zahn-Waxler, C., Radke-Yanow, M., & Brady-Smith, J. (1977). Social perspective-taking and prosocial behaviour. Developmental Psycholo~.13,87-88. Zarbatany, L., Van Brunschot, M., Meadows, K., & Pepper, S. (1996). Effects of friendship and gender on peer group entry. Child Development. 67, 2287-2300. APPENDIX A

Coding Guides and Examples of Transcripts APPENDIX A1

Coding Guide and Examples of Transcripts for Social and Self-Understanding Interview (Part a) - Interpersonal or Social Understanding Coding Guide for Social Understanding Stories (NancylMargie; KennylMark)

A) Third Order Conceotual Role-Takina Questions (Questions 1- 4b)

Questions 1 - 3a

Score of 1 if communicative intent or any mental state verb is mentioned.

Question 1 e.g.: "They were going to do something to her"; "They planned to pick him for the team."

Question 2 e-g.: "Maybe that's the signal saying"; "He agreed with Mark"

Question 3 a) e.g.: "They're going over to talk to her"; "Because they felt sorry for him"

Score is 0 if: a) Response is "I don't knowlcan't explain."

b) Response is a mere repetition or paraphrase of part of the story, "He's usually picked last.", "Because she's new.", "They're friends."

Question 3 b) Justification - Why do ou think thisIHow do you know?

Cateaory Score Example

1. Tangential 0 "I don't know." "I can't explain."

2. Situational1 Behavioural

Response includes only a behavioural or situational explanation, includes direct reference to a) the story or to b) selflpersonal experience but does not refer to mental states or intentional relations; focuses on phvsical action a) Story Reference 1 "Because the story said."

"Because the swingset's there and the girls are on it."

"Because they're smiling,"

"Because he's usually picked last." b) Self Experience 1 "That's what happens to me sometimes."

"Because that's what usually happens."

3. Psychological

Response includes reference to mental states as illustrated by mental state verbs or emotions; such terms imply understanding of intentional relations. - may include references to behaviours that imply intention (e-g., "They were smiling at each other and winking to each other.")

- may include references to communication (e.g., "Because they have never spoken to her.") and perception (e.g., "Because she sees them walking towards her.")

- may include reference to either story or personal experience a) Story Reference 2 "Because she would probably feel left out."

"He doesn't know why he got picked for the team." b) Self Experience 2 "It's something I would do to be nice."

"Because I would feel sorry for for him."

4. Integrated Psychological

Response includes the integration of multiple perspectives including one's own and others' with aspects of the story (integration of personal experience with story).

Intentional actions are integrated with mental stateslfeelings and may also involve moral judgementslreferences to social norms. integration of 3 "Because it's happened to self experience me before and I've seen people who are left out and are the last ones to be picked and you don't want them to feel bad, you don't want to put them on the spot."

Moral judgment 3 "Because it's probably the right thing to do and it would probably be wrong to go over and be mean to her because she's the new girl and you shouldn't do it anyway."

Intentional action1 Mental state "Because when Kenny and Mark were sitting there looking at him weird he didn't know what they were thinking." Question 4 a\

Score

3 = Maybe/Might have

* Some participants said NO at first but then once they started to justify their answer, changed their response - score the final response.

* Some participants didn't answer the YESINO but went directly into the response of "SIHe was thinking..." - score this as YES.

Question 4 b) Justification

Refer to Question 3b

Examples Justifications for Yes and Mavbe Responses to 4 a

Cateaory Score Example

1) Tangential 0 "I don't know"

"I can't explainu

2) Situational1 1 "Because they're smiling/nodding." Behavioural "Because Tom is usually picked last."

"Because they're friends."

3) Psychological 2 "Because she thinks that they want to talk to her."

"Because he knows that they want him on their team." 4) Integrated 3 "If people chose him last before Psychological then he probably has the feeling that they're going to make fun of him again - if that's what people did before."

"She might be thinking if they might want to make friends or they might want to buy her something so she might be unsure about what she is thinking."

Exam~leJustifications for No Res~onsesto 4 a

Cateaory Score Example

1) Tangential 0 "I don't known

"I can't explainu

2) Situational1 1 "Because the story says." Behavioural "Because all they did was wink and smile."

3) Psychological 2 "Because she didn't talk to them."

"Because he just saw Mark and Kenny winking and nodding at each other." 4) Integrated 3 "Because he just saw them Psychological winking and smiling and he doesn't know what they are thinking."

"Because she hasn't talked to them before and she doesn't know if they are good or bad people."

** Note: Sometimes response to 4a and b become integrated - if this happens, just code 4a and b based on content (not order of response).

Total Score = sum of questions 1, 2, 3a, 3b maximum score is 6, scores range from 0-6. If question 4b is included, maximum score is 9, scores range from 0 - 9.

Cateaorical Codes for Section A

Cateaorical Variable for Question 3a - Reference To StorvlSeIflConvention

- a score of 1 in each category if the response includes a reference to:

STORY - any direct reference to the story , e.g., "She's new to the school."; "They were winking and nodding."

SELF - reference to personal experience, e.g., ''This has happened to me before.", "That's something that I would do."

CONV - reference to convention1 social norms, e.g.,. "That's what usually happens."; "Normally, ..."

7 - "I don't known Cateaorical Codina of Question 3a: Emotional Valence of Scenario

Score Exam~le

1- positive scenario "They are going over to make friends."; "He picked him because he felt sorry for him."

2 - negative scenario "They are going over to be meanltease the new girl."; "He picked Tom because they know he's a bad player and they'll make fun of him."

3 - neutral scenario "They are walking towards the new girl to go on the swings."; "They were going to talk tolmeet with her."

B) Em~atheticSensitivitv (Question 5) a) Descriptive emotive words

Score Response

0 "I don't know."

1 Use of simple emotions such as mad, glad, bad, sad, happy, scared, upset.

2 a) Use of psychologicallcognitive and identity related emotions such as shy, confused, uncertain, solitary, nervous, left-out, confused, alone, unwelcomed, nervous, excited.

These emotions are more complex than the basic emotions but do not imply the self-reflection and inner conflict described in the next level of the self-conscious emotions. b) Combination of two different emotions placed in temporal order, or two-part behavioral event involving two distinct feelings, "She felt happy at first then sad." "He would feel sad if they tease him but happy if they treat him nice."

Use of complex/moral and/or self-conscious emotions such as guilty, embarrassed, vulnerable, rebellious, mystified, self- conscious.

Also includes recursive identity related emotions including phrases and emotions that imply self-reflection such as "She felt good about herself - self-confident."

Simultaneous occurrence of two contradicting emotions (mixed valence) "She felt happy but also nervous." or "Half of him was sad and half of him was happy."

Coherent combination of three or more psychological emotions, e.g., "She felt nervous, confused and shy."

b) Justification (Response to Why?)

Cateaory Score Example

Tangential 0 "I don't know."

Situationall 1 "Because he got picked." Behavioural "Because she's new."

Psychological 2 "Because she's all alone with no one to play with." "She's wondering why they're coming over."

"He doesn't know why he got picked for the team."

2) b) Reference to two conflicting feelings but simple explanations.

"He felt good because he was picked but bad because he wasn't."

Integrated 3 "She must feel sad because I've Psychological been a new kid before and I know that you can feel pretty lonely."

"I would feel bad if I was picked over some guys who I knew were better than me." b) Complex explanation of multiple story lines.

"If they were coming to bug her she might feel teased and unwanted but if they were coming to make friends she might feel really good about herself because she's making friends with people."

Total Score = sum of questions 5 a) and b). Maximum score is 6, range 0-6. Cateaorical Codes for Section B

Cateaorical Coding of Question 5a - Valence of Emotions

Score = Raw count of emotions in each category but if same emotion is repeated, only count once.)

BPOS - Number. of positive emotions (good, happy, excited, wanted, needed, hoping, special relieved, pleased, confident, thankful, lucky).

BNEG - Number of negative emotions (bad, sadidepressed, lonely, upset, left-out, anxious/worried, disappointed, shyltimid, confused, used, vulnerable, discouraged, embarrassed, bitter, glum, rebellious, nervous, shakey, suspicious).

BNEUT - Number of neutral emotions (inbetween feeling, wondering, feels nothing, curious, surprised, mysterious, strange, athletic).

- any emotion that is not obviously negative or positive (according to various research) was rated as neutral.

C) Person Perce~tion(Question 6)

Cateaory Score Example

Tangential 0 "I don't know."

Behaviourall 1 Response includes the use of a) Situational physical/behavioural descriptions such as "He was good at soccer.", or "She was blonde", or "He has friends."

b) stereotypical trait phrases or cliches such as "She was nice", or "He was bad." Psychological 2 Response includes the use of: a) a psychological traitfmental state attribution trait such as shy, conceited, thoughtful, respectful, faithful.

b) intentional states such as "She wants to be friends; he likes to help people."

c) combination of stereotypical personality cliches and trait characteristics such as "nice and popular".

d) simple unelaborated contradition such as, "He might be nice but might not be."

e) Reference to behavioural comparisons, "He was better at soccer than everyone else."

Integrated 3 Response includes the use of a) Psychological self-reflexive or recursive mental states, such as "She thinks she's conceited", or "He knows that he's mean" "She has low self-esteem"; "He doesn't care what other people think."

b) complex and/or contradictory psychological traits such as "He was both mean and nice because part of him wanted to be friends with Tom but part of him wanted

-nbn &.a- -6 k:rn I' rrr LU IclanG IUII UI IIIIII. UI "He was one of those people who expects and assumes that things will happen."

c) Reference to psychological comparisons, "She was much more sensitive than the average person."

Total Score = score on question 6. Maximum score is 3, range 0-3.

Cateaorical Codina for Question 6 - Valence of Personality Traits/Descriptions

Score = Raw count of traits in each category but if same trait is repeated, only count once.

CPOS - Number of positive traits (e.g., nice, kind, trustworthy, responsible, understanding, considerate, caring, loving, respectful, outgoing, leader, faithful, enthusiastic)

CNEG - Number of negative traits (e-g., mean, bossy, shyltimid, snobby, rude, bully, troublemaker, deceitful, sneaky, nerdy, show off, follower)

CNEUT - Number of neutral traits (e.g., curious, suspicious, talkative, quiet, normal, unique, competetive, independent, intelligent, tough, loner,)

D) Alternative Explanation (Question 7)

Cateaory Score Example

Tangential 0 a) No

b) Yes, but offers no alternative explanation, e.g., "Yes, there's another way to think about the story but I don't know what it :-II~-1* -..-.Ic.:-. :a U IDIU~II L =ApIaIt I 11. c) Yes, but repeats their response to 3a (i.e., they don't give a different interpretation of the story.)

Situational1 1 Yes and brief explanation with no Behavioral reference to mental states, only physical descriptions. e.g., "The opposite.", "Just different."; "Maybe they are just friends and they like to wink1smiIelnudge. Psychological 2 a) Yes and brief explanation with reference to intentional action.

e.g., "Maybe they were going to be mean.", "Maybe they were going to make friends or something."; "Maybe they want to make fun of the new girl."

b) Yes and brief explanation that includes mental states. e-g., "They'll tease her", "They don't like Tom."

Integrated 3 Yes and offers alternative Psychological explanation referring to concept of intention AND one or more references to mental states (explanation is more complex than for Category 2):

e.g., "Maybe they made a different plan to go over and make friends or something, like she might want to be friends with them or they might want to make friends with her." "The girls might think the new girl is weird and could be really mean and go over to her and say, "Let's go make fun of her," and pick her apart or something."

Total Score = score on question 7. Maximum score is 3, range 0-3.

Cateaorical Codes for Section D

Cateaorical Codina of Question 7 - Emotional Valence of Scenario

Score Example

1 - positive scenario, "They are going over to make friends."; "He picked him because he felt sorry for him."

2 - negative scenario "They are going over to be meadtease the new girl."; "He picked Tom because they know he's a bad player and they'll make fun of him."

3 - neutral scenario "They are walking towards the new girl to go on the swings." Summary Table of Scores

Question Cateaory Maximum Score Ranae

A) Third Order Conceptual Role-Taking 6 Questions

B) Empathetic Sensitivity

C) Person Perception

D) Alternative Explanation

Total Possible Score (without Q 4b) 18 0-1 8 (with Q 4b) 21 0-21 Exam~leof a Transcript Scoring High on the Interpersonal or Social Understanding:Interview

R = Researcher P = Participant, 11 years, 6 months - female.

Social Understanding Story 1

Third Order Conce~tualRole-Taking Why did Nancy smile at Margie? Because they saw the new girl swinging on the swingset. Why did Margie nod? Because she's telling her that she kind of sees her too. Why did Nancy and Margie move off together in the direction of the new girl? Like they might have wanted to tease her or something and they kind have wanted to go together. Why do you think this? How do you know this? Because she's a new person and they've never spoken to her before. Does the new girl have any idea of why Nancy and Margie are walking towards her? Yes, - she might be thinking if they might want to make friends or they might want to buy her about something so she might be unsure about what she is thinking. How do you know that the new girl has an idea of why Nancy and Margie are walking towards her? Because she was wondering why they were approaching her if they have never spoken to her in class.

Em~atheticsensitivitv How do you think the new girl feels? Well if they were coming to bug her she might feel teased and unwanted but if they were coming to make fiiends she might feel like really good about herself because she's making friends with people.

Person Perception Choose a character in the story and describe her. The new girl - well she's new, she probably doesn't have very many friends because she's new and she might want to make fiiends with people in the class and she seems really uneasy about peopie. She might have low self-esteem or something because she doesn't have friends and she really wants to make friends with people and that she might be a person that might not fit in as well as others because she's new and people haven't gotten used to her yet.

D. Alternative Explanations R: Is there another way that you can think about this story? P: Yes, maybe she's planning something to do to them, like they could want to be friends with each other. Like she might want to be friends with them or they might want to make friends with her or it could be the other way around.

SociaI Understanding Stow 2

Third Order Conceutual Role-Taking

Why did Mark smile at Kenny? They might have wanted to tease him or they could have wanted him to feel better about himself by choosing him to be on the team. Why did Kenny nod? Because he's acknowledging that he saw Mark wink. Why did choose Tom to be on the team? Maybe because they wanted to make him feel better. Why do you think this? How do you know this? Because some people can be really generous towards people like they can be really nice and they'll try and help people. Do you think that Tom has any idea of why Kenny chose him to be on the team? He might have, like with the winking and smiling could have planned it a bit better, like to tell him rm going to pick Tom but I don't think that he knows. How do you know this? Because the one boy picked him, like he could of but I just think he might not have known, but he hew that the guy was going to pick him because he was winking but he probably didn't know why exactly.

Em~atheticsensitivity How do you think Tom feels? He could feel real happy that for once he wasn't the last one picked and that somebody's actually noticing and putting him on the team, or he could have felt like maybe they were twins to tease him. R: How would he feel then? P: Maybe kind of used because they were just there to bug him

C. Person Perception R: Choose a character in the story and describe him. P: Kemy - he could have been really nice and generous and concerned about other people maybe he wants to help others, like helpful, considerate, soft-hearted or something when he sees somebody's hurt he might feel that too.

D. Alternative Ex~lanations R: Is there another way that you can think about this story? P: Yes, maybe the boys might have picked him because Tom was a captain and maybe they started being fiiends after this and could have started a friendship. Example of a Transcript Scoring: High on the Interpersonal or Social Understanding Interview

R = Researcher P = Participant, 11 years, 3 months - male.

Social Understanding Stow 1

Third Order Conce~tualRole-Taking Why did Nancy smile at Margie? I think maybe they wanted to talk to her because she hasn't talked to them before. Why did Margie nod? Maybe to say yeah, let's go do it. Why did Nancy and Margie move off together in the direction of the new girl? Well to go talk to her. Why do you think this? How do you know this? Well because it's probably the right thing to do and it would probably be wrong to go over and be mean to her because she's the new ,@rl and you shouldn't do it anyway. Does the new girl have any idea of why Nancy and Margie are walking towards her? I don't think so but she probably hopefully thinks that they want to be new fiiends with her - so she probably has some idea How do you know that the new girl has some idea of why Nancy and Margie are walking towards her? Because she's new to the school and she wants to make new fiiends.

Em~atheticSensitivity How do you think the new girl feels? Well she probably feels alone and shy because she hasn't talked to them and like new to the school and she doesn't want to get involved with the wrong people like they could say something to somebody and they could bite your head off I guess you could say. Does she feel anything else? She might feel disappointed that nobody has talked to her yet.

Person Perce~tion Choose a character in the story and describe her. Nancy - I think that's she's really nice and she's always making lots of friends because she's really nice and people like Nancy because she's got a lot of creativity and imagination and she's an author, well she's not an author yet but she likes to write stories a lot and she's a very nice person and people respect her for that. What do you mean that she's nice? She likes to, like she's not mean to her friends or anything and if there's a problem she tries to compromise and not fight or anything

Alternative Explanations Is there another way that you can think about this story? Well, the girls could be really mean and they could go over to her and like be really mean to her, like they could be nodding because they're like oh, let's go make fun of her then go over and be really mean to her, like pick her apart or something.

Social Understanding Story 2

A. Third Order Conce~tualRole-Taking R Why did Mark smile at Kenny? P: To tell Kenny it's okay to pick Mark. R: Why did Ke~ynod? P: Probably to say yeah, I'll pick him R: Why did Kenny choose Tom to be on the team? P: Because he's never been chosen before and he kind of felt sorry for him. R: Why do you think this? How do you know this? P: Well because, like I said before, it would be the right thing to do, it's kind of, like rd probably do it because I don't think it's right to exclude people from things. R Do you think that Tom has any idea of why K~MYchose him to be on the team? P: Yeah - well he might think wrong because he's never chosen he might think that maybe - he might think that Kemy is going to be mean to him but I don't think Kenny is. R How do you know that Tom has an idea of why Kemy chose him? P: Well because he's never been chosen before and he probably thinks he shouldn't do this because I'm not usually chosen.

B. Em~atheticSensitivity R: How do you think Tom feels? P: He probably feels alone a lot and that people and that he might know that he's not the best but he knows that he's kind of good but people always exclude him from everything so he probably feels really alone and that nobody likes him. R: Does he feel anything else?

P: He might feel like pushed away from the crowd or like in class, like if there's group activities, or if he's in groups like in our classroom people won't ask him for suggestions and if he does give out suggestions people will think oh that's a dumb idea, just keep your mouth shut and stuff like that because I know this happens in a group in our classroom. .

C. Person Perceution R: Choose a character in the story and describe him. P: Mark - because he didn't choose Tom I think he's kind of. ..he thinks he's popular but a lot of people don't hang around him too much and don't like him too much because he excludes people and is mean to some of his friends, and I kind of think that Mark is a troublemaker - he doubledares people and he thinks that people, if they don't know what they're doing they shouldn't be there and stuff. That's about all I can think of.

D. Alternative Exulanations R: Is there another way that you can think about this story? P: Well that the boys were going to play a trick on him, they were going to hit him with the ball or whatever and they were going to make him sit on the bench all the time or make the other players play a trick on him like say oh you can come out and play then say oh no you can't come out. Exam~leof a Transcri~tScoring; Low on the htemrsond or Social Understanding Interview

R = Researcher P = Participant, 12 years, 1 month - female.

Social Understanding; Story 1

Third Order Conce~tualRole-Taking Why did Nancy smile at Margie? Because they're friends. Why did Margie nod? Because she's her friend too. Why did Nancy and Margie move off together in the direction of the new girl? To talk to her. Why do you think this? How do you know this? Because they were smiling. Does the new girl have any idea of why Nancy and Margie are walking towards her? No. How do you know that the new girl doesn't have an idea of why Nancy and Margie are walking towards her? Because she's never spoken to them.

Em~atheticsensitivity How do you think the new girl feels? Shy. Why does she feel shy? Because she's never talked to them.

Person Perce~tion Choose a character in the story and describe her. New Girl - she never tallcs to the other girls, she never says hi or anything.

Alternative Ex~lanations Is there another way that you can think about this story? No. Could there be a different reason why Nancy and Margie are nudging and smiling at each other? P: No. R: So is this the only way to think about the story? P: Yes.

Social Understanding: Story 2

Third Order Conce~tualRole-Taking Why did Mark smile at %MY? Because he wanted Tom to be on the team. Why did &MY nod? To say yes. Why did Kenny choose Tom to be on the team? Because he's good at it Why do you think this? How do you know this? Because they picked him first. Do you think that Tom has any idea of why Kenny chose him to be on the team? No. How do you know that Tom does not have an idea of why &MY chose him? Because Tom is not their friend.

Em~atheticSensitivity How do you think Tom feels? Tom feels good because he thinks that he'll be on the team

Person Perce~tion Choose a character in the story and describe him. Tom - a good person. What do you mean by good person? He's good on the soccer team.

Ntemative Ex~hnations Is there another way that you can think about this story? No. Could there be another reason why Kenny and Mark are winking and smiling at each other? No. R: So is this the only way to think about the story? P: Yes. Example of a Transcript Scoring. Low on the Intemrsonal or Social Understanding: Interview

R = Researcher P = Participant, 11 years, 8 months - male.

Social Understanding: Story 1

Third Order Conce~tualRole-Taking; Why did Nancy smile at Margie? To go over and see the new girl. Why did Margie nod? To say okay, let's go over. Why did Nancy and Margie move off together in the direction of the new girl? To go say hi. Why do you think this? How do you know this? Because they want to be nice to her. Does the new girl have any idea of why Nancy and Margie are walking towards her? No. How do you know that the new girl doesn't have an idea of why Nancy and Margie are walking towards her? Because she wonders what they are doing.

Empathetic Sensitivity How do you think the new girl feels? She feels sad. Why does she feel sad? Because she doesn't know if they are going to come over and talk to her.

Person Perce~tion Choose a character in the story and describe her. Nancy - nice. What do you mean that she's nice? She likes to make new friends and she is friendly. That's about it.

Alternative Explanations Is there another way that you can think about this story? P: They could do or say something mean to her.

Social Understanding Stow 2

Third Order Conce~tualRole-Taking Why did Mark smile at Kenny? To pick Tom for his team. Why did Kemy nod? He said okay. Myciici Kenny choose Tom to be on the team? Because he's usually one of the last people. Why do you think this? How do you know this? Because of the story. Do you think that Tom has any idea of why &MY chose him to be on the team? No. How do you lmow that Tom does not have an idea of why K~MYchose him? Because he just sees them nodding at each other.

Em~atheticSensitivitv How do you think Tom feels? Probably sad because he's one of the last people.

Person Perce~tion Choose a character in the story and describe him. Kenny - a nice person because he told him to pick Tom. Probably has a lot of friends. That's about it.

Alternative Emlanations Is there another way that you can think about this story? Kenny might like Tom but he might pick the better people to try to win. APPENDIX A2

Social and Self-Understanding Interview (Part b) - Intrapersonal or Self-understanding Coding Guide for Self-understanding Interview

Each response to the 6 Self-understanding questions are to be scored according to the following guidelines:

Scale Cateaory Score Example

1. Unable to answer 0 "I don't know." (Tangential)

"I can't explain it."

2. Self-Description Cateaorical Identifications a) Does not include 1 "I'm tall." explanations or interpretations, only "I fight with my brother." descriptive statements.

"I'm just happy." b) May include simple "I like school/sports." judgments/evaluations about objects or events. c) Simple self-judgements "Because I do good in school." that are not followed by "I don't have very many friends." any explanation or just repeat part of the question. No reference to intention, traits or self-reflection.

3. Self-Theory (Comparative Assessments and inter-personal implications)

- Self is defined in relation to others, self-statements are followed by an explanation. Responses imply the beginning of self-reflection. a) Involves simple, 2 "I'm a natural athlete because I was just born that way." p hysical-causal explanations, often "I'm smart because my report card/ based on behaviours, my mom says." preferences and/or may include simple stereotypical, psychological "I have friends because I'm nice." traits with no interpretation b) May include mental "I like my hair." states and simple self- judgements followed . "I like playing sports and I try real by an intentional act. hard."

"I do good in school because I study a lot."

"I have a lot of friends because I share with them." c) May refer to comparisons "I'm the best on the team." between self and other (implicit or explicit "Everyone plays better than I do." comparisons). "I'm smart." "I'm athletic." d) Ablilities are considered "People sometimes call me a goody-two in light of the reactions shoes and stuff like that because I of others - may include rarely get in trouble in class." approval, disapproval or any affective resonse. "I listen to my parents so I don't get into trouble." 4. lntearated Self-Theory (Svstematic beliefs and plans)

a) Integrated Self 3 "I've got lots of friends because (More sophisticated I try to be nice to people and they response - integrates respond well - I think that's why." psychological aspects - mental states of both self and other).

b) Includes an elaborated "I like being the way I am because I seii-theory including a don't think I'm perfect and i don't want combination of thoughts and to be because they say if you're perfect feelings - may include self- in school then in highschool you'll be conflict, contradiction. the opposite to people. I'm just happy the way I am.

"Because I don't like to make fun of people or anything. I don't try to embarrass people because I've had that done and I don't like the way it feels."

Self-judgment is followed by "I don't know a lot of sports. I don't an explanation based on a really like to try them so I'm not combination of knowledge and really not that good when I start out." affect implying greater self- reflection. "I'm smart because I've never gotten under a "C". "I try my hardest that's the most important, that I try my hardest to make my parents proud."

c) May include psychological "Because it's not the outside that comparisons and/or judgments counts, it's the inside and I feel okay based on social norms. about the way I look. To some people good looks are all that matters - I don't care."

Total Score = sum of questions 1-6. Maximum score is 18 (6 x 3)' scores tange itom G-18. Cateaorical Codes for Self-Understandinq

Cate~oricalVariable - Global Self-worth - main reason why participants are happy to be who they are or wish they were different

friends, social "Because I have a lot of friends." school, "Because I get good grades." family, "Because I have a good family." self - mentions specific characteristic of self (e.g., "because I'm pretty"; "because I'm mature." multiple reasons (e.g., because I have lots of friends, good family and do well at school." No main reason (e.g., "I just like being me"; "I just want to be someone else") "I don't know"

Cateaorical Variable - Cateaorical Self

Scale Cate~ory Score Example

1. Personal Agency 1 "I'm good at school because I - sense of self is usually finish all of my derived from self- assignments." judgements. "I look in the mirror and I like what I see."

"I know that I do well in sports."

2. Significant Others 2 "My report card says I do well." sense of self is derived from judgements "My parents say that I am pretty." of significant others (peer, family, teacher) "My friends say that I am a really nice and a good friend."

3. Combination (Self 3 "I'm smart in school because I try hard and Other) and my teacher says I can do anything +h + em..+ ,112,. ,112,. ..vvu1 I,.. 4. Unable to answer 7 "I don't know."

"I don't know why I have so many friends." "I can't explain."

Cateaorical Variable - Definition of Self-Esteem

Response to question, "In your own words, what does self-esteem mean to you." If no answer or don't know I prompted with "What's the first thing that comes to your mind?"

Score Example

Response includes aspect of self (e.g., "How you thinwfeel about yourself or how you act."). Reflects an objective perspective on self (i.e., the "cultural eye")

Response includes self and other (e.g., "It's the way you think about yourself and the way you treat others, you're always putting people down to bring yourself up.")

Response includes personal experience/example (e.g., "I know that I have high self-esteem because I always think I can do something."). Reflects a subjective perspective on self (i.e., the private or "mind's eye").

"I don't know"; "I've never heard the term self-esteem before." APPENDIX B

Instruments APPENDIX B 1

Self-perception Profile for Children (SPPC;Harter, 1985) Self-perception Profile for Children (SPPC) (Harter, 1985)

Administration and Instructions The scale may be administered in groups as well as Individually. After filling out the information at the top of the scale, children are instructed as to how to answer the questions, given below. We have found it best to read the items outloud for 3rd and 4th graders, whereas for 5th graders and older, they can read the items for themselves, after you explain the sample item. Typically, we introduce the scale as a survey and, if time, ask the children to give examples of what a survey is. They usually generate examples involving two kinds of toothpaste, peanut butter, cereal, etc. to which you can respond that in a survey, there are no right or wrong answers, its just what you think, your opinion. In explaining the question format, it is essential that you make it clear that for any given item they only check one box on either side of the sentence. They do not check both sides. (Invariably there will be one or two children who will check both sides initially and thus you will want to have someone monitor each child's sheet at the onset to make certain that they understand that they are only to check one box per item.)

I INSTRUCTIONS TO THE CHILD: We have some sentences here and, as you can see from the top of your sheet where it says "What I am like," we are interested in what each of you is like, what kind of a person you are like. This is a survey, not a test. There are no right or wrong answers. Sincekids are very different from one another, each of you will be putting down something different. First let me explain how these questions work. There is a sample question at the top, marked (a). 1'11 read it outloud and you follow along with me. (Examiner reads sample question.) This question talks about two kinds of kids, and we want to know which kids are most like you. (1) So, what I want you to decide first is whether you are more like the kids on the left side who would rather play outdoors, or whether you are more like the kids on the right side who would rather watch T.V. Don't mark anything yet, but first decide which kind of kid is most like you, and go to that side of the sentence. (2) Now, the second thing 1 want you to think about, now that you have decided which kind of kids a7.e most like you, is to decide whether that is only sort of true for you, or really true for you. If it's only sor: of true, then put an X in the box under sort of true; if it's really true for you, then put an X in that box, under really true. (3) For each sentence you only check one box. Sometimes it will be on one side of the page, another time it will be on the other side of the page, but you can only check one box for each sentence. You don't check both sides, just the one side most like you. (4) OK, that one was just for practice. Now we have some more sentences which I!m going to read out loud. For each one, just check one box, the one that goes with what is true for you, what you are most like. What I Am Like

Age -Birthday Group -- hiontn Day Boy or Girl (circle which)

SAMPLE SENTENCE

Really Sort of Sort of Really True True True True forme forme for me for me

Some kids would rather Other kids would rather play outdoors in their BUT watch T.V. spare time U

Some kids feel that they Other kids worry about are very good at their BUT whether they can do the school work school work assigned to them.

Some kids find it hard to Other kids find it's pretty make friends BUT easy to make friends. I7

Some kids do very well Other kids don? fed that at all kinds of sports BUT they are very good when it comes to sports.

Some kids are happy Other kids are not happy with the way they look BUT with the way they look.

Some kids often do not Other kids usually like u like the way they behave BUT the way they behave.

Some kids are often Other kids are pretty unhappy with themselves BUT pleased with themselves.

Some kids feel like they Other kids aren't so sure are just as smart as BUT and wonder if they are u as other kids their age as smart.

Some kids have alor of Other k~dsdon't have friends BUT very many friends. 279 Really Sort of Sort of Really True True True True for me for me for me for me

Some kids wish they Other kids feel they are could be alot better at BUT good enough at sports. Cl sports

Some kids are happy Other kids wish their with their height and BUT height or weight were weight different.

Some kids usually do Other kids often don't cl the right thing BUT do the right thing.

Some kids don't like the Other kids do like the way they are leading BUT way they are leading their life their life.

Some kids are pretty Other kids can do their slow in finishing their BUT school work quickly. u school work

Some kids would like to Other kids have as many have alot more friends BUT friends as they want.

Some kids think they Other kids are afraid could do well at just BUT they might not do well at u about any new sports sports they haven't ever activity they haven't tried. tried before

Some kids wish their Other kids like their u body was different BUT - .body the way it is.

Some kids usually act Other kids often don't the way they know they BUT act the way they are are supposed to supposed to.

Some kids are happy with Other kids are often not themselves as a person BUT happy with themselves.

Some kids often forget Other kids can what they learn BUT remember things easily.

Some kids are always Other kids usually do aoing thmgs with aiot BUT thinss by themselves. [71 of kiCs Really Sort of Sort of Really True True True True for me for me for me for me

Some kids feel that they Other kids don't feel are better than others BUT they can play as well. 0 El their age at sports 0 0

Some kids wish their Other kids like their physical appearance (how BUT physical appearance the 0 they look) was different way it is. 0 0

Some kids usually get Other kids usually don't in trouble because of BUT do things that get then 0 0 things they do in trouble. 0 0

Some kids like the kind Other kids often wish of person they are BUT they were someone 0 0 else. 0

Some kids do very well Other kids don't do at their classwork BUT very well at their 0 n classwork. 0

Some kids wish that Other kids feel that most more people their age BUT people their age do like 0 liked :hem them. 0 0

In games and sports Other kids usually play some kids usually watch BUT rather than just watch. instead of play 0 0

Some kids wish Other kids like their face something about their BUT and hair the way they 0 face or hair looked are. 0 different Some kids do things Other kids hardly ever they know they BUT do things they know 0 shouldn't do they shouldn't do. 0 0

Some kids are very Other kids wish they happy being the way BUT were different. they are 0

Some kids have trouble Other kids almost figuring out the answers BUT always can figure out 0 in school the answers. 0 0

Some kids are popular Other kids are not very BUT popular. u with others their age nu n n 28 1

Really - Sort of Sort of Really True True True True for me for me for me for me

Some kids don't do well Other kids are good at 0 0 at new outdoor games BUT new games right away.

Some kids think that Other kids think that they are good looking BUT they are not very 0 0 good looking. n

Some kids behave Other kids often find it n n themselves very well BUT hard to behave 0 themselves.

Some kids are not very Other kids think :he way haopy with the way they BUT they do things is fine. do alot of things

Susan Haner. Ph.D.. University of Denver. 1985 APPENDIX B2

Social and Self-Understanding Interview Student Directions Social and Self4Jnderstandi.g Interview Student Directions Directions to Student

This interview will have two sections. In the first section I am going to read you two short stories about some children and their friends. After I finish reading each story, I will ask you some questions about the characters in the story. In the second section I will ask you some questions about some of your answers from the 'What I Am Like" questionnaire I recently gave you in class. Remember that you are participating in a research study and that this has no co~ectionto your school work. I am interested only in your opinions, there are no right or wrong answers. Please do not forget that you have the right to refuse to answer any of the questions and you can stop the interview whenever you wish. All of your responses will be recorded on audiotape and only myself and another research will have access to these tapes. Your responses will remain anonymous because you will be given a subject number and your name will not appear on any of the test material (tapes or papers). Also, if you have any questions during the interview, please do not hesitate to ask. We will now begin the first section of the interview. I am going to be reading to you two short stories about some children when they are with their fiiends at school. I will then ask you some questions after each story about the characters involved. Remember that there are no right or wrong answers. I am only interested in your interpretation or your thoughts about the story. Ready? APPENDIX B3

Social and Self-understanding Interview (Part b) - Intrapersonal Understanding Intrapersonal or Self-understanding Interview Directions to student

We are now ready to start the second part of the interview. As I mentioned before in your classroom, I am a researcher interested in what kind of people grade six students think that they are like. To find this out I asked you to fill out a questionnaire or a survey (not a test) about what kind of person you think you are like. Sometimes however, questionnaires do not provide researchers with enough information and they need to talk to or interview students to find out more about their thoughts. To help me get a better understanding of how grade sixes think about themselves I am going to ask you some questions about some of your answers from the questionnaire I recently gave you in class. I would like to know the reasons why you answered your questionnaire the way that you did. Therefore, for six of the questionnaire questions I will ask you to explain why you answered that particulai question or what is the main reason that caused you to circle that answer. Remember that I am interested only in your opinions, there are no right or wrong answers. Also, remember that you have the right to refuse to answer any question and to stop the interview at any time. Ready? Further Assessment of Self-Conce~t(Justification of Self-Beliefs) To attain a deeper understanding of the child's self-concept and the reasons for herhis self- judgments, following the administration of the SPPC (Harter, 1985), each child is asked six questions based on the child's responses on the SPPC (one question on each domain - item 25,8, 3,34, 17,30). Each question is framed in the following manner: "What is the main reason for why it is really truelsort of true that you are (item content, e-g., good at your school work; don't have many friends, etc.). (SPPC item 25) What is the main reason for why it is really trudsort of true that you: do very well at your classwork OR don't do very well at your classwork? (SPPC item 8) What is the main reason for why it is really truelsort of true that you: have a lot of friends OR don't have very many friends? (SPPC item 3) What is the main reason for why it is really truelsort of true that you: do very well at all kinds of sports OR don't feel that you are very good when it comes to sports? (SPPC item 34) What is the main reason for why it is really trudsort of true that you: think that you are good looking OR think that you are not very good looking? (SPPC item 17) What is the main reason for why it is really trudsort of true that you: usually act the way you know you are supposed to OR often don't act the way you are supposed to? (SPPC item 30) What is the main reason for why it is really trudsort of true that you: are very happy being the way you are OR wish you were different? APPENDIX B4

Teacher's Rating Scale of Child's Actual Behavior (TRF, Harter, 1985) TEACHER'S RATING SCALE OF CHILD'S ACTUAL BEHAVIOR 287 [Parallels the seli-perceptron profile for chtidreni

Class, gradet'group Rater For each child. please indicate what you feel to be hisiher ac:ual competence on each Guestion. in vour opinion. Firs: decide what kind oi child he or she is like. the one described on the left or right. and then indicate wherher this IS lust son oi true or really true for that individual. Thus. for each item. check one oi four boxes.

Really Sort of Sort of Really True True True True

1. a tf This chid is really OR ' This child can't do good at hisiher the school work school work assigned. 2. I) f This child finds it OR For this child 'it's hard to make friends pretty easv. 3. I_) 0 This child does OR This child isn't really well at all very good when it kinds of sports comes to spans. 4. @ This child is OR This child is not good-looking . verv good-looking. 5. ) f This child is usually OR This child is oi:en E well-behaved not well-behaved.

6. C] This child oiten OR This child can CI forgets what s:he remember things u leains OR easilv. 7. 0 This child has alot This child doesn*: 0 E oi friends . have manv friends.

8. ' C This child is better OR This child can't play than others his:her as well. LI D age at sports 9- f This child has a nice OR This child doesn't a c! physical appearance have such a nice physical appearance. 10. This child usually OR This child would be 0 acts appropriately better if s:he acted diiierentlv. 11- This child has OR .This child almost 0 trouble figuring out alwavs can figure out the answers in the answers. school 12 17 0 This child is popular OR This child is not very [I] ) with others hislher popular. age 73- . This child doesnet OR This child is good at do well at new new games right outdoor games awav. 74. 0 This child isn't OR This child is pretty C] a very good looking good-looking. 15. 17 This child often gets OR This child usuallv 17 I3 in trouble because doesn't do things oi things heishe does that get himrher in trouble.

Suwn Haner Untvenlrv or Denver. 1% APPENDIX B5

Social Competence Nomination Form (adapted fiom Ford, 1982) and Peer Likability Scale (adapted itom Matthews & Keating, 1995) Social Competence Nomination Form (adapted fiom Ford, 1982) Directions: In this questionnaire there are descriptions of three different social situations. I would like you to read each story and for each of the names you have been given, write the number (from 1 to 5) which shows how well you think each person on your list could handle each situation (e.g., 1 = the person would not handle the situation well at all, 5 = the person would handle the situation extremely well). There are no right or wrong answers - I just want to know your opinion. All of your answers will be kept completely confidential - no one else at your school will see them. A. Grou~Assimment Situation Everyone in your grade has been given an assignment to make studying Canadian history more fun. For this assignment, groups of five to ten students must put together a skit which shows some important event in Canadian history. Each group must have a director to organize and coordinate the group's efforts. The director must be able to get your group together and cooperating so that the skit will be a good one. How would you rate this person's ability to be a good director?

B. Student Visitor Situation One of your school's librarians has had an operation and is recovering in the hospital. The students in your grade have gotten together and decided to do something for the librarian. The class decides that someone should make a personal visit to the hospital. This person would bring flowers and tell the librarian how much the students miss her and that they hope to see her back at the school soon. How would you rate this person's ability to be a good class representative? C. Peer Counselor Situation School counselors are trying to put together a new program where kids with problems can go to students as well as to adults for help. These students would be called "peer counselors." The adult counselors have asked you and some other students for suggestions. They say they're looking for kids who others feel can really open up to with their problems. They also want the peer counsellors to be good listeners, and to really care about their classmates. How would you rate this person's ability to be a good peer counsellor?

D. Peer-rated likability (Peer Acceptance) (adapted from Matthews & Keating, 1995) For each name on your list, write down the number (from 1 to 5) that represents how much you would enjoy spending time with that person. For example, if I enjoyed spending time with Person X I would write the number 5, if I wasn't sure I would write either 2,3 or 4 and if I didn't enjoy being with Person X, I would write 1. e.g., Person X 4 APPENDIX B6

Mental State Verbs Task (MVT, adapted Itom Astington & Olson,

1990;Booth & Ha& 1994) Mental State Verbs Task

(adapted from Astington & Olson, 1990; Booth & Hall, 1994) In this task, you will read a short story in which someone says, thinks or knows something (which is marked by underlining). Then you will read four alternatives that use other words instead of say, think, or know, and you are to circle the letter beside one of the four sentences that fits best into the story. 1. Kevin is supposed to empty the garbage before he goes to school but he often forgets and his mother has to do it. Today he remembered it. Just as he's leaving for school his mother says, "Did you empty the garbage today? You forgot yesterday." Kevin savs. "Well I did it todav. for -sure." A. Kevin concludes that he emptied the garbage today. B. Kevin claims that he emptied the garbage today. C. Kevin asserts that he emptied the garbage today. D. Kevin assumes that he emptied the garbage today. 2. Jane and Kate are arguing about which is the best place to eat. Jane thinks Harvey's is best but Kate thinks McDonald's is. Kate says that McDonald's is nearer, but Jane still thinks Harvey's is the best one to go to because the burgers taste better. She savs to Kate. "It's true. McDonald's is nearer, but I'd rather go to Harvey's." A. Jane contradicts that McDonald's is nearer. B. Jane doubts that McDonald's is nearer. C. Jane sungests that McDonald's is nearer. D. Jane concedes that McDonald's is nearer. 3. Susan and Eva are planning to go on a picnic. They want to choose a nice day. One morning they wake up early. Eva says, "Shall we go today?" Susan looks out of the window and she savs. "It will be sunnv all dav." A. Susan predicts that it will be sunny all day. B. Susan knows that it will be sunny all day. C. Susan interprets that it will be sunny all day. D. Susan im~liesthat it will be sunny all day. 4. Bob is reading a scene from one of Shakespeare's plays for English homework. He comes to the lines: "...that she may feel how sharper than a serpent's tooth it is to have a thankless child". He knows the words but he doesn't understand it, so he goes to ask his mother. She savs it means "King Lea.wants Cordelia to know how painful it is to have a chiid who is ungrateful." A. Bob's mother understands the lines. B. Bob's mother interprets the lines. C. Bob's mother criticises the lines. D. Bob's mother defines the lines. 5. Ted is studying in his room at home. Suddenly the bookcase wobbles and things move on the shelves. Ted says, "What's going on? Maybe it's an earthquake." At dinner that night Ted's father savs. " There was an earthauake todav. I just heard it on the news." A. Ted's father predicts that there was an earthquake. B. Ted's father concedes that there was an earthquake. C. Ted's father confirms that there was an earthquake. D. Ted's father denies that there was an earthquake. 6. Sally and Joni just finished watching a movie at the movie theatre. They were talking about the movie on their walk home. Sally asked Joni the name of the hero in the movie. During the movie, Joni thought the hero had a neat name. Now, she can not think of his name. A. Joni realizes the name of hero in the movie. B. Joni forzets the name of hero in the movie. C. Joni notices the name of hero in the movie. D. Joni understands the name of hero in the movie. 7. Last week in science class Mr. Jones showed Dave that acid solution turns litmus paper pink. This week there's a test. The first question says "What colour will litmus paper be when you dip it

sc~2Acz?" Dzye+hputs ~&h& :if .IA,.;,!! & A. Dave remembers that it will be pink. B. Dave hv~othesizesthat it will be pink. C. Dave infers that it will be pink. D. Dave observes that it will be pink. 8. Lisa and John are visiting the zoo. They see an animal with long horns. John wonders it if it a gazelle, or perhaps a gnu. Lisa thinks that it is not a mu. Then they see a sign which tells them that the animal is an elk. A. Lisa denies that it is a gnu. B. Lisa doubts that it is a gnu. C. Lisa declares that it is a gnu. D. Lisa believes that it is a gnu. 9. Stephen leaves a candy bar on the kitchen table. While he's out his mother sees it there and puts it into the cupboard. When Stephen comes back he can't find the candy bar. Then he goes up to his sister's room. He sees an empty candy bar wrapper on the floor, just like the one that he'd had. Stephen thinks that his sister must have eaten his candy bar. A. Stephen knows that his sister has eaten his candy bar. B. Stephen infers that his sister has eaten his candy bar. C. Stephen proves that his sister has eaten his candy bar. D. Stephen discovers that his sister has eaten his candy bar. 10. Jason is very good at making all sorts of models. One day Jason's brother makes a model aeroplane but it won't fly properly. Jason thinks that it will fly if he puts a counterweight in the tail, so he goes to find a weight to try out his idea A. Jason remembers that it will fly if... B. Jason discovers that it will fly if. .. C. Jason hwothesizes that it will fly if. .. D. Jason ex~lainsthat it will fly if. ..

11. FUeethsl &XI-J ,..--Tono txmc.. - o~n1nv-k~-.WA"A- the t2!&s-PAf: x?!&t&g, 2 +gt bGd fieyJ r:ght front of her face. She kept her eyes open because she wanted to see this large bird up close. When talking to her friend Sally later, Jane told Sally about the bird. Sally denied that Jane had seen a bird. Jane said. "I know a large bird flew in front of mv face." A. Jane observed the large bird. B. Jane recomized the large bird. C. Jane understood that it was a large bird. D. Jane realized that it was a large bird. 12. Last week in class Mr. Smith reviewed the states in America and their major products. Today, John has a multiple choice test on that topic. One of the questions asks: "What is the major product made in Wisconsin?" After reading only the question, John cannot answer it. Luckily for John, there are four choices. After reading the choices, John knows that the choice "the maior ~roductis dairv goods" is the correct answer. A. John recomized that the major product is dairy goods. B. John realized that the major product is dairy goods. C. John understood that the major product is dairy goods. D. John observed that the major product is dairy goods. 13. Patrick comes home early from school and finds that nobody is home. He looks in the garage and sees that his mother's car is gone. He also sees that his brother Sam's football equipment is not in the utility room where it always is. He knows that his mother is driving Sam to football practice. A. He understands that his mother is driving Sam to football practice. B. He considers that his mother is driving Sam to football practice. C. He concludes that his mother is driving Sam to football practice. D. He imagines that his mother is driving Sam to football practice. 14. Danielle has been working on her math homework for over an hour. She was having trouble with one problem. Finally, she tried to use a different formula and it worked. The teacher graded

her hnmmxrrn& CCa rrnt lfUlOI, hTn.r. nn-:-ll- l-o--.r LA--- 4.- CI- *Ll ---I..l-- --A UYIIII.VUII~ YIIY &us. AWU IV. APVw ,UUIICIIICI NIVW011VW LVUV LllCI VlWUlGlII. A. Now, Danielle recognizes this is the way to do the problem. B. Now, Danielle understands this is the way to do the problem. C. Now, Danielle infers this is the way to do the problem. D. Now, Danielle com~rehendsthis is the way to do the problem. 15. Georgia sees an animal at the zoo that she has seen before. Her friend Cynthia asks her what it is called. Georgia calls it a Cheetah. Cynthia says, "No, it is a Hyena" Georgia knows why she called it a Cheetah and not a Hvena. Both animals have spots and their names sound similar. A. Georgia assumed that this was the reason she called the animal a Cheetah. B. Georgia that this was the reason she called the animal a Cheetah. C. Georgia expected that this was the reason she called the animal a Cheetah. D. Georgia reflected that this was the reason she called the animal a Cheetah. APPENDIX B7

Gates-MacGinitie Vocabulary Test, Level D 516 (Form 4) (MacGinitie & MacGinitie, 1994) Student Directions Directions for the Gates-MacGinitie Vocabulary Test, D 516 Form 4) Once the participants are ready with their test booklets, the researcher says, "Now turn your booklet so the words "Sample Page" are at the top." I will do the work on this page with you. This test measures your knowledge of words or your vocabulary." (Gates-MacGinitie, 1994, pg, 22). Once all the participants have turned to the sample page, the researcher says: Find number V-1 near the top of the page. (Pause). The words next to V-A are a bi~shout. Shout has a line under it. One of the words or group of words below shout means the same, or nearly the same, as shout. The words or groups of words are boy, yell, fun thing to see, game, shoe. The oce that means nearly the same as shout is yell - a big yell means the same as a big shout. So the answer to number V-1 is yell. To show that the answer for V-1 is yell, fill in answer space L in front of the word yell. When you have finished, put your pencil down. Do not make any other marks. @. 22) The researcher checks that each student has marked the correct answer space with a mark and gives individual instruction to any participant who needs it. The researcher then says: Now look at number V-2. (Pause). The words next to V-2 are They will &g. Stay has a line under it. So you should fmd the word or group of words that means most nearly the same as stay. Look at each of the words or groups of words underneath. Find the one that means nearly the same as stay. Then take your pencil and fill in the space in front of that one. When you have finished, put your pencil down. (p. 22) The researcher gives the participants the time to find the answer then says, "They will not go means most nearly the same as They will stay. So you should have marked space R in front of not go. If you marked the wrong answer space, erase the mark carefully, then mark space R." @. 22). The researcher checks that each student has marked the correct answer space and gives individual instruction to any participant who needs it. Also, the researcher will ensure that any participant who has changed an answer has erased the wrong mark thoroughly. Before beginning the test, the researcher also checks that all the participants have marked the answers for sample questions V-1 and V-2 correctly. The researcher holds up a copy of the test booklet and says, "Look up here. (Wait). When I tell you to begin, you should start with number 1 here at the top of the first page. (Point). You should do this fist page, and then continue until it says STOP." (Gates-MacGinitie, 1994, p. 23). The researcher continues to say, "Circle only one letter for each question. If you wish to change an answer, erase your first mark carefully, then circle the answer that you want." To relieve any possible test anxiety, the researcher says, "Don't worry if you don't know all the words. You will do best if you mark an answer for each question, even if you are not sure. Mark the answer you think is right and go on. I cannot help if there are words that you don't know. If you have any questions now before we begin, raise your hand." (Gates-MacGinitie, 1994, p. 23). Directions for the Gates-MacGinitie Vocabulary Test, D 516 (Form 4) (Cont.) The researcher answers any questions then says, 'All right, turn to the first page and begin with number 1." (Gates-MacGinitie, 1994, p. 23). As the participants complete the test, the researcher walks about the room, checking to make sure that all the students have started at the right place and proceed to circle the answers correctly, not looking at the work of other participants. Once the participant has completed the test, the researcher collects the booklets. APPENDIX B8

Gates-MacGinitie Vocabulary Test, Level D 516 (Form 4) (MacGinitie & MacGinitie, 1994) Gates-MacGinitie Vocabulary Test (MacGinitie & MacGinitie, 1994) Level D 96 Sample Page

Vocabulary

V-1. a big shout @ boy O yell @ fun to see O game @ shoe

V-2. They will -stay. 3 eat CDt2.k O not go a plzy O zsk for help 301 VodJuhfy Gates-HacCJinitie Vocabalarp Test, Level D 5/6 -

6. It was inspec+d 11. a new regadlation 0 negkileri o boss O looked at closely a reduction O put in a row O Mymember 0 stanped Owaytosiglql 0 vyed rule

2. It can -pour. 7. He win probably go. 0 flow out 0 with mdty 0 pile up a careiully 0 be'stong be likely Zo 0 %read O wisb not to 0 pose @ promfiy

3. a big -menu 8. a different breed f3. a iast -effort 0 crowd @ kind . . o willed place 0 list @ Iigit wind re~0rcl 0 pktosit @ food 0 bre2th OElenae @size @* restaurant .@break @ ho?e

4. He might interrupt 9. It had swelled. 14. T&ey were convinced. a watc5.carefally 0 SniEed 0 allowed to pass @ inte=ld to Owarmedup 0 sent to jail @ beinah- o gram b&,emDJ O mde to believe @breakin. CD become painfal 0 connect& 0 follow 0 jelled o toldtovote

5. the important labour 10. a strange misforbme 15. the only fractrrre a* 0 a6vedxz-e O bit of bad luck a small number O kind of song 0 pax5 0 mission 0 &..~3; place 0 ending br& 302

Vocabulary

22 an unnsaaZ -region

20. an important captive O dve @ prisoner 3'1, another specimen 42. It is d&uent 0 la- @ def~rzained o not like others

33. It may be displaced. 38. He was summoned. 0 disgrad o put in jail 0- 0 addon 0 annoyed ol&e?ledto @ on show 0 sent for 0 settled 0.mqaSed

I 35. the authoritative one 45. It muit be obtained.

@ attentive a ke?t inside booW ordained 0 0Ecia.l 0 gore a old= used @ ~~ e not horn APPENDIX C

School Letters and Consent Forms APPENDIX C1

Cover Letter to School Principal Cover Letts? to School Principal Dear School Principal: I am a doctoral student at the Ontario Institute for Studies in EducationAJniversity of Toronto and I am conducting a study investigating the possible relationships among social understanding, self concept and social relations in preadolescents. In collaboration with the Research Review Committee of your Board of Education, the Superintendent of Schools has granted me permission to approach you regarding the participation of your school, specifically your ,pde 6 class in this study. It is my intention to ask the permission of the parents of a Grade 6 class within your school to participate in the study. This would occur, upon your own and the parents' written permission in late March or early April, 1997. Specifically, my request is that I may have a total of 5-7 minutes on a date in April, 1997 to introduce myself to the class and to distribute Parent Consent Forms, 50-60 minutes classtime on a date in AprilIMay, 1997 to group administer three questionnaires on self-concept and peer behaviour, a vocabulary subtest and a metacognitive task concerning the understanding of mental state verbs. In addition, for the weeks of (TBA) I request permission to withdraw each student from the classroom for approximately 30 minutes (6 students a day) to conduct individual interviews concerning interpersonal and self understanding either in the school library or an available room outside of class. Furthermore, I request your permission to ask the present Grade 6 classroom teacher to complete a questionnaire that will rate students on their social behaviour within the classroom. This questionnaire will approximately take one hour of the classroom teacher's time. For your inspection, I have enclosed the cover letters, consent forms and test instruments that I plan to distribute to the teachers, parents and students. The letter to the classroom teacher outlines in detail the tentative timeline of the study and the tasks involved in the study. I hope that you consider the study worth the time that it would require of your teachers and students. In return, it would be my pleasure to provide you with a summary of my findings. Please do not hesitate to contact me if you have any questions or concerns regarding my research. I may be reached at the Institute of Child Study, Toronto, through my Thesis Supervisor, Dr. Janet Astington (416-978- 18l3), my parents' home in Caledonia (765-601 1, March-April, 1997), my home in Toronto (416-538-4584) or by e-mail, [email protected]. Thank you very much for your time. Your cooperation would be greatly appreciated.

Sincerely,

Sandra Bosacki, B.Ed., M.Ed. Enclosures APPENDIX C2

Cover Letter to Classroom Teacher Cover Letter to Classroom Teacher

Dear Classroom Teacher: Your principal has granted me permission to approach you with my research request. I am a doctoral student at the Ontario Institute for Studies in Education and I am conducting a study investigating the relationships among social understanding, self-concept and social relations in preadolescents. The required parental consent that is necessary for my study to be implemented will be obtained by asking your students to send home an Informed Consent Form and accompanying Cover Letter that requests Parental Permission for the student to participate in the study. The students will be asked to return the signed Consent Forms to school as soon as possible. Upon attainment of the Parental Consent Forms, the study will proceed consisting of: a) the group adminstration of two questionnaires: the Self Perception Profile for Children (SPPC; Harter, 1985), a modified version of the Social Competence Nomination Form (Ford, 1982); the Gates-MacGinitie Vocabulary Test, Level D 516 (MacGinitie & MacGinitie, 1994), a modified version of the Mental State Verbs Task (Astington & Olson, 1990; Booth & Hall, 1994) and students also will be asked to rate each of their classmates on a 5-point scale as to how much she or he would enjoy spending time with that person (peer-rated likability), requiring 50-60 minutes of class time; b) the withdrawal of each student from class for 20-30 minutes for an interview between myself and the student regarding interpersonal and self-understanding involving questions concerning two socially ambiguous stories and selected responses from the student's completed SPPC, and c) your participation as classroom teacher by completing a 15-item questio~aire (Teacher's Rating Scale of Child's Actual Behaviour, Harter, 1985) for each of your students that rates their classroom social behaviour. The completion of this questio~airewill approximately take one hour of your time. The specific tasks and tentative timeline for the study include the following:

Mon., March, 1997 I will arrive in your class at a time determined by you for an approximate 5-7 minute presentation which will include an introduction, explanation of my purpose and the distribution of the Cover Letters and Informed Consent Forms to the students. The students will be requested to ask their parents to complete the form, if possible, that evening, and return the signed forms the following school morning.

The remaining weeks Days set aside for the collection in March, 1997 of the parent's Consent Form. I request that you collect the signed forms and place them in a sealed envelope until I return to the school.

Last Friday in March I will return to the school to collect the Consent Forms.

Also, I will request your completion of the SPPC Teacher's rating scale for each of your students who participate which will require about one hour of your time. If you are unable to complete these rating scales during the school day due to time constraints, you may wish to complete these questionnaires at home and I will collect them fiom you the following Monday.

Mon., April Beginning on the Monday, I will 1997 group administer the various measures (approx. 60 min classtime).

Mon., April, 1997 Following administration, I will Fri., April, 1997 be individually withdrawing 6 students for 30 minutes to conduct the interviews either in the library or in another designated area outside the classroom. I plan to interview 3 students in the morning and 3 in the afternoon, completing 30 or so interviews by Friday afternoon. In total, I shall be in your classroom for approximately one and a half weeks.

Final Day I will debrief your class by explaining of Data Collection the rationale and purpose of my study. Thank-you letters will also be distributed to students and parents, including an offer to mail out summaries of my research findings once the data has been analyzed.

Through the investigation of the possible co~ectionamong social understanding, self- concept and social relations, the goal of my research is to increase educators' awareness and understanding of the significant role all three concepts occupy in the life of the preadolescent. The possibility that the manner in which preadolescents perceive social information (the understanding of mental states in others or "theory of mind ability") will be related to how they feel about themselves and the extent to which they engage in social relations may also suggest to educators that both the inter- and intrapersonal aspects of social reasoning may play an important role in learning. Furthermore, the difficulty of self-organization and social differentiation may be eased by emphasizing both the cognitive and affective components of education through the promotion of metacognitive activities that foster self-awareness and reflexive thinking regarding one's own thoughts and feelings. For example, metacognitive activities such as self-narration, bibliotherapy and psychodrama may assist in the development of a positive sense of self in relation to both self and other. Moreover, by exploring the possible differences that may occur due to gender within the dynamics of social understanding, I aspire to increase educators' cognizance of the different ways girls and boys interpret and understand social information within the classroom. Despite the aforementioned benefits of participating in this study, due to the sensitive nature of some of the questions involving self-concept and social attitudes, some children may experience anxiety andfor self-doubt. To ensure that this risk is kept to a minimum, I will remind your students of their right to refuse to answer any of the questions and to leave the interview at any time they wish. Students will also have the opportunity to ask questions about the study following the group administration and individual interviews. I believe that both teachers and students will benefit fiom the present study in that they will be encouraged to self-reflect and develop a greater understanding of the self and other as psychological beings. Furthermore, the opportunity to participate in an academic research study will also benefit the students by exposing them to the process involved in conducting doctoral research. To ensure unbiased and valid results, it is crucial that the parents and students participating in this study will not be informed at this time of the particular relationships planned to be explored in this research. The students themselves need only to be aware that their parents will be asked to grant them permission to participate in a research study being conducted by the Ontario Institute for Studies in EducatiodUniversity of Toronto to increase their own and educators' understanding of the inner world of the preadolescent. Finally, my research will not mention the name of the Haldimand-Norfolk R.C.S.S. Board, nor the names of the students, school, principals or teachers involved. I will not have access to any parent or student names, thus guaranteeing the anonymity of all concerned. I thank you very much for your time and cooperation. Please do not hesitate to call me at home (Phone number in Caledonia, 765-601 1; Toronto, (416) 538-4584) if you have any questions or concerns. I can also be reached through my Thesis Supervisor, Dr. Janet Astington at the Institute of Child Study, Toronto, Ontario (phone no. (416) 978-1813). I look forward to working with you on this project, and I anticipate valuable information for parents and teachers as a result of our efforts.

Sincerely,

Sandra Bosacki, B-Ed., M.Ed. APPENDIX C3

DecIaration of Informed Consent for Teacher Declaration of Informed Consent for Teacher

I consent to participate in this study of the relationships among social understanding, self- concept and social relations in Grade 6 students. The main purpose of this research is to increase awareness in both educators and parents of the possible relationships that may exist among the inter- and intrapersonal aspects of social reasoning and social relations within the classroom. Results fiom the proposed study may generate future curriculum development that aims to promote cognitive/affective unity or the personal side of education.

I understand the following statements:

(I)My participation in the study will involve completing a questionnaire that involves rating each of my students who participate with parental consent on their social behaviour observed during school hours and permitting the researcher to administer a selfconcept questionnaire, two peer rating scales, a vocabulary subtest and a mental state verbs task during classtime and to conduct individual interviews with the students regarding interpersonal and self-understanding.

(2) I am free to withdraw from the study, or refuse to answer any questions at any time without penalty.

(3) My name will appear anywhere on the questionnaire; my scores will be identified by subject number only. Therefore my anonymity is assured.

(4) I am free to make any complaints known to Dr. Merlin Wahlstrom, OISJXJT Chair of Curriculum, Teaching and Learning Department (923-6641 ext 2624).

- - - - - Sandra Bosacki Signature of Teacher

Date APPENIDIX c4

Class Introduction to Research Study Class Introduction to Research Study My name is Sandra Bosacki and I am a teacher and a researcher working on my Ph.D. thesis at the University of Toronto. I am interested in studying how children think and feel about themselves and their friends, especially during grade six. The ages of 11 and 12 are important times in your lives and teachers are interested in how your thoughts and feelings affect the way you act at school. Remember that you are experts on your own thoughts and feelings and that this is your chance to help to "teach teachers" on how you think about yourself and your friends. You have been chosen to participate in my study which will involve two parts. The first part involves filling out some surveys in class (short questionnaires) about what kind of a person you think that you are like and what kind of people you think your friends are. You will also be completing some reading vocabulary tasks which are similar to some of your class work. We will complete these tasks together as a class and it will take about one hour of you class time. Tbe second part of the study involves an interview where I will take you to the library and read you two short stories about friends and ask you some questions about the characters in the stories. I will then ask you questions about some of your answers on the questionnaire that deals with what kind of a person you think you are. This interview will take about 20 minutes and will be tape recorded. Your participation in this study is completely voluntary - you do not have to participate if you do not want to. These tasks have nothing to do with your school work or report cards. If you do participate in the study you will remain anonymous, your names will not appear on any of the forms or the cassette tapes. Only myself and my research supervisor will be looking at your responses. You have the right to refuse to answer any of the questions and/or to stop the interview at any time. If you have any questions during the study please feel fiee to ask - especially if you do not understand the instructions or one of the questions. I will be handing out letters to your parents describing the study and a consent form that they must sign in order for you to participate. Your parents are asked to sign the form whether or not they will allow you to be in the study and you are asked to bring the form back to school. I will be back in a few days to collect the consent forms and we will start the data collection next week Does anyone have any questions before I hand out your parents' letters and consent forms? APPENDIX C5

Cover Letter to Parents Cover Letter to Parents Dear ParentfGuardian: I am a doctoral student at the Ontario Institute for Studies in Education of the University of Toronto and I am interested in studying the relationships among social understanding, self-concept and social relations in preadolescents. I was also employed as an Occasional Teacher and Educational Assistant with your child's school board from 1993-1994. This study has been approved by the School Board Research Ethics Committee and is being conducted in partial fdfilhent of the Doctor of Philosophy degree under the supervision of Dr. Janet Astington. I am requesting your permission to allow your child to participate in this study which entails the completion of a questionnaire on self-concept, two peer-rating scales, a reading vocabulary measure, a metacognitive task and an interview regarding social and self understanding. Your child's present classroom teacher will also be participating in this study by rating your child on social behaviours and attitudes observed in the classroom. The Self-perception Profile for Children (SPPC; Harter, 1985), two peer rating scales (Social Competence Nomination Scale, Ford, 1982; and Peer Likability Scale), the Gates- MacGinitie Vocabulary subtest and a metacognitive task involving the understanding of mental state verbs will be administered to your child's class during school hours and will take approximately one hour. Within the week following this group administration of questionnaires, your child will be individually withdrawn from the classroom and, in a quiet area (library), asked to complete a 20 minute interview involving questions based on two social stories written by the researcher and some selected items from your child's responses on the SPPC. The interviews will be tape-recorded and transcribed for future analysis. Furthermore, your child will be given the opportunity to ask questions before, during and after the study and will be free to withdraw from the study at any given time without penalty. Although sell of the measures to be used in the study have been widely used within schools for many years (aside from the two short stories I have written), the fact that this study encourages children to self-reflect may arouse anxiety and feelings of self-doubt in some children. For example, in the Self-perception Profde for Children (Harter, 1985), your child will be asked to check off which sentence describes what kind of person they are like (eg, Some kids are happy with the way they look but other kids are not happy with the way they look). Thus, to ensure that the risk of being exposed to the sensitive nature of some of the questions is kept to a minimu111,I will remind your students of their right to refuse to answer any of the questions and to leave the interview at any time they wish. Students will also have the opportunity to ask questions about the study following the urniin aclminictrstinn 2nd incliviAiia1 intewi~wcT believe that hnth tearherc anrl ctnrl~ntrGll --= ------. --- .. -. "---. - -- -"------" -- w....--- .. - benefit from the present study in that they will be encouraged to self-reflect and develop a greater understanding of the self and otkr as psychologid beings. Furthermore, the opportunity to participate in an academic research study will also benefit the students by exposing them to the process involved in conducting doctoral research. In order for your child to take part in this study, you are requested to read and sign one of the attached Declaration of Informed Consent forms and return it to the school. You may keep the other one for your records. To ensure confidentiality, my study will not contain the name of your child, the school board or any participating schools and teachers. Also, at no time shall I have access to school records containing any student or parent names. The goals of my research are twofold: 1) to increase educators' awareness and understanding of how preadolescent girls and boys think and feel; and 2) to promote future curriculum development that integrates both the emotional and cognitive aspects of education. I aim to make participation in this study a safe and positive experience for your child. If you have any questions or would like a summary of the results when the study is completed, I can be reached at the Department of Cuniculum, Teaching and Learning, Ontario Institute for Studies in Education of the University of Toronto (phone no. (416) 923-6641, Ext 2603), or by e-mail at [email protected]. I can also be reached through my Thesis Supervisor, Dr. Janet Astington at the Institute of Child Study, Toronto, Ontario (phone no. (416) 978-1813). Upon completion, my study will be available for your perusal at the school board office in September, 1998. Thank you very much for allowing your child to participate. Your cooperation is greatly appreciated. Sincerely,

Sandra Bosacki, B.Ed., M.Ed. APPENDIX C6

Declaration of Informed Consent by ParentiGuardian Declaration of Informed Consent by ParentIGuardian I consent to allow my daughterlson (print herhis full name) to participate in this study on the relationships between social understanding, self-concept and social relations. The main purpose of this research is to increase awareness in both educators and parents of the possible relationships that may exist among the inter- and intrapersonal aspects of social reasoning and social relations within the classroom. Results from the proposed study may generate future curriculum development that aims to promote cognitivelaffective unity or the personal side of education. I understand the following statements: (1) My child's participation in the study will involve the in-class completion of: 1) self- concept questionnaire (Harter's Self-Perception Profile for Children, SPPC), 2) two peer rating scales (Social Competence Nomination Scale and Peer Likability Scale), 3) Gates-MacGinitie Vocabulary sub-test and a Mental State Verb Task followed by an interview dealing with social and self understanding involving questions based on two social stories written by the researcher and your child's responses on the SPPC. (2) My child will be given the opportunity to ask questions before, during or after the study and is free to withdraw from the study, or refuse to answer any questions at any time without penalty. (3) My child's name will not appear anywhere on the questionnaires or on the transcripts; herhis responses will be identified by subject number only, ensuring anonpity. (4) I am free to make any complaints lcnown to Dr. Merlin Wahlstrom, OISE Chair of Curriculum, Teaching and Learning Department (923-6641 ext 2624).

Sandra Bosacki Signature of ParentIGuardian

Date

Child's Date of Birth (dimly)

Ages and Gender of Child's Siblings (If any)

Child's First Language

I do not allow my child to participate in the present study.

Sicmatnr~--o----- nf-- -Parent/GiiarAian------. APPENDIX C7

Debriefing Instructions Debriefing Instructions Debriefing: After Classroom Administration of Tasks We are now finished the first part of the study. For the second part of the study, I will be taking each of you separately to the library where I will read you two short stories followed by some questions. I will then ask you questions about some of your answers on the "What I Am Like" questionnaire. The entire interview should take about 20 to 25 minutes. Please remember that you have the right to refuse to answer any of the questions and to stop the interview whenever you wish. Also, your responses will remain anonymous. Are there any questions before I begin to take people to the library?

Debriefing- After Inter- Intra~ersonalInterview We are now finished the study. Thank you for helping me in my study of how girls and boys in grade six think about themselves and their friends. As I mentioned in your class room, understanding how people think and feel is very complicated and teachers and researchers still have a great deal to learn about how the mind works - especially the mind of an 11- or 12-year-old. The main purpose of this study was to design a task to measure how girls and boys understand thoughts and emotions in other people (the short stories that I read to you) and then to use this task to see if this ability to "read other people's minds" affects how you understand your own thoughts and feelings and how you act with your classmates. With your help, the results of my study may help teachers to better understand how grade six students think about their thoughts and feelings. With this new knowledge, teachers may be able to create some classroom activities that can help you to better understand yourself and your friends. Do you have any questions about the study or what I do at the university? After I finish interviewing everyone, I will hand out thank-you letters that you can give to your parents for allowing you to participate in the study. Thanks once again for your cooperation. APPENDIX C8

Thank-You Letter to Parents Thank-You Letter to Parents Dear ParentIGuardian: Thank you for permitting your child to participate in my research study on the relationships among social understanding, self-concept and social relations in preadolescents. This study was conducted as partial requirement for my Ph.D. thesis in Education at the Ontario Institute for Studies in EducationIUniversity of Toronto. The main goal of my research is to advance our knowledge of preadolescents' understanding of themselves and others as psychological beings. More specifically, the purpose of my study is twofold: 1) design a task to assess social underscanding in preadolescence which wiII increase educators awareness and understanding of how preadolescent girls and boys thiilk and feel; and 2) to promote future curriculum development that integrates both the emotional and cognitive aspects of education. It is expected that findings will show that the manner in which preadolescents perceive social information (understand mental states in others or "theory of mind" ability) will be related to how they perceive themselves and the extent to which they engage in social relations. Such findings may suggest that both the inter- and intrapersonal aspects of social reasoning may play an important role in learning. Furthermore, the difficulty of self-organization and social differentiation may be eased by emphasizing both the cognitive and affective components of education through the promotion of metacognitive activitiex that foster self-awareness and reflexive thinking regarding one's own thoughts and feelings. For example, metacognitive activities such as self-narration, bibliotherapy and psychodrama may assist in the development of a positive sense of self in relation to both self and other. Moreover, by exploring the possible differences that may occur due to gender within the dynamics of social understanding, I aspire to increase educators' cognizance of the different ways girls and boys interpret and understand social information within the classroom. If you have any questions or would like a summary of the results when the data has been analyzed, I can be reached at the Department of Curriculum, Teaching and Learning, Ontario Institute for Studies in Education/University of Toronto (phone no. (416) 923-6641, Ext 2603), or by e-mail at [email protected]. I can also be reached through my Thesis Supervisor, Dr. Janet Astington at the Institute of Child Study, Toronto, Ontario (phone no. (416) 978- 18 13). Upon completion, my study will be available for your perusal at the school board office in September, 1998.

Your cooperation has been greatly appreciated.

Sincerely, Sandra Bosacici, B.Ed., M.E~.