Citation
Emotion management from late childhood through adolescence : the school as social context

Material Information

Title:
Emotion management from late childhood through adolescence : the school as social context
Creator:
Schillinger, Susan M
Publication Date:
Language:
English
Physical Description:
viii, 144 leaves : ; 29 cm.

Subjects

Subjects / Keywords:
Adolescents ( jstor )
Anger ( jstor )
Emotional development ( jstor )
Emotional expression ( jstor )
Sadness ( jstor )
Schools ( jstor )
Social interaction ( jstor )
Students ( jstor )
Teachers ( jstor )
Womens studies ( jstor )
Dissertations, Academic -- Educational Psychology -- UF ( lcsh )
Educational Psychology thesis, Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2002.
Bibliography:
Includes bibliographical references (leaves 135-143).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Susan M. Schillinger.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
028835401 ( ALEPH )
50450607 ( OCLC )

Downloads

This item has the following downloads:


Full Text










EMOTION MANAGEMENT FROM LATE CHILDHOOD
THROUGH ADOLESCENCE: THE SCHOOL AS SOCIAL CONTEXT















S By ..L. E

SUSAN M. SCHILLINGER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2002























Copyright 2002

by

Susan M. Schillinger


















For my mother














ACKNOWLEDGMENTS

I want to express my most sincere gratitude and appreciation to the members of my doctoral committee, Dr. Patricia Ashton, my chairperson, Dr. Bridget Franks, my cochair, Dr. David Miller, and Dr. Arthur Newman. Their guidance, encouragement, expertise, and patience, have been invaluable sources of support in the completion of this dissertation.

Second, I am grateful to the principals, teachers, and students of Alachua

Elementary, Idylwild Elementary, Kanapaha Middle, Westwood Middle, Gainesville High School, and P. K. Yonge Developmental Research School who willingly volunteered to participate in the dissertation research and in the pilot study. Without their support, this study would not have come to fruition. I owe them all a debt of gratitude.

Third, I wish to thank Mike for sharing his expertise in the complicated and

difficult data analyses. In addition, I would be remiss if I did not express my appreciation to my employer and my work colleagues for their understanding, their cooperation, and their flexibility during the final phase of this dissertation.

Finally, I am indebted to my family and friends for standing by and supporting me during this endeavor. One individual, in particular, deserves special credit for having lived through this adventure with me. He nurtured me when I most needed it, encouraged me when times were rough, and always remained optimistic. I offer a heartfelt thank you to Gabriel for all his support.















TABLE OF CONTENTS
pae~

ACKNOWLEDGEMENTS........................................... iv

ABSTRACT ....................................................... vii

CHAPTERS

I INTRODUCTION ............................................... 1

Statement of the Problem ...................................... 5
Purpose of the Study ......................................... 6
Research Hypotheses ......................................... 8
Significance of the Study ...................................... 9
Delimitations and Limitations of the Study ......................... 12

2 REVIEW OF THE LITERATURE ................................... 13

Emotion Type............................................... 16
Age Differences in Emotion Regulation .......................... 21
Gender Differences in Emotion Regulation ........................ 28
Interpersonal Relationships and Outcome Expectancies ............... 31

3 METHODOLOGY ............................................... 40

Pilot Study................................................. 40
Dissertation Study........................................... 49
Participants ........................................... 49
Research Design......................................... 51
Research Instrument...................................... 51
Procedure............................................... 53
Scoring and Data Analysis ................................. 54
Reliability of the Dissertation Instrument .......................... 55









4 RESULTS ....................................................... 58

Hypothesis 1-Emotion Management Decisions .................... 59
Hypothesis 2-Self-Efficacy Beliefs ....................... .66
Hypothesis 3--Outcome Expectancies ............................ 69
Summary of Results.......................................... 82

5 DISCUSSION ................................................... 86

Adolescents' Emotion Management Decisions ....................... 86
Theoretical Implications of the Research ........................... 100
Practical Implications of the Research ............................ 102
Recommendations for Future Research ............................ 104
Limitations of the Study....................................... 105
Summary .................................................. 106

APPENDICES

A PILOT STUDY QUESTIONNAIRE ................................. 107

B DISSERTATION QUESTIONNAIRE ................................ 111I

C CONSENT FORMS.............................................. 129

D CHILD ASSENT SCRIPTS........................................ 133

REFERENCES.................................................... 135

BIOGRAPHICAL SKETCH.......................................... 144














Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

EMOTION MANAGEMENT FROM LATE CHILDHOOD
THROUGH ADOLESCENCE: THE SCHOOL AS SOCIAL CONTEXT By

Susan M. Schillinger

May 2002

Chairperson: Patricia T. Ashton
Cochair: Bridget A. Franks
Major Department: Educational Psychology

The purpose of this study was to examine developmental changes in adolescents' reasoning in emotion decisions to dissemble or display anger and sadness in a school context. The study also explored adolescents' self-efficacy beliefs about emotion regulation and their outcome expectancies of emotional expression for themselves and two audience figures, a close friend and a well-liked teacher. One hundred forty-three students in grades 5, 8, and 11 participated in this research.

Analyses of variance revealed two important developments in students' reports of their expression of negative emotion. First, 1I th graders were more likely than 5th graders to report that they would display negative emotion to a best friend rather than to a teacher. Second, I Ith graders had stronger efficacy beliefs in their ability to control their negative emotions than did 5th graders.

In addition, a number of gender differences emerged in the analyses. Females, in contrast with males, indicated they would be able to control their negative emotion more








with a teacher than with a best friend. Females anticipated more understanding and support from their best friend than from a teacher for expressing negative emotion. They also reported that they would feel worse if they did not express their emotion to a best friend. Conversely, males indicated no difference regarding decisions to express emotion with either audience figure. They also indicated they would feel better if they did not display their negative emotion to either audience figure.

Participants in all grades anticipated more teasing and belittling from a friend

than from a teacher for expressing anger. Students did not expect negative consequences from either a teacher or a friend for expressing negative emotion. They reported that expressing sadness, but not anger, would likely upset their friend and expressing either negative emotion at school would probably upset the teacher.

These findings add support to previous research on (a) gender differences

indicating the more intimate nature of adolescent female friendships compared to males' and (b) developmental research indicating that students' beliefs in their emotional selfefficacy, and their tendency to confide in a best friend rather than an adult, increase with age.














CHAPTER 1
INTRODUCTION

There has been increased public attention to the concepts of emotional

intelligence (Goleman, 1995; Salovey & Sluyter, 1997) and emotional competence (Saarni, 1990, 1997a, 1999) in recent years. Both concepts have implications for the effective management of emotional behavior and how children and adolescents learn to regulate their own emotions to serve specific goals. This public concern is especially salient given the escalating number of aggressive incidents in American society in general and the rise in aggressive behaviors and disciplinary infractions in the nation's schools. Some schools have responded by offering courses in "emotional literacy" that teach conflict resolution and anger management skills to children as early as kindergarten (Salovey & Sluyter, 1997). Other programs have been designed to increase selfawareness and emotional regulation skills in adolescents. The empirical research related to the theoretical underpinning of these programs and the evidence demonstrating the effectiveness of these programs remain limited despite the widespread adoption of such programs. Recently, developmental, educational, social psychologists, and others have begun to examine the roles of various social-contextual factors and individual characteristics (i.e., age, sex, level of cognitive development, and temperament) as factors in children's and adolescents' emotion management decisions.

Although the concepts of emotional intelligence and emotional competence imply a certain wisdom about what one should do in a given situation when faced with a








pleasant or distressing task or event, the notion of what is wise or prudent action may depend on a number of factors, not the least of which are the situational, interpersonal, social, or cultural contexts that color one's decision making. These varying contextual influences, combined with the traditional belief that emotions are internal feeling states, difficult to substantiate with objective observation, have created a complex challenge for emotional development researchers who want to better understand the relationship between the external and internal variables that influence a person's decision to express or dissemble his or her emotions.

Barrett and Campos (1987) described "a paradigm shift in theories of emotion" and explicated how "theories are moving away from simple stimulus-response orientations to orientations emphasizing personal meanings, organismic strivings, and the functional [emphasis added] importance of emotions" (p. 555). According to the functionalist approach, emotional regulation is a bidirectional interrelational process that includes an individual's appraisal of the significance of an event, the individual's feelings surrounding the significance of the event itself, and the manner in which the person decides to regulate or express emotion in a given environment. Barrett and Campos emphasized the adaptive function of emotions and partitioned the adaptive function of emotion into three types: the first "concerns behavior-regulatory functions; the second concerns internal-regulatory functions; and the third concerns social-regulatory functions" (p. 566). In essence, emotional regulation is not just the decision to regulate one's own internal feeling states and actions but extends to consideration of how others will respond to one's emotional expression and one's goals or outcome expectations. In brief, as Zeman and Garber (1996) noted, "emotions have important communicational








and interpersonal regulatory consequences. The expression of emotion can affect the establishment and maintenance of interpersonal relationships, and, conversely, the social environment can influence whether individuals regulate or display their emotions" (p. 966).

The School as a Social Context

Between the ages of 5 and 18, children and adolescents spend considerable time in the company of their peers and teachers in a closed, formally strucntured, social environment. Successful school adjustment and subsequent behavioral adaptation are largely determined by a child's sensitivity to a number of implicit classroom and social norms. As children enter school, they must learn how to form and maintain interpersonal relationships with classmates as well as with teachers. These relationships are endowed with differing values, expectations, and rules of conduct (Birch & Ladd, 1996; Juvonen, 1996). Children must learn how to communicate effectively within each type of interpersonal relationship and understand the emotional and behavioral signals that contribute to successful relationship building. An important function of the elementary school is to instill prosocial behaviors in children and teach them the social conventions they will need to progress both socially and academically (Wentzel, 1996).

By the time children have reached the later elementary school grades and enter middle school and adolescence, their skill in understanding emotional display rules and reading nonverbal behavioral signals from peers and teachers is thought to be relatively sophisticated (Gnepp & Hess, 1986; Saarni, 1979, 1988, 1989a). By age 11 or 12, most children have internalized "cultural scripts" (Saarni, 1989b, 1990, 1995; Wierzbicka, 1994) and understand how to regulate their own negative emotion in order to protect








themselves or the feelings of another person. These scripts provide guidelines about the outcome expectancies of emotional expression and the likely sequence of events to come in a given social context.

Several recent studies (Fuchs & Thelen, 1988; Zeman & Garber, 1996; Zeman & Shipman, 1997) have explored the role of parents and peers as socialization agents in children's decisions to manage their negative emotions, in particular, anger and sadness. In general, these studies have found that children and adolescents express these two emotions differently depending on which audience figure (parent or peer) is present. In brief, it appears that the outcome expectancies that children and adolescents have about expressing negative emotion change as a function of the social context, the presence or absence of specific key socialization figures, as well as type of interpersonal relationship. All these factors influence children's and adolescents' decisions to suppress or display what they are feeling.

Although most of this research has been conducted in schools and has relied upon the use of self-disclosure questionnaires and structured interview formats, the social context of the school situation has not been the focus of investigation. Most of the vignettes or stories presented to the research participants concerned emotion eliciting situations that occurred outside the school setting; emotions that might be evoked in interactions in the home with a parent or in recreational settings with peers. The influence of teachers and other instructional or administrative personnel as socialization agents for emotional regulation in school settings has received surprisingly little attention in the emotional development literature. Little is known about the interpersonal relationship that develops between teacher and student and how it influences adolescents' emotion








management decisions and outcome expectancies for expressing or dissembling anger and sadness in the school setting.

Statement of the Problem

A considerable amount of the emotional regulation research has focused on children's understanding of display rules for emotions and how they use that understanding to modify their own emotional expression. Several studies have specifically examined how children and adolescents report they would act in negative affect producing situations. Anger has been the most commonly studied negative emotion in infants, preschool children, elementary school-aged children, adolescents, and adults. A few studies have focused on children's or adolescents' sad emotions or on their attempts to regulate them (Garber, Braafladt, & Zeman, 1991; Glasberg & Aboud, 1982; Levine, 1995; Saarni, 1992; Stapley & Haviland, 1989; Stein & Jewett, 1986; Stein & Levine, 1987). This research has examined coping behaviors and strategies employed by various age groups to deal with negative affect. More recently, researchers have begun to study negative affect from an interactional perspective and have noted differences in the ways that children and adolescents cope with their anger, sadness, or fear in interaction with specific socialization figures (i.e., mothers, fathers, and peers).

Although it is generally agreed that parents and peers are the primary socializers of young children's emotional expression, from the time children begin kindergarten, teachers and other school personnel also become socializers of emotional expression. Given the amount of time that children and adolescents spend in school with classmates, teachers, and other school supervisory authorities, it is surprising that so little research about emotional development and/or emotional regulation has been set in this venue. The








education literature is replete with articles devoted to anger management programs at all levels of the curriculum, conflict resolution and peer mediation programs, and selfawareness programs that teach specific emotional regulation skills to children as early as preschool ages. As Zeman and Shipman (1997) have noted, however, few of these studies have investigated the developmental progression of emotion management through the adolescent years. The current study examined factors that may influence adolescents' decisions to control or display their emotions in one specific social context, the school. These factors include type of audience figure (well-liked teacher and close friend), type of emotion (anger and sadness), age (grade level, 5th, 8th, and 11 th), and sex (female and male).

Purpose of the Study

The purpose of this research was to replicate Zeman and Shipman's (1997) research with adolescents and extend it into a new social context. It examined developmental changes in adolescents' reasoning in emotion decisions to dissemble or display anger and sadness from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (11 th grade). I examined adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures, a close friend and a teacher, in a school context. Previous research (Cole, 1986; Fuchs & Thelen, 1988; Saarni, 1979, 1988, 1993; Zeman & Garber, 1996; Zeman & Shipman, 1997) suggests that both the social context and the presence of an audience figure may influence children's and adolescents' reasoning about outcome expectancies or goals in interpersonal relationships. In their work with elementary school children, Saarni (1979) and Zeman and Garber (1996) found that their









explanations for expressing or masking genuine emotion served four purposes: (a) to maintain one's self-esteem, (b) to avoid punishment or further trouble, (c) to maintain the smooth functioning of an interpersonal relationship, and (d) to comply with social convention norm maintenance or role constraints. All four of these purposes could apply to adolescents' emotion reasoning, as well. Zeman and Shipman's research was among the first to explore the relationship between adolescents' decision making about expression of negative emotion and their outcome expectancies. The outcome expectancies they investigated included "interpersonal expectations (positive and negative), non interpersonal [expectations], instrumental expectations (positive and negative), norm maintenance (cultural rules regarding emotion management), and internalization of emotional experience to feel better" (p. 918). Zeman and Shipman found that outcome expectancies varied as a function of emotion type (anger or sadness), audience figure (parent or peer), age, and gender.

The functionalist emotion regulation theorists have contended that adolescents'

decisions to control or express their negative emotion becomes situationally specific with age (Thompson, 1994) and that adolescents' relationships with their peers in a public setting may alter their decisions to express or dissemble anger or sadness for selfpresentation purposes (Zeman & Shipman, 1997). In short, the nature of the situation is likely to affect their willingness to express vulnerability attributable to sadness or their anger at being slighted or frustrated in pursuit of a desired goal

Because little research has examined the role of teachers and their influence on adolescents' emotional responding, I investigated whether the presence of a well-liked teacher would alter adolescents' decisions to control or dissemble emotion. A well-liked







8

teacher was selected in an attempt to standardize the teacher audience figure and to guard against the degree of "liking-of-the-teacher" as a potential source of measurement error. If the functionalist perspective is accurate, in that emotional regulation is bidirectional and interrelational, then the presence of a well-liked teacher or a best friend should affect adolescents' emotional decision making. The teacher's role as authority figure in a school setting may prove to be a significant factor in this situation.

Research Hypotheses

On the basis of previous research (Davis, 1995; Fuchs & Thelen, 1988; Gnepp & Hess, 1986; Saarni, 1979, 1984, 1988, 1989b, 1992, 1995, 1997b; Underwood, Coie, & Herbsman, 1992; Zeman & Garber, 1996; Zeman & Shipman, 1997) and the theoretical perspective of the functionalist approach to emotional regulation, the following general research hypotheses are proposed:

1. Adolescents' decisions to express or dissemble emotion will vary as a function of the type of audience figure, type of emotion, age, and sex of the participants. It is expected that with age, adolescents' decisions to regulate emotion will interact with the sex of the participant, with emotion type, and with type of audience figure.

2. Adolescents will report greater self-efficacy for regulating sadness than for regulating anger.

3. Adolescents' total outcome expectancy score will vary as a function of emotion type, audience figure, age, and sex of the participant. Significant main effects in the total outcome expectancy scores are predicted for type of audience figure, sex of the participant, and age; significant interaction effects are expected for audience figure and








sex of the participant, audience figure and age of the participant, and sex and age of the participants.

Significance of the Study

The theoretical significance of this research is that it tests and extends the

functionalist contention (Barrett & Campos, 1987; Campos, Campos, & Barrett, 1989; Thompson, 1994) that emotional regulation is multifaceted, socially adaptive and interactive, situationally specific, and goal oriented. This study is important to our understanding of adolescent emotion management decisions because the findings will advance and extend our knowledge about emotional regulation in a previously unexamined social context (i.e., the school). It also investigates the influence of a significant but little examined socialization figure, a teacher. Finally, this research expands our understanding of adolescents' social cognition about impression management in the transition from preadolescence, 5th grade, to late adolescence, 11 th grade.

This research also augments our understanding about a significant practical

problem educators in our middle and secondary schools face, that is, an increase in the number of disciplinary infractions and the escalation of students' displays of negative emotion. As students progress from late elementary through middle school, the structure and climate of the schools change. These social contextual factors may evoke negative emotion in early adolescents due to a mismatch between the developmental needs of the students, and of the organization and climate of the new schools (Eccles & Midgley, 1989; Eccles et al., 1993; Wenz-Gross, Siperstein, Untch, & Widaman, 1997).







10
Recent research in school transitions in early adolescence indicates that students' relationships with teachers "deteriorate" as students move into middle schools and junior high schools (Eccles, Lord, & Buchanan, 1996). This student-teacher relationship decline occurs at a time when young people are in need of adults other than their parents who will be positive emotional role models and in whom they can confide (Wentzel, 1997). Eccles et al. (1996) suggested that early adolescents have two specific psychological needs during this transition to junior high school, an "increasing need for autonomy and participation in decisions regarding [their] experiences and the continuing need for strong social supports and close, trusting relationships with adults" (p. 276). A similar contextual change occurs with the transition from middle school to secondary school. High school students may become increasingly angry or sad and alienated by developmentally inappropriate or rigid teacher classroom rules and may drop out of school (Eccles et al., 1993).

Finally, this study contributes to an on-going effort to expand our knowledge of the relationship between the social context, interpersonal relationships (i.e., teacher, close friend), and individual characteristics (i.e., age, sex) of the participants. It enhances our understanding of the relationship between outcome expectancies and adolescents' decisions to express or dissemble anger and sadness in school situations with peers and teachers.

Zeman and Shipman's (1997) recent research provided some insight into the

developmental transition that occurs in emotion management decision making from 5th grade to 1 Ith grade. Their study, however, failed to find significant age-related differences in emotional disclosure with peers. This finding is puzzling in light of the







11
literature on the changing nature of peer relationships during adolescence. Several studies have shown that peer relationships deepen during this period and that girls, in particular, are likely to express their innermost thoughts and feelings with a best friend. Whether adolescent males and females would disclose their negative emotion to best friends and/or well-liked teachers in a school setting was a question in the current study.

Zeman and Shipman's (1997) research did not address emotional disclosure

within a particular social context. Their interest concerned the interpersonal relationship between the adolescent and a specific audience figure. The social context in their investigation comprised the individual characteristics of the participants (age and sex), emotion type (anger or sadness), and outcome expectancies in interactions with a neutral audience figure (mother, father, or best friend).

Moreover, Zeman and Shipman (1997) adapted their vignettes from previous

research on the display rule knowledge of young children. These vignettes may not have been developmentally appropriate for the participants in their study. The pilot study of the current research was designed to elicit information about situations that produce anger and sadness in a school or school-related context from adolescents. This information was collected, analyzed, and reworked into ecologically valid vignettes for the subsequent dissertation research.

In sum, this study contributes to our theoretical understanding of the complex

emotion management decision making that occurs in the transition from late childhood to late adolescence and extends our knowledge of emotional regulation in this age group. This information will have practical significance as well, because it addresses an educational and classroom management problem extant in middle and secondary schools.








Delimitations of the Study

This research was confined to adolescents' decisions to manage their negative emotions of sadness and anger, their sense of emotional self-efficacy concerning emotional regulation, and their outcome expectancies or goals for expressing or dissembling negative emotion. Although intensity of emotion has been implicated in some decisions to regulate anger, this study did not explicitly examine that aspect of emotion regulation decision making in adolescence. This study also relied on the use of self-report instruments that are subject to impression management (Le., the desire to present oneself in the best possible light) and other well documented limitations.

Limitations and Oualifications of the Study

The participants in this research may or may not have been representative of other preadolescent, early adolescent, and late adolescent individuals in other communities. Therefore, the results of this research should be interpreted with caution and not generalized to other populations in the same or similar age ranges. Similar caution is indicated with regard to assumptions about participants' gender-role behaviors. Stereotypical gender-role suppositions of the reader and of the participants may color both data analysis interpretation and/or participants' responses.

All participants were assumed to be "normal functioning" students. They were not labeled "severely emotionally disturbed" or "emotionally handicapped." Special education students with emotional disorders were not selected to participate in this research.














CHAPTER 2
REVIEW OF THE LITERATURE

According to the functionalist perspective (Barrett & Campos, 1987; Campos, Mumme, Kermoian, & Campos, 1994), individuals' decisions to express or dissemble their emotions are embedded in an extensive sociocultural web of emotional beliefs, cultural display rules for emotional expression, and cultural expectancies about the outcomes of expressing or regulating emotional behavior. The individual cannot be separated from his or her sociocultural environment. In this theoretical framework, children come to learn the meaning of emotions in their culture by engaging in social transactions with significant others. Emotions are not seen simply as intrapersonal feeling states. Instead, they are considered the products of social collaboration. More specifically, "emotions are conceptualized as flexible, contextually bound, and goal directed" (Campos et al., 1994, p. 284). Functionalist theory emphasizes the adaptive qualities of emotion and the significance of the person-environment transaction. From this perspective, individuals will differ in the meaning they attach to different events and to the emotions that arise from interpersonal relationships.

Emotional regulation involves adaptation to new person-environmental

transactions. These adaptations often require the masking of genuine emotion in order to achieve one's goals or to avoid undesirable consequences. Campos et al. (1994) contended that "emotion regulation involves selecting responses acceptable to the social group to which one belongs because emotions take place in a social context" (p. 296).








How do children learn how to regulate their own emotion and behavior within a given social context? Saarni (1990, 1997a, 1999) and Wierzbicka (1994) have suggested that children first learn emotional regulation from their parents in the home environment by internalizing what they call "cultural scripts" for appropriate emotional display. Parents impart information about emotional experience and expression to their children by modeling the appropriate behavior and engaging in conversations with their children about feeling states and the meanings that are culturally attached to them. Saarni (1990) referred to these sociocultural rules and beliefs as "naive theories of emotion" and maintained that these theories are dependent on an individual's level of cognitive complexity and emotional development, "unique family influences," and the social transactions that elicit emotions. She argued: "How we make sense of our feelings reveals the impact of socialization, whether it stems from our families, our peers, social institutions such as schools, the media, or our culture in general" (p. 130). In a similar vein, Gordon (1989) stated that children actively construct their emotional lives and once they understand the meaning of a particular emotion, "children become able to act toward it--magnifying, suppressing, or simulating it in themselves, and evoking or avoiding it in other people" (p. 324).

Reichenbach and Masters (1983) have hypothesized that expression of negative emotions may be more subject to socialization efforts than expressions of positive emotions. As previously mentioned, the expression or display of negative emotion, particularly anger, is disruptive to social functioning. It can threaten interpersonal relationships, it can increase aggressive or competitive behavior, and, in general, in Western cultures, its expression is socially disapproved (Stearns & Stearns, 1986). There is increasing evidence that anger and sadness are socialized differentially as a function of







15
sex, gender-role expectation, socioeconomic class, cultural norms, and historical context (see Brody, 1985, 1993; Brody & Hall, 1993 for reviews; Miller & Sperry, 1987; Stearns & Stearns, 1986). With infants, parents have been found to discourage expression of anger and contempt in females, whereas parents engage in more "emotional-matching" behaviors with male babies. This differential socialization might then create "(a) a mistrust of expressions that indicate such states and (b) an assumption that many [situational] contexts that might induce certain states actually will not because the person has 'risen above them'" (Reichenbach & Masters, 1983, p. 1000).

Although Thompson (1994) has argued that a lack of consensus exists about what constitutes emotional regulation, most functionalist theorists believe that emotional regulation is a two-fold process. It can pertain exclusively to the self-management of an individual's internal feeling states or it can be construed as the management of emotional reactions of others. Ultimately, Thompson noted that "emotion regulation consists of the extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions, especially their intensive and temporal features, to accomplish one's goals" (pp. 27-28). This definition emphasizes the functionalist concern with a person's goals and his or her desired outcomes in intrapsychic self-management, interpersonal relationships, or in person-environmental transactions.

Emotion Type

In this study I investigated the negative emotions of anger and sadness of 5th-, 8th-, and 11 th-grade adolescents. Both emotions involve internal feeling states, a set of physiological reactions, a cognitive component, and a set of culturally prescribed rules for overt expression or dissemblance. This section will focus on the causes of anger and sadness and the distinctions between these two emotions. I will describe how the







16
developmental changes in cognition during late childhood and adolescence contribute to decision-making about emotional regulation.

Of the two emotions, anger has received the most research attention because its public display is usually disruptive to the self and to the harmony of an individual's social environment. According to Averill (1982), "more than most emotions, anger is often condemned as antisocial," and "angry outbursts [of children] are usually punished by parents, teachers, and others in authority" (p. 31). Subjectively, anger is typically regarded as an unpleasant negative state of arousal that occurs as the result of some demeaning or offensive act against the self or against "me and mine" (Lazarus, 1991). The offended individual has to decide whether to remain in the unpleasant state or engage in some activity to ameliorate the negative emotion and return to a neutral or positive emotional state. This decision involves taking into consideration his or her personal desires in conjunction with environmental factors that may interfere with his or her wishes.

Anger may also occur as the outcome of the loss of a valued item, person, or goal. Various researchers (Averill, 1982, 1983; Berkowitz, 1990; Izard, 1991; Whitesell & Harter, 1996; Whitesell, Robinson, & Harter, 1993) have identified precipitating actions or events that may lead to anger in children, adolescents, and adults. These include possible or actual physical injury or pain; possible or actual property damage; frustration or the interruption of some ongoing planned activity; events, actions or attitudes that result in a loss of personal pride, self-esteem, or sense of personal worth; violations of socially accepted rules of conduct, and violations of expectations and wishes important to the individual. It is generally acknowledged that for a situation to provoke anger, the offending act must be cognitively appraised as being intentional versus accidental. If the








act is deemed accidental, then it must be appraised by an individual as having been preventable.

Some theorists (e.g., Averill, 1982; Berkowitz, 1990; Izard, 1991; Lazarus, 1991; Stein & Levine, 1987) have noted that interference with or threats to an individual's goal attainment can be instrumental in causing anger, as well. In brief, there is agreement that anger usually results from intentional threats to an individual's well-being, from an interpretation of those threats as personally meaningful and physiologically arousing, and from interference in an individual's pursuit of a desired goal or emotional end-state.

In adolescence, anger often results from an offense against the self. Whitesell and her colleagues (1993) studied anger-provoking situations for young adolescents (sixth, seventh, and eighth grade) and found that the prototypical causes of anger in this age group "included real or threatened physical or psychological harm inflicted on the self by another, typically involving a violation that [was] judged unfair" (p. 522). Examples of physical harm included being hit, kicked, shoved, or beaten up. Psychological harm resulted from peers or friends telling rumors about one's self, talking behind one's back, teasing, put-downs, and not acting as a friend is expected to act (i.e., being loyal, trustworthy, and keeping secrets). As children progressed from middle-childhood to early adolescence, incidents that produced anger were much more likely to result from psychological harm than from physical or verbal assault. In fact, having rumors told about oneself produced as much anger in this age group as did physical assault.

As Whitesell et al. (1993) noted, the shift from physical causes of anger to

psychological causes coincides with "a qualitative shift in the nature and importance of peer relationships, particularly with regard to issues involving mutuality, confidentiality, loyalty, and trust" (p. 523). This finding has implications for the current research study in








that one of the outcome expectancies under investigation concerned the anticipation of understanding and personal support for adolescents' emotional disclosures. It was expected that participants' outcome expectations in the current study would vary as a function of age, gender, audience figure, and emotion type.

Sadness, in contrast to anger, has received limited research attention. Some

studies have focused on children's understanding of the causes of sadness (Glasberg & Aboud, 1982; Levine, 1995; Stein & Levine, 1987; Rotenberg, Mars, & Crick, 1987-88), but little research has examined "normal" sad affect in adolescents. It is important to differentiate normal sad affect from medically diagnosed depression. Sadness is said to be caused by separation (either physically or psychologically) from loved ones, death (especially the loss of an important family member or friend), disappointment, loneliness, and, similar to anger, the failure to achieve specific valued goals (Izard, 1991). The sadness is temporary, not as long-lasting or as debilitating as clinical depression. Loss and the inability to reinstate a desired goal are the most frequently noted causes of sadness in children and adults (Lazarus, 1991; Stein & Levine, 1987).

In one relevant developmental study, Rotenberg et al. (1987-88) investigated the causes, intensity, motives, and consequences of sadness in first-, third-, fifth-, and seventh-grade children. They found that the majority of their respondents believed that sadness was caused by harm. Younger children indicated that harm to themselves would produce the most sadness, whereas the older children (grades 5 and 7) disclosed that harm to others would cause more sadness. Rotenberg et al. noted that the nature of harm invoking sadness also changed with age and perspective taking ability. First and third graders reported physical (concrete) sources of harm as sadness producing, whereas seventh graders cited instances of psychological harm as the cause of their sadness.








These findings are consistent with research on cognitive development during early adolescence and with Elkind and Bowen's (1979) hypotheses of "the imaginary audience" and the "personal fable" that characterize some adolescent behavior. Many early adolescents become egocentric and preoccupied with what others will think about them. For example, in an investigation into the likelihood of emotional expression and its relationship to the consequences of communicating anger and sadness, Fuchs and Thelen (1988) found that sixth-grade boys were especially reticent to express their sadness to their parents; they expected negative interpersonal outcomes for doing so, especially from their fathers. These early adolescent males were concerned about preserving a positive masculine self-image and conforming to their parents' and society's expectations for appropriate masculine behavior.

Rotenberg et aL (1987-88) also recorded the emergence of one other cause of sadness. They found that prevention of goal achievement or attainment also caused sadness, but that this cause appeared to decrease with age. The seventh graders in their study were the least likely to cite prevention of goal attainment as a cause of sadness. Rotenberg and his colleagues suggested that with increased physical and social abilities, older children were less likely to experience prevention of goal achievement. This finding may have developmental implications for the current research in terms of emotional selfefficacy and adolescents' belief in their ability to regulate their own anger or sadness. Emotional self-efficacy may vary as a function of emotion type. It is also possible that, with age and increased cognitive development, older adolescents may use more strategic emotional regulation strategies that vary as a function of social and interpersonal context.

Some researchers (Averill, 1982, 1983; Berkowitz, 1991; Lazarus, 1991; Levine, 1995; Stein & Jewett, 1986; Stein & Levine, 1987) have argued that the same situation or







20
event may trigger anger or sadness depending on the individual's perception of the event. In other words, negative emotional events or situations contain certain salient dimensions such as intentionality, blame-worthiness, personal goals, desirability of the end-state (or goal), and the permanence or reversibility of the loss. Anger typically occurs if the individual concentrates on the cause of the problem and the possibility of goal reinstatement; sadness usually results if the individual fixates on the consequences of the offending act and the impossibility of goal reinstatement (Stein & Levine, 1987).

Stein and Levine (1987) further differentiated between anger and sadness by characterizing them on an active-passive dimension. Of the two affects, anger is considered to be the more action-oriented and outer-directed emotion. It carries with it a necessity to rid one's self of the source of the problem. Sadness, on the other hand, is thought to be a more passive feeling state that is inner-directed and requires no formal action toward its removal.

These findings also have implications for the current study in that they reflect the multifaceted and complex interaction of emotion type with emotional experience and understanding, cognitive development (i.e., the ability to appraise an event as evoking a specific emotion in one's self or in others), gender, and age differences. Likewise, Zeman and Shipman's (1997) investigation of the influence of social context and outcome expectancies on adolescents' decisions to manage their sadness and anger adds two new factors to consider in attempts to understand inconsistencies in adolescents' emotional regulation.

Age Differences in Emotion Regulation

Several studies have focused on the age at which children can distinguish

between the facial expressions for anger and sadness and their situational determinants







21
(Barden, Zelko, Duncan, & Masters, 1980; Denham & Couchoud, 1990; Reichenbach & Masters, 1983). It is now generally believed that by 6 years of age children show consistency in identifying specific contextual and expressive clues for anger and sadness. Furthermore, it has been suggested that children as young as 4 years old are capable of understanding that "the 'appearance' of an emotional expression on the face does not necessarily have to correspond to the 'reality' of the internally felt emotional state" (Harris, Donnelly, Guz, & Pitt-Watson, 1986, as cited in Saarni, 1988, p. 276). Although some 4-year-olds may be able to make this appearance-reality distinction, by age 6 or 7 most children have mastered this skill. At this age they better understand that a person's facial expression may differ from the emotion being experienced and that, at times, it is necessary to dissemble genuine emotion (Davis, 1995; Saarni, 1979, 1988). Harris and Saarni (1989) have suggested that, with age, children become more adept at reading a situation and adjusting their emotional displays to their social context. In other words, even though children recognize that a negative emotion is occurring in themselves or others, they exert control over the external expression of anger or sadness, as the social situation requires, by using socially or culturally prescribed display rules for the expression of emotion. From Harris and Saarni's perspective, control of overt emotional displays may operate on two levels; an introspective self-reflective level (iLe., the child's thoughts and understanding about negative emotion) that mediates between a reflexive response to a negative emotional trigger and another level of social conventions or influences inherent in the particular situation. Both the child's cognitive maturity and the ability to self-reflect on the antecedents of negative emotion and consequences of behavior influence emotional decision making. It is generally believed that by the time children reach preadolescence, around 11 years of age, these cultural scripts for display








rule usage are well established and well practiced. Preadolescents have the cognitive perspective taking and reflective abilities to think about their emotional experiences and, they have internalized the socialization efforts of parents, peers, and significant others. They are aware that their responses to anger or sadness eliciting events will affect other people.

With increasing age, teenagers become more proficient at masking genuine emotion and substituting a more acceptable or appropriate affect in order to attain personal goals. Their efforts to conceal or control emotional displays become more strategic, and they have a larger repertoire of tactics they can employ to manage negative emotions. Adolescents can reflect on their own cycles of emotion and accompanying behavior in terms of patterns. They understand that one emotion can trigger a second and that both influence their own and others' cognition and behavior (Saarni, 1990, 1992, 1995).

Several studies (Cole, 1986; Davis, 1995; Fuchs & Thelen, 1988, Gnepp & Hess, 1986; Saarni, 1979, 1984, 1988, 1989b, 1992, 1997b; Underwood et al., 1992; Zeman & Garber, 1996; Zeman & Shipman, 1996, 1997, 1998) have examined age differences in emotional regulation and children's understanding about how and when to use culturally acquired display rules to dissemble emotion. The majority of this research has been conducted with children ages 6 through 11 using the disappointing gift paradigm developed by Saarni (1984). This method assesses children's understanding of the social convention that it is better to mask one's disappointment when receiving an inappropriate or "baby" present than to hurt the feelings of the gift giver. Few studies have investigated adolescents' emotional display rule knowledge, or ability, or motivation to regulate their social or interpersonal relationships. It is presumed that this age group is very








knowledgeable about the appropriateness of masking, substituting, minimizing, or maximizing their expression of negative emotion. However, little research has been conducted on adolescents' ability or perhaps, of more importance, their motivation to dissemble negative emotion.

Gnepp and Hess (1986) examined 1st-, 3rd-, 5th-, and 10th-grade children's

understanding of the use of verbal and facial display rules. They found that children as early as first grade are aware that use of a cultural display rule can have prosocial or selfprotective functions. Furthermore, children's knowledge about when, where, and with whom to express or mask genuine emotions increases across the grade school years. Gnepp and Hess noted, however, that the 5th graders' responses did not differ significantly from the 10th graders. It seems noteworthy that teenagers, who have the emotional knowledge and the ability to be proficient at using cultural display rules, "predicted that the story characters would regulate their facial expressions less than half of the time" (p. 106). The authors speculated that the lack of a significant difference between 5th and 10th graders' predictions of story characters' use of facial display rules may have been due to (a) the 10th-graders belief that their peers "either cannot or will not regulate their facial displays of emotion, even if ideally they should" (p. 108) (i.e., an ability versus motivational component) or (b) the simplistic nature of the protocol, that could have produced the lack of difference between groups. Gnepp and Hess chose to limit the number of facial responses from which their participants could select. They suggested that the 10th graders may have wanted to display more "subtle" facial expressions than were available within their coding system. Either one of these hypotheses might have been explored, yet there has been little follow-up research to








investigate this similarity between 5th- and 10th-grade students' reported use of facial display rules.

Saarni's (1988) research in this area is instructive. In a recent study that investigated children's social cognition about the interpersonal consequences of "presenting an emotional front" to adults or peers, she interviewed children from three age groups, 6-7, 10-11, and 13-14. The children were shown a series of emotion eliciting vignettes and were asked how others "are likely to react" when genuine emotion is expressed or when someone tries to mask his or her true feelings. These children were also asked to specify which audience figure (adult or peer) they would prefer as the recipient of the expression of genuine emotion. Moreover, they were asked what they thought the expected outcomes would be for a peer who either constantly dissembled or expressed genuine emotions. Finally, these children were asked how they arrived at a balance in making their emotional decisions. The last question, according to Saarnmi "required children to take into account both unique self-knowledge as well as situational knowledge" (p. 277). In general, with age, children thought that they would be successful in "fooling" the various vignette characters. Older girls preferred to express genuine emotions to peers rather than to adults. Older boys had more cynical reactions regarding whether they could successfully fool a "bully" (over 50% of the 13-14 males thought that a bully would see through a "false front"). Older boys also thought that they would not be able to fool the school principal after participating in a school prank. Taken together, these studies raise important questions that need to be further examined regarding adolescents' decision making about masking or expressing negative emotions. It is not clear whether ability, motivation, or outcome expectancy is most influential in such decision making.







25
With a few exceptions, as noted above, most of the research in the understanding of display rule usage and decisions about whether to display genuine negative affect has centered on the social cognitive development of first- through sixth-grade children. Two recent studies (Papini, Farmer, Clark, Micka, & Barnett, 1990; Zeman & Shipman, 1997) specifically focused on adolescents' decisions to disclose their emotions to parents and peers. Papini et al. focused on 174 junior high school students between the ages of 12 and 15. They found that the females in their study engaged in more emotional self-disclosure to parents and peers than did the males. Furthermore, with age, adolescents chose to confide their feelings to peers rather than to parents. Younger adolescents were more likely to disclose to parents than to peers. Because the focus of their study was neither the use of cultural display rules, nor, more specifically, the negative emotions of anger and sadness, the implications of their findings for the proposed research are somewhat limited.

Papini et al. were more interested in the quality of the parent-child interaction: the openness of parent-child communication, family cohesiveness, and the affective quality of family functioning. Adolescents who perceived their parents to be receptive to their emotional disclosures were more inclined to disclose. Adolescents' emotional selfdisclosures to friends were associated more strongly with their own psychosocial characteristics (i.e., self-esteem and identity status) than with the quality of family functioning. Papini et al. suggested that their findings lend support to the social distancing hypothesis of Steinberg (1989), which predicts that, with age and pubertal maturation, adolescents distance themselves from their parents and associate with other adolescents who are experiencing similar physical changes and emotions.







26
More recently, Zeman and Shipman (1997) examined developmental changes in emotion management decisions across a broader range of adolescents; preadolescents (5th graders), early adolescents (8th graders), and late adolescents (1 I th graders). Their goals were to examine adolescents' decision making about expression of genuine versus dissembled emotion, adolescents' emotional self-efficacy beliefs in their ability to control their own anger and sadness, and the outcome expectancies these participants had for three distinct audience figures (i.e., mother, father, and best friend). The combination of audience figure and the age and gender of the adolescent comprised the social context for expression of negative emotion. The research question was--Would adolescent emotion management decision making vary as a function of the social context?

In general, with age, their participants reported regulating sadness more than anger. Pre-, early, and older adolescents indicated that they felt more emotional selfefficacy for the regulation of sadness than for anger. There appeared to be general agreement in all three age groups that sadness should not be openly expressed to parents or peers. In contrast, adolescents in all age groups indicated that they would feel worse if they did not let their anger out. The research literature has demonstrated a link between children's beliefs in their ability to control a negative emotion and its intensity. In general, the higher the intensity of a negative affective state, the less children perceive it to be within their power to control.

Consistent with one of their hypotheses, Zeman and Shipman (1997) found that the presence of the audience figure did influence adolescents' decisions to display or dissemble their anger or sadness. Eighth-grade adolescents were significantly more likely to report regulating their negative emotions with their mothers than with their fathers or peers; they expected to receive the least interpersonal support from mothers than did








either the younger or older groups. Eighth graders also expressed concern that their negative emotion expression might produce a reciprocal reaction from their mothers and best friends but not from their fathers. This finding lends further support for Steinberg's (1989) social distancing hypothesis. It also supports other research into developmental changes that occur in parent-child relationships during adolescence; decreased cohesion (Papini et al., 1990; Steinberg, 1988, 1989, 1999) increased conflict between early adolescents and parents (Collins & Russell, 1991), and less emotional self-disclosure between parent and child (Buhrmester & Furman, 1987).

With peers, Zeman and Shipman's (1997) findings were somewhat puzzling and inconsistent with findings in previous research. No significant differences were found between 5th-, 8th-, or 11 th-graders' emotion management decisions or in the consequences they expected for expressing genuine negative emotions with best friends. Previous research in peer relationships has indicated that there is a consistent increase of emotional disclosure between close friends (Belle, Burr, & Cooney, 1987; Belle & Longfellow, 1984; Berndt, 1982, 1996; Youniss & Smollar, 1990) beginning in late childhood and early adolescence. Zeman and Shipman theorized that their research protocol may have produced only mild negative emotion that would have been insufficient to demonstrate the age-related changes or concern over the consequences of its expression with peers. They speculated that the nature of their vignettes might not have produced adequately intense anger or sadness in their participants because the story characters were not the emotion provocateurs (i.e., triggers of the emotion experience).

The preceding explanation has significance for the current research study. Zeman and Shipman (1997) adapted their vignettes from previous research into the display rule knowledge of young children. These vignettes may not have been developmentally







28
appropriate for the participants in their study. The pilot study of the current research was designed to elicit information about situations that produce anger and sadness in a school or school-related context from adolescents. This information was gathered, analyzed, and reworked into ecologically valid vignettes for the dissertation study.

To sum up, with age, children become more adept at assessing the situational determinants that influence their emotional regulation decisions. Cognitive maturity (perspective-taking ability), emotion type (anger or sadness), emotional experience, the presence of peers, parents, or unrelated adults, all affect children's outcome expectancies and enter into their decision making processes with respect to genuine emotional expression or dissimulation.

Gender Differences in Emotion Regulation

Zeman and Shipman (1997) found some gender differences in decisions to

express negative emotion as a function of emotion type. Males reported regulating more negative emotion than females and believed it was more important to dissemble their sadness than their anger. Overall, males expected "less understanding and more belittling" in response to emotional disclosure. They believed that they should not express their emotions, more so than did females, and males thought they would feel better if they kept their emotions to themselves. This finding is consistent with the literature on gender differences in emotional expression and with the social support literature. In times of stress, females are more likely to seek interpersonal support and engage in emotional self-disclosure with close friends. Males tend to disclose emotionally when they believe society expects such behavior (Belle et al., 1987).

Other researchers studying emotional expression have found that emotional

regulation decisions are influenced by gender and gender-role socialization (Birnbaum &







29
Croll, 1984; also see Brody, 1985; Brody & Hall, 1993, Deaux & Major, 1987; Shields, 1995, for reviews). Females are usually more emotionally expressive overall than males, but males express anger more overtly and intensely than females do (Diener, Sandvik, & Larsen, 1985; Fujita, Diener, & Sandvik, 1991; Sonnemans & Frijda, 1995). Gender differences in anger expression may also interact with socioeconomic status (Miller & Sperry, 1987), individual characteristics, and the interpersonal context of the emotional exchange (Fabes & Martin, 1991; Fabes, Eisenberg, Smith, & Murphy, 1996; Whitesell & Harter, 1996; Whitesell et al., 1993).

In general, there is some evidence that females both experience and express more happiness, sadness, and fear than males do, whereas males experience and express more anger than females (Brody, Lovas, & Hay, 1995). There is further evidence that the socialization of anger differs by sex; males are encouraged to display their anger when they are provoked (i.e., to use external coping strategies and physical or verbal responses to provocation), whereas females' overt expression of anger is widely discouraged. There is some evidence that females are socialized to use a more "internalizing" coping strategy and tend to turn their anger inward toward the self, especially during adolescence (Renouf& Harter, 1990). In other words, females have been found to use more nonaggressive and indirect strategies to cope with their anger.

Observational studies of gender differences in dyadic interactions between

mothers and their young infants have noted that mothers displayed more positive affect to female infants than to male infants and, in general, demonstrated more emotional expressivity to females (see Brody & Hall, 1993 for a review). Other observations of parent behavior with preschoolers have found similar gender differences in socialization. In one study, Greif Alvarez, and Ulman (as cited in Brody & Hall, 1993) found that








fathers used more emotion words, overall, with their daughters than they did with their sons, and mothers minimized discussing anger with their preschool daughters, but openly discussed angry feelings with their sons.

Fivush (1989) examined mother-child conversations about the past and found, with preschool-aged daughters, mothers rarely spoke about anger but did so with their sons; when preschool-aged daughters had been angered by a peer, mothers emphasized the feeling of sadness because the child's friendship with the preschool peer had been threatened by conflict. Mothers have also been observed to discuss the causes and consequences of emotion more with sons than with daughters and to pay more attention to male toddlers' expressions of anger. In contrast, the angry outbursts of female toddlers were ignored or inhibited by their mothers (Brody & Hall, 1993).

Shields (1995) argued that gender-role socialization within the family and peer

group is significant in that it initiates and sustains the connection between emotion beliefs and values in gender development. On the basis of the work of Deaux and Major (1987), she stated that "gender-related behaviors are influenced by three aspects of the context: expectations of perceivers, self-systems of the actor, and situational clues" (p. 213). Negative emotions are gender-coded. The outward expression of anger is most commonly associated with the masculine role; the internalization of anger and sadness is characteristic of the female role. The most notable difference between the sexes, particularly in adults, is women's reportedly greater conflict about openly expressing anger. In Shield's research, adult women described anger as effective but upsetting and costly to relationships.

To summarize, gender differences in emotion regulation, that is, decisions to express the negative emotions of anger and sadness, have been linked to emotion type,








early gender-role socialization by parents and peers, individual overall emotional expressivity, socioeconomic status, the interpersonal context of emotional expression, expectations of the perceiver, self-system of the individual, and situational clues. According to Saarni (1988), "by middle childhood children are capable of subtle insights into how emotional experience and social context are to be integrated" (p. 290). The current research was designed to build on Saarni's contention that adolescents have attained this level of social cognitive development. It is based on her prediction that adolescents will use their implicit or naive theories of emotion to regulate their own negative emotion and to consider the consequences of their dissembled or genuine emotional displays on others as a function of the social or interpersonal context.

Interpersonal Relationships and Outcome Expectancies

The interpersonal relationships between mothers and adolescents, fathers and adolescents, and between adolescent best friends have been the focus of considerable research in emotional display and expectations for social support. Little research, however, has been conducted on adolescents' emotional expression or dissemblance with a non-parental adult figure. For example, Fuchs and Thelen (1988) conducted a study that specifically focused on interpersonal relationships and the consequences children (Le., first, fourth, and sixth grade) expected for expression of anger and sadness with parents. They found that older children, especially males, held less positive expectancies for genuine expression of emotion with parents than did younger students. This was especially true for males and the expression of sadness.

Overall, males became less emotionally expressive with age and tended to mask their feelings of sadness, especially with their fathers. Fuchs and Thelen (1988) suggested that parental socialization practices are directed toward the suppression of sadness in








boys. On the contrary, girls, in their study, expected to receive more support and fewer negative consequences for genuine expression of both sadness and anger with parents. These findings demonstrated a strong correlation between the likelihood of expression of negative affect and its anticipated consequences, but the focus of the study was limited to the interpersonal context between children and parents; a private world in contrast to the public setting of schools. The social contextual variable was not considered.

Whitesell and Harter (1996) conducted a study varying the interpersonal context as the critical element in early adolescents' decisions to express anger at a personal violation. The sample consisted of two age groups; preadolescents (ages 11 and 12) and young adolescents (ages 13-15). Their focus was on peer relationships and how the changing nature of adolescent friendship status affects decisions to express or dissemble negative emotion. They varied their hypothetical vignettes to depict anger-provoking actions of either best friend or casual acquaintance. Whitesell and Harter found gender differences in the ratings of personal violation ensuing from the offending act of being called "stupid" by best friends or classmates. Girls reported higher ratings of more intense negative emotions (i.e., anger, sadness, and hurt feelings) than did boys and were more relationship oriented in their choice of response strategies. Furthermore, with age, females were more likely than males to report a stronger sense of violation if the angerprovoking action had been the action of a best friend. The older females were also more likely than males to report hurt feelings and a deep sense of sadness when wronged by a best friend.

According to Whitesell and Harter (1996), hurt feelings are a complex blend of negative emotion comprised of significant amounts of both anger and sadness, as well as fear about the potential loss of a valued relationship. With casual acquaintances like








classmates, no significant relationship existed. Consequently, in their study, when offended by a classmate, the most commonly reported negative emotion for both sexes was anger.

Whitesell and Harter's (1996) research underscores how the interaction between age, gender, expectation for a friend's behavior and how the interpersonal context of a relationship (i.e., friendship status) can shape adolescents' interpretation of and responses to negative emotional elicitors. It did not specifically address, however, the various types of outcome expectancies adolescents might consider in response to an emotionally provocative event, nor did it examine decisions to express anger within a given social context if the anger experienced had not been evoked by a best friend or classmate.

Saarni's (1988) investigation featured peers and non-parental adult figures in

emotionally evocative vignettes depicting common situations that occur at home, during play, and at school Her participants were students in Grades 2, 5, and 8. As previously stated, she found age and sex differences in her participants' beliefs about how others would react to their genuine or dissembled emotion. The reported consequences for emotional display varied as a function of the social context described in the vignette. The relationship between the protagonist and the peer or non-parental adult (iLe., an aunt or a male school principal) affected the participants' decisions to express or dissemble genuine emotion as well as their expectations of support, teasing, or punishment.

In one vignette about accidentally setting off a fire alarm in a school setting and being apprehended by the principal, older children (i.e., especially female eighth graders) thought that if they did reveal their fear of punishment to the intervening principal, he would deal with them less harshly. According to Saarni, this anticipated consequence indicated that older children recognized that displaying genuine emotion for a misdeed








might be a strategic and regulated act to escape severe negative punishment. Younger children tended to focus only on the consequences of breaking school rules, and 73% of them expected negative reactions. They were not as likely to believe that genuinely expressed emotion would reduce the principal's wrath.

Saarni's (1988) investigation highlighted the interaction between age, sex, social cognition, and the integration of these variables with the interpersonal and situational context. Although her study provided support for children's understanding of the appropriate use of culturally prescribed display rules for emotional expression, her findings are limited to middle childhood and have not been replicated in a similar study investigating emotion management decision making in an adolescent population. The current study was designed to extend our knowledge of adolescents' outcome expectancies and how these expectations regulate their own emotional expression and influence the emotional experience of specific others in an educational context.

These expectancies include positive supportive reactions from one's parents or peers, negative reactions such as teasing, negative instrumental consequences, such as punishment or losing privileges, upsetting the other person in the relationship and making them feel bad, a belief that one's emotional expression might violate a social rule, and finally, anticipating that not expressing genuine emotion might make one feel worse than if one dissembled it.

Zeman and Shipman (1997) focused on these outcome expectancies and how the social context or interpersonal relationship(s) between adolescents and parents and peers interact with emotional regulation decisions regarding sadness and anger. As noted earlier, Zeman and Shipman incorporated the audience figure from previous research into their investigation of social-contextual influences on outcome expectancies for managing







35
anger and sadness in 5th-, 8th-, and 1 Ith-grade participants. The combination of audience figure and the age and gender of the adolescent comprised the social context for expression of negative emotion. The results confirmed their predictions that the presence of an audience figure (mother, father, or best friend) would influence adolescents' decisions to regulate their negative emotion. With age, adolescents' reported regulating their emotions more in the presence of their parents, especially their mothers, and had very different outcome expectations based on the presence of the specific audience figure.

In general, participants reported that they expected more teasing and belittling from best friends if they genuinely expressed anger or sadness. Furthermore, this group of adolescents anticipated more loss of privileges and negative instrumental consequences from both mothers and best friends in contrast to the reaction of fathers. These findings need to be put in perspective, however. In this study, the emotionally evocative event was simulated by a series of vignettes to which the participant was asked to respond as s/he believed the protagonist would respond. The parent or best friend was not the emotional provocateur depicted in the vignette; they were neutral observers, in whose presence the emotional regulation decision making occurred. The vignettes were deliberately designed to disentangle the influence of the emotional trigger from the expression of anger or sadness. Zeman and Shipman (1997) were interested more in the decision making process and whether the presence of the audience figure influenced adolescents' outcome expectancies than in the type of emotion expressed.

This is an important distinction to make because previous studies on emotional expression have frequently confounded these influences. The current research was designed to separate the effect of the emotional provocateur from the decision making







36
process as Zeman and Shipman did in 1997. It differed from their study however, because the research instrument was created from the responses of the population of interest in the study, 5th-, 8th-, and I Ith-grade adolescents. The final emotion-evoking vignettes were more appropriate and representative of the types of situations that anger or sadden adolescents in school.

As noted earlier, the majority of the research on adolescents' anticipated

consequences of emotional disclosure has focused on interactions between themselves and their parents or their peers, or both. Very little research has been conducted within social contexts like schools in which adolescents express or dissemble their emotions with non-parental authority figures. This is an important area of study because the presence of an adult authority figure might well influence the decision making that occurs in emotional regulation. Teacher-student relationships have received virtually no attention in the emotional development or emotional regulation literature, even though children spend a considerable portion of their waking hours in a social context that creates a different type of interpersonal relationship that may affect outcome expectancies and or consequences for genuine emotional expression.

For example, in a study of school-aged children in the third, fifth, and seventh grades, Underwood et al. (1992) asked their participants to respond hypothetically to vignettes depicting hostile, neutral, and non-hostile events in a school situation. They featured peers and teachers as the emotional provocateurs in the videos. Participants first were asked how they would feel if the depicted event happened to them. Next, the participants were asked about the type of facial expression they would display. Finally, the children were asked about their reasoning for choosing the expression they reported







37
and whether they thought the emotional provocateur (teacher or peer) was being unfair or mean.

Underwood et al. (1992) found that participants responded differently to the peer vignettes than to the authority figure vignettes. Overall, the participants reported that they would feel angry if a peer was the provocateur, and this response appeared to be independent of the intentionality of the act. Underwood et al. had varied the intentionality in their vignettes to reflect hostile, non-hostile, or ambiguous acts. Their participants, however, did not interpret the intentionality as the researchers had conceived it. They perceived non-hostile and ambiguous acts by peers as eliciting anger. The teacher vignettes evoked less anger but more sadness in their participants. These children reported using display rules for anger more often for teachers than for peers and chose more nonconfrontational actions with teachers than with peers. With age, the children reported more masking of angry facial expressions, but only with teachers. Girls reported more masking of anger expression than did boys, but all these findings appeared to be a function of the social context conveyed by the vignettes.

Some unusual age by gender by emotion interactions occurred in this study. For boys, there was a trend, with age, to report using display rules for anger. Seventh-grade boys reported using more display rules for anger than did third graders. For girls, the opposite occurred; seventh-grade girls reported less display rule usage than third-grade girls. A similar developmental trend was reported for dissembling sadness. With age, boys increasingly reported masking their sadness, whereas the opposite was noted for girls. Third-grade girls reported invoking more display rules than did seventh-grade girls.

The research of Underwood et al. (1992) demonstrated third-, fifth-, and seventhgrade children's understanding of socially conveyed display rules and how the nature of








the interpersonal relationship can effect decision making about expressing genuine negative emotion. It should be noted, that in this study, the social and cultural expectations for anger reactions and aggressive displays of behavior differed between researchers and participants. The researchers were employed by a private university whose faculty and students are predominantly Caucasian and middle to upper-middle class; the participants were African American children from urban, low-income families who resided in the university town. According to the authors, all non-African American children were excluded from participation because race was not to be a factor in their investigation. Underwood et al. were puzzled because their peer vignettes evoked more anger responses and retaliatory behavior, overall, even when the provoking incident was ambiguous. They speculated that their participants' interpretation of psychological or physical harm might have stemmed from their social and cultural upbringing and could have influenced this finding (Miller & Sperry, 1987).

Nevertheless, this study is valuable in demonstrating the developmental

differences in third, fifth, and seventh graders' understanding of appropriate emotional disclosure in an understudied social context, the school. It also highlights the differences in interpersonal relationships between peers and teachers and suggests that schools and teachers are important socialization agents of students' negative emotional expression and behavior. It further emphasizes that young adolescents are aware that "it is more dangerous and less socially acceptable to acknowledge [and display] anger toward authority figures than peers" (Underwood et al., 1992, p. 376). However, as previously stated, the problem with this study as with many others, is that display rule awareness and emotional management decisions are confounded with the authority of the emotional







39
provocateur. Adolescents' decisions to express or dissemble genuine emotion may be less constrained when the teacher is a well-liked but neutral audience figure.

Summary

In brief research suggests that the understanding and use of display rules are

related to emotional intensity, age, gender, emotion type, social cognitive development, interpersonal relationships, socialization figures, and the social and cultural factors embedded in various social contexts. In short, with age, children acquire emotional knowledge that they apply to situations with increasing social cognitive ability if the motivation to do so matches their needs and expectations for desirable interpersonal outcomes. Taken together, the results of this literature review indicate a need to investigate adolescents' thinking about their decisions to express or mask their genuine negative emotion in school, to explore their self-efficacy beliefs about their ability to control negative emotion, and to examine their outcome expectations for expressing negative emotion in the presence of a well-liked teacher or a best friend.














CHAPTER 3
METHODOLOGY

The purpose of this research was to examine developmental changes in

adolescents reasoning in emotion decisions to dissemble or display anger and sadness from preadolescence (5th-grade) through middle adolescence (8th-grade) to late adolescence (11 th-grade). I also examined adolescents self-efficacy beliefs about emotional regulation and their outcome expectancies of displaying or suppressing emotion for themselves and two audience figures, a close friend and a teacher, in a school context.

Because little research has examined the role of teachers and their influence on adolescents emotional responding, I investigated whether the presence of a well-liked teacher would influence adolescents decisions to express emotion. This chapter describes the participants, research instruments, data collection procedures, and methods of data analysis used in the pilot study and subsequent dissertation study.

Pilot Study

A pilot study was conducted to identify school-related events or situations that elicited feelings of sadness and anger in 5th, 8th, and 11 th grade students. The questions were open-ended to allow participants to respond freely from their personal experiences at school.








Participants

The participants were drawn from a population of students attending local

elementary, middle, and high schools in Alachua County, Florida. The total population of Alachua County, according to the 2000 U.S. Census, was 217,955 and was comprised of 73.5% White, 19.3% Black or African-American, .2% Native American, 3.5% Asian,

5.7% Hispanic or Latino of any race, 1.4% some other race and, 0.1% two or more races. One hundred thirty six students from five local public schools participated in the pilot study. Thirty-six 5th-grade (13 males and 23 females), 52 8th-grade (17 males and 35 females), and 48 lth-grade (24 males and 24 females) students participated in the first phase of this study. The 5th graders ranged in age from 9 years 10 months to 11 years 8 months (M= 10 years 5 months, SD = 3.19 months). The 8th graders ranged in age from 12 years 9 months to 15 years 2 months (M= 13 years 4 months, SD= 3.78 months). The 11th graders ranged in age from 16 years to 18 years 6 months (M= 16 years 5 months, SD 3.44 months). The ethnic composition of the sample was 64.7% Caucasian, 22.1% African-American, 3.7% Asian-American, and 2.9% Hispanic. Five percent of the students described themselves as multi-ethnic. A detailed description of the sample by grade level is provided in Table 1.

Procedure

The pilot study questionnaire (See Appendix A) was administered to 5th-, 8th-, and 11 th-grade students who had returned their parental consent forms. The students were asked to describe events or situations that produced anger or sadness at school or in a school-related context. They received no compensation for their participation. Students were told that the research was concerned with learning more about common








Table 1

Description of the Pilot Study Sample


Sex 5th % 8th % 11th % Total Male 13 36.1 17 32.7 24 50 54 Female 23 63.9 35 67.3 24 50 82 Total 36 100 52 100 48 100 136 Ethnicity

Asian or Pacific Islander 2 5.5 1 1.9 2 4.2 5 African-American 13 36.1 5 9.6 12 25.0 30 Hispanic 2 3.8 2 4.2 4 White (Non-Hispanic) 19 52.8 39 75.0 30 62.5 88 Multi-Ethnic 2 5.5 4 7.7 1 2.1 7 Other 1 1.9 1 2.1 2 Total 36 100* 52 100* 48 100* 136

*Total percentage has been adjusted by rounding.


feelings that most students experience related to schooling. They were then asked to respond, in writing, to three open-ended questions and instructed to mark the intensity of their emotion on a 4-point Likert-type scale.

The questionnaire requested the student to,'Think about and write down

something that recently happened at school or in a school related event that made you feel angry (or sad)" The second part of the question referred to the intensity of the experience:'l "How angry (or sad) did this make you?' Students were asked to quantify the








intensity of their emotion by selecting a numerical response on a scale ranging from

1 (only a little annoyed or sad) to 4 (extremely angry or sad). All participants were requested to relate at least one specific situation for anger and one specific situation for sadness. In order to remediate any potential residual negative emotion in participants in the pilot study, the questionnaire also asked students to think about and relate an event that made them feel happy or glad. This question was always asked last.

The pilot study information was vital to the broader dissertation study for the

following reasons: first, by collecting student-generated emotion evoking situations, the construct validity of the emotion evoking vignettes of the final assessment instrument was increased. The vignettes used in similar studies of childreds and adolescent decision making about regulation of anger or sadness (Zeman & Garber, 1996; Zeman & Shipman, 1997) were constructed for use with younger children and were later modified for their study with adolescents. Zeman and Shipman hypothesized that perhaps the mild emotion evoked by their vignettes was not sufficient to produce the predicted age differences in emotional disclosure to best friends versus parents.

In another related study, Underwood and her colleagues (1992) were surprised that the anger producing scenarios they created for their research project did not evoke the anger responses they anticipated from their participants. The current pilot study was designed to reduce the likelihood of similar inauthentic or developmentally inappropriate emotional scenarios.

Second, because the next phase of this research involved local students, the

student-generated emotional events were meant to compensate for potential differences in student perception of emotional events that might result from demographic and








geographic particulars. Previous research on this topic has been conducted in differing regions of the country, notably the northern United States. Student? experiences at school and teacher behavior and practices might vary as a function of geographic location. Assessing local students experiences was intended to increase the ecological validity of the emotion evoking vignettes on the dissertation instrument and contribute to more valid and reliable research results.

Data Analysis

Holsti (1969) described content analysis as a'echnique for making inferences by objectively and systematically identifying specified characteristics of messageg'(p. 14). In accordance with the guidelines for this type of analysis, students responses were read, independently, by two reviewers (i.e., the principal investigator and a second independent reviewer) and examined for commonalities. Categories were created from the literature and were derived from "groups of words with similar meaning or connotation" (Weber, 1990, p. 37). The categories were also revised as necessary to be mutually exclusive and exhaustive so that every response was covered by the coding system. Six categories emerged from the analysis and represented six major themes: (a) physical aggression to the self or toward another person (11%), (b) psychological injury to the self by a friend, a teacher, a classmate, or an unnamed other individual (44%), (c) getting a bad grade (5.1%), (d) unfair rules (21.3%), (e) being unjustly or falsely accused of something (8.8%), and (f) being treated like babies (1.5%). Three other categories were created to account for the remaining responses and to comply with the guidelines for content analysis (Holsti, 1969); a denial of anger category (2.2%), a'bot school related'situation category (.7%), and a no response category (2.9%). The raterg percent of agreement for








the anger scenarios was 92% for the 5th-grade responses, 90% for the 8th-grade responses, and 94% for the I Ith-grade responses. All differences were resolved by discussion. The anger categories for the content analysis were consistent with previous research findings for causes of anger in preadolescent and adolescent children.

The tendency to become angry as a result of physical aggression to ones self

decreased with age from a 5.9% response rate from 5th graders to 1.5% for 11 th graders. A noticeable increase in the total number of psychological injury anger situations reported from 5th grade to 8th grade supported Whitesell and her colleagues (1993, 1996) findings with adolescents. The number of anger situations concerning rules that were deemed to be unfair also increased considerably with age; 4.4% for 5th graders, 6.6% for 8th graders, and 10.3% for 11 th graders. See Table 2 for frequency of anger category by grade.

Sadness categories were derived following the same procedure used in the anger content analysis and were informed by a review of the literature. Nine themes emerged from participants responses: (a) physical or psychological separation from friends or loved ones, (b) death, (c) general disappointment, (d) loneliness and alienation, (e) failure to achieve a specific goal, (f) physical harm or injury to ones self or another person, (g) psychological harm or injury to oneds self caused by a friend, classmate, teacher or unnamed other person, (h) psychological harm or injury to another person, and (i) loss of personal property or possessions. As in the anger content analysis, three other categories were created to fulfill the requirements of content analysis; denial of sadness, a'bot school related'category and no response. Interrater percent of agreement for the sadness scenarios was 94% for the 5th-grade responses, 92% for the 8th-grade responses,









Table 2

Frequency Table of Anger Categories by Grade


Grade

Anger Category 5th 8th 11 th Total %


physical aggression to self physical aggression to other(s) psychological injury by a friend psychological injury by a teacher psychological injury by a classmate psychological injury by unnamed other(s) getting a bad grade unfair rules

being unjustly/falsely accused treated like babies denial of anger not school related no response Total


8 4 2

1

4 8 3 4 10 7 6 2 9 2 4 4 1 4 2 6 9 14 2 9 1

2

2 1


2

36 52


and 92% for the I Ith-grade responses. All differences were resolved by discussion. See Table 3 for frequency of sadness category by grade.


10.3

.7

11.2 15.4 12.5 7.4 5.1 21.3 8.8

1.5 2.2 .7

2.9 100








Table 3

Frequency Table of Sadness Categories by Grade


Grade

Sadness Category 5th 8th 1lth Total % physical separation 4 3 1 8 5.9 psychological separation 4 7 1 12 8.8 death (of friend, relative, acquaintance) 3 4 4 11 8.1 general disappointment 1 1 8 10 7.4 failure to achieve a specific goal 3 7 6 16 11.8 physical harm to another person 2 4 6 4.4 psychological harm to self by peers 10 8 3 21 15.4 psychological harm to self by teachers 2 1 1 4 2.9 psychological harm to another person 2 10 13 25 18.4 loss of personal property 2 2 4 2.9 denial of sadness 2 2 3 7 5.1 not school related 2 2 4 2.9 no response 3 3 2 8 5.9 Total 36 52 48 136 100







48
Consistent with the research of Rotenberg et al. (1987-88), there was an age trend in reported sadness situations resulting from psychological harm. Fifth-grade students reported experiencing sadness resulting from some harm to themselves (28% of 5th-grade responses), whereas the number of instances of psychological harm reported by 8th- and 11 th-grade students decreased with age, from 15% to 6% respectively. In contrast, the number of sadness situations incurred by psychological harm directed at a person other than themselves increased with age; from 5.5% in the 5th-grade to 27% in the 1 lth-grade.

Eighth graders were most likely to cite psychological separation from a friend as a cause of sadness in school. Their responses centered around friends transferring to different schools and the break-up of romantic relationships. Eleventh-grade students were most saddened by teasing, bullying, or harassment of other students with limited mental or physical capabilities in the school situation. Thirty-one percent of their responses in this category made reference to this type of incident. This finding is consistent with the literature on decreasing self-involvment during late adolescence and an increase in perspective taking abilities (Steinberg, 1996).

Furthermore, students were more reticent to acknowledge their sad emotion than their anger. The "no response" rate doubled from the questions about anger situations to those about sadness. The "denial of emotion" rate increased from a total of three for the anger questions to seven for situations involving sadness.

Participants responses were tallied by emotion type (anger or sadness) to determine the most frequently mentioned incidents by grade. The categories were compared with recent findings in the literature for similarity of response with other preadolescents and adolescents regarding the causes of anger and sadness (Karniol &








Heiman, 1987; Rotenberg et al., 1987-1988; Tangney et al., 1996; T6restad, 1990; Whitesell et al., 1993) and found to be consistent. The most frequently cited category for each emotion, plus the overall most frequently cited category, were then used to construct four anger and four sadness vignettes for the dissertation study.

Dissertation Study

Participants

The participants in this study were 21 5th graders (10 males and 11 females, 46 8th graders (14 males and 30 females), and 50 1 Ith graders (25 males and 25 females) drawn from P. K. Yonge Developmental Research School. An additional 26 fifth-grade students (10 males and 16 females) were recruited from one of the pilot study elementary schools to augment the fifth-grade student total and maintain approximately equal numbers of participants by grade. In all, 146 students participated in this study. The 5th graders ranged in age from 10 years 2 months to 12 years 2 months (M= 10 years 5 months, SD = 3.66 months). The 8th graders ranged in age from 13 years 1 month to 15 years 2 months (M= 13 years 5 months, SD = 3.5 months). The 1lth graders ranged in age from 16 years 2 months to 17 years 8 months (M= 16 years 7 months, SD = 3.3 months). All students received a free video rental certificate as compensation for their participation.

P. K. Yonge was the school of choice for two reasons; first, the students who

attend P. K. Yonge are demographically representative of the population in North Central Florida. The school uses a lottery system for admission to ensure an equitable distribution of students based on gender, race, socioeconomic status, and academic ability. Second, a majority of P. K. Yonge students commenced their education there, and did not








experience potentially disruptive school transitions from elementary to middle school, and from middle school to high school that could contribute to measurement error. See Table 4 for a description of the dissertation study participants. Table 4

Description of the Dissertation Study Sample



Sex 5th % 8th % 11th % Total Male 20 42.5 14 31.8 25 50.0 59 Female 27 57.5 30 68.2 25 50.0 82 Total 47 100 44 100 50 100 141 Ethnicity

Native American 1 2.1 1 Asian or Pacific Islander 1 2.0 1 African-American 15 31.9 12 27.3 9 18.0 36 Hispanic 1 2.1 3 6.8 2 4.0 6 White (Non-Hispanic) 25 53.2 23 52.3 34 68.0 82 Multi-ethnic 5 10.6 6 13.6 4 8.0 15 Total 47 100* 44 100 50 100 141

*Total percentage has been adjusted by rounding.








Research Design

The study was a 2 (emotion type) X 2 (sex) X 3 (grade level) X 2 (audience figure) mixed model factorial design. Emotion was designated as a within-subjects variable with two types: anger and sadness. Between-subject variables were sex (i.e., male and female) and grade level (i.e., 5th, 8th, and 11 th). This design differs slightly from Zeman and Shipmads (1997) research design. In their study, audience figure was designated as a between-subject variable with three levels (i.e., mother, father, and best friend) "to reduce the possibility of carryover effects" (p. 919). Although their concern about carryover effects was understandable, their design resulted in very small sample sizes in each cell. In the current study, I decided to designate audience figure (i.e., best friend and teacher) as a within-subjects variable to increase the power of the research design.

Research Instrument

Eight short stories (vignettes) depicting the most frequently reported school situations concerning anger and sadness from the pilot study were created for the questionnaire. Four stories described situations that commonly elicited sadness, and four stories featured situations that commonly evoked anger in the pilot study adolescents. Each vignette was written in a third-person narrative and specified the emotion it was intended to evoke. Participants were asked to assume the role of the protagonist in the story and to answer the questions posed on the questionnaire. Each vignette contained two sets of eight questions; one set for each audience figure (i.e., a close friend or a wellliked teacher). In this study, contrary to previous research, the emotional provocateur in the vignette was not the person (either well-liked teacher or best friend) with whom the







52
adolescent had to decide whether to display or dissemble genuine emotion. The audience figure was designed to be a neutral observer in order to assess more accurately adolescents' emotion management decisions and to disentangle the decision making about emotional regulation from the eliciting situation or event. (See Appendix B for the research instrument, Appendix C for the consent forms, and Appendix D for the student assent scripts.)

Emotion expression and feelings of self-efficacy. The first question concerned the decision to express or dissemble their emotion. 1. Would you show how (sad or angry) you feel to your (teacher or best friend)? Participants were instructed to indicate their decision on a 4-point Likert-type scale (1 = definitely would not, 2 = probably would not, 3 = probably would, 4 = definitely would). This scale's wording was changed slightly to reflect the syntax of each question. The second question addressed the participants' emotional self-efficacy (i.e., their belief in their ability to regulate their own emotional response): 2. Do you think you could control showing your (angry or sad) feeling if you wanted to?

Questions 3 through 8 addressed the participants' outcome expectancies, that is, their perceptions of consequences for expressing or dissembling their anger or sadness. The questions and descriptions were adapted from Zeman and Shipman's (1997) research.
Outcome expectancies. The first category, interpersonal support, addressed the

participants' belief that expressing their emotion would be met with a positive, supportive response: 3. Do you think your (teacher or best friend) would be understanding of how (sad or angry) that makes you feel? The second category, negative interpersonal








response, concerned expectations of some sort of negative reaction to participants expressing their feeling: 4. Would your (teacher or best friend) make fun of you or tease you if you showed how (angry or sad) you really felt? The third category, instrumental consequences, referred to the participants' expectation that to express emotion would result in their receiving some sort of negative noninterpersonal consequence from the audience figure: 5. Would there be a negative response from your (teacher or best friend)would he or she punish you in some way-f you showed how (sad or angry) you felt? The fourth category, emotional reaction, reflected the participants' belief that expressing their emotion would result in the audience figure experiencing some negative emotional response: 6. Would showing your (angry or sad) feelings upset your (teacher or best friend) and make (him or her) feel bad? The fifth category, norm maintenance, was defined as adherence to some rule regarding regulation of emotion in specific situations:

7. Do you believe that you should show (sad or angry) feelings to your (teacher or best friend)? The last category, internalization, addressed the participant self-esteem and involved the anticipation that dissembling emotion would result in either a better or worse outcome than might occur if the participants' expressed their emotion to the audience figure: 8. How would you feel if you kept your (sad or angry) feelings inside and did not show them to your (teacher or best friend)? Procedure

Students who had parental permission to participate were assessed in small groups (10 to 20 same-grade students) in sessions of approximately 30 to 45 minutes depending on grade level. I provided directions orally and read aloud a sample vignette. The vignettes were counterbalanced by emotion type across all participants. For the friend








audience figure condition, students were told to think of their same-sex best or closest friend. For the teacher audience figure condition, participants were instructed to think of a particular teacher they liked. Participants were then asked to read the eight vignettes on their own and to answer the questions for each story. Scoring and Data Analyses

Repeated measures analyses of variance (ANOVAs) were conducted on the first two questions: emotion management decisions and self-efficacy regarding emotional regulation. The dependent variable was each students response to the question using the 4-point Likert-type scale summed across the four stories for each emotion type. Scores could range from 4 to 16 for sadness or anger. The within-subjects variables were emotion type (i.e., anger or sadness) and audience figure (i.e., best friend or teacher). The between-subjects variables were the participants sex and age or grade (i.e., 5th, 8th, and 11th).

The scores on the final six questions, the outcome expectancies, were treated as a unified construct and a repeated measures multivariate analysis of variance (MANOVA) using the multivariate criterion of Wilks lambda (A) was conducted on the total outcome expectancy score. The dependent variable was each participants total score on each of the six expectancies, summed by emotion type across all four vignettes. Scores could range from 4 to 16 for each emotion type. Univariate ANOVAs were conducted when significant multivariate effects emerged. Significant main effects were analyzed with ANOVAs, t-tests, and appropriate post hoc procedures. The familywise error rate was controlled with the Shaffer-Holm procedure and the Bonferroni correction for multiple comparisons. All data analyses were performed using SPSS version 10.0 for Windows.









Reliability of the Dissertation Instrument

An internal consistency estimate of reliability was computed for the dissertation instrument. The coefficient alphas (a) for each variable in the analysis ranged from .67 to .84. The overall alpha (a) coefficient for the dissertation instrument was .91. The estimates are listed in Table 5.

The internal consistencies of the scales are generally acceptable given that the alpha coefficient tends to underestimate reliability for questionnaires employing this vignette/scenario approach. Moreover, the questionnaire was administered only once, under different testing situations and location, to 11 different groups of students over a 13-week period. Therefore, two distinct sources of error that can affect the reliability coefficient must be considered. The first, individual differences in mood, level of fatigue, and attitude toward the test, has been shown to exert an influence on internal consistency measures. The second, in spite of attempts to standardize administration of the questionnaire, many variations and concessions had to be accommodated in each testing situation (Borg & Gall, 1989). The dissertation questionnaire was extremely long for fifth graders, and they quickly lost interest in selecting the most appropriate response. Eighthgrade students remained focused for a longer time span than did fifth-grade students, but the length of the questionnaire was a deterrent to their concentration, as well. Eleventhgrade students completed the survey in the least amount of time, but several of them resorted to a strategy of checking the same response for each question without regard to the nature of the vignette, audience figure, or outcome expectancy. Taken together, all these factors are likely to have influenced student responses and the subsequent accuracy of internal consistency estimates of reliability.








Table 5

Internal Consistency Estimates for the Dissertation Instrument


Scales Cronbach Number a of Items

Show Emotion .81 16
Anger to Friend .69 4 Anger to Teacher .73 4 Sadness to Friend .75 4 Sadness to Teacher .76 4 Self-Efficacy .93 16
Anger with Friend .77 4 Anger with Teacher .81 4 Sadness with Friend .76 4 Sadness with Teacher .81 4 Interpersonal Support .80 16
Anger with Friend .67 4 Anger with Teacher .73 4 Sadness with Friend .73 4 Sadness with Teacher .72 4 Negative Interpersonal Response .85 16
Anger with Friend .78 4 Anger with Teacher .72 4 Sadness with Friend .75 4 Sadness with Teacher .72 4 Instrumental Consequences .90 16
Anger with Friend .74 4 Anger with Teacher .67 4 Sadness with Friend .80 4 Sadness with Teacher .74 4








Table rntinued


Scales Cronbach Number o of Items

Emotional Reaction .91 16
Anger with Friend .67 4 Anger with Teacher .71 4 Sadness with Friend .78 4 Sadness with Teacher .82 4 Norm Maintenance .83 16
Anger with Friend .76 4 Anger with Teacher .74 4 Sadness with Friend .81 4 Sadness with Teacher .80 4 Internalization .92 16
Anger with Friend .78 4 Anger with Teacher .74 4 Sadness with Friend .83 4 Sadness with Teacher .84 4 Overall Internal Consistency .91 128














CHAPTER 4
RESULTS

The purpose of this research was to examine developmental changes in

adolescents' reasoning in emotion decisions to express or dissemble anger and sadness from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (1I th grade). I also examined adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures, a close friend and a teacher, in a school context. Zeman and Shipman's (1997) research explored the relationship between adolescents' decision making about expression of negative emotion (anger or sadness) and their outcome expectancies. The six outcome expectancies they investigated were reexamined in this study.

Expectancies, considered as a single construct, were investigated. The first expectancy, interpersonal support, comprised the adolescent's belief that expressing negative emotion would be met with a positive, supportive response. The second, negative interpersonal response, referred to an adolescent's expectation of being teased or made fun of for expressing anger or sadness. The third, instrumental consequences, entailed the student's belief that to express negative emotion in a school context might result in some form of punitive behavior from the audience figure (best friend or teacher). The fourth expectancy, emotional reaction, reflected the adolescent's belief that expressing negative emotion would upset the audience figure (best friend or teacher). The








fifth, norm maintenance, was defined as adherence to an internalized cultural rule regarding expression of a negative emotion in a social context. The sixth and last expectation, internalization, addressed the adolescent's resulting state of well-being and self-esteem if they chose not to display their negative emotion to a close friend or teacher. Zeman and Shipman (1997) found that adolescents' outcome expectancies varied as a function of emotion type (anger or sadness), audience figure (parent or peer), age, and gender. I extended their research by using school as the social context and a well-liked teacher as an audience figure.

This chapter presents the results of the analysis of the data obtained from the

questionnaire that asked students about situations at school that were intended to invoke negative emotions. The statistical procedures used in the data analyses were all performed using the SPSS Graduate Pack 10.0 for Windows.

Hypothesis 1-Emotion Management Decisions

The first research hypothesis proposed that adolescents' decisions to express or dissemble emotion would vary as a function of the type of audience figure (best friend or teacher), type of emotion (anger or sadness), age (grade), and sex of the participants. It predicted that with age, adolescents' decisions to regulate their emotion would interact with the sex of the participant, with emotion type, and with the type of audience figure. The descriptive data are presented in Table 6.

A repeated measures univariate analysis of variance (ANOVA) was conducted on the first question; the student's willingness to display or dissemble their emotion. The dependent variable was the participant's response summed across all four stories for each emotion type. Scores ranged from 4 to 16 for each emotion type. Within-group variables







60

were emotion type with two levels (anger or sadness) and audience figure with two levels (best friend or teacher). The between-group variables were sex (male or female) and grade with three levels (5th, 8th, or 11 th).

A significant main effect emerged for audience figure only, F(1, 129) = 57.86,

p = .000. There were no significant main effects for emotion type, sex, or grade level. An emotion by grade interaction emerged, but was nonsignificant when the Shaffer-Hohnlm procedure was applied to control the familywise error rate. Two other interaction were significant, however; audience figure and sex of the participant, F(l, 129) = 12.72, p = .001, and audience figure and grade, F(2, 129) = 6.93, p = .001. The repeated measures ANOVA results are presented in Table 7.

Audience Figure X Sex interaction. A follow up analysis of variance (ANOVA)

for simple main effects was conducted on the audience figure by sex interaction to further examine the significant interaction effects of the repeated measures ANOVA. The familywise error rate was controlled by the Shaffer-Holm procedure. The simple main effects ANOVA revealed a significant difference on audience type for both sexes; males, F(1, 56) = 7.46, p = .008, females, F(1, 77) = 69.45, p = .000. Female students reported that they would be more likely to display their negative emotion to a friend rather than to a teacher. Males also reported that they were more likely to show their anger or sadness to a friend than a teacher, but the difference was not as large. The descriptive statistics appear in Table 8. The simple main effects ANOVA results are listed in Table 9.









Table 6

Descriptive Statistics for Adolescent: Decision to Express Emotion


Group Sex Grade Mean SD N


Show Anger to a Friend Show Anger to a Teacher Show Sadness to a Friend Show Sadness to a Teacher


Male 5th
8th
1I th Female 5th
8th
11 th Male 5th
8th
11 th Female 5th
8th
llth
Male 5th
8th
11 th Female 5th
8th
11th Male 5th
8th
11th Female 5th
8th
11 th


10.40 11.62
12.04 11.84 11.54 12.68 10.60 10.15 10.29
10.00 9.04 10.28 10.55 10.15 11.46 11.84 12.32 12.76 11.10 9.08 9.67 10.76 8.64 9.52


3.32 2.93 2.24 3.42 2.06 2.51 3.36
2.48 2.79 3.06 2.85 3.13 3.25 3.78 1.89
3.09 2.26 3.06 3.18
3.04 2.14 2.86
2.42 3.66









Table 7

Results of Repeated Measures ANOVA on Adolescentg Decisions to Express Emotion



Type III Sum Mean Source of Squares df Squares F p


Between Subjects
Sex
Grade
Sex X Grade
Error
Within Subjects
Emotion
Emotion X Sex
Emotion X Grade
Emotion X Sex X Grade
Error (Emotion)
Audience
Audience X Sex
Audience X Grade
Audience X Sex X Grade
Error (Audience)
Emotion X Audience
Emotion X Audience X Sex
Emotion X Audience X
Grade
Emotion X Audience X Sex
X Grade
Error (Emotion X Audience)


14.84 50.62
2.52 2823.55


6.05 11.21 21.21 11.14 422.60 354.16 77.84 84.85
4.61 789.66 .28
3.35 6.97

5.01

216.15


1
2
2 129


1 1
2
2 129
1 1
2 2 129
1 1
2


14.84 25.31
1.26
21.89


6.05 11.21 10.60 5.57 3.28 354.16 77.84 42.43 2.31 6.12 0.28 3.35 3.48


2 2.51


.68 1.16
.06



1.85
3.42 3.24 1.70


57.86 12.72 6.93 .38


.17 2.00 2.08


.412 .318 .944



.176 .067 .043 .187


.000 .001 .001 .687


.683 .160 .129


1.50 .228


1.68


Note. a = .05.









Table 8

Descriptive Statistics for Audience by Sex Interaction on Decision to Express Emotion



Group Sex Mean N Show Emotion to Friend Male 11.11 57 Show Emotion to Teacher 10.20 57 Show Emotion to Friend Female 12.15 78 Show Emotion to Teacher 9.67 78


Table 9

Simple Main Effects ANOVA for Audience by Sex Interaction on Decision to Express Emotion



Type III Sum Mean
Source Sex of Squares df Squares F P Audience Male 23.26 1 23.26 7.46 .008 Error 174.61 56 3.12 Audience Female 240.01 1 240.01 69.45 .000 Error 266.11 77 3.46








Audience X Grade interaction. A follow up test for simple main effects was

conducted on the audience by grade interaction to further investigate the differences in decisions to express emotion to either a best friend or a teacher. The descriptive statistics for the simple main effects ANOVA are presented in Table 10. The Shaffer-Holm procedure was used to control the familywise error rate. The ANOVA revealed a significant difference on audience type for 8th-grade, F(1, 44) = 34.59, p = .000, and 11 lth-grade students, F(1, 48) = 47.94, p = .000. In contrast to 5th graders who reported that they would not distinguish between teacher and best friend in expressing emotion, 8th and 11 th graders reported that they would be less likely to display emotion to a teacher than to a best friend. The results of the ANOVA appear in Table 11. Mean scores revealed that with age, adolescents were inclined to express their emotion to a best friend, rather than to a well-liked teacher. Eighth-grade students, however, were least likely to express negative emotion to a teacher.









Table 10

Descriptive Statistics for Audience by Grade Interaction on Decision to Express Emotion



Group Grade Mean N Show Emotion to Friend 5th 11.23 45 Show Emotion to Teacher 10.59 45 Show Emotion to Friend 8th 11.60 41 Show Emotion to Teacher 9.09 41 Show Emotion to Friend 11th 12.24 49 Show Emotion to Teacher 9.94 49



Table 11

Results of ANOVA for Audience by Grade Interaction on Decision to Express Emotion


Type III Sum of Squares

9.34

154.91 129.38


149.62 130.30 130.45


Mean Squares

9.34


F

2.65


p .110


44 3.52

1 129.38 34.59 .000 40 3.74

1 130.30 47.94 .000 48 2.72


Note. a = .0 167.


Grade 5th


Source Audience


Error


Audience


Error


1lth


Audience


Error








Hypothesis 2-Self-Efficacy Beliefs

The second research hypothesis predicted that adolescents would report greater self-efficacy for regulating sadness than for regulating anger. To investigate this claim, a repeated measures univariate analysis of variance (ANOVA) was conducted on the second question (i.e., whether adolescents thought they could control showing their emotion to their best friend or to a teacher they liked). Again, each participant's response was summed across all four stories by emotion type. Scores ranged from 4 to 16 for each emotion. Within-group variables were emotion type and audience figure. The betweengroup variables were sex and grade. The Shaffer-Holm procedure was used to control the familywise error rate. The second research hypothesis was not supported by the data. There was no significant main effect for emotion type, F(1, 128) = 2.51, _p = .115. The descriptive statistics for the analysis are listed in Table 12. The univariate ANOVA results appear in Table 13.

Two other main effects were significant, however. The first was a difference in self-efficacy for grade (age) of the participant, E(2,128) = 4.39, p = .014. Post hoc pairwise comparisons were conducted by grade using the Bonferroni adjustment for multiple comparisons. T-tests revealed that only the I lth-grade students' scores differed significantly from the 5th-grade students' responses, t(128) = -2.81, p = .017. The other comparisons were not significant. The mean scores differed by age indicating that 11 th graders had stronger beliefs in their ability to control their negative emotions than did the 5th graders. Descriptive statistics appear in Table 14.









Table 12

Descriptive Statistics for Adolescents Self-Efficacy Beliefs


Group Sex Grade Mean SD N


Controlling Anger with a Friend Controlling Anger with a Teacher Controlling Sadness with a Friend


Controlling Sadness with a Teacher


Male 5th
8th
1lth
Female 5th
8th
11th Male 5th
8th
11th Female 5th
8th
1Ith Male 5th
8th
11 th Female 5th
8th
11th Male 5th
8th
11 th Female 5th
8th
1lth


11.20 13.23 13.13 11.40 12.48 12.56 11.40 12.69 13.38 12.16 12.56 13.00 11.10 13.46 12.99 11.32
11.44 12.08 10.90 13.15 13.25 11.56
12.15 12.68


3.11 2.89 1.96
2.99 2.64 3.24 3.22 2.84 2.04 2.93 2.89 3.08
3.51 1.90 1.99
3.22 2.59 3.27 3.35 2.73 2.36 2.45 2.82 3.26









Table 13

Results of Univariate ANOVA on Adolescentg Self-Efficacy Beliefs


Type III Sum Mean Source of Squares df Squares F p


Between Subjects
Sex
Grade
Sex X Grade
Error
Within Subjects
Emotion
Emotion X Sex
Emotion X Grade
Emotion X Sex X Grade
Error (Emotion)
Audience
Audience X Sex
Audience X Grade
Audience X Sex X Grade
Error (Audience)
Emotion X Audience
Emotion X Audience X Sex
Emotion X Audience X
Grade
Emotion X Audience X Sex
X Grade
Error (Emotion X Audience)


17.39 232.85 45.93 3394.72


8.54 6.50
.35
5.62 435.16 5.53 8.53 3.50 1.62
135.83 .02 .14
4.10

.33

179.49


17.39 116.43 22.97 26.52


8.54 6.50 .17 2.81 3.40 5.53 8.53 1.75 .81
1.06 .02 .14
2.05

.17


.66 4.39 .87



2.51 1.91 .05
.83


5.21 8.04 1.65
.76


.02 .10 1.46


.420 .014 .423



.115 .169 .951
.440


.024 .005
.196 .468


.911 .754 .236


.12 .889


1.40


Note. a = .05.








Table 14

Descriptive Statistics for Grade Main Effect on Adolescents' Self-Efficacy Beliefs


Grade Mean SD N 5th 11.31 3.01 45 8th 12.55 2.69 44 11th 12.66 2.93 50


The second main effect in the self-efficacy analysis occurred for audience figure, F(1, 128) = 5.21, p = .024. However, because the main effect was modified by an interaction between audience figure and sex, F(1, 128) = 8.04, p = .005, a follow-up oneway ANOVA to investigate simple main effects was conducted. Female students indicated they would be able to control their emotion more with a teacher than with a close friend, F(1, 76) = 17.48, p = .000. There was no significant difference in males' reported ability to control their emotion between audience figures, F(1, 56) = .02, p = .904. Males' mean scores for self-efficacy on control of negative emotion were virtually the same for both their best friend and a teacher. Descriptive statistics for the post hoc analysis are presented in Table 15. Results of the simple main effects ANOVA for the audience by sex interaction appear in Table 16.

Hypothesis 3-Outcome Expectancies

The third research hypothesis investigated the six outcome expectancies and

examined whether adolescents' scores would vary as function of emotion type, audience figure, grade (age), or sex of the participant. It predicted significant main effects in the









Table 15 Descriptive Statistics for the Audience by Sex Interaction on Adolescents Self-Efficacy Beliefs


Sex Male


Audience Friend Teacher Friend Teacher


Female


Mean 12.44 12.46 11.88

12.35


Table 16

Simple Main Effects ANOVA for Audience by Sex Interaction on Adolescentg SelfEfficacy Beliefs



Type III Sum Mean Source of Squares df Squares F p Audience Male .0087 1 .0087 .015 .904 Error 33.491 56 .598 Audience Female 8.416 1 8.416 17.48 .000 Error 36.584 76 .481









total outcome expectancy scores for type of emotion, audience figure, sex of the participant, and age. It further predicted significant interaction effects for audience figure and sex of the participant, audience figure and grade (age) of the participant, and sex and age of the participants.

The scores on the final six questions were treated as a unified construct and a

repeated measures multivariate analysis of variance (MANOVA) using the multivariate criterion of Wilks' lambda (A) was conducted on the outcome expectancy score for each expectancy. Univariate ANOVAs were conducted when significant multivariate effects emerged.

Results of the multivariate analysis revealed significant main effects on the

outcome expectancy score for grade (age), A = .83, [(6, 118) = 1.88, p = .037, emotion, A = .65, F(6, 118) = 10.72, p = .000, and audience figure, A = .41, [(6, 118) = 27.89, p = .000. There was no significant difference on the between-group variable of sex. These main effects were modified, however, by two-way interactions between emotion and audience figure, A = .46, F(6, 118) = 23.141, p = .000, audience figure and sex, A = .78, [(6, 118) = 5.529, p = .000; and audience figure and grade, A = .75, F(12, 238) = 3.018, p = .001. A three-way interaction between audience figure, sex, and grade, A = .81, [(12, 236) = 2.160, p = .014, was nonsignificant when the Shaffer-Hohnlm procedure was applied to control the familywise error rate. Because each main effect variable was involved in an interaction, only the two-way interaction effects were pursued in further data analyses. All follow-up univariate ANOVAs were conducted at a new alpha level of .008 (i.e., .05/6) using the Shaffer-Holm procedure. The MANOVA results are presented in Table 17.









Table 17

MANOVA of Total Outcome Expectancy Scores


Wilks' Hypothesis Error Source A F df df p


Between Subjects

Sex

Grade

Sex X Grade Within Subjects

Emotion

Emotion X Sex

Emotion X Grade

Emotion X Sex X Grade

Audience

Audience X Sex

Audience X Grade

Audience X Sex X Grade

Emotion X Audience

Emotion X Audience X Sex

Emotion X Audience X
Grade
Emotion X Audience X Sex
X Grade Note. a = .05.


2.16 1.89 1.67



10.72 1.62 1.58 .82 27.89 5.53 3.03

2.24 23.14 1.72 .62 .53


118 236 236



118 118 236 236 118 118 236 236 118 118 236


12 236


.052 .037 .076



.000 .149 .101 .621 .000

.000 .001 .014 .000 .122 .825

.892








Audience X Sex interaction. A univariate ANOVA was conducted on the

audience by sex interaction. The results were significant for the following expectancies: interpersonal support, F(1, 123) = 24.29, p = .000, norm maintenance, E(1, 123) = 17.82, p = .000, and internalization, F(1, 123) = 11.09, p = .001. The descriptive statistics appear in Table 18. The ANOVA results are presented in Table 19. Simple main effects tests were conducted on the significant interactions for each gender separately. The familywise error rate was controlled at a = .008 (i.e., .05/6) by using the Shaffer-Hohnlm procedure. The results revealed that females were more likely to expect an understanding supportive reaction from their friend than from a teacher, F(1, 123) = 34.75, p = 000; they believed they should show their emotion to a friend rather than to a teacher, E(1, 123) = 74.64, p = .000; and they indicated that they would feel worse if they did not express or display their emotion to a friend rather than to a teacher, F(1, 123) = 28.12, p = .000. There was no significant difference between audience figures for males on interpersonal support, F(1, 123) = 1.37, p = .244, norm maintenance, F(1, 123) = 2.85, p = .094, or internalization, F(1, 123) = .01, P = .940. These results support the hypothesis that there would be a significant audience by sex interaction effect on adolescents outcome expectancy scores.

Audience Figure X Grade interaction. A univariate ANOVA was conducted on the interaction effect between audience figure and grade. Only the results for the norm maintenance expectancy were significant, F_(2, 123) = 6.42, p = .002. The descriptive statistics appear in Table 20 and the ANOVA results are presented in Table 21.









Table 18

Descriptive Statistics for Audience by Sex Interaction on Outcome Expectancies


Expectancy


Interpersonal Support Norm Maintenance Internalization


Audience


Friend


Teacher


Friend


Teacher


Friend


Teacher


Sex


Male Female Male Female Male Female Male Female Male Female Male Female


Mean 11.43 13.25 11.77 11.72

10.87 11.80

10.37 9.67 8.76 8.00 8.78 9.03









Table 19

Summary Univariate ANOVA Table for Audience by Sex Interaction on Outcome Expectancies



Sum of Mean Expectancy Squares df Squares F P2


Interpersonal Support Error Neg. Interpersonal Response Error Instrumental Consequences Error Emotional Reaction Error Norm Maintenance Error Internalization Error


105.33 532.34

.17

366.31

3.04 270.54

3.62 298.46 79.85 551.22 30.85 342.06


105.33

4.33 .17 2.98

3.04 2.98 3.62

2.43 79.85

4.48 30.85

2.78


24.29 .000



.06 .812 1.38 .242 1.49 .224 17.82 .000 11.09 .001


Note. a = .008.








Table 20

Descriptive Statistics for Audience by Grade Interaction on Outcome Expectancies



Expectancy Audience Grade Mean N


Norm Maintenance Friend 5th 10.83 43 8th 11.31 38 11 th 11.87 48 Teacher 5th 10.47 43 8th 9.43 38 1 Ith 10.16 48


Simple main effects tests were conducted for each grade separately on the significant interaction for norm maintenance. The familywise error rate was controlled at a = .017 (.05/3) using the Shaffer-Holm procedure. Fifth-grade students scores did not differ significantly between the two audience figures, F(1, 123) = 1.19, p = .278. Eighth-grade and 11 th-grade students were more likely to respond that they should show their negative emotions to a friend but not to a teacher, F(1, 123) = 25.94, p = .000, and F(1, 123) =31.28, p = .000, respectively. With age, students were more likely to respond that they should show their negative emotions to a friend but not to a teacher. These findings support the research hypothesis that predicted a significant interaction between audience figure and grade on outcome expectancies.









Table 21

Univariate ANOVA Table for Audience by Grade Interaction on Outcome Expectancies



Sum of Mean Expectancy Squares df Squares F P


Interpersonal Support Error

Neg. Interpersonal Response Error

Instrumental Consequences Error

Emotional Reaction Error

Norm Maintenance Error

Internalization Error

Note. a = .008.


37.40 532.34 26.12 366.31

8.92 270.54

5.65 298.46 57.55 551.22

8.30 342.06


18.70 4.33 13.06 2.98 4.46 2.98 2.83 2.43 28.77 4.48 4.15 2.78


4.32 .015 4.39 .014


2.03


.136


1.17 .315 6.42 .002 1.49 .229


Emotion X Audience interaction. A univariate analysis of variance (ANOVA) was conducted to investigate the interaction between emotion and audience figure. The familywise error rate was controlled by the Shaffer-Holm procedure with a = .008 (i.e., .05/6). The results were significant for the following expectancies: negative interpersonal response, F(1, 123) = 66.48, p = 000, instrumental consequences, F(1, 123) = 11.54,









Table 22

Univariate ANOVA Table for Emotion by Audience Interaction on Outcome Expectancies



Sum of Mean
Expectancy Squares df Squares F p Interpersonal Support 3.27 1 3.27 2.40 .124 Error 167.66 123 1.36 Neg. Interpersonal Response 86.64 1 86.64 66.48 .000 Error 160.29 123 1.30 Instrumental Consequences 15.76 1 15.76 11.54 .001 Error 167.94 123 1.37 Emotional Reaction 12.72 1 12.72 10.70 .001 Error 146.27 123 1.19 Norm Maintenance 1.55 1 1.55 1.19 .277 Error 159.20 123 1.29 Internalization .734 1 .734 .96 .330 Error 94.37 123 .77 Note. a = .008


p =.001, and emotional reaction, F(1, 123) = 10.70, p = .001. There were no significant differences on the other three expectancies. A summary of the findings is presented in Table 22.








To further investigate differences due to the audience figure and emotion

interaction, simple main effects ANOVAs were conducted for each audience figure separately with a = .025. Descriptive statistics for the emotion by audience interaction appear in Table 23. The follow up ANOVA results for the interaction are presented in Table 24. The analyses revealed a significant difference between audience figures on the negative interpersonal response expectancy, F(1, 123) = 86.18, p = .000. Students anticipated more teasing and negative behaviors from their friend(s) than from their teachers for expressing their anger. There was no significant difference in student scores for the teacher audience figure, F(1, 123) = .02, p = .899, for expression of sadness.

Although students did not believe that friends would punish them for expression of either anger or sadness, F(1, 123) = .70, p = .658, the students were less certain about their teachers' responses, F(1, 123) = 28.45, p = .000. The students tended to respond that teachers definitely would not punish their expression of sadness, but teachers probably would not punish their expression of anger.

Finally, the analysis revealed a significant difference in students expectancy

regarding the audience figures emotional reaction to their expression of negative emotion. Students indicated that expressing their sadness, but not their anger, to a friend would probably upset the friend, F(1, 123) = 12.74, p = .000. There was no significant difference between emotion types on this expectancy for teachers, F(1, 123) = .60, p = .447. These results only partially support the hypothesis that adolescent/ expectancies would vary as a function of emotion type and audience figure. Expectancies varied for friends as a function of emotion type, but did not vary for teachers.









Table 23

DescriDtive Statistics for Emotion by Audience


Interaction on Outcome Expectancies


Expectancy

Neg. Interpersonal Response Instrumental Consequences Emotional Reaction


Audience Friend


Teacher


Friend


Teacher


Friend


Teacher


Emotion Anger Sadness Anger Sadness Anger Sadness Anger Sadness Anger Sadness Anger Sadness


Mean 8.42 6.75 6.73 6.75 6.36 6.29 7.07 6.28 7.71 8.24 8.12 8.00


N

129 129 129 129 129 129 129 129 129 129 129 129









Table 24

Results of ANOVA for Emotion by Audience Interaction on Outcome Expectancies



Sum of Mean Source Squares df Squares F p Neg. Interpersonal Response


Friend Error Teacher Error


197.97 259.03

.19

182.31


1

128

1

128


197.97 2.02 .19 1.42


86.18


.000


.02 .899


Instrumental Consequences

Friend Error

Teacher

Error

Emotional Reaction

Friend Error

Teacher

Error


.47 170.03 42.73 165.77



22.39 173.61

1.40 175.10


.47 1.33 42.73 1.30



22.39 1.36 1.40 1.37


.70 .658


28.45 12.74


.000 .000


.60 .442


Note. a = .008.









Summary of Results

The first research hypothesis proposed that adolescents' decisions to express or dissemble their anger or sadness would vary as a function of the type of audience figure, type of emotion, age (grade), and sex of the participant. The type of audience figure was the only variable for which a significant difference emerged. This main effect was modified by significant interactions between audience figure and sex, and audience figure and grade. There were no other significant interactions on the first question. Both male and female students reported that they would be more likely to express their negative emotions to a best friend rather than to a teacher. Female students, however, were much more likely than male students to report they would express their negative emotions to a friend rather than to a teacher. Males indicated they would be more likely to express their anger or sadness to a friend than to a teacher, but the difference between the two was minimal.

Fifth-grade students were almost equally inclined to report they would express negative emotion to either a best friend or a teacher they liked. Eighth-grade students were least likely to express negative emotion to a teacher. In fact, there was a noteworthy decrease between fifth- and eighth-graderg scores on displaying negative emotion to a teacher. Eleventh-grade students also indicated that they would be disinclined to display negative emotion to a teacher. The situation was reversed for the best-friend condition. With age, students indicated that they would be more likely to display their anger or sadness to a friend.

The second research question investigated self-efficacy and predicted that

adolescents would report greater self-efficacy for regulating sadness than anger. There









was no difference in self-efficacy between emotion types so the hypothesis was not supported. There was, however, an increase in reported self-efficacy with age. Eleventhgrade students were far more likely to report that they would be able to control the display of their negative emotions than were fifth-graders. Self-efficacy was also influenced by an interaction with the type of audience figure and sex. Female students indicated that they would be able to control their negative emotion more with a teacher than with a best friend. Male students did not differentiate between the two audience figures and indicated that their ability to regulate their anger or sadness was not necessarily dependent on whether they were in the presence of either a teacher or bestfriend in the school context.

The third hypothesis predicted that the outcome expectancy scores would vary as a function of emotion type, audience figure, age, and sex of the participant. Significant differences did result for emotion type, grade (age), and audience figure, but each of these variables interacted with another. There was no main effect between sexes on any of the expectancies. Audience by sex interactions emerged for interpersonal support, norm maintenance, and internalization. An audience by grade interaction occurred for norm maintenance. Last, an emotion by audience interaction was found for negative interpersonal response, instrumental consequences, and emotional reaction. There were no significant three-way interactions or the predicted age by sex interaction. Interpersonal Support

Females expected their best friend would be more supportive and understanding of the depth of their negative emotion than would a teacher. Males did not differ








significantly between audience figures for interpersonal support. There were no significant differences by grade for this expectation. Negative Interpersonal Response

Students expected more teasing and negative behaviors from their best friend for displaying their anger than they anticipated from their teacher. There was no significant difference between the two audience figures for expression of sadness. Instrumental Consequences

Students did not expect to receive negative consequences or punishment from

their teacher(s) for expressing their anger or sadness in the school context. Furthermore, there was no significant difference between emotion types on expressing emotion with a best friend. Students indicated that they would probably not receive a negative response for expressing either emotion to a friend. Emotional Reaction

Students indicated that revealing their sadness to a friend would probably make the friend feel bad. This was not true for revealing anger to a friend. Students believed that displaying anger would not upset the friend. There was no significant difference between emotion types for expressing negative emotion to a teacher. Students believed that displaying both anger and sadness would be likely to upset the teacher and cause an emotional reaction.

Norm Maintenance

Female students believed that they definitely should show their emotion to a best friend rather than the teacher. Males did not make that distinction between audience figures. Fifth-grade students reported that they should express their anger and sadness to







85
both a best friend and the teacher. Eighth graders were least likely of any other group to report they should display their negative emotion to a teacher. Eleventh-grade students indicated that they believed that they should reveal their negative emotion to a best friend, but not to a teacher.

Internalization

Female students believed they would feel worse if they did not express their

negative emotion to a friend. They also indicated that they would feel better if they did not display their emotion to a teacher. Male students expected to feel better if they did not reveal their emotion to either a best friend or a teacher.














CHAPTER 5
DISCUSSION

This study examined adolescents' decisions about displaying or dissembling negative emotion in the school context in the presence of two audience figures, a best friend and a well-liked teacher. The purpose of the research was to examine developmental changes in adolescents' reasoning in emotion decisions to dissemble or display negative emotion from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (1I th grade). I also examined developmental changes in adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures; a close friend and a teacher, in the school context.

This chapter presents a discussion and interpretation of the results of the study as they relate to the original research hypotheses. It also addresses the theoretical and practical implications of the findings. The chapter concludes with recommendations for future research and limitations of the study.

Adolescents' Emotion Management Decisions Hypothesis 1-Decisions to Express Emotion

The first research hypothesis predicted that adolescents' decision to express or dissemble negative emotion would vary as a function of the type of audience figure, the type of emotion, age, and sex of the participants. It was also expected that with age,







87

adolescents' decisions to regulate their negative emotion would interact with the sex of the participant, with emotion type, and with type of audience figure.

The only significant main effect that emerged was for audience figure. Male and female participants at all grade levels indicated that they would be more likely to express their negative emotions to a best friend rather than to a well-liked teacher. This finding, however, was moderated by an audience figure by grade interaction and an audience figure by sex interaction.

Follow-up analyses of the audience figure by grade interaction revealed that 5th graders were as likely to report that they would express negative emotion to a well-liked teacher as to a friend. In contrast, 8th graders and I Ith graders were less likely to report they would display negative emotion to a teacher than to a friend. This finding is consistent with other developmental research that has shown that with age, adolescents prefer close friends rather than adults as confidants with whom they emotionally selfdisclose (Belle & Longfellow, 1984; Berndt, 1982; Hartup, 1996; Papini et al., 1990). The research literature suggests alternative hypotheses that might account for this finding. The difference could be attributed to developmental changes in social cognition regarding appropriate expression of negative emotion (Saarni, 1979, 1988, 1989a, 1990, 1999), or it might indicate that the climate of middle schools discourages negative emotional display with adults who hold positions of authority (Eccles & Midgley, 1989; Eccles et al., 1996), or both of these factors may have influenced the developmental effect.

This finding is consistent with previous research on emotion management during adolescence (Gnepp & Hess, 1986; Fuchs & Thelen, 1988; Saarni, 1988; Zeman & Garber, 1996, Zeman & Shipman, 1997). Adolescents strive to present themselves in a








socially desirable manner for self-protective reasons. Furthermore, it supports Saarni's (1988, 1990, 1999) contention that with age, individuals master cultural scripts for the expression of emotion as appropriate or not suitable in particular social contexts.

Follow-up analyses of the audience figure by sex interaction revealed that females were more likely than males to report that they would express negative emotion to their best friend rather than to a well-liked teacher. This gender difference in emotion disclosure lends support to the literature on gender differences in adolescent friendships in general, but more particularly, the more intimate nature of adolescent female friendships (Berndt, 1982; Buhrmester & Furman, 1987, Hartup, 1993, 1996). The finding is also consistent with previous research on gender differences in emotional expression (Birnbaum & Croll, 1984; Brody, 1985, 1993; Brody & Hall, 1993, 2000; Fabes & Martin, 1991; Shields, 1995).

The failure to find an emotion by sex interaction was surprising in that researchers have generally reported that males are more likely to express anger rather than sadness. A trend in that direction was evident (p = .07), however, suggesting that with a larger sample this difference may have emerged.
Summary. An interesting developmental finding emerged from the analysis of the first hypothesis. Younger children were as likely to express negative emotion to a teacher as to a best friend, and older adolescents preferred to disclose their negative emotion to a best friend rather than to a teacher. In addition to the developmental findings, an interesting interaction between gender and audience figure was obtained; that is, females were more likely than males to disclose negative emotion to a friend.








Hypothesis 2-Self-efficacy Beliefs

The second research hypothesis predicted that adolescents would report greater self-efficacy for regulating sadness than for regulating anger; however, no difference for self-efficacy by emotion type emerged. Therefore, the hypothesis was not supported. This finding contrasts with Zeman and Shipman's (1997) research. In their study, adolescents indicated a belief in their ability to control expressions of sadness significantly more than expressions of anger.

A possible explanation for the lack of agreement might be the content of the vignettes in the current study. Even though the scenarios were developed from adolescents' self-reports of experiences in school that triggered anger and sadness, the situations described in the vignettes were commonplace occurrences that might not have triggered sufficient intense emotion. Intensity of emotion has been cited as a determining factor in children's and preadolescents' beliefs about their ability to control expression of negative emotion. Younger children, in particular, have reported that they would be unable to control their anger if they were sufficiently provoked and deemed the situation important enough to express their displeasure (Saarni, 1988, 1999).

Although significant differences in the ability to control sadness and anger did not emerge for hypothesis 2, students' beliefs in their ability to control their emotions did increase with age. Fifth graders were significantly less likely to report that they could control their negative emotion with a teacher or a best friend than were 11 th graders. Nonetheless, there was no difference between the 5th graders and 8th graders or between the 8th graders and I Ith graders on this question. Saarni (1990, 1999) has argued that, with age, emotionally competent individuals become more adept in their ability to








regulate emotional expression. The failure to find a significant difference between 5th and 8th graders and 8th and 11th graders suggests a need for further study of this developmental trend.

A significant interaction between audience figure and sex indicates that females perceive themselves as better able to regulate their emotion with a teacher than with a best friend, whereas males do not distinguish between a friend and a teacher. As stated previously, the development of intimate relationships with best friends is an integral feature of adolescence for most young females. These relationships are built on mutual trust, reciprocity, and the ability to share genuine emotions without negative repercussion (Hartup, 1996). Furthermore, both parties hold relatively equal power. Conversely, teacher-student relationships are unequal, unilateral, and characterized by an imbalance of power. Teachers have the ability to embarrass students in a very public social arena. Consequently, a female adolescent's belief in her ability to successfully regulate her genuine emotion with a teacher could become a valuable self-protection strategy. Another possible explanation for this finding might be found in the literature on adolescent females' development. Brown and Gilligan (1992) noted that some females tend to go underground in early adolescence. That is, they resort to silence and dissemblance of their true feelings so as not to stand out in the classroom.

Interpretation of this finding is made with caution due to the paucity of research on emotional self-efficacy. Saarni (1999), who has written extensively on the subject, contends that self-efficacy in emotion management occurs "when we emerge from an emotion-eliciting encounter with a sense of having accomplished what we set out to do" (p. 3). In other words, the individual has attained the requisite capacity and skills to effect







91
a successful emotional transaction. That female students in the current research reported that they could control their negative emotion more effectively with a teacher than with a best friend may have more to do with impression management than with self-efficacy.

A possible explanation for the lack of difference by emotion type on the selfefficacy question is that participants were not personally emotionally involved in the vignettes. Even though they were asked to imagine themselves as the protagonists in the stories, the situations were hypothetical. As Eisenberg and Zhou (2000) have recently argued, "emotional responses to actual events are likely to be much more compelling and meaningful than mild [inductions] that have little relevance to an individual's life" (p. 168). Consequently, participants' actual behavior in a situation in which a personal goal is at stake, may have had little bearing on their responses on the questionnaire. Perhaps students' self-efficacy responses would have differed by emotion type if the situations had been more personally meaningful for them.

SmmMy. Adolescents did not report greater self-efficacy for regulating sadness more than anger; therefore, the second hypothesis was not supported. There was, however, an increase in reported self-efficacy beliefs with age. Eleventh-grade students reported more confidence in their ability to regulate negative emotion than did 5th-grade students. In addition to the developmental finding, an interaction between gender and audience figures occurred; that is, females indicated they would be more likely to control their negative emotion with a teacher than with a best friend. Males did not differentiate between the two audience figures in their self-efficacy beliefs.








Hypothesis 3-Outcome Expectancies

The third research hypothesis predicted that adolescents' responses on the

outcome expectancy questions would vary as a function of emotion type, audience figure, age (grade), and sex of the participants. It further predicted interactions between audience figure and sex, audience figure and grade, and sex and age of the participants. Significant differences did emerge for emotion type, grade (age), and audience figure, but each of these variables interacted with another. There were no significant sex differences for hypothesis #3. However, significant interactions resulted between audience figure and grade, and audience figure and sex. There was no significant interaction between sex and age. The following discussion will focus on each of the outcome expectancies, individually.

Interpersonal support. Females anticipated a more supportive and understanding response to their emotion display from a best friend than from a well-liked teacher. This finding is consistent with research (Belle et al., 1987; Buhrmester & Furman, 1987) that has found that in times of distress, females turn to close friends for comfort and emotional support. It also supports the findings of Zeman and Shipman (1997) on sex differences in expectations of interpersonal support. Males, in contrast, did not differ significantly in their expectancy for support from either audience figure.

Surprisingly, younger adolescents did not differ significantly from late

adolescents in this expectancy. This finding is in contrast with the research of Fuchs and Thelen (1988). They found that older children (especially males) anticipated less positive consequences as a result of their negative emotional expression than did younger children.




Full Text

PAGE 1

EMOTION MANAGEMENT FROM LATE CHILDHOOD THROUGH ADOLESCENCE: THE SCHOOL AS SOCIAL CONTEXT 08Trj 1~OSI — — -By>* J SUSAN M. SCHILLINGER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULHLLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2002

PAGE 2

Copyright 2002 by Susan M. Schillinger

PAGE 3

, it., ,.. t For my mother

PAGE 4

, ACKNOWLEDGMENTS I want to express my most sincere gratitude and appreciation to the members of my doctoral committee, Dr. Patricia Ashton, my chairperson. Dr. Bridget Franks, my cochair. Dr. David Miller, and Dr. Arthur Newman. Their guidance, encouragement, expertise, and patience, have been invaluable sources of support in the completion of this dissertation. Second, I am grateful to the principals, teachers, and students of Alachua Elementary, Idylwild Elementary, Kanapaha Middle, Westwood Middle, Gainesville High School, and P. JL Yonge Developmental Research School who willingly volunteered to participate m the dissertation research and in the pilot study. Without their support, this study would not have come to fruition. I owe them all a debt of gratitude. Third, I wish to thank Mike for sharing his expertise in the complicated and difficult data analyses. In addition, I would be remiss if I did not express my appreciation to my employer and my work colleagues for their understandmg, their cooperation, and their flexibility during the final phase of this dissertation. Finally, I am indebted to my family and friends for standing by and supporting me during this endeavor. One individual, in particular, deserves special credit for having lived through this adventure with me. He nurtured me when I most needed it, encouraged me when times were rough, and always remained optunistic. I offer a heartfelt thank you to Gabriel for all his support. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGEMENTS iv ABSTRACT vii CHAPTERS 1 INTRODUCTION 1 Statement of the Problem 5 Purpose of the Study 6 Research Hypotheses 8 Significance of the Study 9 Delimitations and Limitations of the Study 12 2 REVIEW OF THE LITERATURE 13 Emotion Type 16 Age Differences in Emotion Regulation 21 Gender Differences in Emotion Regulation 28 Interpersonal Relationships and Outcome Expectancies 31 3 METHODOLOGY 40 Pilot Study 40 Dissertation Study 49 Participants 49 Research Design 51 Research Instrument 51 Procedure 53 Scoring and Data Analysis 54 Reliability of the Dissertation Instrument 55 V

PAGE 6

4 RESULTS 58 Hypothesis 1 — ^Emotion Management Decisions 59 Hypothesis 2 — Self-Eflficacy Beliefs 66 Hypothesis 3 — Outcome Expectancies 69 Summary of Results 82 5 DISCUSSION 86 Adolescents' Emotion Management Decisions 86 Theoretical Implications of the Research 1 00 Practical Implications of the Research 102 Recommendations for Future Research 104 Limitations of the Study 1 05 Summary 106 APPENDICES A PILOT STUDY QUESTIONNAIRE 107 B DISSERTATION QUESTIONNAIRE Ill C CONSENT FORMS 129 D CHILD ASSENT SCRIPTS 133 REFERENCES 135 BIOGRAPHICAL SKETCH 144 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EMOTION MANAGEMENT FROM LATE CHILDHOOD THROUGH ADOLESCENCE: THE SCHOOL AS SOCIAL CONTEXT By Susan M. Schillinger May 2002 Chairperson: Patricia T. Ashton Cochair: Bridget A. Franks Major Department: Educational Psychology The purpose of this study was to examine developmental changes in adolescents' reasoning in emotion decisions to dissemble or display anger and sadness in a school context. The study also explored adolescents' self-efificacy beliefs about emotion regulation and their outcome expectancies of emotional expression for themselves and two audience figures, a close friend and a well-liked teacher. One hundred forty-three students in grades 5, 8, and 1 1 participated in this research. Analyses of variance revealed two important developments in students' reports of their expression of negative emotion. First, 1 1th graders were more likely than 5th graders to report that they would display negative emotion to a best friend rather than to a teacher. Second, 1 1th graders had stronger efficacy beliefs in their ability to control their negative emotions than did 5th graders. In addition, a number of gender differences emerged in the analyses. Females, in contrast with males, indicated they would be able to control their negative emotion more vii

PAGE 8

with a teacher than with a best friend. Females anticipated more understanding and support from their best friend than from a teacher for expressing negative emotion. They also reported that they would feel worse if they did not express their emotion to a best friend. Conversely, males indicated no difference regarding decisions to express emotion with either audience figure. They also indicated they would feel better if they did not display their negative emotion to either audience figure. Participants in all grades anticipated more teasing and belittling from a friend than from a teacher for expressing anger. Students did not expect negative consequences from either a teacher or a friend for expressing negative emotion. They reported that expressing sadness, but not anger, would likely upset their friend and expressing either negative emotion at school would probably upset the teacher. These findings add support to previous research on (a) gender differences indicating the more intimate nature of adolescent female friendships compared to males' and (b) developmental research indicating that students' beliefs in their emotional selfeflficacy, and their tendency to confide in a best friend rather than an aduh, increase with age. viii

PAGE 9

CHAPTER 1 INTRODUCTION There has been increased public attention to the concepts of emotional intelligence (Goleman, 1995; Salovey & Sluyter, 1997) and emotional competence (Saami, 1990, 1997a, 1999) in recent years. Both concepts have implications for the effective management of emotional behavior and how children and adolescents learn to regulate their own emotions to serve specific goals. This public concern is especially salient given the escalating number of aggressive incidents in American society in general and the rise in aggressive behaviors and disciplinary infractions in the nation's schools. Some schools have responded by offering courses in "emotional literacy" that teach conflict resohition and anger management skills to children as early as kindergarten (Salovey & Sluyter, 1997). Other programs have been designed to mcrease selfawareness and emotional regulation skills in adolescents. The empirical research related to the theoretical underpinning of these programs and the evidence demonstrating the effectiveness of these programs remain limited despite the widespread adoption of such programs. Recently, developmental, educational, social psychologists, and others have begun to examine the roles of various social-contextual fectors and individual characteristics (i.e., age, sex, level of cognitive development, and temperament) as factors in children's and adolescents' emotion management decisions. Although the concepts of emotional intelligence and emotional competence imply a certain wisdom about what one should do in a given situation when &ced with a 1

PAGE 10

2 pleasant or distressing task or event, the notion of what is wise or prudent action may depend on a number of fectors, not the least of which are the situational, interpersonal, social, or cultural contexts that color one's decision making. These varying contextual influences, combined with the traditional belief that emotions are internal feeling states, difficult to substantiate with objective observation, have created a complex challenge for emotional development researchers who want to better understand the relationship between the external and internal variables that influence a person's decision to express or dissemble his or her emotions. Barrett and Campos (1987) described "a paradigm shift in theories of emotion" and explicated how "theories are moving away from simple stimulus-response orientations to orientations emphasizing personal meanings, organismic strivings, and the functional [emphasis added] importance of emotions" (p. 555). According to the functionalist approach, emotional regulation is a bidirectional interrelational process that includes an individual's appraisal of the significance of an event, the individual's feelings surrounding the significance of the event itself, and the manner in which the person decides to regulate or express emotion in a given environment. Barrett and Campos emphasized the adaptive function of emotions and partitioned the adaptive function of emotion into three types: the first "concerns behavior-regulatory fimctions; the second concerns internal-regulatory functions; and the third concerns social-regulatory functions" (p. 566). In essence, emotional regulation is not just the decision to regulate one's own internal feeling states and actions but extends to consideration of how others will respond to one's emotional expression and one's goals or outcome expectations. In brief, as Zeman and Garber (1996) noted, "emotions have important communicational

PAGE 11

3 and interpersonal regulatory consequences. The expression of emotion can affect the establishment and maintenance of interpersonal relationships, and, conversely, the social enviroiunent can influence whether individuals regulate or display their emotions" (p. 966). The School as a Social Context Between the ages of 5 and 1 8, children and adolescents spend considerable time in the company of their peers and teachers in a closed, formally structured, social environment. Successful school adjustment and subsequent behavioral adaptation are largely determined by a child's sensitivity to a number of implicit classroom and social norms. As children enter school, they must learn how to form and maintain interpersonal relationships with classmates as well as with teachers. These relationships are endowed with differing values, expectations, and rules of conduct (Birch & Ladd, 1996; Juvonen, 1996). CMdren must learn how to communicate effectively within each type of interpersonal relationship and understand the emotional and behavioral signals that contribute to successful relationship building. An important function of the elementary school is to instill prosocial behaviors in children and teach them the social conventions they will need to progress both socially and academically (Wentzel, 1996). By the time children have reached the later elementary school grades and enter middle school and adolescence, their skill in understanding emotional display rules and reading nonverbal behavioral signals from peers and teachers is thought to be relatively sophisticated (Gnepp & Hess, 1986; Saami, 1979, 1988, 1989a). By age 11 or 12, most children have internalized "cultural scripts" (Saami, 1989b, 1990, 1995; Wierzbicka, 1994) and understand how to regulate their own negative emotion in order to protect

PAGE 12

themselves or the feelings of another person. These scripts provide guidelines about the outcome expectancies of emotional e)q)ression and the hkely sequence of events to come in a given social context. Several recent studies (Fuchs & Thelen, 1988; Zeman & Garber, 1996; Zeman & Shipman, 1997) have explored the role of parents and peers as socialization agents m children's decisions to manage their negative emotions, in particular, anger and sadness. In general, these studies have found that children and adolescents express these two emotions differently depending on which audience figure (parent or peer) is present. In brief, it appears that the outcome expectancies that children and adolescents have about expressing negative emotion change as a function of the social context, the presence or absence of specific key socialization figures, as well as type of interpersonal relationship. All these fectors influence children's and adolescents' decisions to suppress or display what they are feeling. Although most of this research has been conducted in schools and has relied upon the use of self-disclosure questionnaires and structured interview formats, the social context of the school situation has not been the focus of investigation. Most of the vignettes or stories presented to the research participants concerned emotion eliciting situations that occurred outside the school setting; emotions that might be evoked in interactions in the home with a parent or in recreational settings with peers. The influence of teachers and other instructional or administrative personnel as socialization agents for emotional regulation in school settings has received surprisingly little attention in the emotional development Uterature. Little is known about the interpersonal relationship that develops between teacher and student and how it influences adolescents' emotion

PAGE 13

5 management decisions and outcome expectancies for expressing or dissembling anger and sadness in the school setting. Statement of the Problem A considerable amount of the emotional regulation research has focused on children's understanding of display rules for emotions and how they use that imderstanding to modify their own emotional expression. Several studies have specifically examined how children and adolescents report they would act in negative affect producing situations. Anger has been the most commonly studied negative emotion in infants, preschool children, elementary school-aged children, adolescents, and adults. A few studies have focused on children's or adolescents' sad emotions or on their attempts to regulate them (Garber, Braafladt, & Zeman, 1991; Glasberg & Aboud, 1982; Levine, 1995; Saami, 1992; Stapley & Haviland, 1989; Stein & Jewett, 1986; Stein & Levine, 1987). This research has examined coping behaviors and strategies employed by various age groups to deal with negative affect. More recently, researchers have begun to study negative affect from an interactional perspective and have noted differences in the ways that children and adolescents cope with their anger, sadness, or fear in interaction with specific socialization figures (i.e., mothers, fathers, and peers). Although it is generally agreed that parents and peers are the primary socializers of yoimg children's emotional expression, from the time children begin kindergarten, teachers and other school personnel also become socializers of emotional expression. Given the amount of time that children and adolescents spend in school with classmates, teachers, and other school supervisory authorities, it is surprising that so little research about emotional development and/or emotional regulation has been set in this venue. The

PAGE 14

6 education literature is replete with articles devoted to anger management programs at all levels of the curriculum, conflict resolution and peer mediation programs, and selfawareness programs that teach specific emotional regulation skills to children as early as preschool ages. As Zeman and Shipman (1997) have noted, however, few of these studies have investigated the developmental progression of emotion management through the adolescent years. The current study examined factors that may influence adolescents' decisions to control or display their emotions in one specific social context, the school These factors mclude type of audience figure (well-liked teacher and close friend), type of emotion (anger and sadness), age (grade level, 5th, 8th, and 1 1th), and sex (female and male). Purpose of the Study The purpose of this research was to replicate Zeman and Shipman's (1997) research with adolescents and extend it into a new social context. It examined developmental changes in adolescents' reasoning m emotion decisions to dissemble or display anger and sadness from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (11th grade). I examined adolescents' self-eflficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures, a close friend and a teacher, in a school context. Previous research (Cole, 1986; Fuchs & Thelen, 1988; Saami, 1979, 1988, 1993; Zeman & Garber, 1996; Zeman & Shipman, 1997) suggests that both the social context and the presence of an audience figure may influence children's and adolescents' reasoning about outcome expectancies or goals in interpersonal relationships. In their work with elementary school children, Saami (1979) and Zeman and Garber (1996) found that their

PAGE 15

7 explanations for expressing or masking genuine emotion served four purposes: (a) to maintain one's self-esteem, (b) to avoid punishment or fiirther trouble, (c) to maintain the smooth functioning of an interpersonal relationship, and (d) to comply with social convention norm maintenance or role constraints. All four of these purposes could ^ply to adolescents' emotion reasoning, as well. Zeman and Shipman's research was among the first to explore the relationship between adolescents' decision making about expression of negative emotion and their outcome expectancies. The outcome expectancies they investigated mcluded "interpersonal expectations (positive and negative), non interpersonal [expectations], instrumental expectations (positive and negative), norm maintenance (cultural rules regarding emotion management), and internalization of emotional experience to feel better" (p. 918). Zeman and Shipman found that outcome expectancies varied as a function of emotion type (anger or sadness), audience figure (parent or peer), age, and gender. The fimctionalist emotion regulation theorists have contended that adolescents' decisions to control or express their negative emotion becomes situationally specific with age (Thompson, 1994) and that adolescents' relationships with their peers in a public setting may alter their decisions to express or dissemble anger or sadness for selfpresentation purposes (Zeman & Shipman, 1997). In short, the nature of the situation is likely to affect their willingness to express vulnerability attributable to sadness or their anger at being slighted or frustrated in pursuit of a desired goal Because little research has examined the role of teachers and their influence on adolescents' emotional responding, I investigated whether the presence of a well-liked teacher would alter adolescents' decisions to control or dissemble emotion. A well-liked

PAGE 16

8 teacher was selected in an attempt to standardize the teacher audience figure and to guard against the degree of "liking-of-the-teacher" as a potential source of measurement error. If the fimctionalist perspective is accurate, m that emotional regulation is bidirectional and interrelational, then the presence of a well-liked teacher or a best friend should affect adolescents' emotional decision making. The teacher's role as authority figure in a school setting may prove to be a significant fector in this situation. Research Hypotheses On the basis of previous research (Davis, 1995; Fuchs & Thelen, 1988; Gnepp & Hess, 1986; Saami, 1979, 1984, 1988, 1989b, 1992, 1995, 1997b; Underwood, Coie, & Herbsman, 1992; Zeman & Garber, 1996; Zeman & Shipman, 1997) and the theoretical perspective of the fimctionalist approach to emotional regulation, the following general research hypotheses are proposed: 1 . Adolescents' decisions to express or dissemble emotion will vary as a fimction of the type of audience figure, type of emotion, age, and sex of the participants. It is expected that with age, adolescents' decisions to regulate emotion will interact with the sex of the participant, with emotion type, and with type of audience figure. 2. Adolescents will report greater self-efficacy for regulating sadness than for regulating anger. 3. Adolescents' total outcome expectancy score will vary as a fimction of emotion type, audience figure, age, and sex of the participant. Significant main effects in the total outcome expectancy scores are predicted for type of audience figure, sex of the participemt, and age; significant interaction effects are expected for audience figure and

PAGE 17

sex of the participant, audience figure and age of the participant, and sex and age of the participants. Significance of the Study The theoretical significance of this research is that it tests and extends the fiinctionalist contention (Barrett & Campos, 1987; Campos, Campos, & Barrett, 1989; Thompson, 1994) that emotional regulation is multifaceted, socially adaptive and interactive, situationally specific, and goal oriented. This study is important to our understanding of adolescent emotion management decisions because the findings will advance and extend our knowledge about emotional regulation in a previously unexamined social context (i.e., the school). It also mvestigates the influence of a significant but little examined socialization figure, a teacher. Finally, this research expands our understanding of adolescents' social cognition about impression management in the transition from preadolescence, 5th grade, to late adolescence, 1 1th grade. This research also augments our understanding about a significant practical problem educators in our middle and secondary schools face, that is, an increase in the mmiber of disciplinary infi^tions and the escalation of students' displays of negative emotioa As students progress from late elementary through middle school, the structure and climate of the schools change. These social contextual fectors may evoke negative emotion in early adolescents due to a mismatch between the developmental needs of the students, and of the organization and climate of the new schools (Eccles & Midgley, 1989; Eccles et al., 1993; Wenz-Gross, Siperstein, Untch, & Widaman, 1997).

PAGE 18

10 Recent research in school transitions in early adolescence indicates that students' relationships with teachers "deteriorate" as students move into middle schools and jimior high schools (Eccles, Lord, & Buchanan, 1996). This student-teacher relationship decline occurs at a time when yoimg people are in need of adults other than their parents who will be positive emotional role models and in whom they can confide (Wentzel, 1997). Eccles et al. (1996) suggested that early adolescents have two specific psychological needs during this transition to junior high school, an "increasing need for autonomy and participation in decisions regarding [their] experiences and the continuing need for strong social supports and close, trusting relationships with adults" (p. 276). A similar contextual change occurs with the transition from middle school to secondary school. High school students may become increasingly angry or sad and alienated by developmentally inappropriate or rigid teacher classroom rules and may drop out of school (Eccles et al., 1993). Finally, this study contributes to an on-going effort to expand our knowledge of the relationship between the social context, interpersonal relationships (i.e., teacher, close friend), and individual characteristics (i.e., age, sex) of the participants. It enhances our understanding of the relationship between outcome expectancies and adolescents' decisions to express or dissemble anger and sadness in school situations with peers and teachers. Zeman and Shipman's (1997) recent research provided some insight into the developmental transition that occurs in emotion management decision making from 5th grade to 1 1th grade. Their study, however, foiled to find significant age-related differences in emotional disclosure with peers. This finding is puzzling in light of the

PAGE 19

11 literature on the changing nature of peer relationships during adolescence. Several studies have shown that peer relationships deepen during this period and that girls, in particular, are likely to express their innermost thoughts and feelings with a best friend. Whether adolescent males and females would disclose their negative emotion to best friends and/or well-liked teachers in a school settmg was a question m the current study. Zeman and Shipman's (1997) research did not address emotional disclosure within a particular social context. Their interest concerned the interpersonal relationship between the adolescent and a specific audience figure. The social context in their investigation comprised the individual characteristics of the participants (age and sex), emotion type (anger or sadness), and outcome expectancies in interactions with a neutral audience figure (mother, fether, or best friend). Moreover, Zeman and Shipman (1997) adapted their vignettes from previous research on the display rule knowledge of young children. These vignettes may not have been developmentally appropriate for the participants in their study. The pilot study of the current research was designed to elicit information about situations that produce anger and sadness in a school or school-related context from adolescents. This information was collected, analyzed, and reworked into ecologically valid vignettes for the subsequent dissertation research. In sum, this study contributes to our theoretical understanding of the complex emotion management decision making that occurs in the transition from late childhood to late adolescence and extends oiuknowledge of emotional regulation in this age group. This information will have practical significance as well, because it addresses an educational and classroom management problem extant in middle and secondary schools.

PAGE 20

12 Delimitations of the Study This research was confined to adolescents' decisions to manage their negative emotions of sadness and anger, their sense of emotional self-efficacy concerning emotional regulation, and their outcome expectancies or goals for expressing or dissembling negative emotion. Although intensity of emotion has been implicated in some decisions to regulate anger, this study did not explicitly exanune that aspect of emotion regulation decision making in adolescence. This study also relied on the use of self-report instruments that are subject to impression management (i.e., the desire to present oneself in the best possible light) and other well documented limitations. Limitations and Qualifications of the Study The participants in this research may or may not have been representative of other preadolescent, early adolescent, and late adolescent individuals in other communities. Therefore, the results of this research should be interpreted with caution and not generalized to other populations in the same or similar age ranges. Similar caution is indicated with regard to assumptions about participants' gender-role behaviors. Stereotypical gender-role suppositions of the reader and of the participants may color both data analysis interpretation and/or participants' responses. All participants were assumed to be "normal functioning" students. They were not labeled "severely emotionally disturbed" or "emotionally handicapped." Special education students with emotional disorders were not selected to participate in this research.

PAGE 21

CHAPTER 2 REVIEW OF THE LITERATURE According to the fimctionalist perspective (Barrett & Campos, 1987; Campos, Mumme, Kermoian, & Campos, 1994), individuals' decisions to express or dissemble their emotions are embedded in an extensive sociocultural web of emotional beliefs, cultural display rules for emotional expression, and cultural expectancies about the outcomes of expressing or regulating emotional behavior. The individual cannot be separated from his or her sociocultural environment. In this theoretical framework, children come to learn the meaning of emotions in their culture by engaging in social transactions with significant others. Emotions are not seen simply as intrapersonal feeling states. Instead, they are considered the products of social collaboration. More specifically, "emotions are conceptualized as flexible, contextually bound, and goal directed" (Campos et al., 1994, p. 284). Functionalist theory emphasizes the adaptive quahties of emotion and the significance of the person-environment transaction. From this perspective, individuals will differ in the meaning they attach to different events and to the emotions that arise from interpersonal relationships. Emotional regulation involves adaptation to new person-environmental transactions. These adaptations often require the masking of genuine emotion in order to achieve one's goals or to avoid undesirable consequences. Campos et aL (1994) contended that "emotion regulation involves selecting responses acceptable to the social group to which one belongs because emotions take place in a social context" (p. 296). 13

PAGE 22

14 How do children learn how to regulate their own emotion and behavior within a given social context? Saami (1990, 1997a, 1999) and Wierzbicka (1994) have suggested that children first learn emotional regulation from their parents in the home environment by internalizing what they call "cultural scripts" for appropriate emotional display. Parents impart information about emotional experience and expression to their children by modeling the appropriate behavior and engaging in conversations with their children about feeling states and the meanings that are culturally attached to them. Saami (1990) referred to these sociocultural rules and beliefs as "naive theories of emotion" and maintained that these theories are dependent on an individual's level of cognitive complexity and emotional development, "unique family influences," and the social transactions that elicit emotions. She argued: "How we make sense of our feelings reveals the impact of socialization, whether it stems from our femilies, our peers, social institutions such as schools, the media, or our culture in general" (p. 130). In a similar vein, Gordon (1989) stated that children actively construct their emotional lives and once they understand the meaning of a particular emotion, "children become able to act toward it-magnifying, suppressing, or simulating it in themselves, and evoking or avoiding it in other people" (p. 324). Reichenbach and Masters (1983) have hypothesized that expression of negative emotions may be more subject to socialization efforts than expressions of positive emotions. As previously mentioned, the expression or display of negative emotion, particularly anger, is disruptive to social functioning. It can threaten interpersonal relationships, it can increase aggressive or competitive behavior, and, in general, in Western cultures, its expression is socially disapproved (Steams & Steams, 1986). There is increasing evidence that anger and sadness are socialized differentially as a function of

PAGE 23

sex, gender-role expectation, socioeconomic class, cultural norms, and historical context (see Brody, 1985, 1993; Brody & HaU, 1993 for reviews; Miller & Sperry, 1987; Steams & Steams, 1986). With infants, parents have been found to discourage expression of anger and contempt in females, whereas parents engage in more "emotional-matching" behaviors with male babies. This differential socialization might then create "(a) a mistmst of expressions that indicate such states and (b) an assumption that many [situational] contexts that might induce certain states actually will not because the person has 'risen above them'" (Reichenbach & Masters, 1983, p. 1000). Although Thompson (1994) has argued that a lack of consensus exists about what constitutes emotional regulation, most functionalist theorists believe that emotional regulation is a two-fold process. It can pertain exclusively to the self-management of an individual's internal feeling states or it can be constmed as the management of emotional reactions of others. Ultimately, Thompson noted that "emotion regulation consists of the extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions, especially their intensive and temporal features, to accomplish one's goals" (pp. 27-28). This definition enphasizes the functionalist concem with a person's goals and his or her desired outcomes in intrapsychic self-management, interpersonal relationships, or in person-environmental transactions. Emotion Type In this study I investigated the negative emotions of anger and sadness of 5th-, 8th-, and 1 Ith-grade adolescents. Both emotions involve internal feeling states, a set of physiological reactions, a cognitive component, and a set of culturally prescribed rules for overt expression or dissemblance. This section will focus on the causes of anger and sadness and the distinctions between these two emotions. I will describe how the

PAGE 24

16 developmental changes in cognition during late childhood and adolescence contribute to decision-making about emotional regulation. Of the two emotions, anger has received the most research attention because its public display is usually disruptive to the self and to the harmony of an individual's social environment. According to Averill (1982), "more than most emotions, anger is often condemned as antisocial," and "angry outbursts [of children] are usually pimished by parents, teachers, and others in authority" (p. 31). Subjectively, anger is typically regarded as an unpleasant negative state of arousal that occurs as the result of some demeaning or offensive act against the self or against "me and mine" (Lazarus, 1991). The offended individual has to decide whether to remain in the unpleasant state or engage in some activity to ameliorate the negative emotion and return to a neutral or positive emotional state. This decision involves taking into consideration his or her personal desires in conjunction with environmental factors that may interfere with his or her wishes. Anger may also occur as the outcome of the loss of a valued item, person, or goal Various researchers (Averill, 1982, 1983; Berkowitz, 1990; Izard, 1991; Whitesell & Harter, 1996; Whitesell, Robinson, & Harter, 1993) have identified precipitating actions or events that may lead to anger in children, adolescents, and adults. These include possible or actual physical injury or pain; possible or actual property damage; frustration or the interruption of some ongoing planned activity; events, actions or attitudes that result in a loss of personal pride, self-esteem, or sense of personal worth; violations of socially accepted rules of conduct, and violations of expectations and wishes important to the individual. It is generally acknowledged that for a situation to provoke anger, the offending act must be cognitively appraised as being intentional versus accidental. If the

PAGE 25

17 act is deemed accidental, then it must be appraised by an individual as having been preventable. Some theorists (e.g., Averill, 1982; Berkowitz, 1990; Izard, 1991; Lazarus, 1991; Stein & Levine, 1987) have noted that interference with or threats to an individual's goal attainment can be instrumental in causing anger, as well. In brief, there is agreement that anger usually results from intentional threats to an individual's well-being, from an interpretation of those threats as personally meaningful and physiologically arousing, and from interference in an individual's pursuit of a desired goal or emotional end-state. In adolescence, anger often results from an offense against the self. Whitesell and her colleagues (1993) studied anger-provoking situations for yoimg adolescents (sbcth, seventh, and eighth grade) and found that the prototypical causes of anger in this age group "included real or threatened physical or psychological harm inflicted on the self by another, typically involving a violation that [was] judged unfair" (p. 522). Examples of physical harm included being hit, kicked, shoved, or beaten up. Psychological harm resulted from peers or friends telling rumors about one's self, talking behind one's back, teasing, put-downs, and not acting as a friend is expected to act (i.e., being loyal, trustworthy, and keeping secrets). As children progressed from middle-childhood to early adolescence, incidents that produced anger were much more likely to resuh from psychological harm than from physical or verbal assault. In feet, having rumors told about oneself produced as much anger in this age group as did physical assauft. As Whitesell et al. (1993) noted, the shift from physical causes of anger to psychological causes coincides with "a qualitative shift in the nature and importance of peer relationships, particularly with regard to issues involving mutuality, confidentiality, loyalty, and trust" (p. 523). This finding has impUcations for the current research study in

PAGE 26

18 that one of the outcome expectancies under investigation concerned the anticipation of understanding and personal support for adolescents' emotional disclosures. It was expected that participants' outcome expectations in the current study would vary as a function of age, gender, audience figure, and emotion type. Sadness, in contrast to anger, has received limited research attention. Some studies have focused on children's imderstanding of the causes of sadness (Glasberg & Aboud, 1982; Levine, 1995; Stein & Levine, 1987; Rotenberg, Mars, & Crick, 1987-88), but Uttle research has examined "normal" sad affect in adolescents. It is important to differentiate normal sad affect from medically diagnosed depression. Sadness is said to be caused by separation (either physically or psychologically) from loved ones, death (especially the loss of an important family member or friend), disappointment, loneliness, and, similar to anger, the &ilure to achieve specific valued goals (Izard, 1991). The sadness is temporary, not as long-lasting or as debilitating as clinical depression. Loss and the inability to reinstate a desired goal are the most frequently noted causes of sadness in children and adults (Lazarus, 1991; Stein & Levine, 1987). In one relevant developmental study, Rotenberg et al. (1987-88) investigated the causes, intensity, motives, and consequences of sadness in first-, third-, fifth-, and seventh-grade children. They foimd that the majority of their respondents believed that sadness was caused by harm. Younger children indicated that harm to themselves would produce the most sadness, whereas the older children (grades 5 and 7) disclosed that harm to others would cause more sadness. Rotenberg et al. noted that the nature of harm invoking sadness also changed with age and perspective taking ability. First and third graders reported physical (concrete) sources of harm as sadness producing, whereas seventh graders cited instances of psychological harm as the cause of their sadness.

PAGE 27

19 These findings are consistent with research on cognitive development during early adolescence and with EDcind and Bowen's (1979) hypotheses of "the imaginary audience" and the "personal feble" that characterize some adolescent behavior. Many early adolescents become egocentric and preoccupied with what others will think about them For example, in an investigation into the likelihood of emotional expression and its relationship to the consequences of commimicating anger and sadness, Fuchs and Thelen (1988) found that sixth-grade boys were especially reticent to express their sadness to their parents; they ejqiected negative interpersonal outcomes for doing so, especially from their fathers. These early adolescent males were concerned about preserving a positive masculine self-image and conforming to their parents' and society's expectations for appropriate masculine behavior. Rotenberg et aL (1987-88) also recorded the emergence of one other cause of sadness. They found that prevention of goal achievement or attainment also caused sadness, but that this cause appeared to decrease with age. The seventh graders in their study were the least likely to cite prevention of goal attainment as a cause of sadness. Rotenberg and his colleagues suggested that with increased physical and social abilities, older children were less likely to experience prevention of goal achievement. This finding may have developmental implications for the current research in terms of emotional selfefificacy and adolescents' belief in their ability to regulate their own anger or sadness. Emotional self-efficacy may vary as a fimction of emotion type. It is also possible that, with age and increased cognitive development, older adolescents may use more strategic emotional regulation strategies that vary as a fimction of social and interpersonal context. Some researchers (Averill, 1982, 1983; Berkowitz, 1991; Lazarus, 1991; Levine, 1995; Stein & Jewett, 1986; Stein & Levine, 1987) have argued that the same situation or

PAGE 28

20 event may trigger anger or sadness depending on the individual's perception of the event. In other words, negative emotional events or situations contain certain salient dimensions such as intentionality, blame-worthiness, personal goals, desirability of the end-state (or goal), and the permanence or reversibility of the loss. Anger typically occurs if the individual concentrates on the cause of the problem and the possibility of goal reinstatement; sadness usually results if the iodividual fixates on the consequences of the ofFendmg act and the impossibility of goal reinstatement (Stem & Levine, 1987). Stein and Levine (1987) further differentiated between anger and sadness by characterizmg them on an active-passive dimension. Of the two affects, anger is considered to be the more action-oriented and outer-directed emotioiL It carries with it a necessity to rid one's self of the source of the problem. Sadness, on the other hand, is thought to be a more passive feeling state that is inner-directed and requires no formal action toward its removal. These findings also have implications for the current study in that they reflect the multifaceted and complex mteraction of emotion type with emotional experience and understanding, cognitive development (i.e., the ability to appraise an event as evoking a specific emotion in one's self or in others), gender, and age differences. Likewise, Zeman and Shipman's (1997) investigation of the influence of social context and outcome expectancies on adolescents' decisions to manage their sadness and anger adds two new factors to consider in attempts to understand inconsistencies in adolescents' emotional regulation. Age Differences in Emotion Regulation Several studies have focused on the age at which children can distmguish between flie fecial expressions for anger and sadness and their situational determinants

PAGE 29

21 (Barden, Zelko, Duncan, & Masters, 1980; Denham & Couchoud, 1990; Reichenbach & Masters, 1983). It is now generally believed that by 6 years of age children show consistency in identifying specific contextual and expressive clues for anger and sadness. Furthermore, it has been suggested that children as young as 4 years old are capable of understanding that "the 'appearance' of an emotional expression on the face does not necessarily have to correspond to the 'reality' of the internally felt emotional state" (Harris, Donnelly, Guz, & PittWatson, 1986, as cited in Saami, 1988, p. 276). Although some 4-year-olds may be able to make this appearance-reality distinction, by age 6 or 7 most children have mastered this skill. At this age they better imderstand that a person's facial expression may differ from the emotion being experienced and that, at times, it is necessary to dissemble genuine emotion (Davis, 1995; Saami, 1979, 1988). Harris and Saami (1989) have suggested that, with age, children become more adept at reading a situation and adjusting their emotional displays to their social context. In other words, even though children recognize that a negative emotion is occurring in themselves or others, they exert control over the extemal expression of anger or sadness, as the social situation requires, by using socially or culturally prescribed display rules for the expression of emotion. From Harris and Saami' s perspective, control of overt emotional displays may operate on two levels; an introspective self-reflective level (Le., the child's thoughts and understanding about negative emotion) that mediates between a reflexive response to a negative emotional trigger and another level of social conventions or influences inherent in the particular situation. Both the child's cognitive maturity and the ability to self-reflect on the antecedents of negative emotion and consequences of behavior influence emotional decision making. It is generally believed that by the time children reach preadolescence, around 1 1 years of age, these cultural scripts for display

PAGE 30

22 rule usage are well established and well practiced. Preadolescents have the cognitive perspective taking and reflective abilities to think about their eniotional experiences and, they have internalized the socialization efforts of parents, peers, and significant others. They are aware that their responses to anger or sadness eliciting events will affect other people. With increasing age, teenagers become more proficient at masking genuine emotion and substituting a more acceptable or appropriate affect in order to attain personal goals. Their efforts to conceal or control emotional displays become more strategic, and they have a larger repertoire of tactics they can employ to manage negative emotions. Adolescents can reflect on their own cycles of emotion and accon^anying behavior in terms of patterns. They understand that one emotion can trigger a second and that both influence their own and others' cognition and behavior (Saami, 1990, 1992, 1995). Several studies (Cole, 1986; Davis, 1995; Fuchs & Theien, 1988, Gnepp & Hess, 1986; Saami, 1979, 1984, 1988, 1989b, 1992, 1997b; Underwood et al, 1992; Zeman & Garber, 1996; Zeman & Shipman, 1996, 1997, 1998) have examined age differences in emotional regulation and children's understanding about how and when to use culturally acquired display rules to dissemble emotion. The majority of this research has been conducted with children ages 6 through 1 1 using the disappointing gift paradigm developed by Saami (1984). This method assesses children's understanding of the social convention that it is better to mask one's disappointment when receiving an inappropriate or "baby" present than to hurt the feehngs of flie gift giver. Few studies have investigated adolescents' emotional display mle knowledge, or ability, or motivation to regulate their social or interpersonal relationships. It is presumed that this age group is very

PAGE 31

23 knowledgeable about the appropriateness of masking, substituting, minimizing, or maximizing their expression of negative emotion. However, little research has been conducted on adolescents' ability or perhaps, of more importance, their motivation to dissemble negative emotion. Gnepp and Hess (1986) examined 1st-, 3rd-, 5th-, and lOth-grade children's imderstandmg of the use of verbal and facial display rules. They found that children as early as first grade are aware that use of a cultural display rule can have prosocial or selfprotective functions. Furthermore, children's knowledge about when, where, and with whom to express or mask genuine emotions increases across the grade school years. Gnepp and Hess noted, however, that the 5th graders' responses did not differ significantly fi-om the 10th graders. It seems noteworthy that teenagers, who have the emotional knowledge and the ability to be proficient at using cultural display rules, "predicted that the story characters would regulate their facial expressions less than half of the time" (p. 106). The authors speculated that the lack of a significant difference between 5th and 10th graders' predictions of story characters' use official display rules may have been due to (a) the lOth-graders behef that their peers "either cannot or will not regulate their facial displays of emotion, even if ideally they should" (p. 108) (i.e., an ability versus motivational component) or (b) the simpUstic nature of the protocol, that could have produced the lack of difference between groups. Gnepp and Hess chose to Umit the number of facial responses from which their participants could select. They suggested that the 10th graders may have wanted to display more "subtle" facial expressions than were available within their coding system. Either one of these hypotheses might have been explored, yet there has been little follow-up research to

PAGE 32

24 investigate this similarity between 5thand lOth-grade students' reported use of fecial display rules. Saami's (1988) research in this area is instructive. In a recent study that investigated children's social cognition about the interpersonal consequences of "presenting an emotional front" to adults or peers, she interviewed children from three age groups, 6-7, 10-11, and 13-14. The children were shown a series of emotion ehciting vignettes and were asked how others "are likely to react" when genuine emotion is expressed or when someone tries to mask his or her true feelings. These children were also asked to specify which audience figure (adult or peer) they would prefer as the recipient of the expression of genuine emotion. Moreover, they were asked what they thought the expected outcomes would be for a peer who either constantly dissembled or expressed genuine emotions. Finally, these children were asked how they arrived at a balance in making their emotional decisions. The last question, accordii^ to Saami "required children to take into account both unique self-knowledge as well as situational knowledge" (p. 277). In general, with age, children thought that they would be successfiil in "fooling" the various vignette characters. Older girls preferred to express genuine emotions to peers rather than to adults. Older boys had more cynical reactions regarding whether they could successfully fool a "bully" (over 50% of the 13-14 males thought that a bully would see through a "felse front"). Older boys also thought that they would not be able to fool the school principal after participating in a school prank. Taken together, these studies raise important questions that need to be fiirther examined regarding adolescents' decision making about maskmg or expressing negative emotions. It is not clear whether ability, motivation, or outcome expectancy is most influential in such decision making.

PAGE 33

25 With a few exceptions, as noted above, most of the research in the understanding of display rule usage and decisions about whether to display genuine negative affect has centered on the social cognitive development of &stthrough sixth-grade children. Two recent studies (Papini, Farmer, Clark, Micka, & Bamett, 1990; Zeman & Shipman, 1997) specifically focused on adolescents' decisions to disclose their emotions to parents and peers. Papini et aL focused on 1 74 junior high school students between the ages of 12 and 15. They found that the females in their study engaged in more emotional self-disclosure to parents and peers than did the males. Furthermore, with age, adolescents chose to confide their feelings to peers rather than to parents. Younger adolescents were more likely to disclose to parents than to peers. Because the focus of their study was neither the use of cultural display rules, nor, more specifically, the negative emotions of anger and sadness, the implications of their findings for the proposed research are somewhat limited. Papini et al. were more interested in the quality of the parent-child interaction: the openness of parent-child communication, &mily cohesiveness, and the aflfective quality of family fimctioning. Adolescents who perceived their parents to be receptive to their emotional disclosures were more inclined to disclose. Adolescents' emotional selfdisclosures to friends were associated more strongly with their own psychosocial characteristics (i.e., self-esteem and identity status) than with the quality of family fimctioning. Papini et al, suggested that their fmdings lend support to the social distancing hypothesis of Steinberg (1989), which predicts that, with age and pubertal maturation, adolescents distance themselves from their parents and associate with other adolescents who are experiencing similar physical changes and emotions.

PAGE 34

26 More recently, Zeman and Shipman (1997) examined developmental changes in emotion management decisions across a broader range of adolescents; preadolescents (5th graders), early adolescents (8th graders), and late adolescents (1 1th graders). Their goals were to examine adolescents' decision making about expression of genuine versus dissembled emotion, adolescents' emotional self-efficacy beliefs in their ability to control their own anger and sadness, and the outcome expectancies these participants had for three distinct audience figures (i.e., mother, father, and best fi-iend). The combination of audience figure and the age and gender of the adolescent comprised the social context for expression of negative emotion. The research question wasWould adolescent emotion management decision making vary as a fimction of the social context? In general, with age, their participants reported regulating sadness more than anger. Pre-, early, and older adolescents mdicated that they felt more emotional selfefficacy for the regulation of sadness than for anger. There appeared to be general agreement in all three age groups that sadness should not be openly expressed to parents or peers. In contrast, adolescents in all age groups indicated that they would feel worse if they did not let their anger out. The research literature has demonstrated a link between children's behefs in their ability to control a negative emotion and its intensity. In general, the higher the intensity of a negative affective state, the less children perceive it to be within their power to control. Consistent with one of their hypotheses, Zeman and Shipman (1997) found that the presence of the audience figure did influence adolescents' decisions to display or dissemble their anger or sadness. Eighth-grade adolescents were significantly more Ukely to report regulating their negative emotions with their mothers than with their fathers or peers; they expected to receive the least interpersonal support fi-om mothers than did

PAGE 35

27 either the younger or older groups. Eighth graders also expressed concern that their negative emotion expression might produce a reciprocal reaction from their mothers and best friends but not from their fathers. This finding lends fiirther support for Steinberg's (1989) social distancing hypothesis. It also supports other research into developmental changes that occur in parent-child relationships during adolescence; decreased cohesion (Papini et al., 1990; Steinberg, 1988, 1989, 1999) increased conflict between early adolescents and parents (Collins & Russell, 1991), and less emotional self-disclosure between parent and child (Buhrmester & Furman, 1987). With peers, Zeman and Shipman's (1997) findings were somewhat puzzling and inconsistent with findings in previous research. No significant differences were foimd between 5th-, 8th-, or 1 Ith-graders' emotion management decisions or in the consequences they expected for expressing genuine negative emotions with best friends. Previous research in peer relationships has indicated that there is a consistent increase of emotional disclosure between close friends (Belle, Burr, & Cooney, 1987; Belle & Longfellow, 1984; Berndt, 1982, 1996; Youniss & SmoUar, 1990) beginning in late childhood and early adolescence. Zeman and Shipman theorized that their research protocol may have produced only mild negative emotion that would have been insufficient to demonstrate the age-related changes or concern over the consequences of its expression with peers. They speculated that the nature of their vignettes might not have produced adequately intense anger or sadness in their participants because the story characters were not the emotion provocateurs (i.e., triggers of the emotion experience). The preceding explanation has significance for the current research study. Zeman and Shipman (1997) adapted their vignettes from previous research into the display rule knowledge of young children. These vignettes may not have been developmentally

PAGE 36

28 appropriate for the participants in their study. The pilot study of the current research was designed to elicit information about situations that produce anger and sadness in a school or school-related context from adolescents. This information was gathered, analyzed, and reworked into ecologically valid vignettes for the dissertation study. To sum up, with age, children become more adept at assessing the situational determinants that influence their emotional regulation decisions. Cognitive maturity (perspective-taking ability), emotion type (anger or sadness), emotional experience, the presence of peers, parents, or unrelated adults, all affect children's outcome expectancies and enter into their decision making processes with respect to genuine emotional expression or dissimulatioa Gender Differences in Emotion Regulation Zeman and Shipman (1997) found some gender differences in decisions to express negative emotion as a function of emotion type. Males reported regulating more negative emotion than females and believed it was more important to dissemble their sadness than their anger. Overall, males expected "less understanding and more belittling" in response to emotional disclosure. They believed that they should not express their emotions, more so than did females, and males thought they would feel better if they kept their emotions to themselves. This fmding is consistent with the literature on gender differences in emotional expression and with the social support literature. In times of stress, females are more likely to seek interpersonal support and engage in emotional self-disclosure with close friends. Males tend to disclose emotionally when they believe society expects such behavior (Belle et al., 1987). Other researchers studying emotional expression have found that emotional regulation decisions are influenced by gender and gender-role socialization (Birnbaum &

PAGE 37

Croll, 1984; also see Brody, 1985; Brody & Hall, 1993, Deaux & Major, 1987; Shields, 1995, for reviews). Females are usually more emotionally expressive overall than males, but males express anger more overtly and intensely than females do (Diener, Sandvik, & Larsen, 1985; Fujita, Diener, & Sandvik, 1991; Sonnemans & Frijda, 1995). Gender differences in anger expression may also interact with socioeconomic status (Miller & Sperry, 1987), individual characteristics, and the interpersonal context of the emotional exchange (Fabes & Martin, 1991; Fabes, Eisenberg, Smith, & Murphy, 1996; Whitesell & Harter, 1996; WhiteseU et al., 1993). In general, there is some evidence that females both experience and express more happiness, sadness, and fear than males do, whereas males experience and express more anger than females (Brody, Lovas, & Hay, 1995). There is further evidence that the sociaUzation of anger differs by sex; males are encouraged to display their anger when they are provoked (i.e., to use external coping strategies and physical or verbal responses to provocation), whereas females' overt expression of anger is widely discouraged. There is some evidence that females are socialized to use a more "internalizing" coping strategy and tend to turn their anger inward toward the self; especially during adolescence (Renouf & Harter, 1990). In other words, females have been found to use more nonaggressive and indirect strategies to cope with their anger. Observational studies of gender differences in dyadic interactions between mothers and their young infants have noted that mothers displayed more positive affect to female infants than to male infants and, in general, demonstrated more emotional expressivity to females (see Brody & HaU, 1993 for a review). Other observations of parent behavior with preschoolers have found similar gender differences in socialization. In one study, Grei^ Alvarez, and Ulman (as cited in Brody & Hall, 1993) found that

PAGE 38

fathers used more emotion words, overall, with their daughters than they did with their sons, and mothers minimized discussing anger with their preschool daughters, but openly discussed angry feelings with their sons. Fivush (1989) examined mother-child conversations about the past and found, with preschool-aged daughters, mothers rarely spoke about anger but did so with their sons; when preschool-aged daughters had been angered by a peer, mothers enq)hasized the feeling of sadness because the child's friendship with the preschool peer had been threatened by conflict. Mothers have also been observed to discuss the causes and consequences of emotion more with sons than with daughters and to pay more attention to male toddlers' expressions of anger. In contrast, the angry outbursts of female toddlers were ignored or inhibited by their mothers (Brody & Hall, 1993). Shields (1995) argued that gender-role socialization within the femily and peer group is significant in that it initiates and sustains the connection between emotion beliefs and values in gender development. On the basis of the work of Deaux and Major (1987), she stated that "gender-related behaviors are influenced by three aspects of the context: expectations of perceivers, self-systems of the actor, and situational clues" (p. 213). Negative emotions are gender-coded. The outward expression of anger is most commonly associated with the masculine role; the internalization of anger and sadness is characteristic of the female role. The most notable difference between the sexes, particularly in adults, is women's reportedly greater conflict about openly expressing anger. In Shield's research, adult women described anger as effective but upsetting and costly to relationships. To summarize, gender differences in emotion regvilation, that is, decisions to express the negative emotions of anger and sadness, have been linked to emotion type.

PAGE 39

31 early gender-role socialization by parents and peers, individual overall emotional expressivity, socioeconomic status, the interpersonal context of emotional expression, expectations of the perceiver, self-system of the individual, and situational clues. According to Saami (1988), "by middle childhood children are capable of subtle insights into how emotional experience and social context are to be integrated" (p. 290). The current research was designed to build on Saami's contention that adolescents have attained this level of social cognitive development. It is based on her prediction that adolescents will use their implicit or naive theories of emotion to regulate their own negative emotion and to consider the consequences of their dissembled or genuine emotional displays on others as a function of the social or interpersonal context. Interpersonal Relationships and Outcome Expectancies The interpersonal relationships between mothers and adolescents, fethers and adolescents, and between adolescent best friends have been the focus of considerable research in emotional display and expectations for social support. Little research, however, has been conducted on adolescents' emotional expression or dissemblance with a non-parental adult figure. For example, Fuchs and Thelen (1988) conducted a study that specifically focused on interpersonal relationships and the consequences children (Le., first, fourth, and sixth grade) expected for expression of anger and sadness with parents. They found that older children, especially males, held less positive expectancies for genuine expression of emotion with parents than did younger students. This was especially true for males and the expression of sadness. Overall, males became less emotionally expressive with age and tended to mask their feelings of sadness, especially with their fathers. Fuchs and Thelen (1988) suggested that parental sociaU2ation practices are directed toward the suppression of sadness in

PAGE 40

32 boys. On the contrary, girls, in their study, expected to receive more support and fewer negative consequences for genuine expression of both sadness and anger with parents. These findings demonstrated a strong correlation between the likelihood of expression of negative affect and its anticipated consequences, but the focus of the study was limited to the interpersonal context between children and parents; a private world in contrast to the public setting of schools. The social contextual variable was not considered. Whitesell and Harter (1996) conducted a study varying the interpersonal context as the critical element in early adolescents' decisions to express anger at a personal violation. The sample consisted of two age groups; preadolescents (ages 1 1 and 12) and young adolescents (ages 13-15). Their focus was on peer relationships and how the changing nature of adolescent friendship status affects decisions to express or dissemble negative emotion. They varied their hypothetical vignettes to depict anger-provoking actions of either best friend or casual acquamtance. Whitesell and Harter found gender differences in the ratings of personal violation ensuing from the offending act of being called "stupid" by best friends or classmates. Girls reported higher ratings of more intense negative emotions (i.e., anger, sadness, and hurt feelings) than did boys and were more relationship oriented in their choice of response strategies. Furthermore, with age, females were more Ukely than males to report a stronger sense of violation if the angerprovoking action had been the action of a best friend. The older females were also more likely than males to report hurt feelings and a deep sense of sadness when wronged by a best friend. ' • According to WhiteseU and Harter (1996), hurt feelings are a complex blend of negative emotion comprised of significant amounts of both anger and sadness, as well as fear about the potential loss of a valued relationship. With casual acquaintances hke

PAGE 41

classmates, no significant relationship existed. Consequently, in their study, when offended by a classmate, the most commonly reported negative emotion for both sexes was ai^er. Whitesell and Harter's (1996) research underscores how the interaction between age, gender, expectation for a friend's behavior and how the interpersonal context of a relationship (i.e., friendship status) can shape adolescents' interpretation of and responses to negative emotional elicitors. It did not specifically address, however, the various types of outcome expectancies adolescents might consider in response to an emotionally provocative event, nor did it examine decisions to express anger within a given social context if the anger experienced had not been evoked by a best friend or classmate. Saami's (1988) investigation featured peers and non-parental aduh figures in emotionally evocative vignettes depicting common situations that occur at home, during play, and at school. Her participants were students in Grades 2, 5, and 8. As previously stated, she found age and sex differences in her participants' beliefs about how others would react to their genuine or dissembled emotion. The reported consequences for emotional display varied as a fimction of the social context described in the vignette. The relationship between the protagonist and the peer or non-parental aduk (i.e., an axmt or a male school principal) affected the participants' decisions to express or dissemble genuine emotion as well as their expectations of support, teasing, or punishment. In one vignette about accidentally setting off a fire alarm in a school setting and being apprehended by the principal, older children (i.e., especially female eighth graders) thought that if they did reveal their fear of punishment to the intervening principal, he would deal with them less harshly. According to Saami, this anticipated consequence indicated that older children recognized that displaying genuine emotion for a misdeed

PAGE 42

34 might be a strategic and regulated act to escape severe negative punishment. Younger children tended to focus only on the consequences of breaking school rules, and 73% of them expected negative reactions. They were not as likely to believe that genuinely expressed emotion would reduce the principal's wrath. Saami's (1988) investigation highlighted the interaction between age, sex, social cognition, and the integration of these variables with the mterpersonal and situational context. Although her study provided support for children's understanding of the appropriate use of culturally prescribed display rules for emotional expression, her findings are limited to middle childhood and have not been replicated m a similar study investigating emotion management decision making in an adolescent population. The current study was designed to extend our knowledge of adolescents' outcome e)q)ectancies and how these expectations regulate their own emotional expression and influence the emotional experience of specific others m an educational context. These expectancies include positive supportive reactions from one's parents or peers, negative reactions such as teasing, negative instrumental consequences, such as punishment or losing privileges, upsetting the other person in the relationship and making them feel bad, a belief that one's emotional expression might violate a social rule, and finally, anticipating that not expressing genuine emotion might make one feel worse than if one dissembled it. Zeman and Shipman (1997) focused on these outcome expectancies and how the social context or interpersonal relationship(s) between adolescents and parents and peers mteract with emotional regulation decisions regarding sadness and anger. As noted earlier, Zeman aixl Shipman incorporated the audience figure from previous research into their investigation of social-contextual influences on outcome expectancies for managing

PAGE 43

35 anger and sadness in 5th-, 8th-, and 1 Ith-grade participants. The combination of audience figure and the age and gender of the adolescent comprised the social context for expression of negative emotion. The results confirmed their predictions that the presence of an audience figure (mother, fether, or best friend) would influence adolescents' decisions to regulate their negative emotion. With age, adolescents' reported regulating their emotions more in the presence of their parents, especially their mothers, and had very different outcome expectations based on the presence of the specific audience figure. In general, participants reported that they expected more teasing and belittling from best friends if they genuinely expressed anger or sadness. Furthermore, this group of adolescents anticipated more loss of privileges and negative instrumental consequences from both mothers and best friends in contrast to the reaction of fathers. These findings need to be put in perspective, however. In this study, the emotionally evocative event was simulated by a series of vignettes to which the participant was asked to respond as s/he believed the protagonist would respond. The parent or best friend was not the emotional provocateur depicted in the vignette; they were neutral observers, in whose presence the emotional regulation decision making occurred. The vignettes were deliberately designed to disentangle the influence of the emotional trigger from the expression of anger or sadness. Zeman and Shipman (1997) were interested more in the decision making process and whether the presence of the audience figure influenced adolescents' outcome expectancies than in the type of emotion expressed. This is an important distinction to make because previous studies on emotional expression have frequently confounded these influences. The current research was designed to separate the effect of the emotional provocateur from the decision making

PAGE 44

36 process as Zeman and Shipman did in 1997. It differed from their study however, because the research instrument was created from the responses of the population of interest in the study, 5th-, 8th-, and 1 Ith-grade adolescents. The fmal emotion-evoking vignettes were more appropriate and representative of the types of situations that anger or sadden adolescents m school. As noted earlier, the majority of the research on adolescents' anticipated consequences of emotional disclosure has focused on interactions between themselves and their parents or their peers, or both. Very little research has been conducted within social contexts like schools in which adolescents express or dissemble their emotions with non-parental authority figures. This is an important area of study because the presence of an aduh authority figure might well influence the decision making that occurs in emotional regulation. Teacher-student relationships have received virtually no attention in the emotional development or emotional regulation literature, even though children spend a considerable portion of their waking hours in a social context that creates a different type of interpersonal relationship that may affect outcome expectancies and or consequences for genuine emotional expression. For example, in a study of school-aged children in the third, fifth, and seventh grades. Underwood et al. (1992) asked their participants to respond hypothetically to vignettes depicting hostile, neutral, and non-hostile events in a school situation. They featured peers and teachers as the emotional provocateurs in the videos. Participants first were asked how they would feel if the depicted event happened to them. Next, the participants were asked about the type of facial expression they would display. Finally, the children were asked about their reasoning for choosing the expression they reported

PAGE 45

37 and whether they thought the emotional provocateur (teacher or peer) was being unfair or mean. Underwood et al. (1992) foimd that participants responded differently to the peer vignettes than to the authority figure vignettes. Overall, the participants reported that they would feel angry if a peer was the provocateur, and this response appeared to be independent of the intentionality of the act. Underwood et al. had varied the intentionality in their vignettes to reflect hostile, non-hostile, or ambiguous acts. Their participants, however, did not interpret the intentionality as the researchers had conceived it. They perceived non-hostile and ambiguous acts by peers as ehciting anger. The teacher vignettes evoked less anger but more sadness in their participants. These children reported usuig display rules for anger more often for teachers than for peers and chose more nonconfrontational actions with teachers than with peers. With age, the children reported more masking of angry fecial expressions, but only with teachers. Girls reported more masking of anger expression than did boys, but all these findings appeared to be a fiinction of the social context conveyed by the vignettes. Some unusual age by gender by emotion interactions occurred in this study. For boys, there was a trend, with age, to report usmg display rules for anger. Seventh-grade boys reported using more display rules for anger than did third graders. For girls, the opposite occurred; seventh-grade girls reported less display rule usage than third-grade girls. A similar developmental trend was reported for dissembling sadness. With age, boys mcreasingly reported masking their sadness, whereas the opposite was noted for girls. Third-grade girls reported invoking more display rules than did seventh-grade girls. The research of Underwood et al. (1992) demonstrated third-, fifth-, and seventhgrade children's understanding of socially conveyed display rules and how the nature of

PAGE 46

38 the interpersonal relationship can effect decision making about expressing genuine negative emotion. It should be noted, that in this study, the social and cultural expectations for anger reactions and aggressive displays of behavior differed between researchers and participants. The researchers were employed by a private university whose faculty and students are predominantly Caucasian and middle to upper-middle class; the participants were African American children from urban, low-income &miUes who resided in the university town. According to the authors, all nonAfrican American children were excluded from participation because race was not to be a &ctor in their investigation. Underwood et al. were puzzled because their peer vignettes evoked more anger responses and retaliatory behavior, overall, even when the provoking incident was ambiguous. They speculated that their participants' interpretation of psychological or physical harm might have stemmed from their social and cultural upbringing and could have influenced this finding (Miller & Sperry, 1987). Nevertheless, this study is valuable in demonstrating the developmental differences in third, fifth, and seventh graders' understanding of appropriate emotional disclosure in an understudied social context, the school. It also highlights the differences in interpersonal relationships between peers and teachers and suggests that schools and teachers are important socialization agents of students' negative emotional expression and behavior. It fiirther emphasizes that young adolescents are aware that "it is more dangerous and less socially acceptable to acknowledge [and display] anger toward authority figures than peers" (Underwood et al., 1992, p. 376). However, as previously stated, the problem with this study as with many others, is that display rule awareness and emotional management decisions are confounded with the authority of the emotional

PAGE 47

39 provocateur. Adolescents' decisions to express or dissemble genuine emotion may be less constrained when the teacher is a well-liked but neutral audience figure. Summary In brie^ research suggests that the understanding and use of display rules are related to emotional intensity, age, gender, emotion type, social cognitive development, interpersonal relationships, socialization figures, and the social and cultural factors embedded in various social contexts. In short, with age, children acquire emotional knowledge that they apply to situations with increasing social cognitive ability if the motivation to do so matches their needs and expectations for desirable interpersonal outcomes. Taken together, the results of this hterature review indicate a need to investigate adolescents' thinking about their decisions to express or mask their genuine negative emotion in school, to explore their self-efficacy beliefs about their ability to control negative emotion, and to examine their outcome expectations for expressing negative emotion in the presence of a wellliked teacher or a best friend.

PAGE 48

CHAPTER 3 METHODOLOGY The purpose of this research was to examine developmental changes in adolescent^ reasoning in emotion decisions to dissemble or display anger and sadness from preadolescence (5th-grade) through middle adolescence (8th-grade) to late adolescence (1 Ith-grade). I also examined adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of displaying or suppressing emotion for themselves and two audience figures, a close friend and a teacher, in a school context. Because little research has examined the role of teachers and their influence on adolescent:^ emotional responding, I investigated whether the presence of a well-liked teacher would influence adolescents? decisions to express emotioiL This chapter describes the participants, research instruments, data collection procedures, and methods of data analysis used in the pilot study and subsequent dissertation study. Pilot Study , A pilot study was conducted to identify school-related events or situations that elicited feelings of sadness and anger in 5th, 8th, and 1 1th grade students. The questions were open-ended to allow participants to respond freely from their personal experiences at school. J ' ' — :^ f" ." " '

PAGE 49

41 Participants The participants were drawn from a population of students attending local elementary, middle, and high schools in Alachua County, Florida. The total population of Alachua County, according to the 2000 U.S. Census, was 217,955 and was conprised of 73.5% White, 19.3% Black or AfricanAmerican, .2% Native American, 3.5% Asian, 5.7% Hispanic or Latino of any race, 1.4% some other race and, 0.1% two or more races. One hundred thirty six students from five local public schools participated in the pilot study. Thirty-six 5th-grade (13 males and 23 females), 52 8th-grade (17 males and 35 females), and 48 1 Ith-grade (24 males and 24 females) students participated in the first phase of this study. The 5th graders ranged in age from 9 years 10 months to 1 1 years 8 months (M= \0 years 5 months, iSD = 3.19 months). The 8th graders ranged in age from 12 years 9 months to 15 years 2 months (M=\3 years 4 months, SD = 3.78 months). The 1 1th graders ranged in age from 16 years to 18 years 6 months (M=\6 years 5 months, SD 3.44 months). The ethnic composition of the sample was 64.7% Caucasian, 22.1% AfricanAmerican, 3.7% AsianAmerican, and 2.9% Hispanic. Five percent of the students described themselves as multi-ethnic. A detailed description of the sample by grade level is provided in Table 1 . ^ t v ' , , ? Procedure The pitot study questionnaire (See Appendbc A) was administered to 5th-, 8th-, and 1 Ith-grade students who had returned their parental consent forms. The students were asked to describe events or situations that produced anger or sadness at school or in a school-related context. They received no compensation for their participation. Students were told that the research was concerned with learning more about common

PAGE 50

42 Table 1 Description of the Pilot Study Sample Sex 5th % 8th % nth % Total Male 13 36.1 17 32.7 24 50 54 Female 23 63.9 35 67.3 24 50 82 Total 36 100 52 100 48 100 136 Ethnicity Asian or Pacific Islander 2 5.5 1 1.9 2 4.2 5 AfricanAmerican 13 36.1 5 9.6 12 25.0 30 Hispanic 2 3.8 2 4.2 4 White (Non-Hispanic) 19 52.8 39 75.0 30 62.5 88 Multi-Ethnic 2 5.5 4 7.7 1 2.1 7 Other 1 1.9 1 2.1 2 Total 36 100* 52 100* 48 100* 136 *Total percentage has been adjusted by rounduig. feelings that most students experience related to schooling. They were then asked to respond, in writing, to three open-ended questions and instructed to mark the intensity of their emotion on a 4-point Likert-type scale. The questionnaire requested the student to,'Think about and write down something that recently h^pened at school or in a school related event that made you feel angry (or sad)!' The second part of the question referred to the intensity of the e^rience:'How angry (or sad) did this make you?' Students were asked to quantify the

PAGE 51

43 intensity of their emotion by selecting a numerical response on a scale ranging from 1 (only a little annoyed or sad ) to 4 (extremely angry or sad) . All participants were requested to relate at least one specific situation for anger and one specific situation for sadness. In order to remediate any potential residual negative emotion in participants in the pilot study, the questionnaire also asked students to think about and relate an event that made them feel happy or glad. This question was always asked last. The pilot study information was vital to the broader dissertation study for the following reasons: first, by collecting student-generated emotion evoking situations, the construct validity of the emotion evoking vignettes of the final assessment instrument was increased. The vignettes used in similar studies of children's and adolescenti^ decision making about regulation of anger or sadness (Zeman & Garber, 1996; Zeman & Shipman, 1997) were constructed for use with younger children and were later modified for their study with adolescents. Zeman and Shipman hypothesized that perhaps the mild emotion evoked by their vignettes was not sufficient to produce the predicted age differences in emotional disclosure to best friends versus parents. In another related study. Underwood and her colleagues (1992) were surprised that the anger producing scenarios they created for their research project did not evoke the anger responses they anticipated from their participants. The current pilot study was designed to reduce the likelihood of sunilar inauthentic or developmentally inappropriate emotional scenarios. Second, because the next phase of this research involved local students, the student-generated emotional events were meant to compensate for potential differences in student perception of emotional events that might result from demographic and

PAGE 52

44 geographic particulars. Previous research on this topic has been conducted in differing regions of the country, notably the northern United States. Student^ experiences at school and teacher behavior and practices might vary as a function of geographic location. Assessing local student^ experiences was intended to increase the ecological validity of the emotion evoking vignettes on the dissertation instrument and contribute to more valid and reUable research results. . ' " > > . v = . Data Analysis Holsti (1969) described content analysis as a'technique for making inferences by objectively and systematically identifying specified characteristics of messaged' (p. 14). In accordance with the guidelines for this type of analysis, students' responses were read, independently, by two reviewers (i.e., the principal investigator and a second independent reviewer) and examined for commonalities. Categories were created from the literature and were derived from "groups of words with similar meaning or connotation" (Weber, 1990, p. 37). The categories were also revised as necessary to be mutually exclusive and exhaustive so that every response was covered by the coding system. Sbt categories emerged from the analysis and represented six major themes: (a) physical aggression to the self or toward another person (1 1%), (b) psychological injury to the self by a friend, a teacher, a classmate, or an unnamed other individual (44%), (c) getting a bad grade (5.1%), (d) unfeir rules (21.3%), (e) being unjustly or falsely accused of something (8.8%), and (f) being treated like babies (1 .5%). Three other categories were created to account for the remaining responses and to comply with the guidelines for content analysis (Holsti, 1969); a denial of anger category (2.2%), a'hDt school related' situation category (.7%), and a no response category (2.9%). The raters' percent of agreement for

PAGE 53

45 the anger scenarios was 92% for the 5th-grade responses, 90% for the 8th-grade responses, and 94% for the 1 Ith-grade responses. All differences were resolved by discussion. The anger categories for the content analysis were consistent with previous research findings for causes of anger in preadolescent and adolescent children. The tendency to become angry as a result of physical aggression to one's self decreased with age from a 5.9% response rate from 5th graders to 1.5% for 1 1th graders. A noticeable increase in the total number of psychological injury anger situations reported from 5th grade to 8th grade supported Whitesell and her colleagues' (1993, 1996) findings with adolescents. The number of anger situations concerning rules that were deemed to be unfair also increased considerably with age; 4.4% for 5th graders, 6.6% for 8th graders, and 10.3% for 1 1th graders. See Table 2 for frequency of anger category by grade. Sadness categories were derived foUowmg the same procedure used in the anger content analysis and were informed by a review of the Uterature. Nine themes emerged from participant^ responses: (a) physical or psychological separation from friends or loved ones, (b) death, (c) general disappointment, (d) loneliness and aUenation, (e) failure to achieve a specific goal, (f) physical harm or injury to one's self or another person, (g) psychological harm or injury to one's self caused by a friend, classmate, teacher or unnamed other person, (h) psychological harm or injury to another person, and (i) loss of personal property or possessions. As in the anger content analysis, three other categories were created to fulfill the requirements of content analysis; denial of sadness, a'bot school related' category and no response. Interrater percent of agreement for the sadness scenarios was 94% for the 5th-grade responses, 92% for the 8th-grade responses.

PAGE 54

46 Table 2 Frequency Table of Anger Categories by Grade Grade Anger Category 5th 8th 11th Total % nhvsical aesression to self 8 4 2 14 10.3 1 1. 1 1. 7 nsvcholoffical iniiirv bv a friend 4 8 3 15 11.2 A in 7 1 S 4 nsvcholoPical iniiirv hv a classmate 6 2 9 17 12 5 D5?vcholoeical iniurv hv unnamed otherT 2 4 4 10 7 4 eettine a bad erade 1 4 2 7 5 1 unfair rules 6 9 14 29 21.3 being unjustly/falsely accused 2 9 1 12 8.8 treated like babies 2 2 1.5 denial of anger 2 1 3 2.2 not school related 1 1 .7 no response 2 2 4 2.9 Total 36 52 48 136 100 and 92% for the 1 Ith-grade responses. All differences were resolved by discussioa See Table 3 for frequency of sadness category by grade.

PAGE 55

47 Table 3 Frequency Table of Sadness Categories by Grade Grade Sadness Category 5th 8th nth Total % L/lJLYalWai o&UCUallUll -J c o Q r%cvf*nr*lr\o'if*ji 1 crf*T*\5if*jitir\n A t 7 17 0.0 J 4 1 1 1 1 S 1 0. 1 ^biibiai uiacijH'^AiiULJXidii 1 1 1 1 o 0 in 7 4 iallUTC LU aCniCVC a SpcClIlC gOai •5 J / if 0 lo 1 1.0 rinv<2iPfi1 nfirm to j*nf*thf*r tv*i"Qf^fi A *r 0 4 4 1 0 1 u Q o •5 J 71 13.4 psychological harm to self by teachers 2 1 1 4 2.9 psychological harm to another person 2 10 13 25 18.4 loss of personal property 2 2 4 2.9 denial of sadness 2 2 3 7 5.1 not school related 2 2 4 2.9 no response 3 3 2 8 5.9 Total 36 52 48 136 100

PAGE 56

48 Consistent with the research of Rotenberg et al. (1987-88), there was an age trend in reported sadness situations resulting from psychological harm. Fifth-grade students reported experiencing sadness resulting from some harm to themselves (28% of 5th-grade responses), whereas the nxmiber of instances of psychological hann reported by 8thand 1 Ith-grade students decreased with age, from 15% to 6% respectively. In contrast, the number of sadness situations incurred by psychological harm directed at a person other than themselves increased with age; from 5.5% in the 5th-grade to 27% in the 1 Ith-grade. Eighth graders were most likely to cite psychological separation from a friend as a cause of sadness in school. Their responses centered around friends transferring to different schools and the break-up of romantic relationships. Eleventh-grade students were most saddened by teasing, bullying, or harassment of other students with Umited mental or physical capabilities in the school situation. Thirty-one percent of their responses in this category made reference to this type of incident. This finding is consistent with the literature on decreasing self-involvment during late adolescence and an increase in perspective taking abilities (Steinberg, 1996). Furthermore, students were more reticent to acknowledge their sad emotion than their anger. The "no response" rate doubled from the questions about anger situations to those about sadness. The "denial of emotion" rate increased from a total of three for the anger questions to seven for situations involving sadness. Participants? responses were tallied by emotion type (anger or sadness) to determine the most frequently mentioned incidents by grade. The categories were con^ared with recent fmdings in the literature for similarity of response with other preadolescents and adolescents regarding the causes of anger and sadness (Kamiol &

PAGE 57

49 Heiman, 1987; Rotenberg et al., 1987-1988; Tangney et al., 1996; Torestad, 1990; Whkesell et al., 1993) and found to be consistent. The most frequently ched category for each emotion, plus the overall most frequently cited category, were then used to construct four anger and four sadness vignettes for the dissertation study. Dissertation Study Participants The participants in this study were 21 5th graders (10 males and 1 1 females, 46 8th graders (14 males and 30 females), and 50 1 1th graders (25 males and 25 females) drawn from P. K. Yonge Developmental Research School. An additional 26 fifth-grade students (10 males and 16 females) were recruited from one of the pilot study elementary schools to augment the fifth-grade student total and maintain approximately equal numbers of participants by grade. In all, 146 students participated in this study. The 5th graders ranged in age from 10 years 2 months to 12 years 2 months (A/ = 10 years 5 months, SD = 3.66 months). The 8th graders ranged in age from 13 years 1 month to 15 years 2 months (A/= 13 years 5 months, SO = 3.5 months). The 1 1th graders ranged in age from 16 years 2 months to 17 years 8 months (M= 16 years 7 months, SD =3.3 months). All students received a free video rental certificate as conq)ensation for their participation. P. K. Yonge was the school of choice for two reasons; first, the students who attend P. K. Yonge are demographically representative of the population in North Central Florida. The school uses a lottery system for admission to ensure an equitable distribution of students based on gender, race, socioeconomic status, and academic ability. Second, a majority of P. K. Yonge students commenced their education there, and did not

PAGE 58

50 experience potentially disruptive school transitions from elementary to middle school, and from middle school to high school that could contribute to measurement error. See Table 4 for a description of the dissertation study participants. Table 4 Description of the Dissertation Study Sample Sex 5th % 8th % 11th % Total Male 20 42.5 14 31.8 25 50.0 59 Female 27 57.5 30 68.2 25 50.0 82 Total 47 100 44 100 50 100 141 Ethnicity Native American 1 2.1 1 Asian or Pacific Islander 1 2.0 1 AfricanAmerican 15 31.9 12 27.3 9 18.0 36 Hispanic 1 2.1 3 6.8 2 4.0 6 White (Non-Hispanic) 25 53.2 23 52.3 34 68.0 82 Multi-ethnic 5 10.6 6 13.6 4 8.0 15 Total 47 100* 44 100 50 100 141 * Total percentage has been adjusted by roundmg.

PAGE 59

51 Research Design The study was a 2 (emotion type) X 2 (sex) X 3 (grade level) X 2 (audience figure) mixed model factorial design. Emotion was designated as a within-subjects variable with two types: anger and sadness. Between-subject variables were sex (i.e., male and female) and grade level (i.e., 5th, 8th, and 1 1th). This design differs slightly from Zeman and Shipmarfs (1997) research design. In their study, audience figure was designated as a between-subject variable with three levels (i.e., mother, fether, and best friend) "to reduce the possibility of carryover effects" (p. 919). Although their concern about carryover effects was imderstandable, their design resulted in very small sample sizes in each cell. In the current study, I decided to designate audience figure (i.e., best friend and teacher) as a within-subjects variable to increase the power of the research design. Research Instrument Eight short stories (vignettes) depicting the most frequently reported school situations concerning anger and sadness from the pilot study were created for the questionnaire. Four stories described situations that commonly elicited sadness, and four stories featured situations that commonly evoked anger in the pilot study adolescents. Each vignette was written in a third-person narrative and specified the emotion it was intended to evoke. Participants were asked to assume the role of the protagonist in the story and to answer the questions posed on the questionnaire. Each vignette contained two sets of eight questions; one set for each audience figure (i.e., a close friend or a wellliked teacher). In this study, contrary to previous research, the emotional provocateur in the vignette was not the person (either well-liked teacher or best friend) with whom the

PAGE 60

52 adolescent had to decide whether to display or dissemble genuine emotion. The audience figure was designed to be a neutral observer in order to assess more accurately adolescents' emotion management decisions and to disentangle the decision making about emotional regulation from the eliciting situation or event. (See Appendix B for the research instrument, Appendix C for the consent forms, and Appendix D for the student assent scripts.) Emotion expression and feelings of self-efificacy. The first question concerned the decision to express or dissemble their emotion. 1 . Would you show how (sad or angry) you feel to your (teacher or best friend) ? Participants were instructed to indicate their decision on a 4-point Likert-type scale (1 definitely would not 2 = probably would not 3 = probably would . 4 = definitely would ). This scale's wording was changed slightly to reflect the syntax of each question. The second question addressed the participants' emotional self-efficacy (i.e., their belief in their ability to regulate their own emotional response): 2. Do you think you could control showing your (angry or sad) feeling if you wanted to? Questions 3 through 8 addressed the participants' outcome expectancies, that is, their perceptions of consequences for expressing or dissembling their anger or sadness. The questions and descriptions were adapted from Zeman and Shipman's (1997) research. Outcome expectancies. The first category, interpersonal support addressed the participants' belief that expressing their emotion would be met with a positive, supportive response: 3. Do vou think vour (teacher or best friend) would be understanding of how (sad or angry) that makes vou feel? The second category, negative interpersonal

PAGE 61

53 response , concerned expectations of some sort of negative reaction to participants' expressing their feeling: 4. Would your (teacher or best friend) make fim of you or tease you if you showed how (angry or sad) you really felt? The third category, instrumental consequences , referred to the participant^ expectation that to express emotion would result in their receiving some sort of negative noninterpersonal consequence from the audience figure: 5. Would there be a negative response from your (teacher or best friend) — would he or she punish you in some way-tf you showed how (sad or angry) you felt? The fourth category, emotional reaction , reflected the participants' belief that expressing their emotion would resuk in the audience figure experiencing some negative emotional response: 6. Would showing your (angry or sad) feelings upset your (teacher or best friend) and make (him or her) feel bad ? The fifth category, norm maintenance, was defined as adherence to some rule regarding regulation of emotion in specific situations: 7. Do you believe that you should show (sad or angry) feelings to your (teacher or best friend) ? The last category, internalization , addressed the participant^ self-esteem and involved the anticipation that dissembling emotion would result in either a better or worse outcome than might occur if the participants' expressed their emotion to the audience figure: 8. How would you feel if you kept your (sad or angry) feelmgs inside and did not show them to your (teacher or best friend)? Procedure Students who had parental permission to participate were assessed in small groups (10 to 20 same-grade students) in sessions of approximately 30 to 45 minutes depending on grade level. I provided directions orally and read aloud a sample vignette. The vignettes were counterbalanced by emotion type across all participants. For the friend

PAGE 62

54 audience figure condition, students were told to think of their same-sex best or closest friend. For the teacher audience figure condition, participants were instructed to think of a particular teacher they liked. Participants were then asked to read the eight vignettes on their own and to answer the questions for each story. Scoring and Data Analyses Repeated measures analyses of variance (ANOV As) were conducted on the first two questions: emotion management decisions and self-efficacy regarding emotional regulation. The dependent variable was each student's response to the question using the 4-point Likert-type scale summed across the four stories for each emotion type. Scores could range from 4 to 16 for sadness or anger. The within-subjects variables were emotion type (i.e., anger or sadness) and audience figure (i.e., best friend or teacher). The between-subjects variables were the participants sex and age or grade (i.e., 5th, 8th, and nth). The scores on the final six questions, the outcome expectancies, were treated as a unified construct and a repeated measures multivariate analysis of variance (MANOVA) using the muftivariate criterwn of Wilksf lambda (A) was conducted on the total outcome expectancy score. The dependent variable was each participants total score on each of the six expectancies, summed by emotion type across all four vignettes. Scores could range from 4 to 16 for each emotion type. Univariate ANOV As were conducted when significant multivariate effects emerged. Significant main effects were analyzed with ANOV As, t-tests, and appropriate post hoc procedures. The familywise error rate was controlled with the Shaflfer-Hohn procedure and the Bonferroni correction for multiple comparisons. All data analyses were performed using SPSS version 10.0 for Windows.

PAGE 63

55 Reliability of the Dissertation Instrument An internal consistency estimate of reliability was computed for the dissertation instrument. The coefficient alphas (a) for each variable in the analysis ranged from .67 to .84. The overall alpha (a) coeflRcient for the dissertation instrument was .91. The estimates are listed in Table 5. The internal consistencies of the scales are generally acceptable given that the alpha coefficient tends to underestimate rehability for questionnaires employing this vignette/scenario approach. Moreover, the questionnaire was administered only once, under different testing situations and location, to 1 1 different groups of students over a 13-week period. Therefore, two distinct sources of error that can affect the rehability coefficient must be considered. The first, individual differences in mood, level of fatigue, and attitude toward the test, has been shown to exert an influence on internal consistency measures. The second, in spite of attempts to standardize administration of the questionnaire, many variations and concessions had to be accommodated in each testing situation (Borg & Gall, 1989). The dissertation questionnaire was extremely long for fifth graders, and they quickly lost interest in selecting the most appropriate response. Eighthgrade students remained focused for a longer time span than did fifth-grade students, but the length of the questionnaire was a deterrent to their concentration, as well. Eleventhgrade students completed the survey in the least amount of time, but several of them resorted to a strategy of checking the same response for each question without regard to the nature of the vignette, audience figure, or outcome expectancy. Taken together, all these factors are likely to have influenced student responses and the subsequent accuracy of internal consistency estimates of reliability.

PAGE 64

Table 5 Internal Consistency Estimates for the Dissertation Instrument Scales Cronbach Number a of Items Show Emotion .81 16 Aneer to Friend .69 4 Anger to Teacher .73 4 Sadness to Friend .75 4 Sadness to Teacher .76 4 Self-Efficacy .93 16 Aneer with Friend X XXX^^X/A TT XbXX X X x^^xxvx .77 4 Aneer with Teacher X x-i.xcn%^x TT x^xx X v^^^ X X 81 4 Sadness with Friend .76 4 Sadness with Teacher .81 4 Interpersonal Support .80 16 Anger with Friend .67 4 Anger with Teacher .73 "l 4 Sadness with Friend .73 4 Sadness with Teacher .72 4 Negative Interpersonal Response .85 16 Anger with Friend .78 4 /\nger wiin i eacner .72 4 Sadness with Friend .75 4 Sadness with Teacher .72 4 Instrumental Consequences .90 16 Anger with Friend .74 4 Anger with Teacher .67 4 Sadness with Friend .80 4 Sadness with Teacher .74 4

PAGE 65

Table ^-continued Scales Cronbach a Number of Items Emotional Reaction .91 16 Anger with Friend .67 4 Anger with Teacher .71 4 Sadness with Friend .78 4 Sadness with Teacher .82 4 Norm Maintenance .83 16 Anger with Friend .76 4 Anger with Teacher .74 4 Sadness with Friend .81 4 Sadness with Teacher .80 4 Internalization .92 16 Anger with Friend .78 4 Anger with Teacher .74 4 Sadness with Friend .83 4 Sadness with Teacher .84 4 Overall Internal Consistency .91 128

PAGE 66

CHAPTER 4 RESULTS The purpose of this research was to examine developmental changes in adolescents' reasoning in emotion decisions to express or dissemble anger and sadness from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (1 1th grade). I also examined adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures, a close friend and a teacher, in a school context. Zeman and Shipman's (1997) research explored the relationship between adolescents' decision making about expression of negative emotion (anger or sadness) and their outcome expectancies. The six outcome expectancies they mvestigated were reexamined in this study. Expectancies, considered as a single construct, were investigated. The first expectancy, interpersonal support, comprised the adolescent's belief that expressing negative emotion would be met with a positive, supportive response. The second, negative interpersonal response, referred to an adolescent's expectation of being teased or made fun of for expressing anger or sadness. The third, instrumental consequences, entailed the student's belief that to express negative emotion in a school context might result in some form of punitive behavior from the audience figure (best friend or teacher). The fourth expectancy, emotional reaction, reflected the adolescent's belief that expressing negative emotion would upset the audience figure (best friend or teacher). The 58

PAGE 67

fifth, norm maintenance, was defined as adherence to an internalized cultiiral rule regarding expression of a negative emotion in a social context. The sixth and last expectation, internalization, addressed the adolescent's resulting state of well-being and self-esteem if they chose not to display their negative emotion to a close friend or teacher. Zeman and Shipman (1997) found that adolescents' outcome expectancies varied as a function of emotion type (anger or sadness), audience figure (parent or peer), age, and gender. I extended their research by using school as the social context and a well-liked teacher as an audience figure. This chapter presents the results of the analysis of the data obtained from the questionnaire that asked students about situations at school that were intended to invoke negative emotions. The statistical procedures used in the data analyses were all performed using the SPSS Graduate Pack 10.0 for Windows. Hypothesis 1 — ^Emotion Management Decisions The first research hypothesis proposed that adolescents' decisions to express or dissemble emotion would vary as a fimction of the type of audieiK^e figure (best friend or teacher), type of emotion (anger or sadness), age (grade), and sex of the participants. It predicted that with age, adolescents' decisions to regulate their emotion would interact with the sex of the participant, with emotion type, and with the type of audience figure. The descriptive data are presented in Table 6. A repeated measures imivariate analysis of variance (ANOVA) was conducted on the first question; the student's willingness to display or dissemble their emotion. The dependent variable was the participant's response summed across all four stories for each emotion type. Scores ranged from 4 to 16 for each emotion type. Within-group variables

PAGE 68

60 were emotion type with two levels (anger or sadness) and audience figure with two levels (best friend or teacher). The between-group variables were sex (male or female) and grade with three levels (5th, 8th, or 1 1th). A significant main effect emerged for audience figure only, F(l, 129) = 57.86, P = .000. There were no significant main effects for emotion type, sex, or grade level. An emotion by grade interaction emerged, but was nonsignificant when the Shafifer-Holm procedure was applied to control the familywise error rate. Two other interaction were significant, however; audience figure and sex of the participant, F(l, 129) = 12.72, p = .001, and audience figure and grade, F(2, 129) = 6.93, p = .001. The repeated measures ANOVA results are presented in Table 7. Audience Figure X Sex interaction. A follow up analysis of variance (ANOVA) for simple main effects was conducted on the audience figure by sex interaction to further examine the significant interaction effects of the repeated measures ANOVA. The familywise error rate was controlled by the Shaffer-Hohn procedure. The simple main effects ANOVA revealed a significant difference on audience type for both sexes; males, F(l, 56) = 7.46, E = .008, females, F(l, 77) = 69.45, p = .000. Female students reported that they would be more likely to display their negative emotion to a friend rather than to a teacher. Males also reported that they were more likely to show their anger or sadness to a friend than a teacher, but the difference was not as large. The descriptive statistics appear in Table 8. The simple main effects ANOVA results are Usted in Table 9.

PAGE 69

61 Table 6 Descriptive Statistics for Adolescent^ Decision to Express Emotion Group Sex Grade Mean SD N Show Anger to a Friend Male 5th 10.40 3.32 20 8th 11.62 2.93 13 nth 12.04 2.24 24 Female 5th 11.84 3.42 25 8th 11.54 2.06 28 • « 11th 12.68 2.51 25 Show Anger to a Teacher Male 5th 10.60 3.36 20 8th 10.15 2.48 13 nth 10.29 2.79 24 Female 5th 10.00 3.06 25 8th 9.04 2.85 28 nth 10.28 3.13 25 Show Sadness to a Friend Male 5th 10.55 3.25 20 8th 10.15 3.78 13 nth 11.46 1.89 24 Female 5th 11.84 3.09 25 8th 12.32 2.26 28 nth 12.76 3.06 25 Show Sadness to a Teacher Male 5th 11.10 3.18 20 8th 9.08 3.04 13 nth 9.67 2.14 24 Female 5th 10.76 2.86 25 8th 8.64 2.42 28 nth 9.52 3.66 25

PAGE 70

62 Table 7 Results of Repeated Measures ANOVA on Adolescents' Decisions to Express Emotion Type III Sum Mean Source of Squares df Squares F p Between Subjects Sex 14.84 1 14.84 .68 .412 Grade 50.62 2 25.31 1.16 .318 Sex X Grade 2.52 2 1.26 .06 .944 Error 2823.55 129 21.89 thin Subjects Emotion 6.05 1 6.05 1.85 .176 Emotion X Sex 11.21 1 11.21 3.42 .067 Emotion X Grade Zl .Z 1 z in 041 Emotion X Sex X Grade 11.14 2 5.57 1.70 .187 Error (Emotion) 422.60 129 3.28 Audience 354.16 1 354.16 57.86 .000 Audience X Sex 77.84 1 77.84 12.72 .001 Audience X Grade 84.85 2 42.43 6.93 .001 Audience X Sex X Grade 4.61 2 2.31 .38 .687 Error (Audience) 789.66 129 6.12 Emotion X Audience .28 1 0.28 .17 .683 Emotion X Audience X Sex 3.35 1 3.35 2.00 .160 Emotion X Audience X Grade 6.97 2 3.48 2.08 .129 Emotion X Audience X Sex X Grade 5.01 2 2.51 1.50 .228 Error (Emotion X Audience) 216.15 129 1.68 Note, a = .05.

PAGE 71

63 Table 8 Descriptive Statistics for Audience by Sex Interaction on Decision to Express Emotion Group Sex Mean N Show Emotion to Friend Male 11.11 57 Show Emotion to Teacher 10.20 57 Show Emotion to Friend Female 12.15 78 Show Emotion to Teacher 9.67 78 Table 9 Simple Main Effects ANOVA for Audience by Sex Interaction on Decision to Express Emotion Source Sex Type III Sum of Squares Mean df Squares Audience Error Audience Error Male Female 23.26 174.61 240.01 266.11 1 23.26 7.46 .008 56 77 3.12 1 240.01 69.45 .000 3.46

PAGE 72

1 64 Audience X Grade interaction. A follow up test for simple main effects was conducted on the audience by grade interaction to further investigate the differences in decisions to express emotion to either a best friend or a teacher. The descriptive statistics for the simple main effects ANOVA are presented in Table 10. The Shaffer-Hohn procedure was used to control the familywise error rate. The ANOVA revealed a significant difference on audience type for 8th-grade, F(l, 44) = 34.59, p .000, and 1 Ith-grade students, F(l, 48) = 47.94, p = .000. In contrast to 5th graders who reported that they would not distinguish between teacher and best friend in expressing emotion, 8th and 1 1th graders reported that they would be less likely to display emotion to a teacher than to a best friend. The results of the ANOVA appear in Table 1 1 . Mean scores revealed that with age, adolescents were inclined to express their emotion to a best friend, rather than to a well-liked teacher. Eighth-grade students, however, were least likely to express negative emotion to a teacher. •1

PAGE 73

65 Table 10 Descriptive Statistics for Audience by Grade Interaction on Decision to Express Emotion Group Grade Mean N Show Emotion to Friend Sth 11.23 45 Show Emotion to Teacher 10.59 45 Show Emotion to Friend 8th 11.60 41 Show Emotion to Teacher 9.09 41 Show Emotion to Friend nth 12.24 49 Show Emotion to Teacher 9.94 49 Table 11 Results of ANOVA for Audience by Grade Interaction on Decision to Express Emotion Grade Source Type III Sum of Squares df Mean Squares F E 5th Audience 9.34 1 9.34 2.65 .110 Error 154.91 44 3.52 8th Audience 129.38 1 129.38 34.59 .000 Error 149.62 40 3.74 11th Audience 130.30 1 130.30 47.94 .000 Error 130.45 48 2.72 Note, g = .0167.

PAGE 74

66 Hypothesis 2 — Self-Efficacv Beliefs The second research hypothesis predicted that adolescents would report greater self-efficacy for regulating sadness than for regulating anger. To investigate this claim, a repeated measures univariate analysis of variance (ANOVA) was conducted on the second question (i.e., whether adolescents thought they could control showing their emotion to their best friend or to a teacher they liked). Again, each participant's response was siunmed across all four stories by emotion type. Scores ranged from 4 to 16 for each emotion. Within-group variables were emotion type and audience figure. The betweengroup variables were sex and grade. The Shaffer-Hohn procedure was used to control the familyAvise error rate. The second research hypothesis was not supported by the data. There was no significant main effect for emotion type, F(l, 128) = 2.51, ^ = .115. The descriptive statistics for the analysis are listed in Table 12. The univariate ANOVA results appear in Table 13. Two other main effects were significant, however. The first was a difference in self-efficacy for grade (age) of the participant, F(2,128) = 4.39, p = .014. Post hoc pairwise comparisons were conducted by grade usmg the Bonferroni adjustment for multiple comparisons. T-tests revealed that only the 1 Ith-grade students' scores differed significantly from the 5th-grade students' responses, t(128) = -2.81, p = .017. The other comparisons were not significant. The mean scores differed by age indicating that 1 1th graders had stronger beliefs in their ability to control their negative emotions than did the 5th graders. Descriptive statistics appear in Table 14.

PAGE 75

67 Table 12 Descriptive Statistics for Adolescents? Self-Efficacy Beliefs Group Sex Grade Mean SD N Controlling Anger with a Friend Male 5th 11.20 3.11 20 8th 13.23 2.89 13 11th 13.13 1.96 24 Female 5th 11.40 2.99 25 8th 12.48 2.64 27 nth 12.56 3.24 25 Controlling Anger with a Teacher Male 5th 11.40 3.22 20 8th 12.69 2.84 13 nth 13.38 2.04 24 Female 5th 12.16 2.93 25 8th 12.56 2.89 27 nth 13.00 3.08 25 Controlling Sadness with a Friend Male 5th 11.10 3.51 20 8th 13.46 1.90 13 nth 12.99 1.99 24 Female 5th 11.32 3.22 25 8th 11.44 2.59 27 nth 12.08 3.27 25 Controlling Sadness with a Teacher Male 5th 10.90 3.35 20 8th 13.15 2.73 13 nth 13.25 2.36 24 Female 5th 11.56 2.45 25 8th 12.15 2.82 27 nth 12.68 3.26 25

PAGE 76

Table 13 Results of Univariate ANQVA on Adolescent^ Self-Efficacy Beliefs 68 Type in Sum Mean Source of Squares df Squares Between Subjects Sex 17.39 1 17.39 .66 .420 Grade 232.85 2 116.43 4.39 .014 Sex X Grade 45.93 2 22.97 .87 .423 Error 3394.72 128 26.52 thin Subiects J Emotion 8.54 1 8.54 2.51 .115 Emotion X Sex 6.50 1 6.50 1.91 .169 Emotion X Grade .35 2 .17 .05 .951 Emotion X Sex X Grade 5.62 2 2.81 .83 .440 Error (Emotion) 435.16 128 3.40 Audience 5.53 1 5.53 5.21 .024 Audience X Sex 8.53 1 8.53 8.04 .005 Audience X Grade 3.50 2 1.75 1.65 .196 Audience X Sex X Grade 1.62 2 .81 .76 .468 Error (Audience) 135.83 128 1.06 Emotion X Audience .02 1 .02 .02 .911 Emotion X Audience X Sex .14 1 .14 .10 .754 Emotion X Audience X Grade 4.10 2 2.05 1.46 .236 Emotion X Audience X Sex X Grade .33 2 .17 .12 .889 Error (Emotion X Audience) 179.49 128 1.40 Note, g = .05.

PAGE 77

Table 14 Descriptive Statistics for Grade Main Eflfect on Adolescents' Self-EfBcacv Beliefs Grade Mean SD N 5th 11.31 3.01 45 8th 12.55 2.69 44 11th 12.66 2.93 50 The second main effect in the self-efficacy analysis occurred for audience figure, F(l, 128) = 5.21, p = .024. However, because the main effect was modified by an interaction between audience figure and sex, F(l, 128) = 8.04, £ = .005, a follow-up oneway ANOVA to investigate simple main effects was conducted. Female students indicated they would be able to control their emotion more with a teacher than with a close fiiend, F(l, 76) = 17.48, p = .000. There was no significant difference in males' reported ability to control their emotion between audience figures, F(l, 56) = .02, p = .904. Males' mean scores for self-efficacy on control of negative emotion were virtually the same for both their best Mend and a teacher. Descriptive statistics for the post hoc analysis are presented in Table 15. Results of the simple mam effects ANOVA for the audience by sex interaction appear m Table 16. Hvpothesis 3 — Outcome Expectancies The third research hypothesis investigated the six outcome expectancies and examined whether adolescents' scores would vary as function of emotion type, audience figure, grade (age), or sex of the participant. It predicted significant main effects in the

PAGE 78

70 Table 15 Descriptive Statistics for the Audience by Sex Interaction on Adolescent^ Self-EflScacy Beliefs Sex Audience Mean N Male Friend 12.44 57 Teacher 12.46 57 Female Friend 11.88 77 Teacher 12.35 77 Table 16 Simple Main Effects ANOVA for Audience by Sex Interaction on Adolescent^ SelfEfficacy Beliefs Type III Sum Mean Source of Squares df Squares F g .0087 .015 .904 .598 8.416 17.48 .000 .481 1 i k Audience Male .0087 1 Error 33.491 56 Audience Female 8.416 1 Error 36.584 76

PAGE 79

71 total outcome expectancy scores for type of emotion, audience figure, sex of the participant, and age. It further predicted significant interaction eflfects for audience figure and sex of the participant, audience figure and grade (age) of the participant, and sex and age of the participants. The scores on the final six questions were treated as a unified construct and a repeated measures multivariate analysis of variance (MANOVA) using the multivariate criterion of Wilks" lambda (A) was conducted on the outcome ejq)ectancy score for each expectancy. Univariate ANOVAs were conducted when significant multivariate effects emerged. Results of the multivariate analysis revealed significant main effects on the outcome expectancy score for grade (age), A = .83, F(6, 118) = 1.88, p = .037, emotion, A = .65, F(6, 118)= 10.72, p = .000, and audience figure, A = .41, F(6, 118) = 27.89, P = .000. There was no significant difference on the between-group variable of sex. These main effects were modified, however, by two-way mteractions between emotion and audience figure, A = .46, F(6, 1 1 8) = 23. 141, p = .000, audience figure and sex, A = .78, F(6, 118) = 5.529, p = .000; and audience figure and grade, A = .75, F(12, 238) = 3.018, p = .001. A three-way interaction between audience figure, sex, and grade, A = .81, F(12, 236) = 2.160, p = .014, was nonsignificant when the Shaffer-Holm procedure was applied to control the familywise error rate. Because each main effect variable was involved in an interaction, only the two-way interaction effects were pursued in further data analyses. All follow-up univariate ANOVAs were conducted at a new alpha level of .008 (i.e., .05/6) using the Shaffer-Hohn procedure. The MANOVA results are presented in Table 17.

PAGE 80

Table 17 MANOVA of Total Outcome Expectancy Scores Source Wilk^ A F Hypothesis df Error df E Between Subjects Sex .90 2.16 6 118 .052 Grade .83 1.89 12 236 .037 Sex X Grade .85 1.67 12 236 .076 Within Subjects • Emotion .65 10.72 6 118 .000 Emotion X Sex .92 1.62 6 118 .149 Emotion X Grade .86 1.58 12 236 .101 Emotion X Sex X Grade .91 .82 12 236 .621 Audience .41 27.89 6 118 .000 Audience X Sex .78 5.53 6 118 .000 Audience X Grade .75 3.03 12 236 (\(\ 1 .001 Audience X Sex X Grade .81 2.24 12 236 .014 Emotion X Audience .46 23.14 6 118 .000 Emotion X Audience X Sex .92 1.72 6 118 .122 Emotion X Audience X Grade .94 .62 12 236 .825 Emotion X Audience X Sex X Grade .95 .53 12 236 .892 Note, g = .05.

PAGE 81

Audience X Sex interaction. A univariate ANOV A was conducted on the audience by sex interaction. The results were significant for the following expectancies: interpersonal support, F(l, 123) = 24.29, p .000, norm maintenance, F(l, 123) = 17.82, P = .000, and internalization, F(l, 123) = 1 1 .09, p = .001 . The descriptive statistics appear in Table 18. The ANOV A results are presented in Table 19. Simple main effects tests were conducted on the significant interactions for each gender separately. The familywise error rate was controlled at a = .008 (Le., .05/6) by using the Shaffer-Holm procedure. The results revealed that females were more likely to expect an understanding supportive reaction fi-om their friend than from a teacher, F(l, 123) = 34.75, p = 000; they believed they should show their emotion to a friend rather than to a teacher, F(l, 123) = 74.64, p = .000; and they indicated that they would feel worse if they did not express or display their emotion to a friend rather than to a teacher, F(l, 123) = 28.12, p = .000. There was no significant difference between audience figures for males on interpersonal support, F(l, 123) = 1.37, p = .244, norm maintenance, F(l, 123) = 2.85, p = .094, or internalization, F(U 123) = .01, p = .940. These results support the hypothesis that there would be a significant audience by sex interaction effect on adolescent^ outcome expectancy scores. Audience Figure X Grade interaction. A imivariate ANOVA was conducted on the interaction effect between audience figure and grade. Only the results for the norm maintenance expectancy were significant, F{2, 123) = 6.42, p = .002. The descriptive statistics appear in Table 20 and the ANOVA results are presented in Table 21.

PAGE 82

Table 18 Descriptive Statistics for Audience by Sex Interaction on Outcome Expectancies Ejq)ectancy Audience Sex Mean N Interpersonal Support Norm Maintenance Internalization Friend Teacher Friend Teacher Friend Teacher Male Female Male Female Male Female Male Female Male Female Male Female 11.43 13.25 11.77 11.72 10.87 11.80 10.37 9.67 8.76 8.00 8.78 9.03 55 74 55 74 55 74 55 74 55 74 55 74

PAGE 83

75 Table 19 Summary Univariate ANOVA Table for Audience by Sex Interaction on Outcome Expectancies Expectancy Sum 01 Squares df Mean Squares F Interpersonal Support 1 1 1 rt^ 11 .UUU Error 1 01 A 11 Neg. Interpersonal Response 1 1 .1 / 1 1 .1 / .uo .olZ Error Z.yo Instrumental Consequences 1 f\A 3.U4 1 1 1 f\A 1 10 1 .JO Error z/U.j4 1 Ol O OB Z.yo Emotional Reaction 3.62 1 3.62 1.49 .224 Error 298.46 123 2.43 Norm Maintenance 79.85 1 79.85 17.82 .000 Error 551.22 123 4.48 Internalization 30.85 1 30.85 11.09 .001 Error 342.06 123 2.78 Note, a = .008.

PAGE 84

76 Table 20 Descriptive Statistics for Audience by Grade Interaction on Outcome Expectancies Expectancy Audience Grade Mean N Norm Maintenance Friend 5th 10.83 43 8th 11.31 38 nth 11.87 48 Teacher 5th 10.47 43 8th 9.43 38 nth 10.16 48 Simple main efifects tests were conducted for each grade separately on the significant interaction for norm maintenance. The familywise error rate was controlled at a = .017 (.05/3) using the Shaffer-Holm procedure. Fifth-grade students' scores did not differ significantly between the two audience figures, F(l, 123) = 1.19, p = .278. Eighth-grade and 1 Ith-grade students were more likely to respond that they should show their negative emotions to a friend but not to a teacher, F(l, 123) = 25.94, p = .000, and F(l, 123) =31.28, p = .000, respectively. With age, students were more likely to respond that they should show their negative emotions to a friend but not to a teacher. These findings support the research hypothesis that predicted a significant interaction between audience figure and grade on outcome expectancies.

PAGE 85

77 Table 21 Univariate ANOVA Table for Audience by Grade Interaction on Outcome Expectancies lixpeciancy Sum of squares Af Ul Mean oquares c U Interpersonal Support 37.40 2 18.70 4.32 .015 Error 532.34 123 4.33 Neg. Interpersonal Response 26.12 2 13.06 4.39 .014 Error 366.31 123 2.98 Instrumental Consequences 8.92 2 4.46 2.03 .136 Error 270.54 123 2.98 Emotional Reaction 5.65 2 2.83 1.17 .315 Error 298.46 123 2.43 Norm Maintenance 57.55 2 28.77 6.42 .002 Error 551.22 123 4.48 Internalization 8.30 2 4.15 1.49 .229 Error 342.06 123 2.78 Note, a = .008. Emotion X Audience interactioa A univariate analysis of variance (ANOVA) was conducted to investigate the interaction between emotion and audience figure. The familywise error rate was controlled by the Shaffer-Hohn procedure with a = .008 (i.e., .05/6). The results were significant for the following expectancies: negative interpersonal response, F(l, 123) = 66.48, p = 000, instrumental consequences, F(l, 123) = 1 1.54,

PAGE 86

Table 22 Univariate ANOVA Table for Emotion by Audience Interaction on Outcome Expectancies Expectancy Sum 01 Squares df Mean Squares F Interpersonal Support 1 07 J.Z / 1 1 "J J.Z / 0 /in 1 "7/1 Error 1 AT lo/.oo Neg. Interpersonal Response 00.04 1 1 00.04 AA /IC DO. 45 nnn .uuu Error 1 AA TO 1 Instrumental Consequences 1 ^ TA 1 J. /O 1 1 1 < TA 1 J. /O 1 1
PAGE 87

79 To further investigate differences due to the audience figure and emotion interaction, simple main effects ANOVAs were conducted for each audience figure separately with a = .025. Descriptive statistics for the emotion by audience interaction appear in Table 23. The follow up ANOVA results for the interaction are presented in Table 24. The analyses revealed a significant difference between audience figures on the negative interpersonal response expectancy, F(l, 123) = 86.18, p = .000. Students anticipated more teasing and negative behaviors from their friend(s) than from their teachers for expressing their anger. There was no significant difference in student^ scores for the teacher audience figure, F(l, 123) = .02, p = .899, for expression of sadness. Although students did not believe that friends would punish them for expression of either anger or sadness, F(l, 123) = .70, p = .658, the students were less certain about their teachers' responses, F(l, 123) = 28.45, p = .000. The students teiKied to respond that teachers definitely would not punish their expression of sadness, but teachers probably would not punish their expression of anger. Finally, the analysis revealed a significant difference in student^ expectancy regarding the audience figured emotional reaction to their expression of negative emotioiL Students indicated that expressing their sadness, but not their anger, to a friend would probably upset the friend, F(l, 123) = 12.74, p .000. There was no significant difference between emotion types on this expectancy for teachers, F(l, 123) = .60, p = .447. These results only partially support the hypothesis that adolescents? expectancies would vary as a function of emotion type and audience figure. Expectancies varied for friends as a fimction of emotion type, but did not vary for teachers.

PAGE 88

80 Table 23 Descriptive Statistics for Emotion by Audience Interaction on Outcome Expectancies Expectancy Audience Emotion Mean N Neg. Interpersonal Response Friend Instrumental Consequences Emotional Reaction Friend Teacher Friend Anger Sadness Teacher Anger Sadness Anger Sadness Anger Sadness Anger Sadness Teacher Anger Sadness 8.42 6.75 6.73 6.75 6.36 6.29 7.07 6.28 7.71 8.24 8.12 8.00 129 129 129 129 129 129 129 129 129 129 129 129

PAGE 89

Table 24 Results of ANOVA for Emotion by Audience Interaction on Outcome Expectancies 81 Source Sum of Squares df Mean Squares F E Neg. Interpersonal Response i Friend ly/.y/ 1 1 ly/.y/ O/^ 1 0 AAA Error 1 TO Z.UZ Teacher 1 O Ay 1 1 o f4T OQO .oyy Error 1 i?T 1 1 1 /IT Instrumental Consequences Friend /IT .4/ 1 /IT .4/ . /U .OJO Error 1 /U.U3 1 TO 1 11 l.ii Teacher 1 /IT "71 42. /i TO /I C Z0.45 AAA .000 Error 165.77 128 1.30 Emotional Reaction Friend 22.39 1 22.39 12.74 .000 Error 173.61 128 1.36 Teacher 1.40 1 1.40 .60 .442 Error 175.10 128 1.37 Note, a = .008.

PAGE 90

Summary of Results The first research hypothesis proposed that adolescents' decisions to express or dissemble their anger or sadness would vary as a function of the type of audience figure, type of emotion, age (grade), and sex of the participant. The type of audience figure was the only variable for which a significant difference emerged. This main effect was modified by significant interactions between audience figure and sex, and audience figure and grade. There were no other significant interactions on the first question. Both male and female students reported that they would be more likely to express their negative emotions to a best fi-iend rather than to a teacher. Female students, however, were much more likely than male students to report they would express their negative emotions to a friend rather than to a teacher. Males indicated they would be more likely to express their anger or sadness to a friend than to a teacher, but the difference between the two was minimal. Fifth-grade students were almost equally inclined to report they would e)q)ress negative emotion to either a best friend or a teacher they liked. Eighth-grade students were least likely to express negative emotion to a teacher. In fact, there was a noteworthy decrease between fifthand eighth-graders' scores on displaying negative emotion to a teacher. Eleventh-grade students also indicated that they would be disinclined to display negative emotion to a teacher. The situation was reversed for the best-friend condition. With age, students indicated that they would be more likely to display their anger or sadness to a friend. The second research question investigated self-efficacy and predicted that adolescents would report greater self-efFicacy for regulating sadness than anger. There

PAGE 91

83 was no difference in self-efficacy between emotion types so the hypothesis was not supported. There was, however, an increase in reported self-efficacy with age. Eleventhgrade students were far more likely to report that they would be able to control the display of their negative emotions than were fifth-graders. Self-efficacy was also influenced by an interaction with the type of audience figure and sex. Female students indicated that they would be able to control their negative emotion nwre with a teacher than with a best fi"iend. Male students did not differentiate between the two audience figures and indicated that their ability to regulate their anger or sadness was not necessarily dependent on whether they were in the presence of either a teacher or bestfriend in the school context. The third hypothesis predicted that the outcome expectancy scores would vary as a fimction of emotion type, audience figure, age, and sex of the participant. Significant differences did result for emotion type, grade (age), and audience figure, but each of these variables interacted with another. There was no main effect between sexes on any of the expectancies. Audience by sex interactions emerged for interpersonal support, norm maintenance, and internalization. An audience by grade interaction occurred for norm maintenance. Last, an emotion by audience interaction was found for negative interpersonal response, instrumental consequences, and emotional reaction. There were no significant three-way interactions or the predicted age by sex interaction. Interpersonal Supp ort Females expected their best friend would be more supportive and imderstanding of the depth of their negative emotion than would a teacher. Males did not differ

PAGE 92

84 significantly between audience figures for interpersonal support. There were no significant differences by grade for this expectation. Negative Interpersonal Response Students expected more teasing and negative behaviors from their best friend for displaymg their anger than they anticipated from their teacher. There was no significant difference between the two audience figures for expression of sadness. Instrumental Consequences Students did not expect to receive negative consequences or punishment from their teacher(s) for expressing their anger or sadness in the school context. Furthermore, there was no significant difference between emotion types on expressing emotion with a best friend. Students indicated that they would probably not receive a negative response for expressing either emotion to a friend. Emotional Reaction Students indicated that revealing their sadness to a friend would probably make the friend feel bad. This was not true for revealmg anger to a friend. Students beUeved that displaying anger would not upset the friend. There was no significant difference between emotion types for expressing negative emotion to a teacher. Students believed that displaying both anger and sadness would be likely to upset the teacher and cause an emotional reactioa Norm Maintenance Female students believed that they definitely should show their emotion to a best friend rather than the teacher. Males did not make that distinction between audience figures. Fifth-grade students reported that they should express their anger and sadness to

PAGE 93

85 both a best friend and the teacher. Eighth graders were least likely of any other group to report they should display their negative emotion to a teacher. Eleventh-grade students indicated that they believed that they should reveal their negative emotion to a best friend, but not to a teacher. Internalization Female students believed they would feel worse if they did not express their negative emotion to a friend. They also indicated that they would feel better if they did not display their emotion to a teacher. Male students expected to feel better if they did not reveal their emotion to either a best friend or a teacher.

PAGE 94

CHAPTER 5 DISCUSSION This study examined adolescents' decisions about displaying or dissembling negative emotion in the school context in the presence of two audience figures, a best Mend and a well-liked teacher. The purpose of the research was to examine developmental changes in adolescents' reasoning in emotion decisions to dissemble or display negative emotion from preadolescence (5th grade) through middle adolescence (8th grade) to late adolescence (11th grade). I also examined developmental changes in adolescents' self-efficacy beliefs about emotional regulation and their outcome expectancies of emotional display for themselves and two audience figures; a close friend and a teacher, in the school context. This chapter presents a discussion and interpretation of the results of the study as they relate to the original research hypotheses. It also addresses the theoretical and practical implications of the findings. The chapter concludes with recommendations for future research and limitations of the study. Adolescents' Emotion Management Decisions Hypothesis 1 — Decisions to Express Emotion The first research hypothesis predicted that adolescents' decision to express or dissemble negative emotion would vary as a function of the type of audience figure, the type of emotion, age, and sex of the participants. It was also expected that with age. 86

PAGE 95

87 adolescents' decisions to regulate their negative emotion would interact with the sex of the participant, with emotion type, and with type of audience figure. The only significant main effect that emerged was for audience figure. Male and female participants at all grade levels indicated that they would be more likely to express their negative emotions to a best fi-iend rather than to a wellliked teacher. This finding, however, was moderated by an audience figure by grade interaction and an audience figure by sex interaction. Follow-up analyses of the audience figure by grade interaction revealed that 5th graders were as likely to report that they would express negative emotion to a well-liked teacher as to a fiiend. In contrast, 8th graders and 1 1th graders were less likely to report they would display negative emotion to a teacher than to a fi-iend. This finding is consistent with other developmental research that has shown that with age, adolescents prefer close fi-iends rather than adults as confidants with whom they emotionally selfdisclose (Belle & Longfellow, 1984; Bemdt, 1982; Hartup, 1996; Papini et al, 1990). The research literature suggests alternative hypotheses that might account for this fitnding. The difference could be attributed to developmental changes in social cognition regarding appropriate expression of negative emotion (Saami, 1979, 1988, 1989a, 1990, 1999), or it might indicate that the climate of middle schools discourages negative emotional display with adults who hold positions of authority (Eccles & Midgley, 1989; Eccles et al., 1996), or both of these &ctors may have influenced the developmental effect. This finding is consistent with previous research on emotion management during adolescence (Gnepp & Hess, 1986; Fuchs & Thelen, 1988; Saami, 1988; Zeman & Garber, 1996, Zeman & Shipman, 1997). Adolescents strive to present themselves in a

PAGE 96

88 socially desirable manner for self-protective reasons. Furthermore, it supports Saami's (1988, 1990, 1999) contention that with age, individuals master cultural scripts for the expression of emotion as appropriate or not suitable in particular social contexts. Follow-up analyses of the audience figure by sex interaction revealed that females were more likely than males to report that they would express negative emotion to their best friend rather than to a well-liked teacher. This gender difference in emotion disclosure lends support to the literature on gender differences in adolescent friendships in general, but more particularly, the more intimate nature of adolescent female friendships (Bemdt, 1982; Buhrmester & Furman, 1987, Hartup, 1993, 1996). The finding is also consistent with previous research on gender differences in emotional expression (Bimbaum & Croll, 1984; Brody, 1985, 1993; Brody & Hall, 1993, 2000; Fabes & Martm, 1991; Shields, 1995). The failure to find an emotion by sex interaction was surprising in that researchers have generally reported that males are more likely to express anger rather than sadness. A trend in that direction was evident (p = .07), however, suggesting that with a larger sample this difference may have emerged. Summary. An interesting developmental fmding emerged from the analysis of the first hypothesis. Younger children were as likely to express negative emotion to a teacher as to a best friend, and older adolescents preferred to disclose their negative emotion to a best friend rather than to a teacher. In addition to the developmental findings, an interesting interaction between gender and audience figure was obtained; that is, females were more likely than males to discbse negative emotion to a friend.

PAGE 97

89 Hypothesis 2 — Self-efficacy Beliefs The second research hypothesis predicted that adolescents would report greater self-efficacy for regulating sadness than for regulating anger; however, no difference for self-efficacy by emotion type emerged. Therefore, the hypothesis was not supported. This fmding contrasts with Zeman and Shipman's (1997) research. In their study, adolescents indicated a belief in their ability to control expressions of sadness significantly more than expressions of anger. A possible explanation for the lack of agreement might be the content of the vignettes in the current study. Even though the scenarios were developed from adolescents' self-reports of experiences in school that triggered anger and sadness, the situations described m the vignettes were commonplace occurrences that might not have triggered sufficient intense emotion. Intensity of emotion has been cited as a determining factor in children's and preadolescents' beliefs about their ability to control expression of negative emotion. Younger children, in particular, have reported that they would be unable to control their anger if they were sufficiently provoked and deemed the situation important enough to express their displeasure (Saami, 1988, 1999). Although significant differences in the ability to control sadness and anger did not emerge for hypothesis 2, students' beliefs in their ability to control their emotions did increase with age. Fifth graders were significantly less likely to report that they could control their negative emotion with a teacher or a best friend than were 1 1th graders. Nonetheless, there was no difference between the 5th graders and 8th graders or between the 8th graders and 1 1th graders on this question. Saami (1990, 1999) has argued that, with age, emotionally competent individuals become more adept in their ability to

PAGE 98

90 regulate emotional expression. The failure to find a significant difference between 5th and 8th graders and 8th and 1 1th graders suggests a need for further study of this developmental trend. A significant interaction between audience figure and sex indicates that females perceive themselves as better able to regulate their emotion with a teacher than with a best fi-iend, whereas males do not distinguish between a fi-iend and a teacher. As stated previously, the development of intimate relationships with best friends is an integral feature of adolescence for most young females. These relationships are built on mutual trust, reciprocity, and the ability to share genuine emotions without negative repercussion (Hartup, 1996). Furthermore, both parties hold relatively equal power. Conversely, teacher-student relationships are unequal, unilateral, and characterized by an imbalance of power. Teachers have the ability to embarrass students in a very public social arena. Consequently, a female adolescent's belief in her ability to successfully regulate her genuine emotion with a teacher could become a valuable self-protection strategy. Another possible explanation for this finding might be found in the literature on adolescent females' development. Brown and Gilligan (1992) noted that some females tend to go underground in early adolescence. That is, they resort to silence and dissemblance of their true feelings so as not to stand out in the classroom. Interpretation of this finding is made with caution due to the paucity of research on emotional self-eflficacy. Saami (1999), who has written extensively on the subject, contends that self-efificacy in emotion management occurs "when we emerge from an emotion-eliciting encounter with a sense of having accomplished what we set out to do" (p. 3). In other words, the individual has attained the requisite capacity and skills to efiect

PAGE 99

91 a successful emotional transaction. That female students in the current research reported that they could control their negative emotion more effectively with a teacher than with a best friend may have more to do with impression management than with self-efficacy. A possible explanation for the lack of difference by emotion type on the selfefficacy question is that participants were not personally emotionally involved in the vignettes. Even though they were asked to imagine themselves as the protagonists in the stories, the situations were hypothetical. As Eisenberg and Zhou (2000) have recently argued, "emotional responses to actual events are likely to be much more compelling and meaningful than mild [inductions] that have Uttle relevance to an individual's Ufe" (p. 168). Consequently, participants' actual behavior in a situation in which a personal goal is at stake, may have had little bearing on their responses on the questionnaire. Perhaps students' self-efficacy responses would have differed by emotion type if the situations had been more personally meaningful for them. Summary. Adolescents did not report greater self-efficacy for regulating sadness more than anger; therefore, the second hypothesis was not supported. There was, however, an increase in reported self-efficacy beliefs with age. Eleventh-grade students reported more confidence in their ability to regulate negative emotion than did 5th-grade students. In addition to the developmental finding, an interaction between gender and audience figures occurred; that is, females indicated they would be more likely to control their negative emotion with a teacher than with a best friend. Males did not differentiate between the two audience figures in their self-efficacy beliefs.

PAGE 100

92 Hypothesis 3 — Outcome Expectancies The third research hypothesis predicted that adolescents' responses on the outcome expectancy questions would vary as a function of emotion type, audience figure, age (grade), and sex of the participants. It further predicted interactions between audience figure and sex, audience figure and grade, and sex and age of the participants. Significant differences did emerge for emotion type, grade (age), and audience figure, but each of these variables interacted with another. There were no significant sex differences for hypothesis #3. However, significant interactions resulted between audience figure and grade, and audience figure and sex. There was no significant interaction between sex and age. The following discussion will focus on each of the outcome expectancies, individually. Interpersonal support. Females anticipated a more supportive and understanding response to their emotion display fi-om a best Mend than fi-om a well-liked teacher. This finding is consistent with research (Belle et al., 1987; Buhrmester & Furman, 1987) that has found that in times of distress, females turn to close friends for comfort and emotional support. It also supports the findings of Zeman and Shipman (1997) on sex differences in expectations of interpersonal support. Males, in contrast, did not differ significantly in their expectancy for support fi-om either audience figure. Surprisingly, yoimger adolescents did not differ significantly fi-om late adolescents in this expectancy. This finding is in contrast with the research of Fuchs and Thelen (1988). They found that older children (especially males) anticipated less positive consequences as a result of their negative emotional expression than did younger children.

PAGE 101

93 Zeman and Shipman (1997) also found age differences in adolescents' expectancies for interpersonal support. Fifth and 1 1th graders in their study anticipated a supportive response from their mothers for expression of negative emotion. Eighth graders were the least likely to expect support for emotional display. It is conceivable that the lack of age differences in the current research for expectations of support was influenced by school socialization practices. Appropriate student behavior and conduct are emphasized daily in most elementary schools. Negative emotional displays are discouraged; prosocial behaviors are promoted. By fifth grade, most students are keenly aware of proper school behavior and the consequences for disciplinary infractions. Taken together, this pattern of school conduct and teacher practices could have influenced students' responses for expectations of interpersonal support. Negative interpersonal response. Students expected more teasing, belittling, and ridicule from their friends for displaying their anger than they did from a teacher. There were no differences on this expectancy between audience figures for expressing sadness. It is puzzling that gender effects were not found for this expectancy. For example, Zeman and Shipman (1997) found that males expected more negative interpersonal responses from their friends than from either parent. Other research has also reported sex differences in this outcome expectation. Fuchs and Thelen (1988) found that males anticipated negative reactions for expressing sadness, especially from their fathers; females, on the contrary, expected more support and fewer negative consequences for genuine expression of both anger and sadness with their parents than did males.

PAGE 102

Research on conflict in adolescents' peer relations (Laursen & Collins, 1994) has found that in particular social contexts (Le., closed-field experimental settings), adolescents anticipate more negative behaviors from their friends than they would if contextual constraints were removed. The social setting, in part, can influence adolescents' behavior in relationships with friends or non-parental adults. The majority of students who participated in the current research attended a local university developmental research school. The number of students in prekindergarten through Grade 12 totals less than 1,000. Classes are normally small (i.e., 25 students or less) and the school asserts that it provides a warm and supportive environment for its teachers and students. The warmth of the academic environment has been cited as an instrumental ^tor in forging successful teacher-student relationships during adolescence (Eccles et al., 1996; Eccles & Midgley, 1989; Wentzel, 1997). It may also enhance students' trust in their teachers' responses to emotional display. Supportive teachers and administrators may contribute to a school climate that fosters an increase in students' self-esteem so that minor instances of teasing or belittling from friends are safely tolerated (Harter, 1996). Instrumental consequences. Students did not anticipate punitive consequences from a teacher for expressing either anger or sadness in the school context. Although the analysis revealed a significant difference between audience figures and between emotion types for a teacher, the findings were not in the anticipated direction. Moreover, students did not expect to be ostracized or shunned by their friend for openly expressing either emotioa This finding is surprising given the recent focus on student conduct codes in

PAGE 103

95 schools, the heightened consequences for violating school conduct rules, and the school district's mandate for student safety. A related study by Underwood et al. (1992) found that students (third, fifth, and seventh graders) would express anger more with a peer than with a teacher. By seventh grade, students recognized that it was "more dangerous and less socially acceptable to [express] anger toward authority figures than peers" (p. 376). Zeman and Shipman (1997) reported an emotion by audience interaction affecting emotional disclosure expectancies. In that study, however, adolescents anticipated fewer punitive responses from fathers than they did from mothers or peers for emotional displays of anger rather than sadness. In the current study, there were no significant differences among audience figures for expression of either negative emotion. It is difficult to reconcile Zeman and Shipman' s fmdings with those of the current study because the audience figures were dissimilar. A possible explanation for participants' lack of distinction between audience figures or emotion type could be the social context and students' interpersonal relationships with the teacher. This is consistent with a functionalist perspective regarding emotional development (Barrett & Campos, 1987) and with the research on emotion management during adolescence (Fuchs & Thelen, 1988; Whitesell & Harter, 1996; Zeman & Shipman, 1997). This particular school's climate may have influenced students' responses such that they did not anticipate a punitive response from a well-liked teacher who was a bystander, not the emotional provocateur. As previously stated, the school is characterized as providing a warm and supportive atmosphere that is meant to enhance teacher-student relationships and students' well being. Students may have

PAGE 104

perceived that it would be safe to emotionally disclose to the teacher if they chose to do so. Furthermore, teacher-student relationships may be stronger and more caring at this school than at many others. This finding could encourage teachers to form good relationships with their students so that they may act as a "safety valve" when students are upset with someone else. Emotional reaction. Students reported that expressing sadness to a friend would be more likely to make the friend feel bad than expressing anger, but they did not make this distinction with a teacher. Zeman and Shipman (1997) found a significant emotion by audience by grade interaction, but they cautioned readers to interpret the interaction with caution because of the small sample size per cell. It is conceivable that the participants in the Zeman and Shipman study were more homogeneous than participants in the current study. Participant demographics and geographical factors may, in part, account for this difference. Zeman and Shipman also found that adolescents' expectations for emotional reaction varied as a function of the audience figure (i.e., best friend, mother, and father) and that this result was moderated by the age of the participant. I found no significant age (grade) effects for this expectancy, but emotion decisions to express anger and sadness varied as a fimction of audience figure. The finding that emotion decisions to express anger and sadness varied as a function of audience figure is consistent with Saami's (1990) contention that the ability "to take into account multiple frames of reference in making good guesses about how someone else is likely to react emotionally is strategic and fimctional" (pp. 160-161). It is strategic because it facilitates social transactions. It is functional because it "promotes

PAGE 105

97 relationship bonds and depth of communication" (p. 161). This skill is a precursor to another milestone of emotional competence; that is, the abihty to understand that one's emotional expressive behavior can affect another individual and to use this information as a guide in one's self-presentation strategies. According to Saami (1990, 1999) individuals can promote self-efificacy in emotionally provocative situations by strategically managing their emotional expressive behavior so that the other will respond in a desirable manner. Students in the current study demonstrated this awareness by making a distinction between the emotional reactions of a friend and a teacher to their expression of sadness or anger. Norm maintenance. Female students were more likely to report that they should display their negative emotion to a friend than to a teacher. Males did not make a distinction between audience figures and reported they were equally likely to express negative emotion with either a friend or a teacher. Zeman and Shipman (1997) did not find this effect; rather they found that males reported having rules about not expressing emotion more than did girls, and they found that with fethers and friends, children indicated that they should not express sadness more than anger. The finding that females distinguished between a teacher and a friend in expressing emotion is consistent with female adolescents' commitment to the strength of aflRliation in relationships (Hartup, 1993, 1996), but it is inconsistent with the aduk literature on the display of anger (Averill, 1982, 1983; Brody, 1993; Brody & Hall, 1993, 2000; Miller, 1991; Shields, 1995). Whereas females are generally considered the more emotionally expressive sex, research reveals that males are far more Ukely to express anger publicly than are females. Males are also less Ukely to openly express sadness for

PAGE 106

98 fear of exposure and other negative reprisals. That the male students in the current research did not distinguish between anger and sadness may be an indication of social maturity or a trend toward less conformity to masculine stereotypes for sex role behavior. Age differences in the understanding of display rules (Le., norm maintenance) also appeared. Fifth graders' responses indicated that they did not distinguish between a friend and a teacher in expressing their negative emotions. In contrast, 8th and 1 1th graders indicated they should reveal negative emotion to a friend, but not to a teacher. These findings are consistent with previous research on the use of display rules in a school context (Underwood et al., 1992). Internalization. A sex by audience interaction emerged for this expectancy. Females reported they would feel worse if they did not express their emotions to a friend, whereas they believed they would feel better if they did not express their anger or sadness to a teacher. Males, on the other hand, indicated they would feel better if they kept their negative emotions to themselves and did not express either emotion to a friend or a teacher. This fmding is somewhat consistent with Zeman and Shipman's (1997) results. They also found sex differences in internalization, but there was no interaction of sex with the audience figure. Saami and von SaUsch (1993) offered four reasons or motivations for emotional dissemblance. The first is to avoid negative outcomes. In the current research, it is conceivable that females believed that emotional dissemblance with a best friend might threaten their relationship. In contrast, males might wish to protect their masculine relationships by dissembling their emotions at school.

PAGE 107

99 The second reason for dissemblance is to protect one's self-esteem or to cope more eflfectively with how one feels. Male adolescents indicated they would feel better if they dissembled their negative emotion. They could have considered this an effective coping strategy or a method of maintaining positive self-worth. Females reported only the intent to dissemble emotion with a teacher. Again, this might be an effective coping strategy or a way to enhance their self-esteem at school The third reason concerns the norms and conventions of the social context. As previously stated, teachers generally discourage the display of negative emotion. Both male and female adolescents demonstrated an understanding of the appropriate cultural display rules in the school context. According to Saami and von Salisch (1993), the fourth reason for emotional dissemblance is to regulate relationship dynamics. This motivation could apply equally to both male and female students. Female adolescents are known to value close relationships and tend to be concerned about issues of trust and genuine emotional display. Gilligan, Lyons, and Hammer (1990) have written that "making connections," especially establishing strong interpersonal relationships with female friends, is a core con^onent of adolescent females' development of the self Conversely, male adolescents may have believed that expressing their emotion might change an existing relationship dynamic with either a friend or a teacher. This interpretation is speculative, of course. Nonetheless, it represents an initial attempt to extend our understanding of the dynamics of adolescents' emotion management decision making. Summary. Several predicted interactions occurred for the third hypothesis; that is, the MANOVA on adolescent outcome expectancies revealed audience by sex

PAGE 108

100 interactions, audience by grade interactions, and an emotion by audience interaction. Gender differences emerged for expectations of interpersonal support, norm maintenance, and internalization. That is, females anticipated more supportive and understanding responses from a best friend rather than a well-liked teacher. Females indicated they should show their negative emotion to a friend rather than to a teacher, and they reported that they would feel worse if they did not disclose their negative emotion to a friend. The reverse was true for displaying negative emotion to a teacher. Males did not differ significantly between audience figures on expectations for interpersonal support or norm maintenance. Moreover, they reported they would feel better if they did not disclose their emotion to either audience figure. Developmental findings emerged for the norm maintenance expectatioa Fifth graders did not make a distinction between audience figures about display rules for expressing negative emotion. In contrast, eleventh graders indicated they should reveal their negative emotion to friend rather than to a well-liked teacher. Students did not anticipate a negative instrumental consequence (i.e., some form of punishment) from their teacher for expressing either emotion. They did, however, indicate that displaying anger to a teacher would upset him or her and cause an emotional reactioa The opposite was true for displaying emotion to a friend. Students believed that expressing sadness, but not anger, would upset the friend. Theoretical Implications The fiinctionalist perspective on emotion regulation contends that individuals' decisions to express or dissemble their emotions are embedded in an extensive socialcultural web of emotion beliefs, internalized cultural display rules, and cultural

PAGE 109

101 expectancies about the outcomes of expressing or regulating emotion behavior in a transactional relationship (Barrett & Campos, 1987; Campos et al., 1984). The individual cannot be separated from his or her sociocultural environment or social context. Emotions are not seen simply as intrapersonal feeling states. They are the products of a relationship between individuals; a social collaboration between individuals within a shared system of beliefs and expectancies. Functionalist theory enq)hasizes the ad^tive qualities of emotion and the significance of the person-environment transaction. From this perspective, individuals will attach different meaning to emotion events and respond in accordance with their personal goals. The participants in this study demonstrated these fimctionalist principles. Students responded to the vignettes in ways that demonstrated a personal understanding of the display rules implicit in school behavioral socialization. Their responses indicated an awareness of the different display rules for interpersonal relationships and for the outcomes of emotion expression for themselves and another person (Le., the audience figure). Their emotion management decisions were adaptive to the presence of the audience figure, type of negative emotion evoked by the vignettes, and the significance of the person-environment transaction. Most fimctionalist theorists believe that emotion regulation is a two-fold process. It can pertain exclusively to the self-management of internal feeling states, or it can be construed as the management of emotional reaction of others. The participants in this study indicated a desire to maintain an hedonic emotional outcome for themselves, but they also demonstrated an awareness that display of negative enK)tion could diaiipt the hedonic state of another person. Clearly, they cared about upsetting the emotional balance

PAGE 110

102 in their interpersonal relationships. In accordance with the functionalist perspective, their personal goals were intimately entwined in the decision making process and gave meaning to their choices to display or dissemble negative emotion. To summarize, this study lends support to the functionalist perspective on emotion and emotion regulation during adolescence. It offers evidence that emotion regulation is a bidirectional, transactional, relationally oriented process that is influenced by the significance to the individual, his or her particular goals, and the social context in which the emotional event occurs. Practical Implications It has been widely reported that the majority of middle school and high school teachers have had minimal exposure to adolescent development courses or theory. This study provides a usefiil framework for understanding adolescent decision making about when and with whom they will regulate negative emotion at school. Although adolescents' reasons for expressing or dissembling negative emotion were not examined in the current study, previous research suggests that self-presentation, impression management, and self-esteem goals are responsible, in part, for their responses. Teacher education courses in adolescents' emotional development during middle school and high school could be improved if the current findings are supported by other research and if they are incorporated into the teacher education curriculum. Second, this study provides some initial evidence that adolescents are hesitant to express their genuine emotions to an authority figure, even a teacher they like and respect. This finding has practical importance for middle school and high school counselors who encourage teenagers to express their anger and sadness, especially in

PAGE 111

/ 103 light of recent school shootings. The developmental trends that were identified in the current research indicate that, with age, students are less inclined to express negative emotion to a teacher and more likely to confide in a friend of the same sex. Students may be reluctant to inform teachers, counselors, or school administrators about situations that may cause discomfort or anguish for themselves or others. Third, this study augments our understanding of social contextual fectors that conflict with adolescents' developmental needs. As previously stated, there has been an increase in the nimiber of disciplinary infractions and the escalation of students' display of negative emotion in middle and secondary schools. As students progress educationally, the organization and climate of the schools become more repressive and imdemocratic (Eccles & Midgley, 1989; Eccles et al., 1993; Wenz-Gross et al., 1997). Students' relationships with teachers and administrators tend to deteriorate. Moreover, high school students may become increasingly angry or saddened by developmentally inappropriate or rigid school and classroom rules. The student-generated responses to the open-ended pilot study questions support this contention. Seventeen percent of 5th graders described situations in which unfair rules were the basis of their anger. By 1 1th grade, the response rate for the unfair rules category escalated to 33% of the total anger situation responses. Even though the student responses in the pilot study were limited to a small sample and may not be representative of the broader secondary school population, secondary educators and administrators would be wise to develop school rules and codes of conduct in conjunction with their student populations.

PAGE 112

104 Recommendations fo r Future Research The results of this study contribute to our understanding about adolescents' emotion management decisions in schools, but it is an initial exploration of a topic that has broad implications for emotional development theory and educational practice. Future research should investigate the cognitive processes adolescents use in making decisions about dissembling or expressing negative emotion in schools. Ideally, this examination would encompass the reasons and methods they employ as well as their outcome expectancies for displays of negative emotion. Further research into gender differences in emotional expression and outcome expectancies should also be pursued. Although the current study did not include the influence of culture or ethnicity as a variable, fiirther investigation of their roles as moderating forces in the expression of negative emotion should not be overlooked. Most of the previous research that has focused on the consequences of adolescents' emotional expression has examined differences between the outcome expectations of parents and peers. Future research should investigate the relationship between adolescents and parents as well as adolescents and teachers for similarity of response on outcome expectancies. It is difficult to generalize the results of this study based on the lack of available research of adolescents' emotion management decisions with teachers and other authority figures. Finally, ecologically vaUd assessment instruments need to be developed for future research in adolescents' emotion decision making. The extant research is based on several researcher-developed questionnaires that bear little resemblance to each other. Ideally, new assessment measures should include the perspective of multiple informants.

PAGE 113

105 employ a variety of methodologies (i.e., a combination of self-reports, observations, and physiological measures), and take the influence of social context into account (Flannery, Torquati, & Lindemeier, 1994). The questionnaire in the current study may have some weaknesses that reduced the validity of the participants' responses. For example, the length of the questionnaire and the uniform format of the questions may have a created a response bias for some participants. In addition, the scenarios may not have been equally appropriate for each of the three diverse age groups. Development of parallel forms of an ecologically valid instrument that more accurately reflect the emotional concerns of late childhood, middle adolescence, and late adolescence is highly recommended. Limitations of the Study The findings of this research extend what we know about adolescent emotion management decisions, but they should be interpreted with caution. Self-reports have many advantages including the efficient use of time and ease of administration and scoring. Moreover, because they reflect the participant's own perspective, they are frequently used to gain the insider's view of what an adolescent is feeling. This insider's perspective, however, is also one of the greatest disadvantages of self-report mstruments and deserves reiteration here. First, participants in this research may have responded in socially desirable ways to enhance their self-image or to conform to social stereotypes for expected behavior. Self-reports are subject to individual bias. Second, participants may have responded to the questionnaire with the specific purpose of pleasing the researcher. Third, and perhaps most important, participants' responses may not reflect their actual behavior.

PAGE 114

106 Discrepancies between behavior in a hypothetical situation and behavior in a real situation are well documented. In short, self-report methodology suffers from ecological validity concerns. Even though the vignettes in the dissertation study were developed from local 5th-, 8th-, and 1 Ith-grade students' responses regarding school situations that evoked anger and sadness, their responses should not be generalized to a broader population. Demographic and geographic differences have been known to influence students' responses and to affect the external validity of a research instrument. The use of an instrument with limited evidence of reliability and validity in conjimction with a small sample size further limits the ^plication of these findings to a broader population. Finally, the teacher and best friend described in the vignettes did not provoke the emotional distress. Consequently, one cannot assume that the students would report that they would respond in a similar feshion if the teacher or friend were the provocateur. Future research should investigate adolescents' emotion management decisions when the teacher or another school official is the causal agent of students' negative emotion. Summary These findings add support to a functionalist perspective that decisions to express or dissemble emotions are relationally oriented, contextiially embedded, and situationally adapted to the goals of the individual. The findings also contribute to our understanding about (a) gender differences in emotional expression, and (b) developmental research indicating that students' beliefs in their emotional self-efficacy, and their tendency to confide in a best friend rather than an adult, increase with age. Future research should include comparisons of parents, teachers, and best friends and the influence of culture on adolescents' emotion management decisions.

PAGE 115

APPENDIX A PILOT STUDY QUESTIONNAIRE Student Data Sheet Please provide the following information. Your responses will be kept totally confidential. DO NOT write your name on this sheet or identifi^ yourself in any way other than to answer the following questions: Please check the correct box for items 1 , 2, and 4. 1 . I am: male I I female I I 2. I am in the: 5th I I 8th I I 1 1th I I grade. 3. My birthdate is: (Month) (Day) (Year) 4. My race/ethnicity is: American Indian, Eskimo or Aleut I I Asian or Pacific Islander I I Black/African-American | | Hispanic | | White (non-Hispanic) | | Multi-ethnic I I Other Please describe 107

PAGE 116

108 Pilot Study Questionnaire la. Please think about and write down something that recently happened to you at school or at a school -related event that made you feel angry. What happened? lb. How angry were you? I would say that I was: a little annoyed Please check the box that best describes the way you felt, feirly angry very angry extremely angry

PAGE 117

109 Pilot Study Questionnaire 2a. Please think about and write down something that recently happened to you at school or at a school -related event that made you feel sad. What happened? 2b. How sad were you? I would say that I was: a little sad Please check the box that best describes the way you felt, feirly sad very sad extremely sad

PAGE 118

110 Pilot Study Questionnaire 3a. Please think about and write down something that recently happened to you at school or at a school -related event that made you feel happy. What happened? 3b. How pleased were you? I would say that I was: a little pleased Please check the box that best describes the way you felt, feirly happy very happy ecstatic

PAGE 119

APPENDIX B DISSERTATION INSTRUMENT Student Data Sheet Please provide the following information. DO NOT write your name on this sheet or identify yotirself in any way other than to answer the following questions: Please check the box that best describes you for items 1, 2, and 4. 1. I am male Q female O 2. I am in the 5th 8th 11th grade. 3. My birth date is (Month) (Day) (Year) 4. My race or ethnicity is Native American, Eskhno, Inuit, or Aleut Q Asian or Pacific Islander I I Black/African-American I I Hispanic I I White (non-Hispanic) I I Multi-ethnic I I Other Please describe 111

PAGE 120

112 Emotion Decision-Making Study Instructions This booklet contains stories about common situations in school that produce anger or sadness in students your age. Please read each story and try to imagine how you would respond if this happened to you. Sample Story --Bill was hoping to be chosen for a special program at school. He tried very hard to do well on all the things that mattered. When the principal announced the names of students who were selected for the program. Bill did not get picked and now he feels sad. Imagine that this situation happened to you. Your best friend, who is the same sex as you are, has heard what happened. Please think about it for a moment and then answer the following questions. There are no right or wrong answers to these questions. Please check the box that best matches the way you would really think, act, or feel. Sample Questions Deiiiiitely Probably Probably Definitely Not Not Would Would 1. Would you show how sad you feel to your best friend? n Definitely Probably Probably Definitely Not Not Could Could 2. Do you think you could control showing your sadness if you wanted to? Definitely Probably Probably Definitely Not Not Would Would 3. Do you think your best friend would be understanding of how sad that makes you feel? Definitely Probably Probably Definitely Not Not Should Should 7. Do you believe that you should show sad feeUngs to your best friend? Please turn to the next page and begin.

PAGE 121

113 Emotion Decision-Making Study Story 1Sam is sitting in class when the principal annoimces that the pep rally for the entire school has been cancelled because some students got mto a fight on school grounds. Sam thinks this is very unfair and it makes him feel angry. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has also heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how angry you feel to your best friend? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your best friend would be understanding of how angry that makes you feel? 4. Do you think your best friend would make fim of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your best friend; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your best friend? 8. How would you feel if you kept your angry feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Would Would 1 1 1 — 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 122

114 Story 1" Sam is sitting in class when the principal announces that the pep rally for the entire school has been cancelled because some students got into a fight on school grou nds. Sam thinks this is very unfair and it makes him feel angry. Again, imagine that this situation happened to you. This tune, however, a teacher you like has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how angry you feel to your teacher? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your teacher would be understanding of how angry that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how angry you really fek? 5. Would there be a negative response fi-om your teacher —would he or she punish you in some way — if you showed how angry you felt? 6. Would showing your anger upset your teacher; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your teacher? 8. How would you feel if you kept your angry feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could n 1 1 1 — 1 Definitely Probably Probably Definitely Not Not Would Would I.J Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 123

Emotion Decision-Making Study 115 Story 2 -Joe is walking down the school hallway between classes and he sees some kids picking on a nd making fun of a handicapped student. This makes him feel sad. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has also seen what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how sad you feel to your best friend? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your best friend would be imderstanding of how sad that makes you feel? 4. Do you think your best friend would make fim of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way — if you showed how sad you felt? 6. Would showing your sadness upset your best friend; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your best friend? 8. How would you feel if you kept your sad feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could 1 — 1 LJ LJ n LJ PrnlMhh/ Prahahlv Dcfinitdy Not Not Would Would 1 — 1 l_J 1 — 1 1 — 1 LJ 1 — 1 LJ Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Bener Better Please turn the page over and answer the next set of questions.

PAGE 124

116 Story 2 ~ Joe is walking down the school hallway between classes and he sees some kids picking on and making fun of a handicapped student. This makes him feel sad. Again, imagine that this situation happened to you. This time, however, a teacher you like h as seen what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation ... 1 . Would you show how sad you feel to your teacher? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your teacher would be understanding of how sad that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your teacher —would he or she pimish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your teacher; make him or her feel bad? 7. Do you believe that you should show your sad feeUngs to your teacher? 8. How would you feel if you kept your sad feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could l_J LJ LJ n LJ Definitely Probably Probably Definitely Not Not Would Would f— 1 LJ LJ n LJ rn LJ Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 125

Emotion Decision-Making Study 117 Story 3 -The teacher has just asked the class a question about the day's lesson. Megan raises her hand and gets called on. She answers the question incorrectly, and the teacher makes fiin of her and says her answer was "stupid." This makes Megan feel angry. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part 1. In this situation . . . 1 . Would you show how angry you feel to your best friend? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your best friend would be understanding of how angry that makes you feel? 4. Do you think your best friend would make fiin of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your best friend; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your best friend? 8. How would you feel if you kept your angry feelmgs inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could Definitely Probably Probably Definitely Not Not Would Would n Definitely Probably Probably Definitdy Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 126

118 Story 3 The teacher has just asked the class a question about the day's lesson. Megan raises her hand and gets called on. She answers the question incorrectly, and the teacher makes fun of her and says her answer was "stupid." This makes Megan feel angry. Again, imagine that this situation happened to you. This time, however, a teacher you like has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how angry you feel to your teacher? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your teacher would be understanding of how angry that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your teacher —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your teacher; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your teacher? 8. How would you feel if you kept your angry feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would L il Definitely Probably Probably Definitely Not Not Could Could 1 — 1 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Would Would 1 — 1 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 127

Emotion Decision-Making Study 119 Story 4 ~ Sally has just overheard one of her classmates tell other students that Sally is really "fat and ugly." It makes Sally feel sad to learn that other people are talking about her behind her back. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how sad you feel to your best friend? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your best friend would be understanding of how sad that makes you feel? 4. Do you think your best friend would make fun of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your best friend; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your best friend? 8. How would you feel if you kept your sad feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definit^y Probably Probably Definitely Not Not Could Could T\pfinitelv Probablv Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 128

120 Story 4 -Sally has just overheard one of her classmates tell other students that Sally is really "fat and ugly." It makes Sally feel sad to learn that other people are talking about her behind her back. Again, imagine that this situation happened to you. This time, however, a teacher vou like has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how sad you feel to your teacher? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your teacher would be understanding of how sad that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your teacher —would he or she punish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your teacher; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your teacher? 8. How would you feel if you kept your sad feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could r-~i 1 1 Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 129

Emotion Decision-Making Study 121 Story 5 ~ Pat is walking down the hallway at school and a group of students points at her and l aughs as she walks by. This behavior makes Pat feel angry. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has seen what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how angry you feel to your best friend? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your best friend would be understanding of how angry that makes you feel? 4. Do you think your best friend would make fim of you or tease you if you showed how angry you really feh? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your best friend; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your best friend? 8. How would you feel if you kept your angry feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could [—1 LJ [— 1 LJ rn i_i n 1 1 Definitely Probably Probably Definitely Not Not Would Would 1 — 1 1 — 1 LJ I—I LJ Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 130

122 Story 5 -Pat is walking down the hallway at school and a group of students points at her and laughs as she walks by. This behavior makes Pat feel angry. Again, imagine that this situation happened to you. This time, however, a teacher you like h as seen what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how angry you feel to your teacher? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your teacher would be understanding of how angry that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your teacher —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your teacher; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your teacher? 8. How would you feel if you kept your angry feelings inside and did not show them to your teacher? Definitety Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could 1 — 1 1 — 1 1 — 1 1 — 1 U Definitely Probably Probably Definitely Not Not Would Would 1 — 1 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 131

Emotion Decision-Making Study 123 Story 6 -Jennifer has just learned that a good friend has stolen her boy friend and all the kids at school knew about it except her. No one told her about it. This makes her feel sad. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how sad you feel to your best friend? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your best friend would be understanding of how sad that makes you feel? 4. Do you think your best friend would make fiin of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your best friend; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your best friend? 8. How would you feel if you kept your sad feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Nol Not Would Would Definitely Probably Probably Definitely iNOi IN 01 Definitely Probably Probably Definitely iNOl INOl WOUlU LJ n LJ LJ Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 132

124 Story 6 -Jennifer has just learned that a good friend has stolen her boy friend and all the kids at school knew about it except her. No one told her about it. This makes her feel sad. Again, imagine that this situation happened to you. This time, however, a teacher vou like has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part II. In this situation . . . 1 . Would you show how sad you feel to your teacher? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your teacher would be understanding of how sad that makes you feel? 4. Do you think your teacher would make fim of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your teacher —would he or she punish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your teacher; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your teacher? 8. How would you feel if you kept your sad feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would 1 1 1 1 1 1 1 n Definitely Probably Probably Definitely Could Could 1 1 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Would Would 1 — 1 1 — 1 r 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 133

Emotion Decision-Making Study 125 Story 7 Daniel is sitting in the cafeteria eating his lunch when someone throws food at him. The food stains his clothes and that makes him angry. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has seen what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part 1. In this situation . . . 1 . Would you show how angry you feel to your best friend? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your best friend would be understanding of how angry that makes you feel? 4. Do you think your best friend would make fim of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way — if you showed how angry you felt? 6. Would showing your anger upset your best friend; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your best friend? 8. How would you feel if you kept your angry feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could 1 — 1 f—l LJ LJ r-| LJ Definitely Probably Probably Definitely Not Not Would Would 1 — 1 1 — 1 LJ 1 — 1 LJ r— 1 LJ Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 134

126 Story 7 Daniel is sitting in the cafeteria eating his lunch when someone throws food at him. The food stains his clothes and that makes him angry. Again, imagine that this situation happened to you. This time, however, a teacher vou like has seen what happened. Please think about it for a naoment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how angry you feel to your teacher? 2. Do you think you could control showing your anger if you wanted to? 3. Do you think your teacher would be imderstanding of how angry that makes you feel? 4. Do you think your teacher would make fiin of you or tease you if you showed how angry you really felt? 5. Would there be a negative response from your teacher —would he or she punish you in some way ~ if you showed how angry you felt? 6. Would showing your anger upset your teacher; make him or her feel bad? 7. Do you believe that you should show your angry feelings to your teacher? 8. How would you feel if you kept your angry feelings inside and did not show them to your teacher? Definiteiy Probably Probably Definitely Not Not Would Would Definitely Probably Probably Definitely Not Not Could Could 1 — 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Would Would f 1 1 1 1 — 1 1 — 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please go on to the next story.

PAGE 135

Emotion Decision-Making Study 127 Story 8 ~ Larry studied very hard for his test and really thought he would get a good grade on it. ^en the teacher handed back his paper, however, Larry saw that he actually just passed the test with a "D." This makes him feel sad. Imagine that this situation happened to you. Your best or closest friend, who is the same sex as you are, has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part I. In this situation . . . 1 . Would you show how sad you feel to your best friend? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your best friend would be understanding of how sad that makes you feel? 4. Do you think your best friend would make fim of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your best friend —would he or she punish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your best friend; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your best friend? 8. How would you feel if you kept your sad feelings inside and did not show them to your best friend? Definitely Probably Probably Definitely Not Not Would Would Ueiinitely rroDaDiy rroDaDiy Definitely Not Not Could Could Definitely Probably Probably Definitely Not Not Would Would n 1 1 n 1 1 n 1 1 n 1 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Please turn the page over and answer the next set of questions.

PAGE 136

128 Story 8 -Larry studied very hard for his test and really thought he would get a good grade on it. When the teacher handed back his paper, however, Larry saw that he actually just passed the test with a "D." This makes him feel sad. Again, imagine that this situation happened to you. This time, however, a teacher you like has heard what happened. Please think about it for a moment and then answer the following questions. Please check the box that best matches the way you would really feel. Part n. In this situation . . . 1 . Would you show how sad you feel to your teacher? 2. Do you think you could control showing your sadness if you wanted to? 3. Do you think your teacher would be understanding of how sad that makes you feel? 4. Do you think your teacher would make fun of you or tease you if you showed how sad you really felt? 5. Would there be a negative response from your teacher —would he or she pimish you in some way ~ if you showed how sad you felt? 6. Would showing your sadness upset your teacher; make him or her feel bad? 7. Do you believe that you should show your sad feelings to your teacher? 8. How would you feel if you kept your sad feelings inside and did not show them to your teacher? Definitely Probably Probably Definitely Not Not Would Would n 1 1 n 1 1 Definitdy Probably Probably Definitely IN 01 All 1/1 1 1 Defiaiteiy Probably Probably Definitely I'lOl VTUUIU 1 1 n r 1 Definitely Probably Probably Definitely Not Not Should Should Definitely Probably Probably Definitely Worse Worse Better Better Thank you for participating in this study.

PAGE 137

APPENDIX C CONSENT FORMS Dear Parent/Guardian, I am a doctoral student in the Department of Educational Psychology at the University of Florida under the supervision of Dr. Patricia T. Ashton. I am conducting research on developmental changes in the emotion decision-making of fifth-, eighth, and eleventh-grade adolescents. The purpose of this study is to examine how decisions to express or conceal emotions change during this age range. The results of this study may help other researchers and teachers better understand adolescents' school behaviors and emotional responses and allow them to design instructional practices accordingly. These results may not directly help your child today, but may benefit fiiture students. With your permission, I would like to ask your son or daughter to volimteer for this research. Each student will be asked to read eight short vignettes (stories) about common school events or situations that typically produce feelings of anger and sadness in adolescent and preadolescent students. Participants will then be asked to answer sixteen questions related to each vignette; but they will not have to answer any question they do not wish to answer. The procedure should take no more than 30 to 45 minutes and will take place during regular school hours. Participants will not have to report their name for any reason and their identity will be kept confidential to the extent provided by law. The only personal information requested will be a demographic profile sheet (that is, the 129

PAGE 138

130 student's age, grade-level, sex, and race). Participation or non-participation in this study will not affect your son or daughter's grade(s) or placement in any program(s) at school. You and your child have the right to withdraw consent for your child's participation at any time without consequence. There are no known risks or immediate benefits to the participants. A free video rental coupon from Blockbuster Video will be compensation for your daughter or son's participation. If you have any questions about this research protocol, please contact me at (352) 375-3854 or my faculty supervisor, Dr. Patricia Ashton, at (352) 392-0723 ext. 226. Questions or concerns about your child's rights should be directed to the UFIRB office. University of Florida, Box 1 12250, Gainesville, FL 32611-2250, (352) 392-0433. I have read the procedure described above. I voluntarily give my consent for my study of adolescents' emotional decision-making. I have received a copy of this description. Susan M. Schillinger daughter/son. , to participate in Susan Schillinger' s Parent / Guardian Date 2nd Parent / Witness Date

PAGE 139

TEACHER CONSENT FORM 131 Dear Teacher, I am a doctoral student in the Department of Educational Psychology at the University of Florida under the supervision of Dr. Patricia T. AshtoiL I am conducting research on developmental changes in the emotional decision-making of fifth-, eighth, and eleventh-grade adolescents. The purpose of this study is to examine how decisions to express or conceal emotions change during this age range. This information will help other researchers and myself to design measures that more accurately reflect students' concerns. The results may also help teachers and school administrators better understand the types of situations that lead to common emotional reactions in adolescents. I am asking you and your students to participate in this study. Each student will be asked to read eight short vignettes (stories) about common school events or situations that typically produce feelings of anger and sadness in adolescent and preadolescent students. Participants will then be asked to answer sixteen questions related to each vignette; but, they will not have to answer any question they do not wish to answer. The questionnaire should take no more than 30 to 45 minutes and will be administered during your class period within regular school hours. Participants will not have to report their name for any reason and their identity will be kept confidential to the extent provided by law. The only personal information requested will be a brief demographic profile sheet (that is, the student's age, grade-level, sex, and race). Since the questionnaire is not part of the school curriculum, administration of the questionnaire will require your cooperation. If a student chooses to participate, he or she will miss regular class time. With your permission, each participating student will be

PAGE 140

132 allowed to make up any missed work. Participation or non-participation in this study should not afifect students' grade(s) or placement in any program(s). Please be assured that this study involves no known risks or immediate benefits to your students. A free video rental coupon from Blockbuster Video will be offered as condensation for their participation. Results of this study will be available upon request. If you have any questions about this research protocol, please contact me at (352) 3753854 or my faculty supervisor, Dr. Patricia Ashton, at (352) 392-0723 ext. 226. Questions or concerns about research participants' rights can be directed to the UFIRB ofiBce, University of Florida, Box 1 12250, Gainesville, FL 3261 1-2250; (352) 392-0433. Susan M. Schillinger I have read the procedure described above. I voluntarily give my consent to participate in Susan Schillinger's study of adolescents' emotional decision-making. I have received a copy of this description. Signature of Teacher Date

PAGE 141

APPENDIX D STUDENT ASSENT SCRIPTS Fifth Grade Student Verbal Assent Script My name is Sue and I am a graduate student in educational psychology at the University of Florida. I would like you to read some short stories about things that happen at school that make some kids your age angry and sad. After you have read the stories, I want you to answer some questions on this form You may stop at any time and you will not have to answer any question that you do not want to answer. Also, you do not have to Avrite your name on the questionnaire. Only you and I will see what you write. You will receive a free video rental coupon for doing this. Do you have any questions? Would you like to do this? Eighth Grade Student Verbal Assent Script My name is Sue Schillinger and I am a graduate student in educational psychology at the University of Florida. I am gouig to ask you to read eight short stories and then answer some questions about these stories. This exercise should take about 30-45 minutes and will happen only twice this month, during your regular English class. If you decide to participate, you may stop at any time and you do not have to answer any questions you do not want to answer. Also, you do not have to write your name on the questionnaire, your answers will be completely secret and anonymous. Takmg part in this study will not affect your grades and you will receive a free video rental coupon from Blockbuster Video for answering these questions. Would you like to participate? 133

PAGE 142

134 Eleventh Grade Student Verbal Assent Script My name is Sue Schillinger and I am a graduate student in educational psychology at the University of Florida. I would like you to read eight short stories and then answer a set of questions about each story. This procedure should take only about 30-45 minutes and will occur only twice this month, during your regular English class. If you choose to participate, you may withdraw from participation at any time with no penalty or consequence. You will not have to answer any question that you do not want to answer. Your answers will be completely confidential and anonymous; that is, you will not have to provide your name for any reason. There are no anticipated risks in this research and you will receive a free video rental coupon from Blockbuster Video as compensation for your participation. Although there is no direct benefit to you as a participant, 1 hope this research will help educators better understand adolescents' emotion decision-making. Are you interested in participating?

PAGE 143

REFERENCES Averill, J. (1982). Anger and aggression: An essay on emotioa New York: Springer-Verlag. Averill, J. (1983). Studies on anger and aggression. Implications for theories of emotion. American Psychologist, 38. 1 145-1 160. Harden, R., Zelko, F., Duncan, S., & Masters, J. (1980). Children's consensual knowledge about the experiential determinants of emotioa Journal of Personality and Social Psychology. 39. 968-976. Barrett, K. C, & Campos, J. J. (1987). Perspectives on emotional development: A fimctionalist approach to emotions. In J. Osofsky (Ed.), Handbook of infant development (2nd ed., pp. 555-578). New York: WUey. Belle, D., Burr, R., & Cooney, J. (1987). Boys and girls as social support theorists. Sex Roles. 17. 657-665. Belle, D., & Longfellow, C. (1984, August). Turning to others: Children's use of confidants. Paper presented at the meeting of the American Psychological Association, Toronto, Ontario, Canada. Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive-neoassociationistic analysis. American Psychologist. 45. 494-503. Bemdt, T. J. (1982). The features and effects of friendships in early adolescence. Child Development. 53. 1447-1460. Bemdt, T. J. (1996). Transitions in friendship and friends' influence. In J. A. Graber, J. Brooks-Gimn, & A. C. Petersen (Eds.), Transitions through adolescence: Interpersonal domains and context (pp. 57-84). Mahwah, NJ: Erlbaum. Birch, S. H., & Ladd, G. W. (1996). Interpersonal relationships in the school environment and children's early school adjustment: The role of teachers and peers. In J. Juvonen & K. R. Wentzel (Eds.), Social motivation: Understanding children's school adjustment (pp. 199-225). New York: Cambridge University Press. Bimbaum, D. W., & CrolL, W. L. (1984). The etiology of children's stereotypes about sex diflferences in emotionality. Sex Roles. 10. 677-691. 135

PAGE 144

136 Borg, W. R., & Gall, M. D. (1989). Educational research (5th ed.) White Plains, NY: Longman. Brody, L. R. (1985). Gender differences in emotional development: A review of theories and research. Journal of Personality. 53. 102-149. Brody, L. R. (1993). On understanding gender differences in the expression of emotion. In S. L. Ablon, D. Brown, E. J. Khantzian, & J. E. Mack (Eds.), Human feelings: Explorations in affect development and meaning (pp. 87-121), Hillsdale, NJ: Analytic Press. Brody, L. R., & Hall, J. (1993). Gender and emotion. In M. Lewis & J. Haviland (Eds.), Handbook of emotions (pp. 447-460). New York: Guilford Press. Brody, L. R., & Hall, J. A. (2000). Gender, emotion, and expression. In M. Lewis & J. M. HavilandJones (Eds.), Handbook of emotions (pp. 338-349). New York: Guilford Press. Brody, L. R., Lovas, S. S., & Hay, D. H. (1995). Gender differences in anger and fear as a function of situational context. Sex Roles. 32. 47-79. Brown, L. M., & Gilligan, C. (1992). Meeting at the crossroads: Women's psychology and girls' development. Cambridge, MA: Harvard University Press. Buhrmester, D., & Furman, W. (1987). The development of companionship and intimacy. Child Development. 58, 1 101-1 1 13. Campos, J. J., Campos, R. G., & Barrett, K. C. (1989). Emergent themes in the study of emotional development and emotional regulation. Developmental Psychology. 25, 394-402. Can^s, J., Mumme, D., Kermoian, R., & Campos, R. (1994). A fimctionalist perspective on the nature of emotion. Monographs of the Society for Research in Child Development. 59 (Serial No. 240). Cole, P. M. (1986). Children's spontaneous control of fecial expression. Child Development. 57. 1309-1321. Collins, W. A., & Russell, G. (1991). Mother-child and fether-child relationships in middle childhood and adolescence: A developmental review. Developmental Review. IL 99-136. Davis, T. L. (1995). Gender differences in masking of negative emotions: Ability or motivation? Developmental Psychology. 31. 660-667.

PAGE 145

137 Deaux, K., & Major, B. (1987). Putting gender into context: An interactive nvodel of gender-related behavior. Psychological Review. 94, 369-389. Denham, S. A., & Couchoud, E. A. (1990). Young preschoolers' understanding of emotions. Child Study Journal 20. 171-192. Diener, E., Sandvik, E., & Larsen, R. J. (1985). Age and sex effects for emotional intensity. Developmental Psychology. 21. 542-546. Eccles, J. S., Lord, S., & Buchanan, C. M. (1996). School transitions in early adolescence: What are we doing to our young people? In J. A. Graber, J. Brooks-Gunn, & A. C. Petersen (Eds.), Transitions through adolescence: Interpersonal domains and context (pp. 25 1-284). Mahwah, NJ: Erlbaum. Eccles, J. S., & Midgley, C. (1989). Stage/environment fit: Developmentally appropriate classrooms for early adolescents. In R. E. Ames & C. Ames (Eds.), Research on motivation in education (Vol. 3, pp. 139-186). San Diego, CA: Academic Press. Eccles, J. S., Midgley, C, Wigfield, A., Buchanan, C. M., Reuman, D., Flanagan, C, & Maclver, D. (1993). Development during adolescence: The impact of stageenvironment fit on young adolescents' experiences in schools and in families. American Psychologist. 48. 90-101. Eisenberg, N., & Zhou, Q. (2000). Regulation fi-om a developmental perspective. Psychological Inquiry. 11. 166-171. Elkind, D., & Bo wen, R. (1979). Imaginary audience behavior in children and adolescents. Developmental Psychology. 15. 38-44. Fabes, R. A., & Martin, C. L. (1991). Gender and age stereotypes of emotionality. Personahty and Social Psychology Bulletin. 17. 532-540. Fabes, R. A., Eisenberg, N., Smith, M. C, & Murphy, B. C. (1996). Getting angry at peers: Associations with liking of the provocateur. Child Development. 67. 942-956. Fivush, R. (1989). Exploring sex differences in the emotional content of motherchild conversations about the past. Sex Roles. 20. 675-691 . Flannery, D., Torquatl, J., & Lindemeier, L. (1994). The method and meaning of emotional expression and experience during adolescence. Journal of Adolescent Research. 9. 8-27. Fuchs, D., & Thelen, M. H. (1988). Children's expected interpersonal consequences of commimicating their affective state and reported likelihood of expression. Child Development. 59. 1314-1322.

PAGE 146

138 Fujita, F., Diener, E., & Sandvik, E. (1991). Gender differences in negative affect and well-being: The case for emotional intensity. Journal of Personality and Social Psychology. 61. 427-434. Garber, J., Braafladt, N., & Zeman, J. (1991). The regulation of sad affect: An mformation-processing perspectiye. In J. Garber & K. A. Dodge (Eds.), The deyelopment of emotion regulation and dysreeulation (pp. 208-240). New York: Cambridge Uniyersity Press. Gilligan, C, Lyons, N. P., & Hammer, T. J. (1990). Making connections: The relational worlds of adolescent girls at Emma Willard school. Cambridge, MA: Harvard Uniyersity Press. Glasberg, R., & Aboud, F. (1982). Keeping one's distance from sadness: Children's self-reports of emotional experience. Deyelopmental Psychology. 18. 287293. Gnepp, J., & Hess, D. L. R. (1986). Children's understanding of verbal and facial display rules. Developmental Psychology. 22. 103-108. Goleman, D. (1995). Emotional mtelligence. New York: Bantam Books. Gordon, S. (1989). The socialization of children's emotions: Emotional culture, competence, and exposure. In C. Saami & P. L. Harris (Eds.), Children's understanding of emotion (pp. 319-349). New York: Cambridge University Press. Greif, E., Alvarez, M., & Ulman, K. (1981, March). Recognizmg emotions in other people: Sex differences in socialization. Paper presented at the biennial meeting of the Society for Research in Child Development, Seattle, WA. Harris, P. L., Donnelly, K., Guz, G., & PittWatson, R. (1986). Children's understanding of the distinction between real and apparent emotioa Child Development. 5L 895-909. Harris, P., & Saami, C. (1989). Children's understanding of emotion. New York: Cambridge University Press. Harter, S. (1996). Teacher and classmate influences on scholastic motivation, selfesteem, and level of voice in adolescents. In J. Juvonen & K. R. Wentzel (Eds.), Social motivation: Understanding children's school adjustment (pp. 1 1-42). New York: Cambridge University Press. Hartup, W. W. (1993). Adolescents and their friends. New Directions for Child Development. 60. 3-22.

PAGE 147

139 Hartup, W. W. (1996). The company they keep: Friendships and their developmental significance. Child Development. 67, 1-13. Holsti, O. R. (1969). Content analysis for the social sciences and humanities. Reading, MA: AddisonWesley. Izard, C. E. (1991). The psvchologv of emotions. New York: Plenum Press. Juvonen, J. (1996). Self-presentation tactics promoting teacher and peer approval: The fimction of excuses and other clever explanations. In J. Juvonen &, K. R. Wentzel (Eds.), Social motivation: Understanding children's school adjustment (pp. 43-65). New York: Cambridge University Press. Kamiol, R., & Heiman, T. (1987). Situational antecedents of children's anger experiences and subsequent responses to adult versus peer provokers. Aggressive Behavior. 13. 109-118. Laursen, B., & Collins, W. A. (1994). Interpersonal conflict during adolescence. Psvchological Bulletin. 115. 197-209. Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press. Levine, L. J. (1995). Young children's understanding of the causes of anger and sadness. Child Development. 66. 697-709. Miller, J. B. (1991). The construction of anger in women and men. In J. Jordan, A. Kaplan, J. B. Miller, I. Stiver, & J. Surrey (Eds.), Women's growth in connection (pp. 181-196). New York: Guilford Press. Miller, P., & Sperry, L. (1987). The socialization of anger and aggression. Merrill-Palmer Quarterly. 33. 1-31. Nelson, HI, W. M., Hart, K. J., &. Finch, Jr., A. J. (1993). Anger in children: A cognitive behavioral view of the assessment-therapy connection. Journal of RationalEmotive & Cognitive Behavior Therapy. 1 1 , 135-150. Papini, D. R., Farmer, F. F., Clark, S. M., Micka, J. C, & Bamett, J. K. (1990). Early adolescent age and gender differences in patterns of emotional self-disclosure to parents and friends. Adolescence. 25. 959-976. Reichenbach, L., & Masters, J. C. (1983). Children's use of expressive and contextual cues in judgments of emotion. Child Development. 54. 993-1004.

PAGE 148

140 Renouf, A. G., & Harter, S. (1990). Low self-worth and anger as components of the depressive experience in young adolescents. Development and Psv chopatholoev. 2. 293-310. Rotenberg, K. J., Mars, K., & Crick, N. (1987-88). Development of children's sadness. Psychology and Human Development, 2, 13-25. Saami, C. (1979). Children's understanding of display rules for expressive behavior. Developmental Psychology. 15, 424-429. Saami, C. (1984). An observational study of children's attempts to monitor their expressive behavior. Child Development. 55, 1504-1513. Saami, C. (1988). Children's understanding of the consequences of dissemblance of nonverbal emotional-expressive behavior. Journal of Nonverbal Behavior. 12. 275294. Saami, C. (1989a). Children's understanding of strategic control of emotional expression in social transactions. In C. Saami & P. L. Harris (Eds.), Children's understanding of emotion (pp. 181-208). New York: Cambridge University Press. Saami, C. (1989b, April). Cognitive capabilities involved in the socialization of emotion: Development in middle childhood. Paper presented at the biennial meeting of the Society for Research in Child Development, Kansas City, MO. Saami, C. (1990). Emotional competence: How emotions and relationships become integrated. In R. Dienstbier (Series Ed.) & R. Thompson (Vol. Ed.), Socioemotional development: Vol. 36. The Nebraska symposium on motivation. 1988 (pp. 1 15-182). Lincohi, NE: University of Nebraska Press. Saami, C. (1992). Children's emotional-expressive behaviors as regulators of others' happy and sad states. New Directions for Child Development. 55. 91-106. Saami, C. (1993). Socialization of emotion. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotion (pp. 435-446). New York: Guilford Press. Saami, C. (1995, March). Coping with aversive feelings. P aper presented at the biennial meeting of the Society for Research in Child Development, Indianapolis, IN. Saami, C. (1997a). Emotional competence and self-regulation in childhood. In P. Salovey & D. J. Sluyter (Eds.), Emotional development and emotional intelligence: Educational implications (pp. 35-66). New York: Basic Books. Saami, C. (1997b, April). Emotion management and strategies for social interaction. Paper presented at the biennial meeting of the Society for Research in Child Development, Washington, DC.

PAGE 149

141 Saami, C. (1999). The development of emotional competence. New York: Guilford Press. Saami, C, & von Salisch, M. (1993). The socialization of emotional dissemblance. In M. Lewis & C. Saami (Eds.), Lvine and deception in everyday life (pp. 106-125). New York: Guilford Press. Salovey, P., & Sluyter, D. J. (1997). Emotional development and emotional intelligence: Educational implications. New York: Basic Books. Shields, S. (1995). The role of emotion beliefs and values in gender devetopment. In N. Eisenberg (Ed.), Social development. Review of Personality and Social Psychology. Vol. 15 (pp. 212-232). Thousand Oaks, CA: Sage. Sonnemans, J., & Frijda, N. H. (1995). The determinants of subjective emotional intensity. Cognition and Emotion. 5, 483-506. Stapley, J. C, & Haviland, J. M. (1989). Beyond depression: Gender differences in normal adolescents' emotional experiences. Sex Roles. 20. 295-308. Steams, C. Z., & Steams, P. N. (1986). Anger: The stmggle for emotional control in America's history. Chicago: University of Chicago Press. Stein, N. L., & Jewett, J. L. (1986). A conceptual analysis of the meaning of negative emotions: Implication for a theory of development. In C. Izard & P. Reed (Ed.), Measuring emotions in infants and children (Vol. 2, pp. 238-267). Cambridge: Cambridge University Press. Stein, N. L., & Levine, L. J. (1987). Thmking about feelings: The development and organization of emotional knowledge. In R. E. Snow & M. Farr (Eds.), Aptitude, learning, and mstmction: Vol. 3. Cognition, connation and affect (pp. 165-197). Hillsdale, NJ: Erlbaum. Steinberg, L. D. (1988). Reciprocal relations between parent-child distance and pubertal maturation. Developmental Psychology. 24. 122-128. Steinberg, L. D. (1989). Pubertal maturation and femily relations: Evidence for the distancing hypothesis. In G. R. Adams, R. Montemayor, & T. Gullotta (Eds.), Advances in adolescent development (Vol. 1, pp. 71-97). Beverly Hills, CA: Sage. Steinberg, L. D. (1999). Adolescence (5th ed.). New York: McGraw-Hill. Tangney, J. P., Hill-Barlow, D., Wagner, P. E., Marschall, D. E., Borenstein, J. K., Sanftner, J., Mohr, T., & Gramzow, R. (1996). Assessing individual differences in constmctive versus destmctive responses to anger across the lifespan. Journal of Personahty and Social Psychology. 70. 780-796.

PAGE 150

142 Thompson, R. A. (1990). Emotion and self-regulation. In R. Dienstbier (Series Ed.) & R. Thompson fVol. FAX Socioemotional development: Vol. 36. The Nebraska Symposium on Motivation. 1988 (pp. 367-467). Lincoln, NE: University of Nebraska Press. Thompson, R. A. (1994). Emotion regulation: A theme in search of definition. In N. A Fox (Ed.), The development of emotion regulation (pp. 25-52). Monographs of tl^ Society for Research in Child Development, 59, 2-3. Torestad, B. (1990). What is anger provoking? A psychophysical study of perceived causes of anger. Aggressive Behavior. 16, 9-26. Underwood, M. K., Coie, J. D., & Herbsman, C. R. (1992). Display rules for anger and aggression in school-age children. Child Development, 63, 366-380. Weber, R. P. (1990). Basic content analysis. (2nd ed.). Newbury Park, CA: Sage. Wenz-Gross, M., Sipersteui, G. N., Untch, A. S., & Widaman, K. F. (1997). Stress, social support, and adjustment of adolescents in middle school. Journal of Early Adolescence, 17, 129-151. Wentzel, K. R. (1996). Social goals and social relationships as motivators of school adjustment. In J. Juvonen & K. R. Wentzel (Eds.), Social motivation: Understanding children's school adjustment (pp. 226-247). New York: Cambridge University Press. Wentzel, K. R. (1997). Student motivation in the middle school: The role of perceived pedagogical caring. Journal of Educational Psychology. 89, 41 1-419. Whitesell, N. R., & Harter, S. (1996). The mterpersonal context of emotion: Anger with close friends and classmates. Child Development. 67. 1345-1359. Whitesell, N. R., Robinson, N. S., & Harter, S. (1993). Coping with angerprovoking situations: Yoimg adolescents' theories of strategy use and effectiveness. Journal of Applied Developmental Psychology, 14, 521-545. Wierzbicka, A. (1994). Emotion, language, and cultural scripts. In S. Kitayama & H. R. Markus (Eds.), Emotion and culture: Empirical studies of mutual influence (pp. 133-196). Washington, DC: American Psychological Association. Youniss, J., & Smollar, J. (1990). Self through relationship development. In H. Bosma & S. Jackson (Eds.) Coping and self-concept in adolescence (pp. 129-148). New York: SpringerVerlag. Zeman, J., & Garber, J. (1996). Display rules for anger, sadness, and pain: It depends on who is watching. Child Development, 67, 957-973.

PAGE 151

143 Zeman, J., & Shipman, K. (1996). Children's expression of negative affect: Reasons and methods. Developmental Psychology. 32, 842-849. Zeman, J., &. Shipman, K. (1997). Social-contextual influences on expectancies for managing anger and sadness: The transition from middle childhood to adolescence. Developmental Psychology. 33, 9 1 7-924. Zeman, J., & Shipman, K. (1998). Influences of social context on children's affect regulation: A fimctionalist perspective. Journal of Nonverbal Behavior. 22. 141-165.

PAGE 152

BIOGRAPHICAL SKETCH Susan M. Schiilinger was bom and raised in Detroit, Michigan. She earned her first degree in dental hygiene from the University of Detroit. After several years in private practice, she returned to school and graduated in 1981 with a Bachelor of Science degree in education from the University of North Carolina at Chapel Hill. In 1988, she completed a Master of Arts degree in counseling from Rollins College in Winter Park, Florida. She has worked in a variety of positions in higher education for twenty years as an instructor and an administrator. Upon completion of her doctoral program in educational psychology, she will resume her career with a state fimded program that provides financial and technical assistance to educationally disadvantaged students. She currently lives in Gainesville, Florida and is the mother of a daughter, Christine. 144

PAGE 153

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Patricia T. Ashton, Chair Professor of Educational Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Bridget A^Franks, Cochar Associate Professor of Educational Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is folly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. M. David Miller Professor of Educational Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is folly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Arthur J. Newman Professor of Educational Leadership, Policy, and Foundations

PAGE 154

This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May 2002 ;;hairman, Educational Psychology Dean, Graduate School