INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 9316189

Adult attachment representations and maternal separation in mothers of firstborn infants

Lutz, Wilma Jean Vargo, Ph.D.

The Ohio State University, 1993

UMI 300 N. ZeebRd. Ann Arbor, MI 48106

ADULT ATTACHMENT REPRESENTATIONS AND MATERNAL SEPARATION ANXIETY IN MOTHERS OF FIRSTBORN INFANTS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Wilma Jean Vargo Lutz, B. S. N., M. S.

The Ohio State University

1993

Dissertation Committee: Approved Ellen Hock Kimberlee Whaley Adviser Human Ecology Andrew Schwebel Department of Family Relations and Human Development ACKNOWLEDGMENTS

The author first acknowledges and expresses her to the women who participated in this study. Their openness in relating information about early family experiences and sharing about becoming and being a mother for the first time provided the foundation necessary and important to the successful completion of this project. Next, the author expresses sincere appreciation to her adviser and mentor, Ellen Hock. Her enduring belief in the author instilled and provided a "secure base" in order to proceed during the most difficult moments of the doctoral program and to successfully complete this study. Her guidance and support throughout the development, implementation, and final completion of this dissertation were invaluable. Special thanks are given to members of the author's dissertation committee, Kim Whaley and Andrew Schwebel, for their helpful suggestions and assistance with the study. Special recognition and thanks are also given to Carol George for her assistance in learning to administer The Adult Attachment Interview, to Mary Main and Erik Hesse for their assistance in learning to use The Adult Attachment Rating and Classification Systems, and to R. Rogers Kobak and Holland Cole Detke for their assistance in learning to use the Adult Attachment Interview Q.-set. The author extends a sincere thank you to Mary Beth Schirtzinger who helped administer some of the attachment interviews and provided interrater reliability for an newly designed interview used in this study. The writer also expresses her appreciation to all of the people who assisted with transcription of the attachment interviews, and is especially indebted to Redonda Engle and Jane Chapman for their persistence and continuing concern with the accuracy of the transcripts. Special appreciation is also extended to Ed Wojniak who provided the second Qrsort ratings for the attachment interviews.

The author extends a very special thank you to her father, Mr. William Vargo, to Aaron, Holly, Erica, Christin and Alison Vargo, and to Mrs. Jane Lutz for their and encouragement throughout her doctoral program. Very special thanks are also given to Benjamin, Zachary, and Oscar for their love, their unique approach to data analysis, and their vitality and tenacity for play. Finally, the author extends deepest appreciation to her husband, Phil, for his continuous proof-reading efforts, numerous trips to the computer center, and assistance in checking over 23,000 pieces of coded data. His love and support made this all possible.

iii VITA

1970 ...... B. S. N.f The Ohio State University, Columbus, Ohio 1970-1971 ...... Staff nurse, University Hospitals Columbus, Ohio 1972 ...... M. S., The Ohio State University Columbus, Ohio 1973 ...... Instructor of Nursing, The Ohio State University, Columbus, Ohio 1973-1974 ...... Nursing Coordinator, Nisonger Center for Developmental Disabilities, Columbus, Ohio 1974-1975 ...... Nursing Consultant, Willson Children's Center, Columbus, Ohio 1976-1977 ...... Instructor of Nursing, Capital University, Columbus, Ohio

1977-1980 ...... Assistant Professor of Nursing, Capital University, Columbus, Ohio 1980-1989 ...... Associate Professor of Nursing Capital University, Columbus, Ohio (Tenured in 1982)

1989-1991 ...... Professor of Nursing, Capital University, Columbus, Ohio

1980-1983 ...... Family Nurse Practitioner, Community Health & Nursing Service, Columbus, Ohio

iv 1981, 1986, 1991 ...... Certification as a Family Nurse Practitioner awarded by the American Nurses' Association 1987-present...... Research associate, Department of Family Relations & Human Development, The Ohio State University, Columbus, Ohio

PUBLICATIONS Lutz, W. J. (1978). Community mental health and retardation. In J. Curry and K. Peppe (Eds.), Mental retardation: Nursing approaches to care. St. Louis: Mosby. Chickerella, B. and Lutz, W. J. (1981). "Professional nurturance: Preceptorships for baccalaureate nursing students," The American loumal of Nursing .81(1). 107-109.

(Article was reprinted in Nursing in transition ed. by T. A. Duespohl, Rockville, MD: Aspen Systems Corporation, 1983) Lutz, W. J. (1986). "Helping children and their families cope with painful procedures," loumal of Pediatric Nursing. 1(1), 24-32.

Lutz, W. (1987). Assessment of the 's health status. In J. Servonsky and S. Opas (Eds.), Nursing Management of Children (pp. 66-109). Monterey, CA: Jones-Bartlett.

Hock, E., Schirtzinger, M. B. & Lutz, W. (1992). Dimensions of family relationships associated with in mothers of young children. Psychology of Women Quarterly. 16. 229-241.

FIELDS OF STUDY

Major Field: Human Ecology Family Relations and Human Development Ellen Hock, Ph. D. (Adviser)

Minor Field: Counseling Psychology Samuel Osipow, Ph. D. (Adviser)

v TABLE OF CONTENTS

ACKNOWLEDGMENTS...... ii VITA...... iv LIST OF TABLES...... viii

CHAPTER PAGE I. INTRODUCTION AND NATURE OF THE PROBLEM...... 1

II. THEORETICAL FOUNDATIONS OF THE STUDY...... 7 Attachment theory ...... 7 Maternal separation anxiety ...... 15 III. REVIEW OF THE RESEARCH LITERATURE...... 19

Adult attachment ...... 19 Maternal separation anxiety ...... 39 Mothers' perceptions of their infants ...... 43 Sum m ary ...... 45 Research questions ...... 47

IV. METHODOLOGY...... 50 Design ...... 50 Sample selection ...... 50 Procedures ...... 51 Measures ...... 52 Data analytic strategies ...... 57 V. RESULTS ...... 60 Description of the sample ...... 60 Measurement of representations of attachment relationships .. 61 Relations between demographic and major study variables .... 76 Maternal separation anxiety ...... 78 Mothers' perceptions of their infants ...... 86 VI. DISCUSSION ...... 92 Measurement of representations of attachment relationships .. 92 Maternal separation anxiety ...... 97 Mothers' perceptions of their infants ...... 104

VII. CONCLUSIONS AND IMPLICATIONS FOR FUTURE RESEARCH...... 108 APPENDICES A. Instruments ...... 113 B. Descriptions of Adult Attachment Classifications and Selected Scales from the Adult Attachment Rating and Classification Systems...... 127 C. list of Abbreviations used for Instruments ...... 132 D. Additional Data Tables...... 134 BIBLIOGRAPHY ...... 137

vii LIST OF TABLES

TABLE PAGE

1. Contingency Analysis of Attachment Classifications Derived from the Adult Attachment Rating and Classification Systems (ARCS) and Adult Attachment Interview Qrset (AAQ) ...... 61 2. Associations between Adult Attachment Classifications and Mother-Father-Peer Scales ...... 63

3. Associations between Mother-Father-Peer Scales and Probable Childhood Experience Scales of the Adult Attachment Rating and Classification Systems ...... 66 4. Associations between Mother-Father-Peer Scales and States of Mind Scales of the Adult Attachment Rating and Classification Systems ...... 70

5. Associations between Adult Attachment Interview Q.-set Dimensions and Mother-Father-Peer Scales ...... 74 6. Associations between Dimensions of Maternal Separation Anxiety and Demographic Variables ...... 76

7. Descriptive Statistics and Paired t-test Results for Dimensions of Maternal Separation Anxiety ...... 78 8. Differences in Dimensions of Maternal Separation Anxiety among Mothers in Secure and Insecure Groups ...... 79 9. Associations between Dimensions of Maternal Separation Anxiety and Mother-Father-Peer Scales ...... 81 10. Associations between Adult Attachment Classifications, Coherency of Mind, and Dimensions of Maternal Separation A nxiety...... 84 11. Descriptive Statistics and Chi-square Results for Aspects of Mothers' Perceptions of their Infants ...... 87

viii 12. Associations between Dimensions of Maternal Separation Anxiety and Mothers' Perceptions of their Infants ...... 90 13. Associations between Dimensions of Maternal Separation Anxiety and Probable Childhood Experience Scales of the Adult Attachment Rating and Classification Systems ...... 135 14. Associations between Dimensions of Maternal Separation Anxiety and States of Mind Scales of the Adult Attachment Rating and Classification Systems...... 136

ix CHAPTER I

INTRODUCTION AND NATURE OF THE PROBLEM

"Healthy mothers, healthy children" is an adage commonly used by clinical practitioners interested in maintaining and promoting health in families. It has several meanings but most often refers to the idea that healthy psychological functioning in mothers contributes to and facilitates optimal development in infants and young children. How do mothers achieve the capacity for healthy psychological functioning? The answer to this question is perplexing since healthy emotional functioning in mothers appears to involve a paradox.

Generally, the capacity to form close, affectionate relationships with children is considered to be one index of healthy maternal functioning.

Mothers are expected to establish and maintain strong emotional ties with their children. This emotional connection, which includes mutuality or the ability to share or join affective states, is believed to foster the formation of a basic sense of and confidence in infants and young children. Moreover, mutuality is believed to be part of the foundation necessary for exploration which fosters the development of autonomy and competence in children

(Bowlby, 1969 & 1973; Mahler, Pine & Bergman, 1975; Sander, 1962; Winnicott, 1975). On the other hand, healthy maternal functioning also involves the

1 capacity to effectively deal with separation and separation-related experiences. That is, mothers are also expected to be able to retain their own sense of psychological separateness and not become overly invested in their children. They must deal with their feelings about closeness to and separation from their children so that the emergence of competence and independent functioning will not be impeded (Mahler, Pine & Bergman, 1975; Sander, 1962; Winnicott, 1975). Therein lies the paradox: How do mothers balance strong feelings of emotional closeness (attachment) with feelings about separation in order to promote autonomy in their children? Although theorists have recognized that regulating feelings of closeness and separation may be difficult for mothers (Benedek, 1970; Mahler, Pine & Bergman, 1975; Levy, 1970; Winnicott, 1975), little is known about how women balance these feelings. The ability to balance closeness and separation is a complex issue for mothers, and arises partially from socialization practices within families and partially from the experience of pregnancy and childbirth. Generally, women are socialized within American families to value close, affectionate relationships with other persons. Attachment to or connectedness with others tend to be emphasized over pursuit of individual interests and personal needs which contribute to the development of autonomy (Chodrow, 1978;

Matlin, 1987). The challenge for women is not which of these should develop, emotional connectedness or autonomy, but rather how to balance needs for closeness with needs for autonomy (Bloom-Feshbach & Bloom-Feshbach, 1987; Choler & Grunebaum, 1981; Choler & Scott, 1987). The ability to balance feelings of closeness and separation often become particularly challenging for women following childbirth. Pregnancy, childbirth, and the early postpartum months represent a period of significant transition for women. Women must not only adapt to many biological and social changes, such as changes in physical appearance and lifestyle, but also experience profound, rapid psychological changes during the process of becoming a mother. One pervasive change that is believed to be an indicator of healthy adjustment during the early months postpart ally is the ability of the mother to integrate her relationship with her child into her personality yet retain her own individuality (Benedek, 1970; Bibring, 1961; Grossman, Eichler, & Winickoff, 1980; Lederman, 1984; Leifer, 1977; Rubin, 1984). In other words, a mother is expected to establish an intimate tie to her new infant, which may involve very intense feelings of , and simultaneously regulate these feelings in order to be able to meet her own needs and to maintain interests separate from her child. Recently, the capacity to form close, affectionate relationships with others has been conceptualized as "adult attachment" and studies have revealed that this construct is associated with indices of emotional health (Kobak & Sceery, 1988; Cole & Kobak, 1990) and healthy parent-child relationships (Bretherton, Biringen, Ridgeway, Maslin, & Sherman, 1989; Haft & Slade, 1989; Main, Kaplan & Cassidy, 1985; Ricks, 1985). Although "adult attachment" has been operationalized in a variety of ways in research, mental representations of attachment relationships with parents which evolve during childhood have been a fundamental component in all current measures of this construct (Hazan & Shaver, 1987: Feeney & Noller, 1990;

Bretherton, et al., 1989; Main, et al., 1985; Ricks, 1985). Some measures 4 emphasize the individual's memories of specific aspects of these relationships, such as parental and rejection (Ricks, 1985). Other measures include these recollections in addition to more general evaluations and/or integration of these relationships from childhood (Bretherton, et al.,

1989; Main, et al., 1985). However, all measures have incorporated questions or items designed to reflect salient attachment-related issues, such as sharing of affection and warmth between parent and child, parental promotion of autonomy, and the ability of the parent and child to cope with separation and/or separation-related experiences (Bretherton, et al., 1989; Hazan & Shaver, 1987; Main, et al.; Ricks; 1985). Thus, in this study, "adult attachment" is conceptualized as mental representations of attachment relationships with one's parents which develop during childhood. Further, the term, representations of childhood relationships, is used to reflect both specific recollections as well as integration of attachment relationships and attachment-related experiences with parents during childhood. At this time, most studies have included either representations based on the adult's memories of specific aspects of attachment relationships and experiences or representations based on integration of the same. Very few studies have incorporated both (Benoit, Zeanah & Barton, 1989; Cole & Kobak, 1990), and the results from these studies have been inconsistent. One study found no relation between specific recollections and integrated representations (Benoit, et al., 1989), whereas the other found an association between the same (Cole & Kobak, 1990). Recently, studies have also found that maternal separation anxiety is

associated with indices of emotional well-being in mothers (Hock & DeMeis, 1990; Hock & Schirtzinger, 1990; McBride& Belsky, 1988) and healthy mother- child relationships (Aber, 1987; Aber& Slade, 1987; McBride & Belsky, 1988).

Maternal separation anxiety, which reflects a mother's feelings and concerns about separation from her child, is proposed to have several origins. However, experiences with relationships during childhood are believed to be

an important source. Through socialization experiences in one's family and the larger cultural setting, a mother forms expectations and acquires standards for the expression of feelings and behaviors associated with the

maternal role, including those that involve closeness to and separation from children (Hock, 1984; Hock, McBride, &Gnezda, 1989). Thus, childhood relationships with one's parents complement and contribute to the expression

of maternal separation anxiety. Since maternal separation anxiety and adult attachment

representations have been found to be related to both emotional well-being

in mothers and indices of healthy mother-child relationships, early family relationships and experiences may be one underlying factor that contributes

to how mothers develop the capacity for healthy psychological functioning. Although maternal separation anxiety and adult attachment representations

appear to be important contributors to emotional well-being and healthy mother-child relationships, virtually nothing is known about the empirical

relationship between them. At this time, two studies have included maternal separation anxiety and adult attachment representations within their investigations (Hock, Schirtzinger, & Lutz, 1992; Hunter, 1990). Therefore, the purpose of this study is twofold: 1) to examine the measurement of adult attachment representations and 2) to explore the relationship between adult attachment representations and maternal separation anxiety in first-time mothers. More specifically, the objectives of the study are:

1. To investigate the equivalence of three approaches to the measurement of representations of attachment relationships. 2. To explore the relationship between representations of

attachment relationships and dimensions of maternal separation anxiety in mothers of infants.

3. To explore the relationship between representations of attachment

relationships and mothers' perceptions of their 6 to 8 week old infants. CHAPTER II THEORETICAL FOUNDATIONS OF THE STUDY

The theoretical foundations for this study are presented in two parts. The discussion begins with a review of attachment theory which provides the foundation for understanding adult attachment, certain aspects of maternal separation anxiety, and mothers' perceptions of their infants. Discussion of the theoretical foundations of maternal separation anxiety follows.

Attachment theory

Background of attachment theory. Attachment theory evolved from

John Bowlby's extensive clinical work with children who experienced loss of or long-term separation from their parents. The theory, which was first proposed in 1958, is intended to be a revision of psychoanalytic theories of object relationships. Attachment theory offers alternative explanations concerning the formation of enduring affectional bonds and relationships (attachments), the effects of separation and loss (or threat of loss) on attachment relationships, and the psychological processes that influence or mediate attachment behaviors and relationships. The theory has eclectic underpinnings, incorporating principles from ethology, evolutionary theory, control systems theory, and cognitive psychology (Bowlby, 1969; 1973

7 & 1980; Bretherton, 1985). Assumptions, definitions, and propositions of attachment theory.

Attachment theory is based on the assumption that human infants are bom with behavioral patterns or tendencies which help promote their survival.

Attachment behaviors constitute one class of innate patterns or tendencies. Although these behaviors are distinct from other survival behaviors, such as feeding and mating, they are equal in importance and significance. Attachment behavior "any is form of behaviour that results in a

person attaining or maintaining proximity to some other clearly identified individual who is conceived as better able to cope with the world" (Bowlby,

1988, pp. 26-27). Various forms of this behavior are expressed when necessary. For example, attachment behavior may be exhibited under conditions of uncertainty, illness, , fatigue, or separation (Stevenson-

Hinde, 1990). However, the underlying function remains constant; i.e., to

protect the person from harm or danger and to foster a sense of trust, confidence, and/or security within the individual (Bowlby, 1969& 1980).

According to attachment theory, infants have the innate propensity to emit signals, such as crying or smiling, to attract adult caregivers and to use specific behaviors, such as following, clinging, or calling, to seek and maintain proximity to caregivers who can provide food, care, and protection.

Infant-caregiver proximity not only depends upon the infant’s ability to utilize proximity-promoting and security-seeking signals and behaviors, but also on the caregiver's complementary propensity to recognize and respond to infant cues as well as use specific behaviors, known as retrieval behaviors, to maintain closeness and proximity to the infant (Bowlby, 1969 & 1979).

During the formation of an attachment relationship, the infant’s signals and behaviors, as well as the caregiver's responses and use of retrieval behaviors, become selective. That is, infant behaviors are expressed in unique ways toward one or more preferred caregivers (attachment figures) and caregiver responses and behaviors are directed uniquely toward a specific child. Thus,

an attachment relationship is best described and conceptualized as a dyadic

rather than an individual phenomenon (Bretherton, 1985& 1987). Bowlby (1969) proposes that infants form attachments to persons who consistently and repeatedly respond to proximity-seeking and security- regulating signals and behaviors in an appropriate or sensitive manner.

Subjectively, appropriate caregiver responses are believed to engender a of trust, confidence, and/or security within the infant. That is, infants develop attachments through interactions in which caregivers are both physically accessible and psychologically available to the infant. These experiences contribute to the formation of asecure base from which the infant explores or ventures into the world and returns from time to time in order to re-establish a sense of felt-security (Ainsworth, 1982). Further, this secure base, which is formed through early experiences with caregivers, is seen as a crucial aspect underlying healthy emotional development and functioning throughout one's life (Bowlby, 1988). Attachment behaviors in adults are considered to be a direct continuation of those in infancy and early childhood. Bowlby (1969) hypothesizes that a strong causal relationship exists between relationships with attachment figures (parents) during childhood and the capacity to form and maintain affectionate bonds and relationships in adulthood. Additionally, he attributes variations in adult attachment relationships to differences in the ways caregivers (parents) performed their roles during the adult's 10

childhood. That is, differences in the availability, accessibility and responsiveness of one's parents during childhood contribute to individual differences in the ways adults perceive, value and form new attachment relationships. The adult's secure base is also considered to be a direct continuation of the one forged in infancy and childhood. Although affectionate relationships which develop during adulthood, such as those between close friends and/or spouses, complement the adult's secure base, early relationships are proposed to have the most pervasive influence (Bowlby, 1969; 1973 & 1988; Ainsworth, 1989). Constructs underlying attachment relationships. Bowlby uses two constructs, behavioral control systems andworking models, to explain how attachment relationships are formed in infancy and how attachment behaviors function within infants, children and adults. Briefly, behavioral control systems are presumed to coordinate and regulate the of specific and the expression of overt behaviors. Working models, also known as representations or representational models, are presumed to organize or integrate behavioral control systems and thus mediate the perceptions, feelings and behaviors that arise in given situations (Bowlby,

1969 & 1973; Crittenden, 1989). A behavioral control system is defined as any set of behaviors which serve the same set-goal or function (Bowlby, 1969). For example, crying, clinging, or smiling may appear to be disparate behaviors, but they are functionally similar. All of them can be used to achieve the set-goal of seeking and/or maintaining proximity to a caregiver. The following example further illustrates the functioning of one behavioral control system, the attachment system. Infants are believed to be able to evaluate how much 11 proximity to or contact with a preferred caregiver is needed in order to feel secure and then compare this need to their present state. If more proximity and felt-security are needed than currently exists, the attachment system activates proximity-seeking behaviors such as smiling, crying, and/or clinging. Once the desired proximity and level of felt-security are attained, the attachment system is deactivated and proximity-seeking behaviors cease. Four major behavioral control systems are presumed to be operative during infancy. These systems are as follows: 1) the attachment system which coordinates proximity-seeking behaviors in order to protect the child and provide a sense of security, 2) the fear/weariness system which coordinates avoidance of people and objects that are potentially dangerous, 3) the affiliative system which coordinates other kinds of social interaction such as play, and 4) the exploratory system which coordinates investigation of the environment (Campos, Barrett, Lamb, Goldsmith, &Stenberg, 1983). These four systems persist throughout childhood, adolescence, and adulthood; however, additional systems emerge during adolescence and adult life. These systems include the mating-reproductive system which coordinates sexual attraction to others, mating, and procreation, and the caregiving system which coordinates the provision of care to one's partner (e. g., spouse) and sharing the provision of care to the young with one's partner (Ainsworth,

1985). Working models can be defined as dynamic, mental representations that include both cognitive and affective components (Bretherton, 1985;

Zeanah & Barton, 1989). Working models, as originally proposed by Bowlby (1969 & 1973), reflect global, unified representations within a person. Although individuals may be aware of their internal working models, these 12 representations tend to operate unconsciously. Working models govern how incoming information is attended to and perceived, determine which affects are experienced, and serve as mental guides for predicting and directing future behaviors (Bowlby, 1980; Zeanah & Anders, 1987). In other words, working models provide rules and rule systems for the direction of perceptions and behaviors as well as the affective appraisal of experiences (Main, et al., 1985).

According to Bowlby (1969 & 1973), each person has two complementary working models, a working model of the world (environmental model) and a working model of the self (organismic model), which are constructed from the individual's experiences with caregivers and other significant persons in his/her environment. A person's working model of the world incorporates representations of one's attachment relationships with caregivers which includes the individual's expectations concerning the availability, accessibility, and responsiveness of preferred caregivers. The working model of the self incorporates the person's perception and appraisal of his/her behavior and feelings, including how acceptable or unacceptable the person is in the eyes of the preferred caregiver.

Bowlby (1973) suggests that working models serve as standards for both the caregiver and child to appraise affective exchanges and each other's behavior. Further, he asserts that these internal standards are inherently predisposed to favor the development of attachment. That is, proximity and affectionate interchanges between the child and caregiver tend to be appraised and felt as pleasurable by both of them, whereas distance

(separation) and expression of rejection are appraised and felt as 13 unpleasurable. The tendency to appraise such experiences as pleasurable fosters the development of attachment between the child and caregiver (Bowlby, 1969). Working models are presumed to become increasingly resistant to dramatic change as they are integrated into one's personality. An enhanced capacity for linking thoughts, feelings, and memories, which evolves during the second year of life, is believed to be the impetus for increased integration. At this time, working models begin to characterize the person's general pattern of relating to others. As the child grows and matures, working models become more complex and better defined but their basic structure does not change. The models become, in essence, "automatic" ways of acting and thinking which contributes to relative stability in how persons construe themselves as individuals and their relationships with others (Bowlby, 1973&

1980).

Recently, Main (1990) has expanded on the concept of conditional behavioral strategies as it applies to working models of attachment in adults. Attachment strategies are viewed as one kind of biologically-based, adaptive strategies which are part of an individual's attachment behavioral control system. These strategies are seen as a way to examine attachment organization across the lifespan, to identify differences in attachment organization in adults, and to enhance understanding of the psychological processes underlying working models of attachment in adults. Compared to the concept of working models, these strategies, as developed by Kobak and colleagues (in press), are believed to provide a more inclusive description of attachment organization. That is, attachment strategies not only reflect a person's working model but also describe his/her adaptation to a caregiver. 14

A secure strategy evolves from experiences in which the caregiver is

physically accessible and psychologically available. This strategy is considered to be a primary strategy in infancy and adulthood because the set- goal of the attachment system (e.g., proximity to the caregiver) is achieved in

response to stressful or fearful conditions (Kobak, Cole, Ferenz-Gillies& Fleming, in press; Main, 1990). Repression and preoccupation strategies tend to evolve from experiences in which the caregiver is either consistently or intermittently inaccessible and unavailable. As Kobak, et al. (in press) explain, "failure to reassure the child of an attachment figure's availability leads to a re-appraisal of inaccessibility and continued activation of the attachment system with accompanying fear and attachment behaviors that no longer serve the function of facilitating access to the caregiver. Since this situation is inherently anxiety provoking, the child must develop secondary strategies to reduce anxiety and maintain the attachment relationship" (pp. 6-7). Repression and preoccupation are considered to be secondary strategies because they involve suppression or manipulation of the primary strategy and alter the output of the attachment system (Main, 1990). Fear and/or anxiety that accompanies failure to achieve the set-goal may be regulated through deactivation or hyperactivation of the attachment system. In adults, deactivation serves to shift one's attention away from attachment- related information, restricts access to attachment memories, and influences both and processing through the exclusion of feelings from awareness or minimization of negative feelings (Kobak, et al., in press; Main, 1990). Hyperactivation functions as a way to reduce fear and/or anxiety through close monitoring of the attachment figure and exaggeration 15

of attachment signals. This secondary strategy may be occasionally effective in eliciting responses from a caregiver and reducing anxiety but comfort, support, and felt-security are experienced inconsistently and/or may never be fully achieved. In adults, hyperactivation serves to perpetuate negative involvement with caregivers and to shift attention to excessive and irrelevant attachment information (Kobak, et al., in press; Main, 1990).

Maternal Separation Anxiety

Maternal separation anxiety is a complex, multidimensional, and multidetermined construct with origins in the mother's personality structure, genetically-determined biases, and role-related beliefs and dictates of her sociocultural background. The construct is defined as an unpleasant emotional state reflecting a mother's concerns and apprehensions about leaving her child. Maternal separation anxiety may be evidenced by feelings of , or which accompany short-term separation experiences (Hock, 1984; Hock, et al., 1989). The early, clinically-based work of psychoanalysts Benedek (1959) and Levy (1943) emphasize important aspects of the mother's personality that play a crucial role in the mother-infant separation process. Levy stresses the attribute of overprotection whereas Benedek focuses on the characteristic of motherliness. Levy (1970) asserts that the intensity of maternal overprotection is governed by intrapsychic and cultural forces. That is, the extent to which a mother needs to maintain closeness to her child is rooted not only within her basic personality structure but also in her earlier experiences with relationships. The degree of intensity of maternal overprotection and its attenuation at certain points influences the mother's 16 ability to balance separation and closeness which is essential for healthy emotional development of the child. According to Benedek (1970), motherliness supplies the matrix for healthy development of children. It facilitates normal, emotional symbiosis between mothers and children as well as prepares mothers for separation from their children, which in turn, fosters autonomy in the child. Benedek asserts that motherliness is derived from a complex process of psychobiologic organization within the individual. Further, she states that motherliness functions as part of the "human mental apparatus" that through evolution became able to modify and integrate early memories and interactive patterns.

Motherliness, in Benedek's view, serves as an instinctual, intuitive guide for a mother's decisions, choices, and reactions.

Maternal separation anxiety is seen as part of motherliness (Benedek, 1970). Intense feelings of separation anxiety, particularly during the first six months after the infant's birth, are considered normal in all mothers.

However, a mother's ego may fall victim to the negative influences of this which distorts the meaning of motherliness. Distortion not only gives rise to behaviors that can interfere with the development of healthy, autonomous functioning in the child but also disrupts the woman's own process of individuation as a mother. Too much separation anxiety, in other words, may interfere with the mother's ability to separate psychobiologically from her child which may lead to difficulty in the re-establishment of her own psychological autonomy. That is, the mother may have difficulty perceiving herself as a person with her own unique needs and may not become secure in her ability to care for and love the child. Specifically, excessive amounts of separation anxiety may give rise to overindulgent, 17 oversolicitous, and extreme overprotective behaviors which undermine a child's attempts to become autonomous. Absence of separation anxiety may be seen in mothers who have weak egos, or whose personality organization is immature. These mothers, according to Benedek (1970), appear to emotionally abandon their children after infancy and never become appropriately attached or connected to their children. Implicitly, Benedek is stating that too much and too little maternal separation anxiety reflects disturbances in the mother's psychological well-being and may interfere with the development of a healthy mother-child relationship. Through discussion of the stages of the separation-individuation process in children, Mahler, Pine, & Bergman (1975) have also reaffirmed the importance of the mother's role in fostering and supporting the development of autonomy in her child. They describe mothers who "appersonate" their infants and discourage independent functioning (Hock, et al., 1989). This idea is similar to Levy's concept of maternal overprotection and Benedek's notion of too much separation anxiety.

Bowlby's (1969, 1973 & 1980) ethological-evolutionary theory of attachment provides a heuristic foundation for understanding the presence of separation anxiety in both infants and mothers. Separation anxiety is viewed as a basic human response to situations or events that intrinsically signal an increased risk for danger. These events need not involve threats of physical harm in order to elicit separation anxiety. They only need be perceived as having the potential to be psychologically harmful or disagreeable. In this way, actual separations as well as threats of separation or abandonment may evoke intense separation anxiety in the individual

(Bowlby, 1988). 18

Since attachment theory was discussed in detail earlier, only a few

salient points will be re-emphasized here. Both mothers and infants are presumed to have genetically-determined, instinctual tendencies to use

signals and behaviors to maintain physical proximity to each other. Efforts to maintain proximity are mutual such that the mothers initiate distance- reducing behaviors on some occasions and children instigate them on others. Further, physical closeness and affectionate interchanges are appraised and

felt as pleasurable whereas as distance (e. g., separation) and expression of rejection are appraised and felt as disagreeable and painful by both (Bowlby, 1973; Hock, et al., 1989). Finally, maternal separation anxiety is influenced by prevailing cultural beliefs and attitudes about mother-infant separation. The ethnic and cultural setting of the individual provides socialization experiences associated

with the maternal role. Through these experiences, a person forms expectations and acquires standards for acceptable maternal behaviors,

including behaviors that involve mother-infant separation, such as employment-related separation and the use of nonmatemal caregivers. These cultural influences, augmented by the forces addressed by both the psychoanalytic and ethological-evolutionary perspectives, determine

individual differences in maternal separation anxiety (Hock, 1984; Hock, et al., 1988; Hock, et al., 1989). CHAPTER III REVIEW OF THE RESEARCH LITERATURE

The review of empirical literature is presented in three parts in this chapter. The discussion begins with a description of the ways in which adult attachment has been conceptualized and operationally defined in research and includes a review of findings based on these approaches. The latter review is particularly important because one of the primary aims of this study is to evaluate the comparability of selected measures of adult attachment representations. A discussion of the empirical literature pertaining to maternal separation anxiety is presented next, followed by a review of research findings for mothers' perceptions of their infants. The latter reviews are particularly relevant because a second aim of the study is to explore the relationship between adult attachment representations, maternal separation anxiety, and perceptions of infants in first-time mothers. The chapter concludes with a summary of the major conceptual issues and the research questions addressed in this study.

Adult Attachment

The conceptualization and operationalization of adult attachment.

Historically, most operational definitions of adult attachment have evolved

19 20 from research focusing on infant attachment relationships. Because this work provides the foundation for understanding certain essential elements in the conceptualization and operationalization of adult attachment representations, it will be discussed briefly. The Strange Situation Behavior Instrument (SSBI) is widely used to assess individual differences in the quality of infant-caregiver attachment relationships. The SSBI is a laboratory-based observational technique consisting of eight 3-minute episodes, which are representative of "every day" situations, administered in a standard order. Infants who are between 12 and 18 months old are exposed to a series of events, including introduction to an unfamiliar playroom as well as separations and reunions involving the baby's parent and a stranger (Ainsworth & Wittig, 1969; Ainsworth, Blehar, Waters, & Wall, 1978). Based on infant reactions during the SSBI, Ainsworth and her colleagues (1978) identified three qualitative patterns of behavioral responses, one secure and two insecure (anxious-avoidant and anxious- ambivalent). Infants with a secure pattern freely explored the unfamiliar playroom and interacted with their mothers and the stranger before separation. When their mothers left the room, many cried or stopped playing to follow or search for them. During reunion episodes, these infants actively greeted and/or sought physical contact with their mothers. Infants with an anxious-avoidant pattern explored the unfamiliar playroom without interacting with their mothers or the stranger before separation. They were more likely to cry when left alone than when the stranger was present. During reunions, they avoided contact with their mothers by ignoring or turning away from them. Those with an anxious-ambivalent pattern resisted 21 interaction with their mothers before separation but tended to remain in close proximity to them. These infants became very distressed when their mothers left the room and were difficult to comfort during reunions. They often showed and actively resisted physical contact and/or interaction with their mothers (Ainsworth, et al., 1978; Ainsworth& Wittig, 1969; Ainsworth, 1982). These three infant attachment patterns were also found to be systematically and consistently related to maternal behaviors observed in the home throughout the infant's first year of life (Ainsworth & Wittig, 1969; Ainsworth, et al., 1978; Main, Tomasini, & Tolan, 1979). Infants with a secure pattern were significantly more likely than those with an insecure pattern to have mothers who demonstrated sensitivity and responsiveness to their signals during episodes of feeding, play, and close body contact. Mothers of anxious-avoidant infants tended to avoid close contact with their infants and to display disinterest in the infant's activities. Mothers of anxious- ambivalent infants tended to either ignore or inconsistently respond to their infant's signals and did not hold their infants often or for long periods of time (Ainsworth, 1982). Subsequently, Ainsworth and her colleagues (1978) proposed that infants internally organize attachment behaviors into clearly identifiable patterns displayed or directed toward a preferred caregiver or a small hierarchy of caregivers. Further, they believe that these internal organizations (working models or representations) are constructed gradually during the course of interactions with the caregiver. These three infant attachment patterns or classifications, then, are hypothesized to reflect the infant's working model of the caregiver's availability, accessibility, and 22 responsiveness based on the history of the infant-caregiver relationship

(Ainsworth, 1982). Although adult attachment has been conceptualized and operationalized in different ways, a person's representations of relationships with caregivers during childhood are a fundamental concept within these approaches. That is, the history of one's experiences and relationship with caregivers continues to be recognized as an important part of the adult’s working model of attachment relationships. In the next section, three different approaches to the conceptualization and operationalization of working models of attachment relationships in adults will be discussed

(attachment style, parent attachment representations, and representations of childhood relationships). A brief description of the approach will be presented first, followed by a review of research findings. Attachment stvle. Drawing upon conceptual similarities in the formation of attachment relationships between infants and their parents and the development of intimate relationships between adults, Hazan& Shaver (1987) formed three attachment styles (secure, avoidant, anxious/ ambivalent). Descriptions of infant attachment classifications from the SSBI were translated into terms appropriate to intimate relationships between adults. Attachment style is intended to specifically refer to an adult's overall, qualitative evaluation of feelings and cognitions concerning intimate relationships in adulthood. Subjects determine their own attachment style by indicating which group of statements best describes their feelings about relationships. A companion measure, labeled as attachment history, was also created. The attachment history consists of questions concerning long-term separations (between parent and child and between both parents) and 23 adjectives to describe the relationship between each parent and the subject during childhood, as well as the parent's relationship with each other. Thus, the attachment history reflects the adult's evaluation of early childhood relationships with each parent and early family memories of the relationship between parents. The trichotomous attachment style and attachment history measures were included as one section in a larger, three-part questionnaire on adult love experiences (Hazan & Shaver, 1987).

Hazan & Shaver (1987) found that attachment styles were associated with differences in intimate or love experiences in adults. For example, subjects with a secure attachment style described their most important love experience as happy, friendly, and trusting. Those with an avoidant style tended to fear intimacy and were prone to and emotional highs and lows. Respondents with an ambivalent style tended to express an obsession with love, a for reciprocation and union, and extreme sexual attraction and jealousy. In addition, they found that the best predictors of adult attachment style were the subjects' perceptions and quality of their relationship with each parent and the parents' relationship with each other. Out of 86 child-parent and parent-parent relationship variables, subjects with a secure attachment style were more likely to report that they had warm relationships with both parents and that the relationship between their parents was affectionate and caring. Subjects with an avoidant attachment style were more likely to describe their relationship with their mothers as cold and rejecting, whereas those with an ambivalent attachment style were more likely to describe their fathers as unfair. Subsequent research based on Hazan& Shaver's work has focused on refining the measurement of adult attachment style (Collins & Reed, 1990; Mikulincer, et al., 1990) and investigating the relationship between attachment style and dimensions of healthy adult functioning in romantic relationships (Collins & Reed, 1990; Feeney& Roller, 1990; Simpson, 1990), in work settings (Hazan & Shaver, 1990), and in separation-related experiences (Mikulincer, et al., 1990). Together, findings from these studies suggest that adults with a secure attachment style tend to have better emotional health and have the capacity to, or do function, more effectively than those with an insecure attachment style. For example, adults with a secure style tend to have higher self-esteem, greater interpersonal orientation, and more positive working models of the world (e. g., more trusting of others, more likely to believe others are altruistic, more confident in the availability of others) than those with insecure attachment styles (Collins & Reed, 1990;

Feeney& Noller, 1990). Adults with a secure style also tend to be more comfortable with intimacy (Collins & Reed, 1990) and tend to experience greater amounts of trust, interdependence, and satisfaction in relationships

(Simpson, 1990). These adults tend to place higher value on and derive more from relationships than work, and are less likely to report symptoms of depression, psychosomatic illness, and physical illness (Hazan &

Shaver, 1990). Further, they are more likely to have lower levels of distress in response to separation experiences (Mikulincer, et al., 1990). Interestingly, these investigations have continued to find that adults with a secure attachment style are more likely to report positive perceptions of their early family relationships than those with an insecure style (Collins & Reed, 1990; Feeney& Noller, 1990; Mikulincer, et al., 1990). Adults with insecure attachment styles tend to have lower self-esteem, lower interpersonal orientation, and less positive working models of the 25 world than those with secure attachment styles (Collins& Read, 1990; Feeney & Noller, 1990). Those with an avoidant style tend to be more uncomfortable with intimacy (Collins & Reed, 1990) and are more likely to express mistrust of others (Feeney & Noller, 1990). Adults with an ambivalent attachment style are least likely to have experienced enduring intimate relationships, and are more likely to report feelings of dependence and a strong desire for commitment in relationships (Feeney & Noller, 1990). Parent attachment representations. Working models of attachment relationships in adults have also been operationalized through interviews designed to reveal the adult's representations of caregiving experiences and/or an attachment relationship with a specific child. For example, the Experiences of Mothering Interview is used to assess caregiving representations in parents of early school-aged children, one dimension of a parent's internal working model of attachment relationships. Scores for parental caregiving representations are formed by averaging the ratings of two 7-point representation scales, secure base and competence (George &

Solomon, 1989). The Parent Attachment Interview, designed for use with parents of toddlers and preschoolers, is used to determine a parent's internal working model of the self as an attachment figure to a specific child. The maternal insight/sensitivity scale associated with this interview is used to reveal individual differences in the quality of the attachment relationship from the parent's perspective (Bretherton, et al., 1989). More specifically, this 9-point rating scale assesses two aspects of a parent-child attachment relationship: 1) the degree to which parents respond sensitively and appropriately to their child's communication, and 2) the extent to which parents display introspection concerning the parent-child relationship as 26

well as the personality and behavior of both the parent and child (Biringen &

Bretherton, 1989). Scores are determined by assessing whether the parent's general statements about child-rearing, the child's relationship with the parent, and the child's relationship with other family members are consistent with specific descriptions of events presented during the interview (Bretherton, 1990).

In both interviews, parents are asked to describe affective aspects of their relationship with their children and how they feel about and manage separation-related issues with their children, such as bedtime or beginning school. Parents are also asked to describe how their parenting practices are like and unlike their own parents (George & Solomon, 1989), or to describe intergenerational similarities and differences between their relationship with their parents and their relationship to their own child (Bretherton, et al., 1989). Thus, the importance of the adult's relationships with his or her family of origin is recognized, but is not the primary focus of in these measures of parent attachment representations. Results pertaining to the parent's relationships during early childhood were only reported in the study describing the Parent Attachment Interview

(Bretherton, et al., 1989). These results were based on comparisons of similarities and differences between the subject's relationship with her mother and with her child. Nineteen of 37 mothers in the sample reported one or more positive aspects from their family of origin that they wanted to carry over to their relationship with their child. These positive aspects included participating in joint activities, providing affection and comfort, understanding, explaining, and encouraging. Similarities in negative aspects arose only when the mothers discussed personality traits. These negative 27 similarities included having a short temper, being easily upset, and lack of self-esteem.

Subjects who described the relationship with their child as different across generations tended to share the same values as those who described the relationship as similar across generations. In contrast to their family of origin, the mothers wanted to provide more physical affection, have better communication, participate in more joint activities, and use different disciplinary practices. Fifteen mothers discussed both positive aspects that they wanted to repeat and negative aspects that they wanted to change in their relationship with their child. Seven mothers talked exclusively about positive aspects whereas fourteen discussed negative aspects without mentioning positive similarities across generations. Only one mother focused exclusively on negative similarities across generations without also indicating a desire for change in her relationship with her child. These findings suggest that variations in relationships during childhood may differentially influence a mother's perception about her current relationship with her child. Ratings of parent caregiving representations have been found to be positively and significantly related to laboratory-based ratings of the child's security of attachment at 6 years of age. Although the overall correspondence between caregiving representations and maternal behaviors in the home was limited, these representations were meaningfully related to maternal behaviors pertaining to the child's autonomy (George& Solomon, 1989). Scores from the maternal insight/sensitivity scale have been found to be positively, significantly related to infant attachment classifications at 18 months and the child's representation of attachment at 37 months of age. 28

These scores have also been found to be significantly related to parental descriptions of the child's attention span and sociability (Bretherton, et al., 1989). Together, these findings suggest that parent attachment representations (specifically, the capacity to be sensitive and function as a secure base) may influence parental perceptions of their child's behavior

and may also contribute to the way in which the child constructs his/her own working model of attachment relationships. Moreover, recollections of

relationships during childhood may also contribute to the formation of and/or to the dynamics within a parent's attachment representation. Representations of childhood relationships. Representations of

childhood relationships with each parent and significant others are the third way in which working models of attachment relationships in adults have

been operationalized. The Adult Attachment Interview (AAI) and the Mother-

Father-Peer Scale (MFP) have been used most often to assess these representations. The adult's working model is usually inferred from responses to the interview questions or items in the questionnaire (Zeanah&

Anders, 1987). Both measures are derived from compatible theoretical foundations.

The basis for the AAI is attachment theory which postulates the construction of complementary working models of the self and of others (attachment figures) from one's history of experiences and relationships during childhood. The basis for the MFP is Epstein's theory of the self-concept which postulates that an individual's self-theory, one part of a person's theory of reality (of self and of the world), consists of a hierarchically organized set of major and minor postulates inductively derived from significant emotional experiences (Ricks, 1985). One of the most basic major postulates within a 29

person's self-theory is self-esteem. This postulate is largely a function of relationships with one's parents during childhood. Individuals with high self-esteem carry within themselves a "loving parent" who is proud and accepting of them. In contrast, individuals with low self-esteem carry a "disapproving parent" who is highly critical of them (Ricks, 1985). Although both measures rely heavily on adult memories of childhood relationships with parents and significant others, these approaches have

several important differences. The MFP provides a purely retrospective assessment of specific memories of parental and peer relationships in childhood. Scores derived from this questionnaire reflect the individual's perception and evaluation of dimensions associated with these early relationships (e. g., acceptance-rejection, independence-overprotection, idealization). Each score reflects a specific representation about one dimension or one part of the dimension (Epstein, 1983; Ricks, 1985). However, these scores do not reflect the extent to which these dimensions may have been integrated into the adult's working model of attachment relationships, do not provide information about the importance of the early relationships to the individual, and do not indicate how these early relationships may influence current functioning of the person.

The AAI is designed to assess an adult'scurrent "state of mind" with respect to attachment relationships. Although subjects describe and evaluate specific childhood relationships, retrospective assessment is not the only, nor the primary, interest in this interview. The main focus is the subject's present perception about the importance of these relationships and experiences, the ease or difficulty with which the person accesses early information, and the extent to which early relationships and experiences are 30 integrated into a coherent working model of attachment relationships. For example, subjects are not only asked to describe and evaluate childhood relationships and separation-related experiences with parents, but are also asked to indicate how these past relationships and experiences influence their current relationship with their parents and with their own child. Ratings and overall classifications are then derived which reflect the individual's willingness to access attachment information, the importance of attachment relationships, and the extent of integration of childhood attachment relationships and experiences (George, et al., in press; Main,

1990). The interview, ratings, and attachment classifications provide more information and insight into the dynamics associated with an adult's working model of attachment relationships than the MFP.

The overall attachment classifications derived from this interview parallel, but are not identical to, Ainsworth’s infant classifications. Both classifications reflect internal working models of attachment relationships constructed from the individual's interactions and experiences with caregivers during childhood. However, the organization behaviorof provides the basis for the infant attachment classifications, whereas the organization ofthought is the basis for adult attachment classifications derived from the AAI (George& Solomon, 1989; Kobak, et al., in press). Studies using the AAI and MFP. Most studies have used either the MFP or the AAI to assess representations of childhood relationships. Two studies have used both to investigate the relation between dysfunctional parent- child relationships and clinical or subclinical disorders. These studies have focused on disorders believed to be associated with disturbances in mother- infant relationships (Benoit, et al., 1989) and in parent-teen relationships 31

(Cole & Kobak, 1990). Interestingly, Cole & Kobak (1990) found significant differences, whereas Benoit, et al. (1989) found no differences, on the MFP Scales between subjects with secure and insecure attachment classifications.

This discrepancy may be related, at least in part, to methodological differences between the two studies. Cole & Kobak (1990) used a newly developed Q.-sort method to determine attachment classifications and used Rick's method for scoring the MFP (described in Chapter IV). Benoit and her colleagues (1989) used Epstein's original scoring method for the MFP and the Adult Attachment Rating and Classification Systems, which is also briefly described in Chapter IV.

Studies have predominantly focused on the association between representations of childhood relationships in adults and indices of child or adolescent functioning, including affect regulation in adolescents (Kobak &

Sceery, 1988), depression and eating disorders in adolescents (Cole & Kobak,

1990), play behavior and performance on problem- solving tasks in 3 to 4 year olds (Crowell & Feldman, 1988), failure to thrive syndrome in infants (Benoit, et al., 1989), and the security of infant attachment relationships (Grossman, Fremmer-Bombik, Rudolph, & Grossman, 1988; Ricks, 1985; Main, et al., 1985; Fonagy, Steele & Steele, 1991). A few studies have explored the relation between representations of childhood relationships and differences in maternal functioning, including the style of assistance provided to preschool children during problem-solving tasks (Crowell & Feldman, 1988), maternal behavior during separation and reunion episodes with preschool children (Crowell & Feldman, 1991), the capacity to attune affectively to infants (Haft& Slade, 1989), maternal (Jacobvitz, Morgan, 32

Kretchmar & Morgan, 1991), and maternal sense of self and selected aspects of mother-infant interaction (Biringen, 1990). Kobak & Sceery (1988) found that representations of childhood relationships were associated with differences in affect regulation in a sample of 53 male and female adolescents in their first year of college. Most students were from middle-class (84%), two-parent families (74%). The AAI was used to assess the adolescents' representations. Based on independent Qr sort ratings, students in the secure/autonomous group were found to be more ego-resilient, less anxious, and less hostile than those in the insecure group. Students in the secure group also reported low levels of personal distress and higher levels of social support from their families. Adolescents in the dismissing group were rated as low in ego-resilience and more hostile than those in the secure and preoccupied groups. These students also reported more and low levels of family support. Adolescents in the preoccupied group were rated as less ego-resilient and more anxious than those in the secure and dismissing groups. These students also reported high levels of personal distress.

Cole & Kobak (1990) found that representations of childhood relationships were associated with depression and eating disorders in a sample of white, middle-class females with a mean age of 18 years, 6 months. Most (82%) came from two-parent families. Both the MFP and AAI were used to assess representations. Depressed subjects reported poor relationships, whereas eating disorder subjects reported more positive relationships with both parents. Depressed subjects tended to report both parents as low in warmth and high in overprotection. Eating disorder subjects tended to report high warmth and support for autonomy in their mothers. Significant, 33 positive associations were also found between depression and the preoccupied attachment classification and between eating disorders and the dismissing attachment classification. Features of depression that were significantly associated with preoccupation included role reversal with both parents, overdisclosure, heightened and exaggerated emotions, and low self- confidence in relationships. Despite self-reports of positive parental relationships, insecure father-daughter relationships (determined by independent Q_-sort ratings) was significantly associated with the dismissing classification. Based on these findings, Cole& Kobak (1990) suggest that dynamics within attachment relationships with parents contribute to the development of specific disorders in adolescents.

Benoit, et al. (1989) used the MFP and AAI to assess representations of childhood relationships in a sample of 50 mothers with hospitalized infants, 25 with failure to thrive syndrome (FTT) and 25 with normal growth patterns. The two groups were matched for race, socioeconomic status, maternal education, infant age and sex. Mothers of FTT infants were found to be significantly more likely to have an insecure attachment representation, particularly preoccupation. Although no significant differences on the MFP scales were found between these two groups, mothers of FTT infants were more likely to have experienced the loss of a significant attachment figure during childhood. These mothers were also more likely to report less help and support from family members and to rate their current relationship with a spouse or partner as more dissatisfying. The investigators propose that insecure models of attachment relationships in mothers reflect a propensity toward disturbances in caregiving relationships, which include interactional difficulties as well as aberrant interpretations and subjective experiences 34

with their infants. This propensity, in combination with infant characteristics and behaviors and contextual factors (such as lack of help and support and stressful or dissatisfying relationships with one's spouse), may contribute to a disordered mother-infant relationship and growth failure in infants (FTT).

Several studies have consistently found significant associations

between representations of childhood relationships and the quality (security) of infant and child attachment. Using the MFP, Ricks (1985) assessed representations and infant attachment classifications in 28 mother-infant dyads living in stable, middle-class families. Mothers of securely attached infants were more likely to report positive recollections of childhood relationships than mothers of insecurely attached infants. These mothers

specifically reported both parents and peers as more accepting and both parents as encouraging of independence. Data from other studies in which the AAI was used to assess representations provide support for Ricks findings. Main, et al. (1985) found that representations of childhood relationships, assessed in both mothers and fathers when their children were 5 to 6 years old, were found to be highly related to the child's security of attachment to each parent in infancy. In two samples of German mothers and infants, Grossman, et al. (1988) found that mothers who valued attachment were more likely to have securely attached infants and those who devalued attachment were more likely to have insecurely attached children. Fonagy, et al. (1991) found that representations of childhood relationships, assessed during the last trimester of pregnancy in a sample of 96 first-time mothers, predicted the security of infant-mother attachment at 12 months 75% of the time. Other studies have also focused on this association but have used other measures to assess early childhood relationships. Results from these studies have been mixed. Morris (1982) used a 160-item clinical interview to collect information about early family relationships of mothers of securely and insecurely attached infants. Objective ratings from the interview revealed

that mothers of insecurely attached infants were more likely to have come from families in which there was an emotionally intense father-daughter relationship (either with her or a sister) and role reversal between mother

and daughter. "Mothers of securely attached infants were more likely to be positively identified with their own mothers and to perceive them as strong, nurturant and emotionally available. The generational boundaries were more clearly drawn and children’s emotional needs were more likely to be

met without jeopardizing parental self-esteem" (p. 316). In an investigation of low income families, Lyons-Ruth, Connell,

Grunebaum, Botein, & Zoll (1984) found that parental warmth in the family of origin failed to distinguish between mothers of securely and insecurely attached infants. Interestingly, mothers of securely attached infants reported more family conflict than those of insecurely attached children. In a study of 64 families, Belsky & Isabella (1988) found no evidence that mothers

of securely attached infants had more positive recollections of early family relationships than those of insecurely attached children. Data from Morris, Ricks, and Main suggest that positive early childhood relationships are associated with, and perhaps contribute to, the mother's capacity to function as a secure base for her child, which in turn may influence the quality of infant attachment relationships. However, the findings from Lyons-Ruth, et al. and Belsky and Isabella, clearly suggest that 36 the nature of this influence is not causal. Findings from two other studies suggest that organization or integration may be as important or more important than recollections of childhood relationships per se. Main & Goldwyn (1984) found that some mothers of securely attached infants had problematic childhood relationships or experiences. These mothers easily recalled early attachment information. Although they expressed anger about early relationships, they also conveyed "" toward their parents; i. e., it seemed that they had psychologically worked through their early negative experiences. Mothers of insecurely attached infants, on the other hand, tended to have more difficulty recalling their early childhood relationships. When these mothers were able to recall early relationships, they tended to idealize their parents and to be inconsistent in their descriptions of their parents. Grossman, et al. (1988) found that mothers who valued attachment but reported negative childhood experiences with attachment figures were able to recall these experiences easily and reflected upon them in a non-defensive manner. Mothers who devalued attachment tended not to recall much about their early relationships, or talked about them in a detached way.

Findings from several studies provide support for the idea that representations of childhood relationships may contribute to differences in the way mothers' perceive and/or interact with their children. Crowell &

Feldman (1988) found that representations of childhood relationships were associated with differences in the interactional behaviors of mothers and preschool children. Three groups were studied during play and problem­ solving sessions in a laboratory setting. Twenty children had behavior problems, 20 had developmental delays, and 24 had no known behavior or 37

developmental problems. Mothers of children with behavior and with developmental problems were more likely to have an insecure attachment representation than mothers of children without problems. Mothers in the secure/autonomous group were more likely to be supportive and helpful during play and to use a style that promoted learning and self-discovery during problem-solving tasks. Mothers in the insecure group were less supportive and helpful. Mothers in the dismissing group tended to use a

directive or controlling style of assistance, whereas those in the preoccupied group tended to use either confusing or controlling styles. These results remained essentially unchanged after the effects of the child's group status (i. e., behavior, delayed, and without problems) were removed. Crowell & Feldman (1991) found that representations of childhood relationships were associated with maternal behavior during separation and reunion episodes with their preschool children. Compared to mothers classified as insecure, those judged as secure were less anxious and thoroughly prepared their children prior to separation. Upon reunion, mothers judged as secure were more responsive, more affectionate, and more likely to correctly interpret their child's behavior and emotional state than those classified as insecure.

Grossman, et al. (1988) found that mothers who valued attachment were more sensitive toward and more accepting of individuality in their infants. During a follow-up visit when the children were 24 months old, these mothers also showed more understanding for developmental problems and individuality of their toddlers. Haft & Slade (1989) found that representations of childhood relationships were systematically related to differences in the capacity to 38

affectively attune to infants in a sample of 15 mothers and their 10 to 13 month old infants. Mothers with a secure/autonomous attachment representation attuned more to their infants than those with an insecure representation. Mothers with a secure/autonomous representation also attuned to a wider range of infant affect. Mothers with a dismissing attachment representation attuned to negative affect, whereas mothers with a preoccupied representation randomly attuned to both positive and negative affect. In a sample of 49 mothers of 6 month old infants, Jacobvitz, et al. (1991) found associations between representations of childhood relationships

(assessed using the MFP), maternal empathy, and maternal sensitivity during mother-infant interaction. Mothers who recalled acceptance by their own mothers during childhood were found to be more sensitive during interaction than those who recalled being rejected by their mothers. Interestingly, mothers who idealized their own mothers were found to beless sensitive during interaction than those who did not idealize their mothers during childhood. Mothers who recalled that their own mothers encouraged independence, as compared to those who recalled maternal overprotection, viewed themselves as more competent, were less likely to respond to their infant's distress with their own distress, and were less likely to interfere with their infant's mood, needs, and interests.

Using an extreme groups design, Biringen (1990) found associations between representations of childhood relationships (assessed using one MFP scale, parental acceptance) and aspects of the mother-infant relationship, such as maternal sensitivity, dyadic physical avoidance, dyadic harmony, and mothers' perceptions of infant responsiveness. Mothers who reported higher 39 levels of parental acceptance during childhood were more sensitive and responsive during interactions with their 11 to 19 month old infants than those who reported lower levels of parental acceptance. Mothers with higher levels of parental acceptance were also less likely to exhibit abrupt handling, physical restraint, and turning away from the infant. Further, these mothers were more likely to be classified as harmonious during interaction than those with lower levels of parental acceptance. Although significant associations were found between representations of parental acceptance and covert anxiety in the mothers, these representations were unrelated to other measures of maternal functioning (self-esteem and overt anxiety). Significant associations were found between representations of parental acceptance and mothers' perceptions of their infant's responsiveness toward them, but were unrelated to other aspects of maternal perceptions, including infant sociability, language competence and psychomotor competence.

Maternal Separation Anxiety

Although a number of studies have found significant associations between maternal separation anxiety and a variety of child characteristics, such as infant temperament (McBride & Belsky, 1988), infant colic (Humphry & Hock, 1989), and infant coping (Hock& Clinger, 1981; Hock, 1984), the primary focus of this study is on separation anxiety as a unique personality attribute of mothers. Several studies have related individual differences in maternal separation anxiety to a variety of maternal behaviors and attributes

(Aber, 1987; Aber & Slade, 1987; Hock, et al., 1988; DeMeis, Hock & McBride,

1986; Hock, et al., 1989; McBride & Belsky, 1988) and recently, studies have revealed aspects of maternal separation anxiety associated with emotional 40

well-being in mothers (Hock & DeMeis, 1990; Hock & Schirtzinger, 1990; Hock, et al., 1992). At the present time, two studies have investigated the association between representations of attachment relationships and maternal separation anxiety (Hock, et al., 1992; Hunter, 1990). Results from several studies have revealed that maternal separation anxiety may guide or influence maternal decisions and behaviors. In a

longitudinal study of 130 first-time mothers, Hock, et al. (1988) found that differences in maternal separation anxiety were related to employment preference and choices about child care. Women who preferred to be employed expressed less maternal separation anxiety (Subscale 1), less concern about the effects of separation on their child (Subscale 2) and less

employment-related concerns than those who preferred to be home (Subscale

3). Among employed women, those who expressed less concern about the

effects of separation on their child tended to enroll their infants in day care centers whereas those with greater employment-related separation concerns tended to use day care centers less often. Employed mothers with fewer

concerns about the effects of separation tended to later enroll their child in

preschool, whereas nonemployed mothers with greater concerns about the

effects of separation were less likely to enroll their child in preschool. In a longitudinal study of 62 well-educated, career-oriented women, DeMeis, et al. (1986) found that maternal separation anxiety (Subscale 1) was highest for all mothers in the early postpartum months then declined significantly over the infant's first year of life. However, differences in the rate and extent of decline were associated with mothers' employment preference. In women who preferred employment, maternal separation anxiety declined earlier and to a greater extent than in those who preferred 41 to be at home. Although no significant differences were found for Subscale 2 (concerns about the effects of separation), employment-related concerns

(Subscale 3) increased markedly in mothers who preferred to be at home and decreased significantly in those who preferred employment. Hock, et al. (1989) found that maternal separation anxiety (Subscale 1) was positively and significantly related to both interview-based assessments of mothers' separation concerns and mothers' reunion behavior in the laboratory. That is, mothers who scored high on Subscale 1 expressed high degrees of concern at the point of separation from their 3 to 4 month old infants, and tended to behave in an anxious or concerned manner when they were reunited with their children. These mothers verbally stated that they missed their child or expressed relief to be with their infant again. They questioned the caregiver extensively about how the infant behaved during their absence and immediately picked up the child or soothed the infant (or both) when he or she was actually content. Aber (1987) and Aber & Slade (1987) found that mothers who expressed greater concern about the effects of separation (Subscale 2) were less likely to behave in ways that promoted the autonomy of their 15 to 24 month old toddlers during a laboratory-based separation-reunion procedure. Mothers who expressed less maternal separation anxiety (Subscale 1) were more likely to behave in ways that promoted autonomy. Differences in dimensions of maternal separation anxiety have also been found to be associated with security of infant attachment. McBride& Belsky (1988) found significant differences between mothers of securely and insecurely attached infants for employment-related concerns (Subscale 3) and marginally significant differences for maternal separation anxiety 42

(Subscale 1). Mothers of securely attached infants had higher levels of employment-related concerns and moderate levels of maternal separation anxiety than those of insecurely attached infants.

Maternal separation anxiety (Subscale 1) has been found to be associated with indices of emotional well-being in mothers. McBride& Belsky (1988) found that mothers who experienced greater levels of maternal separation anxiety (Subscale 1) between 3 and 9 months infant age tended to have higher levels of interpersonal affect than those who experienced lower levels. However, mothers who had greater maternal separation anxiety during this 6 month period of time also tended to have lower self-esteem. In a sample of 164 mothers of 8 month old infants, Hock& DeMeis (1990) found that women with high levels of separation anxiety (Subscale 1) and high levels of career salience were at higher risk for depressive symptomatology. Hock & Schirtzinger (1990), in a sample of 87 mothers of 6 year old children, found that women with low levels of separation anxiety tended to have lower levels of depressive symptomatology than those with high levels of separation anxiety. Mothers with low levels of separation anxiety also had a more positive internal working model of the self and tended to be more healthy with respect to the way they coped with stress. Hock, et al. (1992) found that three variables accounted for 37% of the variance in maternal depressive symptomatology in a sample of 83 mothers of 6 year old children. Maternal separation anxiety (Subscale 1) accounted for the greatest proportion (17%) of the unique variance followed by marital satisfaction (15%) and recollections of father rejection during childhood (4%). 43

Studies that have investigated the association between maternal separation anxiety and representations of attachment relationships (assessed using the MFP) reveal a varied pattern of results. Hock, et al. (1992) found that both mother and father overprotection were significantly and positively correlated with maternal separation anxiety (Subscale 1). Mother and father rejection as well as mother and father acceptance were unrelated to separation anxiety. Hunter (1990), in a sample of 43 mothers of 3 to 4 month old infants, found that representations of attachment relationships predicted scores for only one dimension of maternal separation anxiety (Subscale 1).

High scores on mother overprotection and low scores on father encouraged independence significantly predicted scores for Subscale 1. That is, women who recalled their mothers as overprotective and fathers as discouraging of autonomy were more likely to experience higher levels of maternal separation anxiety. Interestingly, high scores on mother rejection and high scores on mother encouraged independence reached marginal significance in predicting scores for Subscale 1. That is, women who recalled their mothers as rejecting but encouraging of autonomy tended to experience higher levels of maternal separation anxiety.

Mothers' Perceptions of their Infants

To the writer's knowledge, only one study has been completed that specifically examined the association between representations of attachment relationships (as defined in this study) and mothers' perceptions of their infants. As described in the first section of this chapter, Biringen (1990) found significant associations between representations of attachment relationships and mothers' perceptions of their infant's responsiveness 44 toward them. Maternal representations were unrelated to other aspects of the mothers' perceptions, such as sociability, language and psychomotor competence. Results from several studies reveal that perceptions of infants may contribute to or influence maternal functioning, specifically a mother's capacity to function as a secure base for her child. In a longitudinal study of 318 healthy, first-born infants, mothers with positive expectations about newborn behavior were found to have "optimal empathy" (Broussard, 1970; Broussard & Hartner, 1971; Broussard, 1984). That is, these mothers were able to accurately recognize infant cues and responded flexibly to their infant's needs. Mothers with negative perceptions misinterpreted cues and inconsistently responded to infant needs. In a sample of 43 mother-infant dyads, Nover, Shore, Timberlake & Greenspan (1984) found that mothers with distorted perceptions were significantly less likely than those with non-distorted perceptions to be sociable, physically affectionate, responsive, and expressive of overt pleasure during a free play session with their infants in the laboratory. That is, mothers with negative perceptions of their infants tended to be less socially interactive and less psychologically available to their infants. Further, these mothers also tended to interfere with their infant's exploratory play and general interest in the playroom. In a sample of 52 mothers and their 15 month old children, Nuttal,

Stollak, Fitzgerald, & Messe (1985) found that interpersonal perceptual style was significantly related to mother-child interaction during a structured sequence of play activities in a laboratory setting. Mothers with positive 45 perceptual styles used more positive touch and engaged in more mutual play and exchange games with their children than those with negative styles.

Summary

Although adult attachment representations have been operationalized in different ways in research, studies have primarily relied on two measures, the AAI and the MFP, to determine representations. Some studies have included both measures in their investigations but different methods were used to score both the AAI and the MFP. Further, the results from these studies were inconsistent. To the author's knowledge, there have been no studies completed at this time that have examined the equivalence of these approaches to the measurement of adult attachment representations.

An ever-expanding body of research supports the contention that both representations of attachment relationships and maternal separation anxiety may contribute to the development of healthy parent-child relationships and may influence psychological functioning of parents. Although studies have examined associations between representations of attachment relationships and maternal separation anxiety, the findings are inconsistent. Lack of a more systematic relationship between representations of attachment relationships and one or more dimensions of maternal separation anxiety is somewhat surprising in light of the following considerations. Maternal separation anxiety is derived, in part, from early relationship experiences in one's family, and represents beliefs and feelings specifically associated with separation-related issues within the context of mother-child relationships.

According to theoretical predictions, working models of attachment relationships are constructed from the person's history of experiences and 46

relationships with caregivers (parents) and significant others in childhood.

Also, working models are believed to guide one's perception and the expression of one's feelings and behavior. If recollections and/or integration of relationships with parents during childhood reflect attachment representations in adults, then one would expect these representations to be associated with maternal separation anxiety.

Further, since dimensions of maternal separation anxiety have been

found to be associated with security of infant attachment and autonomy- related behavior in mothers, maternal separation anxiety may be a unique aspect of a mother's capacity to function as a secure base for her child.

According to attachment theory, experiences and relationships with parents during childhood have a pervasive influence on the formation of a person's

secure base. One would expect, then, that representations derived from these

early experiences and relationships would be associated with aspects of a mother's secure base, such as maternal separation anxiety. Although only one study was found that specifically explored relations between representations of attachment relationships and mother's

perceptions of their infants, theoretical support exists for this association.

Working models of attachment relationships are presumed to guide one's

perceptions and these models are proposed to evolve from, and to vary based on, one's relationships with caregivers during childhood. Therefore, one would expect representations of attachment relationships to differentially and subjectively influence how a mother perceives her infant and evaluates her relationship with her infant. Research Questions

Based on these conceptual issues, the following research questions directed the design and activities of this study: 1. Do the Mother-Father-Peer (MFP) Scale, the Adult Attachment Rating and Classification Systems (ARCS), and Adult Attachment

Interview d-set (AAQ) provide similar or comparable information about attachment representations in mothers of infants?

a. To what extent do mothers receive identical classifications

using the ARCS and AAQ? b. Which scores for the MFP scales are highly correlated with secure and insecure adult attachment classifications? c. Which scores for the MFP scales are highly correlated with the ratings for the ARCS scales and the scores for the AAQ.

dimensions?

2. Do demographic characteristics vary systematically with each dimension of maternal separation anxiety? (Demographics included mother's age, years of education, occupational prestige ratings, and combined family income.)

3. Do demographic differences exist between mothers classified as

secure/autonomous and those classified as insecure?

4. To what extent do scores for each dimension of maternal separation anxiety remain stable or change between the third trimester of pregnancy and 6 to 8 weeks postpartum?

a. Is stability or change associated with secure/autonomous or

insecure adult attachment classification? 48

b. Is stability or change associated with certain demographic

characteristics or a particular pattern of demographic characteristics? 5. What systematic differences in maternal separation anxiety exist between first-time mothers who are classified as secure/ autonomous and those classified as insecure with respect to attachment relationships? a. Are mothers classified as secure/autonomous more or less

anxious about separation from their infants (Subscale 1) than

those classified as insecure? b. Do mothers classified as secure/autonomous have greater concerns about the effects of separation experiences on their

children (Subscale 2) than those classified as insecure? c. Do mothers classified as secure/autonomous have more or less

employment-related separation concerns (Subscale 3) than

those classified as insecure? 6. Which aspects of representations of attachment relationships are associated with each dimension of maternal separation anxiety? a. To what extent are specific representations of attachment

relationships, as determined by the MFP, related to each dimension of maternal separation anxiety?

b. To what extent is integration of attachment relationships, as determined by adult attachment classification (secure/ autonomous vs. insecure), related to each dimension of maternal

separation anxiety? 49

7. What differences exist in perceptions of infants between mothers classified as secure/autonomous and those classified as insecure? a. Are mothers classified as secure/autonomous more likely to have higher scores on positiveness of perceptions than those

classified as insecure? b. To what extent is positiveness of maternal perceptions associated with each dimension of maternal separation anxiety? CHAPTER IV METHODOLOGY

This chapter begins with a discussion of the research design and the method used to select subjects for this study. The procedure and instruments utilized in the study are presented next, followed by a discussion of the strategies employed to analyze the data.

Design

A segmented longitudinal panel design was used for this study. This strategy, which is a modification of time-ordered, multi-occasion designs, uses two or more data collection points that bracket a crucial or salient transition period for subjects, such as the transition to motherhood (Lazerlere & Klein, 1987; Neale & Leibert, 1986). Two data collection points were used in this study, the third trimester of pregnancy and 6 to 8 weeks postpartum.

Sample Selection

The mothers were randomly selected from a larger sample of subjects

(N=192) who are participants in a study focusing on emotional well-being in mothers of young children funded by The National Institute of Mental Health

(Ellen Hock, Ph. D., principal investigator). Every fifth subject was asked to

50 51 participate in this study. The subjects were recruited from childbirth education classes, obstetrical clinics, private physician offices, a women's health center, and other community resources in Columbus, Ohio.

Procedures

Subjects were contacted by telephone to explain the study and an interview was scheduled for those who indicated a willingness to participate. Before the interview began, each subject was asked to read and sign a consent form (OSU consent form HS-027) to participate. The Adult Attachment Interview, the Maternal Separation Anxiety Scale, and the Mother-Father-Peer Scale (Appendix A) were administered during the third trimester of pregnancy. At this time, subjects were also asked to describe what they expected their infant and their relationship with the infant to be like at 6 to 8 weeks of age. All interviews were conducted in either the subject's home or The Family Relations and Human Development

Laboratory at The Ohio State University. The adult attachment interviews were conducted over a 6-month period of time (December, 1990 through May,

1991) by persons who had extensive interviewing experience in professional work settings, had received training in administration of the interview from C. George, Ph. D., and had completed pilot interviews to become familiar with the interview protocol. Between 6 and 8 weeks postpartum, subjects were contacted by telephone in order to administer the Mother's Perceptions of Her Baby

Interview (Appendix A). At the end of this interview, the subjects' prenatal responses were read to them and they were asked to rate the extent to which their prenatal expectations matched their current perceptions using a 52

7-point scale (7=exactly, l=not at all). The Maternal Separation Anxiety Scale was mailed to subjects within two weeks following the telephone interview with a stamped, self-addressed envelope for them to return the questionnaire.

Measures The Maternal Separation Anxiety Scale CMSASi was used to assess maternal separation anxiety. This instrument is a 35-item, self-administered questionnaire developed by Hock, Gnezda and McBride in 1982. Each item is a 5-point Iikert scale ranging from strongly agree (5) to strongly disagree (1). The MSAS yields scores for three independent dimensions which represent different aspects of maternal separation anxiety. Subscale 1, entitled Maternal Separation Anxiety, represents a mother's 1) level of worry, sadness, and guilt when separated from her child, 2) beliefs about the importance of exclusive maternal care, 3) beliefs that her child prefers her care and is better off in her care, and 4) beliefs about her child's ability to adapt to nonmatemal care. Subscale 2, entitled Perception of Separation Effects on the Child, represents a mother's view of her child's reaction to separation from her and the effects, positive or negative, on the child.

Subscale 3, entitled Employment-related Separation Concerns, represents a mother's attitude about balancing maternal role and career investments that would be associated with mother-child separation (Hock, et al., 1989). The instrument and subscales have high internal consistency.

Cronbach's coefficient alpha for the complete instrument is .88 and for subscales 1, 2, and 3 are .90, .71, and .79 respectively. The test-retest stability for the entire instrument is .75 and for subscales 1, 2, and 3 are .73, .58, and .72 53 respectively. The three-factor structure of the instrument is highly stable over time. Coefficients of congruence over a three month period are 1.0 for factor 1 (maternal separation anxiety), .987 for factor 2 (perception of separation effects on the child), and .996 for factor 3 (employment-related separation concerns) (DeMeis, et al., 1986; Hock, et al., 1989). Both the Mother-Father-Peer Scale (MFP) and the Adult Attachment Interview (AM) were used to assess representations of attachment relationships. The MFP is a 70-item, self-administered, 5-point likert style questionnaire designed to assess three dimensions of recollections of childhood relationships with both parents and peers (Epstein, 1983). These dimensions include 1) acceptance-rejection by mother, father, and peers, 2) independence-overprotection by mother and father, and 3) idealization of mother and father. Test-retest reliability for each scale is acceptable, ranging from .88 to

.93. Internal consistency coefficients for the mother and father independence-overprotection scales are .88 and .82 respectively. Internal consistency coefficients for the mother and father acceptance-rejection and independence-overprotection scales are both .91 and for the peer acceptance- rejection scale, .93. The acceptance-rejection and independence- overprotection scales also have reasonable convergent and divergent validity. For example, the mother encouraged independence, father encouraged independence, mother acceptance, father acceptance, and peer acceptance scales correlate more highly with self-esteem, a sense of competence and a sense of love worthiness than with introversion, , and defensiveness. No additional reliability and validity data 54 have been reported for the parent idealization scales (Epstein, 1983; Ricks, 1985).

Ricks (1985) developed an alternate scoring method in which each bipolar dimension is separated into independent measures, thus deriving twelve scales which represent specific representations (postulates) about relationships during childhood. Five additional scales are derived which serve as indices of the subjects' overall representations (higher-order postulates) of relationships with parents and peers. Both methods (Epstein and Ricks) were used to derive scores from the MFP in this study. The Adult Attachment Interview (AAI1 is a hour-long structured interview designed to determine an adult's attitude toward or "state of mind" regarding attachment. The interview is comprised of 18 questions which assess an individual's understanding of the importance of attachments in his/her life. Subjects are asked to describe their relationships with both parents, siblings and significant others during childhood and then to recall memories in support of these descriptions. They are also asked to recall memories of separations, losses, and physical and emotional upsets as well as how they and their parents "behaved as they did." Questions probe for the subject's level of understanding of and insight into these relationships and experiences as well as their current assessments of the same (George, et al., in press; Main, et al., 1985). Interviews were transcribed verbatim and scored using both the Adult Attachment Rating and Classification Systems (Main& Goldwyn, in press) and the Adult Attachment Interview Qrset (Kobak, 1988). The Adult Attachment Rating and Classification Systems (ARCS) includes seventeen 9-point rating scales and several overall attachment classifications and subcategories. The 55 scales and classifications utilized in this study are described briefly in Appendix B. Transcripts are first rated to determine the subject's probable experience with each parent during childhood and then rated for the individual's current state of mind with respect to attachment. Attachment classifications are assigned to each transcript by matching the patterns of scale ratings to the description for each attachment category (Kobak& Sceery, 1988; Main, et al., 1985; Main, 1990),

All scoring based on the Attachment Rating and Classification Systems was completed by the writer who received training from Maiy Main, Ph. D. during the Summer of 1991. The author correctly classified 13 of 15 (87%) transcripts on a pretest administered by Mary Main. The Adult Attachment Interview Q-set (AAQ) is comprised of 100 statements designed to assess both working models of attachment relationships and organization of thought in interview transcripts. Raters read the interview transcript then provide a Qrsort description by distributing the Qjitems among nine categories ranging from Least

Characteristic (1) to Most Characteristic (9). The final Qjsort description has a forced bell-shaped distribution with 18 items in the rating category of 5

(Neither Characteristic or Uncharacteristic) and 5 items in each of the extreme rating categories (Kobak, et al., in press). Typically, two raters provide Qrsort descriptions for each transcript and composites across raters are then correlated with Qrsorted prototypes for two attachment dimensions (Security-Anxiety and Repression- Preoccupation). When composite reliability (Spearman-Brown formula) falls below .58, a third rater may complete the Qrsort. In transcripts where a third rating is required, composites from the two most consistent Qrsort 56

descriptions are correlated with the dimensional prototypes. Attachment

classifications are assigned by examining the pattern of correlations between

the two attachment dimensions. For example, subjects who have a positive correlation with the Secure-Anxious dimensional prototype and a negative correlation with the Repressed-Preoccupied prototype receive a Secure classification. Subjects who have negative correlations with the Secure- Anxious prototype and positive correlations with the Repressed-Preoccupied prototype receive a Dismissing classification, whereas those who have negative correlations with both dimensional prototypes receive a Preoccupied classification. The Security-Anxiety dimension has been found to effectively differentiate Secure from Anxious subjects. Likewise, the Repression-Preoccupation dimension has been found to differentiate between Dismissing and Preoccupied subjects but more overlap has been reported between these two insecure groups than between the Secure and Anxious groups (Kobak, et al., in press). In this study, two raters provided the Q.-sort descriptions for each subject's transcript. One Qrdescription was completed by the author who had established composite reliability ranging from .58 to .82 on training transcripts from R. Rogers Kobak, Ph. D. The second Qrdescription was completed by a doctoral student trained by the writer. The student established a minimum composite reliability of .58 with the author on at least 5 training transcripts before completing Qrsorts independently. Mother’s Perceptions of Her Babv Interview (MPBI) is a semi­ structured interview consisting of three questions with probes that was specifically designed by the writer to assess the degree of positiveness of mothers' perceptions of their infants in this study. Content analysis of the 57 interview was completed (presented in Chapter V) and the following scoring procedure was developed. Subjects responses to each question were classified as either "positive" or "mixed." The designation of "positive" was assigned when the mother exclusively used positive adjectives or phrases to describe her infant or aspects of the mother-infant relationship; "mixed" was assigned when the mother included any negative or neutral adjectives and phrases in her descriptions. Examples of adjectives and phrases judged as "mixed" include the following: "strong-willed," "demanding," "spoiled," "she's so dependent," "it's frustrating when I can't figure out what he wants," "I feel like I'm the Texaco fuel pump." A value of 1 was assigned to each question with all positive descriptions and a value of 0 was assigned to each question with mixed descriptions. The number of positive and mixed designations were summed to form one score which is intended to reflect the overall degree of positiveness of maternal perceptions. Fifteen interviews were randomly selected and rated by a judge blind to the other measures in the study.

Interrater agreement ranged from 86.7% to 93.3% for the three questions; overall interrater agreement for the interview was 93.3%.

Data Analytic Strategies

Descriptive and correlational analyses were carried out in order to examine the equivalence of the three methods used to measure adult attachment representations in this study. Frequencies and proportions of subjects classified as secure/autonomous, dismissing, and preoccupied were calculated as well as the frequency and proportion of subjects receiving identical adult attachment classifications using the ARCS and AAQ.methods. A coefficient of concordance (kappa) was also computed to evaluate the degree 58

to which the AAQ.& ARCS methods classify subjects in the same manner. The

kappa statistic was developed by Cohen (1968) to specifically measure interobserver reliability of categorical data. In this analytic context, the chi- square statistic serves as an index ofassociation and not necessarily as a

measure of agreement; therefore, use of the kappa statistic is recommended over chi-square analyses when judges make assignments to the same levels of nominal data (Cohen, 1968; Hartmann, 1977). Additionally, the pattern of correlations among the two sets of scores from the MFP and the adult attachment classifications derived from both classification methods were examined as well as the pattern among both sets of MFP scores, ARCS scale ratings, and AAQ. dimensional scores. In other words, scores were compared both between methods (e. g., interview-based versus self-report) and across methods (e. g., Epstein's scoring versus Ricks' scoring methods). Because Qrsort methodology provides ipsative descriptions of an individual, the Spearman-Brown correlations generated for the AAQ. dimensions for each subject were transformed into standard scores for use in statistical analyses. This transformation was done in order to preserve the normal distribution of these scores (Minium& Clarke, 1982). Correlational analyses were also conducted in order to explore the relationship among adult attachment representations, demographic variables, dimensions of maternal separation anxiety, and maternal perceptions of their infants. Since the number of subjects classified as

Dismissing and Preoccupied was small, they were combined into one group for statistical analyses. Chi-square analyses and t-tests were used to reveal significant differences between secure and insecure groups of mothers. The 59 chi-square statistic is useful in making inferences with data that can be reduced to proportions or frequencies; the statistic provides a measure of the discrepancy between expected and observed frequencies (Minium& Clarke, 1982). The t-test statistic is useful in testing differences between means derived from samples that have been randomly selected from a population of observations that follow a normal distribution. The statistic is not markedly affected by violations of the assumptions of normality and homogeneity of variance unless samples are very small or the population distribution is seriously skewed (Minium & Clarke, 1982). CHAPTER V RESULTS

This chapter begins with a brief description of the sample, followed by

a presentation of the results for the study. Findings from analyses conducted in order to examine the equivalence of three adult attachment measures are described first. Next, results are presented from analyses performed to determine associations between selected demographic and major study variables. Following consideration of these findings, results are described for the analyses used to examine relations between maternal separation anxiety, adult attachment representations and maternal perceptions of infants. The chapter concludes with a description of the results of the analyses carried out to explore associations between mothers' perceptions of their infants, adult attachment representations, and maternal separation anxiety. (To facilitate reading of this chapter, a list of abbreviations, complete titles, and the authors of each study measure can be found in Appendix C.)

Description of the sample

The sample consisted of 49 married, primiparous women who were 19 years of age or older. Ninety-six percent (n=47) of the sample was Caucasian whereas 4% (n=2) was Hispanic. Fifty-one percent (n=25) of the women delivered girls and 49% (n=24) delivered boys. The mean age of the mothers was 28.2 years (range: 19 to 36 years, SD=4.095) and their mean level of

60 61

education was 15.1 years (range: 12 to 19 years, SD=2.023). Combined family income ranged from $7,500 to $82,500 with a mean income of $48,520. The revised Duncan Prestige Scale (Stevens & Featherman, 1980) was used to determine socioeconomic status of the family. The mean prestige rating was 41.34 (range: 18 to 80.4, SD=22.988) for mothers and 44.97 (range: 18.5 to 82.5, SD=19.537) for fathers, indicating representation from all levels of socioeconomic status.

Measurement of representations of attachment relationships

The frequency of agreement between the classifications derived from the Adult Attachment Rating and Classification Systems (ARCS) and the Adult Attachment Interview Qrset (AAQ) is presented in Table 1 below. In both adult

attachment classification methods, 39 subjects (79.6%) were classified as secure whereas 10 (20.4%) received insecure classifications. Four (8.2%) were classified as Dismissing and 6 (12.2%) were classified as Preoccupied.

Table 1 Contingency Analysis of Attachment Classifications derived from the Adult Attachment Rating and Classification Systems (ARCS) and Adult Attachment Interview Q-set (AAQ)

ARCS Classifications ______Dismissing Secure Preoccupied AAQ, Classifications Dismissing 4 2 0 Secure 2 39 0 Preoccupied 0 0 6 62

Forty-seven subjects (96%) received identical classifications using these two classification methods. Of the two subjects who received different classifications, both changes occurred between the secure and dismissing attachment groups. Classifications derived from the ARCS and AAQ.methods were significantly and positively correlated (Y=.874, p=.0001). The coefficient of concordance (kappa) for these two sets of classifications was .870, indicating strong agreement between these two methods of measuring representations of attachment relationships (Wilkinson, 1990).

Correlations between the Mother-Father-Peer (MFP) scales, Adult attachment classifications, Probable Childhood Experience scales (ARCS), States of Mind scales (ARCS), and AAQ.dimensions are presented in Tables 2 through 5. When the pattern of associations for the two MFP scoring methods were compared in each of these tables, few differences were found. That is, if significant associations emerged between a bipolar Epstein scale and the adult attachment classifications, one or more of the corresponding Ricks' scales were usually significant. Likewise, if a significant relation was found between an Epstein scale and one of the ARCS scales or an AAQ.dimension, at least one of the Ricks' counterparts were usually significant. Therefore, the remaining discussion in this section will focus primarily on one set of MFP scales, Ricks' scales, and will only include major discrepancies or inconsistencies found between the two MFP scoring methods. 63

Table 2

Associations between Adult Attachment Classifications and Mother-Father-Peer Scales

Secure/Insecure Classification ______ARCS AAQ

Epstein's MFP Scales

Mother acceptance- rejection .176 .299* Mother independence- overprotection .222 .266*

Mother idealization .086 .095 Father acceptance- rejection -.014 .152 Father independence- overprotection .328* .333*

Father idealization .068 .132 Peer acceptance- rejection .225 .123

Rick's MFP Scales

Mother acceptance .162 .282* Mother rejection .177 -.291*

Mother acceptance total .176 .299*

Mother independence .416** .450**

Mother overprotection .094 .051 Mother independence total .222 .266* Table 2 (continued)

Secure/Insecure Classification ARCS AAQ,

Mother idealization .086 .095 Father acceptance .007 .168

Father rejection .034 -.122 Father acceptance total -.014 .152

Father independence .346* .384** Father overprotection -.166 -.122 Father independence total .328* .333*

Father idealization .068 .132 Peer rejection -.230 -.161

Peer acceptance .199 .073 Peer acceptance total .225 .123

*p< .05 **p< .01 ***p< .001 t=trend 65

As illustrated in Table 2, one difference emerged in the pattern of correlations between attachment classifications and the scores for Epstein's and Ricks' scales. Epstein's bipolar scale of Mother independence- overprotection was unrelated, whereas one of the Ricks' scales (Mother independence) was significantly related, to adult attachment classification. With the ARCS method, significant correlations were found between adult attachment status and Mother and Father independence as well as Father independence total. Mothers classified as secure with this method were more likely to recall both parents as encouraging autonomy during childhood than those judged as insecure. With the AAQmethod, significant associations were found for Mother acceptance, Mother rejection, Mother acceptance total, Mother independence, Father independence, and Father independence total.

Mothers classified as secure with this method were more likely to recall both of their parents as encouraging autonomy during childhood, but were also more likely to recall their own mothers as more accepting and less rejecting than those judged as insecure. Table 3

Associations between Mother-Father-Peer Scales and Probable Childhood Experience Scales of the Adult Attachment Rating and Classification Systems

Probable Childhood Experience Scales (ARCS) Loving Loving Rejecting Rejecting Involving Involving Pressured to Pressured to Neglecting Neglecting mother father mother father mother father achieve (M) achieve (Fi mother father

Epstein's MFP Scales

Mother acceptance- rejection . 760*** .262* -.818*** -.208 -. 4 3 7 * * .102 -.083 .095 .099 .024

Mother independence- overprotection .530*** .25 51 -.544*** -.192 -.384* .037 -.023 -.178 .097 .138

Mother idealization .545* .133 -.493*** .024 -.469* .012 -.216 .124 -.105 .091

Father acceptance- rejection .271* .741*** -.234 -.685*** -.157 -.046 .093 -.144 .110 -.433**

Father independence overprotection .288* .487*** -.251* -.487*** -.309* -.104 .201 -.378* -.043 -.139

Father idealization .117 .542*** -.041 -.422** -.264* -.124 -.132 -.162 -.039 -.310*

Peer acceptance- rejection .264* .070 -.385** -.242* -.075 -.105 .191 -.187 -.308* .076

Rick's MFP Scales

Mother acceptance .719*** .255* -.749*** -.189 -.486*** .107 -.052 .090 .085 .096

Mother rejection -.739*** -.248* .822*** .210 .352** -.089 .107 -.091 -.105 .053

Mother acceptance total .760*** .262* -.818*** -.208 -.437** .102 -.083 .095 .099 .024 Table 3 (continued)

Probable Childhood Experience Scales (ARCS) Loving Loving Rejecting Rejecting Involving Involving Pressured to Pressured to Neglecting Neglecting m o th er father m o th er father mother lather achieve (M) achieve (F) m other father

Mother independence .664*** .288* -.686*** -.2631 -.404** -.010 -.058 -.219 .007 .147

Mother overprotection -.193 -.133 .191 .042 .237 -.085 -.029 .068 -.181 -.083

Mother independence total .530*** .2S5t -.544*** -.192 -.384** .037 -.023 -.177 .097 .138

Mother idealization .545*** .133 -.493*** .024 -.469*** .012 -.216 .124 -.105 .091

Father acceptance .320* .749*** -.248c -.633*** -.199 .061 .060 -.094 .111 -.419**

Father rejection -.197 -.669*** .198 .681*** .099 .152 -.119 .182 -.100 .410**

Father acceptance total .27 It .741*** -.234 -.685*** -.157 -.046 .093 -.144 .110 -.433**

Father independence .375** .649*** -.294** -.563*** -.410** -.179 .128 -.232 -.077 -.151

Father overprotection -.043 -.051 .084 .175 .036 -.049 -.225 .435** -.024 .064

Father independence total .288* .487*** -.250t -.487*** -.309* -.104 .201 -.378** -.043 -.139

Father idealization .117 .542*** -.041 -.422** -.2641 -.124 -.132 -.162 -.039 -.310*

Peer rejection -.234 -.038 .387** •253t .137 .273t -.177 •273t .287* -.109

Peer acceptance .2 7 1( .097 -.349** -.209 -.006 .075 .188 -.083 -.302* .035

Peer acceptance total .2 6 4 t .070 -.385** -.242t -.075 -.104 .191 -.187 -.308* .076

*p< .OS **p< .01 ***p< .001 t» trend O' ■ - J As shown in Table 3, several significant associations were found between Ricks' MFP scales and the Probable Childhood Experience scales of the Adult Attachment Rating and Classification Systems (ARCS). High positive correlates for each of the ARCS Loving scales were Acceptance, Acceptance total, Independence and Idealization, whereas a high negative correlate for

each of the Loving scales was Rejection. Likewise, a high positive correlate

of the each of the ARCS Rejecting scales was Rejection, whereas negative correlates were Acceptance, Acceptance total, and Independence. In other words, subjects who recalled both parents as more accepting, less rejecting, and more supportive of autonomy described more active loving memories with each parent during the Adult Attachment Interview (AAI). These subjects also idealized their parents and described fewer experiences with each parent in which her attachment-related concerns were rejected or avoided (e. g., the parent punished the child when hurt or ridiculed the child when distressed). A few moderate associations were found between Ricks' scales and some of the other Childhood Experience scales (ARCS); however, the pattern was more varied between each parent. For example, a positive correlate for the

ARCS Involving mother scale was Mother rejection, and negative correlates were Mother acceptance, acceptance total, independence, independence total, and idealization. No significant correlates emerged for the Involving father scale. Mothers who idealized their own mothers and recalled more maternal acceptance, less maternal rejection, and more maternal support of autonomy during childhood, were less likely to describe involving and/or role reversing experiences with their mothers during the AAI (e. g., demanding that the child center her attention on parental needs, expecting the child to 69

continually please the parent, or consistently drawing the child's attention and concern toward the parent's state of physical and/or psychological well­ being). A positive correlate for the ARCS Pressured to achieve (father) scale was Father overprotection, but no significant correlates were found for the Pressured to achieve (mother) scale. Mothers who recalled more paternal overprotection were more likely to describe experiences in which their fathers pushed them to excel (academically, intellectually, and/or in other areas), and either expected them to care for themselves or assume adult responsibilities as children or adolescents. A positive correlate for the

Neglecting father scale was Father rejection, whereas the negative correlates were Father acceptance, acceptance total, and idealization. No significant

correlates for the Neglecting mother scale were found among any of the MFP

mother scales. Mothers with less paternal idealization, and those who recalled their fathers as more rejecting and less accepting, were more likely to describe memories of their fathers as inattentive, uninvolved, or psychologically inaccessible when present in the home. Together, these results indicate a relatively high degree of consistency between self-report and interview-based ratings of memories forcertain aspects of childhood relationships with parents, namely parental warmth/love and promotion of autonomy. The results also reveal less consistency between these types of measures for other aspects of childhood relationships with parents, such as neglect and role reversal. Table 4 Associations between Mother-Father-Peer Scales and States of Mind Scales of the Adult Attachment Rating and

Classification Systems

States of Mind Scales (ARCS) Idealizing Idealizing Involving anger Involving anger Derogation or Lack of Metacognltlve Passivity of Unresolved Unresolved Coherency mother father______m other______father______attachment recall processes______thought ______loss______trauma _____ o f mind

Epstein's MFP scales

Mother acceptance- rejecdon .234 .061 -.409** .027 .053 -.177 -.055 -.289* .140 -.501**’ .290*

Mother Independence- overprotection .027 -.048 -.198 -.049 .060 -.200 .096 -.231 .122 -.337* .272*

Mother idealization .148 .026 -.352* -.028 -.042 .138 -.014 -.226 .117 -.199 .175 Father acceptance- rejecdon .163 .329* -.131 -.184 .081 -.025 -.170 -.181 -.002 -.329* .068

Father independence- overprotecdon -.145 -.120 -.095 -.210 -.057 -.200 .051 -.203 -.010 -.300* .268*

Father Ideallzadon .146 .256* -.254* -.331* -.091 .ISO -.118 -.150 .010 -.113 .030

Peer acceptance- rejecdon -.068 -.060 -.139 .038 .021 .008 -.186 -.101 .103 -.132 .123

Rick's MFP scales

Mother acceptance .180 .035 -.403** .031 -.015 -.121 -.074 -.243' .141 -.421** .235

Mother rejecdon -.271* -.084 .382** -.021 -.118 .220 .030 .312* -.128 .542**’ -.322*

Mother acceptance total .234 .061 -.409** .027 .053 -.177 -.055 -.289* .140 -.501**’ .290*

Mother independence -.005 -.129 -.322* -.085 -.044 -.26 It .090 -.293* .153 -.411** .379**

Mother ~v| overprotecdon -.060 -.069 -.022 -.012 -.173 .059 -.073 .079 -.045 .137 -.051 o Table 4 (continued)

States of Mind Scales (ARCS) Idealizing Idealizing Involving anger Involving anger Derogation of Lack of MetacogniUve Passivity of Unresolved Unresolved Coherency mother Father mother Father attachment recall processes thoueht loss trauma of mind

Mother independence total .027 -.048 -.198 -.049 .060 -.200 .096 -.231 .122 -.337* .272*

Mother idealization .148 .026 -.352* -.028 -.042 .138 -.014 -.226 .117 -.199 .175

Father acceptance .168 .331* -.165 -.185 .033 -.015 -.188 -.163 .022 -.346* .076

Father rejection -.144 -.300* .085 .167 -.124 .033 .137 .185 .027 .284* -.05 S

Father acceptance total .163 .329* -.131 -.184 .081 -.025 -.170 -.181 -.002 -.329* .068

Father independence -.226 -.061 -.177 -.220 -.176 -.218 .026 -.269* -.144 -.311* .332*

Father overprotection -.031 .157 -.063 .108 -.141 .092 -.066 .023 -.188 .160 -.063

Father independence total -.145 -.120 -.095 -.210 -.057 -.200 .051 -.203 -.010 -.300* .268*

Father Idealization .146 .2S61 -.254* -.331* -.091 .150 -.118 -.150 .010 -.113 .030

Peer rejection .026 .063 .165 .037 -.041 -.052 .189 .190 -.094 .139 -.120

Peer acceptance -.105 -.052 -.101 .111 -.001 -.037 -.165 -.002 .103 -.112 .115

Peer acceptance total -.068 -.060 -.139 .038 .021 .008 -.186 -.101 .103 -.132 .123

*p< .05 **p< .01 ***p< .001 t-tre n d 72

In comparison to the Childhood Experience scales, few significant associations were found overall between Ricks' MFP scales and the States of Mind scales of the Adult Attachment Rating and Classification Systems (Table

4). Interestingly, both Idealizing scales of the ARCS were unrelated to their respective Ricks’ counterparts (Mother idealization and Father idealization).

The pattern of associations between Ricks' scales and the other States of mind scales was slightly varied for each parent. Significant correlates for Mother acceptance were Involving anger (mother) and Unresolved trauma. Significant correlates for Mother rejection, Mother acceptance total and

Mother independence were Involving anger (mother), Passivity of thought, Unresolved trauma, and Coherency of mind. A significant (negative) correlate for Mother independence total was Unresolved trauma. In other

words, subjects who recalled their own mothers as more accepting, less rejecting, and more supportive of autonomy during childhood expressed lower levels of current involving anger toward their mothers and

demonstrated less passivity of thought during the AAI. Further, these mothers were less disorganized and/or less disoriented when discussing traumatic experiences with attachment figures, and evidenced greater integration and/or coherency overall during the attachment interview. Significant correlates of Father acceptance, Father rejection, and Father acceptance total were Idealizing father and Unresolved trauma.

Significant correlates for Father independence and were Unresolved trauma and Coherency of Mind, and a significant correlate for Father Independence total was Unresolved Trauma. In other words, subjects who recalled their own fathers as more accepting, less rejecting, and more supportive of autonomy during childhood were more likely to evidence idealization of their fathers 73 during the AAI. These mothers were also less disorganized and/or less disoriented when discussing traumatic experiences with attachment figures, and evidenced greater integration and/or coherency overall during the attachment interview.

Together, these results indicate consistency betweensome interview- based ratings of "states of mind" and self-report measures intended to reflect specific representations of parents (such as Mother acceptance) and "higher- order" postulates (such as Father acceptance total). However, the data also reveal a discrepancy between interview-based ratings and self-report measures of parent idealization.

As shown in Table 5 on pages 74& 75, no significant associations emerged between Ricks' MFP scales and the Repressed-preoccupied dimension of the Adult Attachment Interview Q-set (AAQ), but almost all of the scales were significantly related to the Secure-anxious dimension. The exceptions were Mother and Father overprotection. As compared to mothers with anxious strategies, those who displayed a secure attachment strategy during the Adult Attachment Interview idealized both parents and recalled both parents as more accepting, less rejecting, and more supportive of autonomy during childhood. 74

Table 5

Associations between Adult Attachment Interview Q.-set Dimensions and Mother-Father-Peer Scales

AAQ. Dimensions Secure-anxious Repressed-preoccupied

Epstein's MFP Scales

Mother acceptance- rejection .455*** .037

Mother independence- overprotection .411** -.024

Mother idealization .282* .239* Father acceptance- rejection .361** .016 Father independence- overprotection .452*** .060 Father idealization .292* .208 Peer acceptance- rejection .056 .051 Rick's MFP Scales

Mother acceptance .424** .083

Mother rejection .450*** .012

Mother acceptance total .455*** .037 Mother independence .488*** .074

Mother overprotection .182 .047

Mother independence total .411** .024

Mother idealization .282* .239* 75

Table 5 (continued)

AAQ. Dimensions Secure-anxious Repressed-preoccupied

Father acceptance .394** .030

Father rejection -.297* .001 Father acceptance total .361** .016 Father independence .557*** -.086

Father overprotection -.114 -.002 Father independence total .452*** -.060

Father idealization .292* .208 Peer rejection -.056 -.149 Peer acceptance .051 -.052 Peer acceptance total .056 .051

*p< .05 **p< .01 ***p< .001 t=trend 76

Relations between demographic and maior study variables

The correlations between demographic variables and each dimension of maternal separation anxiety (MSA) are presented in Table 6 below. Maternal age was significantly and negatively related to both the prenatal and postpartum scores for MSA Subscale 1. Maternal education was significantly and negatively related to both sets of scores for Subscales 1 and 2, as well as the postpartum scores for Subscale 3. Family income was significantly and negatively related to only the prenatal scores for MSA Subscale 3. Maternal occupational prestige was significantly and negatively related to both sets of scores for Subscale 3, as well as the postpartum scores for Subscale 1.

Table 6 Associations between Dimensions of Maternal Separation Anxiety and Demographic Variables

Time 1 Time 2 MSA Subscale 1 2 3 1 2 3

Maternal age -.307* -.256* -.161 -.281* -.200 -.010

Maternal education -.319* -.392** -.250* -.420** -.394** -.404** Family income -.207 -.174 -.300* -.149 -.163 -.183

Maternal occupational prestige rating -.253t -.263* -.305* -.327* -.198 -.365**

*p< .05 **p<.01 ***p<.001 t=trend 77

In other words, younger mothers and those with fewer years of education reported higher levels of maternal separation anxiety during the third trimester of pregnancy and at 6 to 8 weeks postpartum. Mothers who were employed in less prestigious occupations prenatally also reported higher levels of maternal separation anxiety at 6 to 8 weeks. Mothers with fewer years of education reported greater concern about the effects of separation on the child during the third trimester of pregnancy and at 6 to 8 weeks postpartum. Mothers with higher family incomes had lower levels of employment-related separation concerns prenatally, whereas mothers in less prestigious occupations prenatally had higher levels of employment-related separation concerns during the third trimester of pregnancy and at 6 to 8 weeks postpartum.

Results of t-test analyses revealed no significant differences between the secure and insecure groups with respect to education (t= -1.796, p= .080) and combined family income (t= -.092, p=.927). However, the groups differed significantly with respect to age (t= -3.04, p=.004) and occupational prestige (t= -2.66, p=.028). Mothers in the insecure group were younger (mean=25 years, SD=3.396) and employed in less prestigious jobs prenatally (mean=27.24, SD=19.897) than those in the secure group (mean age=29.1 years, SD=3.848; mean occupational prestige=44.956, SD=22.557). 78

Maternal Separation Anxiety

In order to determine the stability of each dimension of maternal separation anxiety, paired t-tests were conducted using the differences between the prenatal and postpartum scores for each MSA Subscale. Results from these analyses are presented in Table 7 below.

Table 7 Descriptive Statistics and Paired t-test Results for Dimensions of Maternal

Separation Anxiety

Time 1 Time 2 Mean SD Mean SD t P

Maternal Separation Anxiety (Subscale 1) 23.483 3.720 23.639 4.607 .371 .707 Perception of Separation Effects on the Child (Subscale 2) 14.878 3.866 15.020 3.843 .378 .707 Employment-Related Separation Concerns (Subscale 3) 23.633 5.310 25.041 5.123 3.131 .003

Although no significant differences were found in the scores for MSA

Subscales 1 and 2, a significance difference was found in the scores for Subscale 3. That is, there were no significant increases or decreases in the mothers' levels of maternal separation anxiety and their perceptions of separation effects on the child. However, the mothers' concerns about employment-related separations were greater at 6 to 8 weeks postpartum. Table 8

Differences in Dimensions of Maternal Separation Anxiety among Mothers in Secure and Insecure Groups

All subjects Secure Insecure Mean SD Mean SD Mean SD t P Time 1 MSA (Prenatal) Maternal Separation Anxiety (Subscale 1) 23.483 3.720 23.274 3.749 24.300 3.677 .775 .442 Perception of Separation Effects on the Child (Subscale 2) 14.878 3.866 14.667 4.009 15.700 3.302 .751 .457 Employment-Related Separation Concerns (Subscale 3) 23.633 5.310 23.462 5.438 24.300 4.990 .442 .661 Time 2 MSA (6 to 8 weeks) Maternal Separation Anxiety (Subscale 1) 23.639 4.607 22.897 4.582 26.533 3.584 2.327 .024 Perception of Separation Effects on the Child (Subscale 2) 15.020 3.843 14.949 3.927 15.300 3.683 .255 .800 Employment-Related Separation Concerns (Subscate 3) 25.041 5.123 25.205 5.302 24.400 4.551 -.440 .662 As shown in Table 8, t-test results revealed no significant differences between the secure and insecure groups with respect to their prenatal and postpartum scores for MSA Subscales 2 and 3. Although no significant differences were found for prenatal scores, significant differences were found in postpartum scores for Subscale 1. That is, mothers' perceptions of separation effects on the child, as well as levels of concern about employment-related separations, were similar in both attachment groups during the third trimester of pregnancy and at 6 to 8 weeks postpartum. Although the levels of maternal separation anxiety were similar in both attachment groups prenatally, mothers in the insecure group reported higher levels of maternal separation anxiety at 6 to 8 weeks postpartum. 81

Table 9 Associations between Dimensions of Maternal Separation Anxiety and Mother-Father-Peer Scales

Time 1 Time 2 ______MSA Subscale 1 2 3 1 2 3

Epstein's MFP Scales

Mother acceptance- rejection -.120 -.076 .138 -.222 -.182 .078 Mother independence- overprotection -.256* -.158 -.018 -.353* -.169 .046 Mother idealization .032 .021 .015 .093 -.056 -.084 Father acceptance- rejection -.071 .080 .078 -.124 .083 .149 Father independence- overprotection -.338* -.049 -.015 -.423** -.039 .113

Father idealization .073 .308* .072 .119 .2511 .148

Peer acceptance- rejection -.151 -.180 -.124 -.206 -.111 -.067 Rick’s MFP Scales

Mother acceptance -.063 -.045 .170 -.103 -.125 .109 Mother rejection .170 .101 -.093 .328* .227 -.040 Mother acceptance total -.120 -.076 .138 -.222 -.182 .078

Mother independence -.2431 -.159 -.054 -.426** -.170 -.011 Mother overprotection .192 .106 -.033 .148 .114 -.103 Mother independence total -.256* -.158 -.018 -.353* -.169 .046

Mother idealization .032 .021 .015 .093 -.056 -.084 82

Table 9 (continued)

Time 1 Time 2 MSA Subscale 1 2 3 1 2 3

Father acceptance -.027 .127 .154 -.065 .108 .189 Father rejection .110 -.026 .008 .174 -.051 -.095 Father acceptance total -.071 .080 .078 -.124 .083 .149 Father independence -.301* .022 .025 -.359* .044 .088

Father overprotection .252t .133 .067 .341* .143 -.103 Father independence total -.338* -.049 -.015 -.423** -.039 .113

Father idealization .073 .308* .072 .119 .251* .148 Peer rejection .144 .135 .112 .214 .136 .060

Peer acceptance -.145 -.209 -.126 -.179 -.076 -.069 Peer acceptance total -.151 -.180 -.124 -.206 -.111 -.067

*p< .05 **p< .01 ***p< .001 t=trend 83

The correlations between each dimension of maternal separation anxiety and the Mother-Father-Peer (MFP) scales are presented in Table 9.

Significant prenatal correlates were found for MSA Subscale 1 (Father independence and Father independence total) and Subscale 2 (Father idealization), whereas no significant prenatal correlates were found for Subscale 3. Significant postpartum correlates were found only for MSA Subscale 1 (Mother rejection, Mother independence and independence total, Father overprotection, Father independence and independence total).

Mothers who recalled their fathers as less supportive of autonomy during childhood reported higher prenatal levels of maternal separation anxiety. Those who idealized their fathers reported greater concern about the effects of separation on the child prenatally. Those who recalled both parents as less supportive of autonomy, as well as recalled their mothers as more rejecting and fathers as overprotective during childhood, reported higher levels of maternal separation anxiety at 6 to 8 weeks postpartum. Because no significant associations were found between MSA Subscale scores and parental acceptance scores of the MFP, correlations between the

MSA Subscale scores and the Loving scales (ARCS Probable Childhood Experience scales) were also examined. As shown in Table 13 (Appendix D), no significant association was between found any of the MSA Subscale scores and the ratings from each Loving scale. The only significant (negative) association was between the prenatal scores for MSA Subscale 1 and the ratings for the Neglecting mother scale. Mothers with higher levels of maternal separation anxiety prenatally described more childhood memories in which their mothers were inattentive, uninvolved, or psychologically inaccessible when present at home. 84

Table 10

Associations between Adult Attachment Classifications, Coherency of Mind, and Dimensions of Maternal Separation Anxiety

Secure/Insecure Classification Coherency of Mind ARCS AAQ ARCS

Time 1 MSA (Prenatal)

Maternal Separation Anxiety (Subscale 1) -.112 -.048 -.056 Perception of Separation- effects on the child (Subscale 2) -.109 -.069 -.023 Employment-Related Separation Concerns (Subscale 3) -.064 .109 -.047 Time 2 MSA (6 to 8 weeks)

Maternal Separation Anxiety (Subscale 1) -.321* -.314* -.311* Perception of Separation- effects on the child (Subscale 2) -.037 -.050 -.154 Employment-Related Separation Concerns (Subscale 3) .064 .134 .079

*p< .05 **p< .01 ***p< .001 t=trend 85

Correlations between each dimension of maternal separation anxiety and adult attachment status are presented in Table 10. As shown in this table, both the prenatal and postpartum scores for MSA Subscales 2 and 3 were unrelated to adult attachment classification. Although the prenatal scores were unrelated, the postpartum scores for Subscale 1 were significantly and negatively related attachment classification. At 6 to 8 weeks postpartum, mothers in the insecure group reported higher levels of maternal separation anxiety than those in the secure group.

Because attachment classifications are global, indirect measures of integration, the correlations between both sets of MSA Subscale scores and the scale ratings for Coherency of Mind (ARCS), were also examined. As also shown in Table 10, only the postpartum scores for Subscale 1 were significantly and negatively related to the Coherency of Mind ratings.

Mothers who evidenced greater integration of attachment-related experiences and relationships reported lower levels of maternal separation anxiety, whereas those who demonstrated less integration of the same reported higher levels of maternal separation anxiety at 6 to 8 weeks postpartum. 86

Mothers' perceptions of their infants

As shown in Table 11 (page 87), 49% of the mothers (n=24) used all positive adjectives and phrases, whereas 51% (n=25) used one or more mixed expressions, in describing their infants and their relationships with them. Of the 25 mothers who used mixed descriptions during the Mothers' Perceptions Her Baby Interview (MPBI), only 3 responded to all interview questions with neutral or negative phrases. Six used mixed expressions in response to two questions, whereas the majority (n=16) gave mixed responses to one question in the interview. All positive expressions were used by 71.4% (n=35) of the subjects to delineate the unique qualities of their infant and their current relationship with the infant, whereas 81.6% (n= 40) used all positive words to depict their infants' feelings toward them. Mixed descriptions were used by 28.6% (n=14) of the subjects to delineate infant qualities and to describe their current relationship with the infant, whereas 18.4% (n = 9) used mixed phrases to depict their infants' feelings toward them. 87

Table 11

Descriptive Statistics and Chi-square Results for Aspects of Mothers' Perceptions of their Infants

Frequencies Secure/Insecure All positive Mixed *2 (df=l) P

Mothers' Perceptions

Unique Qualities 35 14 .452 .501

Perception of the relationship 35 14 .452 .501 Feelings toward mother 40 9 1.134 .287 MPBI total 24 25 .022 .801

Mothers frequently used adjectives such as happy, content, easy-going, cute, beautiful, sweet, adorable, alert, curious and attentive to describe the unique qualities of their infants. Adjectives such as caring, loving, and dose were given most often in descriptions of the mothers' current relationship with their infants. Phrases such as "he/she feels secure," "he/she trusts me,"

"he/she knows me or knows I am his/her caretaker," and "he/she is happy with me," were reported most often in descriptions of the infant's feelings toward the mother. Several patterns emerged when the mothers were asked how the infants' unique qualities have influenced their relationship with the infant. Mothers 88 perceived that easy-going, contented qualities helped to make the relationship with their infant more comfortable and relaxed. Alert, curious qualities were perceived to enhance interaction with their infants and to encourage them to be more responsive and provide more attention to their infants. Generally, positive qualities were perceived to enhance mothers' feelings of connectedness with their infants and to engender positive feelings in themselves (e. g., "makes me feel more secure as a parent," "makes you melt," "makes you feel good"), whereas mixed qualities were perceived to arouse feelings of worry and a desire to better anticipate their infants' needs. Almost all of the mothers' ratings of the degree of match between their prenatal expectations and postpartum perceptions ranged between 5 and 7.

Forty-five subjects (91.8%) rated the degree of match between their prenatal expectations and postpartum perceptions of their infants at 5 or above. Similarly, 44 subjects (89.9%) rated the degree of match between their prenatal expectations and postpartum perceptions of their relationship with their infant at 5 or above. Five mothers (10.2%) indicated that their current perceptions of their infants and their relationships werebetter than expected and five (10.2%) perceived substantial differences (ratings of 4 and below) between their prenatal expectations and current perceptions of the same.

Among those who indicated differences (ratings of 6 and below) between their prenatal expectations and current perceptions of infant qualities, most focused on specific infant responses, such as more or less responsiveness and more or less crying. Three patterns emerged among those who indicated differences between prenatal expectations and current perceptions of their relationship with their infants. First, mothers focused on routines or 89 schedules being more demanding or less established than anticipated. Second, they emphasized unexpected or unanticipated infant responses such as crying only when angiy, being temperamental, not recognizing faces, and not wanting to be held constantly. Third, mothers focused on their beliefs and feelings about the relationship, such as more or less intense feelings of protection, more or less intense feelings of possessiveness, the infant's lack of awareness of the mother as a person, and believing that the relationship started earlier than 6 to 8 weeks of age. In order to retain adequate cell sizes, MPBI total scores of 0, 1, and 2 were combined into one group for chi-square analyses. These results are also presented in Table 11 (page 87). No significant differences were found between the secure and insecure groups with respect to their total MPBI scores. Overall, mothers in the secure group were neither more nor less likely to use positive descriptions than those in the insecure group. Chi- square analyses also revealed no significant differences between these two groups of mothers with respect to their perceptions of the unique qualities of their infants, perceptions of their relationship with their infants, and perceptions of their infant's feelings toward them. Again, mothers in the secure group were neither more nor less likely to use all positive descriptions in response to the same than those in the insecure group. In addition, t-test results revealed no significant differences between these two groups with respect to parental ratings of the match between prenatal and postnatal perceptions of the infant (t=.382, p=.704) and perceptions of the relationship (t= -.261, p=.795). From the mothers' perspective, their prenatal expectations matched their perceptions of the infant and their relationship with the infant at 6 to 8 weeks relatively well. 90

Table 12

Associations between Dimensions of Maternal Separation Anxiety and Mothers' Perceptions of their Infants

Time 1 Time 2 MSA Subscale 1 2 3 1 2 3 Mothers' perceptions

Unique qualities -.355* -.091 .188 -.288* -.104 .104 Perception of relationship -.036 -.127 -.130 -.024 -.163 -.120

Feelings toward mother -.272* -.153 -,244t -.307* .016 -.214 Overall positiveness (MPBI total) -.315* -.176 -.077 -.291* -.128 -.101

*P< .05 **p< .01 ***p< .001 t=trend

Correlations between aspects of mother's perceptions of their infants and dimensions of maternal separation anxiety are presented in Table 12 above. Maternal perceptions were only significantly related to the scores for

MSA Subscale 1. Perceptions of the infant's feelings toward the mother were significantly and negatively related to the postpartum scores for Subscale 1. That is, mothers with lower postpartum levels of maternal separation anxiety were more likely to use all positive phrases to describe their infants' feelings toward them. Perceptions of unique infant qualities, as well as overall positiveness of perception (MPBI total), were significantly and negatively related to both sets of scores for MSA Subscale 1. Mothers with lower prenatal and postpartum levels of maternal separation anxiety were more likely to use all positive phrases to describe their perceptions of the unique qualities of their infants than those with higher levels of separation anxiety.

Additionally, mothers with lower levels of maternal separation anxiety were more positive overall in their perceptions than those with higher levels of separation anxiety. CHAPTER VI DISCUSSION

The discussion is presented in three parts in this chapter according to the primary objectives of the study. The discussion begins with a consideration of the equivalence of three measures of adult attachment representations selected for comparison in this investigation. A discussion of the dimensions of maternal separation anxiety is presented next, and includes consideration of their associations with adult attachment representations. The chapter concludes with a discussion of mothers' perceptions of their infants which incorporates consideration of their relations to adult attachment representations.

Measurement of representations of attachment relationships

Based on the administration of the Adult Attachment Interview during the third trimester of pregnancy, strong concordance was found between two classification methods used to determine adult attachment representations, the Adult Attachment Rating and Classification Systems (ARCS) and the Adult Attachment Interview Q-set (AAQ). Overall, the majority of classifications were identical using both of these methods with few mismatches identified. These findings, which are consistent with Kobak, et al. (in press), suggest that both of these classification methods provide comparable results when

92 93

attempting to determine secure and insecure representations of attachment relationships in mothers expecting their first child.

The results of the study revealed significant associations between both sets of Mother-Father-Peer scales (Epstein's and Ricks') and adult attachment status using the ARCS and AAQ.classification methods. These findings are consistent with Cole & Kobak (1990) who also found significant correlations between Ricks' MFP scales and attachment status using the AAQ, but are inconsistent with Benoit, et al. (1989) who found no relation between

Epstein's MFP scales and attachment status using the ARCS method. In the present study, the Adult Attachment Interview was administered before the Mother-Father-Peer scale, and this sequence of administration may have influenced the mothers' responses to the questionnaire. The Adult Attachment Interview is designed to "shock the psyche," which in turn, is presumed to activate, deactivate, or hyperactivate attachment-related memories (George, personal communication). That is, attachment-related information is believed to be accessible at various levels throughout the interview. Activation or hyperactivation may have persisted in some mothers in the study sample once the attachment interview was completed. Subsequently, the continuation of these processes could have permitted or enhanced access to childhood memories of relationships with parents as they completed the questionnaire. Generally, the pattern of associations between the MFP scales and adult attachment status provides support for theoretical propositions from attachment theory that delineate important contributors to secure and insecure representations of attachment relationships. For example, experiences with accessible and sensitive caretakers are believed to be 94 important to the formation of a secure working model, whereas experiences with inaccessible and insensitive caretakers are believed to contribute to the development of an insecure working model of attachment relationships. In this study, childhood memories of parental warmth and love (acceptance) were significantly, positively associated with a secure attachment representation, and childhood memories of parental rejection were significantly, negatively related to an insecure attachment representation. Interestingly, results from this study also revealed significant associations between attachment status and childhood memories of the extent to which parents encouraged independence. These findings suggest that parental support and/or discouragement of autonomy may be a salient, unique aspect of an adult's working model of attachment relationships. These findings are particularly noteworthy considering that some studies have emphasized parental acceptance and/or rejection as the major conceptual link in differentiating secure from insecure attachment representations

(Biringen, 1990; Jacobvitz, et al., 1991). The findings in the present study suggest that parental support and/or discouragement of autonomy may also be an important conceptual link to consider when delineating components that may differentiate secure and insecure representations of attachment relationships. The MFP scales appear to provide information comparable to some, but not all, of the scales in the Adult Attachment Rating and Classification Systems (ARCS). Overlap of information appears to be the greatest between scales designed to determine probable experiences with parents during childhood, in particular the extent of parental warmth, love, and rejection that the individual may have experienced. In this study, correlations with the highest magnitudes emerged between ratings for the ARCS Loving scales and scores for Ricks' (MFP) parental acceptance scales (range .719 to .760), as well as between ratings for the ARCS Rejecting scales and scores for Ricks' parental rejection scales (range .681 to .822). The next greatest overlap appears to exist between scales designed to determine the extent of parental warmth/love and support of autonomy that may have been experienced during childhood. Correlations with the second highest magnitude emerged between the ratings for the ARCS Loving scales and scores for Ricks' parental independence scales (range .487 to .664). Little to no overlap appears to exist for scales designed to determine other aspects of childhood experiences with parents, such as neglect and role reversal. The direction and magnitude of the correlations between the scores for Ricks' MFP scales and ratings for the other ARCS scales (Neglecting, Involving, and Pressured to Achieve) were found to be varied for each parent and showed no consistent pattern of associations. Little to no overlap also seems to exist between the MFP scales and the ARCS scales designed to determine "states of mind." Few significant associations were found overall between these measures and the pattern of correlations revealed that the degree of consistency was less than that found between the MFP scales and Probable Childhood Experience scales. These findings indicate that these measures may provide information about different aspects or components of adult attachment representations. In other words, the MFP scales and the ARCS "states of mind" scales do not appear to provide comparable information about psychological processes that may influence various kinds of attachment organization in adults. Interestingly, no significant associations were found between the MFP scales and the ARCS ratings for parent idealization, indicating that these measures do not provide equivalent information about this psychological process associated with recall of childhood experiences with parents. These findings may be due, in part, to the way in which idealization is determined in each measure. ARCS ratings of idealization are based on discrepancies between abstract and specific levels of information provided about early experiences with parents and the extent to which an individual persistently attempts to maintain a "perfect" image of the parent during the attachment interview. Although awareness and recognition are considered in assigning the final scale score, ratings are made for persons who are unaware of and do not acknowledge idealization of their parents. With the MFP, there is no independent way to determine the extent to which responses are influenced by attempts to retain a perfect image of the parent or by tendencies to view one's childhood through "rose colored glasses." That is, mothers who disclaim or are unaware of idealization may respond to these items in a similar way as those who acknowledge or recognize idealization of their parents. As a result, very few or no differences may emerge between mothers who do and do not idealize their parents. Although several significant associations were found between the MFP scales and the secure-anxious dimension of the Adult Attachment Interview Qrset (AAQ), none were found for the repressed-preoccupied dimension. These results are somewhat unexpected because the secure-anxious attachment strategy is intended to reflect a secure, and the repressed- preoccupied attachment strategy an insecure, working model of attachment relationships. The results could be due, in part, to the proportions within 97 each attachment group in the study sample. Because most subjects were classified as secure and only 10 were classified as insecure, there were fewer mothers overall who used secondary strategies during the attachment interview. Among those classified as insecure, the mothers were almost equally divided between the dismissing and preoccupied categories. That is, the proportion of mothers using a repressed strategy was almost equal to the proportion using a preoccupied strategy, and the number of mothers using each of these secondary strategies was very small. In general, the MFP appears to provide information best used for validation of certain childhood experiences with parents, such as parental warmth, love, and rejection. This measure also appears to provide additional information about childhood relationships with parents not revealed directly by ratings from the Adult Attachment Rating and Classification Systems, such as parental support or discouragement of autonomy. Although the MFP provides information about an individual's overall evaluation of experiences from childhood (e. g., higher order postulates reflected by the scores for acceptance total and independence total), the MFP does not appear to provide comparable information about psychological processes that may accompany recall of childhood experiences and that may be operative when an individual attempts to access attachment-related information, such as idealization.

Maternal separation anxiety

The mothers in this study are not unlike other first-time mothers with respect to the dimensions of maternal separation anxiety. The results pertaining to demographic characteristics, as well as those for stability, are consistent with and largely replicate findings from other studies (DeMeis, et 98 al., 1986; Hock, et al., 1988; Hunter, 1990; McBride & Belsky, 1988). For example, mothers who had fewer years of education were more likely to report higher levels of maternal separation anxiety, and mothers with lower family incomes were more likely to report greater concerns about the effects of separation on the child. These findings are consistent with those found by McBride and Belsky (1988). Mothers who were anxious about separation prenatally were likely to be anxious about the same 6 to 8 weeks after birth of their infant. Similarly, mothers who had concerns about separation effects prenatally were likely to be concerned about the same after the birth of their infant. The group as a whole, however, had greater concerns about employment-related separations at 6 to 8 weeks postpartum. An increase in concerns about employment- related separations, without concomitant changes in the levels of maternal separation anxiety and concerns about separation effects, has been found in other samples of first-time mothers during the initial phase of the transition to motherhood. More specifically, this pattern has been found in mothers between 2 days and 7 weeks after birth (DeMeis, et al., 1986; Hock, et al., 1988), and in mothers between 4 days and 4 months after the birth of their infants (Hunter, 1990). The lack of change in the levels maternal separation anxiety among the mothers in this sample is in keeping with Benedek's (1970) assertion that separation anxiety remains at an intensified level in women during the first

6 months following the birth of their infants. In the present study, the second administration of the Maternal Separation Anxiety Scale occurred near or at end of maternity leave for the majority of mothers in the sample. At 6 to 8 weeks postpartum, 36 (73.5%) mothers had just returned to work or 99 were planning to return to work within the next week, and 13 (26.5%) had decided not to return to work. The mothers may have been actively examining and/or confronting their feelings about leaving their infants to return to work at this point in time, or may have been re-affirming that the work-related decisions they made were the best or right ones for them. In either situation, these feelings may have been present or elicited again as they completed the questionnaire, which could account for some of the increase in employment-related separation concerns. With one notable exception, the dimensions of maternal separation anxiety were unrelated to childhood memories of relationships with parents.

Mothers' perceptions of the effects of separation on the child, as well as their concerns about employment-related separations, were not associated with childhood memories of parental acceptance, rejection, and encouragement of autonomy. Maternal separation anxiety, however, was associated with these childhood memories. More specifically, this dimension was related to the extent to which the subjects' mothers were rejecting, the degree to which their fathers were overprotective, and the extent to which both parents encouraged autonomy.

These findings are partially consistent with Hock, et al. (1992) who found a significant relation between maternal separation anxiety and childhood memories of parental overprotection. The findings are also somewhat consistent with Hunter (1990) who found that childhood memories did predict levels of maternal separation anxiety, but did not significantly predict concerns about separation effects and concerns about employment- related separations. The memories that predicted maternal separation anxiety 100

included recollections of maternal overprotection and paternal

encouragement of autonomy. Although the number of significant associations is not extensive, the results of the present study generally provide support for the theoretical underpinnings of maternal separation anxiety. The pattern of associations between the MFP scales and maternal separation anxiety suggests that childhood experiences with parents may contribute to or complement the expression of maternal separation anxiety. More specifically, negative or unhealthy experiences, such as rejection, overprotection, or discouragement of autonomy, may contribute in some way to higher levels of maternal separation anxiety 6 to 8 weeks after the birth of an infant.

No differences were found between mothers in the secure and insecure attachment groups with respect to two dimensions of maternal separation anxiety, perceptions of separation effects on the child and concerns about employment-related separations. However, differences did emerge for maternal separation anxiety. Mothers with insecure attachment representations, as compared to those with secure attachment representations, were more likely to have higher levels of maternal separation anxiety 6 to 8 weeks postpartum. The same pattern of associations was found for integration of attachment-related information (i. e., Coherency of Mind). Integration was significantly associated with levels of maternal separation anxiety, but was unrelated to the other two dimensions. Mothers with lower levels of maternal separation anxiety at 6 to 8 weeks postpartum evidenced greater integration of childhood experiences with parents and greater coherency when discussing attachment-related information than those with higher levels of separation anxiety. 101

These findings suggest that a mother's attachment representation and/or the way in which she integrates attachment-related experiences with parents during childhood may also contribute to or complement the expression of maternal separation anxiety once an infant is born. This is strikingly reminiscent of Benedek's assertions about the derivation of maternal separation anxiety. She describes maternal separation anxiety as an aspect of motherliness which functions as part of thehuman mental

apparatus that through evolution became able to modify and integrate early memories and interactive patterns (Benedek, 1970). However, these findings need to be interpreted cautiously. Other variables not included in this study,

such as characteristics of the marital relationship and maternal mental health, may also influence current representations of attachment relationships and may contribute to levels maternal separation anxiety reported by first-time mothers. Further, the nature of the relationship between attachment representations and maternal separation anxiety cannot be assumed to be causal, and the direction of the influence between these two variables remains unclear. The results are derived from correlational

analyses and do not reveal whether attachment representations influence the way in which mothers may experience maternal separation anxiety or whether maternal separation anxiety in some way contributes to mothers'

"current state of mind" with respect to attachment relationships. Despite these limitations, the results present a rather interesting picture about possible linkages among childhood experiences with parents, representations of attachment relationships, and maternal separation anxiety. Mothers who have healthier childhood experiences with parents are more likely to have secure attachment representations and lower levels of 102 maternal separation anxiety 6 to 8 weeks after the birth of their infants.

Mothers who have more unhealthy or more negative childhood experiences with parents are more likely to have insecure attachment representations and higher levels of separation anxiety at the same point in time. Moreover, mothers who recall both parents as supportive of autonomy during childhood are more likely have secure attachment representations andlower levels of maternal separation anxiety 6 to 8 weeks after the birth of their infants.

Mothers who recall both parents as discouraging of autonomy during childhood are more likely to have aninsecure attachment representation and higher levels of maternal separation anxietyat the same point in time. On the basis of these findings, it would seem that parental support and/or discouragement of autonomy may be the most relevant and/or most meaningful conceptual link between adult attachment representations and maternal separation anxiety. That is, autonomy or separation-related experiences with parents during childhood may have not only contributed to the development of working models of attachment relationships, but may have also helped to lay the foundation for the way mothers perceive and manage separation with their own children. In order to further explain this link, two potential pathways are briefly considered below. The discussion is intended to serve as a heuristic basis for understanding the way in which childhood separation-related experiences may have contributed to both working models and beliefs about separation; however, no theoretical and/or additional empirical literature is available at this time to support these ideas.

By definition, mothers with insecure attachment representations tend to have childhood experiences in which parental warmth and love are 103 lacking, and experiences in which their own parents either rejected their attachment or heightened attention toward parental needs and concerns (Kobak, et al., in press; Main, et al., 1985; Main, 1990). The way in which the mothers' parents dealt with or managed autonomy and separation issues is implicit in these experiences. For example, one way to heightened attention to parental needs or concerns is to make the child feel guilty for having interests and activities outside the home; i. e., discourage emotional and physical separation from the parent. One way to reject attachment is to insist that the child deal with emotional distress alone; i. e., force emotional distance or separation from the parent. In both situations, separation is a salient issue, such that the parents may have discouraged separation on a consistent basis or may forced autonomy unnecessarily and/or prematurely.

Mothers with secure attachment representations, on the other hand, tend to have childhood experiences in which their parents provided a secure base for them. If the parents did not provide a secure base, the mothers tend to be forgiving of parent wrong-doings and/or are reflective or coherent about their negative childhood experiences (Kobak, et al., in press; Main, et al., 1985; Main, 1990). One aspect of functioning as secure base is permitting a child to have interests and activities outside the home; i. e., supporting or encouraging physical separation from the parent. Another aspect is being sensitive and responsive to the child's needs and feelings. This can mean being accessible when the child physical and/or emotional closeness with the parent, and not being intrusive when the child desires some physical and/or emotional distance from the parent. In functioning as secure base for a child, separation and autonomy again become salient issues, 104 such that parents need to be able to be to balance their own and their child's needs for closeness and separation.

Mothers' perceptions of their infants

In this study, maternal perceptions of infants were found to be skewed toward positiveness. Most of the mothers in the sample tended to exclusively use positive words and phrases to depict the unique qualities of their infants, their perception of the relationship with their infant, and the infants feelings toward them. Relatively few mothers used neutral or negative expressions more than once or twice to describe these aspects of their perceptions. The tendency to use all positive descriptions is consistent with both Mercer (1986) and Barnard and Eyres (1986) who found that most mothers in their samples rated their infants more positively than the "average baby" at one month of age. This finding is also consistent with Bretherton, et al. (1989) who reported that the majority of mothers in their sample had very positive perceptions of their infants as a person during the first two months after birth, and with Zeanah, et al. (1985) who found that most parents (both mothers and fathers) gave overwhelmingly positive descriptions of infant personalities prenatally and at one month of age.

First-time mothers may be reluctant or unwilling to mention negative characteristics about their new infants because of societal expectations about early mother-child relationships. Generally, mothers are expected to cherish their new infants and to convey excitement and in their babies. That is, they are expected to view their newly born infants in a favorable manner. A mother's emphasis on the negative attributes of her infant may 105 be interpreted by some people as her dislike of the baby or aversion to the infant. Also, first-time mothers may be reluctant or unwilling to report negative information about their new infants because of the implications of attaching uncomplimentary or disparaging labels to their children at an early age. Centering one's attention on the negative attributes of a young infant is often interpreted by clinicians as an early sign of disturbance in a mother-child relationship.

Additionally, the tendency to focus on positive infant qualities and positive characteristics of the relationship may also reflect active identification with the maternal role, or may be one of several ways in which a mother integrates the relationship with her child into her personality. According to Benedek (1970), a mother establishes self-confidence in her motherliness through gratifying experiences in caring for her infant. She writes,"... observing her child whose smile is her reward, whose crying is her punishment, her self-concept expands through the gratifying introject: good, thriving infant = good mother, good self' (p. 117). In other words, perception of a "good, thriving" infant engenders feelings of self-confidence and fosters a positive perception of one's self as a mother. The tendency to emphasize positive qualities of the infant or positive characteristics of the relationship could reflect a new mother's efforts to derive in her new role as a parent and/or may be one way in which she can validate that she is becoming or is a competent, capable parent. No significant differences were found between prenatal expectations and postpartum perceptions of infants among the mothers in this sample.

These findings are consistent with Zeanah, et al. (1985& 1986) and suggest that maternal perceptions may remain stable over a major life transition, 106 such as the birth of an infant. However, 6 to 8 weeks after birth may not have been enough time for more dramatic changes in perception to occur.

Further, the experience of childbirth, in combination with the physical, social, and psychological adjustments that mothers encounter during the postpartum period, may not be sufficient by themselves to generate substantial change in perceptions. In other words, these experiences may not be sufficient to induce re-examination and/or reorganization of the mothers' perspective about selected aspects of the mother-infant relationship. No differences were found between the two attachment groups with respect to the mothers' perceptions. Mothers in the secure group were as likely as those in the insecure groups to use all positive and/or mixed phrases to describe the unique qualities of their infants, their perception of the relationship with their infant, and the infants feelings toward them. These findings are somewhat surprising because working models of attachment relationships are presumed to guide one's perceptions according to attachment theory. The findings suggest that attachment representations may not guide or influence maternal perceptions in a direct, clearly- identifiable manner. Rather, mothers' attachment representations may have an indirect, or perhaps a curvilinear, relation to their perceptions of their infants at 6 to 8 weeks postpartum. Again, "third" variables, or variables excluded from this study, could be exerting an influence on the mothers' current attachment representations and/or their perceptions of their infants. Interestingly, significant associations were found between mothers' perceptions of infants and only one dimension of maternal separation anxiety. Mothers with higher levels of maternal separation anxiety (Subscale 1) were more likely to use mixed descriptions than those with lower levels of separation anxiety. Although these findings suggest that a mother's perceptions of her infant may contribute to her capacity to function as a secure base (e. g., maternal separation anxiety), the nature of this relationship remains unclear. These correlational results do not reveal whether heightened levels of maternal separation anxiety may have influenced the mothers' use of neutral and/or negative descriptions, or whether the presence of mixed perceptions may have contributed in some way to higher levels of maternal separation anxiety at 6 to 8 weeks postpartum. CHAPTER VII CONCLUSION AND IMPLICATIONS FOR FUTURE RESEARCH

This final chapter begins with a summary of the principal findings from the study and includes a brief review of the study's major limitations.

The chapter ends with a discussion of implications for future research.

The findings from this study enhance knowledge about the measurement of attachment representations in adults, and advance the understanding of maternal separation anxiety in mothers of firstborn infants. Three measures were selected and evaluated in order to determine the extent to which they provided equivalent information about attachment representations in mothers expecting their first child. Additionally, associations between these attachment representations, dimensions of maternal separation anxiety, and aspects of mothers' perceptions of their infants were examined at two salient points during the transition to motherhood, the third trimester of pregnancy and 6 to 8 weeks following the birth of the infants.

The study first examined the extent to which two interview-based methods yielded identical information about adult attachment status, and the degree to which a self-report measure provided information comparable to these classification methods. The Adult Attachment Rating and Classification

108 109 Systems (ARCS) and the Adult Attachment Interview Q_-set (AAQ) were found to be strongly concordant with respect to differentiating secure and insecure attachment representations in mothers expecting their first child. The

Mother-Father-Peer (MFP) Scale was found to provide information best used for validating specific experiences with parents, in particular the extent of warmth, rejection, and support of autonomy that the mothers may have experienced during childhood. However, the MFP did not provide comparable information about psychological processes that may influence access to attachment-related information in mothers.

Secondly, the study explored relationships between adult attachment representations, dimensions of maternal separation anxiety, and mothers' perceptions of their infants. Mothers who had healthy experiences with parents during childhood were likely to have secure attachment representations and lower levels of maternal separation anxiety following the birth of their infants. Mothers who had more unhealthy or more negative experiences with parents during childhood were likely to have insecure attachment representations and higher levels of maternal separation anxiety 6 to 8 weeks after the birth of their infants. The extent to which the subjects' own parents encouraged or supported autonomy emerged as an important aspect of childhood experiences that may contribute to the formation of attachment representations and complement the expression of maternal separation anxiety in mothers of infants. Attachment representations were unrelated, whereas maternal separation anxiety was related, to mothers' perceptions of their 6 to 8 weeks old infants. Mothers with lower levels of maternal separation anxiety, as compared to those with 110 higher levels, were more positive overall in their perceptions of their infants.

The results from this study are limited primarily to Caucasian, well- educated women expecting their first child. The findings may not be applicable to mothers from more diverse sociocultural and racial/ethnic backgrounds, or women with more than one child. Although significant correlations emerged between study variables, directional or causal relationships cannot be inferred from the findings. For example, ihe results do not reveal if attachment representations actually produced the differences in mother's levels of maternal separation anxiety, and if levels of maternal separation anxiety directly contributed to differences in the mother's perceptions of their infants. Further, effects from other variables not included in the study, such as characteristics of the marital relationship and maternal mental health, may have influenced the relations between the study variables.

However, the findings have provided a rich foundation from which other studies can be developed. Future investigations might replicate the results for the three measures of adult attachment representations that were of interest in this study, and extend these findings by comparing other kinds of representational measures designed for use in adults, such as attachment style. The findings from this study highlight the importance of using multiple methods to determine attachment representations in adults. Continued use of both interview-based and self-report measures, in particular, would help to identify areas of overlap more precisely and reveal differences among each measure more explicitly. Use of both kinds of measures would also help in more clearly defining the network of Ill relationships among each component that is presumed to comprise an adult's representation of attachment relationships, such as memories of specific aspects of childhood experiences, overall evaluation of childhood relationships with parents, and integration or coherency of attachment- related information. Findings from this study could be extended to a larger sample of

mothers, with special emphasis placed on the inclusion of women with more diverse racial/ethnic and sociocultural backgrounds. Future studies could also explore the relationship between different aspects of adult attachment

representations and other dimensions of maternal personality and/or functioning. For example, investigations might examine associations between aspects of an adult's working model of the self (such as the ability to utilize support from others), other aspects of representations of attachment relationships (such as the relationship between spouses or close friends), maternal separation anxiety, and indices of mental health in mothers.

In the present study, maternal separation anxiety was significantly related to both attachment representations, which were assessed during pregnancy, and mothers' perceptions of their infants, which were assessed when the infants were 6 to 8 weeks old. However, attachment representations were unrelated to the mothers' perceptions. These findings give rise to the possibility that maternal separation anxiety may be a mediating influence between attachment representations and aspects of maternal functioning. Future studies could examine this relation more directly and continue to explore associations among attachment representations, maternal separation anxiety, and aspects of maternal functioning at other points during the infant's first year of life. They could also be examined across other life 112 transitions, such as the birth of a second child and the first child's entry into school. These types of studies may be particularly beneficial in identifying conditions and/or situations that may precipitate change and reorganization of attachment representations, and perhaps could provide a more direct test of Bowlby's supposition that internal working models are resistant to change. In the present study, the Mothers' Perception of Her Baby Interview was used to assess the degree of positiveness of the mothers' perceptions of their 6 to 8 week old infants. This measure could be revised and a scoring procedure could be developed that is similar to the Adult Attachment Rating and Classification Systems or the Adult Attachment Interview Qrset. In this way, this measure could serve as an index of parental attachment representations. Future investigations might use this measure, or other measures of parental attachment representations, in studies designed to more fully understand the relationship between adult attachment representations, parental attachment representations, and parental functioning. Inclusion of maternal separation anxiety in these studies would also assist in determining the extent to which this construct may serve to mediate between attachment representations and aspects of parental functioning. These types of studies could be particularly helpful in further identifying and distinguishing factors that may influence or contribute to healthy psychological functioning in mothers. APPENDIX A

INSTRUMENTS

113 114

THE ADULT ATTACHMENT INTERVIEW

1. Could you start by telling me a bit about the place and family in which you were raised? We'd like to know where you were born and brought up and with whom. What did your parents do for a living? Did you have brothers and sisters?

2. Would you describe your relationship with your parents as a child? We'd like to know from as far back as you can remember. To begin with, I'd like you to choose five adjectives that reflect your childhood relationship with your mother, and then five adjectives that reflect your relationship with your father.

Could you tell me why you chose those adjectives for each parent?

3. To which parent did you feel the closest, and why? Why isn't there this feeling with the other parent?

4. When you were upset as a child, what would you do? Probes: When you were emotionally upset as child, what would you do? When you were hurt a bit physically? When you just needed comfort, support? Can you illustrate with a specific incident?

5. How do you think these experiences with your parents have affected your adult personality? 115

6. Are there any aspects to your early experiences that you feel were a set­ back in your development?

7. Why do you think your parents behaved as they did?

8. Have there been many changes in your relationship with your parents since childhood? I mean from childhood through the present.

9. What is your relationship with your parents like for you now as an adult?

10. Were there any other adults with whom you were close, like parents, as a child? Or any other adults who were especially important to you, even though not parental?

11. Did you know all of your grandparents?

12. What is the first time you remember being separated from your parents? Probes: How did you or they respond? Are there any separations that stand out in your mind? What about the first days of school? We'd like to know more about your responses to these separations.

13. Were your parents ever threatening with you in any way, perhaps for discipline, perhaps jokingly? Some parents have told us, for example, that their parents would threaten to leave them or send them away from home. Probes: When did this happen? Was it frequent? How did you cope with it? Why did you think your parents did this? Do you feel this experience affects you now as an adult? Does it influence your approach to your own child? 116

14. Did you ever feel rejected as a young child? Probes: Could you describe this incident? How did you feel when this happened? Why do you think your parent did those things? Do you think he/she realized he/she was rejecting you? Does this experience affect you now as an adult? Does it influence your approach to your own child?

15. Did you experience the loss of a parent or other close loved one while you were a young child? Probes: Who? At what age? How did this affect you at the time? Do you remember how you felt, how you coped with this loss? Does this loss affect you now as an adult? Do you think it influences your approach to your own child? Have you lost other close persons in adult years? (Same probes)

16. How do you respond now, in terms of feelings, when you separate from your child?

17. Is there any particular thing which you feel you learned above all from your own childhood experiences?

18. What would you your child might have learned from his/her experience of being parented? 117

Questions and probes modified for women expecting their first child:

Original probes (13, 14, 15): Does it influence your approach to your own child? Do you think it influences your approach to your own child?

Modified probes: Do you think it may influence your approach to your own child? In what ways?

Original question (16): How do you respond now, in terms of feelings, when you separate from your child?

Modified question (16): Once your child is born, how might you respond, in terms of feelings, when you separate from your child?

Original question (18): What would you hope your child might have learned from his/her experience of being parented?

Modified question (18): What do you hope your child may learn from his/her experience of being parented? 118

THE MOTHER-FATHER-PEER SCALE

Indicate the extent to which the following statements describe your childhood relationship with the people indicated by using the following scale:

1 2 3 4 5

Strongly Somewhat Uncertain Somewhat Strongly Disagree with Disagree with About Agree with Agree with Statement______Statement______Statement_____ Statement Statement

WHEN I WAS A CHILD, MY MOTHER (or mother substitute):

1. encouraged me to make my own decisions. 2 3 4 5

2. helped me learn to be independent. 2 3 4 5

3. felt she had to fight my battles for me when I had a disagreement with a teacher or friend. 2 3 4 5 4. was close to a perfect parent. 2 3 4 5

5. was overprotective of me. 2 3 4 5

6. encouraged me to do things for myself. 2 3 4 5 7. encouraged me to try things my way. 2 3 4 5

8. had not a single fault that I can think of. 2 3 4 5 9. did not let me do things that other kids my age were allowed to do. 10. sometimes disapproved of specific things I did, but never gave me the impression that she disliked me as a person. 2 3 4 5 11. enjoyed being with me. 2 3 4 5

12. was an ideal person in every way. 2 3 4 5

13. was someone 1 found very difficult to please. 119

WHEN I WAS A CHILD, MY MOTHER (or mother substitute): 14. usually supported me when I wanted to do new and exciting things. 2 15. worried too much that I would hurt myself or get sick. 2 3 4 5

16. was never angry with me. 2 3 4 5

17. was often rude to me. 2 3 4 5

18. rarely did things with me. 2 3 4 5

19. didn’t like to have me around the house. 2 3 4 5

20. and I never disagreed. 2 3 4 5 21. would often do things for me that I could do myself. 2 3 4 5

22. let me handle my own money. 2 3 4 5 23. could always be depended upon when I really needed her help and trust. 24. gave me the best upbringing anyone could ever have. 2 3 4 5

25. did not want me to grow up. 2 3 4 5 26. tried to make me feel better when I was unhappy. 27. encouraged me to express my own opinion. 2 3 4 5 28. never disappointed me. 2 3 4 5

29. made me feel that I was a burden to her. 2 3 4 5

30. gave me the feeling that she liked me as I was; she didn't feel she had to make me over into someone else. WHEN I WAS CHILD, MY FATHER (or father substi 31. encouraged me to make my own decisions. 1 32. helped me learn to be independent. 1 33. felt he had to fight my battles for me when I had a disagreement with a teacher or friend. 1 34. was close to a perfect parent. 1 35. was overprotective of me. 1 36. encouraged me to do things for myself. 1 37. encouraged me to try things my way. 1 38. had not a single fault that I can think of. 1 39. did not let me do things that other kids my age were allowed to do. 1 40. sometimes disapproved of specific things I did, but never gave me the impression that he disliked me as a person. 1 41. enjoyed being with me. 1

42. was an ideal person in every way. 1 43. was someone I found very difficult to please. 1 44. usually supported me when I wanted to do new and exciting things. 1 45. worried too much that I would hurt myself or get sick. 1 46. was never angry with me. 1 47. was often rude to me. 1 48. rarely did things with me. 1

49. didn't like to have me around the house. 1 50. and I never disagreed. 1 121

WHEN I WAS A CHILD, MY FATHER (or father substitute): 51. would often do things for me that I could do myself. 2 3 4 5

52. let me handle my own money. 2 3 4 5

53. could always be depended upon when I really needed his help and trust. 54. gave me the best upbringing anyone could ever have. 2 3 4 5 55. did not want me to grow up. 2 3 4 5 56. tried to make me feel better when I was unhappy. 57. encouraged me to express my own opinion. 2 3 4 5 58. never disappointed me. 2 3 4 5 59. made me feel that I was a burden to him. 2 3 4 5

60. gave me the feeling that he liked me as I was; he didn't feel he had to make me over into someone else.

WHEN I WAS A CHILD, OTHER CHILDREN:

61. liked to play with me. 2 3 4 5 62. were always criticizing me. 2 3 4 5

63. often shared things with me. 2 3 4 5

64. often picked on me and teased me. 2 3 4 5 65. were usually friendly with me. 2 3 4 5

66. would usually stick up for me. 2 3 4 5

67. liked to ask me to go along with them. 2 3 4 5 WHEN I WAS A CHILD, OTHER CHILDREN: 68. wouldn't listen when I tried to say something. 69. were often unfair to me. 70. would often try to hurt my feelings. 123

MATERNAL SEPARATION ANXIETY SCALE PARENT QUESTIONNAIRE

The following statements represent matters of interest and concern to parents. Not all people feel the same way about them. Answer the statements as you are feeling now or think you will feel as your child grows older. Read each statement carefully and circle the number at the right which most closely reflects YOUR degree of agreement or disagreement. Answer all the items without discussing any of them with anyone.

1 2 3 4 5

Strongly Disagree Somewhat Agree Strongly disagree______agree______agree

1. I miss holding or cuddling my child when I am 1 2 3 4 5 away from him/her.

2. My child is happier with me than with babysitters 1 2 3 4 5 or teachers.

3. Children will be afraid in a new place without 1 2 3 4 5 their mother. 4. My life wouldn't be complete without a career. 1 2 3 4 5 5. If a child is independent and outgoing, he/she 1 2 3 4 5 will make friends easily without his/her mother's help.

6. When away from my child, I often if his/ 1 2 3 4 5 her physical needs (dry diapers, enough to eat, etc.) are being met. 7. Holding and cuddling my child makes me feel so 1 2 3 4 5 good that I really miss the physical closeness when I'm away. 8. I am more concerned with my child's physical 1 2 3 4 5 safety than a babysitter or teacher. 124

1

Strongly Disagree Somewhat Agree Strongly disagree agree___ agree

9. It will be difficult for my child to adjust to someone 1 2 3 4 5 else taking care of him/her.

10. I would resent my job if it meant I had to be away 2 3 4 5 from my child.

11. My child will benefit from group experiences 2 3 4 5 (i.e., nursery school, day care, kindergarten) since they will provide him/her social experiences that he/she could not get at home.

12. When I am away from my child, I feel lonely and 2 3 4 5 miss him/her a great deal.

13. Only a m other just naturally knows how to comfort 2 3 4 5 a distressed child.

14. A child is likely to get upset when he/she is left 2 3 4 5 with a babysitter.

15. I have a systematic plan for how I'm going to build 2 3 4 5 my career in the world of work.

16. It is good for my child to spend time away from me 2 3 4 5 so that he/she can learn to deal independently with unfamiliar people and new situations.

17.1 like to have my child close to me most of the time. 2 3 4 5

18. I am naturally better at keeping my child safe than 2 3 4 5 any other person.

19. I believe that my child misses me when I have to let 2 3 4 5 someone else take care of him/her.

20. A career or job brings me a lot of personal 2 3 4 5 satisfaction.

21. Even though my child fusses a bit when I leave, 2 3 4 5 I know he/she will be OK in a few minutes—after I'm out of sight.

22. I don't like to leave my child. 3 4 5 125

1

Strongly Disagree Somewhat Agree Strongly disagree agree agree

23. My child prefers to be with me more than with 2 3 4 5 anyone else.

24. My child is afraid and sad when he/she is not 2 3 4 5 with me.

25. I would not postponing my career in order 2 3 4 5 to stay home with my child.

26. My child needs to spend time away from me in 2 3 4 5 order to a develop a sense of being an individual in his/her own right.

27. When I am separated from my child, I wonder 2 3 4 5 whether my child is crying and missing me.

28. I don't enjoy myself when I’m away from my child. 2 3 4 5 29. I worry that my child is never completely 2 3 4 5 comfortable in an unfamiliar setting if I am not with him/her.

30. Children are very demanding and I often wish I 2 3 4 5 had more time for a career.

31. Exposure to many different people is good for my 2 3 4 5 child.

32. I worry when someone else cares for my child. 2 3 4 5 33. If I could choose between working full-time or 2 3 4 5 staying home with my child, I would want to stay home.

34. There are times in the lives of young children 2 3 4 5 when they need to be with people other than their mothers.

35. When away from my child, I worry about whether 2 3 4 5 or not the babysitter is able to soothe and comfort my child if he/she is lonely or upset. 126

MOTHER'S PERCEPTIONS OF HER BABY INTERVIEW

1. What words best describe the unique qualities of your baby? Probes: Could you tell me why you chose those words? What words also best describe your baby as a person? (use if mother only describes physical qualities or characteristics of her baby)

2. What words best describe your relationship with your baby?

Probes: Could you tell me why you chose those words? In what ways do your baby's unique qualities influence your relationship with him/her?

3. What words best describe the way your baby feels about you? Probe: Could you tell me why you chose those words? APPENDIX B

Descriptions of Adult Attachment Classifications and Selected Scales from the Adult Attachment Rating and Classification Systems

127 128

Descriptions of Adult Attachment Classifications!

Adults classified as Secure/autonomous tend to value attachment relationships with their own parents, others, or in an abstract sense. These individuals describe a variety of childhood experiences as well as objectively describe and acknowledge both positive and negative experiences and emotions. They consider attachment relationships and attachment-related experiences as influential in the development of their adult personality and their present functioning. These adults present a clear, coherent view of attachment relationships and attachment-related experiences. Adults classified as Dismissing tend not to value attachment relationships. These relationships are considered to be of little concern, value or influence in the development of their adult personality. These individuals have difficulty in recalling specific attachment experiences, tend to deny negative experiences and emotions, and tend to discount their importance for present functioning. Adults classified as Preoccupied tend to be enmeshed in their attitude toward attachment. They have difficulty integrating attachment experiences into an overall coherent model and express continuing efforts to please their parents.

1 Descriptions of the Adult Attachment Classifications were compiled from the following sources: Benoit, et al. (1989); Crowell & Feldman (1989 &1991); Fonagy, et al. (1991); Kobak&Sceery (1988); Main (1990); Main, et al. (1985); Main & Goldwyn (in press). 129

Descriptions of Scales for the

Adult Attachment Rating and Classification Systems^

Probable Childhood Experience Scales There are 5 scales designed to estimate the subject's probable experiences with parents and other attachment figures during childhood.

Each parent and/or attachment figure is rated separately on a 9-point scale with 1 representing low scores, 5 reflecting moderate scores, and 9 indicating high scores. 1. The Loving Scale is used to determine the extent to which the

subject appears to have experienced the parent or attachment

figure as loving or unloving during childhood. 2. The Rejecting Scale is used to ascertain the extent to which the

parent or attachment figure rejected or avoided the child’s attachment. 3. The Involving/Role Reversing Scale is used to judge the extent to

which a parent or attachment figure makes him/herself an object

of the child's attention. 4. The Pressured to Achieve Scale is used to evaluate the extent to

which the child was pressured to achieve a particular status,

position or simply pushed to exceed.

^Descriptions of selected scales from the Adult Attachment Rating and Classification Systems were compiled from the following sources: Benoit, et al. (1989); Crowell & Feldman (1989 &1991); Fonagy, et al. (1991); Kobak & Sceery (1988); Main (1990); Main, et al. (1985); Main & Goldwyn (in press). 130

5. The Neglecting Scale is used to determine the extent to which a

parent or attachment figure, when present in the home, is

inattentive, uninvolved, or psychologically inaccessible to the child.

States of Mind Scales

These scales generally refer to mental representations of attachment- related experiences and the organization of information relevant to attachment. There are two categories of scales, States of mind with respect to parents and Overall states of mind. All but 2 of the scales, Unresolved Loss and Unresolved Trauma, are designed to assess organized states of mind. Ratings are assigned using a 9-point scale with 1 indicating low scores, 5 reflecting moderate scores, and 9 representing high scores. Each parent is rated separately using the States of mind with respect to parents scales, whereas one rating is assigned using the Overall states of mind scales. 1. States of Mind Scales with respect to parents

a. The Idealizing Scale is used to evaluate the extent of discrepancy

between the subject's overall view of the parent (semantic or abstract level) and actual experiences that are reported (episodic or specific level). b. The Involving Anger Scale is used to ascertain the extent to which

the subject expresses current involvement in feelings of anger with a parent and/or attachment figure. Overall States of Mind Scales a. The Overall Derogation of Attachment Scale is used to determine the

extent to which the subject expresses attachment relationships as

beneath their consideration. b. The Insistence of Lack of Recall Scale is used to evaluate the extent

to which the subject insists on an inability to recall his or her childhood. c. The Passivity of Thought Processes Scale is used to ascertain the

extent to which the subject is unable to complete ideas or follow

through in thought. d. The Megarognitive Processes Scale is used to determine the extent of

the subject's ability to monitor and report on his/her processes of

thinking. e. The Unresolved loss Scale is used to evaluate the extent to which the

subject's thinking or discourse about the death of an attachment

figure is disorganized or disoriented. f. The Unresolved Trauma Scale is used to ascertain the extent to

which the subject's thinking or discourse about frightening and/or abusive experiences with an attachment figure is disorganized or

disoriented. h. The Coherency of Mind Scale is used to evaluate the extent to which

the subject's thinking or discourse with respect to attachment is

internally consistent. This scale is considered as the "state of mind" with respect to adult attachment. APPENDIX C

LIST OF ABBREVIATIONS USED FOR INSTRUMENTS

132 133

Abbreviation Title qf Instrument Developer! si AAI Adult Attachment Interview George, Kaplan, & Main (in press) AAQ. Adult Attachment Interview Kobak (1988) Qrset

ARCS Adult Attachment Rating and Main & Goldwyn Classification Systems (in press)

MFP Mother-Father-Peer Scale Epstein (1983)

MPBI Mothers' Perceptions of Her Baby Interview Lutz (1990) MSA/MSAS Maternal Separation Anxiety Hock, McBride, & Scale Gnezda(1982) APPENDIX D

ADDITIONAL DATA TABLES

134 135

Table 13

Associations between Dimensions of Maternal Separation Anxiety and Probable Childhood Experience Scales of the Adult Attachment Rating and Classification Systems

Time 1 Time 2 MSA Subscale 1 2 3 1 2 3

Childhood ExDerience Scales

Loving mother -.110 -.066 .035 -.250* -.123 .031 Loving father .014 .174 .158 -.127 .180 .216

Rejecting mother .159 .139 -.034 .277* .224 -.042

Rejecting father -.068 -.183 -.220 .125 -.163 -.261* Involving mother .094 -.206 -.116 .011 -.076 -.162 Involving father .168 -.132 .094 .157 -.146 -.002 Pressured to achieve (mother) .072 -.101 .053 .010 -.067 .070 Pressured to achieve (father) -.070 .016 .159 .157 .092 -.020

Neglecting mother .318* .249* .198 .237 .192 .162 ■

Neglecting father .086 -.317* -.147 .144 k> -.245*

*p< .05 **p< .01 ***p< .001 t=trend 136

Table 14 Associations between Dimensions of Maternal Separation Anxiety and States of Mind Scales of the Adult Attachment Rating and Classification Systems

Time 1 Time 2 MSA Subscale 1 2 3 1 2 3

States of Mind Scales

Idealizing mother .321* .158 .115 .267* .051 .049

Idealizing father .329* .181 .268* .313* .155 .212

Involving anger (mother) -.057 -.066 -.193 .039 .044 -.258*

Involving anger (father) -.083-.066 -.135 .057 .048 -.196

Derogation of attachment .138 .042 -.100 .222 .075 .084 Insistence on lack of recall .104 .104 -.012 .270t .060 -.010

Metacognitive processes -.064 -.49* -.057 .002 -.296 -.021 Passivity of thought processes -.091 -.191 -.074 .098 -.008 -.182 Unresolved loss -.018 -.025 .087 .031 .121 .055

Unresolved trauma .005 .002 -.215 .241* .029 -.238* Coherency of mind -.056 -.023 -.048 -.311* -.154 .079

*p< .05 **p< .01 ***p< .001 t=trend BIBLIOGRAPHY

Aber, J. L (1987). Attachment in toddlerhood. Paper presented at The Ohio State University, Department of Family Relations and Human Development, Columbus, OH. Aber, J. L & Slade, A. (1987). Parental detachment as a defense against separation anxietv: Behavior and representation. Paper presented at The Society for Research in , Baltimore, MD. Ainsworth, M. D. S. (1982). Attachment: Retrospect and prospect. In C. M. Parkes & J. Stevenson-Hinde, The nlace of attachment in human behavior (pp. 3-30). New York: Basic.

Ainsworth, M. D. S. (1985). Attachment across the life span. Bulletin.oflthe New York Academy of Medicine. £1 (9) 792-812.

Ainsworth, M. D. S. (1989). Attachments beyond infancy. American Psychologist. 44 (4), 709-716.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ.: Erlbaum. Ainsworth, M. D. S. & Wittig, B. A. (1969). Attachment and exploratory behavior of one-year olds in a strange situation. In B. M. Foss (Ed.), Determinants of infant behavior. Vol. 4, London: Methuen.

Barnard, K. E.& Eyres, S. J. (1979). Child health assessment. Part 2: The first year of life. (DHEW Publication No. HRA 79-25). Washington, DC: U. S. Government Printing Office. Belsky, J., Rovine, M., & Taylor, D. G. (1984). The Pennsylvania infant and family development project, 3: The origins of individual differences in infant-mother attachment: Maternal and infant contributions. Child Development. 5 5 , 718-728.

Belsky, J. & Isabella, R. (1988). Maternal, infant, and social-contextual determinants of attachment security. In J. Belsky & T. Nezworski (Eds.), Clinical implications of attachment (pp. 41-94). Hillsdale, NJ: Erlbaum.

137 138

Benedek, T. (1959). Parenthood as a developmental phase, loumal of the American Psychoanalytic Association.Z, 389-417.

Benedek, T. (1970). Motherhood and nurturing. In E. J. Anthony& T. Benedek (Eds,), Parenthood; Its psychology and psychopathology (pp. 153-166). Boston: Little, Brown. Benedek, T. (1970). The family as a psychologic field. In E. J. Anthony & T. Benedek (Eds.), Parenthood; Its psychology and psychopathology (pp. 109-136). Boston: Little, Brown. Benoit, D., Zeanah, C. H.,& Barton, M. L (1989). Maternal attachment disturbances in failure to thrive. Infant Mental Health loumal. JLU (3), 185-202.

Bibring, G. L (1961). A study of the psychological process in pregnancy and of the earliest mother-child relationship. The psvcholanalvtic study of the child. Vol. 16. New York: International Universities.

Biringen, Z. (1990). Direct observation of maternal sensitivity and dyadic interactions in the home: Relations to maternal thinking. Developmental Psychology. 26 (2), 278-284.

Biringen, Z. & Bretherton, I. (1989). Parent attachment interview: Sensitivitv/insight scale. Unpublished manuscript, University of Wisconsin, Madison. Blatt, S. J., Wein, S. J., Chevron, E., & Quinlan, D. M. (1979). Parental representations and depression in normal young adults, loumal of Abnormal Psychology.M (4), 388-397.

Bloom-Feshbach, J. & Bloom-Feshbach, S. (1987). Introduction: Psychological separateness and experiences of loss. In J. Bloom-Feshbach & S. Bloom- Feshbach (Eds.), The psychology of separation and loss: Perspectives on development, life transition, and clinical practice (pp. 1-59). San Francisco: Jossey-Bass. Bowlby, J. (1969). Attachment and loss: Vol. I. Attachment. New York: Basic Books. Bowlby, J. (1973). Attachment and loss: Vol. II. Separation, anxiety and anger. New York: Basic Books.

Bowlby, J. (1979). The making and breaking of affectional bonds. London: Travistock. Bowlby, J. (1980). Attachment and loss; Vol. III.Xoss. sadness and depression. New York: Basic Books. 139

Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New York: Basic Books.

Brazelton, T. J. & Cramer, B. G. (1990). The earliest relationship: Parents. infants, and the drama of earlv attachment. New York: Addison- Wesley. Bretherton, I. (1985). Attachment theory: Retrospect and prospect. Ini. Bretherton & E. Waters, (Eds.), Growing points of attachment theory and research (pp. 3-35). Monographs of the Society for Research in Child Development. Serial no. 209, Vol. 50, Nos. 1-2.

Bretherton, I. (1987). New perspectives on attachment relations: Security, communication, and internal working models. In J. D. Osofsky (Ed.), Handbook of infant development. 2nd ed. (pp. 1061-1100). New York: Wiley.

Bretherton, I. (1990). Communication patterns, internal working models, and the intergenerational transmission of attachment relationships. Infant Mental Health loumal. H (3), 237-252.

Bretherton, I., Biringen, Z., Ridgeway, D., Maslin, C., Sherman, M. (1989). Attachment: The parental perspective. Infant Mental Health loumal. IQ (3), 203-221.

Broussard, E. R. (1984). Maternal empathy: Its relation to emerging self­ representations and empathy in infants. In J. Lichtenberg, M. Bomstein, D. Silver (Eds.), Empathy. Vol. 2 (pp. 81-103). Hillsdale, NJ: Erlbaum. Broussard, E. R. (1986). Prospective longitudinal study of firstborn neonates. In L Erlenmeyer-Kimling& N. E. Miller (Eds.), life span research on the prediction of psychopathology (pp. 49-64). Hillsdale, NJ: Erlbaum.

Broussard, E. R. (1970). Psychosocial disorders in children: Early assessment of infants at risk. Continuing Education for the Family Physician. 8 (2), 44-57.

Broussard, E. R. & Hartner, M. S. S. (1971). Further considerations regarding maternal perception of the first-born. In J. Hellmuth (Ed.), Exceptional infant. Vol. 2: Studies in Abnormalities (pp. 432-449). New York: Brunner/Mazel. Campos, J. J., Barrett, K. C., Lamb, M., Goldsmith, H. H., & Stenberg, C. (1983). Socioemotional development. In M. M. Haith & J. J. Campos (Vol. eds.), Infancy and developmental psychobiology. Vol. II (pp. 783-915), P. H. Mussen (Series ed.), Handbook of child psychology. New York: Wiley.

Chodrow, N. (1978). The reproduction of mothering: Psychoanalysis and the sociology of gender. Los Angeles: University of California. 140

Cohen, J. (1968). Weighted kappa: Nominal scale agreement with provision for scaled disagreement or partial credit. Psychological Bulletin. 70 (4), 213-220. Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. 2nd Ed. Hillsdale, NJ: Erlbaum.

Cohler, B. J. & Scott, F. M. (1987). Separation, interdependence, and social relations across the second half of life. In J. Bloom-Feshbach & S. Bloom-Feshbach (Eds.), The psychology of separation and loss: Perspectives on development, life transition, and clinical practice (pp. 165-204). San Francisco: Jossey-Bass.

Cohler, B. J. & Grunebaum, H. U. (1981). Mothers, grandmothers, and daughters: Personality and child care in three-generation families. New York: Wiley. Cole, H. E. & Kobak, R. R. (1990). Attachment and psychopathology in late adolescence: A study of eating disorder and depression. Paper presented at the International Conference on Personal Relationships, Oxford, England. Collins, N. L. & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. loumal of Personality and Social Psychology. (4), 644-663.

Crittenden, P. M. (1988). Relationships at risk. In J. Belsky & T. Nezworski (Eds.), Clinical implications of attachment (pp. 136-174), Hillsdale, NJ: Erlbaum.

Crittenden, P. M. (1989). Internal representational models of attachment relationships. Paper presented at the meeting of the Society of Research in Child Development, Kansas City, MO. Crockenberg, S. B. (1981). Infant , mother responsiveness, and social support influences on the security of mother-infant attachment. Child Development.5 2 , 857-865.

Crowell, J. A. & Feldman, S. S. (1988). Mothers'internal models of relationships and children's behavioral and developmental status: A study of mother-child interaction. Child Development. 59. 1273-1258.

Crowell, J. A. & Feldman, S. S. (1989). Assessment of mothers' working models of relationships: Some clinical implications. Infant Mental Health loumal. in (31. 173-184.

Crowell, J. A. & Feldman, S. S. (1991). Mothers' working models of attachment relationships and mother and child behavior during separation and reunion. Developmental Psychology. 2Z (4), 597-605. 141

DeMeis, D. K., McBride, S. L, & Hock, E. (1986). The balance of employment and motherhood: Longitudinal study of mothers’ feelings about separation from their first-born infants. Developmental Psychology. 2 2 (5), 627-632.

Egeland, B. & Farber, E. A. (1984). Infant-mother attachment: Factors related to its development and changes over time. Child Development. 54. 753-771.

Epstein, S. (1983). The mother-father-peer scale. Unpublished manuscript. University of Massachusetts, Amhurst. Feeney, J. A. & Noller, P. (1990). Attachment style as a predictor of adult romantic relationships. loumal of Personal and Social Psychology. M (2), 281-291. Field, T. M., Dempsy, J. R., Hallock, N. H. & Shuman, H. H. (1978). The mother's assessment of the behavior of her infant. Infant Behavior and Development. 1, 156-167.

Fonagy, P., Steele, H. & Steele, M. (1991). Maternal representations of attachment during pregnancy predict the organization of infant- mother attachment at one year of age. Child Development. 62f 891-905.

George, C. (1990). Personal communication, August 27, 1990.

George, C. (1990). Personal communication, September 19, 1990. George, C. (1990). Personal communication, October 4, 1990. George, C., Kaplan, N., & Main, M. (in press). The adult attachment interview. In M. Main (Ed.), Assessing attachment through discourse, drawings and reunion situations (working title). New York: Cambridge University Press. George, C. & Solomon, J. (1989). Internal working models of caregiving and security of attachment at age six. Infant Mental HealthJournal. JH (3), 222-237.

Grossman, F. K., Eichler, L. S., & Winickoff, S. A. (1980). Pregnancy, birth, and parenthood. San Francisco: Jossey-Bass.

Grossmann, K., Grossman, K. E., Spangler, G., Suess, G.,& Unzner, L. (1985). Maternal sensitivity and newborns’ orientation responses as related to quality of attachment in Northern Germany. In I. Bretherton& E. Waters, (Eds.), Growing points of attachment theory and research (pp. 233-256). Monographs of the Society for Research in Child Development. Serial no. 209, Vol. 50, Nos. 1-2. 142

Grossman, K., Fremmer-Bombik, E, Rudolph, J. & Grossman, K. E (1988). Maternal attachment representations as related to infant-mother attachment and maternal care during the first year. In R. A. Hinde & J. Stevenson-Hinde (Eds.), Relationships within families: Mutual influences (pp. 241-260). New York: Wiley.

Haft, W. L & Slade, A. (1989). Affect attunement and maternal attachment: A pilot study. Infant Mental Health loumal. 10(3). 157-172.

Hartmann, D. P. (1977). Considerations in the choice of interobserver reliability estimates. loumaLof Applied Behavior Analysis. IQ (1), 103-116.

Hazan, C. & Shaver, P. (1987). Romantic love conceptualized as an attachment process, loumal of Personal and Social Psychology. 52 (3). 511-524.

Hazan, C. & Shaver, P. (1990). Love and work: An attachment-theoretical perspective. loumal of Personality and Social Psychology. 59 (2). 270-280. Henderson, S. (1982). The significance of social relationships in the etiology of neuroses. In C. M. Parkes & J. Stevenson-Hinde (Eds.), The place of attachment in human behavior (pp. 205-2311. New York: Basic Books.

Hock, E. (1984). The transition to day care: Effects of maternal separation anxiety on infant adjustment. In R. C. Ainslie (Ed.), The child and the dav care setting: Qualitative variations and development (pp. 183-205). New York: Praeger.

Hock, E. & Clinger, J. B. (1981). Infant coping behaviors: Their assessment and their relationship to maternal attributes. The journal of Genetic Psychology. 138. 231-243.

Hock, E. & DeMeis, D. (1990). Depression in mothers of infants: The role of maternal employment. Developmental Psychology. 26 (2), 285-291.

Hock, E., DeMeis, D., & McBride, S. L (1988). Maternal separation anxiety: Its role in the balance of employment and motherhood in mothers of infants. In A. Gottfried & A. Gottfried (Eds.), Maternal employment and children's development (pp. 191-230). New York: Plenum.

Hock, E., McBride, S. L., & Gnezda, T. (1989). Maternal separation anxiety: Mother-infant separation from the maternal perspective. Child Development. 60. 793-802.

Hock, E. & Schirtzinger, M. B. (1992). Maternal separation anxiety: Its developmental course and relationship to maternal mental health. Child Development. £2, 93-102. 143

Hock, E., Schirtzinger, M. B. & Lutz, W. J. (1992). Dimensions of family relationships associated with depression in mothers of young children. Psychology of Women Quarterly. 16. 229-241.

Humphry, R. A. & Hock, E (1989). Infants with colic: A study of maternal stress and anxiety. Infant Mental Health loumal. 10 (4), 263-272.

Hunter, B. (1990). Recollections of attachment: As related to a mothers' attitudes toward career, motherhood, and separation from her first­ born infant. Unpublished doctoral dissertation. Washington, DC.: Georgetown University. Jacobvitz, D., Morgan, E., Kretchmar, M. & Morgan, Y. (1991). Parenting across three generations: The development of maternal empathy. Paper presented at The Society for Research in Child Development, Seattle, WA. Kobak, R. R. & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation, and representations of self and other. Child Development. 59. 135-146.

Kobak, R. R., Cole, H. E., Ferenz-Gillies, R., & Fleming, W. S. (in press). Adolescent attachment strategies and mother-teen problem-solving: A Qrsort analysis. Child Development.

Larzele, R. E. & Klein, D. M. (1987). Methodology. In M. B. Sussman & S. K. Steinmetz (Eds.), Handbook of marriage and the family (pp. 125-155). New York: Plenum. Lederman, R. P. (1984). Psychosocial adaptations in pregnancy: Assessment of seven dimensions of maternal development. Englewood Cliffs, NJ: Prentice-Hall. Leifer, M. (1977). Psychological changes accompanying pregnancy and motherhood. Genetic Psychology Monographs. 25, 55-96.

Levy, D. M. (1943). Maternal overprotection. New York: Columbia University Press. Levy, D. M. (1970). The concept of maternal overprotection. In E. J. Anthony & T. Benedek (Eds.), Parenthood: Its psychology and psychopathology (pp. 387-409). Boston: Little, Brown. Lyons-Ruth, K., Connell, D., Grunebaum, H., Botein, S. & Zoll, D. (1984). Maternal family history, maternal caretaking and infant attachment in multiproblem families, loumal of Preventive Psychiatry. 2 (3 & 4), 403-425. Mahler, M. S., Pine, F., & Bergman, A. (1975). The psychological birth of the human infant: Symbiosis and individuation. New York: Basic Books. 144

Main, M. (1990). Cross-cultural studies of attachment organization: Recent studies, changing methodologies, and the concept of conditional strategies. Human Development. 48-61.

Main, M. & Goldwyn, R. (1984). Predicting rejection of her infant from mother's representation of her own experience: Implications for the abused-abusing intergenerational cycle. Child Abuse and Neglect, ft, 203-217. Main, M. & Goldwyn, R. (in press). Adult attachment rating and classification systems. In M. Main (Ed.), Assessing attachment through discourse. drawings and reunion situations (working title). New York: Cambridge University Press. Main, M., Kaplan, N.,& Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. In I. Bretherton & E Waters (Eds.), Growing points of attachment theory and research (pp. 66-104). Monographs of the Society for Research in Child Development. Serial no. 209, Vol. 50, Nos. 1-2.

Main, M., Tomasini, L & Tolan, W. (1979). Differences among mothers of infants judged to differ in security. Develonmental Psychology. 15 (4), 472-473.

Main, M. & Weston, D. R. (1982). The quality of the toddler's relationship to mother and to father: Related to conflict behavior and the readiness to establish new relationships. Child Development. 52. 932-940.

Matlin, M. W. (1987). The psychology of women. New York: Holt, Rinehart, & Winston.

Mercer, R. T. (1986). First-time motherhood: Experiences from teens to forties. New York: Springer.

McBride, S. & Belsky, J. (1988). Characteristics, determinants, and consequences of maternal separation anxiety. Developmental Psychology. 2A (3), 407-414.

Mikulincer, M., Florian, V., & Tolmacz, R. (1990). Attachment styles and fear of personal death: A case study of affect regulation. Journal of Personal and Social Psychology. (2), 273-280.

Minium, E W. & Clarke, R. B. (1982). Elements of statistical reasoning. New York: Wiley. Morris, D. (1982). Attachment and intimacy. In M. Fisher & G. Strickler (Eds.). Intimacy (pp. 305-323). New York: Plenum. 145

Neale, J. M. & Leibert, R. M. (1986). Science and behavior An introduction to methods of research. Englewood Cliffs, NJ: Prentice-Hall.

Nover, A., Shore, M. F., Timberlake, E. M., & Greenspan, S. I. (1984). The relationships of maternal perception and maternal behavior: A study of normal mothers and their infants. American loumal of Orthopsychiatry. 54 (2), 210-223.

Nuttal, J. R., Stollak, G. E., Fitzgerald, H. E., & Messe, L. (1985). Maternal perceptual style and mother-infant play behavior. Infant Mental Health loumal. & (4), 195-203.

Olson, S. L, Bates, J. E., & Bayles, K. (1982). Maternal perceptions of infant and toddler behavior: A longitudinal, construct validation study. Infant Behavior and Development. 5, 397-410.

Olson, S. L., Bates, J. E., & Bayles, K. (1989). Predicting long-term developmental outcomes from maternal perceptions of infant and toddler behavior. Infant Behavior and Development. 12, 77-92.

Ricks, M. H. (1985). The social transmission of parental behavior: Attachment across generations. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research (pp. 211-227). Monographs of the Society for Research in Child Development. Serial no. 209, Vol. 50, Nos. 1-2. Rubin, R. (1984). Maternal identity and the maternal experience. New York: Springer.

Sander, L.W. (1962). Issues in early mother-infant interaction, loumal of the American Academy of Child Psychiatry. 2. 141-166.

Schaffer, R. (1977). Mothering. Cambridge, MA: Harvard University Press.

Simpson, J. A. (1990). Influence of attachment styles on romantic relationships, loumal of Personality and Social Psychology. 59 (5). 971-980. Stem-Bruschweiler, N. & Stern, D. (1989). A model for conceptualizing the role of the mother's representational world in various mother-infant therapies. Infant Mental Health loumal. 10 (3). 142-156.

Stem, D. N. (1985). The interpersonal world of the infant. New York: Basic Books. Stevens, G. & Featherman, D. L. (1980). A revised socioeconomic index of occupational status. Unpublished manuscript, University of Wisconsin, Madison. 146

Stevenson-Hinde, J, (1990). Attachment within family systems: An overview. Infant Mental Health Toumal. H (3), 218-227.

Vaughn, B., Gove, F. L, & Egeland, B. (1980). The relationship between out-of home care and the quality of infant-mother attachment at twelve and eighteen months: Stability and change in families under stress. Child Development. 2 tl, 971-975.

Wilkinson, L (1990). SYSTAT: The System for statistics. Evanston, IL: SYSTAT, Inc. Winnicott, D. W. (1975). Through naedriatics to psvcho-analvsis. New York: Basic Books.

Zeanah, C. H. & Anders, T. F. (1987). Subjectivity in parent-infant relationships: A discussion of internal working models. Infant Mental Health loumal. & (3), 237-250.

Zeanah, C. H. & Barton, M. L (1989). Introduction: Internal representations and parent-infant relationships. Infant Mental Health loumal. IQ (3), 135-141.

Zeanah, C. H., Keener, M. A., Stewart, L & Anders, T. F. (1985). Prenatal perception of infant personality: A preliminary investigation. loumal of the American Academy of Child Psychiatry. 24 (2), 204-210.

Zeanah, C. H., Keener, M. A.,& Anders, T. F. (1986). Developing perceptions of temperament and their relation to mother and infant behavior. loumal of Child Psychology and Psychiatry. 22 (4), 499-510.