<<

The Effect of Rejection Sensitivity on Perceptions of Inclusion in Cyberball

THESIS

Presented in Partial Fulfillment of the Requirements for the Degree Master of Arts in the Graduate School of The Ohio State University

By

Courtney Kristine Shade

Graduate Program in

The Ohio State University

2010

Master's Examination Committee:

Robert Arkin, Adviser

Russell Fazio

Lisa Libby

Copyright by

Courtney Kristine Shade

2010

Abstract

Rejection sensitivity (RS) is the tendency to anxiously expect, readily perceive, and overreact to rejection (Downey & Feldman, 1996). The role of RS in perceptions of inclusion in a social interaction (Cyberball) and reactions to various degrees of inclusion were investigated in two studies. Participants played Cyberball (Williams & Jarvis,

2006) and were asked to estimate the percentage of tosses they received in the game. It was hypothesized that, when distracted, participants high in RS would underestimate the degree to which they were included in a situation of equal inclusion (Study 1). It was also hypothesized that participants high in RS would have more accurate perceptions than participants low in RS when participants experienced conditions other than equal inclusion (Study 2). In addition, participants high and low in RS were expected to differ in their affective reactions to the same objective event.

In Study 1, participants were included (in the Cyberball game) equally to the supposed others playing. Half of the participants were distracted while playing Cyberball; the remaining half was able to devote full to the game. Overall, participants’ estimates did not differ as a function either of distraction or RS; however, RS and

ii concern for rejection predicted males’ estimates of inclusion but not females’ estimates.

There were no differences in affective reactions to the game.

In Study 2, participants were either extremely excluded, slightly excluded, equally included, or over-included in Cyberball. Concern for rejection (but not RS or expectations of rejection) predicted the accuracy of participants’ responses: Specifically, concern for rejection predicted females’ perceptions of inclusion in the game. The special relevance of concern for rejection as well as the surprising gender effects are discussed.

iii

Dedicated to Ryan

iv

Acknowledgments

This project does not simply represent me. It is the product of the guidance, support, encouragement, and love I have received from many people in my life. First, I want to thank my adviser, Bob Arkin, for his guidance throughout the project. I would also like to thank Russ Fazio and Lisa Libby for serving on my committee.

I am extremely grateful to Alison Pfent, Kenneth Demarree, Pat Carroll and

Michael Edwards for taking time out of their busy schedules and sharing their expertise with me. I would also like to thank Randi Shedlosky, Matt Braslow, and Jean Hancock for their encouragement and entertainment when I’ve needed it most.

I am also extremely thankful for my dear friends. My new “Ohio family” has been amazing throughout the past two years. I am grateful for their patience, support, and encouragement throughout this endeavor. They have done a wonderful job at distracting me when I need to step back and encouraging me when I need to push through.

Finally, I will be forever grateful for the patience, love, and support of my family.

The support of my parents has carried me for years and I will be forever grateful. Ryan, my amazing fiancé, has been my rock. He has not only calmed me down when I don’t think I can go on, but he reminds me of my true purpose and goals. I know without a doubt that I could not have done this without him and cannot begin to express my gratitude for my family. v

Vita

June 2004 ...... Peters Township High School

2008 ...... B.S. Cognitive Studies, Vanderbilt

University

2008-2009 ...... Graduate Fellow, Department of

Psychology, The Ohio State University

1980 to present ...... Graduate Teaching Associate, Department

of Psychology, The Ohio State University

Fields of Study

Major Field: Psychology

vi

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments ...... v

Vita ...... vi

List of Tables...... x

List of Figures ...... xi

Chapter 1: Introduction ...... 1

Ostracism ...... 1

Cyberball ...... 2

Individual Differences in Reactions to Ostracism ...... 3

Rejection Sensitivity ...... 4

RS as a Self-Fulfilling Prophecy ...... 7

Rejection Sensitivity Questionnaire ...... 7

Overview of the Present Studies ...... 8

Chapter 2: Study 1 ...... 11

Overview ...... 11

Method ...... 12

vii

Results ...... 14

Discussion...... 21

Chapter 3: Study 2 ...... 24

Overview ...... 24

Method ...... 25

Results ...... 27

Discussion...... 36

Chapter 4: General Discussion ...... 39

Study 1...... 39

Study 2...... 40

Use of RS Subscales ...... 41

Gender Differences ...... 42

The Role of Ambiguity in RS ...... 42

Future Directions ...... 43

References ...... 45

End Notes ...... 49

Appendix A: Tables and Figures ...... 50

Appendix B: Rejection Sensitivity Questionnaire...... 61

Appendix C: Study 1: PANAS Items ...... 68

Appendix D: Study 1: Anagram Task Instructions and Solutions ...... 69 viii

Appendix E: Rosenberg Self Esteem Scale ...... 72

Appendix F: Study 2: PANAS items ...... 73

ix

List of Tables

Table A.1. Principal component analysis and for Study 1 PANAS data ...... 51

Table A.2. Principal component analysis and for Study 2 PANAS data ...... 52

x

List of Figures

Figure A.3. Number of anagrams remembered with distraction condition and: A total

RSQ score, B expectation of rejection score, C concern for rejection score...... 53

Figure A.4. Percent estimates predicted by gender and: A RSQ total score, B concern for rejection ...... 54

Figure A.5. Estimate accuracy predicted gender and concern for rejection ...... 55

Figure A.6. Percentage estimates predicted by gender and toss percentage ...... 56

Figure A.7. Percentage estimates predicted by concern for rejection and toss percentage for: A males, B females ...... 57

Figure A.8. Estimate accuracy predicted by gender and concern for rejection ...... 58

Figure A.9. Estimate accuracy predicted by concern for rejection and toss percentage for:

A males, B females ...... 59

Figure A.10. Post-game self-reported expectations of inclusion in Cyberball predicted by percentage of tosses received and: A total RSQ score, B concern for rejection, C expectations of rejection ...... 60

xi

Chapter 1: Introduction

Imagine yourself interacting with a group of unfamiliar people. You carry on a conversation with the others, but begin to notice that you are not as involved as the others. How might you respond to this? Would every person respond in the same way?

Might there even be individual differences in noticing that one is less included in the group? In a world characterized by ambiguous social interactions, a situation such as this is likely to occur often. What are the determinants of people’s responses, both their thoughts about the events and their affective reactions?

The guiding hypothesis in this research is that specific features of Rejection

Sensitivity (RS) uniquely predict reactions to ostracism and inclusion. It is well known that one’s expectations about an event guide one’s interpretations (Bruner, 1957). In this context, I hypothesize that anxious expectations of rejection will affect the tendency to perceive rejection, while the degree of one’s concern about rejection will influence a person’s affective reactions to inclusion or ostracism.

Ostracism

The fundamental need to belong makes ostracism especially relevant to people

(Baumeister & Leary, 1995). The need to belong is believed to be so strong that it is characterized as analogous to hunger or thirst (Baumeister & Leary, 1995). Yet, everyone experiences rejection at times. Williams (1997/2001) has proposed and thoroughly researched a “need-threat model” of ostracism, which is rooted in the assumption of a 1 fundamental need to belong. Greatly simplified, this model states that although not all ostracism situations are the same (see Williams, 2001 for a review of taxonomic dimensions and antecedents of ostracism), ostracism can threaten four basic needs.

People experience losses in their sense of belonging, their perception of control over their environment, their self esteem, and feelings of meaningful existence consequent to ostracism. These needs are uniquely threatened by ostracism rather than other aversive social experiences (Zadro, Williams, & Richardson, 2005). The distress caused by a threat to these needs is evident in an increase of salivary cortisol of rejected participants compared to non-rejected participants (Blackhart, Eckel, & Tice, 2007). The psychological pain elicited by ostracism mirrors the brain’s response to physical pain

(MacDonald & Leary, 2005) and people will take action to restore these threatened needs through a wide range of methods (Williams & Zadro, 2005).

Cyberball

Cyberball is one important tool to induce ostracism in the laboratory (Williams,

Cheung, & Choi, 2000; Williams & Jarvis, 2006). Cyberball is a computer-simulated online ball-tossing game in which participants are led to believe that they are interacting with real people1. They are instructed to toss a “virtual ball” with the other players

(usually strangers to the participant). In reality, the other players are “confederates” and the game is programmed to include or exclude the participant according to the requirements of the experiment. Despite the fact that the ostracism is occurring without the physical presence of others, Cyberball has quick and powerful effects on the four needs described above (Zadro, Williams, & Richardson, 2004). The more ostracism a

2 person experiences in Cyberball, the more negative she will feel after the game, as evidenced by an increase in negative affect and a decrease in the four needs described above (Williams et al., 2000). Indeed, participants feel ostracized even when they are told from the beginning that they are playing with a computer that is programmed to ostracize them (Zadro et al., 2005).

Cyberball has proven to be a very useful method to manipulate ostracism in the laboratory because the experimenter is able to program the degree of inclusion each participant will experience. As such, it is possible to create a game in which the participant is included equally, is included and then suddenly excluded, is excluded the entire game, or any other possible variation. It is also useful because the experimenter has the ability to know with precision the exact amount of rejection a participant has experienced. Thus, one can compare participants’ perceptions of and reactions to ostracism with the exact amount of ostracism they experienced (Williams & Jarvis,

2006).

Individual Differences in Reactions to Ostracism

It is no secret in social psychology that needs and motivations direct a person’s attention toward relevant stimuli and guide behavior in goal-directed ways (e.g., Bruner,

1957, Gardner, Picket, & Brewer, 2000). For example, it has been shown that sensitivity to social information, such as perception of and memory for social events, depends on an individual’s current level of belongingness needs (Gardner et al., 2000). Specifically, people whose belongingness needs have not recently been met have a better memory of social events compared to people whose belongingness needs have recently been met

3

(Gardner et al., 2000). This is analogous to the selective memory for food-related stimuli that occurs when a person is hungry (Atkinson & McClelland, 1948). So, if a person has a decreased sense of belonging, either momentarily or chronically, the social environment is perceived in a very different way than for a person who is not experiencing a threat to his or her belonging needs (Gardner, Pickett, & Brewer, 2000).

Importantly, a person’s response to a rejection depends on more than just the situation. Individual differences that each person brings to a situation also contribute to how a person will respond to ostracism and in what way the person is threatened. For example, people who are socially anxious take longer than non-socially anxious people to recover from the painful experience of ostracism, as evidenced by the perpetuation of negative affect even 45 minutes after ostracism; non-socially anxious people have returned to normal within that passage of time (Zadro, Boland, & Richardson, 2005).

Research up to this point has focused on factors that can affect the interpretation of the ostracism event after it has been recognized (Williams & Zadro, 2005). The ostracism literature has yet to investigate individual differences in the tendency to recognize that an ostracism event has occurred, which is the focus of the present studies.

Rejection Sensitivity

RS is the tendency to anxiously expect, readily perceive, and overreact to (Feldman & Downey, 1994). Importantly, the concept of rejection sensitivity

(Downey & Feldman, 1996) takes expectations and motivations into account. The construct of rejection sensitivity (RS) stems from a long and rich historical background

(see Downey & Feldman, 1996). Most notably, the authors draw from Bowlby’s (1957)

4 work describing internal working models of interactions with important people in one’s life. Specifically, Bowlby asserted that internal working models stem from early experiences and ultimately bias information processing. If a person often experiences rejection, she will come to expect rejection in new situations. Because rejection is an unpleasant experience, the motivation to avoid rejection will become salient and this chronic motivation will bias processing with the goal of identifying and avoiding rejection experiences (Bowlby, 1957; Downey & Feldman, 1996; Feldman & Downey,

1994). Thus, RS is a cognitive-affective processing disposition (Feldman & Downey,

1994).

As a result of a learning history saturated with experiences of rejection, a person high in RS learns to expect rejection in social situations. Thus, compared to a person low in RS, a person high in RS has different expectations regarding the degree to which he or she will be accepted. It is well established that people will often use their expectations to help disambiguate a situation (e.g., Bruner, 1957; Kunda, 1990). When placed in a situation that could be interpreted as either rejecting or not, people high in RS tend to interpret the situation as rejecting while people low in RS will tend to interpret the same situation in a non-rejecting way. For instance, in one of the first investigations into RS,

Downey and Feldman (1996) paired participants with a confederate. Shortly into the experiment, participants were told that the confederate did not want to continue the experiment. Participants high in RS tended to attribute the sudden departure to themselves, and felt rejected, while participants low in RS tended to make external

5 attributions and did not feel rejected. Thus, participants high and low in RS interpreted and reacted to the situation in very different ways because of their automatic attributions.

This is in line with Bruner’s (1957) work on category accessibility. He argued that two people can arrive at very different conclusions after perceiving the same event. This is because perception occurs through categorization. The way in which an event is categorized depends on the relative accessibility of categories for the perceiver. The more accessible a category is for the perceiver, the less input is required to identify the stimulus according to the category. In addition, there is a wider range of input characteristics that are deemed relevant. In the case of a person high in RS, the category of rejection is highly relevant (Downey and Feldman, 1996). Because of this enhanced relevance, less rejection should be needed for a person high in RS compared to a person low in RS to perceive that rejection has occurred. Thus, a person high in RS is more likely to notice rejection in a situation.

The enhanced ability to perceive rejection in the environment suggests a hypervigilance for rejection stimuli, such that people high in RS are more likely to attend to rejection-related stimuli. This hypervigilance results in a defensive motivation system

(Bradley, Cuthbert, and Lang, 1990, 1999; Lang et al., 1990). One revealing study used the startle probe paradigm to investigate hypervigilance for rejection stimuli (Downey,

Mougios, Ayduk, London, & Shoda, 2004). Within this paradigm, it has been found that when a person is startled by a loud and unpleasant sound, participants will exhibit a startle response. The startle response is augmented when participants are viewing a negative image rather than a positive image (Bradley et al., 1990, 1999; Lang et al.,

6

1990). Downey and her colleagues (2004) had participants view a variety of paintings, some of which represented social rejection or acceptance. Compared to participants low in RS, participants high in RS experienced a greater startle response when viewing a painting representing social rejection. This difference between people high and low in RS was not found for paintings representing social acceptance or general positive and negative valence. The implication of this finding is that people high and low in RS respond to rejection cues very differently – even cues as benign as a painting. For people high in RS, rejection cues have an elevated importance and attract attention.

RS as a Self-Fulfilling Prophecy

However, an unfortunate result of the special processing priority for rejection cues is the development of a recursive and maladaptive pattern. By perceiving rejection so readily, a person high in RS will often react to a benign situation in a self-protective manner and ultimately hurt the relationships that she is trying to cultivate and maintain

(Downey, Freitas, Michaelis, & Khouri, 1998). If rejection is experienced often enough,

RS is thought to serve a protective role in defending against potential rejection by giving processing priority to rejection cues. Unfortunately, hypervigilance for rejection cues can be maladaptive, as when there is no real threat of rejection it can lead to the perception of rejection when rejection has not occurred. Thus, relationships that exist can easily become undermined for the sake of avoiding and defending against rejection (Downey &

Feldman, 1996; Romero-Canyas, Downey, Berenson, Ayduk, & Kang, 2010).

Rejection Sensitivity Questionnaire

7

RS is measured with the Rejection Sensitivity Questionnaire (RSQ; Feldman &

Downey, 1994). This questionnaire presents participants with 18 scenarios in which there is the potential for rejection. Following each scenario, participants rate the degree to which they feel concern about the outcome of the scenario and the degree to which they expect to be accepted. The score for each scenario is the product of the concern for rejection and expectations of rejection. The total RSQ score is the average of the score for each scenario. Thus, this scale combines both concern for rejection and expectations of rejection into a single composite score. It is possible, however, to create a score representing each participant’s expectations of and concern for rejection.

Overview of the Present Studies

One purpose of the present studies is to provide the first investigation of RS using the Cyberball paradigm. In addition, the present studies will provide a closer investigation into the psychological components of RS, which are generally regarded as expectations of rejection and concern for rejection. Specifically, I will investigate whether an anxious expectation of rejection or concern for rejection is the driving force behind behaviors that are typical of a person high in RS.

In Study 1, participants were either allowed to devote their full attention to

Cyberball or were distracted from the game and then estimated the percentage of tosses they received in the game. Importantly, all participants received 25% of the tosses in the

4 person game. When distracted from the game participants should rely on their expectations of rejection when forced to make an estimate. I hypothesized that RS would predict estimates of inclusion when participants were unable to devote their attention to

8 the game, such that participants high in RS would underestimate the degree to which they had been included. This is because people high in RS have chronic expectations of rejection and, when distracted, they would of necessity have to call on their expectations to make an estimate of their level of inclusion.

Study 2 was a parametric extension of Study 1 without the distraction manipulation. Participants were either extremely excluded, slight excluded, equally included, or over-included in a Cyberball game and asked to estimate the percentage of tosses they received. Recent work on RS as a defensive motivational system (Downey et al., 2004) indicates that the attention of people high in RS is automatically directed toward rejection-relevant stimuli. I hypothesized that Cyberball would serve as a rejection-relevant stimulus. Thus, participants high in RS should have more accurate estimates than participants low in RS.

Investigating the RSQ. In addition, the present studies are a first step into investigating the ability of the two components of the RSQ to uniquely predict reactions to rejection. Even though the definition and literature on RS discuss both concern for rejection and expectations of rejection, the two constructs have not yet been measured and used independently. Can expectations of rejection and concern for rejection be meaningfully separated? In the present research the scale was divided into its two subscales. Each subscale score was used as a predictor, in addition to the total RS score, in order to determine which aspect of the RSQ best predicts which types of behaviors and responses. Specifically, I hypothesized that concern scores would influence affective responses to inclusion or exclusion the outcome of social interactions should be more

9 relevant for participants with high concern for rejection. I also hypothesized that expectations of rejection would affect participants’ accuracy in perceiving exclusion.

Participants with high expectations of rejection should more readily scan the environment for signs of rejection.

10

Chapter 2: Study 1

Overview

This study was designed to investigate whether people high and low in RS differ in their perceptions of an objectively equal situation. Toward this end, we borrowed the

Cyberball paradigm from ostracism research. Within this paradigm, we are able to measure the exact amount of inclusion each participant experiences. Participants received

25% of the total tosses in a game with 4 players, reflecting equal treatment for all participants. They were then asked to estimate the percentage of total tosses they received at the end of the game. Also, an additional aspect of this study was the introduction of a distraction manipulation, which created a situation in which participants were forced to make inclusion judgments based on their expectations alone.

The chief hypothesis was that people high and low in RS differ in their perception of the degree to which they were included in an objectively equal situation. When participants were unable to give their entire attention to the Cyberball task they should be forced to rely on their expectations of inclusion. In this situation, participants with expectations of rejection should estimate that they received a smaller percentage of tosses than participants with expectations of inclusion.

A second cluster of hypotheses concerned affective responses. We hypothesized that participants high and low in RS would differ in their responses to the same objective

11 event. Participants high in RS should experience a decrease in negative affect following a perception of equal inclusion because the outcome exceeded their expectation.

Participants low in RS were not expected to experience any change from baseline affect.

In addition, the self esteem of participants low in RS should be unaffected by equal inclusion in Cyberball; by contrast, participants high in RS were expected to experience an increase in self esteem when they perceived equal inclusion -- and a decrease in self esteem when they perceived less than equal inclusion.

Participants were recruited for a study of memory and were instructed to memorize the solutions to a set of anagrams. Participants in the distraction condition rehearsed the anagrams while playing Cyberball (ostensibly causing them to divide their attention between that task and the Cyberball game). The remaining participants also solved the anagrams. However, they were not required to continue to consider them once they began to play Cyberball. Perceived rejection was measured by participants' estimates of the percentage of tosses they received during the game.

Method

Participants

One hundred three introductory psychology students (52 females) participated in exchange for course credit.

Procedure

Both studies were programmed using MediaLab software (Jarvis, 2005). Thus, participants completed all procedures on a computer. On arrival, participants were

12 informed that they would complete several questionnaires and they were given an anagram task. Participants were told that the experiment was taking place simultaneously in three other labs in the Psychology building and that they were engaged in a competition with all participants during the session to remember the most anagram solutions.

Rejection Sensitivity Questionnaire. The Rejection Sensitivity Questionnaire

(RSQ; Feldman & Downey, 1994; see Appendix B) was administered at the beginning of the study. This scale presents 18 scenarios, each with the potential for rejection (e.g.,

“You ask someone in class if you can borrow his/her notes”; “You approach a close friend to talk after doing or saying something that seriously upset him/her”). For each scenario participants rated how concerned they felt about the outcome of the scenario as well as their expectation that they would be rejected. The score for each item represents the product of participants’ concern for rejection and expectation of rejection (higher expectation scores indicate expectations of acceptance). The total RSQ score represents the average of scores for the 18 scenarios.

Mood Measures. Various items selected from the PANAS (Watson, Clark, &

Tellegen, 1988; see Appendix C) were administered to assess participants’ mood at the beginning of the study and then again at the end. The PANAS addresses specific at a specific point in time. Each participant rated how he or she felt “RIGHT

NOW.”

Anagram task. After completing the questionnaires, participants were directed to the anagram task. The task consisted of 15 anagrams and solutions (see Appendix D)

13 and participants were instructed to solve each anagram as well as to memorize each solution. The purpose of this task was to serve as a distraction during the Cyberball game.

Anagram recall. Participants were asked to recall each anagram solution.

In the no-distraction condition, anagram recall took place immediately after the anagram task, which was immediately prior to playing Cyberball. For participants in the distraction condition, recall took place after the Cyberball game. Further, for those in the

Distraction condition, participants were instructed (prior to playing Cyberball) to rehearse the anagrams while they were playing the game.

Cyberball. Next, participants were told that their fellow players were other participants in the competition located in other areas of the Psychology building. All participants were included in the game, and each received an average of 25% of the total tosses2.

Anagram memory. The number of anagrams recalled served as a manipulation check and was expected to reflect the level of distraction while playing the Cyberball game3.

Percent estimates. After playing Cyberball, participants were asked “Based on your best estimate, what percentage of tosses did you receive?” This measure was intended to represent participants’ perceptions of inclusion in the game.

Rosenberg Self Esteem Scale. Participants also completed the Rosenberg Self

Esteem Scale (Rosenberg, 1965; see Appendix E) at the end of the study.

Results

Creation of RS Subscales

14

In order to create the two RSQ subscales, scores reflecting concern for rejection and expectation of rejection were created. The concern for rejection subscale represents the average of the degree of concern that each person reported for each potentially rejecting scenario in the original RSQ. The expectation of rejection subscale represents the average of the degree to which each participant expected a positive outcome in each scenario. A smaller expectation value indicates that a person has expectations of rejection, while a higher expectation value indicates that a person has positive expectations of the scenario. Each of the dependent variables of this study was analyzed in three different ways. First, the total RSQ score will be entered as a predictor of each dependent variable. Next, the expectation of rejection score was entered into the same regression model in place of the total RSQ score. Finally, the concern for rejection score was entered as a predictor.

Anagram Memory

First, the total number of anagrams that each participant recalled was submitted to the three regression models with gender, the RS score, condition, and the two-way and three-way interactions of these predictors. There were no significant effects of gender or interactions with gender, p’s > .51.

RSQ total score. The total number of anagrams each participant recalled was submitted to a regression analysis with RS, condition, and the interaction of RS and condition as predictors. There was a main effect of condition on anagram recall.

Participants in the distraction condition recalled fewer anagrams than participants in the no-distraction condition, B = -1.65, t = -2.811, p < .01. RS was not a significant predictor

15 of anagram memory, p = .17, but there was a significant interaction of RS and condition,

B = .613, t = 2.746, p < .01 (see Figure A.3a). Participants high in RS remembered the same number of anagrams regardless of the distraction manipulation while participants low in RS remembered fewer anagrams in the distraction condition. The adjusted R2 of this model was .11.

RS expectation score. The total number of anagrams each participant recalled was submitted to a regression analysis with the RS expectation score, condition, and the interaction of RS expectation and condition as predictors. There was a main effect of condition, B = -1.59, t = -2.64, p < .01. Participants in the no-distraction condition remembered more anagrams than participants in the distraction condition. Neither expectation of rejection nor the interaction of RS expectation and condition were significant, p’s > .11 (see Figure A.3b). The adjusted R2 of this model was .06.

RS concern score. The total number of anagrams each participant recalled was submitted to a regression analysis with the RS concern score, condition, and the interaction of RS concern and condition interaction as predictors. There was a main effect of condition on the number of anagrams recalled, such that participants in the no- distraction condition remembered more anagrams than participants in the distraction condition, B = -1.60, t = -2.69, p < .01. Concern for rejection did not significantly predict anagram memory, p = .38. The interaction between distraction condition and concern for rejection was significant, B = 1.53, t = 2.07, p < .05 (see Figure A.3c). Participants with high concern for rejection remembered the same number of anagrams regardless of

16 distraction condition, while participants with low concern for rejection remembered more anagrams in the no-distraction condition than in the distraction condition.

Percent Estimates

RSQ total score. First, participants’ estimates of the percentage of tosses they received during the game were submitted to a regression model with the RSQ total score, gender, distraction condition, and the two and three-way interactions included. There was a main effect of RS, such that percent estimates decreased as RS increased, B = -3.71, t =

-2.38, p < .05. There was also a significant interaction of gender and RS, B = 2.15, t =

2.23, p < .05 (see Figure A.4a). This interaction was decomposed using a simple slopes analysis. For females, RS was not a significant predictor of estimates of inclusion (B =

.59, t = .91, p = .37). For males, RS was a significant predictor, (B = -1.56, t = -2.20, p <

.05). As RS increased for males, their estimates of inclusion decreased. Thus, RS affected the estimates of males but not females. The remaining predictors were not significant, p’s

> .29. The adjusted R2 of this model was .01.

RS expectation score. Next, participants’ estimates were submitted to the regression model described above with the RS expectation score substituted for the total

RSQ score. Within this model neither expectations of rejection, gender, condition, or the interactions were significant predictors of percent estimates, p’s > .61. The adjusted R2 of this model was -.04.

RS concern score. Next, participants’ estimates of the percentage of tosses they received were submitted to the same regression model described above with the concern for rejection score substituted for the total RSQ score. This analysis revealed a significant

17 effect of concern for rejection on percent estimates, B = -16.66, t = -3.09, p < .01. As concern for rejection increased, percent estimates decreased. The interaction of gender and concern for rejection was significant, B = 8.92, t = 2.84, p < .01 (see Figure A.4b).

This interaction was decomposed using a simple slopes analysis. For females, concern for rejection was not a significant predictor, B = 1.19, t = .64, p = .52. For males, concern for rejection was a significant predictor, B = -7.74, t = -3.06, p < .01. Males’ estimates of inclusion decreased as their concern for rejection increased. Thus, concern for rejection affected the estimates of males but not females. The remaining predictors were not significant, p’s > .32. The adjusted R2 of this model was .07.

Estimate Accuracy

An additional dependent measure was created from participants’ estimates of the percentage of tosses received. Because of the slight variation in the percentage of tosses each person received, a score that indicates participants’ accuracy at estimating the percentage of tosses they received was created. This score was created by subtracting the actual percentage of tosses the participant received from the percent estimate that he or she made. A positive value indicates an overestimation of inclusion and a negative value indicates an underestimation of inclusion. For example, if a person received 25% of the tosses and estimated that she received 23%, her accuracy score would be -2. If a person received 27% of the tosses and estimated that she received 27%, her accuracy score would be 0. The estimate accuracy score was entered into the same regression models described above for participants’ estimation of the percentage of tosses they received, resulting in a total of three analyses.

18

RSQ total score. Participants’ total RSQ score, gender, the distraction condition, and the interaction terms did not significantly predict estimate accuracy, p’s >.24. The adjusted R2 of this model was -.06.

RS expectation score. Likewise, participants’ expectation of rejection, gender, the distraction condition, and the interaction terms did not significantly predict estimate accuracy, p’s > .83. The adjusted R2 of this model was -.12.

. RS concern score. Within this model, concern for rejection significantly predicted estimate accuracy, B = -13.18, t = -2.1, p < .05. There was a marginally significant interaction of gender and concern for rejection, B = 6.77, t = 1.89, p = .06 (see

Figure A.5). This interaction was investigated with simple slopes. Concern for rejection was not a significant predictor of accuracy for females, B = .36, t = .18, p = .86. For males, concern for rejection was a significant predictor of estimate accuracy, B = -6.41, t

= -2.15, p < .05. Males with low concern for rejection overestimated their level of inclusion, while males with higher concern for rejection were more accurate, and in fact slightly underestimated their inclusion. The remaining predictors were not significant, p’s

> .54. The adjusted R2 of this model was .004.

Mood Change

Principal components analysis was used to determine the relationship among the

PANAS items. This analysis revealed three components with an eigenvalue greater than

1. Items with a loading of .40 or higher were included to create each component (see

Table A.1). These three components were labeled General Negative Affect, Involvement,

19 and Anxiety. These components represent the sum of the products of each item and its weight.

First, each change score was submitted to the three regression models with gender, the RS score, condition, and the two-way and three-way interactions of these predictors. There were no significant effects of gender or interactions with gender, p’s >

.10. Next, each change score was submitted to the regression models with the RS score, distraction condition, and the interaction of RS and condition as predictors.

General Negative Affect. For each of the three models, there were no effects of distraction condition (p’s > .15), the RS term (p’s > .14), or the interaction terms (p’s >

.15) on change in negative affect.

Involvement. There were no effects of gender (p’s > .07), distraction condition,

(p’s < .84), the RS term (p’s > .78), or the interaction terms (p’s > .21) on change in interest for each of the three models.

Anxiety. There were also no effects of gender (p’s > .18), distraction condition,

(p’s > .10), the RS term (p’s > .40), or the interaction terms (p’s > .48) on change in anxiety for each of the three models.

Self Esteem

Finally, participants’ self esteem scores were submitted to the same regression models described above. There were no effects of gender, distraction condition, or the interaction terms within these three models, p’s > .14.

20

RSQ total score. There was a main effect of RS on self esteem, such that self esteem decreased as RS increased, B= -.62, t = -2.25, p < .05. The adjusted R2 of this model was .13.

RS expectation score. There was a marginally significant effect of expectation of rejection on self esteem, such that self esteem decreased as expectation of rejection increased, B = 2.25, t = 1.81, p = .07.. The adjusted R2 of this model was .07.

RS concern score. Concern for rejection did not significantly predict self esteem, p = .23. The adjusted R2 of this model was .004.

Discussion

My main hypotheses were not supported. Participants’ estimates of inclusion did not differ as a function of distraction condition. The effects of the distraction manipulation, which was intended to prevent participants from devoting their full attention to Cyberball, are inconclusive. It is impossible to determine if participants in the

Distraction condition were actually distracted from Cyberball because there was not an appropriate manipulation check. An appropriate manipulation check would assess participants’ attention to Cyberball. In addition, the inclusion of a control condition, in which participants experienced a delay in anagram recall but did not play Cyberball, would also be appropriate. The measurement of participants’ anagram memory was not a sufficient manipulation check because it provided information regarding their attention to the anagram task but said nothing about attention to Cyberball. Thus, it is impossible to know if participants in the distraction condition paid less attention to the game than participants in the no-distraction condition.

21

The evidence from participants’ anagram recall in fact suggests that participants high and low in RS reacted to the distraction manipulation differently. Participants high in RS remembered the same amount of anagrams regardless of distraction condition, which could suggest that they were more involved with the anagram task than with

Cyberball. One reason for this could in the competitive nature of the anagram task.

Perhaps participants high in RS became more involved in the competition than participants low in RS. Participants low in RS remembered fewer anagrams in the distraction condition, but this provides no evidence of where they allocated their attention during that time. In addition, there were no effects of distraction condition on percent estimates. In sum, the distraction manipulation provides no evidence to either support or refute my hypothesis that people high and low in RS will report different levels of inclusion when they are forced to make judgments based off their expectations of rejection.

An interesting finding was that the RSQ total score and concern for rejection predicted estimates of inclusion. As RS and concern for rejection increased, estimates of inclusion decreased. Even more interesting was the unexpected interaction of RS and concern for rejection with gender. RS did not significantly predict females’ percent estimates, but it did for males. As RS and concern for rejection increased, males’ estimates of inclusion decreased. Males low in RS and concern for rejection overestimated their level of inclusion, while males high in RS and concern for rejection underestimated their level of inclusion. Perhaps females automatically attend to the game,

22 while only males who are high in RS find relevance in Cyberball. To my knowledge, this effect of gender on perceptions of rejection has not been found in the RS literature.

Another interesting result of this study regards participants’ affective reactions to

Cyberball. Although I hypothesized that participants high in RS would experience a decrease in negative affect after equal inclusion in Cyberball, there were no effects of any of my predictors on changes in levels of general negative affect, involvement, or anxiety.

RS did predict self esteem at the end of the experiment in the direction to be expected from the literature (Downey and Feldman, 1996). Participants high in RS report lower self esteem than participants low in RS.

This study investigated participants’ tendencies to perceive rejection in a situation where rejection did not occur. Study 2 was designed to determine if participants high and low in RS could differ in their perceptions of and reactions to social interactions that do not result in equal inclusion. In order to respond to rejection, a person must first perceive that rejection took place. Participants must be able to notice subtle inequities in the degree to which they were included if they are to perceive and react to rejection. Do people high and low in RS differ in their tendencies to perceive rejection? To address this question, participants were exposed to either extreme exclusion, slight exclusion, equal inclusion, and over inclusion in Cyberball in Study 2.

23

Chapter 3: Study 2

Overview

This study was an extension of Study 1 and was designed to further investigate several questions remaining from Study 1. Study 1 found that RS and concern for rejection predicted males’, but not females’ estimates of inclusion. Males high in RS provided lower estimates of inclusion than males low in RS. The first goal of this study was further investigate this effect to determine if RS predicts differences in perceptions of inclusion in conditions other than equal inclusion. To investigate this question, participants were exposed to varying degrees of rejection in Cyberball (extreme exclusion, slight exclusion, equal inclusion, and over inclusion). It was hypothesized that participants high in RS would be more accurate at estimating the percentage of tosses they receive than participants low in RS. This is because RS serves as a defensive motivational system that directs attention to cues signaling potential rejection. Thus, participants high in RS are more likely to attend to the game and perceive deviations from equal inclusion. In addition, I hypothesized that this effect would be stronger for males than for females.

I again hypothesized that the reactions of participants high in RS would be more extreme and thus that they would experience a more extreme affective change following the game than would participants low in RS. When excluded, participants high in RS should become more upset, more lonely, and less happy. When included, participants 24 high in RS should become more relieved, less upset, and more happy. The PANAS items used in Study 1 were modified to better investigate the phenomenological experience I believe to be occurring during Cyberball. Specifically, I added measures of relief, cheerfulness, joy, and surprise.

Another addition to this study was the measurement of participants’ expectations and feelings of inclusion. After playing Cyberball, participants were asked to rate the degree to which they felt they had been included in the game. Next, participants rated the degree to which they had expected (prior to starting the game) to be included in

Cyberball. I hypothesized that participants would report prior expectations that reflected their perception of inclusion in the game. Another goal of this study was to further explore the predictive utility of the expectation of rejection and concern for rejection subscales of the RSQ.

Method

Participants

Two hundred forty five introductory psychology students (106 females) participated in exchange for course credit.

Procedure

Participants were recruited to participate in an experiment supposedly concerning group dynamics. They were informed that the experiment was taking place in other labs across the Psychology building. Participants first completed several questionnaires, then played Cyberball, and then provided their estimates of inclusion. The Cyberball game was not introduced until it was time to play the game. 25

Rejection Sensitivity Questionnaire. The RSQ (Downey & Feldman, 1994) was used as an indicator of participants’ chronic RS, as well as the concern for rejection and expectation of rejection subscales.

Mood Measures. Participants also completed a version of the PANAS (see

Appendix F), created to test the specific predictions of this study. They were completed at the very beginning of the study and then immediately after providing estimates of inclusion.

Cyberball. As in Study 1, participants played Cyberball. The set-up and instructions for the game were the same as in Study 1. After completing several questionnaires, participants were directed to the game. The independent variable in this study was the manipulated degree of inclusion or exclusion during the Cyberball game.

Participants in the extreme exclusion condition received only 10% of total tosses.

Participants in the slight exclusion condition received 21.1% of total tosses. Participants in the equal inclusion condition received 25.5% of the total tosses (this condition is analogous to Study 1). Participants in the over inclusion condition received 35% of total tosses.

Percent estimates. After completing Cyberball, participants were immediately asked to estimate the percentage of tosses they had received during the game. This estimate served as one indicator of each participant’s perception of inclusion.

Self-reported expectations of inclusion. Next, participants were asked “Before you played the game, how much did you EXPECT to be included?” They provided their response on a 7-point Likert scale that ranged from “Not at all included” to “Very

26 included”. Importantly, the expectancy measure was taken retrospectively, after

Cyberball was finished, to prevent any effect of thinking about expectancies on perceptions of the game.

Feelings of inclusion. Next, participants were asked “How included in the game did you ACTUALLY feel?” Participants provided their response on a 7-point Likert scale that ranged from “Not at all included” to “Very included”. This served as a second indicator of participants’ perceptions of inclusion in Cyberball.

Rosenberg Self Esteem Scale. Participants also completed the Rosenberg Self

Esteem Scale (Rosenberg, 1965) at the end of the study.

Results

The independent variable of this study, the percentage of total tosses received in

Cyberball, was represented as a continuous predictor in a regression model. This model contains participants’ RS score, gender, the percentage of tosses received and the two and three-way interaction terms as predictors. As in Study 1, each dependent variable was analyzed in three ways. First, each was predicted using the entire RSQ score. Next, each was predicted by the concern for rejection score, and then by the expectation of rejection score.

Percent Estimates

RSQ total score. First, participants’ estimates of the percentage of tosses they received were submitted to the model described above. Within this model, there was a main effect of the percentage of tosses received, such that percent estimates increased as toss percentage increased, B = .785, t = 3.09, p < .01. Neither gender nor RS were

27 significant predictors of percent estimates, p’s = .53. The interaction terms also did not significantly predict percent estimates, p = .13. The adjusted R2 of this model was .46.

RS expectation score. Next, participants’ estimates were submitted to the regression model described above with the expectation of rejection score substituted in the place of the total RSQ score. Toss percentage significantly predicted percent estimates, such that percent estimates increased as toss percentage increased, B = 1.01, t

= 2.95, p < .01. Gender and expectations of rejection did not significantly predict percent estimates, p’s > .62. Likewise, the interaction terms were nonsignificant, p’s > .25. The adjusted R2 of this model was .45.

RS concern score. Next, concern for rejection was substituted for the total RSQ score in the model described above. Within this model, gender was a marginally significant predictor of percent estimates, B - -10.49, t = -1.70, p = .09. Concern for rejection and toss percentage did not predict percent estimates, p’s > .11. There was a significant interaction of gender and toss percentage, B = .63, t = 2.38, p < .05 (see Figure

A.6). There was also a significant interaction of gender and concern for rejection, B = -

8.26, t = -2.91, p < .05.

In addition, there was a marginally significant three-way interaction of gender, toss percentage, and concern for rejection, B = .32, t = 1.83, p = .07 (see Figure A.7).

This three-way interaction was decomposed by investigating the effects of toss percentage and concern for rejection for males and females separately. I found that concern for rejection significantly predicted percent estimates for females (B = -6.75, t =

-2.09, p < .05) but not for males (B = 1.51, t = .59, p = .55). For females, percent

28 estimates increased as concern for rejection decreased. When excluded, females high in concern for rejection provide much lower estimates than females low in concern for rejection. There was no difference in estimates for females when they were included. The adjusted R2 of this model was .47.

Estimate Accuracy

Estimate accuracy was again used as a measure of the accuracy of participants’ inclusion estimates. This variable was created by subtracting the percentage of tosses each participant actually received from the estimate of total tosses received that each participant provided. A score of zero indicates perfect accuracy, while a positive score indicates an overestimation and a negative score indicates an underestimation. The regression model described above was used for these analyses, as well.

RSQ total score. RS did not significantly predict estimate accuracy, p = .19.

Gender, the percentage of tosses received, and the interaction terms also did not significantly predict estimate accuracy, p’s > .16. The adjusted R2 of this model was .05.

RS expectation score. Next, the expectation of rejection score was substituted as a predictor of estimate accuracy. Within this model, expectation of rejection did not significantly predict estimate accuracy, p = .24. Once again gender, the percentage of tosses received, and the interaction terms did not significantly predict estimate accuracy, p’s > .25. The adjusted R2 of this model was .04.

RS concern score. Next, the concern for rejection score was substituted as a predictor of estimate accuracy. Gender was a marginally significant predictor of estimate accuracy, such that males provided more accurate estimates, B = -10.49, t = -1.70, p =

29

.09. There was also a marginally significant effect of toss percentage on estimate accuracy, B = -.70, t = -1.85, p = .07. Concern for rejection did not predict estimate accuracy, p = .11.

Importantly, there was a significant interaction of gender and concern for rejection, B = -8.26, t = -2.01, p < .05 (see Figure A.8). This interaction was decomposed by a simple slope analysis. It was found that concern for rejection was not a significant predictor of estimate accuracy for males, p = .55. Concern for rejection did predict the accuracy of females, such that females underestimated their level of inclusion as their concern for rejection increased, B = -6.75, t = -2.09, p < .05.

There was also a marginally significant interaction of toss percentage and concern for rejection, p = .09. In addition, there was a marginally significant three-way interaction of gender, toss percentage, and concern for rejection, B = .32, t = 1.83, t = .07 (see Figure

A.9). Males were quite accurate regardless of concern for rejection or toss percentage, both of which did not significantly predict estimate accuracy, p’s > .55 (see Figure A.9a).

Concern for rejection significantly predicted estimate accuracy for females, such that females with higher concern for rejection were generally more accurate than females with lower concern for rejection, B = -6.75, t = -2.09, p < .05 (see Figure A.7b). In addition, toss percentage was also a significant predictor of estimate accuracy for females, B = .21, t = 2.61, p < .05. Females became less accurate and overestimated more as they received a greater percentage of tosses. The adjusted R2 of this model was .07.

Feelings of Inclusion

30

Next, feelings of inclusion were predicted by the model described above. For each of the three RS models, toss percentage was the only significant predictor of feelings of inclusion, B = .11, t = 5.84, p < .001. The remaining predictors were nonsignificant, p’s >

.17.

Mood Change

Principal components analysis was used to determine the relationship among the

PANAS items. This analysis revealed three components with an eigenvalue greater than

1. Items with a loading of .40 or higher were included (see Table A.2). These three components were labeled as reflecting Positive Affect, Negative Affect, and Surprise.

These three components were created by adding the product of each relevant item and its weight.

Positive Affect. The model described above was used to predict change in positive affect after playing Cyberball.

RSQ total score. Neither gender nor any interactions with gender predicted change in positive affect, p’s > .45. Toss percentage was a significant predictor of change in positive affect, such that participants experiences an increase in positive affect as they received a greater percentage of tosses in Cyberball, B = .13, t = 3.74, p <

.001.. Neither RS nor the interaction of RS and toss percentage significantly predicted change in positive affect, p’s > .09.

RS expectation score. With the expectation of rejection score substituted in the above model, neither gender nor any interactions with gender predicted positive change in positive affect, p’s > .56. Toss percentage was again a significant predictor of

31 change in positive affect, such that participants experienced a greater increase in positive affect as they were more included, B = .23, t = 2.86, p < .01. Neither expectations of rejection nor the interaction of expectations and toss percentage were significant predictors of change in positive affect, p’s > .11.

RS concern score. When the concern for rejection score was substituted in the above model, neither gender nor any of the interactions with gender predicted change in positive affect, p’s > .60. Toss percentage was a significant predictor of change in positive affect scores, such that participants experienced greater increases in positive affect as they were more included in the game, B = .16, t = 3.01, p < .01. Neither concern for rejection nor the interaction of concern for rejection and toss percentage were significant predictors of change in positive affect, p’s > .18.

Negative Affect. Next, the model described above was used to predict change in negative affect after playing Cyberball.

RSQ total score. Neither gender nor any of the interactions with gender predicted change in negative affect, p’s > .15. Toss percentage significantly predicted change in negative affect, B = -.10, t = -2.11, p < .05. Participants experienced a decrease in negative affect as they became more included. Neither RS nor the interaction of RS and toss percentage predicted change in negative affect, p’s > .14.

RS expectation score. When the expectation score was substituted into the above model, gender became a significant predictor, B = -3.64, t = -2.18, p < .05.

Males experienced a greater increase in negative affect than did females. Toss percentage also significantly predicted change in negative affect, such that negative affect decreased

32 as toss percentage increased, B = -.26, t = -2.24, p < .05. Expectation of rejection was also a marginally significant predictor of change in negative affect, B = -3.75, t = -1.87, p

= .06. As expectations of rejection increased, participants experienced a greater decrease in negative affect. The remaining predictors were not significant, p’s > .09.

RS concern score. When the concern for expectation score was substituted into the above model, gender did not significantly predict change in negative affect, p = .72. Toss percentage significantly predicted negative affect change, such that participants experienced a greater decrease in negative affect as they received a greater percentage of tosses, B = -.049, t = -2.16, p < .05. Concern for rejection and the remaining interaction terms did not significantly predict change in negative affect, p’s >

.77.

Surprise. Next, the model described above was used to predict change in surprise scores after playing Cyberball.

RSQ total score. Toss percentage was the only significant predictor of change in surprise scores, B = -.01, t = -1.96, p = .05. Participants reported less surprise as they received a greater percentage of tosses in the game. The remaining predictors were not significant, p’s > .36.

RS expectation score. When the expectation of rejection score was substituted in the above model, none of the predictors was significant, p’s > .45.

RS concern score. When the concern for rejection score was substituted in the above model, toss percentage significantly predicted change in surprise scores, B =

-.02, t = -2.12, p < .05. Participants experienced a greater decrease in surprise as they

33 became more included in the game. The remaining predictors were not significant, p’s >

.13.

Self Esteem

Participants’ self esteem scores were submitted to the same regression model described above. None of the rejection sensitivity terms significantly predicted self esteem after Cyberball, p’s > .20. Toss percentage also did not significantly predict self esteem for any of the RS models, p’s > .38. The remaining predictors also were not significant, p’s > .46.

Expectations of Inclusion

Next, the expectation scores that participants provided after playing Cyberball were predicted with the same model described above. Importantly, participants reported their expectations of inclusion after playing the game and estimating their level of inclusion.

RSQ total score. With participants’ total RSQ score entered into the above model, RS was a significant predictor of post-game expectations of inclusion, such that expectations of inclusion increased as RS increased, B = .11, t = 2.24, p < .05. Toss percentage did not significantly predict expectations, p = .19. There was a significant interaction of RS and the percentage of tosses received, B = -.01, t = -2.25, p < .05 (see

Figure A.10a). Participants low in RS reported that they had higher expectations of inclusion as the level of actual inclusion increased. Participants high in RS, on the other hand, reported lower prior expectations as the degree to which they were included

34 increased. Gender and the interactions with gender were not significant predictors, p’s >

.33. The adjusted R2 of this model was .02.

RS expectation score. Next, the expectation score that was derived from the RSQ was substituted as a predictor of post-game self-reported expectations of inclusion.

Expectations of rejection measured by the RSQ significantly predicted post-game expectations of rejection, B = -.65, t = -2.18, p < .05. Post-game reports of expectations of inclusion in Cyberball increased as RSQ expectation scores decreased (become more negative/expect bad things). Toss percentage was also a significant predictor of expectations, such that expectations of inclusion decreased as the degree of inclusion increased, B = -.04, t = -2.34, p < .05. These effects were qualified by a marginally significant interaction, B = .02, t = 1.91, p = .057 (see Figure A.10b). It appears that participants with low expectations of inclusion were more affected by the inclusion manipulation and participants with higher expectations of inclusion were less affected by the manipulation. Gender and the interactions involving gender were not significant predictors of post-game expectations of inclusion, p’s > .35. The adjusted R2of this model was .02.

RS concern score. Next, the concern for rejection score was substituted into the above model. Concern for rejection did not significantly predict self-reported expectations, p = .12. Toss percentage was a significant predictor of expectations, such that participants reported lower expectations of inclusion as they received a greater percentage of tosses, B = -.03, t = -2.47, p < .05. Importantly, this effect was qualified by an interaction of concern for rejection and toss percentage, B = -.01, t = -1.98, p < .04

35

(see Figure A.10c). The expectations of participants who had a low concern for rejection were relatively stable regardless of the percentage of tosses they received, while participants with high concern for rejection reported lower expectations as toss percentage increased. Again, gender and the interactions with gender were not signficant predictors, p’s > .21. The adjusted R2 of this model was .02.

Discussion

The hypothesis that participants high in RS would provide more accurate estimates of inclusion than participants low in RS was not supported. The estimates of participants high and low in RS did not differ. Further, RS did not predict feelings of inclusion. Thus, particpants high and low in RS perceived their interactions in Cyberball in the same way.

One interesting caveat to this are the results when the concern for rejection subscale was used as a predictor of estimates of inclusion. When concern for rejection was entered into the model, there was an interaction of gender and toss percentage.

Concern for rejection predicted females’, but not males’ estimates of inclusion. Males’ estimates of inclusion were quite accurate regardless of concern for rejection. Females with high concern for rejection underestimated their level of inclusion when they were rejected (while females with low concern for rejection overestimated slightly). Although this finding is interesting, it does not provide conclusive support for my hypothesis that participants high in RS would provide more accurate estimates of inclusion. However, the gender effect is intriguing and should be further pursued.

36

The hypothesis that participants high in RS would experience more extreme affective reactions to their perceptions than participants low in RS was not supported.

Participants high and low in RS alike experienced a decrease in negative affect and anxiety as they received a greater percentage of tosses in the game. In addition, all participants reported an increase in involvement as they became more included.

However, the signficant interaction of gender and toss percentage for change in negative affect indicates that males and females responded to Cyberball in different ways. Toss percentage did not predict change in negative affect for females, but males experienced an increase in negative affect as they received fewer tosses in the game. Thus, although males in general provided accurate estimates of inclusion and females tended to underestimate their level of inclusion, males experienced a more extreme affective reaction as measured by change in negative affect. This finding is not extremely surprising given early work on RS, which found that males and females responded differently to perceived rejection. Upon percieving rejection, males tend to become controlling and jealous, while females tend to feel helpless and become hostile (Downey and Feldman, 1996). However, the exact purpose behind males’ more extreme affective reactions to Cyberball remains unclear.

The findings regarding self esteem were quite surprising. The existing literature on ostracism has consistently found that participants experience a drop in self esteem when they are excluded (Williams, 1997/2001). The literature on RS also finds that people high in RS report lower self esteem than people low in RS (Downey and Feldman,

37

1996). Surprisingly, this study did not find an effect of toss percentage or RS on self esteem.

Another surprising finding from this study regards the expectations of inclusion that participants provided after they had played Cyberball. The hypothesis that post-game expectations would mirror participants’ perceptions of inclusion was not supported.

Although expectations of inclusion were generally high, participants high in RS reported lower prior expectations as the degree to which they were actually included increased.

This result makes the absence of mood differences even more puzzling. Participants high in RS reported a big difference in their expectations of inclusion and their actual level of inclusion, which should lead to exaggerated mood differences for participants high and low in RS. Participants high in RS should experience an increase in negative affect when their high expectations of inclusion were met with extreme exclusion. Similarly, participants high in RS should experience a increase in positive affect when their low expectations of inclusion were met with over-inclusion.

38

Chapter 4

General Discussion

In summary, the results of these two studies do not provide evidence that participants high and low in RS differ in their perceptions of inclusion as well as reactions to Cyberball. However, the hypothesis that the two subscales derived from the

RSQ could be effective predictors of participants’ responses was partially supported because the concern for rejection subscale revealed intriguing gender effects.

Study 1

In Study 1, participants received equal inclusion in Cyberball. Half of the participants were distracted from Cyberball with a memory task and the other half were free to devote their full attention to the game. It was hypothesized that, when distracted, participants would rely on their chronic expectations of inclusion when estimating the percentage of tosses received. Thus, participants high in RS should have estimated that they had been less than equally included. This hypothesis was not supported. Participants high and low in RS provided, on average, quite accurate estimates of inclusion regardless of the distraction manipulation.

There was, however, a surprising effect of gender on estimates of inclusion. For males, RS predicted estimates of inclusion. Males with high concern for rejection provided accurate estimates, while males with low concern for rejection overestimated their inclusion. RS did not predict estimates for females, who were quite accurate 39 regardless of RS. This effect was surprising because, to my knowledge, there are no documented gender effects in the RS literature. However, given the contradictory gender effects of Study 2, the meaning behind these gender differences remains unclear.

In addition, the hypothesis that the affective reactions of participants high and low in RS would differ was not supported. Neither RS nor distraction condition predicted change in negative affect, involvement, or anxiety. There was an effect of RS (but not expectations of rejection or concern for rejection) on self esteem, such that participants high in RS reported lower self esteem. This effect is consistent with existing knowledge of chronic RS tendencies (Downey and Feldman, 1996).

My hypotheses regarding the RSQ subscales also were not supported. I hypothesized that the expectation subscale would predict estimates of inclusion and that the concern subscale would predict reactions to the game. In fact, expectations of inclusion did not significantly predict any of participants’ responses. The concern for rejection subscale predicted participants’ perceptions of the game but not their reactions to the game.

Study 2

In Study 2, participants were either extremely excluded, slightly excluded, equally included, or over-included in a game of Cyberball. The hypothesis that participants high in RS would provide more accurate estimates of inclusion was only partially supported.

As in Study 1, the concern for rejection subscale predicted participants’ estimates of inclusion. Importantly, there was a significant interaction of gender and concern for rejection. Concern for rejection significantly predicted estimate accuracy for females but

40 not for males. Females with higher concern for rejection were more accurate than females with lower concern for rejection. Females with lower concern for rejection tended to overestimate their level of inclusion.

In addition, there was little evidence that participants high and low in RS experienced different reactions to Cyberball. All participants, regardless of RS, experienced an increase in positive affect and decrease in negative affect as they experienced greater inclusion. In addition, surprise scores were not affected by either RS or level of inclusion in the game. There also were not effects of either RS or toss percentage on self esteem. Thus, participants did not react differently to varying degrees of inclusion as a function of RS.

Use of RS Subscales

The hypothesis that expectations of rejection would predict estimates of inclusion and concern for rejection would predict reactions to the game was not supported. In fact, expectations of rejection did not significantly predict any responses and, within the context of this experiment, may have merely introduced “noise”. However, there appears to be some utility in separating the subscales of the RSQ. Within the context of this experiment, the concern for rejection subscale predicted participants’ responses most often, even more than the total RSQ score. Although it was not expected for concern for rejection to predict perceptions of rejection, perhaps concern for rejection represents the personal relevance of a situation of rejection. Thus, people with a high concern for rejection are more likely to closely monitor the social environment and provide accurate

41 estimates of inclusion. Importantly, the concern for rejection subscale revealed gender differences.

Gender Differences

This study revealed some interesting gender differences regarding perceptions of rejection in Cyberball. In Study 1, RS and concern for rejection significantly predicted males’ estimates of inclusion. In Study 2, concern for rejection significantly predicted females’ estimates of inclusion as well as change in negative affect for males. At present,

I have no explanation for the lack of consistency of gender effects. In their initial foray into investigating RS, Downey and Feldman (1996) discovered gender differences in reactions to perceived rejection. Males tend to become controlling and jealous, while females tend to feel helpless and become hostile. However, there are no documented gender differences in the tendency to perceive rejection to my knowledge.

The Role of Ambiguity in RS

This investigation concerned the effects of RS on reactions to an unambiguously rejecting or accepting situation and the results mainly did not support my hypotheses. The most likely reason for this in the ambiguity, or lack thereof, in Cyberball. RS is a cognitive-affective processing disposition with the goal of identifying and avoiding situations of rejection (Feldman & Downey, 1994). Some recent scholars have begun to conceptualize this disposition as a schema, which guides attention, elaboration, and memory of expectation-congruent material (Baldwin, 2005; Mor & Inbar, 2009). This schema is chronically accessible, allowing rejection to be readily primed in social situations and even to distract from the task at hand (Berenson et al., 2009; Downey et al,

42

2004). The effects of RS are far-reaching, as outlined in the Introduction. However, schemas have their most dramatic effects on perception in ambiguous situations because they guide the processing and interpretation of incoming information in line with expectations (White and Carlston, 1983). Perhaps when a situation is unambiguously rejecting or accepting, people high and low in RS do not respond differently. In the case of Cyberball, it is likely that people who are low in RS are able to perceive rejection just as readily as people high in RS because it is unambiguous.

The special attention that rejection cues receive for people high in RS, as well as overreactions that are not warranted by the situation, likely only occurs in situations that are ambiguous and could be interpreted in either a threatening or benign manner. This was not the case in the present studies. In sum, participants high and low in RS perceive and respond to varying degrees of rejection in Cyberball in much the same way, as evidenced by their estimations of inclusion and affective reactions. Future research on this topic will further investigate the unique role that expectations of rejection

(particularly the violation of expectations) play in reactions to social interactions of both rejection and inclusion, as well as the unique effects of concern for rejection.

Future Directions

Although my main hypotheses were not supported, the results of these two studies suggest several interesting future directions. Perhaps most interesting are the results regarding gender and the RS subscales. It is apparent that expectations of rejection and concern for rejection predict different responses. The concern for rejection subscale most often mirrored the predictive ability of the total RSQ score. Contrary to my hypotheses,

43 expectations of rejection did not predict estimate accuracy. In addition, it appears that concern for rejection is a potential method of distinguishing gender differences, although the results from these studies are mixed. Males with low concern for rejection overestimated their level of inclusion in Study 1, when they were equally included. In

Study 2, concern for rejection significantly predicted estimates for females but not for males. Future work should continue to investigate the unique interaction of gender and concern for rejection.

Importantly, these hypotheses should be investigated within a new experimental setting, one that includes greater ambiguity in the perception of social cues than afforded by the Cyberball task. Although Cyberball is a very useful manipulation of exclusion and produces quick effects, it may not be best suited for a study of individual differences in the perception of inclusion because of the near total absence of ambiguity in judging inclusion and ostracism. This lack of ambiguity could help to explain why participants high and low in RS did not differ in their perceptions of inclusion. Future work can productively explore the present hypotheses in a more ambiguous social setting, where both the perceptions of inclusion and ostracism can be fully articulated and the importance of ambiguity of the stimulus can be explored.

44

References

Atkinson, J. W., & McClelland, D. C. (1948). The effect of different intensities of the

hunger drive on thematic apperception. Journal of Experimental Psychology, 38,

643-658.

Baumeister, R.F., & Leary, M.R. (1995). The need to belong: Desire for interpersonal

attachments as a fundamental human motivation. Psychological Bulletin, 117,

497-529.

Berenson, K. R., Gyurak, A., Ozlem, A., Downey, G., Garner, M. H., Mogg, K., Bradley,

B. P., & Pine, D. (2009). Rejection sensitivity and disruption of attention by

social threat cues. Journal of Research in Personality, 43, 1064 – 1072.

Blackhart, G. C., Eckel, L. A., & Tice, D. M. (2007). Salivary cortisol in response to

acute social

rejection and acceptance by peers. Biological Psychology, 75, 267-276.

Bowlby, J. (1957). An ethological approach to research on child development. British

Journal of Medical Psychology, 30 (4), 230-40.

Bradley, M., Cuthbert, B., & Lang, P. (1990). Startle reflex modification: or

attention? Psychophysiology, 27, 513–522.

Bradley, M., Cuthbert, B., & Lang, P. (1999). Affect and the startle reflex. In M.E.

Dawson, A.M. Schell, & A.H. Boehmelt (Eds.), Startle modification: Implications

45

for neuroscience, cognitive science, and clinical science (pp. 157–183). New

York: Cambridge University Press.

Bruner, J. S. (1957). On perceptual readiness. Psychological Review, 64, 123-152.

Downey, G., & Feldman, S. (1996). The implications of rejection sensitivity for intimate

relationships. Journal of Personality and Social Psychology, 70, 1327-1343.

Downey, G., Freitas, A., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophecy

in close relationships. Rejection sensitivity by romantic partners. Journal of

Personality and Social Psychology, 75, 545-560.

Downey, G., Mugious, V., Ayduk, O., London, B., & Shoda, Y. (2004). Rejection

sensitivity and the defensive motivational system: Insights from the startle

response to rejection cues. Psychological Science, 15, 668-673.

Feldman, S., & Downey, G. (1994). Rejection sensitivity as a mediator of the impact of

childhood exposure to family violence on adult attachment behavior.

Development and Psychopathology, 6, 231-247.

Gardner, W. L., Pickett, C. L., & Brewer, M. B. (2000). Social exclusion and selective

memory: How the need to belong affects memory for social information.

Personality and Social Psychology Bulletin, 26, 486-496.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. (Chapter

4: The naïve analysis of action.)

Jarvis, W. B. G. (2005). MediaLab (version 2006). New York: Empirisoft.

Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480-

498.

46

Lang, P., Bradley, M., & Cuthbert, B. (1990). Emotion, attention, and the startle reflex.

Psychological Review, 97, 377–395.

MacDonald, G. & Leary M. R. (2005). Why does social exclusion hurt? The relationship

between social and physical pain. Psychological Bulletin, 131, 202-223.

Mor, N. & Inbar, M. (2009). Rejection sensitivity and schema-congruent information

processing biases. Journal of Research in Personality, 43, 392 – 398.

Romero-Canyas, R., Downey, G., Berenson, K., Ayduk, O., & Kang J. (2010).

Rejection sensitivity and the rejection-hostility link in romantic relationships.

Journal of Personality, 78, 119-148.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton

University Press.

White, J. D. & Carlston, D. E. (1983). Consequences of schemata for attention,

impression, and recall in complex social interactions. Journal of Personality

and Social Psychology, 45, 538 – 549.

Williams, K. D. & Jarvis, B. (2006) Cyberball: A program for use in research on

interpersonal ostracism and acceptance. Behavior Research Methods, 38, 174-

180.

Williams, K. D. (2001). Ostracism: The power of silence. New York, NY: Guilford

Publications

Williams, K. D., & Zadro, L. (2005). Ostracism: An indiscriminate early detection

system. In K. D. Williams, J. P. Forgas, & W. von Hippel (Eds.), The social

47

outcast: Ostracism, social exclusion, rejection, and (pp. 19-34). New

York: Psychology Press.

Williams, K. D., & Zadro, L. (2005). Ostracism: An indiscriminate early detection

system. In K. D. Williams, J. P. Forgas, & W. von Hippel (Eds.), The social

outcast: Ostracism, social exclusion, rejection, and bullying (pp. 19-34). New

York: Psychology Press.

Williams, K. D., Cheung, C. K. T., & Choi, W. (2000). CyberOstracism: Effects of being

ignored over the Internet. Journal of Personality and Social Psychology, 79, 748-

762.

Williams, K.D. (1997). Social ostracism. In R. M. Kowalski (Ed.), Aversive interpersonal

behaviors (pp. 133-170). New York: Plenum.

Zadro, L., Boland, C., & Richardson, R. (2006). How long does it last? The persistence of

the effects of ostracism in the socially anxious. Journal of Experimental Social

Psychology, 42, 692–697.

Zadro, L., Williams, K. D., & Richardson, R. (2004). How low can you go? Ostracism by

a computer lowers belonging, control, self-esteem, and meaningful existence.

Journal of Experimental Social Psychology, 40, 560-567.

Zadro, L., Williams, K. D., & Richardson, R. (2005). Riding the “O” train: Comparing

the effects of ostracism and verbal dispute on targets and sources. Group

Processes and Interpersonal Relations, 8, 125-143.

48

End Notes

1 http://www1.psych.purdue.edu/~willia55/Announce/cyberball.htm

2Although Cyberball is easily programmable to create the inclusion situation desired, the actual percentage of tosses that each participant received varied. This is because it is impossible to program the participant’s tosses. Thus, some participants did not receive exactly 25% of the tosses.

3In retrospect, anagram memory was not an appropriate manipulation check for the purposes of this study. The purpose of the anagram task was to distract participants from playing Cyberball, but I have no measure of attention to Cyberball. As such, it is impossible to determine if the anagram truly served to distract participants from

Cyberball, thus rendering my conditions meaningless. One fix for this problem would be the addition of a control condition, where participants recalled the anagrams after a delay that did not involve Cyberball. See Study 1 Discussion.

49

Appendix A: Tables and Figures

50

Component General Negative Affect Involvement Anxiety Active .763 Afraid .781 Alert .718 Angry .756 Ashamed .838 Attentive .752 Determined .735 Distressed .766 Enthusiastic .840 Excited .820 Guilty .804 Hostile .762 Inspired .762 Interested .787 Irritable .663 Jittery .466 Nervous .567 .505 Proud .761 Regretful .704 Sad .719 Scared .797 Strong .764 Upset .850

Table A.1. Principal component analysis and for Study 1 PANAS data.

51

Component

Positive Affect Negative Affect Surprise angry .738 attentive .607 cheerful .828 confident .760 disgusted .651 distressed .617 enthusiastic .779 happy .874 hostile .716 inspired .669 interested .756 joyful .827 lonely .540 proud .742 relaxed .570 relieved .543 sad .666 surprised .453 -.695 timid .525 upset .762

Table A.2. Principal component analysis and for Study 2 PANAS data.

52

9 8.5 8 7.5 7 6.5 6 5.5 High RS 5 4.5 Low RS

4 Anagrams Remembered Anagrams Distraction Distraction Absent Present

Condition a) 9

8 High 7 Expectatio ns 6 Low 5 Expectatio ns Anagrams Remembered Anagrams 4 Distraction Absent Distraction Present b) Condition 9 8 High 7 Concern 6 Low Concern 5

Anagrams Remembered Anagrams 4 Distraction Present Distraction Absent c) Condition Figure A.3. Number of anagrams remembered with distraction condition and a) total RSQ score; b) expectation of rejection score; and c) concern for rejection score as predictors.

53

35 33 31 29 27 25 Female 23 Male

Percent Estimates Percent 21 19 17 15 Low RS High RS

Rejection Sensitivity a)

35 33 31 29 27 25 Female 23 Male

Percent Estimates Percent 21 19 17 15 Low Concern High Concern

Rejection Sensitivity b) Figure A.4 Percent estimates predicted by gender and a) RSQ total score and b) concern for rejection. 54

10

8

6

4 Female Male

2 Estimate Accuracy Estimate 0 Low Concern High Concern

-2 Rejection Sensitivity

-4

Figure A.5. Estimate accuracy predicted gender and concern for rejection.

55

50 45 40 35 30 25 20 Linear (Male) 15 Percent Estimates Percent Linear (Female) 10 5 0 Extreme Slight Control Over (35%) (10%) (21%) (25.5%) Percentage of Tosses Received

Figure A.6. Percentage estimates predicted by gender and toss percentage.

56

50 45 40 35 30 25 20 Linear (Concern Low)

15 Percent Estimates Percent 10 Linear (Concern High) 5 0 Extreme Slight Control Over (10%) (21%) (25.5%) (35%) Percentage of Tosses Received a)

50 45 40 35 30 25 20 Linear (Concern Low)

15 Percent Estimates Percent 10 Linear (Concern High) 5 0 Extreme Slight Control Over (10%) (21%) (25.5%) (35%) Percentage of Tosses Received b) Figure A.7. Percentage estimates predicted by concern for rejection and toss percentage for a) males and b) females.

57

5

0 Low Concern High Concern -5 Male

-10 Female Estimate Accuracy Estimate

-15

-20 Concern for Rejection

Figure A.8. Estimate accuracy predicted by gender and concern for rejection.

58

15

10

5 Linear (Concern Low) 0 Linear (Concern High)

Estimate Accuracy Estimate Extreme Slight (21%) Control Over (35%) (10%) (25.5%) -5

-10 Percentage of Tosses Received a)

15

10

5 Linear (Concern Low) 0 Linear (Concern High)

Estimate Accuracy Estimate Extreme Slight (21%) Control Over (35%) (10%) (25.5%) -5

-10 Percentage of Tosses Received b) Figure A.9. Estimate accuracy predicted by concern for rejection and toss percentage for a) males and b) females.

59

5

4.5

4 Linear (Low RS)

3.5 Linear (High RS) Extreme Slight Control Over

Expectations of Inclusionof Expectations (10%) (21%) (25.5%) (35%) Percentage of Tosses Received a) 5.5

5

4.5 Linear (Low 4 Expectations) Linear (High 3.5 Expectations)

Expectations of Inclusionof Expectations Extreme Slight Control Over (10%) (21%) (25.5%) (35%) Percentage of Tosses Received b)

5

4.5 Linear (Low 4 Concern) Linear (High 3.5 Concern) Expectations of Inclusionof Expectations Extreme Slight Control Over (10%) (21%) (25.5%) (35%) c) Percentage of Tosses Received Figure A.10. Post-game self-reported expectations of inclusion in Cyberball predicted by percentage of tosses received and a) total RSQ score; b) concern for rejection; and c) expectations of rejection. 60

Appendix B

Rejection Sensitivity Questionnaire

Each of the items below describes things college students sometimes ask of other people.

Please imagine that you are in each situation. You will be asked to answer the following questions:

1) How concerned or anxious would you be about how the other person would

respond?

2) How do you think the other person would be likely to respond?

1. You ask someone in class if you can borrow his/her notes.

How concerned or anxious would you be over very unconcerned very concerned

whether or not the person would want to lend 1 2 3 4 5 6

you his/her notes?

I would expect that the person would willingly very unlikely very likely

give me his/her notes. 1 2 3 4 5 6

61

2. You ask your boyfriend/girlfriend to move in with you.

How concerned or anxious would you be over very unconcerned very concerned

whether or not the person would want to move 1 2 3 4 5 6

in with you?

I would expect that he/she would want to move very unlikely very likely

in with me. 1 2 3 4 5 6

3. You ask your parents for help in deciding what programs to apply to.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your parents would want to 1 2 3 4 5 6

help you?

I would expect that they would want to help very unlikely very likely

me. 1 2 3 4 5 6

4. You ask someone you don’t know well out on a date.

How concerned or anxious would you be over very unconcerned very concerned

whether or not the person would want to go 1 2 3 4 5 6

out with you?

I would expect that the person would want to very unlikely very likely

go out with me. 1 2 3 4 5 6

5. Your boyfriend/girlfriend has plans to go out with friends tonight, but you really

62 want to spend the evening with him/her, and you tell him/her so.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your boyfriend/girlfriend would 1 2 3 4 5 6

decide to stay in?

I would expect that the person would willingly very unlikely very likely

choose to stay in. 1 2 3 4 5 6

6. You ask your parents for extra money to cover living expenses.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your parents would help you 1 2 3 4 5 6

out?

I would expect that my parents would not mind very unlikely very likely

helping me out. 1 2 3 4 5 6

7. After class, you tell your professor that you have been having some trouble with a section of the course and ask if he/she can give you some extra help.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your professor would want to 1 2 3 4 5 6

help you out?

I would expect that my professor would want very unlikely very likely

to help me out. 1 2 3 4 5 6

63

8. You approach a close friend to talk after doing or saying something that seriously upset him/her.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your friend would want to talk 1 2 3 4 5 6

with you?

I would expect that he/she would want to talk very unlikely very likely

with me to try to work things out. 1 2 3 4 5 6

9. You ask someone in one of your classes to coffee.

How concerned or anxious would you be over very unconcerned very concerned

whether or not the person would want to go? 1 2 3 4 5 6

I would expect that the person would want to go very unlikely very likely

with me. 1 2 3 4 5 6

10. After graduation, you can’t find a job and ask your parents if you can live at home for a while.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your parents would want you to 1 2 3 4 5 6

come home?

I would expect I would be welcome at home. very unlikely very likely

1 2 3 4 5 6

64

11. You ask your friend to go on a vacation with you over Spring Break.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your friend would want to go 1 2 3 4 5 6

with you?

I would expect that he/she would want to go very unlikely very likely

with me. 1 2 3 4 5 6

12. You call your boyfriend/girlfriend after a bitter argument and tell him/her you want to see him/her.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your boyfriend/girlfriend would 1 2 3 4 5 6

want to see you?

I would expect that he/she would want to see very unlikely very likely

me. 1 2 3 4 5 6

13. You ask a friend if you can borrow something of his/hers.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your friend would want to loan 1 2 3 4 5 6

it to you?

I would expect that he/she would willingly very unlikely very likely

loan me it. 1 2 3 4 5 6

65

14. You ask your parents to come to an occasion important to you.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your parents would want to 1 2 3 4 5 6

come?

I would expect that my parents would want to very unlikely very likely

come. 1 2 3 4 5 6

15. You ask a friend to do you a big favor.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your friend would do this 1 2 3 4 5 6

favor?

I would expect that he/she would willingly do very unlikely very likely

this favor for me. 1 2 3 4 5 6

16. You ask your boyfriend/girlfriend if he/she really loves you.

How concerned or anxious would you be over very unconcerned very concerned

whether or not your boyfriend/girlfriend would 1 2 3 4 5 6

say yes?

I would expect that he/she would answer yes very unlikely very likely

sincerely. 1 2 3 4 5 6

17. You go to a party and notice someone on the other side of the room and then

66 you ask them to dance.

How concerned or anxious would you be very unconcerned very concerned

over whether or not the person would want to 1 2 3 4 5 6

dance with you?

I would expect that he/she would want to very unlikely very likely

dance with me. 1 2 3 4 5 6

18. You ask your boyfriend/girlfriend to come home to meet your parents.

How concerned or anxious would you be very unconcerned very concerned

over whether or not your boyfriend/girlfriend 1 2 3 4 5 6

would want to meet your parents?

I would expect that he/she would want to very unlikely very likely

meet my parents. 1 2 3 4 5 6

67

Appendix C

Study 1: PANAS Items

On each of the following screens you will see a word that describes a different feeling or emotion.

For each word indicate to what extent you feel that way RIGHT NOW.

Active Determined Inspired Regretful

Afraid Distressed Interested Sad

Alert Enthusiastic Irritable Scared

Angry Excited Jittery Strong

Ashamed Guilty Nervous Upset

Attentive Hostile Proud

68

Appendix D

Study 1: Anagram Task Instructions and Solutions

Distraction condition

We are studying the effects of distracters on memory.

After completing the following anagram task, you will play an online ball-tossing game with players from the other labs in Lazenby and the Psychology Building as a

"distracter" from memorizing the anagrams. After the game you will be asked to recall as many of the anagrams as you can.

Anagrams are words that can be scrambled to make a different word (e.g., "listen" can be scrambled to spell "silent").

The anagrams will vary in degree of difficulty. After you enter your answer, you can press "Enter" to reveal the correct answer. If you choose to, you can give up on any given anagram by pressing the "Esc" key - this will show you the answer and then you can continue to the next one by pressing "Continue".

69

Remember, the person from each session who remembers the most anagrams will receive a prize.

Hint: You may want to mentally rehearse the anagrams while you play the game.

No-distraction condition

We are studying the effects of distracters on memory.

You are being asked to memorize words and also solve anagrams. You are only being asked to remember the ANSWERS to the anagrams. After completing the anagram task, you will be asked to remember as many of the solutions as possible. You will then play an online ball-tossing game with players from the other labs in Lazenby and the Psychology

Building as a part of a different experiment.

Anagrams are words that can be scrambled to make a different word (e.g., "listen" can be scrambled to spell "silent").

The anagrams will vary in degree of difficulty. After you enter your answer, you can press "Enter" to reveal the correct answer. If you choose to, you can give up on any given anagram by pressing the "Esc" key - this will show you the answer and then you can continue to the next one by pressing "Continue".

Remember, the person from each session who remembers the most anagrams will receive a prize.

70

Hint: You may want to mentally rehearse the solutions while you go through the anagrams.

Admirer (Married) Team (Meat) Meteor (Remote) Cents

Angered (Enraged) Aches (Chase) Inapt (Paint) (Scent)Wonder

Leaf (Flea) Aloft (Float) Carthorse (Downer)

Canoe (Ocean) Excitation (Intoxicate) (Orchestra) Wolves (Vowels)

Sweat (Waste) Decimate (Medicate) Boost (Boots) Article (Recital)

Observe (Verbose) Catalogue (Coagulate) Broth (Throb) Seaside (Disease)

Below (Elbow) Study (Dusty) Cause (Sauce) Adverts (Starved)

71

Appendix E

Rosenberg Self Esteem Scale

The following is a list of statements dealing with your general feelings about yourself.

The following is a list of statements dealing with your general feelings about yourself. If you strongly agree, click “Strongly Agree”. If you agree with the statement, click

“Agree”. If you disagree, click “Disagree”. If you strongly disagree, click “Strongly

Disagree”.

1. I feel that I'm a person of worth, at least on an equal plane with others. 2. I feel that I have a number of good qualities. 3. All in all, I am inclined to feel that I am a failure. 4. I am able to do things as well as most other people. 5. I feel I do not have much to be proud of. 6. I take a positive attitude toward myself. 7. On the whole, I am satisfied with myself. 8. I wish I could have more respect for myself. 9. I certainly feel useless at times. 10. At times I think I am no good at all.

72

Appendix F

Study 2: PANAS Items

On each of the following screens you will see a word that describes a different feeling or emotion.

For each word indicate to what extent you feel that way RIGHT NOW.

Angry Distressed Interested Relieved

Attentive Enthusiastic Joyful Sad

Cheerful Happy Lonely Surprised

Confident Hostile Proud Timid

Disgusted Inspired Relaxed Upset

73