<%BANNER%>

The Roles of Emotion Processes and Hypothetical Scenarios in Predicting Antisocial Behavior

Permanent Link: http://ufdc.ufl.edu/UFE0021816/00001

Material Information

Title: The Roles of Emotion Processes and Hypothetical Scenarios in Predicting Antisocial Behavior
Physical Description: 1 online resource (54 p.)
Language: english
Creator: Clemans, Katherine H
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: adolescence, aggression, anger, attributions, emotion, hostile, hypothetical, regulation, vignettes
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The current study investigated the association of attributions and responses made to hypothetical vignettes, as well as trait levels of anger and anger regulation, with reported antisocial behavior in a sample of early adolescents. A sample of 413 6th graders completed two hypothetical social problem-solving tasks as well as a larger survey assessing anger, anger regulation, aggression, delinquency, and association with delinquent friends. Hierarchical regressions indicated that anger and anger regulation were much stronger predictors of antisocial behavior and delinquent friends than were hypothetical aggressive responses. Surprisingly, hypothetical hostile attributions were not related to hypothetical aggressive responses or to any antisocial outcome variable. Additional multivariate ANCOVAs conducted on attribution-response pattern groups within each vignette indicated that participants who made nonhostile attributions but still generated aggressive responses for one of the two vignettes evidenced higher levels of antisocial tendencies than those with nonhostile attributions and nonaggressive responses. The efficacy of using hypothetical vignettes to predict antisocial behavior in real life is discussed, as well as the need to consider characteristics of both the vignettes and of participants when conducting research in this area.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Katherine H Clemans.
Thesis: Thesis (M.S.)--University of Florida, 2007.
Local: Adviser: Graber, Julia A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021816:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021816/00001

Material Information

Title: The Roles of Emotion Processes and Hypothetical Scenarios in Predicting Antisocial Behavior
Physical Description: 1 online resource (54 p.)
Language: english
Creator: Clemans, Katherine H
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: adolescence, aggression, anger, attributions, emotion, hostile, hypothetical, regulation, vignettes
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The current study investigated the association of attributions and responses made to hypothetical vignettes, as well as trait levels of anger and anger regulation, with reported antisocial behavior in a sample of early adolescents. A sample of 413 6th graders completed two hypothetical social problem-solving tasks as well as a larger survey assessing anger, anger regulation, aggression, delinquency, and association with delinquent friends. Hierarchical regressions indicated that anger and anger regulation were much stronger predictors of antisocial behavior and delinquent friends than were hypothetical aggressive responses. Surprisingly, hypothetical hostile attributions were not related to hypothetical aggressive responses or to any antisocial outcome variable. Additional multivariate ANCOVAs conducted on attribution-response pattern groups within each vignette indicated that participants who made nonhostile attributions but still generated aggressive responses for one of the two vignettes evidenced higher levels of antisocial tendencies than those with nonhostile attributions and nonaggressive responses. The efficacy of using hypothetical vignettes to predict antisocial behavior in real life is discussed, as well as the need to consider characteristics of both the vignettes and of participants when conducting research in this area.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Katherine H Clemans.
Thesis: Thesis (M.S.)--University of Florida, 2007.
Local: Adviser: Graber, Julia A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021816:00001


This item has the following downloads:


Full Text





THE ROLES OF EMOTION PROCESSES AND HYPOTHETICAL SCENARIOS IN
PREDICTING ANTISOCIAL BEHAVIOR























By

KATHERINE HALE CLEMANS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2007

































2007 Katherine Hale Clemans









ACKNOWLEDGMENTS

This research was sponsored by a grant to Dr. Gilbert J. Botvin from the National Institute

on Drug Abuse (P50DA-07656). Thanks are extended to Drs. Julia Graber, Gilbert Botvin,

Jeanne Brooks-Gunn and Tracy Nichols for the use of the dataset. I would also like to extend a

special thanks to my thesis committee members, Drs. Julia Graber, Susan Bluck and Catherine

Cottrell, for their valuable assistance and advice.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ...............................................................................................................3

LIST OF TABLES .............. ......... ........... ........................................6

LIST O F FIG U RE S ................................................................. 7

ABSTRAC T .........................................................................................

CHAPTER

1 IN T R O D U C T IO N ................................................................................................... . 9

Social Inform action Processing ..................................... .............. .......................
The Influence of Em option ....................................................................12
M ethodological Issues ........................................ ........................................ ........................14
The Present Study .............. ............................ ........................... ..... 15

2 M E T H O D S .......................................................................................................19

D e sig n .................................................................................................................. ...............19
P articip an ts ................ ........ .... ...... ............................................. .................................. 19
P ro c e d u re s .............................................................................................2 0
M e a su re s .................................. .................................................................................................2 1
H hypothetical V ignettes ....................................................................... ..21
H o stile attrib u tio n s ............................................................................................. 2 1
H hypothetical aggressive responses ..................................................................... 22
Survey M measures ....................................................... 22
A n g e r ........................................................................................................2 2
A nger reduction.............................................. 23
A aggression ..................................................................................................... ..... 23
D elin qu en cy ....................................................... 2 3
Friends' delinquency ............................................................................. 23
Demographic Characteristics.................................. 24

3 R E SU L T S .............. ... ................................................................25

Demographic Associations ................................. ............................ ..........25
Associations Among Core Constructs ............................................................. .............26
Attribution-Response Patterns ..... ........... ........ .......... ........28
L u n c h T ab le .....................................................................................................................2 9
Hallway ............... ...... ............. ............. ...............30

4 DISCU SSION ......... ......... ..... ....... ..... ...................... ........ 41




4









APPENDIX HYPOTHETICAL VIGNETTE SCRIPTS...................................... ............... 48

L IS T O F R E F E R E N C E S ...................................................................................... ....................4 9

B IO G R A PH IC A L SK E T C H ............................................................................... .....................54









LIST OF TABLES


Table page

3-1 Sum m ary statistics on continuous variables.................................. ............................. 31

3-2 Correlations am ong study variables .............................................. ............................. 32

3-3 Hierarchical regression analysis for variables predicting reported rates of aggression .........34

3-4 Hierarchical regression analysis for variables predicting reported rates of delinquency .......35

3-5 Hierarchical regression analysis for variables predicting reported rates of delinquent
frie n d s ..........................................................................3 6

3-6 Attribution-response pattern group membership...... ....................... ..............37

3-7 Univariate AN OVA s for lunch groups........................................................ ............... 38









LIST OF FIGURES


Figure p e

3-1 Attribution-response pattern differences in antisocial outcome variables for lunch table
and hallw ay vignettes........... .................................................................. ........ .. ....... .. 39

3-2 Lunch table attribution-response pattern differences in delinquency by race.....................40









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

THE ROLES OF EMOTION PROCESSES AND HYPOTHETICAL SCENARIOS IN
PREDICTING ANTISOCIAL BEHAVIOR

By

Katherine Hale Clemans

December 2007

Chair: Julia A. Graber
Major: Psychology

The current study investigated the association of attributions and responses made to

hypothetical vignettes, as well as trait levels of anger and anger regulation, with reported

antisocial behavior in a sample of early adolescents. A sample of 413 6th graders completed two

hypothetical social problem-solving tasks as well as a larger survey assessing anger, anger

regulation, aggression, delinquency, and association with delinquent friends. Hierarchical

regressions indicated that anger and anger regulation were much stronger predictors of antisocial

behavior and delinquent friends than were hypothetical aggressive responses. Surprisingly,

hypothetical hostile attributions were not related to hypothetical aggressive responses or to any

antisocial outcome variable. Additional multivariate ANCOVAs conducted on attribution-

response pattern groups within each vignette indicated that participants who made nonhostile

attributions but still generated aggressive responses for one of the two vignettes evidenced higher

levels of antisocial tendencies than those with nonhostile attributions and nonaggressive

responses. The efficacy of using hypothetical vignettes to predict antisocial behavior in real life

is discussed, as well as the need to consider characteristics of both the vignettes and of

participants when conducting research in this area.









CHAPTER 1
INTRODUCTION

Social Information Processing

Adolescence is a time of numerous behavioral changes, during which temporary increases

in aggression and antisocial activity are often observed (Moffitt, 1993). In addition to the fact

that aggressive behavior merits serious attention and preventive efforts on its own, for a subset of

individuals, aggression in early adolescence can foreshadow more serious antisocial behavior in

adulthood (Coie & Dodge, 1998; Moffitt, Caspi, Rutter, & Silva, 2001). Therefore, the study of

factors that influence aggressive and antisocial tendencies remains an important area of research

in this age group. The present study investigates and compares the ability of attributions and

responses made to hypothetical scenarios, as well as anger-related emotional processes, to

predict rates of antisocial behavior in real life.

When an individual is trying to make sense of a social situation, it is the perception of

others' intentions, regardless of those others' actual intentions, that guides the individual's

response (Berkowitz, 1990; Dodge & Coie, 1987). Most research on attributions of intent is

conducted using a social information-processing (SIP) framework (Crick & Dodge, 1994;

Dodge, 1986; Orobio de Castro, Veerman, Koops, Bosch, & Monshouwer, 2002), which

attempts to explain how cognitive and emotional factors influence an individual's understanding

and conception of the social world. These models describe a series of steps in which social

information is encoded, interaction goals are recognized, and possible response alternatives are

generated and evaluated, resulting in a reaction to a given social situation. Atypical processing

during any step may lead to socially inappropriate behavior, such as aggression (Crick & Dodge,

1994; Lemerise & Arsenio, 2000).









Because of the potentially overwhelming amount of information to which an individual

must attend in any given social scenario, (s)he will often try to simplify the amount of cognitive

processing needed by applying heuristic shortcuts, such as the selective attention to particular

cues, or a reliance on past experiences to inform interpretation of the current event (Crick &

Dodge, 1994; VanOostrum & Horvath, 1997). Hostile attribution biases may arise through

deficiencies in cue utilization, or the capacity to integrate attention information into cognition

(Dodge, 1980). For instance, Dodge and Frame (1982) found that aggressive children falsely

recalled hostile cues as taking place in past social situations more often than did their

nonaggressive peers, partially supporting the conclusion that aggressive children were selectively

attending to hostile cues at the expense of other social signals.

In a study on boys' reactions to ambiguous social situations, Dodge (1980) found that

when aggressive participants made hostile attributions of intent to the peer provocateur, they

responded aggressively 60% of the time. When this same group attributed benign intent to the

provocateur, however, they still reported that they would respond aggressively 26% of the time.

The results of Dodge's study suggest that other characteristics of individuals, besides a tendency

to interpret hostile intent, play a role in the formation of aggressive responses (VanOostrum &

Horvath, 1997).

Researchers often attempt to study the provocation and initiation of aggression as well as

components of the SIP model through the use of hypothetical and role-play vignettes (e.g.,

Dodge, 1990; Vitaro & Pelletier, 1991). These tasks usually present a problematic social

situation or a conflict-resolution scenario and ask the participants to react or to answer questions

about how they would react in real life. Research of this type allows the benefit of assessing

beliefs and emotions at the time of conflict. However, no matter how salient and similar to the









participant's life a particular hypothetical scenario may be, the participants are always aware of

the task's "make-believe" quality. This remains the case in most role-play tasks as well, which,

in research with children and adolescents, may use an adult confederate to interact with the target

child (e.g., Borbely, Graber, Nichols, Brooks-Gunn, & Botvin, 2005). As a result, it is possible

that important aspects of the real-life social scenario are not being captured through hypothetical

and role-play investigative techniques.

Hypothetical social problem-solving tasks can often elicit differences in response

strategies between aggressive and nonaggressive individuals, particularly in cases where the

provocateur's intentions are ambiguous (Crick & Dodge, 1996; Orobio de Castro et al., 2002).

For instance, Tremblay and Ewart (2005) found modest correlations between self-reported

aggression and various aggressive responses to hypothetical vignettes of provoking situations in

a sample of college students. VanOostrum & Horvath (1997) found that hostile attributions to

ambiguous peer scenarios significantly predicted both aggressive hypothetical responses and

actual aggressive behavior in a sample of high school-aged boys. However, results have been

inconsistent across studies, with some suggesting that these relationships only hold for highly

aggressive, all-male samples (Crain, Finch, & Foster, 2005; Orbio de Castro et al., 2002).

Furthermore, there is evidence that children's utilization of response strategies in natural social

settings is frequently different from the answers they give to hypothetical social tasks (see Vitaro

& Pelletier, 1991, for a review). Vitaro and Pelletier found that maladjusted children behaved

more aggressively than their well-adjusted peers to provocation by another child; however, no

differences were observed in the two groups' responses to a hypothetical vignette depicting a

similar social scenario. The authors hypothesized that even though the maladjusted children may

have been aware of socially acceptable behavioral responses, they became unable to act









accordingly when confronted with real social problem situations because other factors, such as

"prohibitive emotional arousal" (p. 514), prevented the proper utilization of these strategies. The

influences of these other factors, specifically anger and anger regulation, are of interest in the

present study.

The Influence of Emotion

Though authors of SIP models have continually mentioned the importance of emotional

factors in social cognition (Crick & Dodge, 1994), the function and effects of emotion processes

within the SIP framework have received relatively little attention. Lemerise and Arsenio (2000)

are an exception, offering a reformulation of Crick and Dodge's model that addresses the

interaction between cognition and emotion processes in the interpretation of social cues and the

generation of behavior. This new model views emotion processes, such as emotionality and

temperamental factors, emotion regulation, and background moods, as influencing each step of

the social cognitive process. For instance, a particularly anxious child may weigh response

outcomes differently than a fearless one (Saltaris, 2002).

The influence of emotion on decision-making strategies depends on two main factors: the

strength and valence of a particular experience of emotion, often referred to as arousal or

emotionality; and the ability to consciously regulate or control the effect of a particular

emotional experience. Efficient emotion regulation in challenging situations allows the

individual to engage in more effortful evaluation of response alternatives, instead of jumping to a

quick conclusion (Lemerise & Arsenio, 2000, Saarni, 1999). Being a "good regulator" may allow

an individual to consider multiple perspectives of a situation instead of jumping to a quick

conclusion (Saarni, 1999).

Individual differences in arousal levels and regulatory abilities are related to differences

in both prosocial and problem behavior. Studies by Eisenberg and colleagues (Eisenberg et al.,









2001; Eisenberg et al., 1996) have shown that elementary school-aged children who are high in

emotionality but low in regulatory abilities display the highest frequency of conduct problems

and socially inappropriate behavior. One explanation of these findings is that these emotional

factors may cause deficiencies in multiple steps of the social-information-processing cycle, so

that in any given social situation, these children are prone to engage in more abnormal social

cognitive processes (Lemerise & Arsenio, 2002). The intensity of emotional experiences and the

ability to regulate those experiences will also influence the qualities an individual notices about a

social encounter and the meaning ascribed to those qualities. In situations of ambiguous intent,

the influence of emotion on decision-making may be heightened due to a lack of clearly

interpretable cues on which to base the formation of cognitive reasons for a particular response

(Lemerise & Arsenio, 2002).

Many types of emotional events can be experienced during the course of a social

interaction, but one emotion thought to be particularly related to aggression is anger (Dearing et

al., 2002). Research has demonstrated that high anger levels can contribute to aggressive

behavior (Cornell, Peterson, & Richards, 1999; Nichols, Graber, Brooks-Gunn, & Botvin, 2006).

The emotional experience of anger is characterized by "physiological arousal and cognitions of

antagonism" (Novaco, 1994, p. 32); trait anger is defined as an enduring propensity to become

angry (Spielberger, Jacobs, Russell, & Crane, 1983). Anger and the ability to effectively regulate

anger are related, since an individual with a higher baseline anger level will have a much tougher

job, relatively, of controlling his/her temper. Nichols et al. (2006) found aggression rates to be

positively correlated with anger and negatively correlated with an overall scale of self-control.

However, emotion arousal and emotion regulation are distinct factors that contribute uniquely to









antisocial behavior (Eisenberg et al., 2001; Lengua, 2003), suggesting that skill at effortful

emotion regulation may vary widely within individuals with high anger levels.

Though the influence of emotion regulation factors on behavior has received much

attention in recent years (Eisenberg, Morris, & Spinrad, 2005), research on anger regulation

specifically has been scarce (Zeman, Shipman, & Suveg, 2002). Results have been mixed, with

some studies finding no direct link between anger regulation and aggression in younger children

(Dearing et al., 2002). In particular, there is little research about this relationship in adolescents.

The present study included a scale that assesses control strategies relating explicitly to

experiences of anger, and as such was more specific than scales assessing general regulation

strategies.

Methodological Issues

Most of the research on the relationships between hostile attribution of intent and

aggressive behavior has been conducted on preschool and elementary-school-aged children; less

attention has been paid to the relationship of these processes in adolescents, especially non-

clinical or non-incarcerated samples, even though this type of research on adolescents is

necessary to examine the persistence and stability of these relationships over the life-span

(VanOostrum & Horvath, 1997). The relationship of hypothetical aggressive responses to

reported rates of aggression in real life is also unclear, especially in this age group (Crain et al.,

2005); more research with mixed-sex, non-referred adolescent samples is needed. Thus, the

present study seeks to determine how well aggressive responses in hypothetical scenarios relate

to reports of aggressive behavior in real life in a non-clinical sample of early adolescents.

In addition, while research relating ambiguously hostile hypothetical vignettes to real-life

behavior have typically focused only on aggression, it is possible that the tendency to

overattribute hostile intentions to social interactions with peers may manifest itself in other









forms. Aggressive behavior is highly related to other forms of antisocial behavior, including

delinquent acts such as vandalism and theft (Moffitt et al., 2001). While previous studies have

investigated hostile attributions and the SIP framework in samples of "delinquent" individuals,

this term has usually been used to differentiate between normal and referred or incarcerated

samples, and thus the label represents aggressive behavior in addition to other antisocial acts

(Guerra & Huessmann, & Zelli, 1990). Few, if any, studies to date have included measures of

specific delinquent behaviors, or have attempted to separate these from specifically aggressive

behaviors. Therefore, one aim of the present study is to investigate whether the predictive ability

of hypothetical vignettes extends to delinquent behavior as well.

Attribution processing style may also be related to the antisocial behavior of one's peers.

For instance, a tendency to perceive hostility in social situations may lead individuals to select

peers for whom this attribution style is considered acceptable. Alternatively, exposure to

antisocial peers may lead to increased hostile attributions through modeling of others (Bosson &

Johnson, 2006). Though investigation of these mechanisms is beyond the scope of the present

study, individuals' friendship with aggressive and/or delinquent peers is included as an outcome

variable in order to assess its associative relationship with responses to the hypothetical

vignettes.

The Present Study

The present investigation examines responses to hypothetical vignettes, emotions, and

self-reported aggression, delinquency, and peer delinquency in a diverse sample of 6th graders.

Based on previous research on the SIP model (Orobio de Castro et al., 2002), a significant

relationship is expected between real-life rates of aggression and aggressive responses to the

hypothetical social problem-solving vignettes, such that more frequent aggressive responding

will lead to higher reported rates of aggressive behavior. Furthermore, it is expected that this









relationship will exist between rates of aggression and hostile attributions of intent in the

hypothetical vignettes as well.

A second purpose of the present study is to investigate the influence of emotion-related

variables on the relationship between responses to hypothetical vignettes and real-life aggression.

The function of emotion in social cognition can be viewed as a marker that signals important

features of the social interaction and offers direction for cognitive processes and behavior

(Damasio, 1994). Anger and hostile attribution should be related in this sense, because if one

feels high levels of anger (due to either high baseline levels or insufficient regulation), this would

be an emotional signal that (s)he is in a threatening situation; therefore, (s)he should attribute

hostile intent to the other person or people involved. A resultant prediction of this study is that

mediated relationships between anger and hostile attributions and anger reduction and hostile

attributions should both be observed, such that greater levels of trait anger and poorer ability to

regulate anger should increase attributions of hostile intent, which in turn should lead to more

aggressive responses in the hypothetical scenarios and more antisocial behavior in real life. A

series of hierarchical regressions in accordance with the procedures of Baron and Kenney (1986)

will be used to show the mediating impact of anger processes on the relationship between hostile

attributions and hypothetical aggressive responses, as well as on the relationship between hostile

attributions and reported antisocial outcome variables.

Anger and anger regulation skills may explain the association between hypothetical

responses and real-life antisocial behavior, but they also may explain unique variance in

antisocial behavior unrelated to the hypothetical scenarios (Zeman et al., 2002). Though anger

levels and anger reduction ability should be moderately negatively correlated (Eisenberg et al.,

2005), unique variance contributions to antisocial behavior are also expected from each. It is









hypothesized that this unique variance would represent differences in situational characteristics

and emotional salience between hypothetical and real-life peer conflicts. Hierarchical regressions

on each antisocial outcome variable, with anger processes, and hypothetical vignette responses

entered in separate steps, will assess the relative variance explained by each variable.

Although examinations of individual hypothetical vignettes is rare in the literature to date

(Orobio de Castro et al., 2002), one of the benefits of using these vignettes in SIP research is that

they allow for the investigation of specific attribution-response relationships as they relate to a

single scenario. Therefore, exploratory multivariate ANCOVAs will also be conducted for each

hypothetical vignette to assess the relationship of individual attribution-response patterns with

levels of real-life antisocial tendencies. It is possible that these additional analyses will uncover

new relationships between hypothetical attributions and responses not seen in the more

traditional regression analyses.

In addition to the age of the participants, several other demographic characteristics may

play a role in the differential formation of hostile attributions and the expression of aggression.

As stated previously, overall rates of aggression appear to be higher in adolescent males than in

adolescent females, suggesting that males and females may utilize emotional factors differently

in hostile or ambiguous peer situations (Crick, 1997). Family background and community

characteristics may also play a role in displays of aggression. For instance, aggression in

African-American urban communities may serve adaptive functions unique to that environment,

and therefore not be considered as socially inappropriate as it would be in other contexts (Coie &

Dodge, 1998). Because of the wealth of research addressing the relationships between these

factors and the cognitive, emotional and behavioral variables in question (see Coie & Dodge,









1998, for review), differences in gender and demographic characteristics will not be a central

focus of this study, but will be tested in preliminary analyses.









CHAPTER 2
METHODS

Design

The current investigation consists of a sub-study of a randomized clinical trial designed

to evaluate a school-based drug abuse and violence prevention program. A total of 42 New York

City middle schools took part in the full study, which was approved by the Institutional Review

Board for the Protection of Human Subjects in Research (IRB) at Weill Medical College, Cornell

University. Fourteen parochial and three public schools in the larger program agreed to

participate in supplementary data collection activities. The high relative percentage of parochial

schools in this sample is a result of the recruitment of only the smallest schools from within the

larger study for the supplementary data collection due to its intensive nature. Data for this study

is drawn from the pre-trial assessment at the beginning of the participants' 6th grade year in order

to avoid confounding effects from the intervention program.

Participants

Only students who completed both the survey measures used in the full study and the

supplementary videotaped activities were included in the current sample (N= 413). The mean

age was 11.63 years (SD = .49 years; range = 9.64 to 13.86 years). Girls made up 50.4% of the

sample (n = 208). Racial subgroups included African American (49.6%), Latino (25.4%),

Caucasian (17.9%), and other (6.8%). Fifty-three percent of the sample lived in two-parent, non-

blended families, while 23.5% lived with single mothers and 23.5% lived in other household

configurations. Parochial school students made up 69.7% of the sample; all other students

attended public school.









Procedures

Parental consent for the full clinical trial was obtained through a passive consent

procedure, which provided a comprehensive description of the investigation and the self-report

surveys and allowed parents the opportunity to object to their child's participation. The

subsample of 17 schools received a secondary passive consent form that explained the

supplementary data collection activities, including the hypothetical peer-interaction tasks.

Parental objection to the primary consent form precluded student participation in the

supplementary tasks.

An ethnically diverse team of three to five data collectors presented the self-report survey

on two separate days during regularly scheduled class periods. A standardized protocol similar to

ones used in previous research on drug use and delinquent behaviors was used (see Botvin et al.,

1994). To maintain the quality of the self-report data, identification codes rather than names were

used to emphasize the confidential nature of the questionnaire and students were assured about

the confidentiality of their responses. To enhance the validity of the self-report data, carbon

monoxide breath samples were collected using a variant of the bogus pipeline procedure

developed by Evans and colleagues (Evans, Hansen, & Mittlemark, 1977). This measure is

primarily used for questions pertaining to cigarette smoking, although research has shown that

bogus pipeline procedures can also increase the validity of other self-reported problem behaviors

(Tourangeau, Smith, and Rasinski, 1997).

Supplementary activities took place on a third day of data collection in the subset of 17

schools. Participating students completed a series of activities and answered questions about

their reactions to hypothetical vignettes describing ambiguous peer interactions in an interview

format. Data collectors were trained undergraduate and master degree students, and primarily









female, African-American and Hispanic, although at least one male data collector was sent to

each school.

Students completed the hypothetical vignette tasks individually in a private classroom

space provided by the school. A data collector read relevant instructions and confidentiality

information to each student before beginning the tasks. Students were informed that they were to

respond to these scenarios as they would in real life. Responses to open-ended questions were

transcribed for later coding. Two hypothetical social problem solving tasks (Dodge, 1990)

dealing with ambiguously aggressive peer interactions were selected from the larger set of tasks

for analysis in the present study. In the first hypothetical scenario, the participant is instructed to

imagine that he/she is denied access to a school lunch table for an unknown reason. In the second

hypothetical scenario, a student walking toward the participant in a school hallway yells out an

insult which may or may not have been directed at the participant. For both hypothetical tasks,

the participant was asked what he/she thought was going on in the mind of the provocateur as

well as what his/her desired and actual responses to the situation would be. Complete scripts for

the social problem solving tasks are included in Appendix A.

Measures

Hypothetical Vignettes

Hostile attributions. Answers to both the attribution and response questions of the

hypothetical social problem-solving vignettes used a coding scale developed for the larger study

(Graber et al., 2001; Nichols et al., 2001). Participants' attributions of the provocateurs' intent

were coded from open-ended responses at the end of the vignettes in which they indicated what

they thought was going on in the mind of the hypothetical peers while the events were taking

place. Attributions of intent were coded as hostile when they included intimidation (e.g., "she

wanted to start trouble," "they want to embarrass me"); attention-seeking (e.g., "she was trying









to be popular," "he was trying to show off in front of his friends"); or personal problems (e.g.,

"he was mad," "he must be jealous of me," "they're racist"). For comparison purposes, other

possible responses included self-blaming attributions (e.g., "he thinks I'm a loser") and

situational attributions (e.g., "somebody else really was sitting there"). Interrater agreement was

89% (K = .80) for the lunch table vignette and 87% (K = .84 ) for the hallway vignette. Hostile

attribution responses for each question were summed to produce a hostile attribution score with a

range of 0-2.

Hypothetical aggressive responses. Aggressive responses were coded from open-ended

responses at the end of the vignettes in which participants indicated what they would do or say to

the other kids if this event happened to them. Responses were coded as aggressive if they

included physical force, such as pushing or shoving; verbal confrontation, such as the use of

threats, insults, ridicule, or sarcasm; or non-verbal hostile gestures, such as giving nasty looks.

Examples of nonaggressive responses included walking away or questioning why the seat was

unavailable. Interrater agreement was 90% (K = .88) for the lunch table vignette and 89% (K =

.86 ) for the hallway vignette. Aggressive responses for each question were summed to produce

a hypothetical aggressive response score with a range of 0-2.

Survey Measures

Anger. The seven-item anger subscale (a = .74) from the Buss and Perry (1992)

Aggression Questionnaire was used to assess trait levels of anger. Students were asked to rate

how well a series of statements fit them. Items included "I sometimes feel like a powder keg

ready to explode" and "Some of my friends think I'm a hothead." Response categories ranged

from 1 (Really Not True for Me) to 5 (Really True for Me). Items were averaged such that higher

scores indicate greater anger.









Anger reduction. Anger reduction skills were assessed with a six-item subscale (a = .81)

created for the larger clinical trial (Epstein, Botvin, Diaz, Baker, & Botvin, 1997). Participants

were asked how often they engaged in a series of activities when they felt really angry. Items

included "Count to ten," "Take a few deep breaths," and "Tell myself this isn't worth fighting

over (it's no big deal.)" Response categories ranged from 1 (Never) to 5 (Always). Items were

averaged such that higher scores indicated greater skill at conscious anger reduction.

Aggression. Self-reported rates of aggression were assessed using 10 items from the

aggression scale (a = .92) of the Youth Self Report (Achenbach & Edelbrock, 1986). Students

were asked how many times in the past month they had engaged in incidents of aggressive

behavior. Items included "Yelled at someone (you were mad at)," "Told someone off," "Pushed

or shoved someone on purpose," and "Hit someone." Response categories were on a 5-point

scale. Response options included 1 (Never), 2 (Once), 3 (2-3 times), 4 (4-5 times), and 5 (More

than 5 times). Responses were rescored to a zero-baseline and then summed to produce an

overall aggression score with a range of 0-40.

Delinquency. Delinquent acts were measured using a 10-item scale (Elliot, Huizinga, &

Menard, 1989, a = .86) that asked how often students had engaged in delinquent behaviors in the

past year. Examples included "purposely damaged or destroyed property or things that did not

belong to you" and "taken something from a store when a clerk wasn't looking." Response

options included 1 (Never), 2 (Once), 3 (2-3 times), 4 (4-5 times), and 5 (More than 5 times).

Responses were rescored to a zero-baseline and then summed to produce an overall delinquency

score with a possible range of 0-40.

Friends' delinquency. Friends' participation in delinquent acts was measured using a

seven-item scale (Elliot et al., 1989, a = .88) that assessed how many of the participants' friends









had engaged in delinquent behaviors in the past year. Examples included "ruined or damaged

something on purpose that wasn't theirs" and "broken into some place to steal something."

Response options referred to the portion of the participant's friends and included 1 (None), 2

(Less than half), 3 (About half), 4 (More than half), and 5 (All or almost all). Responses were

rescored to a zero-baseline and then summed to produce an overall friends' delinquency score

with a possible range of 0-40.

Demographic Characteristics

For facilitation of regression analyses, single mother households were collapsed into the

"other" group of family structure and used as a comparison group for the two-parent variable

(Nichols et al., 2006).









CHAPTER 3
RESULTS

Table 3-1 presents full sample means and standard deviations for each of the seven

continuous variables employed in this study. Tests of normality were run on the distribution of

each variable. Because hypothetical aggressive responses, self-reported aggression, delinquency,

and friends' delinquency were positively skewed (Zskewness = 6.86 to 17.88), these variables were

transformed using the square root function in all subsequent analyses. Distributions of other

variables did not require transformation.

Demographic Associations

Table 3-2 presents bivariate correlations between study variables. Among demographic

variables, family structure was significantly associated with all three antisocial variables; living

in a two-parent, non-blended family served as a protective factor against exhibiting high rates of

antisocial behavior and against associating with antisocial peers. Being African-American was

positively correlated with higher rates of aggressive behavior, and being male was positively

correlated with higher rates of delinquent behavior, but both correlations were small in size,

indicating no meaningful association. It is interesting to note that no gender differences existed

for aggressive behavior in this sample. In addition, attending public school (as opposed to

parochial school) was associated with an increased rate of hostile attributions.

Because of the categorical nature of each demographic variable, chi-square analyses were

conducted as a further test of demographic associations. Family structure and school type were

associated such that participants who lived in two-parent families were less likely to attend

public school (x2 = 18.80, p < .001). Family structure and school type were also both confounded

with race: African-American participants were less likely to live in two-parent, non-blended

families, and public schools contained greater proportions of African-American students and









smaller proportions of Caucasian participants than would be expected by chance (x2 = 64.27, p <

.001). Gender was mildly associated with being African-American or Latino (there were more

African-American girls than boys and more Latino boys than girls). The gender differences were

considered an anomaly of the sample and not relevant to any other findings.

Associations Among Core Constructs

As expected, the three outcome variables (aggression, delinquency, and friends'

delinquency) were highly colinear, and aggression and delinquency especially exhibited similar

relationships with hypothetical and emotion-related variables. Also as predicted, hypothetical

aggressive responses were related to reported aggression, though this association was weak, r =

.16, p < .01. Hypothetical aggressive responses were also weakly associated with increased rates

of delinquent behavior and, to a lesser extent, association with delinquent friends.

Surprisingly, hostile attribution of intent was not significantly associated with

hypothetical aggressive responses or with reported rates of aggression, delinquency or having

delinquent friends. In addition, hostile attributions were not correlated with either anger variable.

One of the hypotheses of this study was that emotional processes would mediate the relationship

between hostile attribution and aggression. However, because no significant correlations existed

between hostile attribution and any of the emotion and aggression variables, this model was not

examined, and hostile attribution was dropped from subsequent analyses.

Anger levels and anger reduction were not meaningfully associated, r = -.09, p = .07. The

small magnitude of this relationship supports the conceptual distinction between emotion arousal

and emotion regulation (Eisenberg et al., 2005). As expected, both anger variables were

moderately correlated with all antisocial variables. High anger levels were associated with

greater rates of aggression, delinquency, and delinquent friends, while good anger reduction

skills predicted lower rates of each.









A hierarchical multiple regression analysis was conducted to test the overall variance

explained by both hypothetical aggressive responses and anger-related variables in rates of

reported aggression, delinquency, and delinquent friends. Because of their associations with one

or more of the outcome variables, gender, family structure (two-parent family vs. other), and race

(African-American vs. other) were entered into all models in step 1 as covariates. School type

was dropped as a covariate due to its lack of association with any outcome variable (Stevens,

1996). Due to the fact that correlational analyses showed anger processes to be more strongly

associated with antisocial behavior and delinquent friends than were hypothetical aggressive

responses, the model was structured to first account for the combined association of anger and

anger regulation and then test further contributions to the relationship explained by hypothetical

aggressive responses. Accordingly, in each regression, anger and anger regulation were entered

in step 2; and hypothetical aggressive responses were entered in step 3.

Tables 3-3, 3-4, and 3-5 present the results of these analyses. After controlling for gender,

family structure, and race, anger and anger regulation both remained significant predictors of all

three outcome variables, explaining 23% of the total variance in reported aggression, 20% of the

total variance in reported delinquency, and 14% of the total variance in delinquent friends.

Hypothetical aggressive responses also remained a significant predictor of each outcome variable

upon addition to the model; however, it only explained an additional 1% of variance in

aggression, 3% of variance in delinquency, and 1% of variance in delinquent friends. Reversing

the order of steps 2 and 3, so that hypothetical responses were entered into the regression before

anger variables, did not substantially change the variance explained by any of the predictors.

Based on these analyses, it appears that simply examining levels of hostile attributions or

aggressive responses in these hypothetical vignettes provided little information about antisocial









behavior occurring in real life, and that emotional characteristics like anger levels and regulation

may be better indicators of antisocial behavior in normative samples. However, hypothetical

scenarios such as the ones used in this study continue to be widely used in research on child and

adolescent samples. Hence, additional analyses were conducted to further explore potential

associations of hypothetical vignette responses with antisocial behaviors in real life.

Attribution-Response Patterns

One approach that goes beyond simply investigating overall levels of attributions and

responses is to instead look at differences in the patterns of responses that participants give to

individual scenarios. For instance, in any given sample, only a portion of participants will

respond aggressively to a particular scenario, and of those that do, not everyone will have

attributed hostility to the provocateur. It is possible that separating participants into groups based

on response patterns will uncover information about differences in social information-processing

and real-life behavior even when hostile attributions show no association with hypothetical

responses in the sample as a whole. This approach allowed for the examination of situational

characteristics of the vignettes that may have resulted in different attribution or response

outcomes from vignette to vignette.

For each vignette, four attribution-response pattern groups were created: nonhostile-

nonaggressive, nonhostile-aggressive, hostile-nonaggressive, and hostile-aggressive. Group

distributions are displayed in Table 6. A chi-square analysis comparing group membership in the

lunch table vignette to the hallway vignette was not significant (x2 = 14.22, p < .115), indicating

that membership in a particular group did not stay stable across the vignettes.

A multivariate analysis of covariance was conducted for each vignette using group

membership as the independent variable and a multivariate construct of "antisocial tendencies,"

consisting of aggression, delinquency, and friends' delinquency, as the dependent variable.









Gender, family structure, and race were included as covariates in each analysis, and for

exploratory purposes, interactions between pattern group membership and each covariate were

included in the analyses as well. Games-Howell error corrections were employed for post-hoc

tests due to the large disparities in group sizes.

Lunch Table

The multivariate ANCOVA for the lunch table vignette groups was significant, Wilks' X

=.95, F(9, 961.48)= 2.09, p < .05, partial r2 = .02, as were univariate tests for each individual

outcome variable (Table 7). Post-hoc means comparisons revealed that the nonhostile-aggressive

group had significantly higher rates of aggression, delinquency, and association with delinquent

friends than did the nonhostile-nonaggressive group (mean differences = .34 to .47 SD; Figure

3-1).

In addition to expected multivariate main effects of family structure (Wilks' X = .96,

F(3,395)= 5.22,p < .01, partial f2 = .04) and race (Wilks' = .97, F(3,395) = 4.66, p < .01,

partial r2 = .03), a multivariate effect also existed for the interaction between lunch table group

membership and race, Wilks' X= .93, F(9, 961.48) = 3.50, p < .001, partial r2 = .03. Univariate

tests revealed that the multivariate effect was driven by a significant interaction effect for

delinquency only (F[3, 397] = 4.70, p < .01, partial r2 = .03), such that within non-African

Americans, nonhostile-aggressive participants were significantly higher in delinquency than all

other groups; however, this pattern did not hold for African Americans (Figure 3-2). One issue is

that large standard errors existed for the hostile-aggressive groups in particular, possibly due to

their small size (N= 10 for African Americans; N= 11 for non-African Americans), which may

have prevented further significant differences from becoming evident.









Hallway

The multivariate ANCOVA for the hallway vignette groups was not significant (Wilks'

S= .97, F(9, 961.48) = 1.51, p = .140, partial r2 = .01), and its pattern of results did not replicate

those of the lunch table scenario (Figure 1). Follow-up univariate ANCOVAs were not

significant for any of the three outcome variables. In addition, no group membership x

demographic variable interactions existed for the hallway vignette.










Table 3-1 Summary statistics on continuous variables
Measure M SD
Hostile attributions .84 .64
Hypothetical aggressive responses .49 .55
Anger level 2.50 .90
Anger reduction 2.79 1.03
Self-reported aggression 12.97 10.18
Delinquency 3.73 4.70
Friends' delinquency 4.17 4.81











Table 3-2 Correlations among


1 2 3 4 5 6
1 Gender
2 Public school .00
3 African Americana -.10* .34**
4 Caucasiana .04 -.29** n/a
5 Latinoa .10* -.03 n/a n/a
6 Two-parent, non-blended .07 -.14** -.26** .19** .09
family
7 Hostile attributions .01 .14** -.02 .01 .04 -.03
8 Hypothetical aggressive -.01 -.01 .05 .01 -.10* .01
responses
9 Anger -.05 -.01 -.03 .05 .01 -.01
10 Anger regulation -.06 -.05 -.03 .05 .01 .03
11 Reported rates of aggression .04 .04 .11* -.03 -.09 -.15**
12 Delinquency .10* .05 .08 -.08 -.04 -.15**
13 Friends' delinquency .01 .06 .07 -.09 .03 -.19**
Note. For student sex, 1 = boys. p < .05. **p < .01. a Correlations were not applicable due to
covariance caused by variable coding.


study variables











Table 3-2 Continued
7 8 9 10 11 12
1 Gender
2 Public school
3 African Americana
4 Caucasiana
5 Latino
6 Two-parent, non-blended
family
7 Hostile attributions
8 Hypothetical aggressive .04
responses
9 Anger -.02 .06
10 Anger regulation -.03 -.08 -.09
11 Reported rates of aggression -.02 .16** .40** -.31**
12 Delinquency -.02 .20** .38** -.29** .75**
13 Friends' delinquency -.07 .13* .33** -.22** .58** .61**











Table 3-3 Summary of hierarchical regression analysis for variables predicting
reported rates of aggression


Variable
Step 1
Gender
Family structure
Race
Step 2
Gender
Family structure
Race
Anger
Anger reduction
Step 3
Gender
Family structure
Race
Anger
Anger reduction
Hypothetical aggressive responses
Note. ** p <.01.


AR
.03**


AR2


.26** .23**


.27** .01**


B SEB


.17 .15
-.40 .15
.23 .15

.17 .13
-.35 .13
.26 .13
.62 .07
-.38 .06

.17 .13
-.36 .13
.24 .13
.61 .07
-.37 .06
.32 .12


.06
-.13**
.08

.06
-.12**
.09
.38**
-.27**

.06
-.12**
.08
.37**
-.26**
.12**










Table 3-4 Summary of hierarchical regression analysis for variables predicting reported
rates of delinquency
Variable AR AR2 B SEB 3
Step 1 .04**
Gender .28 .12 .11*
Family structure -.35 .12 -.14**
Race .14 .12 .06
Step 2 .24** .20**
Gender .28 .11 .12**
Family structure -.31 .11 -.13**
Race .16 .11 .07
Anger .49 .06 .36**
Anger reduction -.28 .05 -.24**
Step 3 .27** .03**
Gender .28 .10 .12**
Family structure -.32 .11 -.13**
Race .14 .11 .06
Anger .48 .06 .35**
Anger reduction -.27 .05 -.23**
Hypothetical aggressive responses .35 .10 .16**
Note. ** p <.01.










Table 3-5 Summary of hierarchical regression analysis for variables predicting reported
rates of delinquent friends
Variable AR AR2 B SE B
Step 1 .04**
Gender .05 .12 .02
Family structure -.44 .12 -.18**
Race .05 .12 .02
Step 2 .18** .14**
Gender .05 .11 .02
Family structure -.41 .11 -.17**
Race .07 .11 .03
Anger .42 .06 .31**
Anger reduction -.21 .05 -.18**
Step 3 .19** .01*
Gender .05 .10 .02
Family structure -.41 .11 -.17**
Race .06 .11 .02
Anger .41 .06 .31**
Anger reduction -.20 .05 -.18**
Hypothetical aggressive responses .21 .10 .09*
Note. *p<.05. **p< .01.










Table 3-6 Attribution-response pattern group membership
Lunch table scenario Hallway scenario
Group N % N %
Nonhostile-nonaggressive 256 62.0 105 25.4
Nonhostile-aggressive 79 19.1 41 9.9
Hostile-nonaggressive 57 13.8 201 48.7
Hostile-aggressive 21 5.1 66 16.0










Table 3-7 Univariate ANOVAs for lunch table groups
Aggression Delinquency Delinquent Friends
Source df F partial n2 df F partial 12 df F partial 12


Group Membership (GM)
Covariates
Gender

Family Structure
Race
Covariate Interactions


3 2.82*

1 .01


3 5.29**

1 2.64


7.20**
3.78


GMx Gender 3 .38 .00 3
GM x Family Structure 3 .90 .00 3
GMx Race 3 1.98 .02 3
Error 397 (2.13) 397
Note.Values in parentheses represent mean square errors. *p < .05. **p < .01.


3 2.75*

1 .47


7.54**
5.10*

.10
.71
4.70**
(1.35)


15.43**
.71

1.32
1.57
1.26
(1.38)














A 1.0
0.8
0.6
S0.4
S0.2
S0.0
-0.2
S-0.4
-0.6
-0.8
-1.0


0 Lunch table
SHallway


Nonhostile Nonaggressive Nonhostile Aggressive Hostile NonAggressive Hostile Aggressive


B 1.0
0.8
S0.6
S0.4
0.2
0.0
= -0.2
S -0.4
S-0.6
-0.8
-1.0





C 1.0
S0.8
0.6
4 0.4
0 o.2
0.0
-0.2
S-0.4
S-0.6
-0.8
-1.0


Nonhostile Nonaggressive Nonhostile Aggressive Hostile NonAggressive Hostile Aggressive


0 Lunch table
* Hallway


















0 Lunch table
* Hallway


Nonhostile Nonaggressive Nonhostile Aggressive Hostile NonAggressive Hostile Aggressive


Figure 3-1 Attribution-response pattern differences in antisocial outcome variables for lunch
table and hallway vignettes. Error bars represent 95% confidence intervals. A)
Attribution-response patterns for aggression. B) Attribution-response patterns for
delinquency. C) Attribution-response patterns for delinquent friends.













1.0
0.8 o
S0.6
S0.4
0. &2s OAfrican
American
0.0
S-0.2 Other

-0.4 -
-0.6
-0.8
-1.0
Nonhostile No nho s tile Hostile Ho s tile Aggre s s ive
Aggres sive Aggressive Nonaggres sive



Figure 3-2 Lunch table attribution-response pattern differences in delinquency by race. Error
bars represent 95% confidence intervals.









CHAPTER 4
DISCUSSION

This study investigated the relationship of anger and anger regulation to the associations

between aggressive responses to hypothetical social scenarios and reported rates of aggression in

real life. Because previous research has tended to focus on the application of SIP-based

hypothetical research techniques to real-life aggression only, delinquent behavior and association

with delinquent friends were also included as outcome variables of interest in order to give a

fuller picture of real-life antisocial behavior. A moderate, yet significant, link was demonstrated

between hypothetical aggression and the three kinds of reported antisocial tendencies. This

finding is similar to previous research using hypothetical and role-play scenarios to assess

aggression (e.g., Tremblay & Ewart, 2005). However, both anger and anger regulation had

stronger associations with all three antisocial variables than did hypothetical aggressive

responses, and these two emotion processes jointly accounted for nearly all of the explained

variance in reported aggression, delinquency, and delinquent friends.

Combined with the finding that neither emotional process was significantly associated

with hypothetical aggressive responses, the results of this study lend support to Lemerise &

Arsenio's (2000) reformulation of the social-information-processing model and suggest that

hypothetical social problem-solving tasks may not capture something crucial about the emotion

processes that contribute to aggression in real life. This supports previous work proposing that

children can "know" appropriate responses to hypothetical situations, but that characteristics

about real-life encounters which differ from hypothetical scenario may generate an impulsive

response which prohibits this knowledge from being implemented (Vitaro & Pelletier, 1991).

Therefore, future researchers using these types of tasks to investigate or evaluate aggressive

behavior need to take their potential limitations into consideration.









In addition, this study assessed the prevalence of hostile intent attributed to the

provocateur in these problem-solving tasks and its association with anger processes and

aggression in both the hypothetical scenarios and in real life. Unexpectedly, hostile attributions

were not related to any antisocial outcome variable. More surprisingly, they were also not related

to hypothetical aggressive responses, which were assessed using the same vignettes. This result

was a departure from prior research on hostile attributions, which has demonstrated a

pronounced association with aggressive behavior (Orobio de Castro et al., 2002). However,

much of this work has compared highly aggressive, clinically referred or incarcerated samples to

less aggressive control groups. It is possible that within normal spectrums of aggressive behavior

in non-referred samples, the SIP model of hostile attributions -- aggressive behavior is not an

accurate representation of cognitive processes taking place during these scenarios.

Hostile attributions were also not associated with either of the two anger variables. The

questions contained in the Buss and Perry (1992) anger scale conceptually assess trait anger, or

enduring tendencies to become angry across contexts (Spielberger et al., 1983), rather than state

anger, which is transiently aroused and assessed during specific situations. The specific

association between trait anger and hostile attribution biases in adolescents has not been

conclusively studied, although research in adults has demonstrated a link between the two (Epps

& Kendall, 1995). Orobio de Castro, Merk, Koops, Veerman and Bosch (2005) found a mild

correlation between emotion regulation skills and interpretation of intent in hypothetical

vignettes, but their sample included a subset of highly aggressive boys referred for behavior

problems. It is possible that the results of the current study reflect the existence of a weaker

association between these constructs in non-clinical, mixed-gender adolescent samples.









An alternative explanation for our findings is that the failure to replicate previous

research on the association between hostile attribution and aggression, as well as the weakness of

the relationship observed between hypothetical aggression and real-life antisocial behavior, may

be due to the small number of hypothetical scenarios assessed in this study (two), resulting in a

restricted range of possible response scores. Prior work on hostile attribution and aggression

typically has presented 7 to 12 scenario vignettes to each participant (e.g., Crain et al., 2005;

Orobio de Castro et al., 2005). Adding more scenarios to the continuous measures used in this

study might have resulted in stronger statistical associations between hostile attribution and other

constructs. However, other research using small sets of scenarios suggests that the nonsignificant

relationship observed between hostile attribution and reported antisocial tendencies may not just

be a result of low variance in the measures. Vitaro and Pelletier (1991) presented children with

four hypothetical vignettes of ambiguous provocation and three actual ambiguous provocations

by a trained peer-confederate. Results indicated that children classified as maladjusted (i.e.,

aggressive and rejected by peers) responded to the peer-confederate provocations with more

verbal attacks than did well-adjusted participants; however, no difference was seen between the

two groups for the hypothetical vignettes. Additional research by these authors (Vitaro, Pelletier,

& Contu, 1989) found that maladjusted children also attributed more negative intentions to

provocateurs in peer-confederate interactions than in hypothetical vignettes. This research

suggests that, small measure set or not, the lack of association in the current study between

hostile attributions and reported rates of aggression may be a reflection of the larger issue of

emotional processing differences between hypothetical scenarios and real-life situations.

Though a surprisingly small relationship was found between hypothetical and real-life

tendencies when responses were summed across vignettes, one of the benefits of hypothetical









vignette research methods remains the ability to investigate attributions and responses as they

relate to a specific hypothetical incident. Because of the importance of the outcomes which these

studies are trying to accomplish the identification and prevention of youth aggression (e.g.,

Hudley & Graham, 1993) it may be worthwhile to reexamine the predictive relevance of these

types of hypothetical measures through additional methodological techniques. It was therefore of

interest to determine whether any useful predictive information about real-life antisocial

behavior could be obtained from investigating attribution-response patterns within individual

vignettes. This attribution-response pattern approach has been underutilized in research to date:

SIP research using group comparisons has usually based group membership on levels of real-life

aggression or participant characteristics such as peer rejection (Lansford et al., 2006; Losel,

Bliesener, & Bender, 2007; Orobio de Castro et al., 2002).

Separating participants into groups based on their attribution-response patterns (hostile or

nonhostile; aggressive or nonaggressive) yielded significant differences between the nonhostile-

nonaggressive and nonhostile-aggressive groups in the lunch table vignette, such that nonhostile-

aggressive participants (approximately 19% of the sample) were higher in aggressive and

delinquent behavior and reported that they had more delinquent friends. An interaction effect

between group membership and race further suggested that this pattern of results was mainly due

to higher rates of antisocial tendencies among the non-African American members of this group.

Nonhostile-aggressive response patterns may represent a particularly problematic group of

adolescents, who seem to be arriving at aggressive solutions to social interactions without having

a "reason" to do so (i.e., feeling threatened by the perceived hostility of a peer).

However, no such relationship was found for the hallway scenario, in which attribution-

response group membership did not yield any significant predictive information about antisocial









tendencies in real life. The wide difference in pattern distribution between the two vignettes,

especially in attributions of hostility, is likely explained by the fact that while both vignettes

could be classified as ambiguous, they differed in their levels of ambiguity: In the hallway

scenario, while it remains unclear whether the provocateur was addressing his/her remarks to the

participant or to another student, the actual statement the participant makes ("Hey, geek. Yeah, I

mean you, nerd") contains more obvious hostility than the statement made in the lunch table

scenario ("You can't sit there, that seat's taken"). As such, the majority of participants thought

that the hallway provocateur was hostile, while the lunch table provocateur was not. The

ambiguity level of the vignettes may be an important predictor of SIP processing in normative,

nonreferred individuals, and as such may be a useful inclusion in future research using these

populations.

Further investigation of antisocial behavior differences among attribution-response

pattern groups should include possible personality measures which might account for

participants' differing reactions to the hypothetical vignettes. For instance, empathy- and fear-

related characteristics might explain why more antisocial adolescents in this sample tended to act

aggressively in the absence of perceived threat in the lunch table scenario (Andershed, Kerr,

Stattin, & Levander, 2002). Social measures such as peer rejection or the desire to increase one's

social status may also play a role in adolescents' decisions to act aggressively in the absence of

hostility (Guerra, Asher, & DeRosier, 2004; Rose, Swenson, & Waller, 2004).1 These variables

could potentially be incorporated at various steps of the SIP model to explain why attributions of

hostile intent do not usually result in aggressive responses in this sample.

Hypothetical tasks like the ones described in this study are often used in prevention and

intervention efforts for youth with behavior problems (Sukhodolsky, Golub, Stone, & Orban,









2005). A greater understanding of the relationship between hypothetical and real-life responses

to stressful social situations, as well as the factors that may differ between the two, may help to

improve the efficacy of these programs. However, it is important to note that effect sizes of the

attribution-response pattern analyses were still quite small, mimicking the amount of variance

explained when continuous versions of the hypothetical variables were analyzed. The utility of

hypothetical scenarios at telling us about actual antisocial behavior is expected to be moderate at

best, suggesting the need to consider other characteristics of adolescents in addition to their

performance in these tasks.

An additional limitation of the study may be that the distinction between reactive

aggression (aggression in response to provocation) and proactive aggression (aggressive

behavior for instrumental gain) was not assessed. Though the two are correlated, reactive

aggression has been shown to have a much stronger relationship with hostile attribution of intent

(Crick & Dodge, 1996). Assessing rates of real-life aggression in this manner may have

contributed to a more detailed understanding of the effects of hostile attribution in this study, and

any future research on this topic should take the distinction between subtypes of aggression into

consideration.

While it was not the initial purpose of this study to investigate demographic differences

in aggression, hostile attributions or emotion processes, the lack of gender effects in these

constructs is of interest. Traditionally, boys have been viewed as more aggressive than girls;

however, a growing amount of research suggests that this is difference is narrowing (Odgers &

Moretti, 2002). Although the inclusion of gender in this study was to serve as a possible control

variable, more research is needed to specifically investigate the presence or absence of gender

differences in social-emotional-cognitive pathways of this type.









The sample used in this study was drawn exclusively from schools within an urban

setting. Aggression levels are higher in urban settings relative to rural settings (Farrel, Sullivan,

Esposito, Meyer, & Valois, 2005), which may reflect differences in cultural views about the

appropriateness of aggression in certain contexts (Tolan, Gorman-Smith, & Henry, 2003).

However, the fact that anger variables showed a stronger association with antisocial tendencies

than did hypothetical variables implies that emotion processes are contributing to variability in

rates of antisocial behavior over and above the implementation of what may be considered a

culturally-appropriate retaliatory response. Still, the results of this study may not be

generalizable to other samples of early adolescents due to particular characteristics of the urban

setting.

The current research is thought to contribute to the collective understanding of hostile

attributions and anger processes and their role in aggression in early adolescence. In addition, the

results of this study have implications for the future use of hypothetical vignettes to assess

aggression in individuals of this age. Analysis of pattern groups, in particular, is a

methodological technique that can easily be applied to existing data to evaluate and compare the

predictive efficacy of individual vignettes. Besides the inclusion of samples with different

demographic characteristics, research in this area should assess the relationships of these core

constructs over time.









APPENDIX
HYPOTHETICAL VIGNETTE SCRIPTS

Task: Lunch Table

Instruction: "Imaging that you get your lunch at school and then walk over to a table. You want

to sit at this table. Several other kids are already seated there and there is one empty seat.

As you begin to sit down, one of the other kids says, 'You can't sit there. It's taken.' A

couple of other kids laugh."

Question 1: "So you were not able to sit at the table. What do you think was going on in the

minds of the kids at the table when this happened?"

Question 2: "What would you do or say to the other kids if this happened to you?

Task: Hallway

Instruction: "Imagine that you are walking down the hallway at your school with two other kids

on the way to lunch when you see another boy/girl coming toward the three of you from

the other end of the hallway. There are lots of kids in the hallway. This other kid yells

out, 'Hey, geek. Yeah, I mean you, nerd.' Some of the other kids start laughing."

Question 1: "So some kids are laughing. What do you think was going on in the mind of the

boy/girl when he/she said this?"

Question 2: "What would you do or say to the boy/girl if this happened to you?"









LIST OF REFERENCES


Achenbach, T. M., & Edelbrock, C. S. (1986). Teacher's report form. Burlington, VT.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173-1182.

Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive-
neoassociationistic analysis. American Psychologist, 45, 494-503.

Borbely, C. J., Graber, J. A., Nichols, T., Brooks-Gunn, J., & Botvin, G. J. (2005). Sixth graders'
conflict resolution in role plays with a peer, parent, and teacher. Journal of Youth and
Adolescence, 34, 279-291.

Bosson, J. K., & Johnson, A. B. (2006). Interpersonal chemistry through negativity: Bonding by
sharing negative attitudes about others. Personal Relationships, 13, 135-150.

Botvin, G. J., Schinke, S. P., Epstein, J. A., & Diaz, T. (1994). Effectiveness of culturally
focused and generic skills training approaches to alcohol and drug abuse prevention among
minority youths. Psychology of Addictive Behavior, 8, 116-127.

Buss, A. H., & Perry, M. (1992). The aggression questionnaire. Journal ofPersonality and
Social Psychology, 63, 452-459.

Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon & N.
Eisenberg (Eds.), Handbook of childpsychology: Vol. 3. Social, emotional, and personality
development (5th ed., pp. 779-862). New York: John Wiley & Sons.

Cornell, D. G., Peterson, C. S., & Richards, H. (1999). Anger as a predictor of aggression among
incarcerated adolescents. Journal of Consulting and Clinical Psychology, 67, 108-115.

Crain, M. M., Finch, C. L., & Foster, S. L. (2005). The relevance of the social information
processing model for understanding relational aggression in girls. Merrill-Palmer
Quarterly, 51, 213-249.

Crick, N. R. (1997). Engagement in gender normative versus nonnormative forms of aggression:
Links to social-psychological adjustment. Developmental Psychology, 33, 610-617.

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.

Crick, N. R., & Dodge, K. A. (1996). Social information-processing mechanisms in reactive and
proactive aggression. Child Development, 67, 993-1002.

Damasio, A. R. (1994). Descartes' error: Emotion, reason, and the human brain. New York:
Avon Books.









Dearing, K. F., Hubbard, J. A., Ramsden, S. R., Parker, E. H., Relyea, N., Smithmyer, C. M., &
Flanagan, K. D. (2002). Children's self-reports about anger regulation: Direct and indirect
links to social preference and aggression. Merrill-Palmer Quarterly, 48, 308-336.

Dodge, K. A. (1980). Social cognition and children's aggressive behavior. Child Development,
51, 162-170.

Dodge, K. A. (1986). A social information processing model of social competence in children. In
M. Perlmutter (Ed.), Minnesota Symposium in Child Psychology (Vol. 18, pp. 77-125).
Hillsdale, NJ: Earlbaum.

Dodge, K. A. (1990). The Child Development Project. Nashville, TN: Vanderbilt University.

Dodge, K. A., & Coie, J. D. (1987). Social information-processing factors in reactive and
proactive aggression in children's peer groups. Journal ofPersonality and Social
Psychology, 53, 1146-1158.

Dodge, K. A., & Frame, C. L. (1982). Social cognitive biases and deficits in aggressive boys.
Child Development, 53, 620-635.

Eisenberg, N., Cumberland, A., Spinrad, T. L., Fabes, R. A., Shepard, S. A., Reiser, M., Murphy,
B. C., Losoya, S. H., Guthrie, I. K. (2001). The relations of regulation and emotionality to
children's externalizing and internalizing problem behavior. Child Development, 72,
1112-1134.

Eisenberg, N., Fabes, R. A., Guthrie, I. K., Murphy, B. C., Maszk, P., Holmgren, R., & Suh, K.
(1996). The relations of regulation and emotionality to problem behavior in elementary
school children. Development and Psychopathology, 8, 141-162.

Eisenberg, N., Morris, A. S., & Spinrad, T. L. (2005). Emotion-related regulation: The construct
and its measurement. In D. M. Teti (Ed.), Handbook of research methods in developmental
science (pp. 81-100). Malden, MA: Blackwell Publishers.

Elliot, D., Huizinga, D., & Menard, S. (1989). Multiple problem youth: Delinquency, substance
use, and mental health problems. New York: Springer-Verlag.

Epps, J., & Kendall, P. C. (1995). Hostile attributional bias in adults. Cognitive Therapy and
Research, 19, 159-178.

Epstein, J. A., Botvin, G. J., Diaz, T., Baker, E., and Botvin, E. M. (1997). Reliability of social
and personal competence measures for adolescents. PsychologicalReport, 81, 449-450.

Evans, R. I., Hansen, W. B., & Mittelmark, M. B. (1977). Increasing the validity of self-reports
of behavior in a smoking in children investigation. Journal ofApplied Psychology, 62,
521-523.









Farrell, A. D., Sullivan, T. N., Esposito, L. E., Meyer, A. L. & Valois, R. F. (2005). A latent
growth curve analysis of the structure of aggression, drug use, and delinquent behaviors
and their interrelations over time in urban and rural adolescents. Journal ofResearch on
Adolescence, 15, 179-204.

Graber, J. A., Nichols, T. R., Luppino, C., Galen, B., Brooks-Gunn, J., Schinke, S., & Botvin, G.
J. (2001). Life .\kill Training: Coding Manual for Videotape Role-Plays. Center for
Children & Families, Teachers College, Columbia University, New York.

Guerra, V. S., Asher, S. R., & DeRosier, M. E. (2004). Effect of children's perceived rejection
on physical aggression. Journal ofAbnormal Child Psychology, 32, 551-563.

Guerra, N. G., Huesmann, L. R., & Zelli, A. (1990). Attributions for social failure and aggression
in incarcerated delinquent youth. Journal ofAbnormal Child Psychology, 18, 347-355.

Hudley, C., & Graham, S. (1993). An attributional intervention to reduce peer-directed
aggression among African-American boys. Child Development, 64, 124-138.

Lansford, J. E., Malone, P. S., Dodge, K. A., Crozier, J. C., Pettit, G. S., & Bates, J. E. (2006). A
12-year prospective study of patterns of social information processing problems and
externalizing behaviors. Journal ofAbnormal Child Psychology, 34, 715-724.

Lengua, L. J. (2003). Associations among emotionality, self-regulation, adjustment problems,
and positive adjustment in middle childhood. AppliedDevelopmental Psychology, 24,
595-618.

Lemerise, E. A., & Arsenio, W. F. (2000). An integrated model of emotion processes and
cognition in social information processing. ChildDevelopment, 71, 107-118.

Losel, F., Bliesener, T., & Bender, D. (2007). Social information processing, experiences of
aggression in social contexts, and aggressive behavior in adolescents. Criminal Justice and
Behavior, 34, 330-347.

Moffit, T. E. (1993). Adolescence-limited and life-persistent antisocial behavior: A
developmental taxonomy. Psychological Review, 100, 674-701.

Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in antisocial
behavior. Cambridge: Cambridge University Press.

Nichols, T. R., Graber, J. A., Brooks-Gunn, J., & Botvin, G. J. (2006). Sex differences in overt
aggression and delinquency among urban minority middle school students. Applied
Developmental Psychology, 27, 78-91.

Nichols, T. R., Graber, J. A., Byrne, C., Luppino, C., Reigada, L., Caskey, E., Brooks-Gunn, J.,
& Botvin, G. J. (2001). Life .\kill Training: Coding Manual for Social Problem Solving:
Adolescent Version. Center for Children & Families, Teachers College, Columbia
University, New York.









Novaco, R. B. (1994). Anger as a risk factor for violence among the mentally disordered. In J.
Monahan & H. J. Steadman (Eds.), Violence and mental disorder: Developments in risk
assessment (pp. 21-59). Chicago: University of Chicago Press.

Odgers, C. L., & Moretti, M. M. (2002). Aggressive and antisocial girls: Research update and
challenges. International Journal ofForensic Mental Health, 1(2), 103-119.

Orobio de Castro, B., Merk, W., Koops, W., Veerman, J. W., & Bosh, J. D., (2005). Emotions in
social information processing and their relationships with reactive and proactive
aggression in referred aggressive boys. Journal of Clinical Child and Adolescent
Psychology, 34, 105-116.

Orobio de Castro, B., Veerman, J. W., Koops, W., Bosch, J. D., & Monshouwer, H. J. (2002).
Hostile attribution of intent and aggressive behavior: A meta-analysis. Child Development,
73, 916-934.

Rose, A. J., Swenson, L. P., & Waller, E. M. (2004). Overt and relational aggression and
perceived popularity: Developmental differences in concurrent and prospective relations.
Developmental Psychology, 40, 378-387.

Saarni, C. (1999). The development of emotional competence. New York: Guilford.

Saltaris, C. (2002). Psychopathy in juvenile offenders: Can temperament and attachment be
considered as robust developmental precursors? Clinical Psychology Review, 22: 729-52.

Spielberger, C. D., Jacobs, G. A., Russell, S. F., & Crane, R. J. (1983). Assessment of anger: The
State-Trait Anger Scale. In J. Butcher & C. D. Spielberger (Eds.), Advances in personality
assessment (2nd ed., pp. 159-187). Hillsdale, NJ: Erlbaum.

Stevens, J. (1996). Applied multivariate statistics for the social sciences. Mahwah, NJ: Lawrence
Erlbaum Associates, 362-392.

Sukhodolsky, D. G., Golub, A., Stone, E. C., & Orban, L. (2005). Dismantling anger control
training for children: A randomized pilot study of social problem-solving versus social
skills training components. Behavior Therapy, 36, 15-23.

Tolan, P. H., Gorman-Smith, D., & Henry, D. B. (2003). The developmental ecology of urban
males' youth violence. Developmental Psychology, 39, 274-291.

Tourangeau, R., Smith, T. W., & Rasinski, K. A. (1997). Motivation to report sensitive behaviors
on surveys: Evidence from a bogus pipeline experiment. Journal of Applied Social
Psychology, 27, 209-222.

Tremblay, P. F., & Ewart, L. A. (2004). The Buss and Perry Aggression Questionnaire and its
relations to values, the Big Five, provoking hypothetical situations, alcohol consumption
patterns, and alcohol expectancies. Personality andIndividualDifferences, 38, 337-346.









VanOostrum, N., & Horvath, P. (1997). The effects of hostile attribution on adolescents'
aggressive responses to social situations. Canadian Journal of School Psychology, 13,
48-59.

Vitaro, F., & Pelletier, D. (1991). Assessment of children's social problem-solving skills in
hypothetical and actual conflict situations. Journal ofAbnormal Child Psychology, 19,
505-518.

Vitaro, F. Pelletier, D., & Contu, S. (1989). Impact of a negative social experience on the social
reasoning process of aggressive-rejected children. Perceptual andMotor .\k//1, 69,
371-382.

Zeman, J., Shipman, K, & Suveg, C. (2002). Anger and sadness regulation: predictions to
internalizing and externalizing symptoms in children. Journal of Clinical Child &
Adolescent Psychology, 31, 393-398.









BIOGRAPHICAL SKETCH

Katherine Clemans is a third-year graduate student in developmental psychology at the

University of Florida. She received a Bachelor of Arts in psychology and a certificate in human

development from Duke University in 2002. She has also completed work toward a Master of

Arts in Liberal Studies degree from Dartmouth College. Her research focuses on emotional

correlates of aggression and moral judgment in adolescents and young adults. Currently she is a

co-investigator of the APEX (Adolescent Peer Experiences) Study, which examines peer

relationships and their influence on psychosocial development during the middle school years.





PAGE 1

THE ROLES OF EMOTION PROCESSES AND HYPOTHETICAL SCENARIOS IN PREDICTING ANTISOCIAL BEHAVIOR By KATHERINE HALE CLEMANS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007 1

PAGE 2

2007 Katherine Hale Clemans 2

PAGE 3

ACKNOWLEDGMENTS This research was sponsored by a grant to Dr. Gilbert J. Botvin from the National Institute on Drug Abuse (P50DA-07656). Thanks are extende d to Drs. Julia Graber, Gilbert Botvin, Jeanne Brooks-Gunn and Tracy Nichols for the use of the dataset. I would also like to extend a special thanks to my thesis committee members, Drs. Julia Graber, Susan Bluck and Catherine Cottrell, for their valuable assistance and advice. 3

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 LIST OF TABLES................................................................................................................. ..........6 LIST OF FIGURES.........................................................................................................................7 ABSTRACT.....................................................................................................................................8 CHAPTER 1 INTRODUCTION................................................................................................................. ...9 Social Information Processing.................................................................................................. 9 The Influence of Emotion....................................................................................................... 12 Methodological Issues.......................................................................................................... ..14 The Present Study.............................................................................................................. .....15 2 METHODS...................................................................................................................... .......19 Design.....................................................................................................................................19 Participants.............................................................................................................................19 Procedures..................................................................................................................... ..........20 Measures.................................................................................................................................21 Hypothetical Vignettes....................................................................................................21 Hostile attributions...................................................................................................21 Hypothetical aggressive responses...........................................................................22 Survey Measures.............................................................................................................22 Anger........................................................................................................................22 Anger reduction........................................................................................................23 Aggression................................................................................................................23 Delinquency.............................................................................................................23 Friends delinquency................................................................................................23 Demographic Characteristics...........................................................................................24 3 RESULTS...................................................................................................................... .........25 Demographic Associations.....................................................................................................25 Associations Among Core Constructs....................................................................................26 Attribution-Response Patterns................................................................................................28 Lunch Table.....................................................................................................................29 Hallway............................................................................................................................30 4 DISCUSSION................................................................................................................... ......41 4

PAGE 5

5 APPENDIX HYPOTHETICAL VIGNETTE SCRIPTS...............................................................48 LIST OF REFERENCES...............................................................................................................49 BIOGRAPHICAL SKETCH.........................................................................................................54

PAGE 6

LIST OF TABLES Table page 3-1 Summary statistics on continuous variables...........................................................................31 3-2 Correlations among study variables.......................................................................................3 2 3-3 Hierarchical regression an alysis for variables predicting reported rates of aggression.........34 3-4 Hierarchical regression an alysis for variables predicting reported rates of delinquency.......35 3-5 Hierarchical regression an alysis for variables predicti ng reported rates of delinquent friends................................................................................................................................36 3-6 Attribution-response pattern group membership....................................................................37 3-7 Univariate ANOVAs for lunch groups...................................................................................38 6

PAGE 7

LIST OF FIGURES Figure page 3-1 Attribution-response pattern differences in antisocial outcome variables for lunch table and hallway vignettes.........................................................................................................3 9 3-2 Lunch table attributi on-response pattern differences in delinquency by race........................40 7

PAGE 8

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science THE ROLES OF EMOTION PROCESSES AND HYPOTHETICAL SCENARIOS IN PREDICTING ANTISOCIAL BEHAVIOR By Katherine Hale Clemans December 2007 Chair: Julia A. Graber Major: Psychology The current study investigated the associati on of attributions and responses made to hypothetical vignettes, as well as trait levels of anger and ange r regulation, with reported antisocial behavior in a sample of early adolescents. A sample of 413 6th graders completed two hypothetical social problem-solving tasks as well as a larger survey a ssessing anger, anger regulation, aggression, delinquency, and association with delinquent friends. Hierarchical regressions indicated that anger and anger regulation were much st ronger predictors of antisocial behavior and delinquent friends than were hypothetical aggre ssive responses. Surprisingly, hypothetical hostile attributions were not related to hypothetical aggressive responses or to any antisocial outcome variable. Additional multivariate ANCOVAs conducted on attributionresponse pattern groups within each vignette indicated that participants who made nonhostile attributions but still ge nerated aggressive responses for one of the two vignettes evidenced higher levels of antisocial tendencie s than those with nonhostile attributions and nonaggressive responses. The efficacy of using hypothetical vignette s to predict antisocial behavior in real life is discussed, as well as the need to consid er characteristics of both the vignettes and of participants when conducting research in this area. 8

PAGE 9

CHAPTER 1 INTRODUCTION Social Information Processing Adolescence is a time of numerous behavior al changes, during which temporary increases in aggression and antisocial activity are often observed (Moffitt, 1993). In addition to the fact that aggressive behavior merits serious attention and preventive e fforts on its own, for a subset of individuals, aggression in early adolescence can foreshadow more serious antisocial behavior in adulthood (Coie & Dodge, 1998; Moffitt, Caspi, Ru tter, & Silva, 2001). Therefore, the study of factors that influence aggressive and antisocial te ndencies remains an important area of research in this age group. The present study investigates and compares the ability of attributions and responses made to hypothetical scenarios, as we ll as anger-related em otional processes, to predict rates of antisocial behavior in real life. When an individual is trying to make sense of a social situation, it is the perception of others intentions, rega rdless of those others actual intentions, that guides the individuals response (Berkowitz, 1990; Dodge & Coie, 1987). Most research on attributions of intent is conducted using a social information-proces sing (SIP) framework (Crick & Dodge, 1994; Dodge, 1986; Orobio de Castro, Veerman, Koops, Bosch, & Monshouwer, 2002), which attempts to explain how cognitive and emotiona l factors influence an individuals understanding and conception of the social world. These models describe a series of steps in which social information is encoded, interaction goals are re cognized, and possible res ponse alternatives are generated and evaluated, resulting in a reaction to a given social situ ation. Atypical processing during any step may lead to socially inappropria te behavior, such as aggression (Crick & Dodge, 1994; Lemerise & Arsenio, 2000). 9

PAGE 10

Because of the potentially overwhelming amount of information to which an individual must attend in any given social scenario, (s)he will often try to simplify the amount of cognitive processing needed by applying heuristic shortcuts, such as the selective attention to particular cues, or a reliance on past experiences to inform interpretation of the current event (Crick & Dodge, 1994; VanOostrum & Horvath, 1997). Hos tile attribution bias es may arise through deficiencies in cue utilization, or the capacity to integrate atte ntion information into cognition (Dodge, 1980). For instance, Dodge and Frame (1982) found that aggressive children falsely recalled hostile cues as taking place in past social situations more often than did their nonaggressive peers, partially suppo rting the conclusion that aggres sive children were selectively attending to hostile cues at the e xpense of other social signals. In a study on boys reactions to ambiguous so cial situations, D odge (1980) found that when aggressive participants made hostile attributions of intent to the peer provocateur, they responded aggressively 60% of the time. When th is same group attributed benign intent to the provocateur, however, they still reported that they would respond aggressively 26% of the time. The results of Dodges study suggest that other ch aracteristics of individu als, besides a tendency to interpret hostile intent, play a role in the formation of aggressive responses (VanOostrum & Horvath, 1997). Researchers often attempt to study the provoca tion and initiation of a ggression as well as components of the SIP model th rough the use of hypothetical and role-play vignettes (e.g., Dodge, 1990; Vitaro & Pe lletier, 1991). These tasks usuall y present a problematic social situation or a conflict-resolution scenario and ask th e participants to react or to answer questions about how they would react in re al life. Research of this type allows the benefit of assessing beliefs and emotions at the time of conflict. However, no matter how salient and similar to the 10

PAGE 11

participants life a particular hypo thetical scenario may be, the pa rticipants are always aware of the tasks make-believe quality. This remains the case in most role-play tasks as well, which, in research with children and adolescents, may use an adult confederate to in teract with the target child (e.g., Borbely, Graber, Nichols, Brooks-Gunn, & Botvin, 2005). As a result, it is possible that important aspects of the real-life social scenario are not being captured through hypothetical and role-play investigative techniques. Hypothetical social problem-solving task s can often elicit differences in response strategies between aggressive a nd nonaggressive indivi duals, particularly in cases where the provocateurs intentions are ambiguous (Crick & Dodge, 1996; Orobio de Castro et al., 2002). For instance, Tremblay and Ewart (2005) found modest correl ations between self-reported aggression and various aggressive responses to hypothetical vignettes of provoking situations in a sample of college students. VanOostrum & Ho rvath (1997) found that ho stile attributions to ambiguous peer scenarios signifi cantly predicted both aggressi ve hypothetical responses and actual aggressive behavior in a sample of hi gh school-aged boys. However, results have been inconsistent across studies, with some suggestin g that these relationshi ps only hold for highly aggressive, all-male samples (Crain, Finch, & Foster, 2005; Orbio de Castro et al., 2002). Furthermore, there is evidence that childrens utilization of response strategies in natural social settings is frequently different from the answers they give to hypothetical so cial tasks (see Vitaro & Pelletier, 1991, for a review). Vitaro and Pe lletier found that maladjusted children behaved more aggressively than their well-adjusted peers to provocation by another child; however, no differences were observed in the two groups responses to a hypothetical vignette depicting a similar social scenario. The authors hypothesized that even though the ma ladjusted children may have been aware of socially acceptable beha vioral responses, they became unable to act 11

PAGE 12

accordingly when confronted with real social problem situations because other factors, such as prohibitive emotional arousal (p. 514), prevented the proper utilization of these strategies. The influences of these other factors, specifically anger and anger regulation, are of interest in the present study. The Influence of Emotion Though authors of SIP models have continually mentioned the importance of emotional factors in social cognition (C rick & Dodge, 1994), the function and effects of emotion processes within the SIP framework have received relativ ely little attention. Lemerise and Arsenio (2000) are an exception, offering a reformulation of Crick and Dodges model that addresses the interaction between cognition and em otion processes in the interpre tation of social cues and the generation of behavior. This ne w model views emotion processes, such as emotionality and temperamental factors, emotion regulation, and background moods, as influencing each step of the social cognitive process. For instance, a particularly anxious child may weigh response outcomes differently than a f earless one (Saltaris, 2002). The influence of emotion on decision-making st rategies depends on two main factors: the strength and valence of a particular experience of emotion, often referred to as arousal or emotionality; and the ability to consciously regul ate or control the effect of a particular emotional experience. Efficient emotion regul ation in challenging situations allows the individual to engage in more e ffortful evaluation of response alternatives, instead of jumping to a quick conclusion (Lemerise & Arsenio, 2000, Saar ni, 1999). Being a good regulator may allow an individual to consider multiple perspectives of a situation instead of jumping to a quick conclusion (Saarni, 1999). Individual differences in arousal levels and regulatory abilities are related to differences in both prosocial and problem be havior. Studies by Eisenberg and colleagues (Eisenberg et al., 12

PAGE 13

2001; Eisenberg et al., 1996) have shown that elementary school-a ged children who are high in emotionality but low in regulatory abilities disp lay the highest frequency of conduct problems and socially inappropriate behavi or. One explanation of these fi ndings is that these emotional factors may cause deficiencies in multiple steps of the social-information-processing cycle, so that in any given social situa tion, these children are prone to engage in more abnormal social cognitive processes (Lemerise & Arsenio, 2002). Th e intensity of emotiona l experiences and the ability to regulate those experien ces will also influence the qualities an individual notices about a social encounter and the meaning ascribed to t hose qualities. In situati ons of ambiguous intent, the influence of emotion on decision-making ma y be heightened due to a lack of clearly interpretable cues on which to base the formati on of cognitive reasons for a particular response (Lemerise & Arsenio, 2002). Many types of emotional events can be e xperienced during the c ourse of a social interaction, but one emotion thought to be particularly related to aggression is anger (Dearing et al., 2002). Research has demonstrated that high anger levels can contribute to aggressive behavior (Cornell, Peterson, & Richards, 1999; Nichols, Gr aber, Brooks-Gunn, & Botvin, 2006). The emotional experience of anger is characte rized by physiological arousal and cognitions of antagonism (Novaco, 1994, p. 32); tra it anger is defined as an enduring propensity to become angry (Spielberger, Jacobs, Russell, & Crane, 1983). Anger and the ab ility to effectively regulate anger are related, since an indivi dual with a higher baseline ange r level will have a much tougher job, relatively, of controlling his/her temper. Ni chols et al. (2006) found aggression rates to be positively correlated with anger and negatively correlated with an overall scale of self-control. However, emotion arousal and emotion regulation ar e distinct factors that contribute uniquely to 13

PAGE 14

antisocial behavior (Eisenberg et al., 2001; Le ngua, 2003), suggesting that skill at effortful emotion regulation may vary widely with in individuals with high anger levels. Though the influence of emotion regulation factors on behavior has received much attention in recent years (Eise nberg, Morris, & Spinrad, 2005), research on anger regulation specifically has been scarce (Zeman, Shipman, & Suveg, 2002). Results have been mixed, with some studies finding no direct link between ange r regulation and aggression in younger children (Dearing et al., 2002). In pa rticular, there is little research ab out this relationshi p in adolescents. The present study included a scal e that assesses control strate gies relating explicitly to experiences of anger, and as such was more sp ecific than scales assessing general regulation strategies. Methodological Issues Most of the research on the relationships between hostile attri bution of intent and aggressive behavior has been conducted on preschool and elementa ry-school-aged children; less attention has been paid to the relationship of these processes in adolescents, especially nonclinical or non-incarcerated samples, even t hough this type of research on adolescents is necessary to examine the persistence and stabili ty of these relationships over the life-span (VanOostrum & Horvath, 1997). The relationship of hypothetical aggressive responses to reported rates of aggression in real life is also unclear, especially in this age group (Crain et al., 2005); more research with mixed-sex, non-referr ed adolescent samples is needed. Thus, the present study seeks to determine how well aggressi ve responses in hypothetical scenarios relate to reports of aggressive behavior in real life in a non-clinical sample of early adolescents. In addition, while research relating ambiguously hostile hypothet ical vignettes to real-life behavior have typically focu sed only on aggression, it is pos sible that the tendency to overattribute hostile intentions to social interac tions with peers may mani fest itself in other 14

PAGE 15

forms. Aggressive behavior is highly related to other forms of antisocial behavior, including delinquent acts such as vandalism and theft (Mo ffitt et al., 2001). While previous studies have investigated hostile attr ibutions and the SIP framework in samples of delinquent individuals, this term has usually been used to differentia te between normal and referred or incarcerated samples, and thus the label represents aggressi ve behavior in addition to other antisocial acts (Guerra & Huessmann, & Zelli, 1990). Few, if any, studies to date have included measures of specific delinquent behaviors, or have attempted to separate these from specifically aggressive behaviors. Therefore, one aim of the present study is to investigat e whether the predictive ability of hypothetical vignettes extends to delinquent behavior as well. Attribution processing style may also be relate d to the antisocial behavior of ones peers. For instance, a tendency to perceive hostility in social situations may lead individuals to select peers for whom this attribution style is c onsidered acceptable. Alternatively, exposure to antisocial peers may lead to increased hostile at tributions through modelin g of others (Bosson & Johnson, 2006). Though investigation of these mech anisms is beyond the scope of the present study, individuals friendship with aggressive and/ or delinquent peers is included as an outcome variable in order to assess its associative relati onship with responses to the hypothetical vignettes. The Present Study The present investig ation examines responses to hypot hetical vignettes, emotions, and self-reported aggression, deli nquency, and peer delinquency in a diverse sample of 6th graders. Based on previous research on the SIP model (Orobio de Castro et al., 2002), a significant relationship is expected between real-life rates of aggression a nd aggressive responses to the hypothetical social problem-solving vignettes, such that more frequent aggressive responding will lead to higher reported rates of aggressive be havior. Furthermore, it is expected that this 15

PAGE 16

relationship will exist between rates of aggression and hostile attributions of intent in the hypothetical vignettes as well. A second purpose of the present study is to in vestigate the influence of emotion-related variables on the relationship betw een responses to hypothetical vi gnettes and real-l ife aggression. The function of emotion in social cognition can be viewed as a marker that signals important features of the social interaction and offers direction for cognitive processes and behavior (Damasio, 1994). Anger and hostile attribution should be related in this sense, because if one feels high levels of anger (due to either high baseline levels or in sufficient regulation), this would be an emotional signal that (s)h e is in a threatening situation; therefore, (s)he should attribute hostile intent to the other person or people involved. A resultant pred iction of this study is that mediated relationships between anger and host ile attributions and a nger reduction and hostile attributions should both be observed, such that greater levels of trait anger and poorer ability to regulate anger should increase attrib utions of hostile intent, which in turn should lead to more aggressive responses in the hypothe tical scenarios and more antisoc ial behavior in real life. A series of hierarchical regressi ons in accordance with the procedures of Baron and Kenney (1986) will be used to show the mediating impact of a nger processes on the relationship between hostile attributions and hypothetical aggres sive responses, as well as on the relationship between hostile attributions and reported an tisocial outcome variables. Anger and anger regulation skills may expl ain the association between hypothetical responses and real-life antisocial behavior, but they also may explain unique variance in antisocial behavior unrelated to the hypothetical scenarios (Zeman et al., 2002). Though anger levels and anger reduction ability should be moderately negatively correlated (Eisenberg et al., 2005), unique variance contributions to antisocial behavior are also expected from each. It is 16

PAGE 17

hypothesized that this unique vari ance would represent differences in situational characteristics and emotional salience between hypoth etical and real-life peer conf licts. Hierarchical regressions on each antisocial outcome variable, with anger processes, and hypothetical vignette responses entered in separate steps, will assess the relative variance explained by each variable. Although examinations of individual hypothetical vignette s is rare in the literature to date (Orobio de Castro et al., 2002), one of the benefits of using thes e vignettes in SIP research is that they allow for the investigation of specific attribution-response rela tionships as they relate to a single scenario. Therefore, exploratory multivar iate ANCOVAs will also be conducted for each hypothetical vignette to assess the relationship of individual attr ibution-response patterns with levels of real-life antisocial te ndencies. It is possible that thes e additional analyses will uncover new relationships between hypothetical attribut ions and responses not seen in the more traditional regression analyses. In addition to the age of the participants, several other demographic characteristics may play a role in the differential formation of hos tile attributions and the expression of aggression. As stated previously, overall rate s of aggression appear to be highe r in adolescent males than in adolescent females, suggesting th at males and females may utiliz e emotional factors differently in hostile or ambiguous peer situations (Crick, 1997). Fam ily background and community characteristics may also play a role in disp lays of aggression. For instance, aggression in African-American urban communities may serve adaptive functions unique to that environment, and therefore not be considered as socially inappr opriate as it would be in other contexts (Coie & Dodge, 1998). Because of the wealth of research addressing the relationships between these factors and the cognitive, emoti onal and behavioral variables in question (see Coie & Dodge, 17

PAGE 18

1998, for review), differences in gender and dem ographic characteristics will not be a central focus of this study, but will be te sted in preliminary analyses. 18

PAGE 19

CHAPTER 2 METHODS Design The current investigation consists of a substudy of a randomized clin ical trial designed to evaluate a school-based drug abuse and violen ce prevention program. A total of 42 New York City middle schools took part in the full study, which was approved by the Institutional Review Board for the Protection of Human Subjects in Re search (IRB) at Weill Medical College, Cornell University. Fourteen parochia l and three public schools in the larger program agreed to participate in supplementary data collection activities. The high relative percentage of parochial schools in this sample is a result of the recruitment of only the sm allest schools from within the larger study for the supplementary data collection due to its intensive nature. Data for this study is drawn from the pre-trial assessment at the beginning of the participants 6th grade year in order to avoid confounding effects from the intervention program. Participants Only students who completed both the survey measures used in the full study and the supplementary videotaped activities were included in the current sample ( N = 413). The mean age was 11.63 years (SD = .49 years; range = 9. 64 to 13.86 years). Girls made up 50.4% of the sample ( n = 208). Racial subgroups included Afri can American (49.6%), Latino (25.4%), Caucasian (17.9%), and other (6.8 %). Fifty-three percent of the sample lived in two-parent, nonblended families, while 23.5% liv ed with single mothers and 23.5% lived in other household configurations. Parochial school students made up 69.7% of the sample; all other students attended public school. 19

PAGE 20

Procedures Parental consent for the full clinical tr ial was obtained through a passive consent procedure, which provided a comprehensive descri ption of the investigat ion and the self-report surveys and allowed parents the opportunity to object to their childs participation. The subsample of 17 schools received a secondary passive consent form that explained the supplementary data collection activities, incl uding the hypothetical pe er-interaction tasks. Parental objection to the primary consent form precluded student pa rticipation in the supplementary tasks. An ethnically diverse team of three to five data collectors presented the self-report survey on two separate days during regular ly scheduled class periods. A sta ndardized protocol similar to ones used in previous research on drug use and delinquent behaviors was used (see Botvin et al., 1994). To maintain the quality of the self-report da ta, identification codes rather than names were used to emphasize the confidentia l nature of the questionnaire and students were assured about the confidentiality of their responses. To enhanc e the validity of the self-report data, carbon monoxide breath samples were collected usi ng a variant of the bogus pipeline procedure developed by Evans and colleagues (Evans, Hansen, & Mittlemark, 1977). This measure is primarily used for questions pertaining to ci garette smoking, although res earch has shown that bogus pipeline procedures can also increase the va lidity of other self-repo rted problem behaviors (Tourangeau, Smith, and Rasinski, 1997). Supplementary activities took place on a third da y of data collection in the subset of 17 schools. Participating students completed a seri es of activities and answered questions about their reactions to hypothetical vignettes describing ambiguous peer interactions in an interview format. Data collectors were trained undergradu ate and master degree students, and primarily 20

PAGE 21

female, African-American and Hispanic, although at least one male data collector was sent to each school. Students completed the hypothetical vignette ta sks individually in a private classroom space provided by the school. A data collector read relevant inst ructions and confidentiality information to each student before beginning the tasks. Students were informed that they were to respond to these scenarios as they would in real life. Responses to ope n-ended questions were transcribed for later coding. T wo hypothetical soci al problem solving tasks (Dodge, 1990) dealing with ambiguously aggressive peer interactions were selected from the larger set of tasks for analysis in the present study. In the first hypothetical scenario, the participan t is instructed to imagine that he/she is denied access to a sc hool lunch table for an unknown reason. In the second hypothetical scenario, a student wa lking toward the participant in a school hallway yells out an insult which may or may not have been directed at the participant. For both hypothetical tasks, the participant was asked what he/she thought was going on in the mind of the provocateur as well as what his/her desired and actual responses to the situati on would be. Complete scripts for the social problem solving task s are included in Appendix A. Measures Hypothetical Vignettes Hostile attributions Answers to both the attributi on and response questions of the hypothetical social problem-solving vignettes used a coding scale developed for the larger study (Graber et al., 2001; Nichols et al ., 2001). Participants attributions of the provocateurs intent were coded from open-ended responses at the end of the vignettes in which they indicated what they thought was going on in the mind of the hypot hetical peers while the events were taking place. Attributions of intent were coded as hostile when they included intimidation (e.g., she wanted to start trouble, they want to embarrass me); attention-seeking (e.g., she was trying 21

PAGE 22

to be popular, he was trying to show off in fr ont of his friends); or personal problems (e.g., he was mad, he must be jealous of me, t heyre racist). For comparison purposes, other possible responses included self-blaming attr ibutions (e.g., he thinks Im a loser) and situational attributions (e.g., som ebody else really was sitting ther e). Interrater agreement was 89% ( = .80) for the lunch table vignette and 87% ( = .84 ) for the hallway vignette. Hostile attribution responses for each question were summed to produce a hostile attribution score with a range of 0-2. Hypothetical aggressive responses Aggressive responses we re coded from open-ended responses at the end of the vignett es in which participants indicated what they would do or say to the other kids if this event ha ppened to them. Responses were coded as aggressive if they included physical force, such as pushing or shovi ng; verbal confrontation, such as the use of threats, insults, ridicule, or sarcasm; or non-verbal hostile gest ures, such as giving nasty looks. Examples of nonaggressive responses included wa lking away or questioning why the seat was unavailable. Interrater agreement was 90% ( = .88) for the lunch table vignette and 89% ( = .86 ) for the hallway vignette. Aggressive responses for each question were summed to produce a hypothetical aggressive respons e score with a range of 0-2. Survey Measures Anger The seven-item anger subscale ( = .74) from the Buss and Perry (1992) Aggression Questionnaire was used to assess trait levels of anger. Students were asked to rate how well a series of statements fit them. Item s included I sometimes feel like a powder keg ready to explode and Some of my friends th ink Im a hothead. Response categories ranged from 1 (Really Not True for Me) to 5 (Really True for Me). Items were averaged such that higher scores indicate greater anger. 22

PAGE 23

Anger reduction Anger reduction skills were assessed with a six-item subscale ( = .81) created for the larger clinical trial (Epstein, Botvin, Diaz, Bake r, & Botvin, 1997). Participants were asked how often they engaged in a series of activities when they felt really angry. Items included Count to ten, Take a few deep breaths and Tell myself this isnt worth fighting over (its no big deal.) Respons e categories ranged from 1 (Nev er) to 5 (Always). Items were averaged such that higher scores indicated greater skill at conscious anger reduction. Aggression Self-reported rates of aggression were assessed using 10 items from the aggression scale ( = .92) of the Youth Self Report (A chenbach & Edelbrock, 1986). Students were asked how many times in the past month th ey had engaged in incidents of aggressive behavior. Items included Yelled at someone (you were mad at), Told someone off, Pushed or shoved someone on purpose, and Hit some one. Response categories were on a 5-point scale. Response options included 1 (Never), 2 (O nce), 3 (2-3 times), 4 (4-5 times), and 5 (More than 5 times). Responses were rescored to a zero-baseline and then summed to produce an overall aggression score with a range of 0-40. Delinquency. Delinquent acts were measured using a 10-item scale (Elliot, Huizinga, & Menard, 1989, = .86) that asked how often students had engaged in delinquent behaviors in the past year. Examples included purposely damaged or destroyed property or things that did not belong to you and taken something from a store when a clerk wasnt looking. Response options included 1 (Never), 2 (Once), 3 (2-3 times ), 4 (4-5 times), and 5 (More than 5 times). Responses were rescored to a zero-baseline a nd then summed to produce an overall delinquency score with a possible range of 0-40. Friends delinquency Friends participation in delinque nt acts was measured using a seven-item scale (Elliot et al., 1989, = .88) that assessed how many of the participants friends 23

PAGE 24

had engaged in delinquent behaviors in the past year. Examples included ruined or damaged something on purpose that wasnt theirs and b roken into some place to steal something. Response options referred to the portion of the participants friends and included 1 (None), 2 (Less than half), 3 (About half), 4 (More than ha lf), and 5 (All or almost all). Responses were rescored to a zero-baseline and then summed to produce an overall frie nds delinquency score with a possible range of 0-40. Demographic Characteristics For facilitation of regression analyses, singl e mother households were collapsed into the other group of family structure and used as a comparison group for the two-parent variable (Nichols et al., 2006). 24

PAGE 25

CHAPTER 3 RESULTS Table 3-1 presents full sample means and standard deviations for each of the seven continuous variables employed in this study. Tests of normality were run on the distribution of each variable. Because hypothetical aggressive responses, self-reported aggression, delinquency, and friends delinquency we re positively skewed ( zskewness = 6.86 to 17.88), these variables were transformed using the square root function in al l subsequent analyses. Distributions of other variables did not require transformation. Demographic Associations Table 3-2 presents bivariate correlations between study variables. Among demographic variables, family structure was significantly associated with all three antisocial variables; living in a two-parent, non-blended family served as a protective factor against exhibiting high rates of antisocial behavior and against associating with antisocial peers. Being African-American was positively correlated with higher rates of aggre ssive behavior, and being male was positively correlated with higher rates of delinquent beha vior, but both correlations were small in size, indicating no meaningful associatio n. It is interesting to note that no gender differences existed for aggressive behavior in this sample. In addition, attending public school (as opposed to parochial school) was associated with an in creased rate of hostil e attributions. Because of the categorical nature of each de mographic variable, chi-square analyses were conducted as a further test of demographic associ ations. Family structur e and school type were associated such that participants who lived in two-parent families were less likely to attend public school (2 = 18.80, p < .001). Family structure and school type were also both confounded with race: African-American participants were less likely to live in two-parent, non-blended families, and public schools contained greater proportions of African-American students and 25

PAGE 26

smaller proportions of Caucasian particip ants than would be expected by chance ( 2 = 64.27, p < .001). Gender was mildly associated with being African-American or Latino (there were more African-American girls than boys and more Latino boys than girls). The gender differences were considered an anomaly of the sample a nd not relevant to any other findings. Associations Among Core Constructs As expected, the three outcome variab les (aggression, delinquency, and friends delinquency) were highly colinear, and aggression and delinquency especially exhibited similar relationships with hypothetical a nd emotion-related variables. Al so as predicted, hypothetical aggressive responses were rela ted to reported aggression, though this association was weak, r = .16, p < .01. Hypothetical aggressive re sponses were also weakly associated with increased rates of delinquent behavior and, to a lesser ex tent, association with delinquent friends. Surprisingly, hostile attribution of intent was not significantly associated with hypothetical aggressive responses or with reported rates of aggression, delinquency or having delinquent friends. In addition, hostile attributions were not correlated with either anger variable. One of the hypotheses of this study was that emotional processes would me diate the relationship between hostile attribution and aggression. However, because no significant correlations existed between hostile attribution and a ny of the emotion and aggression variables, this model was not examined, and hostile attribution wa s dropped from subsequent analyses. Anger levels and anger reduction we re not meaningfully associated, r = -.09, p = .07. The small magnitude of this relations hip supports the conceptual dis tinction between emotion arousal and emotion regulation (Eisenberg et al., 2005). As expected, both a nger variables were moderately correlated with all antisocial variab les. High anger levels were associated with greater rates of aggression, delinquency, and delinquent friends, while good anger reduction skills predicted lo wer rates of each. 26

PAGE 27

A hierarchical multiple regression analysis was conducted to test the overall variance explained by both hypothetical aggr essive responses and anger-rel ated variables in rates of reported aggression, delinquency, and delinquent frie nds. Because of their associations with one or more of the outcome variables, gender, family structure (two-parent fam ily vs. other), and race (African-American vs. other) were entered into a ll models in step 1 as covariates. School type was dropped as a covariate due to its lack of association with any outcome variable (Stevens, 1996). Due to the fact that correlational analyses showed anger processes to be more strongly associated with antisocial behavior and delinque nt friends than were hypothetical aggressive responses, the model was structured to first acco unt for the combined association of anger and anger regulation and then test further contributions to the rela tionship explained by hypothetical aggressive responses. Accordingly, in each regr ession, anger and anger re gulation were entered in step 2; and hypothetical aggressive responses were entered in step 3. Tables 3-3, 3-4, and 3-5 present the results of these analyses. After controlling for gender, family structure, and race, anger and anger regula tion both remained significant predictors of all three outcome variables, explaini ng 23% of the total variance in reported aggression, 20% of the total variance in reported delinquency, and 14% of the total variance in delinquent friends. Hypothetical aggressive re sponses also remained a significant predictor of each outcome variable upon addition to the model; however, it only expl ained an additional 1% of variance in aggression, 3% of variance in delinquency, and 1% of variance in delinquent friends. Reversing the order of steps 2 and 3, so th at hypothetical responses were en tered into the regression before anger variables, did not substa ntially change the variance expl ained by any of the predictors. Based on these analyses, it appears that simply examining levels of hostile attributions or aggressive responses in these hypothetical vignettes pr ovided little informa tion about antisocial 27

PAGE 28

behavior occurring in real life, and that emoti onal characteristics like anger levels and regulation may be better indicators of antisocial behavior in normative samples. However, hypothetical scenarios such as the ones used in this study continue to be wide ly used in research on child and adolescent samples. Hence, additional analyses were conducted to furt her explore potential associations of hypothetical vi gnette responses with antisoc ial behaviors in real life. Attribution-Response Patterns One approach that goes beyond simply investig ating overall levels of attributions and responses is to instead look at di fferences in the patterns of responses that participants give to individual scenarios. For instance, in any gi ven sample, only a portion of participants will respond aggressively to a partic ular scenario, and of those that do, not everyone will have attributed hostility to the provocat eur. It is possible th at separating participan ts into groups based on response patterns will uncover information about differences in social information-processing and real-life behavior even wh en hostile attributions show no association with hypothetical responses in the sample as a whole. This appr oach allowed for the examination of situational characteristics of the vignettes that may have resulted in different at tribution or response outcomes from vignette to vignette. For each vignette, four attribution-res ponse pattern groups were created: nonhostilenonaggressive, nonhostile-aggressiv e, hostile-nonaggressive, and hostile-aggressive. Group distributions are displayed in Ta ble 6. A chi-square analysis co mparing group membership in the lunch table vignette to the hallway vignette was not significant ( 2 = 14.22, p < .115), indicating that membership in a particular group di d not stay stable across the vignettes. A multivariate analysis of covariance was conducted for each vignette using group membership as the independent variable and a multivariate construct of antisocial tendencies, consisting of aggression, delinquency, and friends delinquency, as the dependent variable. 28

PAGE 29

Gender, family structure, and race were include d as covariates in each analysis, and for exploratory purposes, interactions between pattern group membersh ip and each covariate were included in the analyses as well. Games-Howell error corrections were employed for post-hoc tests due to the large disparities in group sizes. Lunch Table The multivariate ANCOVA for the lunch table vignette groups was significant, Wilks = .95, F (9, 961.48) = 2.09, p < .05, partial 2 = .02, as were univariate tests for each individual outcome variable (Table 7). Post-hoc means co mparisons revealed that the nonhostile-aggressive group had significantly higher rates of aggression, delinquency, and association with delinquent friends than did the nonhostile-nonaggressive group (mean differences = .34 to .47 SD ; Figure 3-1). In addition to expected multivariate main effects of family structure (Wilks = .96, F (3,395) = 5.22, p < .01, partial 2 = .04) and race (Wilks = .97, F (3,395) = 4.66, p < .01, partial 2 = .03), a multivariate effect also existed for the interaction between lunch table group membership and race, Wilks = .93, F (9, 961.48) = 3.50, p < .001, partial 2 = .03. Univariate tests revealed that the multivariate effect was driven by a significant interaction effect for delinquency only ( F [3, 397] = 4.70, p < .01, partial 2 = .03), such that within non-African Americans, nonhostile-aggressive pa rticipants were significantly higher in delinquency than all other groups; however, this pattern did not hold fo r African Americans (Figure 3-2). One issue is that large standard errors exis ted for the hostile-aggressive groups in particular, possibly due to their small size ( N = 10 for African Americans; N = 11 for non-African Americans), which may have prevented further significant di fferences from becoming evident. 29

PAGE 30

Hallway The multivariate ANCOVA for the hallway vignette groups was not significant (Wilks = .97, F (9, 961.48) = 1.51, p = .140, partial 2 = .01), and its pattern of results did not replicate those of the lunch table s cenario (Figure 1). Follow-up univariate ANCOVAs were not significant for any of the three outcome vari ables. In addition, no group membership x demographic variable interactions existed for the hallway vignette. 30

PAGE 31

Table 3-1 Summary statistics on continuous variables Measure M SD Hostile attributions .84 .64 Hypothetical aggressive res ponses .49 .55 Anger level 2.50 .90 Anger reduction 2.79 1.03 Self-reported aggression 12.97 10.18 Delinquency 3.73 4.70 Friends' delinquency 4.17 4.81 31

PAGE 32

Table 3-2 Correlations among study variables 1 2 3 4 5 6 1 Gender 2 Public school .00 3 African Americana .10* .34** 4 Caucasiana .04 -.29** n/a 5 Latinoa .10* -.03 n/a n/a 6 Two-parent, non-blended family .07 -.14** -.26** .19** .09 7 Hostile attributions .01 .14** -.02 .01 .04 -.03 8 Hypothetical aggressive responses -.01 -.01 .05 .01 -.10* .01 9 Anger -.05 -.01 -.03 .05 .01 -.01 10 Anger regulation -.06 -.05 -.03 .05 .01 .03 11 Reported rates of aggression .04 .04 .11* -.03 -.09 -.15** 12 Delinquency .10* .05 .08 -.08 -.04 -.15** 13 Friends' delinquency .01 .06 .07 -.09 .03 -.19** Note For student sex, 1 = boys. p < .05. **p < .01. a Correlations were not applicable due to covariance caused by variable coding. 32

PAGE 33

Table 3-2 Continued 7 8 9 10 11 12 1 Gender 2 Public school 3 African Americana 4 Caucasiana 5 Latinoa 6 Two-parent, non-blended family 7 Hostile attributions 8 Hypothetical aggressive responses .04 9 Anger -.02 .06 10 Anger regulation -.03 -.08 -.09 11 Reported rates of aggression -.02 .16** .40** -.31** 12 Delinquency -.02 .20** .38** -.29** .75** 13 Friends' delinquency -.07 .13* .33** -.22** .58** .61** 33

PAGE 34

Table 3-3 Summary of hierarchical regr ession analysis for variables predicting reported rates of aggression Variable R R2 B SE B Step 1 .03** Gender .17 .15 .06 Family structure -.40 .15 -.13** Race .23 .15 .08 Step 2 .26** .23** Gender .17 .13 .06 Family structure -.35 .13 -.12** Race .26 .13 .09 Anger .62 .07 .38** Anger reduction -.38 .06 -.27** Step 3 .27** .01** Gender .17 .13 .06 Family structure -.36 .13 -.12** Race .24 .13 .08 Anger .61 .07 .37** Anger reduction -.37 .06 -.26** Hypothetical aggressive responses .32 .12 .12** Note. ** p < .01. 34

PAGE 35

Table 3-4 Summary of hierarch ical regression analysis for variables predicting reported rates of delinquency Variable R R2 B SE B Step 1 .04** Gender .28 .12 .11* Family structure -.35 .12 -.14** Race .14 .12 .06 Step 2 .24** .20** Gender .28 .11 .12** Family structure -.31 .11 -.13** Race .16 .11 .07 Anger .49 .06 .36** Anger reduction -.28 .05 -.24** Step 3 .27** .03** Gender .28 .10 .12** Family structure -.32 .11 -.13** Race .14 .11 .06 Anger .48 .06 .35** Anger reduction -.27 .05 -.23** Hypothetical aggressive responses .35 .10 .16** Note. ** p < .01. 35

PAGE 36

Table 3-5 Summary of hierarch ical regression analysis for variables predicting reported rates of delinquent friends Variable R R2 B SE B Step 1 .04** Gender .05 .12 .02 Family structure -.44 .12 -.18** Race .05 .12 .02 Step 2 .18** .14** Gender .05 .11 .02 Family structure -.41 .11 -.17** Race .07 .11 .03 Anger .42 .06 .31** Anger reduction -.21 .05 -.18** Step 3 .19** .01* Gender .05 .10 .02 Family structure -.41 .11 -.17** Race .06 .11 .02 Anger .41 .06 .31** Anger reduction -.20 .05 -.18** Hypothetical aggressive responses .21 .10 .09* Note. p <.05. ** p < .01. 36

PAGE 37

37 Table 3-6 Attribution-response pattern group membership Lunch table scenario Hallway scenario Group N % N % Nonhostile-nonaggressive 256 62.0 105 25.4 Nonhostile-aggressive 79 19.1 41 9.9 Hostile-nonaggressive 57 13.8 201 48.7 Hostile-aggressive 21 5.1 66 16.0

PAGE 38

Table 3-7 Univariate ANOVAs for lunch table groups Aggression Delinquency Delinquent Friends Source df F partial 2 df F partial 2 df F partial 2 Group Membership (GM) 3 2.82* .02 3 5.29** .04 3 2.75* .02 Covariates Gender 1 .01 .00 1 2.64 .01 1 .47 .00 Family Structure 1 7.20** .02 1 7.54** .02 1 15.43** .04 Race 1 3.78 .01 1 5.10* .01 1 .71 .00 Covariate Interactions GM x Gender 3 .38 .00 3 .10 .00 3 1.32 .01 GM x Family Structure 3 .90 .00 3 .71 .01 3 1.57 .01 GM x Race 3 1.98 .02 3 4.70** .03 3 1.26 .01 Error 397 (2.13) 397 (1.35) 397 (1.38) Note Values in parentheses represent mean square errors. p < .05. ** p < .01. 38

PAGE 39

A B C -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0Nonhostile NonaggressiveNonhostile AggressiveHostile NonAggressiveHostile AggressiveAggression (z-score) Lunch table Hallway -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0Nonhostile NonaggressiveNonhostile AggressiveHostile NonAggressiveHostile AggressiveDelinquency (z-score) Lunch table Hallway -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0Nonhostile NonaggressiveNonhostile AggressiveHostile NonAggressiveHostile AggressiveDelinquent Friends (z-score) Lunch table Hallway Figure 3-1 Attribution-response pattern differences in antisocia l outcome variables for lunch table and hallway vignettes. Error bars represent 95% confidence intervals. A) Attribution-response patterns for aggressi on. B) Attribution-response patterns for delinquency. C) Attribution-respons e patterns for delinquent friends. 39

PAGE 40

40 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0Nonhostile Aggressive Nonhostile Aggressive Hostile Nonaggressive Hostile AggressiveDelinquency (z-score) African American Other Figure 3-2 Lunch table attribut ion-response pattern differences in delinquency by race. Error bars represent 95% confidence intervals.

PAGE 41

CHAPTER 4 DISCUSSION This study investigated the re lationship of anger and anger re gulation to the associations between aggressive responses to hypothetical social scenarios and reported rates of aggression in real life. Because previous re search has tended to focus on the application of SIP-based hypothetical research techniques to real-life aggression only, delinque nt behavior and association with delinquent friends were also included as outcome variables of interest in order to give a fuller picture of real-life antisocial behavior. A moderate, yet significant, link was demonstrated between hypothetical aggression an d the three kinds of reported antisocial tendencies. This finding is similar to previous research using hypothetical and role-pla y scenarios to assess aggression (e.g., Tremblay & Ewart, 2005). Howe ver, both anger and anger regulation had stronger associations with all three antisocial variables th an did hypothetical aggressive responses, and these two emotion processes join tly accounted for nearly all of the explained variance in reported aggression, de linquency, and delinquent friends. Combined with the finding that neither em otional process was significantly associated with hypothetical aggressive responses, the results of this study lend support to Lemerise & Arsenios (2000) reformulation of the social-information-pro cessing model and suggest that hypothetical social problem-solving tasks may not capture something crucial about the emotion processes that contribute to aggression in real life. This supports previous work proposing that children can know appropriate responses to hypothetical situ ations, but that characteristics about real-life encounters which differ from hypothetical scenario may generate an impulsive response which prohibits this knowledge from being implemented (Vitaro & Pelletier, 1991). Therefore, future researchers using these types of tasks to investigate or evaluate aggressive behavior need to take their pote ntial limitations into consideration. 41

PAGE 42

In addition, this study assessed the prevalence of hostile intent attributed to the provocateur in these problem-solving tasks a nd its association with anger processes and aggression in both the hypothetical scenarios and in real life. Unexpectedly, hostile attributions were not related to any antisocial outcome variable More surprisingly, they were also not related to hypothetical aggressive responses, which were assessed using the same vignettes. This result was a departure from prior research on hostile attributions, which has demonstrated a pronounced association with aggre ssive behavior (Or obio de Castro et al., 2002). However, much of this work has compared highly aggressi ve, clinically referred or incarcerated samples to less aggressive control groups. It is possible that within normal sp ectrums of aggressive behavior in non-referred samples, the SIP model of hostile attributions aggressive behavior is not an accurate representation of cognitive processe s taking place during these scenarios. Hostile attributions were also not associated with either of the two anger variables. The questions contained in the Buss a nd Perry (1992) anger scale concep tually assess trait anger, or enduring tendencies to become a ngry across contexts (Spielberger et al., 1983), rather than state anger, which is transiently aroused and asse ssed during specific situations. The specific association between trait anger and hostile attribution biases in adolescents has not been conclusively studied, although research in adults has demonstrated a link between the two (Epps & Kendall, 1995). Orobio de Castro, Merk, K oops, Veerman and Bosch (2005) found a mild correlation between emotion regulation skills and interpretation of intent in hypothetical vignettes, but their sample included a subset of highly aggressive boys referred for behavior problems. It is possible that th e results of the current study refl ect the existence of a weaker association between these constructs in nonclinical, mixed-gender adolescent samples. 42

PAGE 43

An alternative explanation for our findings is that the failure to replicate previous research on the association betw een hostile attribution and aggression, as well as the weakness of the relationship observed between hypothetical aggression and real -life antisocial behavior, may be due to the small number of hypothetical scenario s assessed in this study (two), resulting in a restricted range of possible response scores. Prior work on hostile attr ibution and aggression typically has presented 7 to 12 scenario vignettes to each par ticipant (e.g., Crain et al., 2005; Orobio de Castro et al., 2005). Adding more scenarios to the continuous measures used in this study might have resulted in strong er statistical associations betw een hostile attribution and other constructs. However, other resear ch using small sets of scenario s suggests that th e nonsignificant relationship observed between hosti le attribution and reported antis ocial tendencies may not just be a result of low variance in the measures. Vitaro and Pelletier (1991) presented children with four hypothetical vignettes of ambiguous provoc ation and three actual ambiguous provocations by a trained peer-confederate. Resu lts indicated that children cl assified as maladjusted (i.e., aggressive and rejected by peer s) responded to the peer-confederate provocations with more verbal attacks than did well-adjusted particip ants; however, no difference was seen between the two groups for the hypothetical vignettes. Additional research by th ese authors (Vitaro, Pelletier, & Contu, 1989) found that maladjus ted children also attributed more negative intentions to provocateurs in peer-confederate interactions than in hypothetical vignettes. This research suggests that, small measure set or not, the l ack of association in the current study between hostile attributions and re ported rates of aggression may be a reflection of the larger issue of emotional processing differences between hypothe tical scenarios and re al-life situations. Though a surprisingly small relationship was found between hypothetical and real-life tendencies when responses were summed across vi gnettes, one of the benefits of hypothetical 43

PAGE 44

vignette research methods remains the ability to inves tigate attributions an d responses as they relate to a specific hypothetical in cident. Because of the importance of the outcomes which these studies are trying to accomplish the identifi cation and prevention of youth aggression (e.g., Hudley & Graham, 1993) it may be worthwhile to reexamine the predictive relevance of these types of hypothetical measures th rough additional methodological tech niques. It was therefore of interest to determine whether any useful predictive informa tion about real-life antisocial behavior could be obtained from investigating attri bution-response patterns within individual vignettes. This attribution-respons e pattern approach has been underutilized in rese arch to date: SIP research using group comparisons has usually based group memb ership on levels of real-life aggression or participant characteristics such as peer rejection (Lansf ord et al., 2006; Lsel, Bliesener, & Bender, 2007; Or obio de Castro et al., 2002). Separating participants into groups based on th eir attribution-response patterns (hostile or nonhostile; aggressive or nonaggre ssive) yielded signifi cant differences between the nonhostilenonaggressive and nonhostile-aggressi ve groups in the lunch table vignette, such that nonhostileaggressive participants (approximately 19% of the sample) were higher in aggressive and delinquent behavior and reported that they had more delinquent friends. An interaction effect between group membership and race further suggested that this pa ttern of results was mainly due to higher rates of antis ocial tendencies among the non-African American members of this group. Nonhostile-aggressive response patterns may repres ent a particularly problematic group of adolescents, who seem to be arri ving at aggressive solutions to social interactions without having a reason to do so (i.e., feeling threatened by the perc eived hostility of a peer). However, no such relationship was found for the hallway scenario, in which attributionresponse group membership did not yield any significant predictiv e information about antisocial 44

PAGE 45

tendencies in real life. The wide difference in pattern distribution between the two vignettes, especially in attributions of hostility, is likely explained by the fact that while both vignettes could be classified as ambiguous, they differed in their levels of ambiguity: In the hallway scenario, while it remains unclear whether the pro vocateur was addressing his/her remarks to the participant or to another student the actual statement the partic ipant makes (Hey, geek. Yeah, I mean you, nerd) contains more obvious hostility than the stat ement made in the lunch table scenario (You cant sit there, th at seats taken). As such, the majority of participants thought that the hallway provocateur was hostile, wh ile the lunch table pr ovocateur was not. The ambiguity level of the vignettes ma y be an important predictor of SIP processing in normative, nonreferred individuals, and as su ch may be a useful inclusion in future research using these populations. Further investigation of antisocial behavior differen ces among attribution-response pattern groups should include possible pers onality measures which might account for participants differing reactions to the hypothetica l vignettes. For instance empathyand fearrelated characteristics might explai n why more antisocial adolescents in this sample tended to act aggressively in the absence of perceived threat in the lunch ta ble scenario (Andershed, Kerr, Stattin, & Levander, 2002). Social measures such as peer rejection or the desire to increase ones social status may also play a ro le in adolescents decisions to act aggressively in the absence of hostility (Guerra, Asher, & DeRosier 2004; Rose, Swenson, & Waller, 2004).1 These variables could potentially be incorporated at various steps of the SIP model to expl ain why attributions of hostile intent do not usually result in aggressive responses in this sample. Hypothetical tasks like the ones described in this study are often used in prevention and intervention efforts for youth with behavior problems (Sukhodolsky, Golub, Stone, & Orban, 45

PAGE 46

2005). A greater understanding of the relationship between hypothetical and real-life responses to stressful social situations, as well as the f actors that may differ between the two, may help to improve the efficacy of these programs. However, it is important to note that effect sizes of the attribution-response pattern analyses were still quite small, mimicking the amount of variance explained when continuous versions of the hypothe tical variables were an alyzed. The utility of hypothetical scenarios at telling us about actual antis ocial behavior is expect ed to be moderate at best, suggesting the need to cons ider other characteri stics of adolescents in addition to their performance in these tasks. An additional limitation of the study may be that the distinction between reactive aggression (aggression in response to provocat ion) and proactive aggression (aggressive behavior for instrumental gain) was not a ssessed. Though the two are correlated, reactive aggression has been shown to have a much stronger relationship with hostile attribution of intent (Crick & Dodge, 1996). Assessing ra tes of real-life aggression in this manner may have contributed to a more detailed und erstanding of the effects of hostil e attribution in this study, and any future research on this topic should take th e distinction between subt ypes of aggression into consideration. While it was not the initial purpose of this study to investigate de mographic differences in aggression, hostile attributions or emotion processes, the l ack of gender effects in these constructs is of interest. Tradit ionally, boys have been viewed as more aggressive than girls; however, a growing amount of rese arch suggests that this is di fference is narrowing (Odgers & Moretti, 2002). Although the inclusion of gender in this study was to serve as a possible control variable, more research is needed to specifica lly investigate the presen ce or absence of gender differences in social-emotional-cognitive pathways of this type. 46

PAGE 47

The sample used in this study was drawn exclusively from schools within an urban setting. Aggression levels are higher in urban settings relative to rural settings (Farrel, Sullivan, Esposito, Meyer, & Valois, 2005), which may reflect differences in cultural views about the appropriateness of aggression in certain cont exts (Tolan, Gorman-Smith, & Henry, 2003). However, the fact that anger variables showed a stronger association with antisocial tendencies than did hypothetical variables implies that emo tion processes are contributing to variability in rates of antisocial behavior ove r and above the implementation of what may be considered a culturally-appropriate retaliato ry response. Still, the resu lts of this study may not be generalizable to other samples of early adolescents due to particul ar characteristics of the urban setting. The current research is thought to contribu te to the collective understanding of hostile attributions and anger processes a nd their role in aggression in ear ly adolescence. In addition, the results of this study have implications for th e future use of hypotheti cal vignettes to assess aggression in individuals of this age. Analysis of pattern groups, in particular, is a methodological technique that can ea sily be applied to existing data to evaluate and compare the predictive efficacy of individual vignettes. Besides the inclusion of samples with different demographic characteristics, rese arch in this area should assess the relationships of these core constructs over time. 47

PAGE 48

48 APPENDIX HYPOTHETICAL VIGNETTE SCRIPTS Task: Lunch Table Instruction: Imaging that you get your lunch at school and then walk over to a table. You want to sit at this table. Several other kids are already seated there and there is one empty seat. As you begin to sit down, one of the other kids says, You cant sit ther e. Its taken. A couple of other kids laugh. Question 1: So you were not able to sit at the table. What do you think was going on in the minds of the kids at the table when this happened? Question 2: What would you do or say to th e other kids if th is happened to you? Task: Hallway Instruction: Imagine that you are walking down th e hallway at your school with two other kids on the way to lunch when you see another boy /girl coming toward the three of you from the other end of the hallway. Th ere are lots of kids in the hallway. This other kid yells out, Hey, geek. Yeah, I mean you, nerd. Some of the other kids start laughing. Question 1: So some kids are laughing. What do you think was going on in the mind of the boy/girl when he/she said this? Question 2: What would you do or say to the boy/girl if this happened to you?

PAGE 49

LIST OF REFERENCES Achenbach, T. M., & Edelbrock, C. S. (1986). Teachers report form Burlington, VT. Baron, R. M., & Kenny, D. A. (1986). The moderato r-mediator variable di stinction in social psychological research: Con ceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 1173-1182. Berkowitz, L. (1990). On the formation and re gulation of anger and aggression: A cognitiveneoassociationistic analysis. American Psychologist, 45, 494-503. Borbely, C. J., Graber, J. A., Nichols, T., Brooks -Gunn, J., & Botvin, G. J. (2005). Sixth graders' conflict resolution in role plays with a peer, pare nt, and teacher. Journal of Youth and Adolescence, 34 279-291. Bosson, J. K., & Johnson, A. B. (2006). Interp ersonal chemistry through negativity: Bonding by sharing negative attitudes about others. Personal Relationships, 13 135-150. Botvin, G. J., Schinke, S. P., Epstein, J. A., & Diaz, T. (1994). Effectiveness of culturally focused and generic skills training approach es to alcohol and dr ug abuse prevention among minority youths. Psychology of Addictive Behavior, 8, 116-127. Buss, A. H., & Perry, M. (1992) The aggression questionnaire. Journal of Personality and Social Psychology, 63, 452-459. Coie, J. D., & Dodge, K. A. (1998). Aggression and antisocial behavior. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (5th ed., pp. 779-862). New York: John Wiley & Sons. Cornell, D. G., Peterson, C. S., & Richards, H. (1999). Anger as a pred ictor of aggression among incarcerated adolescents. Journal of Consulting and Clinical Psychology, 67, 108-115. Crain, M. M., Finch, C. L., & Foster, S. L. ( 2005). The relevance of th e social information processing model for understanding relational aggression in girls. Merrill-Palmer Quarterly, 51 213-249. Crick, N. R. (1997). Engagement in gender nor mative versus nonnormative forms of aggression: Links to social-psychological adjustment. Developmental Psychology, 33 610-617. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social informationprocessing mechanisms in childrens social adjustment. Psychological Bulletin, 115, 74-101. Crick, N. R., & Dodge, K. A. (1996). Social info rmation-processing mechanisms in reactive and proactive aggression. Child Development, 67, 993-1002. Damasio, A. R. (1994). Descartes error: Emotion, reason, and the human brain. New York: Avon Books. 49

PAGE 50

Dearing, K. F., Hubbard, J. A., Ramsden, S. R., Parker, E. H., Relyea, N., Smithmyer, C. M., & Flanagan, K. D. (2002). Childrens self-reports about anger regulation: Direct and indirect links to social preference and aggression. Merrill-Palmer Quarterly, 48, 308-336. Dodge, K. A. (1980). Social cognition and children s aggressive behavior. Child Development, 51, 162-170. Dodge, K. A. (1986). A social information processi ng model of social competence in children. In M. Perlmutter (Ed.), Minnesota Symposium in Child Psychology (Vol. 18, pp. 77-125). Hillsdale, NJ: Earlbaum. Dodge, K. A. (1990). The Child Development Project. Nashville, TN: Vanderbilt University. Dodge, K. A., & Coie, J. D. (1987). Social in formation-processing f actors in reactive and proactive aggression in childrens peer groups. Journal of Personality and Social Psychology, 53, 1146-1158. Dodge, K. A., & Frame, C. L. (1982). Social co gnitive biases and deficits in aggressive boys. Child Development, 53, 620-635. Eisenberg, N., Cumberland, A., Spinrad, T. L., Fabe s, R. A., Shepard, S. A., Reiser, M., Murphy, B. C., Losoya, S. H., Guthrie, I. K. (2001). The relations of regulation and emotionality to childrens externalizing and in ternalizing problem behavior. Child Development, 72, 1112-1134. Eisenberg, N., Fabes, R. A., Guthrie, I. K., Murphy, B. C., Maszk, P., Holmgren, R., & Suh, K. (1996). The relations of regulat ion and emotionality to problem behavior in elementary school children. Development and Psychopathology 8, 141-162. Eisenberg, N., Morris, A. S., & Spinrad, T. L. (2005). Emotion-related regulation: The construct and its measurement. In D. M. Teti (Ed.), Handbook of research methods in developmental science (pp. 81-100). Malden, MA: Blackwell Publishers. Elliot, D., Huizinga, D., & Menard, S. (1989). Multiple problem youth: Delinquency, substance use, and mental health problems. New York: Springer-Verlag. Epps, J., & Kendall, P. C. (1995). Hos tile attributional bias in adults. Cognitive Therapy and Research, 19, 159-178. Epstein, J. A., Botvin, G. J., Diaz, T., Baker, E ., and Botvin, E. M. (1997). Reliability of social and personal competence measures for adolescents. Psychological Report, 81, 449-450. Evans, R. I., Hansen, W. B., & Mittelmark, M. B. (1977). Increasing the validity of self-reports of behavior in a smoking in children investigation. Journal of Applied Psychology, 62 521-523. 50

PAGE 51

Farrell, A. D., Sullivan, T. N., Esposito, L. E., Me yer, A. L. & Valois, R. F. (2005). A latent growth curve analysis of th e structure of aggression, drug use, and delinquent behaviors and their interrelations over time in urban and rural adolescents. Journal of Research on Adolescence, 15, 179-204. Graber, J. A., Nichols, T. R., Luppino, C., Ga len, B., Brooks-Gunn, J., Schinke, S., & Botvin, G. J. (2001). Life Skills Training: Coding Manual for Videotape Role-Plays. Center for Children & Families, Teachers College, Columbia University, New York. Guerra, V. S., Asher, S. R., & DeRosier, M. E. (2004). Effect of childre ns perceived rejection on physical aggression. Journal of Abnormal Child Psychology, 32 551-563. Guerra, N. G., Huesmann, L. R., & Zelli, A. (1990). Attributions for social failure and aggression in incarcerated delinquent youth. Journal of Abnormal Child Psychology, 18 347-355. Hudley, C., & Graham, S. (1993). An attribut ional intervention to reduce peer-directed aggression among African-American boys. Child Development, 64, 124-138. Lansford, J. E., Malone, P. S., Dodge, K. A., Crozie r, J. C., Pettit, G. S., & Bates, J. E. (2006). A 12-year prospective study of patterns of so cial information processing problems and externalizing behaviors. Journal of Abnormal Child Psychology, 34 715-724. Lengua, L. J. (2003). Associations among emotiona lity, self-regulation, adjustment problems, and positive adjustment in middle childhood. Applied Developmental Psychology, 24, 595-618. Lemerise, E. A., & Arsenio, W. F. (2000). An integrated model of emotion processes and cognition in social in formation processing. Child Development, 71, 107-118. Lsel, F., Bliesener, T., & Bender, D. (2007). Social information proc essing, experiences of aggression in social contexts, and ag gressive behavior in adolescents. Criminal Justice and Behavior, 34 330-347. Moffit, T. E. (1993). Adolescence-limited a nd life-persistent an tisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701. Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in antisocial behavior. Cambridge: Cambridge University Press. Nichols, T. R., Graber, J. A., Brooks-Gunn, J., & Botvin, G. J. (2006). Sex differences in overt aggression and delinquency among urba n minority middle school students. Applied Developmental Psychology, 27, 78-91. Nichols, T. R., Graber, J. A., Byrne, C., L uppino, C., Reigada, L., Caskey, E., Brooks-Gunn, J., & Botvin, G. J. (2001). Life Skills Training: Coding Manual for Social Problem Solving: Adolescent Version. Center for Children & Families, Teachers College, Columbia University, New York. 51

PAGE 52

Novaco, R. B. (1994). Anger as a risk factor fo r violence among the mentally disordered. In J. Monahan & H. J. Steadman (Eds.), Violence and mental disorder : Developments in risk assessment (pp. 21-59). Chicago: University of Chicago Press. Odgers, C. L., & Moretti, M. M. (2002). Aggressive and antisocial girls: Research update and challenges. International Journal of Fo rensic Mental Health, 1 (2), 103-119. Orobio de Castro, B., Merk, W., Koops, W., Veerman, J. W., & Bosh, J. D., (2005). Emotions in social information processing and their re lationships with reactive and proactive aggression in referred aggressive boys. Journal of Clinical Child and Adolescent Psychology, 34, 105-116. Orobio de Castro, B., Veerman, J. W., Koops, W. Bosch, J. D., & Monshouwer, H. J. (2002). Hostile attribution of intent and aggressive behavior: A meta-analysis. Child Development, 73, 916-934. Rose, A. J., Swenson, L. P., & Waller, E. M. (2004). Overt and re lational aggression and perceived popularity: Developmental differences in concurrent and prospective relations. Developmental Psychology, 40 378-387. Saarni, C. (1999). The development of emotional competence. New York: Guilford. Saltaris, C. (2002). Psychopathy in juvenile o ffenders: Can temperament and attachment be considered as robust developmental precursors? Clinical Psychology Review 22: 729-52. Spielberger, C. D., Jacobs, G. A., Russell, S. F ., & Crane, R. J. (1983). Assessment of anger: The State-Trait Anger Scale. In J. Bu tcher & C. D. Spielberger (Eds.), Advances in personality assessment (2nd ed., pp. 159-187). Hillsdale, NJ: Erlbaum. Stevens, J. (1996). Applied multivariate statistics for the social sciences. Mahwah, NJ: Lawrence Erlbaum Associates, 362-392. Sukhodolsky, D. G., Golub, A., Stone, E. C., & Orban, L. (2005). Dismantling anger control training for children: A randomized pilot study of social problem-solving versus social skills training components. Behavior Therapy, 36, 15-23. Tolan, P. H., Gorman-Smith, D., & Henry, D. B. (2003). The developmental ecology of urban males youth violence. Developmental Psychology, 39, 274-291. Tourangeau, R., Smith, T. W., & Rasinski, K. A. (1997). Motivation to report sensitive behaviors on surveys: Evidence from a bogus pipeline experiment. Journal of Applied Social Psychology, 27, 209-222. Tremblay, P. F., & Ewart, L. A. (2004). The Buss and Perry Aggression Questionnaire and its relations to values, the Big Five, provoking hypothetical situations, alcohol consumption patterns, and alcohol expectancies. Personality and Individual Differences, 38, 337-346. 52

PAGE 53

53 VanOostrum, N., & Horvath, P. (1997). The e ffects of hostile attr ibution on adolescents aggressive responses to social situations. Canadian Journal of School Psychology, 13, 48-59. Vitaro, F., & Pelletier, D. (1991). Assessment of childrens social problem-solving skills in hypothetical and actual conflict situations. Journal of Abnormal Child Psychology, 19, 505-518. Vitaro, F. Pelletier, D., & Contu, S. (1989). Impact of a negative social experience on the social reasoning process of aggre ssive-rejected children. Perceptual and Motor Skills, 69, 371-382. Zeman, J., Shipman, K, & Suveg, C. (2002). Ange r and sadness regulati on: predictions to internalizing and externalizing symptoms in children. Journal of Clinical Child & Adolescent Psychology, 31, 393-398.

PAGE 54

BIOGRAPHICAL SKETCH Katherine Clemans is a third-year graduate student in developmental psychology at the University of Florida. She received a Bachelor of Arts in psychology a nd a certificate in human development from Duke University in 2002. She ha s also completed work toward a Master of Arts in Liberal Studies degree from Dartmout h College. Her research focuses on emotional correlates of aggression and moral judgment in adolescents and young adults Currently she is a co-investigator of the APEX (Adolescent Pe er Experiences) Study, which examines peer relationships and their influence on psychosoc ial development during the middle school years. 54