Citation
Emotional and cognitive reactions to hypothetical moral trangressions in aggressive and nonaggressive boys in Taiwan

Material Information

Title:
Emotional and cognitive reactions to hypothetical moral trangressions in aggressive and nonaggressive boys in Taiwan
Creator:
Liao, Shu-Tai, 1950-
Publication Date:
Language:
English
Physical Description:
x, 147 leaves : ; 29 cm.

Subjects

Subjects / Keywords:
Blame ( jstor )
Child development ( jstor )
Child psychology ( jstor )
Empathy ( jstor )
Guilt ( jstor )
Human aggression ( jstor )
Morality ( jstor )
Shame ( jstor )
Social psychology ( jstor )
Transgression ( jstor )
Dissertations, Academic -- Foundations of Education -- UF ( lcsh )
Foundations of Education thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 134-146).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Shu-Tai Liao.

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:
030025787 ( ALEPH )
41875560 ( OCLC )

Downloads

This item has the following downloads:


Full Text











EMOTIONAL AND COGNITIVE REACTIONS TO HYPOTHETICAL MORAL TRANSGRESSIONS IN AGGRESSIVE AND NONAGGRESSIVE
BOYS IN TAIWAN














By

SHU-TAI LIAO


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


1998















ACKNOWLEDGbENTS


The completion of this study was made possible by the contribution of a number of people. I wish to express my sincere gratitude to Dr. Patricia Ashton, chairperson of my doctoral committee, for her continuous guidance, encouragement, and careful scrutiny of my work throughout the years of my doctoral study and this project. Without her assistance and expertise, I would never have completed this study. I also wish to thank Dr. Shari Ellis, Dr. Bridget Franks, and Dr. Jin-Wen Hsu, for their valuable counsel and support.

My appreciation is also extended to my family, my

parents, my husband, and two sons, for their encouragement and sacrifices.
















TABLE OF CONTENTS


Page
ACKNOWLEDGNENTS ....................................... ii

LIST OF TABLES ........................................ . V

ABSTRACT .............................................. ix

CHAPTERS

1 INTRODUCTION .................................... 1


Statement of the Problem . Purpose of the Study ..... Significance of the Study Review of Literature ..... Summary...................


2 DESIGN OF THE STUDY ............................. 33

Research Participants ........................... 33
Measures ........................................ 34
Procedures ...................................... 38
Data Analysis ................................... 39
Limitations of the Study.......................... 40
Pilot Study ..................................... 41
Revising Interview Questions and Stories......... 52


3 RESULTS .........................................

Descriptive Data ...............................
Differences Between Aggressive and Nonaggressive
Boys ..........................................
Differences Between the Sixth and Third Graders..

4 DISCUSSION AND RECOMMENDATIONS ..................

Overview of the Findings ........................
Discussion of Results ............................
Recommendations for Further Research ............
Recommendations for Practice ....................


54 55

58 77 99 99
100 117 119


iii


.......................
.......................
.......................
.......................
.......................










APPENDICES

A TEACHER CHECKLIST ............................... 121

B REVISED STORIES ................................. 122

C INTERVIEW QUESTIONS WITH REVISIONS .............. 124

D CONSENT FORMS .................................... 127

E INTERCORRELATION MATRIX .......................... 131

REFERENCES ............................................ 134

BIOGRAPHICAL SKETCH ................................... 147















LIST OF TABLES


Table Page

1 Differences between Aggressive and Nonaggressive
Children in Reactions to Moral Transgression ..... 43

2 Differences between Third and Sixth Graders in
Reactions to Moral Transgression ................. 46

3 Correlations Between Scores on Interview
Questions.......................................... 49

4 Means and Standard Deviations on Each Variable
for Sixth Graders ................................ 56

5 Means and Standard Deviations on Each Variable
for Third Graders ................................ 57

6 ANCOVA Test on Story Completion for Sixth- and
Third-Grade Aggressive and Nonaggressive Boys
with IQ as Covariate ............................. 59

7 MANCOVA Test of Reactive Feelings for Aggressive
and Nonaggressive Boys in Grade Six and Three .... 61

8 Univariate F-Tests of Reactive Feelings for
Aggressive and Nonaggressive Boys in Grade Six
and Three ........................................ 61

9 Frequency and Fisher's Exact Test for Cognitive
Coping Strategy for Aggressive and Nonaggressive
Boys in Sixth Grade .............................. 64

10 Frequency and Fisher's Exact Test for Cognitive Coping Strategy for Aggressive and Nonaggressive
Boys in Third Grade .............................. 64

11 MANOVA of Intention to Compensate for Aggressive and Nonaggressive Boys in Grade Six and Three .... 66









12 Univariate F-Tests of Intention to Compensate for
Aggressive and Nonaggressive Boys in Grade Six
and Three ........................................ 66

13 MANOVA of Intensity of Disturbance for Aggressive
and Nonaggressive Boys in Grade Six and Three ... 67

14 Univariate F-Tests of Intensity of Disturbance for
Aggressive and Nonaggressive Boys in Grade Six
and Three ........................................ 67

15 MANOVA of Length of Disturbance for Aggressive and
Nonaggressive Boys in Grade Six and Three ........ 69

16 Univariate F-Tests of Length of Disturbance for
Aggressive and Nonaggressive Boys in Sixth- and
Third-Grade ...................................... 69

17 MANOVA of Empathic Reactions for Aggressive and
Nonaggressive Boys in Grade Six and Three......... 71

18 Univariate F-Tests of Empathic Reactions for Sixth
and Third-Grade Boys ............................. 71

19 MANOVA of Outcome Expectations for Aggressive and
Nonaggressive Boys in Grade Six and Three ........ 72

20 Univariate F-Tests of Outcome Expectations for
Aggressive and Nonaggressive Boys in Grade Six
and Three ........................................ 72

21 MANOVA of Beliefs about the Legitimacy of
Aggression for Aggressive and Nonaggressive Boys
in Grade Six and Three ........................... 74

22 Univariate F-Tests of Beliefs about the Legitimacy
of Aggression for Aggressive and Nonaggressive
Boys in Grade Six and Three ...................... 7

23 MANOVA of Attribution of Self-Responsibility for
Aggressive and Nonaggressive Boys in Grade Six
and Three ........................................ 76

24 Univariate E-Tests of Attribution of SelfResponsibility for Aggressive and Nonaggressive
Boys in Grade Six and Three ...................... 76









25 MANOVA of Attributing Responsibility to Victim for
Aggressive and Nonaggressive Boys in Grade Six and
Three ............................................ 78

26 Univariate F-Tests of Attributing Responsibility
to Victim for Aggressive and Nonaggressive Boys
in Grade Six and Three ........................... 78

27 ANOVA Results for Testing Age Differences in
Story Completion ................................. 80

28 MANOVA Results for Testing Age Differences in
Reactive Feelings ................................ 81

29 Results of Univariate F-Tests for Age Differences
in Reactive Feelings ............................. 81

30 Frequency and Fisher's Exact Test for Age
Differences in Cognitive Coping Strategy for
Aggressive Boys .................................. 83

31 Frequency and Fisher's Exact Test for Age
Differences in Cognitive Coping Strategy for
Nonaggressive Boys ............................... 83

32 MANOVA Results of Age Differences in Intention to
Help and Apologize ............................... 85

33 Results of Univariate F-Tests for Age Differences
in Intention to Help and Apologize ............... 85

34 MANOVA Results of Age Differences in Intensity of
Disturbance ...................................... 87

35 Results of Univariate F-Tests for Age Differences
in Intensity of Disturbance. ....................... 87

36 MANOVA Results of Age Differences in Length of
Disturbance ........... *......................... 88

37 Results of Univariate F-Tests for Age Differences
in Length of Disturbance. .......................... 88

38 MANOVA Results of Age Differences in Empathic
Reactions ...................... .................. 90

39 Results of Univariate E-Tests for Age Differences
in Empathic Reactions............................. 90


vii









40 MANOVA Results of Age Differences in Outcome
Expectations for Aggression ...................... 92

41 Results of Univariate F-Tests for Age Differences
in Outcome Expectations for Aggression ........... 92

42 MANOVA Results of Age Differences in Beliefs about
the Legitimacy of Aggression .................... 94

43 Results of Univariate F-Tests for Age Differences
in Beliefs about the Legitimacy of Aggression ... 94 44 MANOVA Results of Age Differences in Attribution
of Self-Responsibility ........................... 95

45 Results of Univariate F-Tests for Age Differences
in Attribution of Self-Responsibility ............ 95

46 MANOVA Results of Age Differences in Attributing
Responsibility to Victim ......................... 97

47 Results of Univariate F-Tests for Age Differences
in Attributing Responsibility to Victim .......... 97
























I


viii















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

EMOTIONAL AND COGNITIVE REACTIONS TO HYPOTHETICAL
MORAL TRANSGRESSIONS IN AGGRESSIVE AND NONAGGRESSIVE
BOYS IN TAIWAN

By

Shu-Tai Liao

December 1998

Chairperson: Patricia Ashton Major Department: Foundations of Education

The purpose of this study was to identify differences between aggressive and nonaggressive boys and between sixthand third-grade boys in their emotional and cognitive reactions to hypothetical moral transgressions. One hundred aggressive and one hundred nonaggressive boys in the third and sixth grades in Taiwan selected on the basis of teacher ratings were interviewed individually. The boys read three semiprojective stories describing explicit misbehavior and then answered questions concerning how they would feel and think if they were the perpetrators of the misbehavior.

Results indicated that boys in all groups, regardless of age and aggressiveness, reported that they would feel high levels of negative emotions and low levels of positive emotions and they believed that aggression is inappropriate. However, they differed in the strength of their feelings,









outcome expectations, and responsibility attributions. For the sixth graders, nonaggressive boys reported that they would feel higher levels of sadness, guilt, shame, and empathy, have a stronger intention to compensate, would be more intensely disturbed, expect a less favorable outcome, and attribute less responsibility to victims, when compared to aggressive boys. Among the third graders, nonaggressive boys reported that they would expect less favorable consequences for transgressions and attribute more responsibility to self, relative to aggressive peers.

Age differences were also found. Younger aggressive boys reported that they would feel higher levels of happiness and pride than older boys. The older aggressive boys reported that they would expect more disapproval from friends for the aggression than did younger aggressive boys. The older nonaggressive boys said that they would feel higher levels of shame and a stronger intention to remedy their misbehavior than did the younger boys. Both aggressive and nonaggressive older boys said they would be more intensely disturbed longer and attribute more responsibility to self and victims, compared to the younger boys.

Results are discussed in terms of their relationship to previous studies. Finally, some suggestions for further study and practice are proposed.















CHAPTER 1
INTRODUCTION

Statement of the Problem

Childhood aggression has been a problem in all

societies (Crowell, 1987). Epidemiological studies reported by the Institute of Medicine in 1989 showed that in the United States between 2% and 6% of youths commit clinically severe antisocial behavior; that is about 1.3 to 3.8 million youths (Kazdin, 1994). The increasing rate of crime committed by children and youths is a serious problem in Taiwan as well. According to official report (Justice Department of the Republic of China, 1996), three trends in youth crime can be identified. First, juvenile crime is increasing rapidly. From 1986 to 1995, the juvenile crime events referred to district courts increased by 80%. In 1986, 18% of total crimes were committed by youths; in 1995, the percentage had increased to 30%. The second trend is the age distribution of crime is downward. From 1991 to 1995, the percent of juvenile crimes committed by adolescents aged 12 to 13 increased from 6.7% to 9.1%. For youths aged 13 to 14 the percentage increased from 12.2% to 15.7%. The third trend is that the number of violent crimes committed by







2

youths increased dramatically. During a 1-year period, from 1994 to 1995, injury cases climbed 19.7%, threatening increased 24.6%, and robbery grew by 69.4%.

There is strong evidence that aggression in

individuals tends to be relatively stable over time and situation. A number of studies have shown that early aggressiveness or antisocial conduct is predictive of later serious antisocial behavior or criminality in adolescence, adulthood, and even in the next generation for certain groups of children (Ghodsian, Fogelman, Lambert, & Tibbenham, 1980; Huesmann, Eron, Lefkowitz, & Walder, 1984; Loeber & Dishion, 1983; Pulkkinen & Pitkanen, 1993). From a 22-year longitudinal study, Eron, Huesmann, Dubow, Romanoff, and Yarmel (1986) concluded that

By the time a child is 8 years old, characteristic ways
of behaving aggressively or nonaggressively have already
been established. Aggression as a problem-solving behavior is learned very early in life, and it is
learned very well; the payoff is tremendous. (p. 261)

On the basis of experimental studies, Geen (1990) concluded that "the original act of aggression appears to facilitate subsequent aggression instead of diminishing it" (p. 189). Aronson, Wilson, and Akert (1994) also proposed that people tend to justify their aggression in ways that eventually produce more aggressive behavior.









Obviously, the consistency and repetition of

aggression make it a serious social problem. If we want to alleviate the problem, we need to reduce its recurrence in addition to preventing its occurrence. Thus, knowledge of why some children persist in aggressive behavior, why aggression may produce more aggression, and how to interrupt the continuation of aggression is crucial for alleviating this problem and should not be neglected.

Although researchers and theorists have long been

interested in the processes involved in the development of aggression, most attention has been directed to the causes of aggressive behavior. For example, frustration, physiological factors, observational learning, failure in anger control, and hostile intention attribution have been emphasized as the origins of aggression (Geen, 1990). Fewer studies have been conducted on the maintenance of aggression compared to those on the causes of aggression. Among the few, some researchers (Ferguson, Stegge, & Damhuis, 1990; Regan, 1972; Tangney, 1991, 1992) have argued that affective reactions to moral transgression influence the recurrence of the wrongdoing, and others (Crick & Dodge, 1994; Dodge, 1993; Perry, Perry, & Rasmussen, 1986; Slaby & Guerra, 1989) have stressed cognitive reactions as contributors to the repetition of aggressive behavior. From these contrasting









perspectives several questions emerge: What do aggressive children feel and think after wrongdoing? Do aggressive and nonaggressive children differ in their reactions to transgressions? We need more research to learn how children react to their transgressions and how those reactions influence their subsequent behavior.

Purpose of the Study

Cognitive-developmental researchers have focused on

the role of social-cognitive variables in the development of aggression in children. They have identified outcome expectancies (Perry et al., 1986), beliefs about the legitimacy of aggression (Slaby & Guerra, 1988), and attributions of responsibility (McGraw, 1987) as important factors that affect the development of aggression. In contrast, some researchers with a social-emotional emphasis have proposed that empathy plays an important role in inhibiting aggression (Miller & Eisenberg, 1988), and others have suggested that guilt inhibits aggression (Tangney, 1991).

These cognitive and emotional consequences of

aggression have not been combined in a single study. To provide insight into the relationships among cognitive and emotional factors in the development of aggression, this study will examine the emotions and cognitions that









aggression elicits. The following questions will be explored:

1. Do aggressive and nonaggressive children feel different emotions after moral transgression?

2. Do aggressive and nonaggressive boys differ in the intensity of their disturbance after transgression?

3. Does the transgression disturb nonaggressive children longer than aggressive children?

4. Do aggressive and nonaggressive boys differ in their empathic reactions to the victim of aggression?

5. Are aggressive children more likely to attribute responsibility to others while nonaggressive children attribute responsibility to themselves?

6. Are aggressive children more likely to hold a positive outcome expectancy for transgression than nonaggressive children?

7. Are aggressive children more likely to believe that aggression is a legitimate response than nonaggressive children?

8. Are there age differences in aggressive and nonaggressive children's emotional and cognitive reactions to transgression?









Significance of the Study

Theoretical Implications

Aggressive and nonaggressive children may differ in their emotional and cognitive responses to transgression. Identifying what aggressive children feel and think after aggression may help us explain why some children become aggressive and some do not and why for some people aggressive behavior is stable across the life span (Eron et al., 1986).

Guilt has been conceived as an important moral

emotion, but until recently researchers have not focused their attention on the development of guilt in children. From the extant literature, we still know little about children's reactions to their guilt feelings, how children cope with the emotion of guilt, or the effects of this emotion on aggressive children's later behaviors. This study may help answer these questions.

Learning theorists have emphasized that the

consequences of behavior are the key to explaining the recurrence of a behavior (Bandura, Ross, & Ross, 1963; Schunk, 1991). However, they did not attend to how children interpret the consequences of their behavior; that is, what the consequences mean from children's perspectives. Our theoretical understanding of the impact of behavior would be









improved if researchers included in their analysis of the consequences of behavior the thoughts and feelings that children's behavior elicits. This study focuses on children's reactions. The results should reveal more about what misbehavior means for children than studies that have not included children's feelings and thinking. Practical Implications

Aggressive behavior has become a prevalent and serious social problem. Identification of the factors that influence aggression and the design of effective intervention programs are goals shared by scholars, educators, and policy makers in many countries. To eliminate aggressive behavior, it is critical to have prevention and intervention programs. Because children's emotions and thoughts following transgression could influence the recurrence of the misbehavior, understanding what children feel and think in such situations is necessary for the design of effective interventions.

If the results of this study show that aggressive

children do not experience guilt or other negative emotions after their misbehavior but nonaggressive children do, it would be important to increase aggressive children's experience of those emotions after aggression. If aggressive and nonaggressive children differ in their attributions of







8

responsibility for misbehavior, then an intervention program to modify the attributions of aggressive children should be developed. If results reveal that aggressive children believe that wrongdoing can help them reach their goals, intervention strategies should include modifying children's beliefs about the effectiveness of aggression.

Review of Literature

Emotional Reactions to Transgression

Only a few studies have been conducted on children's emotional reactions following transgression. However, the recent research on guilt, shame, and empathy, which have been conceived as important moral affects (Tangney, 1991; Zahn-Waxler, Kochanska, Krupnick, & McKnew, 1990; ZahnWaxler & Robinson, 1995), provides information on children's emotional responses to their wrongdoing.

Guilt. Guilt refers to thoughts and feelings of

remorse and responsibility for causing negative outcomes that violate social or moral norms. According to Tangney (1990, 1991), guilt is a negative self-evaluative emotion. The focus of the evaluation is a "bad" behavior committed by oneself. The feelings of remorse and responsibility typically motivate this person to remedy his/her transgression. Zahn-Waxler et al. (1990) emphasized the importance of guilt in moral behavior: "As the main affect









in conscience, it checks aggressive impulses and encourages people to undo harms, hence restoring social harmony" (p. 51).

Williams and Bybee (1994) asked 240 children in three grades about the events that made them feel guilty. The researchers found that 90.6% of 5th graders, 93.3% of 8th graders, and 92.3% of 11th graders reported feeling guilty after transgression. Ferguson et al.(1990) found that 96% of the second and fifth graders in their study reported experiencing guilt after a moral transgression. Perry, Perry, Bussey, English, and Arnold (1980) indicated that the more serious that children thought their misbehavior was, the more serious self-punishment they adopted. These research findings reveal that, following misbehavior, children typically experience guilt or negative emotions.

Research on adult transgression has shown that guilt is typically experienced after one behaves aggressively (Okel & Mosher, 1968; Tangney, 1992), that there is an inverse relationship between guilt and hostility and aggression (Mosher, Mortimer, & Grebel, 1968; Schill & Schneider, 1970), that guilt leads to altruistic acts in persons who cause harm (Regan, 1971), and that after a person causes harm to others, he/she is more willing to do something altruistic. These findings are consistent with the









belief that guilt motivates amending action (Ferguson et al., 1990; Regan, 1971; Tangney, 1991) and inhibits prohibited behavior (Mosher, 1965). In light of this research suggesting that guilt inhibits aggression, it is reasonable to propose that aggressive children do not inhibit their aggression by guilt as other children do.

There are several possible reasons why aggressive children may not be influenced by guilt. First, these children might not feel negative emotions or only feel a low level of guilt that does not affect their behavior. Or, children may experience complicated emotions instead of guilt. Zahn-Waxler and Robinson (1995) suggested that once a child has hurt another person, many feelings may be evoked. In aggressive situations the child may feel empathy but also have hostile feelings, enjoy hurting the other, and be more affectively aroused.

Another possibility is that aggressive children feel more shame but not guilt after a transgression than nonaggressive children. Researchers (Tangney, Wagner, Burggraf, Gramzow, & Fletcher, 1991) have found that shame is positively related to aggression. Finally, aggressive children may get rid of uncomfortable feelings fast by justifying their harmful act or focusing their attention on something else.









Shame. Shame is a global painful or humiliating

feeling resulting from a negative evaluation of oneself for failure to live up to standards or goals in the eyes of others (Lazarus, 1991; Lewis, 1992). Experiences of shame can be aroused by transgression of moral standards or social conventions or by personal failure (Ferguson et al., 1990; Tangney, 1992). In shame, the focus of self-evaluation is the entire self--the entire self is painfully scrutinized and negatively evaluated. In this situation, shame experience is usually accompanied by a sense of exposure, a sense of being small, and a sense of worthlessness and powerlessness. Shame feeling, thus, drives a person to hide self from others (Tangney, 1990, 1991; Tangney et al. 1992).

Researchers have addressed the issue of shame and its behavioral correlates. Lewis (1992) and Scheff and Retzinger (1991) argued that feelings of shame can elicit humiliated fury and then transform into hostility, anger, or blind rage and consequently into aggressive behavior; however, shame also can cause withdrawal from situations or persons associated with shameful feelings. Scheff (1987) also suggested that rage may become a new source of shame and then produce a shame-rage-shame-rage spiral. Empirical studies have revealed that shame-prone individuals are more likely to experience anger or hostility than their less









shame-prone peers and to be more aggressive (Tangney, Wagner, Barlow, Marschall, & Gramzow, 1996; Tangney, Wagner, Fletcher, & Gramzow, 1992). For example, in a sample of 363 fifth-grade children, Tangney et al. (1991) found that for the fifth-grade males, proneness to shame was positively related to teacher reports of aggression.

Accordingly, shame can be aroused after transgression, and it can lead to aggression for some children. It is reasonable to suspect that shame is related to the repetition of aggression.

Empathy. Empathy refers to a vicarious affective response caused by others' emotional, physical, or psychological states. Researchers have suggested that empathy is the basis of moral motivation. Empirical evidence has shown that empathy is positively associated with prosocial behavior (Batson et al., 1988) and negatively related to aggressive behavior (Ellis, 1982; Richardson, Hammock, Smith, Gardner, & Signo, 1994).

In their review of literature on the relation between empathy and aggression, Miller and Eisenberg (1988) found a low to modest negative relation between empathic responsiveness and aggressive behavior. However, the negative relation was significant when empathy was assessed by questionnaire and by picture/story method if preschool









children were excluded but not significant when other empathy measures were used. That is, this relation could be influenced by age, methods of assessing empathy, and criteria for aggression.

Feshbach (1983) claimed that empathy is an inhibitor of aggression because empathy fosters prosocial behaviors that are incompatible with aggression and because the pain of the victim of aggression may elicit distress in the bystander or even in an aggressor with empathic capability. An intervention program for empathy training has been found effective in reducing the occurrence of aggression (Feshbach, 1983).

Empathy also has been described as substantially

related to the development and regulation of guilt and shame. According to Hoffman (1982), empathy occurs in response to the distress of others, whereas guilt results from the combination of empathic response and the awareness

of being the cause of others' distress; thus, empathy and guilt are proposed to have a similar developmental path with empathy as a prerequisite of the development of guilt. Tangney (1991) argued that both empathy and guilt are otheroriented, whereas shame is self-oriented; thus, empathy should be positively associated with guilt after a transgression but negatively related to shame. This







14

suggestion has been supported by empirical studies (Tangney, 1991). Therefore, when children harm others empathetic responses may be evoked, and these feelings could influence their subsequent behaviors and other emotions.

Regulation. In recent years, researchers have

identified individual differences in the dynamic features of emotion, including the intensity, persistence, modulation, onset and rise time, range, and lability of and recovery from emotional responses (Thompson, 1994). Thompson indicated that these dynamic features of emotion are substantially influenced by emotion regulation processes, which he defined as "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). Thus, understanding children's emotion regulation is crucial in explaining the qualities and dynamics of their emotional experiences.

Franko, Powers, Zuroff, and Moskowitz's (1985) study demonstrated that children as young as 6 years can describe conscious strategies for dealing with negative affect. Of 32 children, aged 6 to 11, who participated in their study, 92% showed active rather than passive attempts at selfregulation of emotion. Thus, it is reasonable to expect that









children may employ self-regulation strategies to relieve their feelings of discomfort such as guilt, shame, or fear.

Children may employ behavioral or cognitive responses to deal with negative affect (Franko et al., 1985). Lazarus (1991) suggested that the action tendency with guilt is to expiate by apology or making amends and the tendency in shame is to hide, to avoid having one's personal failure observed by others.

Cognitive Reactions to Aggression

Some researchers have explored the cognitive reactions of children following moral transgression. Their research shows that the beliefs children hold about the outcomes of their misbehavior and the ways they explain and justify their wrongdoing may influence the maintenance of the behavior.

Outcome expectancies. Outcome expectancies refer to

the rewarding or punishing consequences one anticipates for performing a behavior (Perry et al., 1986). Social cognitive theorists have emphasized that from observing the effects of acts of aggression individuals develop outcome expectancies for aggression (Bandura, 1973). Crick and Dodge (1994) stated that "outcome expectancies can serve an excitatory or an inhibiting function depending on whether the outcomes







16

expected for particular behaviors are positive or negative" (p. 89).

Research has demonstrated that, compared to

nonaggressive children, aggressive children tend to be more confident that aggression will produce positive consequences, such as obtaining a tangible reward, reducing aversive treatment by others, and increasing self-esteem (Perry et al., 1986; Slaby & Guerra, 1988; Xie, 1991; Zhang, 1990). Thus, aggressive children might develop expectancies that support future aggression from experiencing positive outcomes of their behavior.

Legitimacy. Theorists have emphasized the important role children's beliefs about the appropriateness of aggression play in the persistency of aggression. Slaby and Guerra (1988) argued that from experiences children may develop the belief that aggression is an acceptable behavior in many situations. Huesmann and Eron (Huesmann & Eron, 1984; Eron, 1994) have proposed that social behavior is controlled to a great extent by cognitive scripts, or schema. These scripts are encoded rehearsed, stored, and retrieved in one's mind as are other intellectual behaviors. An aggressive strategy or script must be accepted or considered as appropriate by an individual before it can be encoded. Advancing this theory, Huesmann and Guerra (1997)







17

suggested that through socialization processes an individual develops normative beliefs, which refer to a person's cognitive standards about the acceptability of a behavior. These beliefs set the range of allowable and prohibited behavior. According to this theory, once an aggressive script is stored in one's memory, one may retrieve it, whenever one faces interpersonal problems or is in a situation similar to the one in which the script is encoded. When a script is retrieved, the individual evaluates its appropriateness in light of existing internalized norms and considers the likely consequences. A child with weak or with no internalized prohibitions against aggression, or who believes it is normative to behave in this way, is much more likely to respond with aggression.

Research findings show that aggressive individuals are more likely to believe that aggression is a legitimate response (Eron, Guerra, & Huesmann, 1997; Huesmann & Guerra, 1997; Slaby & Guerra, 1988; Zhang, 1990). In a longitudinal study, Huesmann and Guerra (1997) concluded that "children tended to approve more of aggression as they grew older and that this increase appeared to be correlated with an increase in aggressive behavior" (p. 413). In addition, Guerra and Slaby (1990) found that changing adolescent offenders' belief that aggression is legitimate was more









effective in reducing aggression than changing other cognitive skills and beliefs included in their intervention program. All these findings suggest that children's beliefs about the legitimacy of aggression may influence their subsequent social behavior.

Attributions of responsibility. Aronson (1995)

suggested that people tend to justify their cruelty. When people clearly harm others, they feel dissonant because their behavior is not consistent with their belief that they, are good or decent people. In this situation, the best way to resolve the dissonance is to maximize the culpability of

the victim and convince themselves that the victim deserved the injury.

Bandura (1991) proposed a theory to explain how people use mechanisms of moral disengagement to avoid moral sanction. According to him, people translate moral reasoning into moral behavior through self-regulation. The selfregulation exercises self-sanction in accordance with the positive or negative consequences of a particular behavior. The mechanism of self-regulation needs to be activated to operate. Conversely, people may use psychological mechanisms to disengage in the exercise of moral agency to avoid negative feelings from self-sanction. These mechanism of moral disengagement include describing their injurious







19

behavior as serving a valued social or moral purpose, using euphemistic language to mask aggression, comparing their detrimental conduct to more culpable behavior, displacing responsibility, diffusing responsibility, disregarding or distorting consequences of their harmful actions, and dehumanizing the victims.

Researchers (Graham, Doubleday, & Guarino, 1984;

Graham, Weiner, & Banish-Weiner, 1995; McGraw, 1987) have found that after transgression if people consider that the situation is out of their control or they attribute responsibility for the misbehavior to others, they are less likely to feel guilt. Research by Bandura, Barbaranelli, Caprara, and Pastorelli (1996) provided evidence that moral disengagement fosters harmful conduct by promoting cognitive and affective reactions conducive to aggression.

Therefore, after a transgression children might

develop attributions to justify their wrongdoing or reduce self-blame, and these reactions might have a substantial impact on the recurrence of the transgression. These cognitions are important for understanding how aggressive and nonaggressive children differ in their reactions to their own misbehavior.









The Role of Intelligence in Maintenance of Children's
Aqqression

Researchers have claimed that intelligence also has an impact on the persistence of aggression. In their longitudinal study, Huesmann, Eron, and Yarmel (1987) found a bidirectional relation between intelligence and aggression; that is, early IQ was related to early aggression, and early aggression may have impeded the development of intelligence afterward. After reviewing research literature on the relation of IQ to delinquency, Hirschi and Hindelang (1977) proposed that IQ has an impact on the likelihood of delinquency independent of race and class. Gibson and West (1970) also claimed that IQ is an effective predictor of delinquency independent of socioeconomic status. According to Miller (1972), all kinds of deviant behavior increase sharply among children with IQ. lower than 90.

The Role of Impulsiveness in Maintenance of Children's
Aqgression

Impulsive children tend to act out the first strategy that comes to mind. Camp (1977) postulated that impulsive children are limited in their ability to inhibit their first responses to stimuli and are deficient in maintaining sustained response inhibition. Thus, impulsive children are less likely to think of alternative strategies in









interpersonal situations. Berkowitz (1974, 1993) empasized that in many instances aggressive behavior is an impulsive or involuntary response to internal and external stimulation. He proposed that due to intense emotional agitation or personality, some people do not stop to think about what might happen if they victimize others. They have an urge to attack an available target. A number of studies (Atkins, Stoff, Osborne, & Brown, 1993; Luengo, Carrillo-dela-Pena, Otero, & Romero, 1994; Martin et al., 1994) have found that aggression is associated with poor impulse control. Luengo et al. (1994) also demonstrated that impulsivity is related to a future increase in antisocial behavior.

In summary, researchers agree that intelligence and impulsiveness may influence children's maintenance of aggression. Therefore, a study of aggressive children should take these two constructs into account. Age and Children's Reactions to Transgression

Researchers have claimed the existence of age

differences in children's development of guilt, shame, and empathy. Mascolo and Fischer (1995) proposed a theory of children's developmental trajectories of pride, shame, and guilt based on appraisal patterns, which produce these selfevaluation emotions. According to Mascolo and Fischer,









children as young as 6 to 8 years compare their concrete traits with those of their peers, and at the age of 10 to 12 children begin to experience shame about abstract or general personality characteristics. For guilt development, children

by about 10 to 12 years of age are able to feel guilty about violation of general moral rules. Therefore, children age 10 to 12 seem to have mature moral and self-evaluative emotions.

Hoffman (1982) maintained that age-related improvement

in children's cognitive ability is conducive to older children's experience of guilt or empathy in a broad range of situations. Young children, at the third stage of empathy development, according to Hoffman's theory, can experience empathy for others' feelings, whereas children of late

childhood, the fourth stage of empathy development, can also respond empathically to others' life conditions. Comparing

first-, third-, and fifth-grade students, Thompson and Hoffman (1980) found that older children reported feeling more guilt, more concern for the victim's welfare, and relied more on principles of right and wrong in explaining their feelings of guilt; younger children were more concerned about detection and punishment and were more likely to report feelings of happiness as a result of the wrongdoing.







23

In regard to cognitive reactions, Huesmann and Guerra (1997) found developmental differences in children's normative beliefs about aggression. For first graders, normative beliefs were not stable; approval of aggression showed its largest increase between the first and second grades. It appears that the first 2 years of elementary school are an important period for children's development of normative beliefs. In the fourth grade, normative beliefs about aggression become stable and predict later aggressive behavior.

Therefore, younger and older elementary school

children may differ in the situations in which they feel shame, guilt, and empathy and in beliefs about aggression. Comparing age differences will allow us to learn more about how children react to their misbehavior and the impact of these reactions on their subsequent development of aggressive behavior.

Shame and Guilt in Chinese Culture

Influenced by the belief that Western societies are guilt oriented and Asian societies are shame cultures, psychologists, sociologists, and anthropologists have been interested in the meanings and implications of guilt and shame in various cultures. Chinese culture has been considered a shame culture. Many studies have been conducted







24

to investigate whether feelings of shame and guilt as basic emotions are different in Chinese and Western people and how guilt and shame are used as a mechanism of social and behavioral control in Chinese society. However, due to the difficulty of cross-cultural study, research findings are inconsistent.

Shame and qcuilt-oriented discipline. Wilson (1981), a sociologist, proposed a theory of moral behavior in Chinese society from a social political perspective. According to him, China is a closed and group-oriented society that stresses social order and stability while discouraging individual independence and self-direction. Respect for authority and fixed rules is emphasized. In such an environment, indoctrination of group-centered values has been the focus of moral training, and physical punishment and love-withdrawal have been used as a means of affective manipulation. The outcomes of this type of training are fear, shame or guilt, and deference. When comparing one's own behavior with the ideals manifested by authority or with other members in one's group, Chinese people are vulnerable to feelings of shame. Shame anxiety thus operates as a major dynamic in Chinese group interactions.

From 2 years of observation and interview, Fung (1994) concluded that shame continues to play an important role in









disciplining young children in contemporary modernized Taipei, Taiwan. Parents manipulated the immature sense of shame in their children as a major means of teaching their children right from wrong and transmitting cultural values. Parents believe that this socialization protects children from disgrace-shame, especially from being condemned by persons outside the family. Thus, although shaming implies threats of abandonment or ostracism, it may teach children how to be part of a society, that is how to be included rather than to be set apart. Findings of Her's (1990) phenomenological analysis also suggest the importance of shame as a socialization mechanism in a group-oriented society. She concluded that shame in Chinese culture functions as a confirmation of group membership more than a self-actualization process.

Although researchers have found that shame was used by Chinese parents, one study did not show this result. Stevenson, Chen, and Lee (1992) asked mothers and children in a village in Taiwan what the parents would do if their child got a poor grade or misbehaved. Neither mothers nor children responded that the parents would use shaming techniques. The researchers suggested that perhaps shame is used more frequently in public settings such as the









classroom rather than in the family and by other agents outside the family.

Besides using shame to discipline, Chinese parents are described as stronger in authoritarian control and lovewithdrawal techniques than American parents (Chao, 1994; Kriger & Kroes, 1972). Eberhard (1967) pointed out that love-withdrawal and punishment also facilitate children's shame and conformity. According to him, in societies using more punishment or love-withdrawal techniques, children feel they are the recipient of actions of others in authority; thus, the most important motive for these children is to avoid shame and the anger of others.

In a review of literature on Chinese socialization

processes, Ho (1986) pointed out that threatening, scolding, shaming, and punishment have been used frequently with older Chinese children. However, these techniques are adopted less often by modern Chinese parents. In reality, a variety of socialization techniques rather than a single pattern are used by Chinese parents.

Accordingly, some Chinese parents may adopt shaming as a major discipline technique and others may use different methods. However, even if Chinese parents do not use shaming socialization, their children still appear to be more likely to develop a shame personality than children from Western









cultures, through the use of love-withdrawal, physical punishment, and authoritarian and shaming punishment in school (Stevenson et al., 1992; Wilson, 1981).

Meanings of shame and quilt for the Chinese. Theorists of discrete emotion (Izard, 1977) have maintained the existence of basic emotions, including shame and guilt, which are consistent across cultures. In contrast, social constructionist theories of emotion have posited that an individual's emotional patterns reflect the influences and functions of the social community (Lazarus, 1991). Researchers have been interested in whether the connotative meanings of a given emotion differ between cultures. A number of studies have been conducted comparing the meanings of shame and guilt between Chinese and Western people, and both commonality and differences have been reported.

Studies have shown differences in the ways that Chinese and Americans view shame. Using the semantic differential technique, which minimizes the use of culturally loaded statements by employing opposite adjective pairs to assess subjective experience across cultures, Marsella, Murray, and Golden (1974) found that Caucasian Americans rated shame as more low, weak, dull, rounded, relaxed, and stale than did Chinese Americans. For the Caucasian Americans, shame was less identifiable or









understood, although both groups experienced guilt as an uncomfortable feeling. The researchers stated that shame may be used as a technique of social control more often in Asian societies and children consequently learn to read the cues of shame more readily.

Research by Cheng and Page (1995) showed that Chinese participants rated guilt as a more potent force and viewed guilt in a more negative way than did the American participants. Cheng and Page indicated that the results were not unexpected because Chinese culture often uses shame and guilt as a means of controlling the behavior of individuals and Chinese people may experience guilt more strongly and unpleasantly than Americans.

On the other hand, some research suggests commonality between Chinese and Western people in meanings of shame and guilt. Bond and Hwang (1986) argued that Chinese and Americans experience emotions in the same way; they only differ in emotional expressions due to different display rules.

In their study, Hong and Chiu (1992) asked Chinese students in Hong Kong to recall either a guilt or a shame incident. They found that more guilt-eliciting events than shame-eliciting events were associated with violating a moral norm, holding personal responsibility for the









violation, and absence of an audience. This result is consistent with findings from studies for Western people. However, a high proportion of shame incidents were recalled, including violating a moral norm, absence of an audience, and having personal responsibility, showing that guilt and shame incidents were not mutually exclusive.

As indicated previously, guilt has been found to be an inhibitor of aggression from research findings in Western societies. In Taiwan, Su (1975) and Hong (1985) also found that guilt was negatively related to aggression for middle school students.

Johnson et al. (1987) compared the meanings of guilt and shame for American, Korean, and Taiwanese participants. The Taiwanese were highest on both the guilt and shame measures. However, consistency across national groups was also found. There was consistency among the groups on whether individual test items loaded on the guilt or the shame factors and on the values of the loadings. The crosscultural similarities led the researchers to "question the belief that Asian and Occidental societies differ in the degree to which guilt vs shame are used as mechanisms for social control" (p. 357).

Some anthropologists have questioned the dichotomy of shame and guilt cultures. Eberhard (1967) indicated that









There are hardly any pure shame or pure guilt cultures,
that shame and guilt both are utilized to ensure
socialization of the individual, and if a culture is
called a shame culture this should only indicate that
shame is a more prominent agent than guilt, not an absence
of guilt. (p. 3)

From a cultural study Singer (1953) concluded that "the sense of guilt and the sense of shame are found in most cultures" (p. 79). He believed that people in industrialized and preindustrialized societies do not differ in guilt sanction or shame sanction but in the complex of values, beliefs, and practices.

In a broad and in-depth comparison of Chinese and

American cultures, Hsu (1965) proposed that the guilt-shame contrast is better understood as an expression of the contrast between a repression-dominated and a suppressiondominated way of social control. In control by repression the rules of behavior are mostly within the individual's mind. Violation of the rule is more likely to call forth the feeling of guilt than anything else, with reference to the superego. In control by suppression, the rules of behavior are more strongly dependent upon the external situation. Violating rules is more likely to lead to the feeling of shame, than anything else, when one thinks of one's fellow human beings. "It is not alleged that any society can be wholly governed by only one of these mechanisms of control" (p. 177). He also indicated that the mechanism of repression









is more characteristic of Western societies, whereas the mechanism of suppression is more characteristic of Asian societies such as China.

Thus, it is reasonable to conclude that there is no pure shame society or pure guilt society. Both guilt and shame are present in Chinese and Western societies. The nature of shame as an uncomfortable affect and the antecedents of shame are the same across these cultures. The differences are that with different social structures, values, and practices, Chinese people have more chances to experience shame and, consequently, are more sensitive to shame, experience a more intensive response to shame, and rely more on external resources for controlling behavior than Western people. This by no means implies that guilt is unimportant for Chinese people or that they feel less guilt than they feel shame. In terms of moral emotions, Chinese people feel guilt when they violate a moral norm, and guilt seems to function as an inhibitor of prohibited behavior; the feeling of shame facilitates learning right from wrong and provides aversive feedback when one acts inappropriately. Both guilt and shame are critical for moral development in Chinese people.









Summary

Childhood aggression is a serious problem in most

industrialized societies. The research literature shows that aggression is relatively consistent across time and situations for aggressive persons from an early age. In order to alleviate the problem of aggression, it is necessary to interrupt the continuity of aggression in addition to preventing it. Researchers have identified the following social-cognitive variables in the development of aggression in children: outcome expectancies, legitimate beliefs, and justification. Other researchers have identified the social-emotional variables of empathy and guilt as influential in reducing aggression. Shame and fear have been conceived as important for Chinese children to prohibit inappropriate behavior. The purpose of this study is to identify the differences between aggressive and nonaggressive children in their cognitive and emotional reactions to their own transgression.














CHAPTER 2
RESEARCH DESIGN

The purpose of this study was to explore the

differences between aggressive and nonaggressive boys in their emotional and cognitive reactions to their own misbehavior, in order to better understand why aggression is repetitive in some boys but not in others. In this chapter, I describe the research participants, the procedures, the measures and the analyses used in this study.

Research Participants

Participants were 100 third- and 100 sixth-grade boys from four public elementary schools in a medium-sized city in central Taiwan. One school was in a recent-developed residential district serving a predominantly middle-class community with about 3,600 students. The second school was located in a suburban area serving a lower-middle-class neighborhood with approximate 2,000 students. The other two schools, with about 1,600 students each, were in the downtown area to serve socioeconomically diverse classes.

Classes in these schools were used as the unit to

recruit participants. Aggressive and nonaggressive boys were selected from these classrooms on the basis of teacher ratings. The raw scores of the teacher ratings were









transformed into standard scores. A boy was identified as aggressive if his z score rating of aggression was above 1.3 and his raw score was above 18, from a scale with possible scores ranging from 8 to 40. A total of 50 aggressive boys were selected in each third and sixth grades.

The number of nonaggressive boys was the same as the number of aggressive boys. They were selected from the boys whose z score derived from the teacher rating was below

-0.25 and whose raw score was at least 8 points lower than the raw scores of the aggressive boys in his class. Whenever possible, nonaggressive boys were selected from boys who met these criteria and were matched with one aggressive boy in his class on socioeconomic status. If no nonaggressive boy in a class matched the aggressive boy on SES, the nonaggressive boy who had the SES score closest to the aggressive boy in his class was selected. Socioeconomic status has been conceived as an influential factor in children's aggression (e.g., Eron et al., 1997; Farrington, 1991; Offord, Boyle, & Racine, 1991). Reducing the differences between the two groups in SES was used to avoid confounding the results of this study.

Measures

Aqqression

Teachers completed the 8 items on the aggression factor of the Teacher Checklist (Coie & Dodge, 1988):









threatens or bullies others; uses physical force; starts fights; overreacts to accidental hurts with anger and fighting; gets angry easily and strikes back; gets other kids to gang up on a peer; says mean things to peers; claims other children are to blame in a fight (Appendix A). Teachers rated the aggression of each boy in their class whose parents consented to participation in this study on a 5-point scale from 1 (not at all) to 5 (very much), with scores ranging from 8 to 40, and higher scores indicating more perceived aggression.

Reactions Following Transgressions

Children who met the criteria for selection read three semiprojective stories describing explicit misbehavior and then were interviewed concerning how they would feel and

think were they the perpetrators of the misbehavior. Each story depicted an incident involving boys of the same age as the participant. One of the boys in each story commits a transgression, in the first story cheating at checkers, in

the second stealing another child's toy, and in the third neglecting to help another after an accident (Appendix B).

Stories with similar content were used in other studies of the development of children's moral emotions (Dunn, Brown, & Maguire, 1995; Kochanska, 1991; Krevans & Gibbs, 1996; Thompson & Hoffman, 1980).









First, the children were asked to complete the story, explaining how the protagonist thinks and feels and what would happen next? The story completion is designed to assess the intensity of the child's guilt (Hoffman & Saltzstein, 1967) and extent of reparatory acts (Kochanska, 1991). All children then were asked to assume the roles of the wrongdoers and to answer questions designed to assess their reactions to misbehavior in each story. Interview questions (Appendix C) were designed to measure feelings, emotion regulation, empathy, responsibility attribution, outcome expectations, and beliefs about the legitimacy of the transgression.

The interviewer showed participants a scale with five bars of increasingly greater height. Boys were asked to point to the bar that illustrated the intensity of their feelings, with the highest bar representing very much and a blank showing not at all. For questions asking children "would you (do) . . .?" the interviewer showed them a 5point scale representing "Definitely not," "Probably not," "Not sure," "probably yes," and "Definitely yes." For the two questions on outcome expectancy the interviewer presented children a 4-point scale and requested children to choose their responses from the scale (see Appendix C).









Socioeconomic Status

The Chinese version of Hollingshead's two-factor index of social position developed by Ying-Shin Yang (1985) was used to compute children's socioeconomic status. By this method, parents' education was divided into five levels and occupation into six categories. A formula was used to calculate each boy's socioeconomic status: Index of SES = index of parents' education x 4 + index of parents' occupation x 7. Fathers' information was used for this calculation. When fathers' information was not available, mothers' information was used instead. Impulsiveness

The 16 Impulsiveness Questionnaire developed by

Eysenck, Easting, and Pearson (1984) was used to assess participants' impulsiveness. This questionnaire contains 69 items, but only the 21 items that measure impulsiveness were used. This measurement is applicable for children from the ages of 7 to 15, and the reliability of the impulsiveness scale for boys was .74 in the Eysenck et al. (1984) study of 633 British boys.

Intelligence

The scaled scores of the Vocabulary and Block Design subtests of the Wechsler Intelligence Scale for Children (WISC) were used to assess the participants' general







38

intellectual abilities. These two subtests have been used as a measure of general intellectual ability in previous research (e.g., Klaczynski & Gordon, 1996). According to Sattler (1989), the pro-rated IQ scores derived from these two subtests are highly correlated with the WISC full-scale score (r = .90). The vocabulary subtest, which asks children to explain word meaning, is an excellent indicator of crystallized intelligence. The block design subtest, which requires participants to reproduce three-dimensional designs, is a measure of the ability to apply reasoning to solve spatial problems. In this study, these two subsets of the Chinese version of the third edition of the WISC were used. The original WISC-III has been translated into Chinese, revised, and normed in Taiwan.

Procedures

Teachers of the third and sixth grades in the four

elementary schools were invited to participate in this study and asked to sign the teacher consent form (see Appendix D). Before selecting the aggressive and nonaggressive children,

parental consent forms (see Appendix D) that included the SES questions were given to the boys in the third and sixth grades by their teachers in the four elementary schools. Parents of the boys were asked to sign the consent form and respond to the SES questions if they agreed to allow their









sons to participate in this study. In total, 88.6% of the parents agreed to their sons' participation.

Teachers of these classes were asked to rate each

participant on the Teacher Checklist. I then interviewed the boys who met the selection criteria. In 5 days after the interview these children were asked to complete the impulsiveness questionnaire. Two days to 1 week after the interview, these boys took the intelligence test.

Data Analysis

Coding of Participants' Responses

The responses to the story completion were scored on a 6-point scale representing the extent of reparatory acts (Kochanska, 1991).

Score 1: No resolution of transgression; no discomfort or any action.

Score 2: External resolution of transgression;

external sources or circumstances intervene; victim may retaliate.

Score 3: Attempt to address the transgression; trying to alleviate distress of the victim, but the action is incomplete.

Score 4: Attempt at reparation: giving up something valued to make up for the wrongdoing or making reparation (beyond mere apology or confession).

Score 5: Reparation and attempt at reconciliation.









Score 6: Full resolution of transgression; carrying out a complete and relatively excessive effort to make reparation and to alleviate physical or emotional consequences of wrongdoing.

Statistical Analysis

The scores obtained for each question across the three stories were added together. Scores from one interview question were treated as scores for one variable. MANCOVA and MANOVA procedures were used to analyze data whenever appropriate. Fishers' exact test was also used to analyze the categorical data.

Limitations of the Study

Because of the research design, there are some

limitations of this study. First, because the interview questions are not all free-response format, data collected in this study may not show the full range and dynamics of children's reactions. Another limitation comes from the use of children's responses to semiprojective stories as their reactions to transgressions rather than responses to real life incidents. Although this method is used pervasively, researchers have not examined the validity of the results. Finally, children's emotions and thoughts probably interact with each other. This study treats emotion and cognition separately; thus, the study cannot determine the joint









influences of the two aspects on children's aggressive behavior.

Pilot Study

A pilot study was conducted to determine whether the interview questions proposed for the study were appropriate for use with boys in the third and sixth grades in Taiwan. The interview questions were revised on the basis of the boys' reactions during the interview and the results of the t-tests comparing the responses of the aggressive and nonaggressive boys.

Participants

Seven aggressive (four sixth graders and three third

graders) and five nonaggressive (three sixth graders and two third graders) boys participated in this pilot study. Participants were recruited from two sixth-grade and two third-grade classes at an inner-city elementary school in central Taiwan. They were selected based on teacher ratings and peer nominations. A boy was identified as aggressive when his teacher rating was above the median for his class, his peer nomination was above the 80th percentile, and he had at least twice as many aggressive as prosocial nominations. The nonaggressive children were selected from children with a teacher rating below the class median, peer nomination below the 80th percentile, and SES matching one of the aggressive boys in his class.









Procedures

Two sixth-grade teachers and two third-grade teachers were asked to respond to the teacher checklist for each boy in their classes. Boys in these classes were asked to answer the peer nomination questions and referred to a list of other boys in their class. After selecting the participants according to the criteria described above, I interviewed each boy individually, using prepared questions. Children's replies were written down and tape-recorded. Then, the story completion was coded by another rater and me. Finally, the pilot study data were analyzed by t-tests and Pearson correlations.

Results

Differences between aqqggressive and nonaggressive boys.

The pilot study data were analyzed by I-tests to compare aggressive and nonaggressive boys' responses. Table 1 shows the means on each question for both groups and results of the t-tests. Because the sample size was small, it was unlikely that differences would be statistically significant, but differences between group means could reveal whether the direction of differences is consistent with the study's hypotheses. On the 26 interview questions, differences of group means for 15 were in the expected direction: fear, guilt, sadness, shame, time (time thinking









Table 1

Differences between Aggressive and Nonaggressive Children in Reactions to Moral Transgression


Variable


Fear
Agg
Nagg
Guilt
Agg
Nagg
Sad
Agg
Nagg
Happy
Mg Nagg
Shame
Agg Nagg
Proud
Agg
Nagg
Disturb days
Agg
Nagg
Disturb eat
Agg
Nagg
Time
Agg
Nagg
Apology
Agg
Nagg
After Apology
Agg
Nagg
Hide
Agg
Nagg
After Hide Agg
Nagg


9.57
11.00

8.42 9.60

7.42 9.00

3.14 3.80

8.85 9.80

3.14 3.00

8.17 8.25 5.50 11.50

13.66 18.25

10.42 12.20

7.00 6.75

4.00 6.40


8.50


Prob> IT 0.4620 0.5506


0.4153 0.2286 0.5712


0.4241 0.9599


0.0074 0.4343 0.3625


0.9110 0.1607









Table 1--continued

Variable N

Help
Agg 7 Nagg 5
After Help
Agg 4 Nagg 4
Empathy A
Agg 7 Nagg 5
Empathy B Agg 6 Nagg 5
Empathy C Agg 7 Nagg 5
Friend
Agg 6 Nagg 5
Gain
Agg 6 Nagg 5
Enjoy
Agg 7 Nagg 5
Bad
Agg 6
Nagg 4
Responsibility
Agg 7 Nagg 5
Blame
Agg 7 Nagg 5
Other Respn
Agg 7 Nagg 5
Other Blame
Agg 7
Nagg 4


M


9.71 11.60

4.50 6.25

11.00
12.40

10.83
12.00

11.29 12.80 10.00 11.60

10.66 11.60 5.00 5.60 12.33 12.25

12.14 12.80

10.85 11.80

4.14 4.40

3.28
4.25


Prob > T

0.3246 0.2215 0.2768 0.3395 0.3168 0.0883 0.1880 0.6966 0.9517 0.5855


0.4484


0.77


0.0963









persisted), disturbing eating, apology, help, after help, Empathy A (victim upset), Empathy B (victim sad), Empathy C (feel bad about other's hurt), friend like/not like, gain/not gain, and self-blame. For two questions data were in the opposite direction: happy and other blame. The mean values were very similar for five questions: proud, disturbing days, enjoy, bad, and other's responsibility. Only two aggressive children answered that they would hide after a wrongdoing; therefore, only these two boys were asked the "after hide" question, and data were not available for calculating differences in responses to this variable.

Differences between third and sixth graders. The second set of t-tests were conducted to analyze age differences between third and sixth graders. Table 2 presents the means on each question for the two grades and the results of the t-tests. The sixth graders scored higher than the third graders on seven questions: guilt, shame, disturb days, time, after help, and Empathy C (feeling), and bad; these were in the expected direction. On four questions sixth graders scored lower than third graders: disturb eating, apology, help, and Empathy B. There were no apparent grade differences in the answers of the other 15 questions (differences between the group means were less than 0.5 or the R values higher than .70).









Table 2


Differences between Moral Transgression


Third and Sixth Graders


in Reactions to


Variable N M Prob> II


Fear 3rd
6th
Guilt 3rd 6th
Sad 3rd 6th
Happy 3rd 6th
Shame 3rd 6th
Proud 3rd 6th
Disturb days 3rd 6th
Disturb eat 3rd 6th
Time 3rd 6th
apology 3rd
6th
After Apology 3rd 6th
Hide 3rd
6th
After Hide 3rd 6th


10.20 10.14

7.80 9.71

8.20
8.00

3.20 3.57

8.20 10.00

3.20 3.00

6.67 8.86

8.67 7.57

12.50 17.80

11.80 10.71


0.9769 0.3215


0.8053


0.5074 0.2684 0.2550


0.1911 0.7127


0.4072 0.5820


0.8814


0.7023


7.00 6.67

4.60
5.28

8.00 9.00


Thir an Sixth Graders in Reactions to









Table 2--continued


Variable N

Help
3rd 5 6th 7
After Help 3rd 5 6th 3
Empathy A 3rd 5 6th 7
Empathy B 3rd 5 6th 6
Empathy C 3rd 5 6th 7
Friend
3rd 4 6th 7
Gain
3rd 4 6th 7
Enjoy
3rd 5 6th 7
Bad
3rd 3 6th 7
Responsibility 3rd 5 6th 7
Blame
3rd 5 6th 7
Other Respn 3rd 5 6th 7
Other Blame 3rd 4 6th 7


M


11.20 10.00

4.60 6.67

11.80 11.53

12.20 10.67 11.00
12.57

11.00 10.57

10.75
11.28

5.60 5.00

11.67 12.57

12.20 12.57

11.00
11.43

4.40 4.14 3.75 3.57


Prob> IT 0.5371


0.1521 0.7792 0.1997


0.2978 0.6835


0.4812 0.6966


0.5329 0.7591 0.7335


0.7506 0.7756









Correlations between scores on interview questions.

Table 3 shows the Pearson correlations between scores on the interview questions. It reveals that most of the significan relationships were consistent with expectations drawn from extant literature. However, there are two exceptions: the negative relationships between after help and Empathy B (-0.73) and between after apology and time (-0.91) were not in the expected direction. Children's responses to the question, "How long would you keep thinking about this incident," ranged from several hours to their whole life, making it difficult to categorize their answers. In the revised questionnaire children will be asked to indicate on the scale of five bars of increasingly greater height to point to the bar that illustrates the length of time they would think about the incident with the highest bar representing several years and a blank indicating not at all. Again, the "after hide" question consists of the responses of only two children. Revision of these interview questions seems to be needed.

Ratings of story completion. The researcher and another doctoral student rated participants' story completion narratives on a 6-level scale developed by Kochanska to measure children's intensity of guilt. First, we read and discussed the criterion for each level. Then, we












Table 3


Correlations Between Scores on Interview


fear guilt sad hapy sham proud both unea time aplgy aapl hide ahi


0.70* 0.84** 0.20

0.82** 0.24

-0.05


fear guilt sad hapy sham proud dday eat time aplgy aapl

hide ahi

help

ahp empa emph empc frnd

gain enj bad rspn blam orspn oblam


0.62* 0.29 0.77** 0.11

0.72** 0.19

-0.05 -0.15

0.09


0.13 0.32

0.21

-0.10

0.66*

-0.48


0.67* 0.36

0.63

0.22 0.52

-0.08

0.27


0.58 0.68

0.59 0.36

0.57

-0.53

0.65

0.40


0.47

0.44

0.32 0.07

0.39 0.08 0.27 0.19

0.24


0.51

-0.02

0.31

-0.02

0.45 0.54

-0.75

0.24

-0.91*

-0.07


0.39

0.13

0.22 0.60

0.25

-0.22

0.35

0.74* 0.56

-0.12

0.06


-1.00

1.00

-1.00

1.00

-1.00



-1.00

-1.00

1.00

-1.00

-1.00

1.00


em Am.. ...ie~k


Questions












Table 3--continued


help ahp empa empb empc frnd gain enj bad rspn blam orsp obla


0.42 0.50

0.49

-0.24

0.42 0.15 0.17 0.27

0.21

0.86*

-0.19

-0.29

-1.00


0.27 0.30 0.38

0.04 0.63

-0.29

0.41 0.63 0.09

*-0.31

0.54 0.50



-0.35


0.55 0.70* -0.27

0.86**-0.34

-0.38


-0.24

-0.12

-n 15

0.89**


0.70*

0.69*

0.52

0.53 0.39

0.21 0.05

0.41 0.52

0.73**

0.03 0.27 1.00

0.60*

-0.48


fear guilt sad hap shamn proud dday eat time aplgy

aapl hid ahi help

ahp empa empc
ampc frnd

gain


0.27 0.78** 0.24 0.25 0.88** 0.33 0.31 0.73** 0.51 0.23 0.26 -0.04 0.21 0.82** 0.58 0.28 0.01 0.06

-0.01 0.57 0.26 0.44 0.53 0.70* 0.36 0.92** 0.35 0.54 0.63* 0.20

-0.26 -0.05 0.00 0.29 0.35 0.34 1.00 1.00 -1.00 0.56 0.58 0.41

-0.73* 0.23 0.42 0.71* 0.76** 0.20

0.37 0.71*

0.39


0.24 0.23 0.75* 0.64* 0.63* 0.19 0.30 0.75**-0.29 0.65* 0.68* 0.72* -0.44 -0.29 0.54 -0.17 0.74* 0.54 0.56 -0.14 0.07 0.24 0.52 0.15 0.05 0.34 0.06 -0.02 0.62* -0.25 0.59 0.56 0.55 -0.23 0.03

-0.03 -0.03 0.13 0.09 0.12 -0.06 -0.23

0.29 -0.02 0.06 0.17 0.12 0.01 0.23 0.29 0.36 0.49 0.22 0.36 0.38 0.67* 0.51 0.13 0.63 0.39 0.62 -0.05 0.21 0.41 -0.17 0.12 0.82** 0.65* -0.25 0.02

-0.61 0.25 0.47 0.26 0.09 0.41 0.24 0.06 0.75** 0.25 -0.15 0.16 0.54 0.62* 1.00 1.00 . -1.00 1.00 -1.00 -1.00 0.51 -0.47 0.14 0.81** 0.65* -0.40 -0.02 0.21 0.04 0.42 -0.06 -0.24 0.28 0.42 0.48 0.14 0.33 0.74** 0.82**-0.16 -0.04 0.50 0.01 0.22 0.35 0.64* -0.26 0.03 0.65* -0.04 0.56 0.73** 0.76**-0.16 0.08 0.63* -0.23 0.37 0.24 0.48 -0.24 0.21

-0.52 0.42 0.49 0.65* -0.74**-0.41

-0.14 -0.31 -0.16 0.84** 0.65*


enjy


bad

rspn blam orspn oblam









coded children's responses separately. We agreed on the coding of 91.17% of the children's responses (31 of the total 34). One answer of a boy in each group could not be coded because the boys failed to provide sufficient information; therefore, the total number of scores was 34, not 36.

For the first story, all five of the nonaggressive boys scored 2 (relying on external resolution). Among the seven aggressive children, four of them scored 2, two scored

3 (confess or apology), and one scored 4 (reparation). For the second story, all children received scores of 2, except one aggressive boy whose score was 4. For the third story, one nonaggressive child scored 3, and all the other nonaggressive boys scored 2. Among the aggressive children, two scored 3, four scored 4.

This suggests that nonaggressive children may rely more heavily on adults' solutions or are more afraid of retaliation from victims than aggressive children. The results seem inconsistent with findings from previous studies. The reason for the unexpected results may be because the stories used in this pilot study differed slightly from those used in Kochanska's (1991) study, as explained in the section that follows.









Revising Interview Questions and Stories

Three interview questions require children to answer whether after apologizing, hiding, or helping they would still feel discomfort (for example, after you apologize, would you still feel discomfort or not at all?). Children

were expected to reply "Yes" or "No" to these questions, but some children spontaneously responded with the degree of discomfort they would still feel after apologizing and helping. Therefore, the interview questions about feelings after apologizing and helping were changed to ask children to indicate whether the uneasy feelings would remain after their actions; that is, they were requested to choose their answer from "Definitely not" to "Definitely yes" on a 5point scale, as other "Would you (do) . . . ?" questions. On the other hand, because most children reported that they would not hide after this incident, questions about intention to hide and feelings after hiding were omitted.

For the question "What would you think to make you feel better after you did what XXX did?", children only answered "No" at first. Then I began to try prompting the children's response. After I prompted, some excuses for wrongdoing emerged in some children's responses (some still said "No"). For example, after the answer "No," I asked the boys, "Would you think that you are a better chess player, that you should win, or that moving two checkers won't be a







53

big deal?" One boy, then, gave the answer that Andy did not concentrate on playing checkers so he lost. Because the original question seems too broad for children to answer and because giving children prompts may create different effects on different children, 5 answers were provided and children were asked to choose 1 from the 5 (see Appendix C). These 5 answers represented 5 coping strategies, including comparing with others' wrongdoing, dehumanizing others, blaming others, thinking about compensation, and shifting attention.

One possible reason for the unexpected results on the story completion task is that the wrongdoers in the second and third stories need money to repair their wrongdoing. Children may fear that the reparation will cost their parents' money and they could be punished. In the revised stories no money is necessary to compensate for the wrong done and no audience is present during the misbehavior.















CHAPTER 3
RESULTS

The purpose of this study was to understand the

differences between aggressive and nonaggressive boys in their affective and cognitive reactions following misbehavior. Specifically, the study was designed to investigate whether after misbehavior aggressive and nonaggressive boys differ in the following reactions: types of feelings, intensity of feelings, length of emotional arousal, outcome expectations for aggression, legitimate beliefs in aggression, and attributions of responsibility. Differences between sixth- and third-grade boys in these reactions were also explored in this study.

After interviewing 200 participants and testing their intelligence and impulsiveness, a variety of statistical methods were conducted to answer the research questions. Chapter three presents the results and findings of the analyses of data examined in this study.

Statistical methods used in the data analyses included MANOVA, MANCOVA, ANCOVA, and Fisher's exact test. MANCOVA was planned as the major technique to analyze continuous data of this study. The use of MANCOVA needs to meet two









assumptions: First, the set of dependent variables is significantly correlated with the set of covariates, and, second, the regression of dependent variable(s) on the covariate(s) is equal across groups. The SPSS MANOVA program was used to examine data to determine whether it was appropriate to use MANCOVA (Stevens, 1996) to compare groups in this study. Because the two planned covariates, IQ and impulsiveness, were not highly correlated with most variables, MANCOVA was appropriate only in the analyses of the story completion and reactive feelings with IQ as covariate; impulsiveness was not used as a covariate. MANOVA was used for testing for group differences with multiple dependent variables. Each multivariate analysis was followed by univariate F-tests, in order to locate the variable(s) contributing to the multivariate difference. In analyzing data on cognitive coping strategy, Fisher's exact test was employed because these data were categorical and from a small sample.

Descriptive Data

Table 4 shows the means and standard deviations on each variable in this study (except the categorical variable, cognitive coping strategy) for the aggressive and nonaggressive boys in the sixth grade. These data for the third graders are displayed in Table 5. The correlation matrixes for all the variables are presented in Appendix E.









Table 4

Means and Standard Deviations on Each Variable for Sixth Graders


Aggressive Nonaggressive


Variable M N SD M N SD I.Q. 76.48 46 19.08 85.33 49 15.61 Completion 5.62 47 1.24 6.59 49 1.97 Fear 11.35 48 2.36 12.08 49 2.17 Guilt 11.29 48 2.68 12.47 49 2.06 Sad 8.73 48 2.83 10.27 49 2.46 Happy 4.21 48 1.66 3.69 49 1.04 Shame 10.65 48 2.51 12.18 49 2.38 Proud 3.81 48 1.38 3.57 49 0.98 Apology 13.23 47 2.15 13.94 49 1.45 After apology 6.21 47 2.99 7.47 49 3.37 Help 13.04 48 1.77 13.92 49 1.61 After help 6.38 47 3.01 6.59 49 2.81 Disturb life 10.23 48 3.11 11.43 49 2.67 Disturb two days 11.98 48 2.25 12.73 49 1.88 Disturbed time 10.85 48 2.24 11.31 49 2.05 Victim upset 11.94 48 2.13 12.41 49 2.19 Victim sad 11.00 48 2.29 11.88 49 2.16 Empathy feel 11.87 47 2.23 13.04 49 1.90 Friend 11.23 48 0.95 11.29 49 0.87 Gain 10.47 47 1.67 11.18 49 0.95 Enjoy 5.96 48 2.58 5.14 49 2.00 Bad 13.19 48 1.59 13.55 49 1.62 Self responsibility 13.31 48 1.68 13.43 49 1.44 Self blame 12.85 48 1.82 13.22 49 1.77 Victim responsibility 6.23 48 2.57 5.24 49 1.56 Victim blame 5.48 48 1.99 4.53 49 1.62









Table 5
Means and Standard Deviations on Each Variable for Third Graders

Aggressive Nonaggressive


Variable H N SD M N SD I.Q. 5.55 47 14.40 61.00 47 15.42 Completion 5.20 44 1.34 5.46 48 1.17 Fear 11.00 48 2.62 11.37 49 2.32 Guilt 11.17 48 2.57 12.02 49 2.15 Sad 9.73 48 3.47 10.29 49 3.03 Happy 5.19 48 2.63 4.20 49 1.62 Shame 9.54 48 3.53 10.59 49 3.21 Proud 4.90 48 2.15 3.98 49 1.71 Apology 13.46 48 2.25 14.22 49 1.46 After apology 4.74 47 1.89 5.08 49 2.37 Help 13.50 48 2.25 14.00 49 1.41 After help 4.36 45 1.82 4.67 49 2.03 Disturb life 8.68 47 3.95 9.49 49 3.64 Disturb two days 10.02 47 3.94 10.08 49 3.50 Disturbed time 9.96 48 2.87 9.94 49 2.90 Victim upset 12.44 48 2.20 12.39 49 2.48 Victim sad 11.13 48 3.12 11.59 49 2.65 Empathy feel 11.65 48 2.65 12.35 49 2.31 Friend 10.62 48 1.33 10.94 49 1.21 Gain 10.71 48 1.30 11.33 49 0.85 Enjoy 6.42 48 3.11 5.94 49 3.08 Bad 12.75 48 2.32 13.51 49 1.86 Self responsibility 12.36 47 2.45 13.52 48 1.71 Self blame 11.71 48 2.21 12.37 49 2.10 Victim responsibility 5.04 48 2.81 4.53 49 1.76 Victim blame 4.31 48 1.69 4.16 49 1.48









Differences Between Aqgressive and Nonaggressive Boys

Because there were age differences, data from the

sixth and third graders were separated when comparing the two aggressiveness groups; that is, the sixth-grade aggressive boys were only compared with the nonaggressive boys in the sixth grade, and the third-grade aggressive boys were compared only with nonaggressive boys in the same grade. This section presents results of the statistical analysis on each variable explored in this study, with results of the sixth graders described prior to those of the third graders.

Story Completion

Story completion scores were from open-ended

questions, such as "what would happen next," with a higher score reflecting higher motivation to compensate. Two sixth graders and five third graders did not reply to all three of the stories, and three boys in sixth grade and six in third grade did not have IQ scores; therefore, only 89 third-grade and 95 sixth-grade boys were included in the analysis of the story completion scores.

Because the IQ scores were significantly correlated

with story completion scores, an ANCOVA was performed on the story completion scores for both sixth and third graders, using IQ as the covariate. The results show (see table 6) that aggressive and nonaggressive boys in the sixth grade













Table 6

ANCOVA Test on Story Completion for Sixth- and Third-Grade Aqqressive and Nonaggqqressive Boys with 0IQ as Covariate


Sum of Mean
Source Squares df Square F p


6th grade

Aggression 12.396 1 12.396 4.753 .032 IQ 16.593 1 16.593 6.362 .013

3rd grade

Aggression 0.182 1 0.182 0.124 .725 IQ 9.216 1 9.216 6.277 .014









differed in their responses to the story completion (E =

4.753, R = .032). The nonaggressive boys were more likely to have higher motivation and would put more effort into compensating for their misbehavior.

For third graders, an ANCOVA was also conducted on story completion scores with IQ as the covariate. No difference in story completion was found between aggressive and nonaggressive boys (E = .124, R = .725). Reactive Feelings

Six interview questions were used to assess the six

feelings after wrongdoing examined in this study. Boys were asked how fearful (guilty, ashamed, sad, happy, and proud respectively) would they feel if they did what the protagonist did in each story. As can be seen in Table 7, the MANCOVA results show that the overall reactive feelings of the aggressive and nonaggressive boys in the sixth grade were significantly different, with IQ controlled (E = 2.661, R = .020). The univariate F-tests (see Table 8) and descriptive data (see Table 4) revealed that aggressive boys reported lower levels of sadness (E = 10.092, p = .002), shame (F = 7.590, p = .007), and guilt (E = 4.426, p = .038) than nonaggressive boys after wrongdoing. They did not differ in their feelings of fear, happiness, and pride.









Table 7

MANCOVA Test of Reactive Feelings for Aggressive and Nonaggressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade
Agg .845 2.661 6 87 .020 IQ .704 6.101 6 87 .000

3rd grade
Agg .929 1.101 6 86 .368 IQ .813 3.307 6 86 .006


Table 8


Univariate F-Tests of Reactive Feelings for Aggressive and Nonaggqqressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F R


6th grade
Fear 5.407 1 5.407 1.090 .299 Guilt 24.506 1 24.506 4.426 .038 Sad 69.326 1 69.326 10.092 .002 Happy 7.424 1 7.424 3.786 .055 Shame 43.547 1 43.547 7.590 .007 Proud .214 1 .214 .161 .689

3rd grade
Fear .498 1 .498 .082 .776 Guilt 8.685 1 8.685 1.606 .207 Sad 3.257 1 3.257 .303 .583 Happy 13.379 1 13.379 2.952 .089 Shame 13.876 1 13.876 .244 .268 Proud 11.535 1 11.535 3.382 .069







62

For the third graders, the MANCOVA did not indicate a significant overall difference (E = 1.101, R = .368) between aggressive and nonaggressive boys in their feelings following misbehavior. The univariate analyses did not show any differences in these feelings.

The means for the four negative feelings were much higher than means for the two positive feelings. This difference indicates that after misconduct boys in each group, no matter what their aggressiveness and grade, tended to feel much more fear, sadness, guilt, and shame than happiness and pride.

Emotion Reulation

This section contains information on boys' cognitive coping strategies, intention to help and apologize to the victim, intensity, and length of disturbance following misconduct.

Cognitive coping strategy. One interview question asked boys what would they think to help themselves feel better after the accident happened and asked them to select one of five coping strategies. Boys' answers were assigned categorical scores, and the contingency tables have cells with expected frequency of less than 5; hence Fisher's exact test was used to analyze data from this question.









As shown in Table 9, more boys in both aggressiveness groups reported that they would use the fourth strategy, thinking how to compensate for their wrongdoing, than selected other strategies. No boys in either group selected the first (comparing with others' detrimental conduct) or second strategies (devaluing the victim). Data analysis yielded significant differences between sixth-grade aggressive and nonaggressive boys on all three stories (x2= 6.939, p = .031; 12= 13.172, p = .001; 12= 9.444, p = .009, respectively). More nonaggressive boys were likely to think of how to compensate for their wrongdoing, whereas more aggressive boys were likely to diffuse responsibility (strategy three) or try not to think about the event (strategy five).

The results for third graders are presented in Table 10. Most of the boys in both groups would think of a way to compensate the victim; some would try not to think about the accident, and a few boys would try to justify their wrongdoing. No difference was found between the two groups in their choice of cognitive coping strategy for each of the three stories (X2= 2.907, R = .406; 12= 2.378, p = .498; X2 = 1.508, R = .825, respectively).

Intention to help and apologize to the victim. Boys were asked would they help or apologize to the victims in









Table 9


Frequency and Fisher's Exact Test for Cognitive Coping


Strategy for Aggqqressive and Grade


Strategy

1 2 3 4 5 x2

Story 1
Agg 0 0 4 34 10 6.939 .031
Nonagg 0 0 0 44 5

Story 2
Agg 0 0 6 31 9 13.172 .001
Nonagg 0 0 1 47 1

Story 3
Agg 0 0 8 31 8 9.44 .009
Nonagg 0 0 2 44 2



Table 10

Frequency and Fisher's Exact Test for Cognitive Coping Strategy for Aggqqressive and Nonagqqressive Boys in ThirdGrade

Strategy

1 2 3 4 5 x_2

Story 1
Agg 3 0 4 33 7 2.907 .406
Nonagg 2 0 1 35 11

Story 2
Agg 1 0 2 34 10 2.378 .498
Nonagg 0 0 4 38 7

Story 3
Agg 2 1 7 31 6 1.508 .825
Nonagg 1 0 7 34 7


usummy ft...v.f mvy 49% - w--









order to understand these boys' intention to repair their misconduct. Results of the MANOVA showed that sixth-grade aggressive boys were different from nonaggressive boys in their intention to make reparations to the victim (E =

4.125, R = .019). The univariate analyses revealed a difference in their intention to help the victim (E = 6.705, p = .011), but not in their intention to apologize to the victim (E = 3.575, p = .062).

For the third graders, no overall difference was found between aggressive and nonaggressive boys (E = 2.121, p = .126). However, a univariate analysis revealed a difference in their intention to help the victim (E = 3.972, p = .049).

Intensity of disturbance. Boys were asked if they

still felt discomfort after apologizing and after helping the victim and if these incidents would disturb their daily life. These three questions were used to assess the intensity of the boys' disturbance due to these incidents.

MANOVA results yielded a significant overall

difference between sixth-grade aggressive and nonaggressive boys (E =3.406, p = .021). The univariate E-tests and descriptive data revealed that nonaggressive boys were more likely to be disturbed in daily life (E = 4.760, R = .032) and to continue to feel discomfort after their apology (E =

4.574, p = .035), as displayed in Table 13 and 14.









Table 11

MANOVA of Intention to Compensate for Aagressive and Nonaqggressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df


6th grade .919 4.125 2 94 .019 3rd grade .957 2.121 2 94 .126


Table 12


Univariate F-Tests of Intention and Nonaqqressive Boys in Grade


to Compensate for Aqgressive Six and Three


Sum of Mean
Source Squares df Square F p


6th grade
Apology 11.914 1 11.914 3.575 .062 Help 19.306 1 19.306 6.705 .011

3rd grade
Apology 14.233 1 14.233 3.972 .049 Help 6.062 1 6.062 1.724 .192


Six and Three









Table 13

MANOVA of Intensity of Disturbance for Aggressive and Nonaqggressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade .899 3.406 3 91 .021 3rd grade .984 .472 3 89 .702


Table 14

Univariate F-Tests of Intensity of Disturbance for Agressive and Nonaqqressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F R


6th grade
After Apology 45.344 1 45.344 4.574 .035 After Help .954 1 .954 .111 .739 Disturb Life 39.983 1 39.983 4.760 .032

3rd grade
After Apology 1.609 1 1.609 .342 .560 After Help 1.911 1 1.911 .507 .478 Disturb Life 15.134 1 15.134 1.074 .303









The MANOVA did not show any difference in the thirdgrade boys' responses to the questions about the intensity of their disturbance (F = .472, R = .702). The univariate tests also did not show any differences.

Length of disturbance. Two interview questions were designed to assess the length of disturbance elicited by wrongdoing: If these incidents would disturb them for a couple of days, and how long they would keep thinking about these events. As presented in Table 15, no significant difference in length of disturbance was found between sixthgrade aggressive and nonaggressive boys (E = 1.653, p = .197). The univariate tests did not reveal any significant difference in the likelihood of being disturbed for two days or in the length of time disturbed (see Table 16).

Third-grade boys in the two aggressiveness groups did not differ in length of disturbance after their wrongdoing (F = .048, p = .953). No differences were found in the univariate analyses.

Empathic Reactions

Boys were asked how upset and how sad would they think the victim would be to determine boys' cognitive empathy. Another question, how bad would they feel if they saw the victim was hurt, was used to assess boys' emotional empathy. Data on empathic reactions from the three interview










Table 15
MANOVA of Length of Disturbance for Aggressive and Nonaqgressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade .966 1.653 2 94 .197 3rd grade .999 .048 2 93 .953


Table 16

Univariate F-Tests of Length of Disturbance for Aggressive and Nonaggressive Boys in Sixth- and Third-Grade


Sum of Mean
Source Squares df Square E R


6th grade
Days Disturbed 13.841 1 13.841 3.219 .076 Time Thinking 4.953 1 4.953 1.073 .303

3rd grade
Days Disturbed .087 1 .087 .006 .937 Time Thinking .716 1 .716 .087 .769









questions were analyzed by NANOVA. As illustrated in Table 17, a significant overall difference on the three questions was found for the sixth-grade aggressive and nonaggressive boys (F = 2.732, R = .048). The univariate analyses (see Table 18) showed that nonaggressive boys tended to feel a higher level of empathy when seeing the other boy upset due to their misconduct (F = 7.637, R = .007). There were no differences in these boys' estimate of their victim's upset and sadness.

For third graders, there was no overall difference

between the aggressive and nonaggressive groups (E = 1.110, R = .349). Univariate analyses also showed no differences between them in cognitive and emotional empathy. Outcome Expectations

Boys were asked two questions to assess their outcome expectations: What do you think your friend would think about your behavior and do you think what you did would help you get what you like. As shown in Table 19, results of the MANOVA indicate that outcome expectations were different for aggressive and nonaggressive sixth-grade boys (F = 3.277, R = .038). In univariate F-tests, it was found that aggressive boys were more likely to believe that aggression can help them get what they want (F = 6.751, p = .011); they did not









Table 17


MANOVA of Empathic


Nonaaaressive Boys


Reactions for Aggqqressive and in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade .918 2.732 3 92 .048

3rd grade .964 1.149 3 93 .334







Table 18

Univariate F-Tests of Empathic Reactions for Sixth and Third-Grade Boys


Sum of Mean
Source Squares df Square F p


6th grade
Victim Upset 4.873 1 4.873 1.037 .311 Victim Sad 15.065 1 15.065 3.127 .080 Empathy 32.754 1 32.754 7.637 .007

3rd grade
Victim Upset .060 1 .060 .011 .917 Victim Sad 5.284 1 5.284 .631 .429 Empathy 11.919 1 11.919 1.932 .168


MAOAo Rwti









Table 19

MANOVA of Outcome Expectations for Aqqressive and Nonaqressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade .932 3.377 2 93 .038 3rd grade .925 3.835 2 94 .025


Table 20

Univariate F-Tests of Outcome Expectations for Aqqressive and Nonaqqressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F p

6th grade
Friend 0.064 1 0.064 .077 .782 Object 12.284 1 12.284 6.751 .011

3rd grade
Friend 2.387 1 2.387 1.472 .228 Object 9.267 1 9.267 7.676 .007









differ in their belief that their friends would like what they did in the story (F = .077, R = .782).

For third-grade boys, the pattern was the same as for the sixth graders. There was an overall difference between aggressive and nonaggressive boys' outcome expectations for their misbehavior (F = 4.726, p = .011). The aggressive boys tended to believe that aggression could help them obtain what they want (F = 11.193, R = .003). The boys did not differ in their belief that their friend would like their wrongdoing (F = 1.472, p = .228). Legitimacy Beliefs

Data were gathered on boys' legitimacy beliefs from their answers to two questions: How happy would you feel with your win (toy and the movie) and how bad do you think it was to do this? As presented in Table 21, there was no overall significant difference between sixth-grade aggressive and nonaggressive boys in their beliefs about the legitimacy of aggression (F = 1.600, p = .207). Univariate analyses also showed no differences in levels of happiness elicited by what boys obtained through misbehavior and no difference in their evaluation of how bad their behavior would be (see Table 22).

There was no overall difference in beliefs about the legitimacy of aggression between third-grade aggressive and









Table 21

MANOVA of Beliefs about the Legitimacy of Aggression for Aqqressive and Nonaggressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df


6th grade .967 1.600 2 97 .207 3rd grade .968 1.571 2 97 .213


Table 22

Univariate F-Tests of Beliefs about the Legitimacy of Aqgression for Aqqressive and Nonagqqressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F p

6th grade
Enjoy 16.125 1 16.125 3.028 .085 Bad 3.204 1 3.204 1.240 .268

3rd grade
Enjoy 5.538 1 5.538 .577 .450 Bad 14.013 1 14.013 3.175 .078









nonaggressive boys (E = 1.571, p = .213) indicated in the MANOVA test. Univariate analyses revealed no differences in the two groups of boys in their enjoyment of their gains through misbehavior or in the evaluation of their own behavior.

Responsibility Attribution

Self-responsibility. Boys were asked how responsible

they would feel and how much they would be to blame for what happened to assess their feelings of responsibility. Results of the MANOVA show that there was no overall difference between sixth-grade aggressive and nonaggressive boys' selfresponsibility attribution (E = .604, p = .549). Also, there were no differences in their thoughts about how responsible they were or how much the victim should be blamed. These results are displayed in Table 23 and 24.

For the third graders, an overall difference was found between aggressive and nonaggressive boys in selfresponsibility attribution (E = 3.544, p = .009). Univariate analyses show that the aggressive boys felt less responsible for the incident than nonaggressive boys (E = 7.153, p = .009). No difference was found in their thoughts about how much the victim should be blamed (E = 2.037, p = 157).

Victim responsibility. Data were collected from two questions to assess boys' beliefs about the victim's









Table 23

MANOVA of Attribution of Self-Responsibility for Agqressive and Nonaggressive Boys in Grade Six and Three


Wilks' Hypothesis Error
Source A F df df R


6th grade .987 .604 2 94 .549 3rd grade .928 3.544 2 92 .033


Table 24


Univariate F-Tests of Attribution of Self-Responsibility for Aqgressive and Nonaqqressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F p


6th grade
Self-responsibility .327 1 .327 .134 .716 Self blame 3.325 1 3.325 1.031 .313

3rd grade
Self-responsibility 31.907 1 31.907 7.153 .009 Self blame 9.475 1 9.475 2.037 .157







77

responsibility: How responsible was the victim and how much was the victim to blame for what happened? As displayed in Table 25, sixth-grade aggressive and nonaggressive boys were significantly different in their beliefs about the victims' responsibility (F = 3.436, p = .036). Univariate analyses revealed that sixth-grade aggressive boys were likely to attribute more responsibility to the victim than nonaggressive boys (g = 5.220, p = .025). Aggressive boys also tended to think the victim should be blamed more than the nonaggressive boys did (F = 6.639, p = .012)(see Table 26).

There was no significant difference between thirdgrade aggressive and nonaggressive boys in attribution of victims' responsibility (F = .654, R = .522). Further analysis showed no differences in how responsible they felt the victim was (F = 1.161, R = .284) or how much the victim should be blamed (F = .215, p = .644).

Differences Between Sixth and Third Graders Story Completion

To examine if there was a difference between sixthand third-grade boys in their responses to the story

completion, an ANOVA was performed. No significant difference was found between sixth- and third-grade










Table 25


MANOVA of Attributing Responsibility to Victim Aggqqressive and Nonaqqressive Boys in Grade Six


for and Three


Wilks' Hypothesis Error
Source A E df df R


6th grade .932 3.436 2 94 .036 3rd grade .986 .654 2 94 .522


Table 26

Univariate F-Tests of Attributing Responsibility to Victim for Aggressive and Nonaqgressive Boys in Grade Six and Three


Sum of Mean
Source Squares df Square F R


6th grade
Victim Responsibility 23.491 1 23.491 5.220 .025 Victim Blame 21.817 1 21.817 6.639 .012

3rd grade
Victim Responsibility 6.333 1 6.333 1.161 .284 Victim Blame .540 1 .540 .215 .644







79

aggressive boys in their story completion scores (E = 2.321, R = .131), as shown in Table 27.

A second ANOVA was conducted on the data from the nonaggressive boys. The analysis yielded a significant difference between the sixth- and third-grade nonaggressive boys (E = 11.850, R = .001). The sixth graders were more likely to report that they would be more motivated and would put more effort into compensating for their misbehavior than did the third graders.

Reactive Feelings

To determine whether the reactive feelings the

aggressive boys in the sixth grade reported they would feel after misbehavior were significantly different from those reported by the third-grade aggressive boys, a MANOVA procedure was conducted (see Table 28). An overall significant difference was found (F = 2.861, p = .014). Results of univariate analyses showed that the aggressive boys in the third grade tended to report that they would feel higher levels of happiness (F = 4.751, p = .032) and pride (F = 8.658, p = .004) than the aggressive boys in the sixth grade. They did not differ in feelings of fear, guilt, shame, and sadness. The results are summarized in Table 29.

For the nonaggressive boys, the results of the MANOVA also revealed a significant overall difference between the










Table 27

ANOVA Results for Testing Age Differences in Story Completion


Sum of Mean
Source Squares df Square F R


Aggressive 3.866 1 3.866 2.321 .131 Nonaggressive 31.154 1 31.154 11.850 .001









Table 28

MANOVA Results for Testing Age Differences in Reactive Feelings


Wilks' Hypothesis Error
Source A F df df R


Aggressive .838 2.861 6 89 .014 Nonaggressive .868 2.304 6 91 .041


Table 29

Results of Univariate F-Tests for Age Differences in Reactive Feelings


Sum of Mean
Source Squares df Square F p


Aggressive
Fear 3.010 1 3.010 .485 .488 Guilt .375 1 .375 .054 .816 Sad 24.000 1 24.000 2.392 .125 Happy 23.010 1 23.010 4.751 .032 Shame 29.260 1 29.260 3.115 .081 Proud 28.167 1 28.167 8.658 .004

Nonaggressive
Fear 12.500 1 12.500 2.474 .119 Guilt 4.939 1 4.939 1.115 .294 Sad 0.010 1 0.010 .001 .971 Happy 6.378 1 6.378 3.432 .067 Shame 62.082 1 62.082 7.768 .006 Proud 4.082 1 4.082 2.096 .151









sixth- and third-grade boys (F = 2.304, R = .041). The univariate F-tests yielded a significant difference in the feeling of shame (F = 7.768, R = .006), but not in feelings of fear, guilt, sadness, happiness, or pride. The group means indicate that the older nonaggressive boys tended to report that they would feel more shame than did the younger nonaggressive boys after their own misbehavior.

Emotion Reulation

Cognitive coping strategy. Fisher's exact test was used to examine the differences between the sixth- and third-grade boys' responses to five categories of cognitive coping strategies (comparing to others' misbehavior, dehumanizing others, blaming others, thinking about compensation, and shifting attention) following transgression. Table 30 and 31 display the frequency of each strategy the boys in each group reported that they would use after wrongdoing. The results of Fisher's exact test are presented in Table 30 and 31. No significant difference was found between the sixth- and third-grade aggressive boys for the three stories.

Fisher's exact test revealed an age difference for the nonaggressive boys. As can be seen in Table 31, for the second (X2 = 7.253, p = .027) and third (12= 7.828, p = .050) stories, more sixth graders reported that they would









Table 30

Frequency and Fisher's Exact Test for Age Differences in Cognitive Coping Strategy for Aggressive Boys

Strategy

1 2 3 4 5 _2

Story 1
6th grade 0 0 4 34 10 3.534 .316
3rd grade 3 0 4 33 7

Story 2
6th grade 0 0 6 31 9 3.181 .365
3rd grade 1 0 2 34 10

Story 3
6th grade 0 0 8 31 8 3.352 .501
3rd grade 2 1 7 31 6


Table 31

Frequency and Fisher's Exact Test for Age Differences in Cognitive Coping Strategy for Nonaggressive Boys


Strategy

1 2 3 4 5 x2 R


Story 1
6th grade 3rd grade

Story 2
6th grade 3rd grade

Story 3
6th grade 3rd grade


6.275 7.253 7.828


.099



.027 .050









employ a strategy focusing on compensation than did the third graders. More third graders reported that they would apply strategies of justification or would shift their attention from the incidents.

Intention to help and apologize to the victim. The MANOVA procedure did not yield a significant difference between the sixth- and third-grade aggressive boys in their intention to help or apologize to the victim (E = .658, R = .520), as shown in Table 32. Also, the univariate tests did not indicate any significant differences between the two age groups (see Table 33).

For the nonaggressive boys, MANOVA results showed no overall difference between the sixth and third graders in their intention to apologize or help (E = .469, p = .627). The univariate analyses also showed no age differences among the nonaggressive boys.

The means for intention to help and apologize for both group approached the highest possible score, 15. This information indicated that most boys in each group reported that they would help and apologize to the victims if they had hurt the victims.

Intensity of disturbance. The results of the MANOVA indicated a significant difference between the sixth- and third-grade aggressive boys in the overall intensity of










Table 32


MANOVA Results Apologize


of Age Differences in Intention to Helo and


Wilks' Hypothesis Error
Source A F df df


Aggressive .986 .658 2 92 .520 Nonaggressive .990 .469 2 95 .627


Table 33

Results of Univariate F-Tests for Age Differences in Intention to Help and Apologize


Sum of Mean
Source Squares df Square F p


Aggressive
Apology 1.195 1 1.195 .247 .621 Help 5.442 1 5.442 1.315 .254

Nonaggressive
Apology 2.000 1 2.000 .944 .334 Help .163 1 .163 .071 .790


of . A ae . . ........... ... .... ....... .. ... l and









their disturbance (E = 5.579, R = .002), as displayed in Table 34. Univariate tests indicated that the two age groups differed in the likelihood of continuing to feel discomfort after helping (fE = 14.171, R = .000) and apologizing (E = 5.964, p = .017) to the victim (see Table 35). The group means revealed that the sixth-grade aggressive boys reported that they were more likely to remain distressed after apologizing and helping than did the third-grade aggressive boys.

Significant age differences in the intensity of

disturbance were also presented among the nonaggressive boys (F = 8.235, p = .000). The results of the univariate analyses and group means revealed that the older boys reported that they were more likely to continue feeling discomfort after apologizing (E = 16.436, p = .000) and helping (F = 14.959, p = .000) and also more likely to be disturbed in their daily life (E = 9.037, p = .003).

Younger boys' responses to these questions were

similar regardless of their aggressive status. A majority of younger boys responded that they would not feel discomfort after they had apologized to or helped the victim.

Length of disturbance. A MANOVA was conducted to

examine age differences in the length of disturbance due to misconduct. As can be seen in Table 36, a significant










Table 34

MANOVA Results of Aqe Differences in Intensity of Disturbance


Wilks' Hypothesis Error
Source A F df df R


Aggressive .837 5.579 3 86 .002 Nonaggressive .792 8.235 3 94 .000


Table 35

Results of Univariate F-Tests for Age Differences in Intensity of Disturbance


Sum of Mean
Source Squares df Square F p


aggressive
After apology 36.202 1 36.202 5.964 .017 After help 90.401 1 90.401 14.171 .000 Disturb life 47.193 1 47.193 3.828 .054

nonaggressive
After apology 139.684 1 139.684 16.436 .000 After help 90.163 1 90.163 14.959 .000 Disturb life 92.092 1 92.092 9.037 .003










Table 36

MANOVA Results of Age Differences in Length of Disturbance


Wilks' Hypothesis Error
Source A F df df R


Aggressive .908 4.659 2 92 .012 Nonaggressive .815 10.813 2 95 .000






Table 37

Results of Univariate F-Tests for Age Differences in Length of Disturbance


Sum of Mean
Source Squares df Square F p


Aggressive
Days Disturbed 91.032 1 91.032 8.893 .004 Time Think 28.122 1 28.122 4.299 .041 Nonaggressive
Days Disturbed 172.449 1 172.449 21.805 .000 Time Think 45.806 1 45.806 7.242 .008









difference was found between the sixth- and third-grade aggressive boys (E = 4.659, p = .012). Results of univariate analyses and descriptive data indicated that the sixth graders reported that they were more likely to be bothered for two or more days (F = 8.893, p = .004) and to keep thinking about these incidents for a longer period of time (E = 4.299, p = .041) than did the aggressive boys in the third grade (see Table 37).

For nonaggressive boys, overall age differences were found (E = 10.813, p = .000) through MANOVA procedure. Univariate analyses and the group means revealed that compared to the third graders, the sixth graders were more likely to report that they would be disturbed by the events for two or more days (F = 21.805, p = .000) and to continue to think about these incidents (E = 7.242, p = .008). Empathic Reactions

A MANOVA procedure was conducted to examine age

differences in boys' empathic reactions to the victims of their wrongdoing (see Table 38). No significant overall difference was found between the sixth- and third-grade aggressive boys (F = .820, R = .486). The univariate E-tests yielded no difference in these two groups' estimate of their victims' upset and sadness or in their empathic emotions (see Table 39).










Table 38

MANOVA Results of Age Differences in Empathic Reactions


Wilks' Hypothesis Error
Source A E df df R


Aggressive .974 .820 3 91 .486 Nonaggressive .958 1.359 3 94 .260







Table 39

Results of Univariate F-Tests for Age Differences in Empathic Reactions


Sum of Mean
Source Squares df Square E R


Aggressive
Victim Upset 5.473 1 5.473 1.157 .285 Victim Sad .037 1 .037 .005 .943 Empathy 1.218 1 1.218 .202 .654 Nonaggressive
Victim Upset .010 1 .010 .002 .966 Victim Sad 2.000 1 2.000 .342 .560 Empathy 11.796 1 11.796 2.640 .108




Full Text

PAGE 1

EMOTIONAL AND COGNITIVE REACTIONS TO HYPOTHETICAL MORAL TRANSGRESSIONS IN AGGRESSIVE AND NONAGGRESSIVE BOYS IN TAIWAN By SHU-TAI LIAO 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 1998

PAGE 2

ACKNOWLEDGMENTS The con^letion of this study was made possible by the contribution of a number of people. I wish to express my sincere gratitude to Dr. Patricia Ashton, chairperson of my doctoral committee, for her continuous guidance, encouragement, axid careful scrutiny of my work throughout the years of my doctoral study and this project. Without her assistance and expertise, I would never have completed this study. I also wish to thank Dr. Shari Ellis, Dr. Bridget Franks, and Dr. Jin-Wen Hsu, for their valuable counsel and support . My appreciation is also extended to my feuaily, my parents, my husband, and two sons, for their encouragement and sacrifices. ii

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDOffiNTS ii LIST OF TABLES V ABSTRACT ix CHAPTERS 1 INTRODUCTION 1 Statement of the Problem 1 Purpose of the Study 4 Significance of the Study 6 Review of Literature 8 S\immary 32 2 DESIGN OF THE STUDY 33 Research Participants 33 Measures 34 Procedures 38 Data Analysis 39 Limitations of the Study 40 Pilot Study 41 Revising Interview Questions and Stories 52 3 RESULTS 54 Descriptive Data 55 Differences Between Aggressive and Nonaggressive Boys 59 Differences Between the Sixth and Third Graders.. 77 4 DISCUSSION AND RECOMMENDATIONS 99 Overview of the Findings 99 Discussion of Results 100 Recommendations for Further Research 117 Recommendations for Practice 119 iii

PAGE 4

APPENDICES A TEACHER CHECKLIST 121 B REVISED STORIES 122 C INTERVIEW QUESTIONS WITH REVISIONS 124 D CONSENT FORMS 12" E INTERCORRELATION MATRIX 131 REFERENCES 134 I BIOGRAPHICAL SKETCH 147 iv

PAGE 5

LIST OF TABLES Table Page 1 Differences between Aggressive and Nonaggressive Children in Reactions to Moral Transgression 43 2 Differences between Third and Sixth Graders in Reactions to Moral Transgression 46 3 Correlations Between Scores on Interview Questions 49 4 Means and Standard Deviations on Each Variable for Sixth Graders 56 5 Means and Standard Deviations on Each Variable for Third Graders 57 6 ANCOVA Test on Story Completion for Sixthand Third-Grade Aggressive and Nonaggressive Boys with IQ as Covariate 59 7 MANCOVA Test of Reactive Feelings for Aggressive and Nonaggressive Boys in Grade Six and Three .... 61 8 Univariate F-Tests of Reactive Feelings for Aggressive and Nonaggressive Boys in Grade Six and Three gl 9 Frequency and Fisher' s Exact Test for Cognitive Coping Strategy for Aggressive and Nonaggressive Boys in Sixth Grade 64 10 Frequency and Fisher's Exact Test for Cognitive Coping Strategy for Aggressive and Nonaggressive Boys in Third Grade 64 11 MANOVA of Intention to Compensate for Aggressive and Nonaggressive Boys in Grade Six and Three .... 66 V

PAGE 6

12 Univariate F-Tests of Intention to Compensate for Aggressive and Nonaggressive Boys in Grade Six and Three 66 13 MANOVA of Intensity of Disturbance for Aggressive and Nonaggressive Boys in Grade Six and Three ... 67 14 Univariate F-Tests of Intensity of Disturbance for Aggressive and Nonaggressive Boys in Grade Six and Three 67 15 MANOVA of Length of Disturbance for Aggressive and Nonaggressive Boys in Grade Six and Three 69 16 Univariate F-Tests of Length of Disturbance for Aggressive and Nonaggressive Boys in Sixthand Third-Grade 69 17 MANOVA of Empathic Reactions for Aggressive and Nonaggressive Boys in Grade Six and Three 71 18 Univariate F-Tests of Empathic Reactions for Sixth and Third-Grade Boys 71 19 MANOVA of Outcome Expectations for Aggressive and Nonaggressive Boys in Grade Six and Three 72 20 Univariate F-Tests of Outcome Expectations for Aggressive and Nonaggressive Boys in Grade Six and Three 72 21 MANOVA of Beliefs about the Legitimacy of Aggression for Aggressive and Nonaggressive Boys in Grade Six and Three 74 22 Univariate F-Tests of Beliefs about the Legitimacy of Aggression for Aggressive and Nonaggressive Boys in Grade Six and Three 7/1 23 MANOVA of Attribution of Self -Responsibility for Aggressive and Nonaggressive Boys in Grade Six and Three 7g 24 Univariate F-Tests of Attribution of SelfResponsibility for Aggressive and Nonaggressive Boys in Grade Six and Three 76 vi

PAGE 7

25 MANOVA of Attributing Responsibility to Victim for Aggressive and Nonaggressive Boys in Grade Six and Three 78 26 Univariate F-Tests of Attributing Responsibility to Victim for Aggressive and Nonaggressive Boys in Grade Six and Three 78 27 ANOVA Results for Testing Age Differences in Story Con^letion 80 28 MANOVA Results for Testing Age Differences in Reactive Feelings 81 29 Results of Univariate F-Tests for Age Differences in Reactive Feelings 81 30 Frequency and Fisher' s Exact Test for Age Differences in Cognitive Coping Strategy for Aggressive Boys 83 31 Frequency and Fisher' s Exact Test for Age Differences in Cognitive Coping Strategy for Nonaggressive Boys 83 32 MANOVA Results of Age Differences in Intention to Help and J^ologize 85 33 Results of Univariate F-Tests for Age Differences in Intention to Help and Apologize 85 34 MANOVA Results of Age Differences in Intensity of Disturbance 87 35 Results of Univariate F-Tests for Age Differences in Intensity of Disturbance 87 36 MANOVA Results of Age Differences in Length of Disturbance gg 37 Results of Univariate F-Tests for Age Differences in Length of Disturbance gg 38 MANOVA Results of Age Differences in Empathic Reactions 39 Results of Univariate F-Tests for Age Differences in Empathic Reactions 90 vii

PAGE 8

40 MANOVA Results of Age Differences in Outcome Expectations for Aggression 92 41 Results of Univariate F-Tests for Age Differences in Outcome Expectations for Aggression 92 42 MANOVA Results of Age Differences in Beliefs about the Legitimacy of Aggression 94 43 Results of Univariate F-Tests for Age Differences in Beliefs about the Legitimacy of Aggression ... 94 44 MZ^OVA Results of Age Differences in Attribution of Self-Responsibility 95 45 Results of Univariate F-Tests for Age Differences in Attribution of Self-Responsibility 95 46 MANOVA Results of Age Differences in Attributing Responsibility to Victim 97 47 Results of Univariate F-Tests for Age Differences in Attributing Responsibility to Victim 97 I viii

PAGE 9

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 EMOTIONAL AND COGNITIVE REACTIONS TO HYPOTHETICAL MORAL TRANSGRESSIONS IN AGGRESSIVE AND NONAGGRESSIVE BOYS IN TAIWAN By Shu-Tai Liao December 1998 Chairperson: Patricia Ashton Major Department: Foundations of Education The purpose of this study was to identify differences between aggressive and nonaggressive boys and between sixthamd third-grade boys in their emotional and cognitive reactions to hypothetical moral transgressions. One hundred aggressive and one hundred nonaggressive boys in the third and sixth grades in Taiwan selected on the basis of teacher ratings were interviewed individually. The boys read three semiprojective stories describing explicit misbehavior and then answered questions concerning how they would feel and think if they were the perpetrators of the misbehavior. Results indicated that boys in all groups, regardless of age and aggressiveness, reported that they would feel high levels of negative emotions and low levels of positive emotions and they believed that aggression is inappropriate. However, they differed in the strength of their feelings. ix

PAGE 10

outcome expectations , and responsibility attributions . For the sixth graders , nonaggressive boys reported that they would feel higher levels of sadness, guilt, sheune, and empathy, have a stronger intention to compensate, would be more intensely disturbed, expect a less favorsdale outcome, and attribute less responsibility to victims, when compared to aggressive boys . Among the third graders , nonaggressive boys reported that they would expect less favorable consequences for transgressions and attribute more responsibility to self, relative to aggressive peers. Age differences were also found. Younger aggressive boys reported that they would feel higher levels of happiness and pride than older boys . The older aggressive boys reported that they would expect more disapproval from friends for the aggression than did younger aggressive boys . The older nonaggressive boys said that they would feel higher levels of shame and a stronger intention to remedy their misbehavior than did the younger boys. Both aggressive and nonaggressive older boys said they would be more intensely disturbed longer and attribute more responsibility to self and victims , compared to the younger boys . Results are discussed in terms of their relationship to previous studies. Finally, some suggestions for further study and practice are proposed. X

PAGE 11

CHAPTER 1 INTRODUCTION Statement of the Problem Childhood aggression has been a problem in all societies (Crowell, 1987) . Epidemiological studies reported by the Institute of Medicine in 1989 showed that in the United States between 2% and 6% of youths commit clinically severe antisocial behavior; that is about 1.3 to 3.8 million youths (Kazdin, 1994) . The increasing rate of crime committed by children and youths is a serious problem in Taiwan as well. According to official report (Justice Department of the Republic of China, 1996) , three trends in youth crime can be identified. First, juvenile crime is increasing rapidly. From 1986 to 1995, the juvenile crime events referred to district courts increased by 80% . In 1986, 18% of total crimes were committed by youths; in 1995, the percentage had increased to 30% . The second trend is the age distribution of crime is downward. From 1991 to 1995, the percent of juvenile crimes committed by adolescents aged 12 to 13 increased from 6.7% to 9.1%. For youths aged 13 to 14 the percentage increased from 12.2% to 15.7%. The third trend is that the n\imber of violent crimes committed by 1

PAGE 12

youths increased dramatically. During a 1-year period, from 1994 to 1995, injury cases climbed 19.7%, threatening increased 24.6%, and robbery grew by 69.4%. There is strong evidence that aggression in individuals tends to be relatively stable over time and situation. A number of studies have shown that early aggressiveness or antisocial conduct is predictive of later serious antisocial behavior or criminality in adolescence, adulthood, and even in the next generation for certain groups of children (Ghodsian, Fogelman, Lambert, & Tibbenham, 1980; Huesmann, Eron, Lefkowitz, & Walder, 1984; Loeber & Dishion, 1983; Pulkkinen & Pitkanen, 1993) . From a 22 -year longitudinal study, Eron, Huesmann, Dubow, Romanoff, and Yarmel (1986) concluded that By the time a child is 8 years old, characteristic ways of behaving aggressively or nonaggressively have already been established. Aggression as a problemsolving behavior is learned very early in life, and it is learned very well; the payoff is tremendous, (p. 261) On the basis of experimental studies, Geen (1990) concluded that "the original act of aggression appears to facilitate subsequent aggression instead of diminishing it" (p. 189) . Aronson, Wilson, and Akert (1994) also proposed that people tend to justify their aggression in ways that eventually produce more aggressive behavior.

PAGE 13

3 Obviously, the consistency and repetition of aggression make it a serious social problem. If we want to alleviate the problem, we need to reduce its recurrence in addition to preventing its occurrence. Thus, knowledge of why some children persist in aggressive behavior, why aggression may produce more aggression, and how to interrupt the continuation of aggression is crucial for alleviating this problem and should not be neglected. Although researchers and theorists have long been interested in the processes involved in the development of aggression, most attention has been directed to the causes of aggressive behavior. For exan^le, frustration, physiological factors, observational learning, failure in anger control, and hostile intention attribution have been emphasized as the origins of aggression (Geen, 1990) . Fewer studies have been conducted on the maintenance of aggression cootpared to those on the causes of aggression. Among the few, some researchers (Ferguson, Stegge, & Damhuis, 1990; Regan, 1972/ Tangney, 1991, 1992) have argued that affective reactions to moral transgression influence the recurrence of the wrongdoing, and others (Crick & Dodge, 1994; Dodge, 1993; Perry, Perry, & Rasmussen, 1986; Slaby & Guerra, 1989) have stressed cognitive reactions as contributors to the repetition of aggressive behavior. From these contrasting

PAGE 14

4 perspectives several questions emerge: What do aggressive children feel and think after wrongdoing? Do aggressive and nonaggressive children differ in their reactions to transgressions? We need more research to learn how children react to their tramsgressions and how those reactions influence their subsequent behavior. Purpose of the Study Cognitive-developmental researchers have focused on the role of social-cognitive variables in the development of aggression in children. They have identified outcome expectancies (Perry et al . , 1986), beliefs about the legitimacy of aggression (Slaby & Guerra, 1988) , and attributions of responsibility (McGraw, 1987) as important factors that affect the development of aggression. In contrast, some researchers with a social -emotional emphasis have proposed that empathy plays an important role in inhibiting aggression (Miller & Eisenberg, 1988) , and others have suggested that guilt inhibits aggression (Tangney, 1991) . These cognitive and emotional consequences of aggression have not been combined in a single study. To provide insight into the relationships among cognitive and emotional factors in the development of aggression, this study will examine the emotions and cognitions that

PAGE 15

aggression elicits . The following questions will be explored : 1 . Do aggressive and nonaggressive children feel different emotions after moral transgression? 2 . Do aggressive and nonaggressive boys differ in the intensity of their disturbance after transgression? 3 . Does the transgression disturb nonaggressive children longer than aggressive children? 4 . Do aggressive and nonaggressive boys differ in their empathic reactions to the victim of aggression? 5. Are aggressive children more likely to attribute responsibility to others while nonaggressive children attribute responsibility to themselves? 6. Are aggressive children more likely to hold a positive outcome es^ectancy for transgression than nonaggressive children? 7 . Are aggressive children more likely to believe that aggression is a legitimate response than nonaggressive children? 8 . Are there age differences in aggressive and nonaggressive children' s emotional and cognitive reactions to transgression?

PAGE 16

6 Significance of the Study Theoretical Implications Aggressive and nonaggressive children may differ in their emotional and cognitive responses to transgression. Identifying what aggressive children feel and think after aggression may help us explain why some children become aggressive and some do not and why for some people aggressive behavior is stable across the life span (Eron et al. , 1986) . Guilt has been conceived as an important moral emotion, but until recently researchers have not focused their attention on the development of guilt in children. From the extsmt literature, we still know little about children's reactions to their guilt feelings, how children cope with the emotion of guilt, or the effects of this emotion on aggressive children's later behaviors. This study may help answer these questions. Learning theorists have emphasized that the * consequences of behavior are the key to explaining the recurrence of a behavior (Bandura, Ross, & Ross, 1963/ Schunk, 1991) . However, they did not attend to how children interpret the consequences of their behavior; that is, what the consequences mean from children's perspectives. Our theoretical understanding of the impact of behavior would be

PAGE 17

7 improved if researchers included in their analysis of the consequences of behavior the thoughts and feelings that children's behavior elicits. This study focuses on children's reactions. The results should reveal more about what misbehavior means for children than studies that have not included children's feelings and thinking. Practical Implications Aggressive behavior has become a prevalent and serious social problem. Identification of the factors that influence aggression and the design of effective intervention programs are goals shared by scholars, educators, and policy makers in many countries. To eliminate aggressive behavior, it is critical to have prevention and intervention progreuns . Because children' s emotions and thoughts following transgression could influence the recurrence of the misbehavior, understanding what children feel and think in such situations is necessary for the design of effective interventions . If the results of this study show that aggressive children do not experience guilt or other negative emotions after their misbehavior but nonaggressive children do, it would be important to increase aggressive children' s experience of those emotions after aggression. If aggressive and nonaggressive children differ in their attributions of

PAGE 18

8 responsibility for misbehavior, then an intervention program to modify the attributions of aggressive children should be developed. If results reveal that aggressive children believe that wrongdoing can help them reach their goals, intervention strategies should include modifying children' s beliefs edjout the effectiveness of aggression. Review of Literature Emotional Reactions to Transgression Only a few studies have been conducted on children' s emotional reactions following transgression. However, the recent research on guilt, shame, and empathy, which have been conceived as important moral affects (Tangney, 1991; Zahn-Waxler, Kochanska, Krupnick, & McKnew, 1990; ZahnWaxler & Robinson, 1995), provides information on children's emotional responses to their wrongdoing. Guilt . Guilt refers to thoughts and feelings of remorse and responsibility for causing negative outcomes that violate social or moral norms . According to Tangney (1990, 1991), guilt is a negative self-evaluative emotion. The focus of the evaluation is a "bad" behavior committed by oneself. The feelings of remorse and responsibility typically motivate this person to remedy his/her transgression. Zahn-Waxler et al . (1990) en^hasized the importance of guilt in moral behavior: "As the main affect

PAGE 19

9 in conscience, it checks aggressive impulses and encourages people to undo harms, hence restoring social harmony" (p. 51) . Williams and Bybee (1994) asked 240 children in three grades about the events that made them feel guilty. The researchers found that 90.6% of 5th graders, 93.3% of 8th graders, and 92.3% of 11th graders reported feeling guilty after transgression. Ferguson et al. (1990) found that 96% of the second and fifth graders in their study reported experiencing guilt after a moral transgression. Perry, Perry, Bussey, English, and Arnold (1980) indicated that the more serious that children thought their misbehavior was, the more serious self -punishment they adopted. These research findings reveal that, following misbehavior, children typically experience guilt or negative emotions . Research on adult transgression has shown that guilt is typically experienced after one behaves aggressively (Okel & Mosher, 1968; Tangney, 1992), that there is an inverse relationship between guilt and hostility and aggression (Mosher, Mortimer, & Grebel, 1968; Schill & Schneider, 1970) , that guilt leads to altruistic acts in persons who cause harm (Regan, 1971) , and that after a person causes harm to others, he/she is more willing to do something altruistic. These findings are consistent with the

PAGE 20

10 belief that guilt motivates amending action (Ferguson et al., 1990; Regan, 1971; Tangney, 1991) and inhibits prohibited behavior (Mosher, 1965) . In light of this research suggesting that guilt inhibits aggression, it is reasonable to propose that aggressive children do not inhibit their aggression by guilt as other children do. There are several possible reasons why aggressive children may not be influenced by guilt. First, these children might not feel negative emotions or only feel a low level of guilt that does not affect their behavior. Or, children may experience complicated emotions instead of guilt. ZeQin-Waxler and Robinson (1995) suggested that once a child has hurt another person, many feelings may be evoked. In aggressive situations the child may feel empathy but also have hostile feelings, enjoy hurting the other, and be more affectively aroused. Another possibility is that aggressive children feel more shame but not guilt after a transgression than nonaggressive children. Researchers (Tangney, Wagner, Burggraf, Gramzow, & Fletcher, 1991) have found that shame is positively related to aggression. Finally, aggressive children may get rid of uncomfortable feelings fast by justifying their harmful act or focusing their attention on something else .

PAGE 21

11 Shame . Shsune is a global painful or h\iiailiating feeling resulting from a negative evaluation of oneself for failure to live up to standards or goals in the eyes of others (Lazarus, 1991; Lewis, 1992) . Experiences of shame can be aroused by transgression of moral standards or social conventions or by personal failure (Ferguson et al., 1990; Tangney, 1992) . In shame, the focus of self -evaluation is the entire self — the entire self is painfully scrutinized and negatively evaluated. In this situation, shame experience is usually accompanied by a sense of exposure, a sense of being small, and a sense of worthlessness and powerlessness . Shame feeling, thus, drives a person to hide self from others (Tangney, 1990, 1991; Tangney et al . 1992). Researchers have addressed the issue of shame and its behavioral correlates. Lewis (1992) and Scheff and Retzinger (1991) argued that feelings of shame can elicit humiliated fury and then transform into hostility, anger, or blind rage and consequently into aggressive behavior; however, shame also can cause withdrawal from situations or persons > . " associated with shameful feelings. Scheff (1987) also suggested that rage may become a new source of shame and then produce a shame -rage -shame -rage spiral. Empirical studies have revealed that shame-prone individuals are more likely to experience anger or hostility than their less

PAGE 22

shame-prone peers and to be more aggressive (Tangney, Wagner, Barlow, Marschall, & Gramzow, 1996; Tangpney, Wagner, Fletcher, & Gramzow, 1992) . For example, in a sample of 363 fifth-grade children, Tangney et al . (1991) found that for the fifth-grade males, proneness to shame was positively related to teacher reports of aggression. Accordingly, shame can be aroused after transgression, and it can lead to aggression for some children. It is reasonable to suspect that shame is related to the repetition of aggression. Empathy . Empathy refers to a vicarious affective response caused by others' emotional, physical, or psychological states. Researchers have suggested that empathy is the basis of moral motivation. Empirical evidence has shown that empathy is positively associated with prosocial behavior (Batson et al . , 1988) and negatively related to aggressive behavior (Ellis, 1982; Richardson, Hammock, Smith, Gardner, & Signo, 1994) . In their review of literature on the relation between empathy and aggression. Miller and Eisenberg (1988) found a low to modest negative relation between empathic responsiveness and aggressive behavior. However, the negative relation was significant when empathy was assessed by questionnaire and by picture/story method if preschool

PAGE 23

children were excluded but not significant when other empathy measures were used. That is, this relation could be influenced by age, methods of assessing empathy, and criteria for aggression. Feshbach (1983) claimed that empathy is an inhibitor of aggression because empathy fosters prosocial behaviors that are incompatible with aggression and because the pain of the victim of aggression may elicit distress in the bystander or even in an aggressor with empathic capability. An intervention program for empathy training has been found effective in reducing the occurrence of aggression (Feshbach, 1983) . Empathy also has been described as substantially related to the development and regulation of guilt and shame. According to Hoffman (1982) , empathy occurs in response to the distress of others, whereas guilt results from the combination of empathic response and the awareness of being the cause of others' distress; thus, empathy and guilt are proposed to have a similar developmental path with empathy as a prerequisite of the development of guilt. Tangney (1991) argued that both empathy and guilt are otheroriented, whereas shame is self -oriented; thus, empathy should be positively associated with guilt after a transgression but negatively related to shame. This

PAGE 24

14 suggestion has been supported by empirical studies (Tangney, 1991) . Therefore, when children harm others empathetic responses may be evoked, and these feelings could influence their stibsequent behaviors and other emotions . Regulation . In recent years, researchers have identified individual differences in the dynamic features of emotion, including the intensity, persistence, modulation, onset and rise time, range, and lability of euid recovery from emotional responses (Thompson, 1994) . Thompson indicated that these dynamic features of emotion are substantially influenced by emotion regulation processes, which he defined as "the extrinsic and intrinsic processes responsible for monitoring, evaluating, euad modifying emotional reactions, especially their intensive and temporal features, to accomplish one's goals" (pp. 27-28). Thus, understanding children' s emotion regulation is crucial in explaining the qualities and dynamics of their emotional experiences . Franko, Powers, Zuroff, and Moskowitz's (1985) study demonstrated that children as young as 6 years can describe conscious strategies for dealing with negative affect. Of 32 children, aged 6 to 11, who participated in their study, 92% showed active rather than passive attempts at selfregulation of emotion. Thus, it is reasonable to expect that

PAGE 25

15 children may employ self-regulation strategies to relieve their feelings of discomfort such as guilt, shame, or fear. Children may employ behavioral or cognitive responses to deal with negative affect (Franko et al . , 1985). Lazarus (1991) suggested that the action tendency with guilt is to expiate by apology or making amends and the tendency in sheune is to hide, to avoid having one's personal failure observed by others . Cognitive Reactions to Aggression Some researchers have explored the cognitive reactions of children following moral transgression. Their research shows that the beliefs children hold about the outcomes of their misbehavior and the ways they explain and justify their wrongdoing may influence the maintenance of the behavior . Outcom e expectancies . Outcome expectancies refer to the rewarding or punishing consequences one anticipates for performing a behavior (Perry et al., 1986). Social cognitive theorists have emphasized that from observing the effects of acts of aggression individuals develop outcome expectancies for aggression (Bandura, 1973) . Crick and Dodge (1994) stated that ^^outcome expectancies can serve an excitatory or an inhibiting function depending on whether the outcomes

PAGE 26

16 expected for particular behaviors are positive or negative" (p. 89) . Research has demonstrated that, compared to nonaggressive children, aggressive children tend to be more confident that aggression will produce positive consequences, such as obtaining a tangible reward, reducing aversive treatment by others, and increasing self-esteem (Perry et al., 1986; Slaby & Guerra, 1988; Xie, 1991; Zhang, 1990) . Thus, aggressive children might develop expectancies that support future aggression from experiencing positive outcomes of their behavior. Legitimacy . Theorists have emphasized the important role children's beliefs about the appropriateness of aggression play in the persistency of aggression. Slaby and Guerra (1988) argued that from experiences children may develop the belief that aggression is an acceptable behavior in many situations. Huesmann and Eron (Huesmann & Eron, 1984; Eron, 1994) have proposed that social behavior is controlled to a great extent by cognitive scripts, or schema. These scripts are encoded rehearsed, stored, and retrieved in one's mind as are other intellectual behaviors. An aggressive strategy or script must be accepted or considered as appropriate by an individual before it can be encoded. Advancing this theory, Huesmann and Guerra (1997)

PAGE 27

17 suggested that through socialization processes an individual develops normative beliefs, which refer to a person's cognitive standards about the acceptability of a behavior. These beliefs set the range of allowable and prohibited behavior. According to this theory, once an aggressive script is stored in one's memory, one may retrieve it, whenever one faces interpersonal problems or is in a situation similar to the one in which the script is encoded. When a script is retrieved, the individual evaluates its appropriateness in light of existing internalized norms and considers the likely consequences . A child with weak or with no internalized prohibitions against aggression, or who believes it is normative to behave in this way, is much more likely to respond with aggression. Research findings show that aggressive individuals are more likely to believe that aggression is a legitimate response (Eron, Guerra, & Huesmann, 1997; Huesmann & Guerra, 1997; Slaby & Guerra, 1988; Zhang, 1990). In a longitudinal study, Huesmann and Guerra (1997) concluded that "children tended to approve more of aggression as they grew older and that this increase appeared to be correlated with an increase in aggressive behavior" (p. 413) . In addition, Guerra and Slaby (1990) found that changing adolescent offenders' belief that aggression is legitimate was more

PAGE 28

effective in reducing aggression than changing other cognitive skills and beliefs included in their intervention program. All these findings suggest that children's beliefs about the legitimacy of aggression may influence their subsequent social behavior. Attributions of responsibility . Aronson (1995) suggested that people tend to justify their cruelty. Mien people clearly harm others, they feel dissonant because their behavior is not consistent with their belief that they are good or decent people. In this situation, the best way to resolve the dissonance is to maximize the culpability of the victim and convince themselves that the victim deserved the injury. Bandura (1991) proposed a theory to explain how people use mechanisms of moral disengagement to avoid moral sanction. According to him, people translate moral reasoning into moral behavior through self -regulation . The selfregulation exercises self-sanction in accordance with the positive or negative consequences of a particular behavior. The mechanism of self -regulation needs to be activated to operate. Conversely, people may use psychological mechanisms to disengage in the exercise of moral agency to avoid negative feelings from self -sanction . These mechanisms of moral disengagement include describing their injurious

PAGE 29

19 behavior as serving a valued social or moral purpose, using euphemistic language to mask aggression, comparing their detrimental conduct to more culpable behavior, displacing responsibility, diffusing responsibility, disregarding or distorting consequences of their harmful actions, and dehumanizing the victims . Researchers (Graham, Doubleday, & Guarino, 1984; Graham, Weiner, & Banish-Weiner , 1995; McGraw, 1987) have found that after transgression if people consider that the situation is out of their control or they attribute responsibility for the misbehavior to others, they are less likely to feel guilt. Research by Bandura, Barbaranelli , Caprara, and Pastorelli (1996) provided evidence that moral disengagement fosters harmful conduct by promoting cognitive and affective reactions conducive to aggression. Therefore, after a transgression children might develop attributions to justify their wrongdoing or reduce self -blame, and these reactions might have a substantial impact on the recurrence of the transgression. These cognitions are important for understanding how aggressive and nonaggressive children differ in their reactions to their own misbehavior. *-

PAGE 30

The Role of Intelligence in Maintenance of Children^ s Aggression Researchers have claimed that intelligence also has an impact on the persistence of aggression. In their longitudinal study, Huesmann, Eron, euid Yarmel (1987) found a bidirectional relation between intelligence and aggression; that is, early IQ was related to early aggression, and early aggression may have impeded the development of intelligence afterward. After reviewing research literature on the relation of IQ to delinquency, Hirschi and Hindelang (1977) proposed that IQ has an impact on the likelihood of delinquency independent of race and class. Gibson and West (1970) also claimed that IQ is an effective predictor of delinquency independent of socioeconomic status. According to Miller (1972), all kinds of deviant behavior increase sharply among children with IQ, lower than 90 . The Rol e of Impulsiveness in Maintenance of Children^ s Aggression Impulsive children tend to act out the first strategy that comes to mind. Camp (1977) postulated that is^ulsive children are limited in their ability to inhibit their first responses to stimuli and are deficient in maintaining sustained response inhibition. Thus, impulsive children are less likely to think of alternative strategies in

PAGE 31

interpersonal situations. Berkowitz (1974, 1993) empasized that in many instances aggressive behavior is an impulsive or involuntazry response to internal and external stimulation. He proposed that due to intense emotional agitation or personality, some people do not stop to think about what might happen if they victimize others . They have an urge to attack an availeible target. A number of studies (Atkins, Stoff, Osborne, & Brown, 1993; Luengo, Carrillo-dela-Pena, Otero, & Romero, 1994; Martin et al., 1994) have found that aggression is associated with poor impulse control. Luengo et al. (1994) also demonstrated that impulsivity is related to a future increase in antisocial behavior . In summary, researchers agree that intelligence and impulsiveness may influence children' s maintenance of aggression. Therefore, a study of aggressive children should take these two constructs into account. Age and Children' s Reactions to Transgression Researchers have claimed the existence of age differences in children's development of guilt, shame, and empathy. Mascolo and Fischer (1995) proposed a theory of children's developmental trajectories of pride, shame, and guilt based on appraisal patterns, which produce these selfevaluation emotions. According to Mascolo and Fischer,

PAGE 32

22 children as young as 6 to 8 years compare their concrete traits with those of their peers, and at the age of 10 to 12 children begin to experience shame about abstract or general personality characteristics. For guilt development, children by about 10 to 12 years of age are able to feel guilty about violation of general moral rules. Therefore, children age 10 to 12 seem to have mature moral and self -evaluative emotions . Hoffman (1982) maintained that age-related improvement in children' s cognitive edaility is conducive to older children' s experience of guilt or empathy in a broad range of situations. Young children, at the third stage of empathy development, according to Hoffman's theory, can ei^erience en^athy for others' feelings, whereas children of late childhood, the fourth stage of empathy development, can also respond eii5>athically to others' life conditions. Comparing first-, third-, and fifth-grade students, Thompson and Hoffman (1980) found that older children reported feeling more guilt, more concern for the victim's welfare, and relied more on principles of right and wrong in explaining their feelings of guilt; younger children were more concerned about detection and punishment and were more likely to report feelings of happiness as a result of the wrongdoing .

PAGE 33

23 In regard to cognitive reactions, Huesmann and Guerra (1997) found developmental differences in children's normative beliefs about aggression. For first graders, normative beliefs were not stable; approval of aggression showed its largest increase between the first and second grades. It appears that the first 2 years of elementary school are an important period for children' s development of normative beliefs. In the fourth grade, noinaative beliefs about aggression become stable and predict later aggressive behavior . Therefore, younger and older elementary school children may differ in the situations in which they feel shame, guilt, and empathy and in beliefs about aggression. Comparing age differences will allow us to learn more about how children react to their misbehavior and the impact of these reactions on their subsequent development of aggressive behavior. Shame and Guilt in Chinese Culture Influenced by the belief that Western societies are guilt oriented and Asian societies are shame cultures, psychologists, sociologists, and anthropologists have been interested in the meanings and implications of guilt and shame in various cultures. Chinese culture has been considered a shame culture. Many studies have been conducted

PAGE 34

24 to investigate whether feelings of shame and guilt as basic emotions are different in Chinese and Western people and how guilt and shame are used as a mechanism of social and behavioral control in Chinese society. However, due to the difficulty of cross-cultural study, research findings are inconsistent . Shame and guilt-oriented discipline. Wilson (1981) , a sociologist, proposed a theory of moral behavior in Chinese society from a social political perspective. According to him, China is a closed and group-oriented society that stresses social order and stability while discouraging individual independence and self -direction . Respect for authority and fixed rules is emphasized. In such an environment, indoctrination of group-centered values has been the focus of moral training, and physical punishment and love -withdrawal have been used as a means of affective manipulation. The outcomes of this type of training are fear, shame or guilt, and deference. When comparing one's own behavior with the ideals manifested by authority or with other members in one's group, Chinese people are vulnerable to feelings of shame. Shame anxiety thus operates as a major dynamic in Chinese group interactions. From 2 years of observation and interview, Fung (1994) concluded that shame continues to play an important role in

PAGE 35

25 disciplining young children in contemporary modernized Taipei , Taiwan . Parents manipulated the immature sense of shame in their children as a major means of teaching their children right from wrong and transmitting cultural values . Parents believe that this socialization protects children from disgrace-shame, especially from being condemned by persons outside the family. Thus, although shaming implies threats of abandonment or ostracism, it may teach children how to be part of a society, that is how to be included rather than to be set apart. Findings of Her' s (1990) phenomenological analysis also suggest the importance of shame as a socialization mechanism in a group-oriented society. She concluded that shame in Chinese culture functions as a confirmation of group membership more than a self -actualization process . Although researchers have found that shame was used by Chinese parents, one study did not show this result. Stevenson, Chen, and Lee (1992) asked mothers and children in a village in Taiwan what the parents would do if their child got a poor grade or misbehaved. Neither mothers nor children responded that the parents would use shaming techniques. The researchers suggested that perhaps shame is used more frequently in public settings such as the

PAGE 36

26 classroom rather than in the family and by other agents outside the family. . Besides using shame to discipline, Chinese parents are described as stronger in authoritarian control and lovewithdrawal techniques than American parents (Chao, 1994; Kriger & Kroes, 1972) . Eberhard (1967) pointed out that love -withdrawal and punishment also facilitate children' s shame and conformity. According to him, in societies using more punishment or love-withdrawal techniques, children feel they are the recipient of actions of others in authority; thus, the most important motive for these children is to avoid sheune and the anger of others . In a review of literature on Chinese socialization processes. Ho (1986) pointed out that threatening, scolding, shaming, and punishment have been used frequently with older Chinese children. However, these techniques are adopted less often by modern Chinese parents. In reality, a variety of socialization techniques rather than a single pattern are used by Chinese parents. Accordingly, some Chinese parents may adopt shaming as a major discipline technique and others may use different methods. However, even if Chinese parents do not use shaming socialization, their children still appear to be more likely to develop a shame personality than children from Western

PAGE 37

27 cultures, through the use of love-withdrawal , physical punishment, and authoritarian and shaming punishment in school (Stevenson et al., 1992; Wilson, 1981). Meanings of shaune and guilt for the Chinese. Theorists of discrete emotion (Izard, 1977) have maintained the existence of basic emotions, including sheuae and guilt, which are consistent across cultures. In contrast, social constructionist theories of emotion have posited that an individual's emotional patterns reflect the influences and functions of the social community (Lazarus, 1991) . Researchers have been interested in whether the connotative meanings of a given emotion differ between cultures. A niimber of studies have been conducted comparing the meanings of shame and guilt between Chinese and Western people, and both commonality amd differences have been reported. Studies have shown differences in the ways that Chinese and Americans view shame. Using the semantic diffej^ential technique, which minimizes the use of culturally loaded statements by employing opposite adjective pairs to assess siibjective experience across cultures, Marsella, Murray, and Golden (1974) found that Caucasian Americans rated shame as more low, weak, dull, rounded, relaxed, and stale than did Chinese Americans. For the Caucasian Americans, shame was less identifiable or

PAGE 38

28 understood, although both groups experienced guilt as an uncomfortable feeling. The researchers stated that shame may be used as a technique of social control more often in Asian societies and children consequently learn to read the cues of shame more readily. Research by Cheng and Page (1995) showed that Chinese participants rated guilt as a more potent force and viewed guilt in a more negative way than did the American participants. Cheng and Page indicated that the results were not unexpected because Chinese culture often uses shame and guilt as a means of controlling the behavior of individuals and Chinese people may experience guilt more strongly and unpleasantly than Americans. On the other hand, some research suggests commonality between Chinese and Western people in meanings of shame and guilt. Bond and Hwang (1986) argued that Chinese and Americans experience emotions in the same way; they only differ in emotional expressions due to different display rules . In their study. Hong and Chiu (1992) asked Chinese students in Hong Kong to recall either a guilt or a shame incident. They found that more guilt-eliciting events than shame-eliciting events were associated with violating a moral norm, holding personal responsibility for the

PAGE 39

29 violation, and eibsence of an audience. This result is consistent with findings from studies for Western people. However, a high proportion of shame incidents were recalled, including violating a moral norm, absence of an audience, and having personal responsibility, showing that guilt and shame incidents were not mutually exclusive. As indicated previously, guilt has been found to be an inhibitor of aggression from research findings in Western societies. In Taiwan, Su (1975) and Hong (1985) also found that guilt was negatively related to aggression for middle school students . Johnson et al . (1987) compared the meanings of guilt and shame for American, Korean, and Taiwanese participants. The Taiwanese were highest on both the guilt and shame measures. However, consistency across national groups was also found. There was consistency among the groups on whether individual test items loaded on the guilt or the shame factors and on the values of the loadings. The crosscultural similarities led the researchers to "question the belief that Asian and Occidental societies differ in the degree to which guilt vs shame are used as mechanisms for social control" (p. 357) . Some anthropologists have questioned the dichotomy of shame and guilt cultures. Eberhard (1967) indicated that

PAGE 40

30 There are hardly any pure shame or pure guilt cultures, that shame and guilt both are utilized to ensure socialization of the individual, and if a culture is called a shame culture this should only indicate that shame is a more prominent agent than guilt, not an absence of guilt, (p. 3) From a cultural study Singer (1953) concluded that "the sense of guilt and the sense of shame are found in most cultures" (p. 79) . He believed that people in industrialized and preindustrialized societies do not differ in guilt sanction or shame Scuiction but in the complex of values, beliefs, and practices. In a broad and in-depth comparison of Chinese and American cultures, Hsu (1965) proposed that the guilt-shame contrast is better understood as an expression of the contrast between a repression-dominated and a suppressiondominated way of social control. In control by repression the rules of behavior are mostly within the individual's mind. Violation of the rule is more likely to call forth the feeling of guilt than anything else, with reference to the superego. In control by suppression, the rules of behavior are more strongly dependent upon the external situation. Violating rules is more likely to lead to the feeling of shame, than anything else, when one thinks of one's fellow human beings. "It is not alleged that any society can be wholly governed by only one of these mechanisms of control" (p. 177) . He also indicated that the mechanism of repression

PAGE 41

31 is more characteristic of Western societies, whereas the mechanism of suppression is more characteristic of Asian societies such as China. Thus, it is reasonable to conclude that there is no pure shame society or pure guilt society. Both guilt and shame are present in Chinese and Western societies. The nature of shame as an uncomfortable affect and the antecedents of shame are the same across these cultures. The differences are that with different social structures, values, and practices, Chinese people have more chances to experience shame and, consequently, are more sensitive to shame, experience a more intensive response to shame, and rely more on external resources for controlling behavior than Western people. This by no means implies that guilt is unimportant for Chinese people or that they feel less guilt than they feel shcune. In terms of moral emotions, Chinese people feel guilt when they violate a moral norm, and guilt seems to function as an inhibitor of prohibited behavior; the feeling of shame facilitates learning right from wrong and provides aversive feedback when one acts inappropriately. Both guilt and shame are critical for moral development in Chinese people.

PAGE 42

Summary Childhood aggression is a serious problem in most industrialized societies . The research literature shows that aggression is relatively consistent across time and situations for aggressive persons from an early age. In order to alleviate the problem of aggression, it is necessary to interrupt the continuity of aggression in addition to preventing it. Researchers have identified the following social-cognitive variedales in the development of aggression in children: outcome expectancies, legitimate beliefs, and justification. Other researchers have identified the social -emotional variables of empathy and guilt as influential in reducing aggression. Shame and fear have been conceived as important for Chinese children to prohibit inappropriate behavior. The purpose of this study is to identify the differences between aggressive and nonaggressive children in their cognitive and emotional reactions to their own transgression.

PAGE 43

CHAPTER 2 RESEARCH DESIGN The purpose of this study was to explore the differences between aggressive and nonaggressive boys in their emotional and cognitive reactions to their own misbehavior, in order to better understand why aggression is repetitive in some boys but not in others. In this chapter, I describe the research participants, the procedures, the measures and the analyses used in this study. Research Participants Participants were 100 thirdand 100 sixth-grade boys from four pviblic elementary schools in a medium-sized city in central Taiwan. One school was in a recent-developed residential district serving a predominantly middle-class community with about 3,600 students. The second school was located in a suburban area serving a lower-middle-class neighborhood with approximate 2,000 students. The other two schools, with about 1,600 students each, were in the downtown area to serve socioeconomically diverse classes . Classes in these schools were used as the unit to recruit participants . Aggressive and nonaggressive boys were selected from these classrooms on the basis of teacher ratings. The raw scores of the teacher ratings were 33

PAGE 44

34 transformed into standard scores . A boy was identified as aggressive if his z score rating of aggression was above 1 . 3 and his raw score was above 18, from a scale with possible scores ranging from 8 to 40. A total of 50 aggressive boys were selected in each third and sixth grades . The noimber of nonaggressive boys was the same as the number of aggressive boys. They were selected from the boys whose z score derived from the teacher rating was below -0.25 and whose raw score was at least 8 points lower than the raw scores of the aggressive boys in his class. Whenever possible, nonaggressive boys were selected from boys who met these criteria and were matched with one aggressive boy in his class on socioeconomic status . If no nonaggressive boy in a class matched the aggressive boy on SES, the nonaggressive boy who had the SES score closest to the aggressive boy in his class was selected. Socioeconomic status has been conceived as an influential factor in children's aggression (e.g., Eron et al . , 1997; Farrington, 1991; Offord, Boyle, & Racine, 1991). Reducing the differences between the two groups in SES was used to avoid confounding the results of this study. Measures Aggression Teachers completed the 8 items on the aggression factor of the Teacher Checklist (Coie & Dodge, 1988) :

PAGE 45

35 threatens or bullies others; uses physical force; starts fights; overreacts to accidental hurts with anger and fighting; gets angry easily and strikes back; gets other kids to gang up on a peer; says mean things to peers; claims other children are to blame in a fight (Appendix A) . Teachers rated the aggression of each boy in their class whose parents consented to participation in this study on a 5-point scale from 1 (not__at_all) to 5 ( very much ) , with scores ranging from 8 to 40, and higher scores indicating more perceived aggression. Reacti ons Following Transgressions Children who met the criteria for selection read three semiprojective stories describing explicit misbehavior and then were interviewed concerning how they would feel and think were they the perpetrators of the misbehavior. Each story depicted an incident involving boys of the same age as the participant. One of the boys in each story commits a transgression, in the first story cheating at checkers, in the second stealing another child's toy, and in the third neglecting to help another after an accident (Appendix B) . Stories with similar content were used in other studies of the development of children's moral emotions (Dunn, Brown, & Maguire, 1995; Kochanska, 1991; Krevans & Gibbs, 1996; Thompson & Hoffman, 1980) .

PAGE 46

36 First, the children were asked to complete the story, explaining how the protagonist thinks and feels and what would happen next? The story completion is designed to assess the intensity of the child' s guilt (Hoffman & Saltzstein, 1967) and extent of reparatory acts (Kochanska, 1991) . All children then were asked to assume the roles of the wrongdoers and to answer questions designed to assess their reactions to misbehavior in each story. Interview questions (i^pendix C) were designed to measure feelings, emotion regulation, empathy, responsibility attribution, outcome expectations, and beliefs about the legitimacy of the transgression. The interviewer showed participants a scale with five bars of increasingly greater height. Boys were asked to point to the bar that illustrated the intensity of their feelings, with the highest bar representing very much and a blank showing not at all . For questions asking children "would you (do) . . .?" the interviewer showed them a 5point scale representing "Definitely not," "Probably not," "Not sure," "probably yes," and "Definitely yes." For the two questions on outcome expectancy the interviewer presented children a 4 -point scale and requested children to choose their responses from the scale (see Appendix C) .

PAGE 47

37 Socioeconomic Status The Chinese version of Hollingshead' s two-factor index of social position developed by Ying-Shin Yang (1985) was used to compute children's socioeconomic status. By this method, parents' education was divided into five levels and occupation into six categories . A formula was used to calculate each boy's socioeconomic status: Index of SES = index of parents' education x 4 + index of parents' occupation x 7. Fathers' information was used for this calculation. When fathers' information was not available, mothers' information was used instead. Impu 1 s i vene s s The Ig Impulsiveness Questionnaire developed by Eysenck, Easting, and Pearson (1984) was used to assess participants' impulsiveness. This questionnaire contains 69 items, but only the 21 items that measure impulsiveness were used. This measurement is applicable for children from the ages of 7 to 15, and the reliability of the impulsiveness scale for boys was .74 in the Eysenck et al . (1984) study of 633 British boys. Intelligence The scaled scores of the Vocabulary and Block Design subtests of the Wechsler Intelligence Scale for Children (WISC) were used to assess the participants' general

PAGE 48

38 intellectual abilities . These two subtests have been used as a measure of general intellectual ability in previous research (e.g., Klaczynski & Gordon, 1996). According to Sattler (1989) , the pro-rated IQ scores derived from these two subtests are highly correlated with the WISC full-scale score (r = .90) . The vocabulary subtest, which asks children to explain word meaning, is an excellent indicator of crystallized intelligence. The block design svibtest, which requires participants to reproduce three-dimensional designs, is a measure of the ability to apply reasoning to solve spatial problems. In this study, these two subsets of the Chinese version of the third edition of the WISC were used. The original WISC-III has been translated into Chinese, revised, and normed in Taiwan. Procedures Teachers of the third and sixth grades in the four elementary schools were invited to participate in this study and asked to sign the teacher consent form (see J^pendix D) . Before selecting the aggressive and nonaggressive children, parental consent forms (see Appendix D) that included the SES questions were given to the boys in the third and sixth grades by their teachers in the four elementary schools. Parents of the boys were asked to sign the consent form and respond to the SES questions if they agreed to allow their

PAGE 49

sons to participate in this study. In total, 88.6% of the parents agreed to their sons' participation. Teachers of these classes were asked to rate each participant on the Teacher Checklist. I then interviewed the boys who met the selection criteria. In 5 days after the interview these children were asked to complete the impulsiveness questionnaire. Two days to 1 week after the interview, these boys took the intelligence test. Data Analysis Coding of Participants' Responses The responses to the story completion were scored on a 6-point scale representing the extent of reparatory acts (Kochanska, 1991) . Score 1 : No resolution of transgression ; no discomfort or any action. Score 2: External resolution of transgression; external sources or circumstances intervene; victim may retaliate . Score 3: Attempt to address the transgression; trying to alleviate distress of the victim, but the action is incomplete . Score 4 : Attempt at reparation : giving up something valued to make up for the wrongdoing or making reparation (beyond mere apology or confession) . Score 5: Reparation and attempt at reconciliation.

PAGE 50

40 Score 6: Full resolution of transgression; carrying out a complete and relatively excessive effort to make reparation and to alleviate physical or emotional consequences of wrongdoing. Statistical Analysis The scores obtained for each question across the three stories were added together. Scores from one interview question were treated as scores for one variable. lANCOVA and MANOVA procedures were used to analyze data whenever appropriate. Fishers' exact test was also used to analyze the categorical data. Limitations of the Study Because of the research design, there are some limitations of this study. First, because the interview questions are not all free-response format, data collected in this study may not show the full range and dynamics of children's reactions. Another limitation comes from the use of children' s responses to semiprojective stories as their reactions to transgressions rather than responses to real life incidents. Although this method is used pervasively, researchers have not examined the validity of the results. Finally, children's emotions and thoughts probably interact with each other. This study treats emotion and cognition separately; thus, the study cannot determine the joint

PAGE 51

41 influences of the two aspects on children' s aggressive behavior . Pilot Study A pilot study was conducted to determine whether the interview questions proposed for the study were appropriate for use with boys in the third and sixth grades in Taiwan. The interview questions were revised on the basis of the boys' reactions during the interview and the results of the t-tests comparing the responses of the aggressive and nonaggressive boys . Participants Seven aggressive (four sixth graders and three third graders) and five nonaggressive (three sixth graders and two third graders) boys participated in this pilot study. Participants were recruited from two sixth-grade and two third-grade classes at an inner-city elementary school in central Taiwan. They were selected based on teacher ratings and peer nominations. A boy was identified as aggressive when his teacher rating was above the median for his class, his peer nomination was above the 80th percentile, and he had at least twice as many aggressive as prosocial nominations. The nonaggressive children were selected from children with a teacher rating below the class median, peer nomination below the 80th percentile, and SES matching one of the aggressive boys in his class.

PAGE 52

42 Procedures Two sixth-grade teachers and two third-grade teachers were asked to respond to the teacher checklist for each boy in their classes . Boys in these classes were asked to answer the peer nomination questions and referred to a list of other boys in their class . After selecting the participants according to the criteria described aibov&, I interviewed each boy individually, using prepared questions. Children's replies were written down and tape-recorded. Then, the story completion was coded by another rater and me. Finally, the pilot study data were analyzed by t-tests cuid Pearson correlations . Results Differences between aggressive and nonaggressive boys. The pilot study data were analyzed by t-tests to compare aggressive and nonaggressive boys' responses. Table 1 shows the means on each question for both groups and results of the t-tests. Because the sample size was small, it was unlikely that differences would be statistically significant, but differences between group means could reveal whether the direction of differences is consistent with the study's hypotheses. On the 26 interview questions, differences of group means for 15 were in the expected direction: fear, guilt, sadness, shame, time (time thinking

PAGE 53

43 Table 1 Differences between Aggressive and Nonaagressive Children in Reactions to Moral Transgression Variable N M Prob> |T| Fear Nagg 7 5 9.57 11.00 0.4620 Guilt Nagg 7 5 8.42 9.60 0.5506 Sad Nagg 7 5 7.42 9.00 0.4153 Happy Nagg 7 5 3.14 3.80 0.2286 Shame Nagg 7 5 8.85 9.80 0.5712 Proud Nagg 7 5 3.14 3.00 0.4241 Disturb days Nagg 6 4 8.17 8.25 0.9599 Disturb eat Nagg 6 4 5.50 11.50 0.0074 Time Nagg 3 4 13.66 18.25 0.4343 Apology Nagg 7 5 10.42 12.20 0.3625 After Apology Nagg 3 4 7.00 6.75 0.9110 Hide Nagg 7 5 4.00 6.40 0.1607 After Hide Nagg 0 2 8.50

PAGE 54

Table 1 — continued Variable M Prob > iTl Help Agg 7 Nagg 5 After Help Agg 4 Nagg 4 Empathy A Agg 7 Nagg 5 Empathy B Agg 6 Nagg 5 Empathy C Agg 7 Nagg 5 Friend Agg 6 Nagg 5 Gain Agg 6 Nagg 5 Enjoy Agg 7 Nagg 5 Bad Agg 6 Nagg 4 Responsibility Agg 7 Nagg 5 Blame Agg 7 Nagg 5 Other Respn Agg 7 Nagg 5 Other Blame Agg 7 Nagg 4 9.71 11.60 4.50 6.25 11.00 12.40 10.83 12.00 11.29 12.80 10.00 11.60 10.66 11.60 5.00 5.60 12.33 12.25 12.14 12.80 10.85 11.80 4.14 4.40 3.28 4.25 0.3246 0.2215 0.2768 0.3395 0.3168 0.0883 0.1880 0.6966 0.9517 0.5855 0.4484 0.77 0.0963

PAGE 55

45 persisted) , disturbing eating, apology, help, after help, Empathy A (victim upset) , Empathy B (victim sad) , Empathy C (feel bad about other's hurt), friend like/not like, gain/not gain, and self -blame. For two questions data were in the opposite direction: happy and other blame. The mean values were very similar for five questions: proud, disturbing days, enjoy, bad, and other's responsibility. Only two aggressive children answered that they would hide after a wrongdoing; therefore, only these two boys were asked the "after hide" question, and data were not available for calculating differences in responses to this variable. Differences between third and sixth graders . The second set of t-tests were conducted to euialyze age differences between third and sixth graders. Table 2 presents the means on each question for the two grades and the results of the t-tests. The sixth graders scored higher than the third graders on seven questions: guilt, shame, disturb days, time, after help, and Empathy C (feeling), and bad; these were in the expected direction. On four questions sixth graders scored lower than third graders: disturb eating, apology, help, and Empathy B. There were no apparent grade differences in the answers of the other 15 questions (differences between the group means were less than 0 . 5 or the E values higher than .70) .

PAGE 56

46 Table 2 Differences between Third and Sixth Graders in Reactions to Moral Transgression Variable IS Prob> iTl Fear 3rd 5 6th 7 Guilt 3rd 5 6th 7 Sad 3rd 5 6th 7 Happy 3rd 5 6th 7 Shame 3rd 5 6th 7 Proud 3rd 5 6th 7 Disturb days 3rd 3 6th 7 Disturb eat 3rd 3 6th 7 Time 3rd 2 6th 5 i^ology 6th 7 After Apology 3rd 4 6th 3 Hide 3rd 5 6th 7 After Hide 3rd 1 6th 1 10.20 10.14 7.80 9.71 8.20 8.00 3.20 3.57 8.20 10.00 3.20 3.00 6.67 8.86 8.67 7.57 12.50 17.80 11.80 10.71 7.00 6.67 4.60 5.28 8.00 9.00 0.9769 0.3215 0.8053 0.5074 0.2684 0.2550 0.1911 0.7127 0.4072 0.5820 0.8814 0.7023

PAGE 57

Table 2 — continued 47 Variable If Prob> I T I Help 3rd 5 6th 7 After Help 3rd 5 6th 3 Empathy A 3rd 5 6th 7 Empathy B 3rd 5 6th 6 Empathy C 3rd 5 6th 7 Friend 3rd 4 6th 7 Gain 3rd 4 6th 7 Enjoy 3rd 5 6th 7 Bad 3rd 3 6th 7 Responsibility 3rd 5 6th 7 Blame 3rd 5 6th 7 Other Respn 3rd 5 6th 7 Other Blame 3rd 4 6th 7 11.20 10.00 4.60 6.67 11.80 11.53 12.20 10.67 11.00 12.57 11.00 10.57 10.75 11.28 5.60 5.00 11.67 12.57 12.20 12.57 11.00 11.43 4.40 4.14 3.75 3.57 0.5371 0.1521 0.7792 0.1997 0.2978 0.6835 0.4812 0.6966 0.5329 0.7591 0.7335 0.7506 0.7756

PAGE 58

48 Correlations between scores on interview questions . Table 3 shows the Pearson correlations between scores on the interview questions. It reveals that most of the significan'^ relationships were consistent with expectations drawn from extant literature . However , there are two exceptions : the negative relationships between after help and Empathy B (-0.73) and between after apology and time (-0.91) were not in the expected direction. Children's responses to the question, "How long would you keep thinking about this incident," ranged from several hours to their whole life, making it difficult to categorize their answers. In the revised questionnaire children will be asked to indicate on the scale of five bars of increasingly greater height to point to the bar that illustrates the length of time they would think about the incident with the highest bar representing several years and a blank indicating not at all. Again, the "after hide" question consists of the responses of only two children. Revision of these interview questions seems to be needed. Ratings of sto ry completion. The researcher and another doctoral student rated participants' story completion narratives on a 6-level scale developed by Kochanska to measure children's intensity of guilt. First, we read and discussed the criterion for each level. Then, we

PAGE 59

49 Table 3 Correlations Between Scores on Interview Questions fear guilt sad hapy sh2uii proud both unea tiaa aplgy aapl hide ahi 0.70* 0.84** 0.20 0.62* 0.29 0.13 0.67* 0.58 0.47 0.51 0.39 -1.00 9»iilt 0.82** 0.24 0.77** 0.11 0.32 0.36 0.68 0.44 -0.02 0.13 1.00 Mid -0.05 0.72** 0.19 0.21 0.63 0.59 0.32 0.31 0.22 -1.00 *i*Py -0.05 -0.15 -0.10 0.22 0.36 0.07 -0.02 0.60 1.00 0.09 0.66* 0.52 0.57 0.39 0.45 0.25 -1.00 P^oud -0.48 -0.08 -0.53 0.08 0.54 -0.22 . 0.27 0.65 0.27 -0.75 0.35 -1.00 0.40 0.19 0.24 0.74* -1.00 0.24 -0.91* 0.56 1.00 »P^9y -0.07 -0.12 -1.00 0.06 -1.00 •hi iMlp •hp fxad grain enj bad rspn blam orspn oblam

PAGE 60

50 Table 3 — continued help ahp enpa en^b en^c fmd gain enj bad rspn blam orsp obla fear 0.42 0.27 0.70* 0.27 0.78** 0.24 0.24 0.23 0.75* 0.64* 0.63* 0.19 0.30 guilt 0.50 0.30 0.69* 0.25 0.88** 0.33 0.75**-0.29 0.65* 0.68* 0.72* -0.44 -0.29 sad 0.49 0.38 0.52 0.31 0.73** 0.51 0.54 -0.17 0.74* 0.54 0.56 -0.14 0.07 hap -0.24 0.04 0.53 0.23 0.26 -0.04 0.24 0.52 0.15 0.05 0.34 0.06 -0.02 sham 0.42 0.63 0.39 0.21 0.82** 0.58 0.62* -0.25 0.59 0.56 0.55 -0.23 0.03 proud 0.15 -0.29 0.21 0.28 0.01 0.06 -0.03 -0.03 0.13 0.09 0.12 -0.06 -0.23 dday 0.17 0.41 0.05 -0.01 0.57 0.26 0.29 -0.02 0.06 0.17 0.12 0.01 0.23 eat 0.27 0.63 0.41 0.44 0.53 0.70* 0.29 0.36 0.49 0.22 0.36 0.38 0.67* time 0.21 0.09 0.52 0.36 0.92** 0.35 0.51 0.13 0.63 0.39 0.62 -0.05 0.21 aplgy 0.86**-0.31 0.73** 0.54 0.63* 0.20 0.41 -0.17 0.12 0.82** 0.65* -0.25 0.02 aapl -0.19 0.54 0.03 -0.26 -0.05 0.00 -0.61 0.25 0.47 0.26 0.09 0.41 0.24 hid -0.29 0.50 0.27 0.29 0.35 0.34 0.06 0.75** 0.25 -0.15 0.16 0.54 0.62* ahi -1.00 1.00 1.00 1.00 -1.00 1.00 1.00 -1.00 1.00 -1.00 -1.00 *^lP -0.35 0.60* 0.56 0.58 0.41 0.51 -0.47 0.14 0.81** 0.65* -0.40 -0.02 hp -0.48 -0.73* 0.23 0.42 0.21 0.04 0.42 -0.06 -0.24 0.28 0.42 •°P* 0.71* 0.76** 0.20 0.48 0.14 0.33 0.74** 0.82**-0.16 -0.04 0.37 0.71* 0.50 0.01 0.22 0.35 0.64* -0.26 0.03 fznd gain •njy bad blSM orspn oblaun 0.39 0.65* -0.04 0.56 0.73** 0.76**-0.16 0.08 0.63* -0.23 0.37 0.24 0.48 -0.24 0.21 -0.52 0.42 0.49 0.65* -0.74**-0.41 -0.14 -0.31 -0.16 0.84** 0.65* 0.55 0.70* -0.27 -0.24 0.86**-0.34 -0.12 -0.38 -n 15 0.89**

PAGE 61

51 coded children's responses separately. We agreed on the coding of 91.17% of the children's responses (31 of the total 34) . One answer of a boy in each group could not be coded because the boys failed to provide sufficient information; therefore, the total mimber of scores was 34, not 36. For the first story, all five of the nonaggressive boys scored 2 (relying on external resolution) . Among the seven aggressive children, four of them scored 2, two scored 3 (confess or apology) , and one scored 4 (reparation) . For the second story, all children received scores of 2, except one aggressive boy whose score was 4. For the third story, one nonaggressive child scored 3, and all the other nonaggressive boys scored 2. Among the aggressive children, two scored 3, four scored 4. This suggests that nonaggressive children may rely more heavily on adults' solutions or are more afraid of retaliation from victims than aggressive children. The results seem inconsistent with findings from previous studies. The reason for the unexpected results may be because the stories used in this pilot study differed slightly from those used in Kochanska' s (1991) study, as explained in the section that follows.

PAGE 62

52 Revising Interview Questions and Stories Three interview questions require children to answer whether after apologizing, hiding, or helping they would still feel discomfort (for example, after you apologize, would you still feel discomfort or not at all?) , Children were expected to reply "Yes" or "No" to these questions, but some children spontaneously responded with the degree of discomfort they would still feel after apologizing and helping. Therefore, the interview questions about feelings after apologizing and helping were changed to ask children to indicate whether the uneasy feelings would remain after their actions; that is, they were requested to choose their emswer from "Definitely not" to "Definitely yes" on a 5point scale, as other "Would you (do) . , . ?" questions. On the other hand, because most children reported that they would not hide after this incident, questions about intention to hide and feelings after hiding were omitted. For the question "What would you think to make you feel better after you did what XXX did?", children only answered "No" at first. Then I began to try prompting the children's response. After I prompted, some excuses for wrongdoing emerged in some children's responses (some still said "No"). For example, after the answer "No," I asked the boys, "Would you think that you are a better chess player, that you should win, or that moving two checkers won't be a

PAGE 63

53 big deal?" One boy, then, gave the answer that Andy did not concentrate on playing checkers so he lost. Because the original question seems too broad for children to answer and because giving children prompts may create different effects on different children, 5 answers were provided and children were asked to choose 1 from the 5 (see i^pendix C) . These 5 answers represented 5 coping strategies, including coo^aring with others' wrongdoing, dehumanizing others, blaming others, thinking about compensation, and shifting attention. One possible reason for the unexpected results on the story completion task is that the wrongdoers in the second and third stories need money to repair their wrongdoing. Children may fear that the reparation will cost their parents' money and they could be punished. In the revised stories no money is necessary to con^jensate for the wrong done and no audience is present during the misbehavior.

PAGE 64

CHAPTER 3 RESULTS The purpose of this study was to understand the differences between aggressive and nonaggressive boys in their affective and cognitive reactions following misbehavior. Specifically, the study was designed to investigate whether after misbehavior aggressive and nonaggressive boys differ in the following reactions: types of feelings, intensity of feelings, length of emotional arousal, outcome expectations for aggression, legitimate beliefs in aggression, and attributions of responsibility. Differences between sixthand third-grade boys in these reactions were also explored in this study. After interviewing 200 participants and testing their intelligence and impulsiveness, a variety of statistical methods were conducted to answer the research questions. Chapter three presents the results and findings of the analyses of data examined in this study. Statistical methods used in the data analyses included MANOVA, MANCOVA, ANCOVA, and Fisher's exact test. MANCOVA was planned as the major technique to analyze continuous data of this study. The use of MANCOVA needs to meet two 54

PAGE 65

55 assumptions: First, the set of dependent variables is significantly correlated with the set of covariates, and, second, the regression of dependent variable (s) on the covariate(s) is equal across groups. The SPSS MANOVA program was used to examine data to determine whether it was appropriate to use MANCOVA (Stevens, 1996) to compare groups in this study. Because the two planned covariates, IQ and impulsiveness, were not highly correlated with most variables, MANCOVA was appropriate only in the analyses of the story coii5>letion and reactive feelings with IQ as covariate; iii5>ulsiveness was not used as a covariate. MANOVA was used for testing for group differences with multiple dependent variables. Each multivariate analysis was followed by univariate F-tests, in order to locate the variable (s) contributing to the multivariate difference. In analyzing data on cognitive coping strategy. Fisher's exact test was employed because these data were categorical and from a small sample . Descrip tive Data Table 4 shows the means and standard deviations on each variable in this study (except the categorical variable, cognitive coping strategy) for the aggressive and nonaggressive boys in the sixth grade. These data for the third graders are displayed in Table 5. The correlation matrixes for all the variables are presented in Appendix E.

PAGE 66

56 Table 4 Means a nd Standard Deviations on Each Variable for Sixth Graders Aggressive Nonaggressive Variable a N SD M N SD I.Q. 76 .48 46 19 .08 85 .33 49 15 . 61 Completion 5 .62 47 1 .24 6 .59 49 1 . 97 Fear 11 .35 48 2 .36 12 .08 49 2 .17 Guilt 11 .29 48 2 .68 12 .47 49 2 . 06 Sad 8 .73 48 2 .83 10 .27 49 2 . 46 Happy 4 .21 48 1 .66 3 .69 49 1 . 04 Shame 10 .65 48 2 .51 12 .18 49 2 . 38 Proud 3 .81 48 1 .38 3 .57 49 0 98 J^ology 13 .23 47 2 .15 13 . 94 49 1 45 After apology 6 .21 47 2 .99 7 . 47 49 3 37 Help 13 .04 48 1 .77 13 . 92 49 1 . 61 After help 6 .38 47 3 .01 6 .59 49 2 81 Disturb life 10 .23 48 3 . 11 11 43 4Q £. . D / Disturb two days 11 .98 48 2 .25 12 . 73 49 1 fin Disturbed time 10 85 48 2 24 11 Victim upset 11 94 48 2 13 12 41 49 O £^ 1 Q Victim sad 11. 00 48 2. 29 11. 88 49 2. 16 Empathy feel 11. 87 47 2. 23 13. 04 49 1. 90 Friend 11. 23 48 0. 95 11. 29 49 0. 87 Gain 10. 47 47 1. 67 11. 18 49 0. 95 Enjoy 5. 96 48 2. 58 5. 14 49 2. 00 Bad 13. 19 48 1. 59 13. 55 49 1. 62 Self responsibility 13. 31 48 1. 68 13. 43 49 1. 44 Self blame 12. 85 48 1. 82 13. 22 49 1. 77 Victim responsibility 6. 23 48 2. 57 5. 24 49 1. 56 Victim blame 5. 48 48 1. 99 4. 53 49 1. 62

PAGE 67

57 Table 5 Means and Standard Deviations on Each Variable for Third Graders Aggressive Nonaggressive Variable 1 If N SD M N SD I-Q. 5 .55 47 14 .40 61 .00 47 15 .42 Completion 5 .20 44 1 .34 5 .46 48 1 .17 Fear 11 .00 48 2 .62 11 .37 49 2 .32 Guilt 11 .17 48 2 .57 12 .02 49 2 .15 Sad 9 .73 48 3 .47 10 .29 49 3 .03 Happy 5 .19 48 2 .63 4 .20 49 1 .62 Sheune 9 .54 48 3 .53 10 .59 49 3 .21 Proud 4 .90 48 2 .15 3 .98 49 1 .71 i^ology 13 .46 48 2 .25 14 .22 49 1 .46 After apology 4 .74 47 1 .89 5 .08 49 2 .37 neJ.p 13 .50 48 2 .25 14 .00 49 1 .41 Axter help 4 .36 45 1 .82 4 .67 49 2 .03 uiscuro ±jLre 8 . 68 47 3 .95 9 .49 49 3 .64 Disturb two days 10 .02 47 3 .94 10 .08 49 3 .50 Disturbed time 9 96 48 2 .87 9 .94 49 2 .90 Victim upset 12 44 48 2 20 12 .39 49 2 48 Victim sad 11. 13 48 3 12 11 59 49 2 65 Empathy feel 11. 65 48 2. 65 12 35 49 2 31 10 . 62 48 1. 33 10. 94 49 1. 21 Gain 10. 71 48 1. 30 11. 33 49 0. 85 Enjoy 6. 42 48 3. 11 5. 94 49 3. 08 Bad 12. 75 48 2. 32 13. 51 49 1. 86 Self responsibility 12. 36 47 2. 45 13. 52 48 1. 71 Self blame 11. 71 48 2. 21 12. 37 49 2. 10 Victim responsibility 5. 04 48 2. 81 4. 53 49 1. 76 Victim blame 4. 31 48 1. 69 4. 16 49 1. 48

PAGE 68

58 Differences Between Aggressive and Nonaggressive Boys Because there were age differences, data from the sixth and third graders were separated when comparing the two aggressiveness groups; that is, the sixth-grade aggressive boys were only compared with the nonaggressive boys in the sixth grade, and the third-grade aggressive boys were compared only with nonaggressive boys in the same grade. This section presents results of the statistical analysis on each variable explored in this study, with results of the sixth graders described prior to those of the third graders. Story Completion Story completion scores were from open-ended questions, such as "what would happen next," with a higher score reflecting higher motivation to compensate. Two sixth graders and five third graders did not reply to all three of the stories, and three boys in sixth grade and six in third grade did not have IQ scores; therefore, only 89 third-grade and 95 sixth-grade boys were included in the analysis of the story completion scores. Because the IQ scores were significantly correlated with story completion scores, an ANCOVA was performed on the story completion scores for both sixth and third graders, using IQ as the covariate. The results show (see table 6) that aggressive and nonaggressive boys in the sixth grade

PAGE 69

59 Table 6 ANCOVA Test on Story Completion for Sixthand Third-Grade Aggressive and Nonaggressive Boys with IQ as Covariate Sum of Mean Source Squares df Square F 6th grade Aggression 12.396 1 12.396 4.753 .032 IQ 16.593 1 16.593 6.362 .013 3rd grade Aggression 0.182 1 0.182 0.124 .725 IQ 9.216 1 9.216 6.277 .014

PAGE 70

60 differed in their responses to the story completion (F = 4.753, E = .032) . The nonaggressive boys were more likely tr have higher motivation and would put more effort into con^ensating for their misbehavior. For third graders, an ANCOVA was also conducted on story completion scores with IQ as the covariate. No difference in story completion was found between aggressive and nonaggressive boys (F = .124, p = .725). Reactive Feelings Six interview questions were used to assess the six feelings after wrongdoing examined in this study. Boys were asked how fearful (guilty, ashamed, sad, happy, and proud respectively) would they feel if they did what the protagonist did in each story. As can be seen in Table 7, the MANCOVA results show that the overall reactive feelings of the aggressive and nonaggressive boys in the sixth grade were significantly different, with IQ controlled (F = 2.661, E = .020). The univariate F-tests (see Table 8) and descriptive data (see Table 4) revealed that aggressive boys reported lower levels of sadness (F = 10.092, p = .002), shame (F = 7.590, p = .007), and guilt (F = 4.426, p = .038) than nonaggressive boys after wrongdoing. They did not differ in their feelings of fear, happiness, and pride. 1

PAGE 71

61 Table 7 MANCOVA Test of Reactive Feelings for Aggressive and Nonaaaressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df 6th grade Agg .845 2.661 6 87 .020 IQ .704 6.101 6 87 .000 3rd grade Agg .929 1.101 6 86 .368 IQ .813 3.307 6 86 .006 Table 8 Univari ate F-Tests of Reactive Feelings for Aggressive and Nonaggressive Bovs in Grade Six and Three S\im of Mean Source Squares df Square F C 6th grade Fear 5 .407 1 5 .407 1 .090 .299 Guilt 24 .506 1 24 .506 4 .426 .038 Sad 69 .326 1 69 .326 10 .092 .002 Happy 7 .424 1 7 .424 3 .786 .055 Shame 43 .547 1 43 .547 7 .590 .007 Proud .214 1 .214 .161 .689 3rd grade Fear 498 1 498 .082 .776 Guilt 8. 685 1 8 685 1 606 .207 Sad 3. 257 1 3 257 303 .583 Happy 13. 379 1 13. 379 2 952 .089 Shame 13. 876 1 13. 876 244 .268 Proud 11. 535 1 11. 535 3. 382 .069

PAGE 72

62 For the third graders, the MANCOVA did not indicate a significant overall difference (F = 1.101, e = -368) between aggressive and nonaggressive boys in their feelings following misbehavior. The univariate analyses did not show any differences in these feelings. The means for the four negative feelings were much higher than means for the two positive feelings. This difference indicates that after misconduct boys in each group, no matter what their aggressiveness and grade, tended to feel much more fear, sadness, guilt, and shame than happiness and pride. Emotion Regulation This section contains information on boys' cognitive coping strategies, intention to help and apologize to the victim, intensity, and length of disturbance following misconduct. Cognitive copi ng strategy . One interview question asked boys what would they think to help themselves feel better after the accident happened and asked them to select one of five coping strategies. Boys' answers were assigned categorical scores, and the contingency tables have cells with expected frequency of less than 5; hence Fisher's exact test was used to analyze data from this question.

PAGE 73

63 As shown in Table 9, more boys in both aggressiveness groups reported that they would use the fourth strategy, thinking how to compensate for their wrongdoing, than selected other strategies . No boys in either group selected the first (comparing with others' detrimental conduct) or second strategies (devaluing the victim) . Data analysis yielded significant differences between sixth-grade aggressive and nonaggressive boys on all three stories (x^ = 6.939, E = .031; = 13.172, p = -001; = 9.444, p = .009, respectively) . More nonaggressive boys were likely to think of how to compensate for their wrongdoing, whereas more aggressive boys were likely to diffuse responsibility (strategy three) or try not to think about the event (strategy five) . The results for third graders are presented in Table 10. Most of the boys in both groups would think of a way to compensate the victim; some would try not to think about the accident, and a few boys would try to justify their wrongdoing. No difference was found between the two groups in their choice of cognitive coping strategy for each of the three stories (x^ = 2.907, p = .406; x^ = 2.378, p = .498; x^ = 1.508, E = .825, respectively). Intention to h elp and apologize to the victim . Boys were asked would they help or apologize to the victims in

PAGE 74

64 Table 9 Frequency and Fisher' s Exact Test for Cognitive Coping Strategy for Aggressive and Nonaggressive Boys in Sixth Grade Strategy 12 3 4 5 S Agg Nonagg 0 0 0 4 0 0 34 44 10 6.939 5 .031 Story 2 Agg Nonagg 0 0 0 6 0 1 31 47 9 13.172 1 .001 Story 3 Agg Nonagg 0 0 0 8 0 2 31 44 8 9.44 2 .009 Table 10 Frequency and Fisher' s Exact Test for Cognitive Cooing Strategy for Aggressive and Nonaggressive Boys in ThirdGrade Strategy 1 2 3 4 5 X2 Story 1 Agg Nonagg 3 0 2 0 4 1 33 7 2.907 35 11 .406 Story 2 Agg Nonagg 1 0 0 0 2 4 34 10 2.378 38 7 .498 Story 3 Agg Nonagg 2 1 1 0 7 7 31 6 1.508 34 7 .825

PAGE 75

order to understand these boys' intention to repair their misconduct. Results of the MANOVA showed that sixth-grade aggressive boys were different from nonaggressive boys in their intention to make reparations to the victim (F = 4.125, p = .019) . The univariate analyses revealed a difference in their intention to help the victim (F = 6.705, P = .011), but not in their intention to apologize to the victim (F = 3.575, p = .062). For the third graders, no overall difference was found between aggressive and nonaggressive boys (F = 2.121, p = .126) . However, a univariate analysis revealed a difference in their intention to help the victim (F = 3.972, p = .049) . Intensity of disturbance . Boys were asked if they still felt discomfort after apologizing and after helping the victim and if these incidents would disturb their daily life. These three questions were used to assess the intensity of the boys' disturbance due to these incidents. MANOVA results yielded a significant overall difference between sixth-grade aggressive and nonaggressive boys (F =3.406, p = .021). The univariate F-tests and descriptive data revealed that nonaggressive boys were more likely to be disturbed in daily life (F = 4.760, p = .032) and to continue to feel discomfort after their apology (F = 4.574, p = .035), as displayed in Table 13 and 14.

PAGE 76

66 Table 11 MANOVA of Intention to Compensate for Aggressive and Nonaggressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df e 6th grade .919 4.125 2 94 .019 3rd grade .957 2.121 2 94 .126 Table 12 Univariate F-Tests of Intention to Compensate for Aggressive and Nonaggressive Boys in Grade Six and Three Slim of Mean Source Squares df Square F 6th grade Apology 11 .914 1 11, .914 3, .575 .062 Help 19, .306 1 19, .306 6, .705 .011 3rd grade Apology 14, .233 1 14. .233 3. .972 .049 Help 6. .062 1 6. .062 1. .724 .192

PAGE 77

67 Table 13 MANOVA of Intensity of Disturbance for Aggressive and Nonagqressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df 6th grade .899 3.406 3 91 .021 3rd grade .984 .472 3 89 .702 Table 14 Univariate F-Tests of Intensity of Disturbance for Aggressive and Nonaaaressive Bovs in Grade Six and Three Sum of Mean Source Squares df Square 6th grade After i^ology 45.344 1 45.344 4.574 .035 After Help .954 1 .954 .111 .739 Disturb Life 39.983 1 39.983 4.760 .032 3rd grade After Apology 1.609 1 1.609 .342 .560 After Help 1.911 1 1.911 .507 .478 Disturb Life 15.134 1 15.134 1.074 .303

PAGE 78

68 The MANOVA did not show any difference in the thirdgrade boys' responses to the questions about the intensity of their disturbance (F = .472, e = .702). The univariate tests also did not show any differences. Length of disturbance . Two interview questions were designed to assess the length of disturbance elicited by wrongdoing: If these incidents would disturb them for a couple of days, and how long they would keep thinking about these events. As presented in Table 15, no significant difference in length of disturbance was found between sixthgrade aggressive and nonaggressive boys (F = 1.653, p = .197) . The univariate tests did not reveal any significant difference in the likelihood of being disturbed for two days or in the length of time disturbed (see Table 16) . Third-grade boys in the two aggressiveness groups did not differ in length of disturbance after their wrongdoing (F = .048, E = .953) . No differences were found in the univariate analyses. Empathic Reactions Boys were asked how upset and how sad would they think the victim would be to determine boys' cognitive empathy. Another question, how bad would they feel if they saw the victim was hurt, was used to assess boys' emotional empathy. Data on empathic reactions from the three interview

PAGE 79

69 Table 15 MRNOVA of Length of Disturbance for Aggressive and Nonaggressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df 6th grade ,966 1.653 2 94 .197 3rd grade .999 .048 2 93 .953 Table 16 Univariate F-Tests of Length of Disturbance for Aggressive and Non aggressive Boys in Sixthand Third-Grade Siom of Mean Source Squares df Square J g 6th grade Days Disturbed 13.841 1 13.841 3.219 .076 Time Thinking 4.953 1 4.953 1.073 .303 3rd grade Days Disturbed .087 1 .087 .006 .937 Time Thinking .716 1 .716 .087 .769

PAGE 80

70 questions were analyzed by MANOVA. As illustrated in Table 17, a significant overall difference on the three questions was found for the sixth-grade aggressive and nonaggressive boys (F = 2.732, p = .048). The univariate analyses (see Table 18) showed that nonaggressive boys tended to feel a higher level of enrpathy when seeing the other boy upset due to their misconduct (F = 7.637, p = .007) . There were no differences in these boys' estimate of their victim's upset and sadness . For third graders, there was no overall difference between the aggressive and nonaggressive groups (F = 1.110, E = .349). Univariate analyses also showed no differences between them in cognitive and emotional empathy. Outcome Expectations Boys were asked two questions to assess their outcome expectations: What do you think your friend would think about your behavior and do you think what you did would help you get what you like. As shown in Table 19, results of the MANOVA indicate that outcome expectations were different for aggressive and nonaggressive sixth-grade boys (F = 3.277, p = .038). In univariate F-tests, it was found that aggressive boys were more likely to believe that aggression can help them get what they want (F = 6.751, p = .011); they did not

PAGE 81

71 Table 17 MANOVA n-F Fmpa thic Reactions for Aggressive and Nonaggressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df 6th grade .918 2.732 3 92 .048 3rd grade .964 1.149 3 93 .334 Table 18 Univariate F-Tes ts of Empathic Reactions for Sixth and Third-Grade Boys Sum of Mean Source Squares df Square 6th grade Victim Upset 4.873 1 4.873 1.037 .311 Victim Sad 15.065 1 15.065 3.127 .080 .007 Empathy 32.754 1 32.754 7.637 3rd grade Victim Upset .060 1 .060 .011 917 Victim Sad 5.284 1 5.284 .631 429 Empathy 11.919 1 11.919 1.932 .igs

PAGE 82

72 Table 19 MANOVA of Outcome Expectations for Aggressive and Nonaggressive Boys in Grade Six and Three Source Wilks' A F Hypothesis df Error df 6th grade .932 3 .377 2 93 .038 3rd grade .925 3 .835 2 94 .025 Table 20 Univariate F-Tests of Outco m e Expectations for Aggres and Nonaggressive Bovs in Grade Six and Three Source Svun of Squares df Mean Square F 6th grade Friend Object 0.064 12.284 1 1 0.064 12.284 .077 6.751 .782 .011 3rd grade Friend Object 2.387 9.267 1 1 2.387 9.267 1.472 7.676 .228 .007

PAGE 83

73 differ in their belief that their friends would like what they did in the story (F = .077, e = .782). For third-grade boys, the pattern was the same as for the sixth graders. There was an overall difference between aggressive and nonaggressive boys' outcome expectations for their misbehavior (F = 4.726, p = .011) . The aggressive boys tended to believe that aggression could help them obtain what they want (F = 11.193, p = .003) . The boys did not differ in their belief that their friend would like their wrongdoing (F = 1.472, p = .228). Legitima cy Beliefs Data were gathered on boys' legitimacy beliefs from their answers to two questions: How happy would you feel with your win (toy and the movie) and how bad do you think it was to do this? As presented in Table 21, there was no overall significant difference between sixth-grade aggressive and nonaggressive boys in their beliefs about the legitimacy of aggression (F = 1.600, p = .207). Univariate analyses also showed no differences in levels of happiness elicited by what boys obtained through misbehavior and no difference in their evaluation of how bad their behavior would be (see Table 22) . There was no overall difference in beliefs about the legitimacy of aggression between third-grade aggressive and

PAGE 84

74 Table 21 MANOVA of Beliefs about the Legitimacy of Aggression for Aggressive and Nonaggressive Boys in Grade Six and Three Wilks' Hypothesis Error Source A F df df p 6th grade .967 1.600 2 97 .207 3rd grade .968 1.571 2 97 .213 Table 22 Univari ate F-Tests of Beliefs about the Legitimacy of Aggression for Aggressive and Nonaggressive Boys in Grade Six and Three Sum of Mean Source Squares df Square F p 6th grade Enjoy 16.125 1 16.125 3.028 .085 Bad 3.204 1 3.204 1.240 .268 3rd grade Enjoy 5.538 1 5.538 .577 .450 Bad 14.013 1 14.013 3.175 .078

PAGE 85

75 nonaggressive boys (F = 1.571, e = -213) indicated in the MANOVA test. Univariate analyses revealed no differences in the two groups of boys in their enjoyment of their gains through misbehavior or in the evaluation of their own behavior . Responsibility Attribution Self -responsibility . Boys were asked how responsible they would feel and how much they would be to blame for what happened to assess their feelings of responsibility. Results of the MANOVA show that there was no overall difference between sixth-grade aggressive and nonaggressive boys' selfresponsibility attribution (F = .604, p = .549). Also, there were no differences in their thoughts about how responsible they were or how much the victim should be blamed. These results are displayed in Table 23 and 24. For the third graders, an overall difference was found between aggressive and nonaggressive boys in selfresponsibility attribution (F = 3.544, p = .009). Univariate analyses show that the aggressive boys felt less responsible for the incident than nonaggressive boys (F = 7.153, p = .009) . No difference was found in their thoughts about how much the victim should be blamed (F = 2.037, p = 157) Victim responsibilit-y Data were collected from two questions to assess boys' beliefs about the victim's

PAGE 86

76 Table 23 and Nonaaaressive Bovs 0£ o in exi-KesTaonsxciiity rc Grade Six and Three >r Aaqressive Wilks' Hypothesis Error Source A F df S 6th grade .987 .604 2 94 .549 3rd grade .928 3 .544 2 92 .033 Table 24 Univariate F-Tests of Attri bution of Self-Responsibility for Aggressive and Nonaaares sive Bovs in Grade Six and Three Source Sum of Mean Squares df Square 6th grade Self-responsibility .327 1 .327 .134 716 Self blame 3.325 1 3.325 1.031 .313 3rd grade Self-responsibility 31.907 1 31.907 7.153 .009 Self blame 9.475 1 9.475 2.037 .157

PAGE 87

77 responsibility: How responsible was the victim cmd how much was the victim to blame for what happened? As displayed in Table 25, sixth-grade aggressive and nonaggressive boys were significantly different in their beliefs about the victims' responsibility (F = 3.436, p = .036). Univariate analyses revealed that sixth-grade aggressive boys were likely to attribute more responsibility to the victim than nonaggressive boys (F = 5.220, p = .025) . Aggressive boys also tended to think the victim should be blamed more than the nonaggressive boys did (F = 6.639, p = .012) (see Table 26) . There was no significant difference between thirdgrade aggressive and nonaggressive boys in attribution of victims' responsibility (F = .654, p = .522). Further analysis showed no differences in how responsible they felt the victim was (F = 1.161, p = .284) or how much the victim should be blamed (F = .215, p = .644) . Differe nces Between Sixth and Third Graders Storv Completion To examine if there was a difference between sixthand third-grade boys in their responses to the story completion, an ANOVA was performed. No significant difference was found between sixthand third-grade

PAGE 88

78 Table 25 MANOVA of Attributing Responsibility to Victim for Aggressive and Nonaaqressive Boys in Grade Six and Three Source Wilks' A I Hypothesis df Error df 6th grade .932 3.436 2 94 .036 3rd grade .986 .654 2 94 .522 Table 26 Univa riate F-Tests of Attributing Responsibility to Victim for Aggressive and Nona ggressive Bovs in Grade Six and Three Sum of Mean Source Squares df Square 6th grade Victim Responsibility 23.491 1 23.491 5.220 .025 Victim Blame 21.817 1 21.817 6.639 !oi2 3rd grade Victim Responsibility 6.333 1 6.333 1.161 .284 Victim Blame .540 1 .540 .215 .644

PAGE 89

aggressive boys in their story completion scores (F = 2.321, E = .131) , as shown in Table 27. A second ANOVA was conducted on the data from the nonaggressive boys . The analysis yielded a significant difference between the sixthand third-grade nonaggressive boys (F = 11.850, e = .001) . The sixth graders were more likely to report that they would be more motivated and would put more effort into compensating for their misbehavior than did the third graders . Reactive Feelings To determine whether the reactive feelings the aggressive boys in the sixth grade reported they would feel after misbehavior were significantly different from those reported by the third-grade aggressive boys, a MANOVA procedure was conducted (see Table 28) . An overall significant difference was found (F = 2.861, p = .014) . Results of univariate analyses showed that the aggressive boys in the third grade tended to report that they would feel higher levels of happiness (F = 4.751, p = .032) and pride (F = 8.658, p = .004) than the aggressive boys in the sixth grade. They did not differ in feelings of fear, guilt, shame, and sadness. The results are stumaarized in Table 29. For the nonaggressive boys, the results of the MANOVA also revealed a significant overall difference between the

PAGE 90

80 Table 27 ANOVA Results for Testing Age Differences in Story Completion Sum of Mean Source Scpiares df Square F p Aggressive 3.866 1 3.866 2.321 .131 Nonaggressive 31.154 1 31.154 11.850 .001

PAGE 91

81 Table 28 MANOVA Results for Testing Age Differences in Reactive Feelings Wilks' Hypothesis Error Source A F df df Aggressive .838 2.861 6 89 .014 Nonaggressive .868 2.304 6 91 .041 Table 29 Results of Univa riate F-Tests for Age Differences in Reactive Feelings Sum of Mean Source Squares df Square 1! B Aggressive Fear 3 .010 1 3 .010 .485 .488 Guilt .375 1 .375 .054 .816 Sad 24 .000 1 24 .000 2 .392 .125 Happy 23 .010 1 23 .010 4 .751 .032 Shame 29 260 1 29 .260 3 .115 .081 Proud 28 167 1 28 .167 8 .658 .004 Nonaggr e s s i ve Fear 12. 500 1 12 .500 2 .474 .119 Guilt 4. 939 1 4 .939 1 115 .294 Sad 0. 010 1 0 010 001 .971 Happy 6. 378 1 6 378 3 432 .067 Shame 62. 082 1 62. 082 7, 768 .006 Proud 4. 082 1 4. 082 2. 096 .151

PAGE 92

82 sixthand third-grade boys (F = 2.304, p = .041) . The univariate Ftests yielded a significant difference in the feeling of shame (F = 7.768, g = .006), but not in feelings of fear, guilt, sadness, happiness, or pride. The group means indicate that the older nonaggressive boys tended to report that they would feel more shame than did the younger nonaggressive boys after their own misbehavior. Emotion Regulation Cognitive copin g strategy . Fisher's exact test was used to examine the differences between the sixthand third-grade boys' responses to five categories of cognitive coping strategies (con^jaring to others' misbehavior, dehumanizing others, blaming others, thinking about compensation, and shifting attention) following transgression. Table 30 and 31 display the frequency of each strategy the boys in each group reported that they would use after wrongdoing. The results of Fisher's exact test are presented in Table 30 and 31. No significant difference was found between the sixthand third-grade aggressive boys for the three stories. Fisher's exact test revealed an age difference for the nonaggressive boys. As can be seen in Table 31, for the second (x' = 7.253, p = .027) and third (x'= 7.828, p = .050) stories, more sixth graders reported that they would

PAGE 93

83 Table 30 Frecmen cv and Fisher' s Exact Test for Age Differences in Cognitive Coping Strategy for Aggressive Boys Strategy 1 2 3 4 5 Story 1 6th grade 0 0 4 34 10 3.534 .316 3rd grade 3 0 4 33 7 Story 2 6th grade 0 0 6 31 9 3.181 .365 3rd grade 1 0 2 34 10 Story 3 6th grade 0 0 8 31 8 3.352 .501 3rd grade 2 1 7 31 6 Table 31 Frecfuency and Fisher's Ex act Test for Age Differences in Cognitive Copin g Strategy for Nonaggressive Bovs Strategy Story 1 6th grade 0 0 0 44 5 6.275 099 3rd grade 2 0 1 35 11 Story 2 6th grade 0 0 1 47 1 7.253 027 3rd grade 0 0 4 38 7 Story 3 6th grade 0 0 2 44 2 7.828 .050 3rd grade 1 0 7 34 7

PAGE 94

84 employ a strategy focusing on compensation than did the third graders. More third graders reported that they would apply strategies of justification or would shift their attention from the incidents , Intention to help and apologize to the v ar-t-im The MANOVA procedure did not yield a significant difference between the sixthand third-grade aggressive boys in their intention to help or apologize to the victim (F = .658, p = .520), as shown in Table 32. Also, the univariate tests did not indicate any significant differences between the two age groups (see Table 33) . For the nonaggressive boys, MANOVA results showed no overall difference between the sixth and third graders in their intention to apologize or help (F = .469, p = .627). The univariate analyses also showed no age differences among the nonaggressive boys . The means for intention to help and apologize for both group approached the highest possible score, 15. This information indicated that most boys in each group reported that they would help and apologize to the victims if they had hurt the victims. Intensity of disturbance. The results of the MANOVA indicated a significant difference between the sixthand third-grade aggressive boys in the overall intensity of

PAGE 95

85 Table 32 Apologize Source Wilks' A F Hypothesis df Error df S Aggressive .986 .658 2 92 .520 Nonaggressive .990 .469 2 95 .627 Table 33 Results of Univariate F-Tests for Age Differences in Intention to Help and Apologize Stun of Mean Source Squares df Square Aggressive Apology 1.195 1 1.195 .247 .621 Help 5.442 1 5.442 1.315 .254 Nonaggressive ;^ology 2.000 1 2.000 .944 .334 Help .163 1 .163 .071 .790

PAGE 96

86 their disturbance (F = 5.579, e = .002), as displayed in Table 34 . Univariate tests indicated that the two age groups differed in the likelihood of continuing to feel discomfort after helping (F = 14.171, p = .000) and apologizing (F = 5.964, E = .017) to the victim (see Table 35). The group means revealed that the sixth-grade aggressive boys reported that they were more likely to remain distressed after apologizing and helping than did the third-grade aggressive boys . Significant age differences in the intensity of disturbance were also presented among the nonaggressive boys (F = 8.235, E = .000) . The results of the univariate analyses and group means revealed that the older boys reported that they were more likely to continue feeling discomfort after apologizing (F = 16.436, p = -000) and helping (F = 14.959, p = .000) and also more likely to be disturbed in their daily life (F = 9.037, e = .003). Younger boys' responses to these questions were similar regardless of their aggressive status. A majority of younger boys responded that they would not feel discomfort after they had apologized to or helped the victim. Length of disturbance . A MANOVA was conducted to examine age differences in the length of disturbance due to misconduct. As can be seen in Table 36, a significant

PAGE 97

87 Table 34 MANOVA Results of Aae Differences in Intensity of Disturbance Wilks' Hypothesis Error Source A F df df Aggressive .837 5.579 3 86 .002 Nonaggressive .792 8.235 3 94 .000 Table 35 Results of Uni variate F-Tests for Aae Differences in Intensity of Disturbance Sum of Mean Source Squares df Square aggressive After apology 36.202 1 36.202 5.964 .017 After help 90.401 1 90.401 14.171 .000 Disturb life 47.193 1 47.193 3.828 .054 nonaggressive After apology 139.684 1 139.684 16.436 .000 After help 90.163 1 90.163 14.959 .000 Disturb life 92.092 1 92.092 9.037 .003

PAGE 98

88 Table 36 MANOVA Results of Age Differences in Length of Disturbance Wilks' Hypothesis Error Source A F df df g Aggressive .908 4.659 2 92 .012 Nonaggressive .815 10.813 2 95 .000 Table 37 Results of Univariate F-Tests for Age Differences in Length of Disturbance Svm of Mean Source Squares df Square Aggressive Days Disturbed 91.032 1 91.032 8.893 .004 Time Think 28.122 1 28.122 4.299 .041 Nonaggressive Days Disturbed 172.449 1 172.449 21.805 .000 Time Think 45.806 1 45.806 7.242 .008

PAGE 99

89 difference was found between the sixthand third-grade aggressive boys (F = 4.659, e = .012) . Results of univariate analyses and descriptive data indicated that the sixth graders reported that they were more likely to be bothered for two or more days (F = 8.893, p = .004) and to keep thinking about these incidents for a longer period of time (F = 4.299, E = .041) than did the aggressive boys in the third grade (see Table 37) . For nonaggressive boys, overall age differences were found (F = 10.813, e = .000) through MANOVA procedure. Univariate analyses and the group means revealed that compared to the third graders, the sixth graders were more likely to report that they would be disturbed by the events for two or more days (F = 21.805, p = .000) and to continue to think about these incidents (F = 7.242, e = .008). Empathic Reactions A MANOVA procedure was conducted to examine age differences in boys' empathic reactions to the victims of their wrongdoing (see Table 38) . No significant overall difference was found between the sixthand third-grade aggressive boys (F = .820, p = .486). The univariate F-tests yielded no difference in these two groups' estimate of their victims' upset and sadness or in their eo^athic emotions (see Table 39) .

PAGE 100

90 Table 38 MANOVA Results of Age Differences in Empathic Reactions Wilks' Hypothesis Error Source A F df df Aggressive .974 .820 3 91 .486 Nonaggressive .958 1,359 3 94 .260 Table 39 , ! . Result s of Univariate F-Tests for Age Differences in Empathic Reactions Source Stun of Squares df Mean Square F S Aggressive Victim Upset 5.473 1 5.473 1.157 .285 Victim Sad .037 1 .037 .005 .943 Empathy 1.218 1 1.218 .202 .654 Nonaggressive Victim Upset .010 1 .010 .002 .966 Victim Sad 2.000 1 2.000 .342 .560 Empathy 11.796 1 11.796 2.640 .108

PAGE 101

91 For nonaggressive boys, the result of the MANOVA did not show an overall significant difference between the two age groups (F = 1.359, p = .260). The univariate analyses revealed no age differences in nonaggressive boys' cognitive and affective empathic reactions . Outcome Expectations As displayed in Table 40, MANOVA results indicated that the sixthand third-grade aggressive boys were significantly different in their outcome expectations for aggression (F = 4.471, p = .014) . The age difference was found in friends' approval (F = 6.518, p = .012), with the sixth-grade boys reporting that they were more likely to believe that their friends would not like their wrongdoing than the third-grade boys (see Table 41) . For the nonaggressive boys, the MANOVA results revealed that there was no significant difference in the overall outcome expectations between nonaggressive boys in the sixth and third grades (F = 2.075, p = .131) . No significant differences were found in the univariate tests. Belief s about the Le gi t-ima cy of Aggression Age differences in the boys' beliefs about the legitimacy of aggression were tested by MANOVA procedure. The results did not show a significant difference between the sixthand third-grade aggressive boys (F = .613, p =

PAGE 102

92 Table 40 MANOVA Results of Age Differences in Outcome Expectations for Aggression Wilks' Hypothesis Error Source A F df df Aggressive .911 4.471 2 92 . .014 Nonaggressive . 958 2 . 075 2 95 . 131 Table 41 Results of Uni variate F-Tests for Age Differences in Outcome Expectations for Aggression Sum of Mean Source Squares df Square Aggressive Friend 8.809 1 8.809 6.518 .012 Object 1.371 1 1.371 .614 .435 Nonaggressive Friend 2.949 1 2.949 2.650 .107 Object .500 1 .500 .614 .435

PAGE 103

93 .544), as displayed in Table 42. Univariate F-tests yielded no significant differences in these boys' report of their enjoyment of what they would gain from transgression or in their beliefs about the legitimacy of their behavior (see Table 43) . The results of the MANOVA show that the sixthand third-grade nonaggressive boys did not differ in their beliefs about the legitimacy of aggression (F = 1.219, p = .300) . The univariate F-tests did not reveal any significant differences . Responsibility Attribution Self -respon g -i > >ili ty . To examine whether there was a difference between the sixth and third graders in attribution of responsibility to self, a MANOVA test was conducted. A significant difference was found between the sixthand third-grade aggressive boys (F = 4.170, p = .018), as shown in Table 44. Univariate F-tests indicated that the older aggressive boys would attribute more responsibility (F = 4.880, p = .030) and blame (F = 7.914, p = .006) to themselves for the incidents than would the younger aggressive boys (see Table 45) . For nonaggressive boys, an overall difference between the sixthand third-grade boys was found in attributing

PAGE 104

94 Table 42 MANOVA Results of Age Differences in Beliefs about the LegitimacY of Aggression Wilks' Hypothesis Error Source A F df df Aggressive .987 .613 2 93 .544 Nonaggressive .975 1.219 2 97 .300 Table 43 Results of Univariate F-Tests for Age Differences in Beliefs about the Legitimacy of Aggression Svm of Mean Source Squares df Square Aggressive Enjoy 5.042 1 5.042 .616 .435 Bad 4.594 1 4.594 1.160 .284 Nonaggr e s s i ve Enjoy 15.520 1 15.520 2.296 .133 .040 1 .040 .013 .908

PAGE 105

95 Table 44 MRNOVA Results of Age Differences in Attribution of SelfResponsibility Wilks' Hypothesis Error Source A F df df p Aggressive .917 4.170 2 92 .018 Nonaggressive .893 5.610 2 94 ,005 Table 45 Results of Univariate F-Tests for Age Differences in Attribution of Self -Responsibility Sum of Mean Source Squares df Square Aggressive Self-responsibility 21.468 1 21.468 4.880 .030 Self blame 32.692 1 32.692 7.914 .006 Nonaggressive Self -responsibility ,206 1 ,206 ,082 .775 Self blame 20.167 1 20,167 5,399 .022

PAGE 106

96 For nonaggressive boys, an overall difference between the sixthand third-grade boys was found in attributing responsibility to self (F = 5.610, p = .005) . Univariate analyses showed that the older boys reported that they would think they should be blamed more than did the younger boys (F = 5.399, p = .022) . Victim-responsibility . As can be seen in Table 46, the MANOVA results indicated that the sixthand third-grade aggressive boys differed in attributing responsibility to victims (F = 4.809, p = .010). The univariate tests revealed that the older aggressive boys reported that they would attribute more responsibility (F = 4.676, p = .033) and blame (F = 9.587, p = .003) to the victims than would the younger aggressive boys after wrongdoing (see Table 47) . For the nonaggressive boys, no overall age difference was found in attributing responsibility to victims (F = 2.260, p = .110). However, the univariate F-tests found a significant difference between the two age groups, with the older nonaggressive boys reporting that they would attribute more responsibility to the victims for these incidents than would the younger boys (F = 4.524, p = .036) . Summary In siimmary, boys in all groups reported that they would feel high levels of negative emotions and low levels of positive emotions, and they believed that aggression is

PAGE 107

97 Table 46 MANOVA Results of Age Differences in Attributing Responsibility to Vir->iiTi Wilks' Hypothesis Error Source A F df df Aggressive .906 4.809 2 93 .010 Nonaggressive .955 2.260 2 95 .110 Table 47 Result s of Univariate F-Tests for Age Differences in Attributing Responsibility to Victim Sum of Mean Source Squares df Square Aggressive Victim responsibility 33.844 1 33.844 4.676 .033 Victim blame 2.667 1 32.667 9.587 .003 Nonaggr e s s i ve Victim responsibility 12.500 1 12.500 4.524 .036 Victim blame 3.306 1 3.306 1.375 .244

PAGE 108

98 inappropriate. However, boys in different groups differed in the strength of their feelings and beliefs. Among the sixth graders, nonaggressive boys reported they would feel higher levels of sadness, guilt, shame, and empathy, have a stronger intention to compensate, be more intensely disturbed, expect less favorable outcomes, and attribute less responsibility to victims, when compared to aggressive boys. Among the third graders, nonaggressive boys reported they would expect less favorable consequences for transgressions and attribute more responsibility to self than did their aggressive peers. Age differences were also found. Younger aggressive boys reported that they would feel higher levels of happiness and pride than did the older boys. The older aggressive boys reported that they would expect more disapproval from friends due to aggression than would the younger aggressive boys. The older nonaggressive boys reported that they would feel higher levels of shame and a stronger intention to remedy their misbehavior than did the younger boys. Older boys said that they would be disturbed more intensively and for a longer period of time, and they would attribute more responsibility to self and victims, compared to the younger boys.

PAGE 109

CHAPTER 4 DISCUSSION AND RECOMMENDATIONS This chapter is organized into three major sections. In the first section I siimmarize the findings of this study. In the second I discuss these findings, and in the final section I propose suggestions for further study and practice . Overview of the Findings The aim of this study was to identify differences between aggressive and nonaggressive boys cuid between sixthand third-grade boys in emotional and cognitive reactions following transgressions . One hundred aggressive boys and one hundred nonaggressive boys, selected for participation on the basis of their teachers' rating of their aggressiveness, were interviewed individually. The boys read three semiprojective stories describing explicit misbehavior and then answered questions concerning how they would feel and think if they were the perpetrators of the misbehavior. The results of the data analysis indicated that regardless of their age or aggressiveness, the boys reported that they would feel high levels of negative emotions and 99

PAGE 110

100 low levels of positive emotions and they believed that aggression is inappropriate. However, they differed in the strength of their feelings, outcome expectations, and responsibility attributions. The nonaggressive boys in the sixth grade reported they would feel higher levels of sadness, guilt, shame, and empathy, had a stronger intention to compensate, were more intensely disturbed, expected a less favorable outcome, and attributed less responsibility to victims, when compared to aggressive boys in the sixth grade. Nonaggressive boys in the third grade reported that they would expect less favorable consequences for transgressions and attribute more responsibility to self, compared to their aggressive peers. Age differences were also found. Younger aggressive boys reported that they would feel higher levels of happiness and pride than did the older boys. The older aggressive boys reported that they would expect more disapproval from their friends because of their aggression than did the younger aggressive boys. The older nonaggressive boys reported that they would feel higher levels of shame and a stronger intention to remedy their misbehavior than did the younger boys. The older aggressive and nonaggressive boys both reported that they would be more intensely disturbed and for a longer period and they would attribute more responsibility to self and victims, compared to the younger boys.

PAGE 111

101 Discussion of Results Research Question I The first research question was whether the reactive emotions following transgression are the same for the aggressive and nonaggressive boys. It was found that boys in each group experienced high levels of negative feelings (guilt, shame, sadness, and fear) and low levels of positive emotions (happiness and pride) in this study. However, the sixth-grade nonaggressive boys reported relatively higher levels of sadness, shame, and guilt than their aggressive peers. Thus, in terms of reactive feelings, transgression seemed to be a more negative experience for the nonaggressive than for aggressive sixth graders. Of the six feelings explored in the present study, the older aggressive and nonaggressive boys showed the greatest difference in sadness. Researchers have emphasized the important role that sadness and fear play in younger children's sociomoral development. According to Kochanska (1993) , proneness to distress and certain levels of fear after wrongdoing could facilitate the development of conscience and guilt. In addition, Rothbart et al . (1994) found that sadness is positively correlated with empathy and guilt in 6to 7 -year-old children. Cole, Barrett, and ZahnWaxler (1992) found a relationship between young children's

PAGE 112

102 sadness and reparative behavior during their mishaps . The higher levels of sadness found in the nonaggressive boys in this study are consistent with these studies in demonstrating that sadness is an important emotion in facilitating sociomoral development. In contrast, Rothbart et al. (1994) suggested that sadness and fear are correlated with prosocial traits but not aggression. Further study is needed to clarify (a) whether sadness and fear only relate to prosocial behavior or to both aggressive euid prosocial traits and (b) how sadness and fear influence prosocial/aggressive behavior. Suilt has been conceived as an important moral affect. Previous studies have found a negative correlation between guilt and aggression (Rothbart et al., 1994; Su, 1975; Tangney, 1991; Tangney et al . , 1992). In this study, the higher level of guilt reported by older nonaggressive children, relative to aggressive peers, is consistent with these findings. It is reasonable to expect a negative relation between guilt and aggression, as guilt elicits remorse, distress, and self -sanction and consequently inhibits the occurrence of aggression. Recently, Fraser (1996) proposed that the relationship between guilt and aggression differs for different age and gender groups. In her study, guilt and aggression correlated

PAGE 113

103 negatively for secondand third-grade boys but were positively correlated for fifthand sixth-grade boys. Fraser suggested that social learning, especially outcome expectation, may influence the effectiveness of guilt in controlling aggression in older boys . Thus , although the inhibitory effects of guilt have been found consistently in previous research as well as in this study, whether there ^ are factors that mediate the relationship between guilt and aggression merits further research. Research findings on the relationship between shame and aggression are inconclusive. Tangney and her colleagues (Tangney, 1991; Tangney et al., 1991, 1992, 1996) reported positive correlations between shame and aggression in children, whereas Fraser (1996) found a negative correlation among boys in late childhood. Findings of the present research that the sixth-grade nonaggressive boys would experience a higher degree of shame than their aggressive peers, conflict with findings from Tangney et al. but are consistent with the results of Fraser' s study. Researchers have argued that shame helps to regulate sociomoral behavior (Ferguson & Stegge, 1995; Scheff, 1988). Ferguson and Stegge proposed that feelings of shame alert a ' person to rejection by the surrounding environment. They stated that "shame regulates daily interaction by inhibiting j

PAGE 114

104 arrogance, promoting hvimility, and fostering conformity or deference to standards of conduct valued by the group" (p. 181) . However, other researchers have maintained that shame can elicit humiliation and then evoke anger or hostility and consequently cause aggressive behavior (Lewis, 1992; Scheff, 1987; Tangney et al., 1996, 1992). Researchers (Cole et al., 1992; Fraser, 1996) have suggested that a moderate level of guilt may be most helpful for inhibiting aggression. As shame is a self -evaluative emotion like guilt, perhaps shame also regulates aggression most effectively at a moderate level . From these contrasting perspectives several questions arise: In what situations would shame be most likely to elicit anger and hostility? Do different levels of shame function differently in children' s behavior? Are there optimal levels of shame that facilitate self-regulation in social behavior rather than elicit anger and hostility? Further work on these questions seems crucial for advancing the understanding of shame as an inhibitor of aggression. In addition to inducing aggression through anger and hostility, shame has been found to be associated with a variety of psychological consequences that may interfere with sociomoral development. Researchers have indicated that shame proneness is correlated with lower self-esteem

PAGE 115

105 (Tangney, Burggraf , & Wagner, 1995) , conformity to norms exterior to self (Scheff , 1988) , lower empathy, and use of selfjustification . For Chinese boys shame may be effective in controlling transgression. From the findings of this study, several questions arise: Why does shame seem to facilitate rather than hinder self-regulation and the inhibition of norm violation in Taiwanese boys? Are the effects of shame the same on children in Western culture and in Asian cultures? What are the long-term effects of shame on people's behavior? Does shame cause short-term compliance at the cost of long-term maladjustment (Abell & Gecas, 1997)? Research Question II The second question addressed the differences between aggressive and nonaggressive boys in the intensity of disturbance. The nonaggressive boys in the sixth grade reported that they were more likely than aggressive boys to be bothered in daily life and to continue to feel discomfort even after apologizing. The results support the contention that emotional sensitivity to wrongdoing is important for the development of sociomoral and self-evaluative functioning (Cole et al . , 1992). One possible reason that nonaggressive boys were more likely to continue to feel discomfort after apologizing may

PAGE 116

106 be that they viewed the harm they caused another person as serious and believed that an apology was not sufficient to repair the harm. Another possible reason is that they may have thought an apology does not repair the harm to the victim. Aggressive boys might not continue to feel bad about their behavior because they had done something to condensate the victim. Research Question III The third question was whether aggressive and nonaggressive boys differ in the length of their discomfort. No differences were found between the two groups. Why the sixth-grade nonaggressive boys believed that they would not be disturbed longer, given that they felt significantly higher levels of negative feelings, is not clear. Data were unavailable to determine whether the nonaggressive boys would have repaired or dealt with their misconduct sooner than the aggressive boys, whether boys in both groups would shift their attention in the same length of time, or whether it was difficult for children to estimate the length of time they would feel discomfort. Research Question IV This question explored whether aggressive and nonaggressive boys differ in their empathic responses to their wrongdoing. Researchers have found a negative

PAGE 117

107 relationship between empathy and aggressive behavior (Ellis, 1982; Miller & Eisenberg, 1988; Richardson et al., 1994) and reduction of aggression through empathy training (Feshbach, 1983) . The sixth-grade nonaggressive boys reported higher empathic feelings than their aggressive peers, in this study. This result is consistent with previous findings. Researchers have proposed several mecheuiisms through which empathy may result in less aggression in en^athic children than less-empathic children. Feshbach (1997) postulated that empathic children are more likely to examine a conflict situation from the perspective of another person. This should result in a decrease in misunderstanding and lessen conflict and aggression. Furthermore, when observing pain and distress in a victim of aggression, an empathic observer may experience vicarious distress, even if the observer causes the aggression. Thus, the painful consequence of an aggressive act should function as an inhibitor of the perpetrator's aggressive tendencies. According to Tangney (1991) , both empathy and guilt are other-oriented emotions and are positively correlated with each other. When children harm others, an empathic response may be evoked, accompanied by feelings of guilt. The distress of guilt may reduce the occurrence of aggression. Therefore, empathy should regulate aggressive behavior.

PAGE 118

108 In this study, aggressive and nonaggressive boys in the third grade did not show differences in en5>athic responses to the victim of their misbehavior. In a review article, Eisenberg and Miller (1987) indicated that the positive correlation between empathy and prosocial behavior is weaker for younger children than for older children. Empathy may not be stable or mature in younger children. Research Question V This question addressed whether aggressive and nonaggressive boys differ in their attributions of responsibility for a transgression. Results of this study revealed that aggressive boys in the sixth grade would attribute more responsibility and blame to the victims than would their nonaggressive peers; among the third graders, aggressive boys reported that they would attribute less responsibility to themselves than did the nonaggressive boys . In previous research, Bandura et al.(1996) proposed that diffusing responsibility is a moral disengagement mechanism, which could foster harmful conduct. Other researchers (Graham et al . , 1984, 1995; McGraw, 1987) have found that after transgression if individuals attribute responsibility to others they are less likely to feel guilt. Wang (1983) indicated that the essential factor for the

PAGE 119

development of delinquent gangs in Taiwan was a selfdefensive mechanism against feelings of guilt and fear. Thus, results of this study are consistent with previous findings in supporting the argument that aggressive boys are more likely to use a moral disengagement mechanism to reduce self -censure than are nonaggressive peers . However, different research findings have been reported. Graham et al. (1995) found that relative to aggressive boys, nonaggressive boys were more likely to hide the true causes of transgressions for which they are responsible, and this difference was greater eunong older children. According to Graheun, aggressive boys may not be aware of how to disguise the truth, or they may not care to mask the truth. Thus, more studies are needed to confirm the relation between attribution of responsibility and aggression and to determine how these attributions of responsibility influence aggression. In this study, the third-grade aggressive boys attributed less responsibility to themselves than did nonaggressive boys. At an early age, children learn that their personal responsibility for a negative event may result in their punishment (Caprara, Pastorelli, & Weiner, 1997) . Children in the third grade possess adequate knowledge about the relationship of responsibility and its

PAGE 120

110 consequences. Graham et al . (1995) suggested that childhood aggression is linked to harsh discipline. Thus, if harsh discipline is typical in the families of aggressive children, young aggressive boys may attribute less responsibility to self in order to avoid punishment. Research Question VI This question explored whether aggressive boys are more likely to hold a positive outcome expectancy for transgression than nonaggressive boys. Results revealed that aggressive boys, in both age groups, were more likely to believe that transgression could help them gain what they want. These findings are consistent with results of other studies (Cuddy & Frame, 1991; Perry et al., 1986; Slaby & Guerra, 1988; Xie, 1991; Zhang, 1990). The aggressive boys appeared to enjoy the immediate and direct rewards from aggression and not to consider the possible long-term negative consequences. Perry et al . (1980) found that children who had been attributed with prosocial traits failed to anticipate negative outcomes but employed more self -punishment after misbehavior than did children not attributed with prosocial traits. Thus, the positive results of wrongdoing seem to be more salient or attractive for some children. Knowledge of why aggressive boys are more likely to be reinforced by the immediate

PAGE 121

Ill rewards of wrongdoing but nonaggressive boys are not and how children develop expectancy beliefs would be very helpful for designing intervention programs. On the other hand, given the high levels of negative emotions reported by the aggressive boys in this study, it is reasonable to wonder whether the outcome expectancy for rewards exerted more influence on aggressive boys than did the negative emotions . Research Question VII The seventh question was whether aggressive boys are more likely to believe that aggression is a legitimate response than are nonaggressive boys . No significant differences were found between the aggressive and nonaggressive boys in either sixth or third grades. In previous studies, it has been found that children approving more of aggression tend to behave more aggressively (Eron et al., 1997; Guerra & Slaby, 1990; Huesmann & Guerra, 1997; Slaby & Guerra, 1988; Zhang, 1990) . Results of this study conflict with previous findings. Most of the children in this study, regardless of their aggressiveness, indicated that the protagonists' behaviors were inappropriate. One possible reason for this result is that Asian cultures emphasize conforming to group rules in discipline training. In Taiwan, behavior standards are taught repeatedly in daily life by parents and in the classroom as a stibject (health

PAGE 122

112 and morality) by teachers. Therefore, children in Taiwan may have learned that cheating, stealing, and hurting others (the incidents in the three stories) are not acceptable from an early age. Research Question VIII This question addressed age differences in aggressive and nonaggressive children' s emotional and cognitive reactions to their own wrongdoing. Among the affective reactions, four differences were found. First, the younger aggressive boys reported that they would feel higher levels of happiness and pride them the older boys after wrongdoing. Piaget (1965) claimed that children aged 5 to 10 tend to use objective consequences as criteria to judge the rightness of behavior. Similarly, Harris (1989) and Ferguson and Stegge (1995) indicated that children younger than about 8 -years -old do not apply social norms to infer reactive feelings of misbehavior but focus on outcome of actions. Besides using consequences to evaluate their behavior, younger children find the positive results of their behavior to be more salient and attractive than the negative outcomes (Cuddy & Frame, 1991; Perry et al . , 1980). They focus especially on the gains and successes of their conduct. For example, Nunner -Winkler and Sodian (1988) found

PAGE 123

113 that children younger than 8 years ascribed positive feelings to story characters who had treuisgressed. Another age difference in emotional reactions was that the older nonaggressive boys reported that they would feel a higher level of shame than did the younger boys . Findings of this study are consistent with Eraser's (1996) results. Eraser found that older boys, aged 11 to 12, reported a higher level of shame than did younger boys, aged 8 to 9. According to Fraser, pre-adolescents are particularly sensitive to the evaluations of others. They tend to take evaluations from others as an estimate of their worth as people. This tendency to focus on self rather than on actions is typical of the experience of shame (Tangney, 1990, 1991) . Some researchers have found that young children, around 8 years old, responded with shame after wrongdoing only in the presence of an audience (Ferguson & Stegge, 1995/ Harris, 1989). In the stories used in the present study no witness appeared. This may be another reason why younger boys reported that they would feel less shame than the older nonaggressive boys did. Social-cultural factors provide an alternative ea^lanation. Socialization experiences are considered crucial to the development of shame (Abell & Gecas, 1997;

PAGE 124

114 Barrett, 1995; Ferguson & Stegge, 1995). Older Chinese children receive more shame discipline from parents and teachers than do younger boys; thus, the older boys may be more susceptible to feelings of shame after wrongdoing. More experiences with shame may lead the older nonaggressive boys to conform more to adults' expectations and social norms. The third age difference was that older boys reported that they would be more intensely disturbed and that their discomfort would last longer than the younger boys did. With increasing age, the connection between social norms and emotion becomes stronger (Ferguson & Stegge, 1995; Harris, 1989); therefore, when violating a norm, older children may react with more intense feelings and these feelings may last longer. Children in late childhood are able to experience the same degree of guilt and shame reported by adults (Ferguson & Stegge, 1995) . In addition, one feature of younger children's cognition is focusing their attention on the here and now. Negative emotions due to their wrongdoing may occupy young children's thoughts for a shorter time than for older children . The fourth age difference in emotional reactions was that nonaggressive sixth graders reported higher motivation to compensate the victim for the wrongdoing than did the

PAGE 125

115 younger boys. Following wrongdoing, most people respond with guilt. With guilt, a person actively seeks to control or reduce the negative consequences of the action. Reparation is one way of exerting this control. Therefore, to remedy one's misconduct is an action tendency of guilt. The sixthgrade nonaggressive boys reported that they would experience higher levels of guilt, (although this difference was not statistically significant) , and would attribute more responsibility to self, which zoight account for their stronger intention to compensate the victim. Graham (1988) indicated that when children get older the link between emotion and intended action becomes stronger. Other researchers (Ferguson & Stegge, 1995; Harris, 1989) have suggested that older children understand that making amends helps mitigate guilt, but children under age 8 have difficulty grasping this idea. The stronger intention to remedy misconduct in older children is consistent with these developmental trends. Age differences were also found in children' s reports of their cognitive reactions to their own transgressions. One difference was that sixth-grade aggressive boys believed that their friends would disapprove of their misbehavior more than the younger aggressive boys did. The reason for the difference may be that sixth graders are more concerned

PAGE 126

116 about their friends' approval. According to Berndt (1979), conformity to peers reaches a higher level in sixth graders than in third graders . Younger children tend to be more self -centered . An alternative explanation is that sixth graders are able to integrate more social learning experiences, which is considered to be one of the origins of outcome expectations (Perry, Perry, & Weiss, 1989) . Older aggressive boys may have more experience in being rejected by peers and/or receive more negative consequences due to aggression. These experiences combined with other social learning experiences lead them to hold a more realistic expectation for aggression than younger aggressive boys. Another age difference in children' s cognitive reactions was that the older aggressive boys reported that they would attribute more responsibility and blame to both self and victims than younger aggressive boys, whereas the older nonaggressive boys reported that they would assign more blame to self than did the younger nonaggressive boys. Burger (1981) suggested that people's self -protective motives influence responsibility attributions. Research on children' s coping behavior has shown that younger children use an avoidance strategy frequently, whereas older children are more capable of admitting problems (Aldwin, 1994;

PAGE 127

117 Altshuler, 1989) . This could be an alternative explanation for why the younger boys would attribute less responsibility or blame to self than the older boys. According to Wang (1983) , the key factor in adolescent gang crime is that these adolescents use self-defense mechanisms to reduce the distress from guilt and fear. As proposed by Bandura et al. (1996), diffusing and displacing responsibility are two kinds of moral disengagement mechanisms . The older aggressive boys may have better self defensive skills to protect themselves; thus, they would attribute more responsibility and blame to others to protect themselves from painful self -sanction . Recommendations for Further Research Researchers disagree on the relationship between shame and aggression. Further research is needed to answer the following questions : In what situations does shame inhibit aggression, and what are the long-term effects of discipline based on shame . Longitudinal research is of crucial iii^ortance, because it would allow researchers to confirm whether children with different reactions to transgressions develop different aggressive tendencies later. It may advance our understanding of what developmental changes account for the increasing differences between aggressive and nonaggressive

PAGE 128

boys in their reactions to transgressions from age 8 to 11 . In addition, this type of study may help us understand how children develop guilt, shame, and empathy. Outcome expectations for aggression seem to be an important factor for differentiating between aggressive and nonaggressive boys . To guide the development of interventions to reduce aggression, further research is needed to determine how outcome expectations develop, why some boys hold more favorable expectations for aggression than others, and whether the positive results of transgression are of particular salience for younger boys. The behavior regulating function of negative emotions develops from the feelings of distress that these emotions arouse. If children get rid of these painful feelings quickly, their effects on inhibiting aggression would be diminished. This study only explored children's attribution of responsibility; other coping strategies and moral disengagement mechanisms children use to reduce self-censurc are not clear from the present study. Further studies are needed to address this issue. Impulsiveness has been found to be positively related to aggression in previous studies (Atkins et al . , 1993; Luengo et al . , 1994; Martin et al . , 1994). However, impulsiveness was not correlated with most of the variables

PAGE 129

119 examined in this study. Perhaps impulsiveness moderates the relationship between guilt or shame and aggression. This possibility should be explored in future research. Additional attention should be directed to cultural differences in future research. Although previous research has indicated that emotions of guilt and shsune are elicited by similar events for adults in Western and Asian cultures, this question has not been examined in children. Further research is needed to determine whether meanings of guilt and shame are the same for Western and Asian children and whether the rate of development of moral emotions and cognition is the same for children in both cultures. Recommendations for Practice Findings of this study lead to the following recommendations for how to treat children's transgressions. The focus is on preventing children from developing positive outcome eaq>ectations for aggression, helping children know what behavior is inappropriate, and on in^roving their social problem solving skills. 1. After a child's transgression, adults should show their disapproval clearly enough for the child to understand the behaviors that are not acceptable. Help the child keep in mind that any behavior that hurts another person is unacceptable .

PAGE 130

120 2 . Adults should discuss with children the possible damage caused by children's goals and discuss appropriate ways for children to reach their goals . By doing this , children may be more alert to the possible harm to other people when they act out. 3. To the extent of children's ability, adults should have children make reparations for their tremsgressions . This may teach them the negative consequences for themselves and encourage them to take responsibility after wrongdoing. 4. Adults should discuss others' feelings with children. Understanding and attention to others' feelings may foster the development of empathy as well as guilt.

PAGE 131

APPENDIX A TEACHER CHECKLIST 1 . He threatens or bullies others in order to get his own way. 2 . He uses physical force in order to dominate other kids . 3 . He starts fights with peers . 4. He overreacts to accidental hurts with anger and fighting. 5. He gets angry easily and strikes back when threatened or teased. 6 . He gets other kids to gang up on a peer . 7 . He says mean things to peers . 8. He claims other children are to blame in a fight and feels they started the trouble. Peer Nomination 1 . Who starts fight most often? 2 . Who works well with others? 3. Who disrupts the group? 4. Who is helpful to other students? 5 . Who gets angry easily? 121

PAGE 132

APPENDIX B REVISED STORIES 1 . Andy and Robert are good friends . They play checkers together at school very often. One rainy day, during recess, they play checkers again. When Andy seems to be winning the game, another boy comes over and asks Andy to lend him an eraser. When Andy leaves to get the eraser, Robert notes that nobody is watching him. He moves two checkers quickly. Andy does not see the change and he just continues to go along with the game when he returns. Finally, Robert wins the game . 2. David just gets a toy car on his birthday. After school the next day, Bob is invited to David's house to see the toy car. This car can be steered by remote and has real headlights and a siren. It looks really neat. Bob asks David to let him play with it, but David refuses. Then, David begins to watch TV. He is totally attracted by a TV cartoon so that he does not pay attention to Bob. Bob takes this opportunity to put the car in his bookbag. After a little while. Bob goes home with his bookbag and the toy car. 3 . Mark is hurrying to a movie one evening . Six-yearold Peter, living in Mark's neighborhood, is riding a 122

PAGE 133

123 bicycle around a corner. It is too late when Mark sees Peter. He crashes into Peter and makes a big noise. Peter begins to cry. Mark knows that no others are there, and he leaves quickly to hurry to his movie.

PAGE 134

APPENDIX C INTERVIEW QUESTIONS WITH REVISIONS Questions enclosed in parentheses were used in the pilot study. Story con^letion: 1. What will happen next? Feelings : 1. How would you feel if you did what XXX did? 2. How guilty would you feel if you did what XXX did? Why? 3. How much shame would you feel if you did what XXX did? Why? 4. How much fear would you feel if you did what XXX did? Why? 5. How happy would you feel if you did what XXX did? Why? 6. How sad would you feel if you did what XXX did? Why? 7. How proud would you feel if you did what XXX did? Why? Regulation : 1. What would you think to feel better after you did what XXX did? I would think that other boys do this very often and I just did the same thing that others do. I would think that XXX is not a good guy and he deserves to be treated like this. I would think that I did nothing wrong, XXX should be more careful when he plays chess (after he plays toy car, when he rides bicycle) . 124

PAGE 135

125 I would think that I should do something to make up for what I did . I would not think about this. 2 . Would you apologize to XXX? After apologizing would you still feel discomfort? (Would you apologize to XXX? After apologizing would you still feel discomfort or not at all?) 3. (Would you avoid XXX after this incident? When you avoid XXX do you still feel discomfort or not at all?) This question was not included in the revised questionnaire . 4. Would you help xxx after this incident? After helping XXX, would you still feel discomfort? (Would you help XXX after this incident? After help XXX, would you get rid of all of your discomfort feeling?) 5. How long would you keep thinking about this incident? 6. Would this incident stay in your mind and disturb you for a couple of days or more days? 7. Would this incident disturb your study, eating, or sleeping? Empathy : 1. How upset do you think XXX would be? 2. How sad do you think XXX would be? 3. How bad would you feel if you saw XXX was hurt? Outcome expectancy: 1. What do you think your friends would think if you did this? Very sure they would like what I did. Pretty sure they would like what I did. Pretty sure they would not like what I did. Very sure they would not like what I did.

PAGE 136

126 2 . Do you think doing this would help you get something you like? Very sure I could get things I like by doing this . Pretty sure I could get things I like by doing this . Pretty sure I could not get things I like by doing this. Very sure I could not get things I like by doing this. Legitimacy: 1. How happy would you feel with your win? (Story 1) How happy would you feel with the toy? (Story 2) Would you enjoy the movie very much? (Story 3) 2. How bad do you think it was to do this? Responsibility attribution: 1. How responsible would you feel for what happened? 2. How much would you be to blame for what happened? 3. How responsible was XXX for what happened? 4. How much would XXX be to blame for what happened?

PAGE 137

APPENDIX D CONSENT FORMS PARENT CONSENT FORM Dear Parent: My name is Ms. Shu-Tai Liao, a graduate student in Educational Psychology at the University of Florida in the United States. I am conducting a research project under the supervision of Dr. Patricia Ashton. The purpose of this study is to understand boys' reactions to their own misbehavior. The results of this study may help teachers and parents better understand how to deal with boys' misbehavior to reduce the reoccurrence of the behaviors. Your child's class has been selected to be involved in this study. I am writing to ask your permission for your child's participation. If your child participates in this study, his aggressive behaviors will be rated by his teacher. If his scores meet the criteria for selecting aggressive or nonaggressive children used in this study, he will be asked to take an intelligence test, answer a questionnaire to measure his impulsiveness, and he will be interviewed. The test and the questionnaire will take about 25 minutes and the interview will take approximately 35 minutes. In the interview, your child and I will read three stories, and then your child will be asked what he would feel and think if he had done what the persons in the stories did. With your permission I would like to audiotape this interview. Only I will have access to the tape, which I will personally transcribe. Upon completion of this study, the tape will be erased. Your child will be asked to give his name on the tests and interview for matching purposes. I will replace his name with a code number. Your child's identity and all data obtained in this study will be kept confidential to the extent provided for by law. Participation or nonparticipation in this study will not affect your child's grades or placement in any programs. There are no known risks or immediate benefits to the participants. No compensation is offered for participation. You and your child may withdraw consent for your child's participation at any time, without consequence. Your child will not have to answer any questions that he does not wish to answer. Any questions or concerns regarding your child's rights as a participant should be directed to the UFIRB office at Box 1 12250, University of Florida, Gainesville, FL 3261 1-2250, U. S. A. For any inquiries concerning the procedures, please contact me at my Taiwan telephone number (04)321-1460. The results will be available in September. If you are 127

PAGE 138

128 interested in the results, please call me then at this telephone number, and I will mail you the results and implications of this study. AUTHORIZATION: I have read the procedure described above and I have received a copy of this description. I agree to allow my child, , to participate in Ms. Shu-Tai Liao's study of the emotional and cognitive reactions to transgressions of aggressive and nonaggressive boys. Parent/Guardian Date 2nd Parent/Guardian Researcher's Signature Date Date

PAGE 139

129 TEACHER CONSENT FORM Dear Teacher: My name is Shu-Tai Liao. I am a graduate student in Educational Psycliology at the University of Florida in the United States. As part of my dissertation research, I am invhing you to participate in a study focusing on understanding the reactions of boys to their own misbehavior. If you agree to participate in this study, you will be asked to complete the items on the aggression subscale of the Teacher Checklist for each boy who volimtarily participates in this study in your class. It will take approximately 40 minutes to complete these checklists. Your ratings will be kept confidential to the extent provided by law. Only I will have access to your ratings, and I will replace your name with a code number. The parents will be told that you will provide me with ratings of their child's aggressive behaviors and the ratings of the other children in your class who participate in the study, if they agree to permit their child to participate in the study. There are no anticipated risks, compensation or other direct benefits to you as a participant in this study. You are free to withdraw your consent to participate at any time without consequence. If you have any questions regarding participation in this study, please don't hesitate to contact me at my Taiwan telephone number (04) 321-1460. You may also contact my supervisor, Dr. Patricia Ashton, Professor at University of Florida, at (352)392-0724 Ext. 243. Questions or concerns about research participants' rights may be directed to the UFIRB office. University of Florida, Box 1 12250, Gainesville, FL 3261 1, U. S. A., ph (352) 392-0433. Very sincerely, Shu-Tai Liao AUTHORIZATION: I have read the procedure described above. I have received a copy of this description and hereby agree to participate in Shu-Tai Liao's study of the emotional and cognitive reactions to transgressions of aggressive and nonaggressive boys. Teacher's Signature Date

PAGE 140

130 , : V pi , VERBAL ASSENT SCRIPT My name is Ms. Liao, and I am a graduate student at the University of Florida in the United States. I am inviting you to take part in my study. If you join this study, you will have a chance to be interviewed. In the interview, you and I will read three stories together. Then I will ask you questions about what you would feel and think if you had done what the persons in the stories did. If you join the study, you can stop at any time and you do not have to answer any questions that you do not want to answer. Your grades will not be affected if you answer my questions or you decide not to answer them. Are you willing to join my study?

PAGE 141

APPENDIX E INTERCORRELATION MATRIX Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Aggression (1) .19** -.35** .20** .12 .21** .17* -.20** .21** -.17* I.Q. (2) -.09 .37** .23** .24** -.07 -.25** .31** -.39** Impulsiveness (3) -.15* .05 -.07 -.02 .11 -.04 .07 Completion (4) .18* .21** .07 -.20** .31** -.13 Fear (5) .61** .62** -.33** .62** -.28** Guilt (6) .54** -.19* .57** -.24** Sad (7) -.24** .57** -.13 Happy (8) -.42** .57** Shame (9) -.37** Proud (10) ^ology (11) After apology (12) Help (13) After help (14) Disturb life (15) Disturb two days (16) Disturbed time (17) Victim upset (18) Victim sad (19) Empathy feel (20) Friend (21) Gain (22) Enjoy (23) Bad (24) Self responsibility (25) Self blame (26) Victim responsibility (27) Victim blame (28) 131

PAGE 142

132 Variable (11) (12) (13) (14) (15) (16) (17) (18) (19) Aggression ( 1 ) .20** .14 .19** .05 .14* .06 .07 .05 .13 I.Q. (2) -.01 .21** -.08 .36** .16* .30** .27** .06 .11 Impulsiveness (3) -.03 -.08 -.02 -.08 -.04 -.09 -.01 -.05 .02 Completion ( 4 ) . 15 . 18* .14 .27** .13 .18* .16* -.06 -.01 Fear (5) .20** .22** .19** .30** .35** .39** .51** .35** .43** Guilt (6) .29** .14 .37** .23** .33** .33** .47** .31** .36** Sad (7) .23** . 14 .35** .12 .26** .27** .36** .29** .41** Happy (8) -.32** -.16* -.12 -.12 -.27** -.39** -.35** -.07 -.13 Shame (9) .25** .20** .30** .29** .31** .35** .49** .33** .40** Proud (10) -.20** -.17** -.02 -.14 -.09 -.27** -.28** -.06 n7 ^ology (11) -.04 .36** -.16* .13 .20** .18* .01 .06 . Uo . U / . 11 Help (13) -.10 .28** .24** .14 .12 .13 After help (14) .16* .21** .31** .04 .13 Disturb life (15) .61** .58** .14* .20** Disturb two days(l) .55** .14 .22** Disturbed time (17) .39** .41** Victim upset (18) .71** Victim sad (19) Empathy feel (20) Friend (21) Gain (22) Enjoy (23) Bad (24) Self responsibility (25) Self blame (26) Victim responsibility (27) Victim blame (28)

PAGE 143

133 Variable (20) (21) (22) (23) (24) (25) (26) (27) (28) . 20** > . 26** 19 . x^ 1 «5* . Xw . 17* . 13 . 16* . 16* 1 7* . X / " 1 ft* 9n** 9n** . zu ^9** j^ 09 OS — nft . XX 1 9 . xz 01 . ux nn . uo ni . ux 9ft** 1 R* . xO" n9 . uz — 9^** 1 Q** 90** . Zu " " 9^** 0*3 07 Fear ( 5 ) Rft** 99** 1 1 . XX — 97** AQit* ft1 ** . ox 0*3 OR All i 1 1( ft ) 9ft** . ZD" t\A * * >s9* * AO** — O^** — OQ . uy 1 Q** 99 * * — Q9* * A Q* * A <;** . uo — 11 — . XX nappy ^ o / — 1 ft* . X — 9 ts* * R1 ** . OX^ " — ^9** — 9A ** . XX 1 (\* . xo^ 66** . J J" * 1 Q* . XO" ClA * * >^^* * RQ** . OO^ ^ — 1 o . xu — 1 A — . X4 — . * — 99* * — 1 O* * . " — , O il 4 4 . ZD** OQ** . zo** Ann T octi7 / 1 1 \ '39** "30* * — 1 Q* . XC 9Q* * . Zy " " 9^** 1 A . X4 _ 1 ^ — . X J _ 1 A . X4 . X4 n Q . U 7 — 99 * * . XX . XX 1 R* . XO* . UX no . UZ O Air it 1 T* . X / ^ O C 4 4 O Q 4 4 n c . Uo n ft . uy After helnflt . \JO — n7 _ on* * 91 ** rti . UX n c . Ub Disturb lifel5\ '^ft** 1 Q** HQ . UO — 91 ** . ZX" " 1 O* * . xy ^ ^ on* * O il 4 4 . J4** 1 A . J* ^^9ii>UXU AjfS \XO/ . JO" " 9fl* * . U J . zy** . Z /** . Uo . Oo '31 ** . JX " " 1 ^-k — 91 * * . ZX^ ^ . 40** . jy** . U J A A . U4 . 07 . 40** . 35** . 50** . 08 . 02 Victim sad/l) R7** 99** 1 "7* . X / " ~ . Uo .51** . 02 . 00 r ^1 1 oa\ liiHipatny reeizu) . 34** .20** -.27** . 61** .55** . 63** -.05 -.11 Friend (21) .26** -.27** .48** .33** .34** -.07 -.13 Gain (22) -.27** .27** .21** .12 -.23** -.19** Enjoy (23) -.44** -.40** -.27** .11 .14 Bad (24) ,60** .55** -.08 -.11 Self responsbility (5) .62** -.05 -.12 blame (26) -.03 -.06 Victim resposibilit(27) .74** Victim blame (28) *. Correlation is significant at the 0.05 level. **. Correlation is significcmt at the 0.01 level.

PAGE 144

REFERENCES Abell, E., & Gecas, V, (1997). Guilt, shame, and family socialization. Journal of Family Issues , 18 . 99-123. Aldwin, C. M. (1994) . Stress, coping, and development: An integrative perspective. New York: Guilford Press. Altshuler, J. A. , & Ruble, D. N. (1989). Developmental changes in children' s awareness of strategies for coping with uncontrollable stress. Child Development. 60. 13371349. Aronson, E. (1995). The social animal (7th ed.). New York : Freeman . Aronson, E., Wilson, T., & Akert, R, (1994). Social psychology: The heart and the mind. New York : HarperCollins . Atkins, M. S., Stoff, D. M. , Osborne, M. L. , & Brown, K. (1993) . Distinguishing instrumental and hostile aggression: Does it make a difference? Journal of Abnormal Child Psvchology. 21. 355-365. Bandura, A. (1973) . Aggression: A social learning analysis . Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1991) . Social cognitive theory of moral thought and action. In W. M. Kurtines & J. L, Gewirtz (Eds . ) , Handbook of moral behavior and development: Theory, research and applications (Vol. 1, pp. 71-129). Hillsdale, N J : Erlbaum . Bandura, A., Barbaranelli , C, Caprara, G. V., & Pastorelli, C. (1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology. 71. 364-374. Bandura, A., Ross, D., & Ross, S. A. (1963). Vicarious reinforcement and imitative learning. Journal of Abnormal and Social Psvchology. 67. 601-607. 134

PAGE 145

135 Barrett, K. C. (1995) . A functionalist approach to shame and guilt. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions; The psycholocry of shame, guilt, embarrassment, and pride (pp. 25-63) . New York: Guilford Press . Batson, C. D., Dyck, J. L. , Brandt, J. R. , Batson, J. G., Powell, A. L., McMaster, M. R. , & Griffitt, C. (1988). Five studies testing two new egoistic alternatives to the empathy-altruism hypothesis . Journal of Personality & Social Psychology. 55. 52-77. Berkowitz, L. (1974) . Some determinants of impulsive aggression: Role of mediated associations with reinforcements for aggression. Psychological Review. 81. 165-176. Berkowitz, L. (1993) . Aggression: Its causes, consequences . and control . Philadelphia: Temple University Press . Berndt, T. J. (1979) . Developmental changes in conformity to peers and parents . Developmental Psychology, 15. 608-616. , . Bond, M. H., & Hwang, K. K. (1986). The social psychology of Chinese people. In M. H. Bond (Ed.), The psychology of the Chinese people (pp. 213-266) . Hong Kong: Oxford University Press . Bryant, B. K. (1982) . An index of empathy for children and adolescents. Child Development, 53. 413-425. Burger, J. M. (1981) . Motivational biases in the attribution of responsibility for an accident: A meta-analysis of the defensive-attribution hypothesis . Psychological Bulletin. 90. 496-512. Camp, B. W. (1977) . Verbal mediation in young aggressive boys. Journal of Abnormal Psychology. 86. 145153. Caprara, G. V., Pastorelli, C, & Weiner, B. (1997). Linkages between causal ascriptions, emotion, and behavior. International Journal of Behavioral Development. 20. 153162.

PAGE 146

136 Chao, R. K. (1994) . Beyond parental control and authoritarian parenting style: Understanding Chinese parenting through the cultural notion of training. Child Development . 65 . 1111-1119. Cheng, H. P., & Page, R. C. (1995). A comparison of Chinese (in Taiwan) and American perspectives of love, guilt, and anger. Journal of Mental Health Counse ling. 17. 210-219. Coie, J., & Dodge, K. (1988). Multiple sources of data on social behavior and social status in the school: A crossage comparison. Child Development. 59, 815-829. Cole, P. M. , Barrett, K. C, & Zahn-Waxler, C. (1992). Emotion displays in two-year-olds during mishaps . Child Development . 63 . 314-324. Crick, N. R. , & Dodge, K. A. (1994). A review and reformulation of social information-processing mechanisms in children's social adjustment. Psychological Bulletin. 115, 74-101. Crowell, D. H. (1987) . Childhood aggression and violence: Contemporary issues. In D. H. Crowell, I. M. Evan, & C. R. O'Donnell (Eds.), Childhood aggression and violence (pp. 19-52). New York: Pleniim Press. Cuddy, M. E., & Frame, C. (1991). Comparison of aggressive and nonaggressive boys' self-ef f icacy and outcome es^ectancy beliefs. Child Study Journal, 21, 135-152. Dodge, K. A. (1980). Social cognition and children's aggressive behavior. Child Development, 51, 162-170. Dodge, K. A. (1993) . Social -cognitive mechanisms in the development of conduct disorder and depression. Annual Review of Psychology, 44, 559-584. Dunn, J. , Brown, J. R. , & Maguire, M. (1995) . The development of children's moral sensibility: Individual differences and emotion understanding. De ve 1 opmen tal Psychology. 31. 649-659. Eberhard, W. (1967) . Guilt and sin in traditional China . Berkeley and Los Angeles: University of California Press .

PAGE 147

137 Eisenberg, N. , & Miller, P. A. (1987). The relation of empathy to prosocial and related behaviors . PsYcholoqical Bulletin. 101. 91-119. Ellis, P. L. (1982) . Empathy: A factor in antisocial behavior. Journal of Abnormal Child Psychology. 10. 123-134. Eron, D. E. (1994) . Theories of aggression: From drives to cognitions. In L. R. Huesmann (Ed.), Aggressive behavior: Current perspectives (pp. 3-11) . New York: Plenum Press . Eron, D. E., Guerra, N. , & Huesmann, L. R. (1997). Poverty and violence . In S . Feshbach & J . Zagrodzka (Eds . ) , Aggression: Biological, developmental and social perspectives (pp. 139-154) . New York: Plenum Press. Eron, L. D., Huesmann, L. R. , Dxibow, E., Romanoff, R. , & Yarmel, P. W. (1986) . Aggression and its correlates over 22 years. In D. H. Crowell, L. M. Evans, & C. R. O'Donnell (Eds.), Childhood aggression and violence (pp. 249-262). New York: Plenum Press. Eysenck, S. B. 6., Easting, 6., & Pearson, P. R. (1984). Age norms for impulsiveness, venture someness and empathy in children. Personality and Individual Differences, 5^ 315-321. Farrington, D. P. (1991). Childhood aggression and adult violence : Early precursors and later-life outcomes . In D . J . Pepler & K . H . Rvibin (Eds . ) , The development and treatment of childhood aggression (pp. 5-29) . Hillsdale, NJ: Erlbaiim . Ferguson, T. J., & Stegge, H. (1995). Emotional states and traits in children: The case of guilt and shame. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions : The psychology of shame, guilt, embarrassment and pride (pp. 147-197). New York: Guilford Press. Ferguson, T. J., Stegge, H. , & Damhuis, I. (1990). Guilt and shame experiences in elementary school-age children. In P. J. D. Drenth, J. A. Sergeant, & R. J. Takens (Eds . ) , European perspectives in psychology (Vol . 1 . , pp . 195-218) . Chichester, England: Wiley.

PAGE 148

138 Feshbach, N. D. (1983). Learning to care: A positive approach to child training and discipline. Journal of Clinical Child Psychology. 12. 266-271. Feshbach, N. D. (1997) . Empathy: The formative years-Implications for clinical practice. In A. C. Bohart & L . S . Greenberg (Eds . ) , Empathy reconsidered: New directions in psychotherapy (pp. 33-59) . Washington, DC: American Psychological Association . Franko, D. L. , Powers, T. A., Zuroff, D. C, & Moskowitz, D. S. (1985) . Children and affect: Strategies for self -regulation and sex differences in sadness . American Journal of Orthopsychiatry . 55 . 210-219. Fraser, S. (1996) . Guilt and shame in children: Relationships with empathy and aggression. Unpublished master's thesis, Simon Fraser University, British Columbia, Canada . Fung, H. H. T. (1994). The socialization of shame in young Chinese children (Taiwan) . The University of Chicago. Dissertation Abstracts International. 55-03 B. 1200. Geen, R. G. (1990) . Human aggression . Milton Keynes, English: Open University Press. Ghodsian, M. , Fogelman, K. , Lambert , L. , & Tibbenhsun, A. (1980) . Changes in behavior ratings of a national sample of children. British Journal of Social and Clinical Psychology. 19. 247-256. Gibson, H. B., & West, D. J. (1970). Social and intellectual handicaps as precursors of early delinquency. British Journal of Criminology. 10. 21-32. Graham, S. (1988). Children's developing understanding of the motivational role of affect: An attributional analysis. Cognitive Development. 3. 71-88. Graham, S., Doubleday, C, & Guarino, P. A. (1984). The development of relations between perceived controllability and the emotions of pity, anger, and guilt. Child Development. 55. 561-565. Graham, S., Hudley, C, & Williams, E. (1992). Attributional and emotional determinants of aggression euaong

PAGE 149

139 African-American and Latino young adolescents. Developmental Psychology. 28. 731-740. Graham, S., Weiner, B. , & Benish-Weiner , M. (1995). An attributional analysis of the development of excuse giving in aggressive and nonaggressive African American boys . Developmental Psychology. 31. 274-284. Guerra, N. G. , & Slaby, R. G. (1990). Cognitive mediators of aggression in adolescent offenders : 2 . Intervention. Developmental Psychology. 26, 269-277. Harris, P. L. (1989) . Children and emotion: The development of psychological understanding. Oxford, England: Basil Blackwell. Her, E. H. (1991) . A phenomenological explication of shame in a shame culture: A cross-cultural perspective (China, Taiwan) . Southern Illinois University at Carbondale. Dissertation Abstracts International. 52-05 A, 1571. Hirschi, T., & Hindelang, M. J. (1977). Intelligence and delinquency: A revisionist review. American Sociological Review . 42 . 571-587. Ho, D. Y. F. (1986) . Chinese patterns of socialization: A critical review. In M. H. Bond (Ed.), The psychology of the Chinese people (pp. 1-37) . Hong Kong: Oxford University Press . Hoffman, M. L. (1982) . Development of prosocial motivation: Empathy and guilt. In N. Eisenberg-Berg (Ed.), Development of prosocial behavior (pp. 281-313) . San Diego, CA: Academic Press. Hoffman, M. L., & Saltzstein, H. D. (1967). Parent discipline and the child' s moral development . Journal of Personality and Social Psychology, 5. 45-57. Hong, Y. Y. , & Chiu, C. Y. (1992). A study of the comparative structure of guilt and shame in Chinese society. The Journal of Psychology, 126, 171-179. Hong, Z. H. (1985). A study of aggressive behavior in students of middle school . Unpiiblished master's thesis, Graduate Institute of Education of National Taiwan Normal University, Taipei, Taiwan. (In Chinese.)

PAGE 150

140 Hsu, F. L. K. (1963) . Clan, caste, and club: A comparative study of Chinese. Hindu and American ways of life . Princeton, NJ: D. Van Nostrand. Huesmann, L. R. , & Eron, L. D. (1984). Cognitive processes and the persistence of aggressive behavior. Aggressive Behavior. 10. 243-251. Huesmann, L. R. , Eron, L. D., Lefkowitz, M. M. , & Walder, L. O. (1984) . Stability of aggression over time and generations. Developmental Psychology. 20. 1120-1134. Huesmann, L. R. , Eron, L. D., & Yarmel, P. W. (1987). Intellectual functioning and aggression. Journal of Personality and Social Psychology. 52, 232-240. Huesmann, L. R. , & Guerra, N. G. (1997). Children's normative beliefs about aggression and aggressive behavior. Journal of Personality and Social Psychology, 72, 408-419. Izard, C. E. (1977) . Human emotions. New York: Plentim Press . Johnson, R. C, Danko, 6. P., Huang, Y. H., Park, J. Y., Johnson, S. B., & Nagoshi, C. T. (1987). Guilt, shame, and adjustment in three cultures. Personality and Individual Differences . 8 . 357-364. Justice Department of the Repviblic of China. (1996) . The condition and analysis of adolescents and children crime in the Republic of China. 1995. Taipei, Justice Department of the Republic of China. Kazdin, A. (1994) . Interventions for aggressive and antisocial children. In L. Eron, J. Gentry, & P. Schlegel (Eds . ) , Reason to hope: A psychological perspective on violence and vouth (pp. 341-382) . Washington, DC: American Psychological Association. Klaczynski, P., & Gordon. D. H. (1996). Everyday statistical reasoning during adolescence and young adulthood: Motivational, general ability, and developmental influences. Child Development. 67. 2873-2891. Kochanska, G. (1991) . Socialization and temperament in the development of guilt and conscience. Child Development. 62. 1379-1392.

PAGE 151

141 Kochanska, G. (1993) . Toward a synthesis of parental socialization and child temperament in early development of conscience. Child Development, 64. 325-347. Krevans, J., & Gibbs, J. C. (1996). Parents' use of inductive discipline: Relations to children's empathy and prosocial behavior. Child Development. 67. 3263-3277. Kriger, S. F. , & Kroes, W. H. (1972). Child-rearing attitudes of Chinese, Jewish, and Protestant mothers. Journal of Social Psychology, 86. 205-210. Lazarus, R. S. (1991) . Emotion and adaptation. New York: Oxford University Press. Lewis, M. (1992) . Shame: The exposed self. New York: Free Press. Loeber, R. , & Dishion, T. (1983). Early predictors of male delinquency: A review. Psychological Bulletin. 93. 6899. Luengo, M. A., Carrillo-de-la-Pena, M. T. , Otero, J. M. , & Romero, E. (1994). A short-term longitudinal study of impulsivity and antisocial behavior. Journal of Personality and Social Psychology. 66. 542-548. Marsella, A. J., Murray, M. D., & Golden, C. (1974). Ethnic variations in the phenomenology of emotions: I. Shame. Journal of Cross -Cultural Psychology. 5. 312-328. Martin, C. S., Earleywine, M. , Blackson, T. C, Vanyukov, M. M. , Moss, H. B., & Tarter, R. E. (1994). Aggressivity, inattention, hyperactivity, and impulsivity in boys at high and low risk for substance abuse. Journal of Abnormal Child Psychology. 22. 177-203. Mascolo, M. F., & Fischer, K. W. (1995). Developmental transformations in appraisals for pride, shame, and guilt. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology o f shame. gii-iH-, «>i«K arrassment . and pride (pp. 64-113). New York: Guilford Press. McGraw, K. M. (1987) . Guilt following transgressi An attribution of responsibility approach. Journal of Personality and Social Psychology. 53. 247-256.

PAGE 152

142 Miller, L. C. (1972) . School Behavior Check List: An inventory of deviant behavior for elementary school children . Journal of Consulting and Clinical Psycholog y. 38. 134-144. Miller, P. A. , & Eisenberg, N. (1988). The relation of empathy to aggressive and externalizing/antisocial behavior. Psvchological Bulletin. 103. 324-344. Mosher, D. L. (1965) . Interaction of fear and guilt in inhibiting unacceptable behavior. Journal of Consulting Psvchologv. 29. 161-167. Mosher, D. L. , Mortimer, R. L. , & Grebel, M. (1968). Verbal aggressive behavior in delinquent boys . Journal of Abnormal Psychology. 73. 454-460. Nunner-Winkler , G. , & Sodian, B. (1988). Children's understanding of moral emotions. Child Development. 59. 1323-1338. Offord, D. R., Boyle, M. H., & Racine, Y. A. (1991). The epidemiology of antisocial behavior in childhood and adolescence. In D . J. Pepler & K. H. Rtibin (Eds.), The development and treatment of childhood aggression (pp. 3154) . Hillsdale, NJ: Erlbaum. Okel, E., & Mosher, D. L. (1968). Changes in affective states as a function of guilt over aggressive behavior. Journal of Consulting and Clinical Psychology. 37. 265-270. Perry, D. 6., Perry, L. C, Bussey, K. , English, D., & Arnold, G. (1980). Processes of attribution and children's self -punishment following misbehavior. Child Development, 51. 545-551. Perry, D. G. , Perry, L. C, & Rasmussen, P. (1986). Cognitive social learning mediators of aggression. Child Development. 57. 700-711. Perry, D. G. , Perry, L. C, & Weiss, R. J. (1989). Sex differences in the consequences that children anticipate for aggression. Developmental Psychology, 25. 312-319. Piaget, J. (1965) . The moral judgment of the child. New York: Free Press.

PAGE 153

143 Pulkkinen, L., & Pitkanen, T. (1993) . Continuities in aggressive behavior from childhood to adulthood. Aggressive Behavior. 19. 249-263. Quiggle, N, L. , Garber, J., Panak, W. F. , & Dodge, K. A. (1992) . Social information processing in aggressive and depressed children. Child Development. 63, 1305-1320. Regan, J. W. (1971) . Guilt, perceived injustice, and altruistic behavior. Journal of Personality and Social Psychology. 18. 124-132. Richardson, D. R. , Hammock, G. S., Smith, S. M. , Gardner, W. , & Signo, M. (1994). Empathy as a cognitive inhibitor of interpersonal aggression. Aggressive Behavior. 20. 275-289. Rothbart, M. K. , Ahadi, S. A., & Hershey, K. L. (1994) . Temperament and social behavior in childhood. Merrill-Parlmer Quarterly. 40, 21-39. Sattler, J. M. (1989) . Assessment of children (3rd ed.). San Diego, CA: Sattler. Scheff , T. J. (1987) . The shame-rage spiral: A case study of an interminable quarrel. In H. B. Lewis (Ed.), The role of shame in symptom formation (pp. 109-149) . Hillsdale, N J : Er Ibaum . Scheff, F. J. (1988) . Shame and conformity: The deference emotion system. American Sociological Review, 53, 395-406. Scheff, T. J., & Retzinger, S. M. (1991). Emotions and violence : Shame and rage in destructive conflicts . Lexington, MA: Lexington Books. Schill, T., & Schneider, L. (1970). Guilt and selfreport of hostility. Psychological Reports, 27, 713-714. Schunk, D. H. (1991). Learning theory: An educational perspective . New York: Macmillan . Singer, M. B. (1953) . Shame cultures and guilt cultures. In G. Piers & M. B. Singer, Shame and guilt: A psychoanalytic and a cultural study. Springfield, IL: Charles C. Thomas.

PAGE 154

144 Slaby, R. G. , & Guerra, N. 6.(1988). Cognitive mediators of aggression in adolescent offenders : 1 . Assessment. Developmental Psychology. 24, 580-588. Slaby, R. G. , & Guerra, N. G. (1989) . Evaluative factors in social problem solving by aggressive boys. Journal of Abnormal Child Psychology. 17. 277-288. Stevens, J. (1996) . Applied multivariate statistics for the social sciences (3rd ed.). Mahwah, NJ: Erlbaum. Stevenson, H, W. , Chen, C, & Lee, S. (1992). Chinese families. In J. L. Roopnarine & D. B. Carter (Eds.), Parentchild socialization in diverse cultures (pp. 17-33) . Norwood , N J : Ablex . Su, C. W. (1975) . Parental child-rearing practices as related to moral behavior in adolescence, Acta Psychologica Taiwanica, 17. 109-24. (In Chinese.) Tangney, J. P. (1990) . Assessing individual differences in proneness to shame and guilt: Development of the self-conscious affect and attribution inventory. Journal of Personality and Social Psychology. 59. 102-111 Tangney, J. P. (1991) . Moral affect: The good, the bad, and the ugly. Journal of Personality and Social Psychology . 61 . 598-607. Tangney, J. P. (1992) . Situational determinants of shame and guilt in young adulthood. Personality and Social Psychology Bulletin. 18, 199-206. Tangney , J . P . , Burggraf , S . A . , & Wagner , P . E . (1995) . Shame -pronenes s , guilt-proneness , and psychological symptoms. In J. P. Tangney & K. W. Fischer (Eds.), Selfconscious emotions: The psychology of shcune, guilt, embarrassment and pride (pp. 343-367) . New York: Guilford Press . Tangney , J . P . , Wagner , P . E . , Barlow , D . H . , Marschall, D. E., & Gramzow, R. (1996). Relation of shame and guilt to constructive versus destructive responses to anger across the life span. Journal of Personality and Social Psychology. 70. 797-809. Tangney, J. P., Wagner, P. E., Burggraf, S. A., Gramzow, R. , & Fletcher, C. (1991, June). Children^ s shame-

PAGE 155

145 proneness. but not guilt-proneness . is related to emotional and behavioral maladiustment . Poster presented at the meeting of the American Psychological Society, Washington, DC. Tangney , J . P . , Wagner , P . E . , Fletcher , C . , & Gramzow, R. (1992) . Shame into anger? The relation of shame and guilt to anger and self -reported aggression. Journal of Personality and Social Psychology , 62 . 669-675. Thompson, R. A. (1994) . Emotion regulation: A theme in search of definition. In N. A. Fox (Ed.), The development of emotion regulation: Biological and behavioral considerations . Monographs of the Society for Research in Child Development. 59 (2-3, Serial No. 240). Thompson, R. A., & Hoffman, M. L. (1980). Empathy and the development of guilt in children. Developmental Psvchologv. 16. 155-156. Wang, P. W. (1983) . A study of juvenile delinquency ir Taiwan: An application of differential opportunity theory. Dissertation Abstracts International. 44(1-A). 289. Williams, C, & Bybee, J. (1994). What do children feel guilty about? Developmental and gender differences. Developmental Psychology, 30. 617-623. Wilson, R. W. (1981) . Moral behavior in Chinese society: A theoretical perspective. In R. W. Wilson, S. L. Greenblatt, & A. A. Wilson (Eds.), Moral behavior in Chinese society (pp. 1-20) . New York: Praeger. Xie, S. F. (1991) . A study of the cognitive processes of interpersonal problem solving, aggressive beliefs and aggression. Unpublished master's thesis. Graduate Institute of Educational Psychology and Counseling, National Taiwan Normal University, Taipei, Taiwan. (In Chinese.) Yang, Y. S. (1985). Field-dependent and fieldindependent sixth graders and their moral judgement. Unpublished master' s thesis, Graduate Institute of Education, National Taiwan Normal University. Taipei, Taiwan. (In Chinese) . Zahn -Waxier, C, Kochanska, C, Krupnick, J., & McKnew, D. (1990). Patterns of guilt in children of

PAGE 156

146 depressed and well mothers . Developmental PsYcholoay. 26. 51-59. Zahn-Waxler, C, & Robinson, J. (1995). Empathy and guilt: Early origins of feelings of responsibility. In J. P. Tangney & K. W. Fischer (Eds.), Self-conscious emotions: Th^ psychology of shame, quilt, embarrassment, and pride (pp. 143-173). New York: Guilford Press. Zhang, D. M. (1990) . Cognitive variables which influence aggression and then social status in elementary school children. Unpublished master's thesis, Graduate Institute of Psychology, National Taiwan University, Taipei, Taiwan. (In Chinese.)

PAGE 157

BIOGRAPHICAL SKETCH ' "... * i Shu-Tai Liao was born in Taichung, Taiwan, on November 1, 1950. In 1973 she graduated from Taiwan Normal University with a bachelor's degree in education. In this year she started to teach at Taichung Teachers College. In 1989, she completed her master of arts in education degree in early and middle childhood education at the Ohio State University. In the fall of 1994, she began the doctoral progrsun in educational psychology at the University of Florida, majoring in child development. She will return to Taiwan to resume her teaching career after finishing her doctoral progreun. She has been married since 1977, and she is the mother of two sons . 147

PAGE 158

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. Patj^icia Ashton, Chair Professor of Foundations of Education , 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. Assistant Professor of 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, Associate" Professor of Foundations of Education 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. ilssistant Professor of Foundations of Education

PAGE 159

This dissertation was siibmitted 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. December 1998 . ^ lhairman , Foundations of Education Dean, Graduate School