INTEGRATING SOCIAL COMPETENCE, SELF ESTEEM, AND ANGER AS PREDICTORS OF AGGRESSION; A LONGITUDINAL EXAMINATION OF YOUNG, URBAN ADOLESCENTS By ALLISON METZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2013 Allison Metz
To my wonderful family
4 ACKNOWLEDGMENTS I thank my chair and committee f or their mentoring and guidance through this process. I also thank my family for their support and encouragement to achieve this goal.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 The Social Competence Model ................................ ................................ ............... 12 Measuring Social Competence ................................ ................................ ......... 14 Self Esteem ................................ ................................ ................................ ...... 16 Anger ................................ ................................ ................................ ................ 17 The Present Study ................................ ................................ ................................ .. 19 2 METHOD ................................ ................................ ................................ ................ 20 Participant s ................................ ................................ ................................ ............. 20 Procedure ................................ ................................ ................................ ............... 21 Measures ................................ ................................ ................................ ................ 24 Outcome Measures ................................ ................................ .......................... 24 Predictors ................................ ................................ ................................ ......... 24 Self Report Measures ................................ ................................ ...................... 27 Covariates ................................ ................................ ................................ ........ 27 Analysis Plan ................................ ................................ ................................ .......... 27 3 RESULTS ................................ ................................ ................................ ............... 30 Regressions Predicting Concurrent Aggression ................................ ..................... 31 Regressions Predicting Change in Aggression from Sixth to Seventh Grade ......... 32 Regressions Predicting Change in Aggression from Seventh to Eighth Grade ....... 33 4 DISCUSSION ................................ ................................ ................................ ......... 43 Limitations ................................ ................................ ................................ ............... 46 Implications ................................ ................................ ................................ ............. 48 APPENDIX SOCIAL COMPETENCE SUBGROUPS ................................ ................................ 50 LIST OF REFERENCES ................................ ................................ ............................... 55
6 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 60
7 LIST OF TABLES Table page 2 1 Attrition table ................................ ................................ ................................ ...... 29 3 1 Means and standard deviations f or primary variables ................................ ....... 36 3 2 Correlations of primary variables ................................ ................................ ....... 37 3 3 Regression analysis summary for social competence indicator s at sixth grade predicting concurrent aggression ................................ ............................. 38 3 4 Regression analysis summary for social competence indicators at seventh grade predicting concurrent aggression ................................ ............................. 39 3 5 Regression analysis summary for social competence indicators at eighth grade predicting concurrent aggression ................................ ............................. 40 3 6 Regression analysis summary for social competence indicators predicting change in aggression from sixth to seventh grades ................................ ............ 41 3 7 Regression analysis summary for social competence indicators predicting change in aggressio n from seventh to eighth grades ................................ ......... 42
8 LIST OF FIGURES Figure page A 1 Descriptive comparison of social competence subgroups on aggression .......... 54
9 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 INTEGRATING SOCIAL COMPETENCE, SELF ESTEEM, AND ANG ER AS PREDICTORS OF AGGRESSION; A LONGITUDINAL EXAMINATION OF YOUNG, URBAN ADOLESCENTS By Allison Metz December 2013 Chair: Julia A. Graber Major: Psychology This study examined whether social competence, self esteem, and anger were predictors of direct aggression cross sectionally and longitudinally; moderating effects were also tested. Participants were 453 young, urban, ethnically diverse adolescents drawn from the control group of a larger randomized clinical trial of the Life Skills Training program a school based drug prevention program adapted to target prevention of violence and aggression. Data were collected at baseline (6 th grade), one year follow up (7 th grade), and two year follow up (8 th grade), and included surveys to assess aggression, se lf esteem and anger, and coded videotaped role plays, used to assess social competence, specifically communication and conflict resolution skills, at each time point. A series of hierarchical regressions were used to test the effects of social competence, self esteem, and anger on aggression, both cross sectionally and longitudinally. Additional hierarchical regressions were used to test moderating effects. The main findings of the study were that high anger and low self esteem consistently predict aggressi on in this sample, both cross sectionally and over time, with the exception of eighth grade self esteem, which was not predictive of concurrent
10 aggression. Anger, self esteem, and aggression were not associated with social competence, and anger and self es teem did not interact with one another or with the social competence variables. These results contribute to the available literature on the relationships between anger, self esteem, and aggression, particularly by providing support for the low self esteem hypothesis, which has been met with mixed findings in the literature. In addition, these results suggest that there is a need for standardized methods for assessment of social competence. The development of standardized methods for social competence assess ment is important as prevention programs, such as Life Skills Training, commonly teach these skills. Thus, it would be useful to examine whether social competence is salient to the success of the programs and to individual outcomes.
11 CHAPTER 1 INTRODUCTIO N Despite a wealth of studies on the etiology of aggression during childhood and adolescence, as well as several successful prevention initiatives, aggression continues to be a problem during adolescence. Aggression has broadly been defined as a set of beh aviors intended to cause harm to another individual when it is understood that that person wants to avoid said harm (Joireman, Anderson, & Strathman, 2003). Early adolescence is a particularly important period for the study of aggressive behaviors, as this stage of development has been linked to overall increases in aggressive and antisocial behavior. Although rates of aggression are higher in boys than girls, during adolescence, aggressive behavior has been found to increase for both genders for many indiv iduals (Dodge, Coie, & Lynam, 2006; Moffitt, 1993). Furthermore, research suggests that patterns of aggressive behavior during adolescence are linked to many other problems that can impact current functioning, including academic achievement, social skills deficits, and psychosocial problems (Moffitt, 1993; Crick & Dodge, 1994; Fraser Galinsk, Smokowski, Day, Terzian, Rose, & Guo, 2005; Webster Stratton & Lindsay, 1999). Many research studies and prevention programs have aimed to prevent the development of these negative trajectories through a focus on enhancing social competence via social skills training. Other individual factors, such as self esteem and trait anger, have been identified as predictors of aggression as well and could serve as potential tar gets of intervention (Fraser, et al., 2005; Graber, Nichols, Lynne, Brooks Gunn, & Botvin, 2006; Prinstein & La Greca, 2004). However, fewer studies have examined how these factors may interact in predicting adolescent aggression over time.
12 Hence, the pur pose of the present study is to investigate the interaction between early esteem in predicting overt aggressive behavior at 6th grade and over time using the social competence model. This study is uniqu e as it incorporates individual differences in emotion (trait anger) and self esteem as moderators of competence in communication and conflict resolution in predicting aggression, and examines positive development as well as aggression in a young, urban, m inority sample of adolescents. It was expected that low social competence in communication would predict aggression, and that this relationship might be moderated by conflict resolution skills, anger, and self esteem such that low social competence in comm unication would be most strongly predictive of aggression in those adolescents with low conflict resolution skills, high anger, and low self esteem. The Social Competence Model Competence has been defined as the achievement of the psychosocial functioning ability to meet new developmental demands ( Masten & Curtis, 2000) As such, competence is not a singular construct but varies depending on the domain being assessed. Domains of compet ency commonly examined in adolescence include social competence as indicated by success in peer relationships or social interactions, behavioral competence often indicated by rule and law abiding conduct (Masten, Coatsworth, Neemann, Gest, Tellegen, & Gar mezy, 1995), and positive self evaluation as indicated by self esteem. Competence in many domains remains relatively stable through childhood, though findings vary regarding the stability of competence during the transition to and progression through adol escence (Masten et al., 1995; Monahan & Steinberg, 2011). Monahan and Steinberg (2011) suggest that these mixed findings
13 may be due to changes in the nature of social competence, in particular, suggesting that pubertal changes and transitions into new sch ool environments may accentuate individual differences in social competence during early adolescence. Furthermore, domains of competence likely influence one another in that accentuation of individual differences in social competence may impact changes in behavioral competence during early adolescence. Behavior competence, like law abiding behavior, is often defined by the absence of this behavior such that aggressive behaviors or related behavior problems may be considered one indicator of low competence It is important to note, however, that although some studies have found relationships between indicators of social competence and behavioral competence others have not found these associations. These mixed finding may be attributed to differences in def ining and measuring both behavioral and social competence. For example, Masten et al. (1995) measured behavioral competence by trouble with the law (arrests, etc.) whereas other studies that found links between social and behavioral competence focused on a ggression as the indicator of behavior (Fraser et al., 2005; Webster Stratton & Lindsay, 1999). As aggressive behavior does not always result in problems with the law, legal based indicators of behavior may be too narrow in scope. As indicated, social sk ills training, aimed at increasing social competence, is commonly used in treatment programs for aggressive individuals, as well as in a variety of prevention programs, including those targeting violence and aggression ( Botvin, 2004); these programs includ e training on social competence skills such as effective communication and conflict resolution Such approaches are based on findings that social competence has been linked to aggression, and that training on specific aspects
14 of social competence can offse t development of aggression and other problem behaviors. Although the literature is modest, there is empirical evidence to suggest that acquired social competencies or skills may be protective against adolescent risk taking, including aggressive and delinq uent behavior, and that conflict resolution skills in particular may be protective for these behaviors (LaRusso & Selman, 2011). The majority of these studies, however, have focused on high school age adolescents, leaving questions about the links between risk behaviors and social competencies like general communication and conflict resolution skills in young adolescent samples. Very few studies, however, have examined whether conflict resolution skills in particular have this buffering effect on aggression Furthermore, the literature that does address conflict resolution as it relates to aggression looks specifically at either parent adolescent conflict or at conflict in romantic relationships and findings have been mixed (e.g., Branje, van Doorn, van der Valk, & Meeus, 2009; Ha, Overbeek, Cillessen, & Engels, 2012). Thus, the present study contributes to this literature by examining the relationship between aggression and conflict resolution skills in a variety of scenarios (parent, teacher, and peer). The social competence model predicts healthier behavioral trajectories among those with better communication and conflict resolution skills. In keeping with this model, the present study examines social competence, particularly focused on competence in commun ication and conflict resolution skills as predictors of aggression. Measuring Social Competence As with behavioral competence, social competence is measured in a variety of ways. Common methods include the use of self report, and report by parents, peers or teachers. There are a number of problems with these methods, however. First, the
15 importance of evaluating social competence within varying social contexts is being increasingly recognized (Wright, Nichols, Graber, Brooks Gunn, & Botvin, 2004). For ex necessarily mean that those conflict resolution skills will transfer to other situations. To ariety of situations must be examined. Second, many studies of competence rely on self report data. Self report instruments generally ask about expected responses to social self reporting of p ossible behavior may not, on it s own, provide the most accurate representation of social competence that is possible to achieve. Hence, one approach to assessing adolescent social competence is through direct observation; for example, by observing adolescents engage in role plays of conflict situations. Role play tasks can be used to provide an additional element of realism to social competence data, as they are less influenced by perceptual biases and because they put participants in the situations that they have to respond to on the spot. In this way, role play tasks allow more direct, naturalistic observation for the purposes of data collection than other methods are able to, while also providing enough control to collect more standa rd data (Foster, Inderbitzan, & Nangle, 1993). It is important to note that since role plays are still hypothetical situations, they are not assessments of real world app ropriate or inappropriate responses on the spot (Foster et al., 1993; Wright et al., 2004). Although role play tasks have proven to be an effective way of studying social
16 com petence (Borbely, Graber, Nichols, Brooks Gunn, & Botvin, 2005; Wright et al., 2004). The present study will examine social competence using observed measures of communication and conflict resolution skills exhibited while adolescents try to resolve or n egotiate a conflict with someone else. It is expected that adolescents with low social competence will have higher rates of engagement in aggressive behavior than those with high social competence, and that competence in social interactions may buffer agai nst the development of aggressive behaviors over time. Self Esteem As indicated, self esteem is often conceptualized as a dimension of competence ( Baumeister, Campbell, Krueger, & Vohs, 2003; Harter, 1999; Graber et al., 2006). To use the simplest definit ion, self esteem is the evaluative component of self concept. The role of self esteem in predicting aggression has received a resurgence of attention in the last few decades, with inconsistent findings resulting in two opposing perspectives (for a review, see Ostrowsky, 2010). On the one hand, researchers have found evidence for what has been referred to as the threatened egotism hypothesis, which suggests that high self esteem is most strongly predictive of aggressive behavior, particularly when the indivi esteem is threatened or directly disputed, and when it occurs in combination with narcissism (e.g. Baumeister, Bushman, & Campbell, 2000; Baumeister, Smart, & Boden, 1996; Bushman & Baumeister, 1998) On the other hand, when not linked to factors like narcissism, high self esteem may be related to esteem may be a reflection of higher perceived competence, and therefore, low self esteem reflects lower perceived competence. Support for the low self esteem hypothesis,
17 which proposes that aggression and other antisocial behaviors are a manifestation of esteem, has been found in several prior studies (e.g. Donnellan, Trzesniewski, Robins, Moffitt & Caspi, 2005; Sutherland & Shepherd, 2002; Lochman & Dodge, 1994) It is important to note that self esteem decreases for many young and mid adolescents (e.g., Simmons & Blyth, 1987), perhaps as a result of problems in successfully adapting to the tran sitions and new social demands that occur during this period and in turn, to problems in achieving competence (Graber et al., 2006). Findings regarding self esteem and social competence fit this conceptualization, suggesting that low self esteem is related to lower social competence. For example, in a study of first year college students, Crocker and Luhtanen (2003) found that low levels of self esteem were predictive of social problems. Lochman and Lampron (1986) also found relationships between self estee m and social competence in aggressive and non aggressive boys, with highly aggressive boys having poorer self esteem and deficits in social problem solving skills. In line with these findings, it is expected that self esteem will be related to both observe d social competence and aggression, with adolescents who report low self esteem also showing low social competence, and higher reports of aggression. Anger Notably, the ability to develop competence within and across domains is potentially facilitated or impeded by individual characteristics that each adolescent brings to new developmental challenges. Several studies have identified the propensity for anger as an individual factor that predicts aggression (Wilkowski & Robinson, 2008) and appears to incre ase during early adolescence (e.g., Nichols et al., 2006). Since Speilberger and
18 his colleagues adapted the state trait personality theory to anger (e.g. Spielberger, 1988; Spielberger, Krasner, & S olomon, 1988; del Barrio, Aluja, & Spielberger, 2004), the construct has been understood to include two distinct, but related, elements: state/situational anger and trait anger. Situational anger is a transitory, emotional physiological response that occurs in response to an immediate situation. Trait anger, howev er, refers to a stable dimension of personality that is characterized by a tendency to experience state anger frequently and at heightened levels (Spielberger, 1972; Deffenbacher, Oetting, Lynch, & Morris, 1996). While situational anger may commonly co occ ur with aggression, trait anger has been found to be more strongly and consistently associated with negative outcomes such as substance abuse, antisocial and aggressive behavior, and deficits in social competence, as well as diminished psychological and ph ysical well being (e.g. Deffenbacher et al., 1996; Fives, Kong, Fuller, & DiGiuseppe, 2011; Mahon, 2000; Nichols et al., 2006 ). As such, trait anger adolescent social comp etence and the development of aggression. Runions and Keating (2010) found that anger had a moderating effect on the association between social information processing and aggression. Additionally, in a study with preschoolers, children high in aggression were found to have higher anger and were less compliant and more noncompliant as compared to socially competent children (Kotler & McMahon, 2002). Similarly, researchers have demonstrated that social problem solving deficits relate to anger as well as to a ggression in children and adolescents (see the following articles for a more complete review: Fives et al., 2011; Webster Stratton & Lindsay, 1999; Keltikangas Jarvinen & Kangas, 1988). Altogether,
19 considerable research supports the links between social sk ill deficits, anger, and aggressive behavior. The Present Study In the present study we investigated the relationships between social competence, self esteem, anger, and aggression in a sample of urban minority youth. Urban minority youth are of particular interest in this study as aggression is found to be common in this population and several studies show higher rates of aggression among urban minority adolescents (Williams, Fredland, Han, Campbell, & Kub, 2009; Reingle, Maldonado Molina, Jennings, & Komr o, 2012 ) Nonetheless, urban minority youth remain an underrepresented population in the literature on change in predictors of aggression over time. In sum, it was expected that low social competence would predict aggression and that resolution skills, an ger, and self esteem would relate to both aggression and social competence. It was also expected that resolution skills, anger, and self esteem might moderate the relationship between observed social competence and aggression such that low social competenc e would be most strongly predictive of aggression in those early adolescents with low conflict resolution skills, low self esteem, and high anger. These hypotheses were tested both cross sectionally at 6 th 7 th and 8 th grades and longitudinally across 6 th 7 th and 8 th grades. This study is unique in its focus on communication competence and resolution skills, and its examination of the potential buffering effect of self esteem in predicting positive development versus problem behaviors in a young, urban ad olescent sample.
20 CHAPTER 2 METHOD Participants Participants were 453 young adolescents drawn from the control group of a larger randomized clinical trial targeting prevention of violence and aggression. Data were collected at baseline (6 th grade), one year follow up (7 th grade), and two year follow up (8 th grade). Given that the present study includes only those participants in the control group of the larger study and as a result of attrition, the sample dropped to 278 at 7 th grade and 245 at 8 th grade (Table 2 1). Mean age was 11.65 years (SD = .49) at 6 th grade. Approximately half the sample was female (50.4%). Most of the students were African American (46.7%) or Latino/Hispanic (24.5%), with the rest of the sample, reporting Caucasian (15.6%), or ot her (13.2%, including Asian and Pacific Islander, multiracial, and other/not specified) for race/ethnicity. A majority of the students in the present sample lived in non blended two parent homes (70% in 6 th grade). The rest of the sample lived in other hou sehold configurations (30%). The present investigation is a sub study from a larger randomized clinical trial evaluating a school based Life Skills Training program focused on preventing violence and aggression. A total of 42 public and parochial schools participated in the larger study. All schools participated in baseline data collection and annual surveys; in addition, half of the schools received the Life Skills Training prevention program for three years, while the other half (the control group) rece ived a five session drug prevention course consisting of only informational components rather than interactional. Archival public school records of all participating schools showed that the majority (88%) of schools had greater than 65% student eligibility for free or reduced lunch. The
21 sub study was conducted to examine social skills, and due to the intensive nature of this supplemental data collection only the smallest schools (those with fewer than 150 sixth grade students) were recruited, most of which were parochial schools. Thus the majority of the students in the current sample attended parochial schools (69.7%), which served the same population as the public schools in the larger study. Procedure Parental consent for the larger study was obtained us ing a waiver for active Letters in English and Spanish were sent to parents regarding both the survey and supplemental data collection; letters gave parents information a bout the nature of the study and provided them with an opportunity to disallow their child from participating. Parents and students were told that the project was a study designed to help develop health programs at the school. In the consent form for the p arents, general descriptions were included which explained the types of items that would be on the survey (behaviors, attitudes, and beliefs regarding smoking, drinking, and violence). The consent form was distributed to students in the schools and was als o mailed directly to consent document. In addition, a confidentiality certificate was obtained from the National Institute of Health in order to protect participants fr om any legal action that could be taken against them from admitting to some of the behaviors reported in the survey. In the larger randomized trial, there were 477 Negative Consents from parents during the first year of data collection. Students whose par ents responded that they did not want them to participate did not complete any of the data collection activities. In
22 addition, negative assents (or verbal refusals) from the students ranged from 4 68 throughout middle school. The same consent and protocol procedures were used in each wave of data collection. Annually in 6 th 7 th and 8 th grades, the self report survey was administered on two separate days during regular class time by an ethnically diverse team of three to five data collectors. Data collecti on followed a standardized protocol similar to those used in previous research (e.g. Botvin, Schinke, Epstein, & Diaz, 1994). To maintain the confidential nature of the questionnaire, identification codes were used in place of names on the surveys and stud ents were assured about the confidentiality of their responses. In addition to self report data, carbon monoxide samples were collected at each assessment, utilizing a variant of the bogus pipeline procedure (Evans, Hansen, & Mittlemark, 1977). This measur e was used to increase the validity of questions regarding cigarette smoking, but studies have shown that bogus pipeline procedures can increase the validity of other self reported problem behaviors as well (Tourangeau, Smith, & Rasinski, 1997). The third day of data collection occurred at each assessment point only in the subset of 17 schools that agreed to participate in the supplementary activities. For the sub study, general instructions and consent information were first read to students in the classr oom as a group. Then students were individually escorted to a private room provided by the school for data collection where they were read specific instructions about the role play tasks. Students completed videotaped role plays one on one with a confedera te. Students were asked to engage in interaction as they would in real life based on the given prompts. Data collectors/confederates for these activities were
2 3 trained undergraduate and graduate students who were primarily African American and Hispanic and female. At least one male data collector was sent to each school as confederates in role plays were matched by gender with participants. Students completed seven role play tasks at each assessment point. The order of presentation of the role play tasks was randomized across students. In the present study, we focus on the three tasks assessing skills in conflict resolution due to the salience to aggression. These included a role play in which the confederate acted as a parent, one in which the confederate acted as a teacher, and one in which the confederate acted as a peer. In the parent task, students were given the following allowance. You get $7.50 a week now, and you think y ou should get at least $10 a teacher and I have just given you a lower grade on your report than you feel you her. You want to go to the plays followed the same format and used the same prompts at each assessment point. Role play tasks were coded for a range of behavioral and emotion al responses (Graber et al., 2001). Coders were trained to at least 85% agreement (within one point for Likert type scales or exact agreement for categorical coding) with the gold standard codes; 20% of codes were checked by the gold standard coder for on going reliability.
24 Measures Outcome Measures Aggression (past month) reported rates of aggression were measured at all times of assessment using ten items from the aggression scale of the Youth Self Report (Achenbach & Edelbrock, 1986). St udents were asked how many times in the past month they engaged in each example of overtly aggressive behavior. es were on a 5 point scale from 1 (Never), 2 (Once), 3 (2 3 times), 4 (4 5 times), and 5 (More than 5 times). Items were rescored onto a 0 to 4 scale and then summed to create an overall aggression score for each time of assessment where higher scores indi cate more aggressive behaviors in the past month (6 th grade = .92; 7 th grade = .93; 8 th grade = .94). Predictors Social Competence: Communication. The ways in which students were able to express themselves during the conflict resolution role play tas ks was coded separately for each role play in each year using a global score for communication coded on a 1 5 Likert type scale (1 = not at all competent, 5 = highly competent). To assess the degree to which a participant exhibits overall skill in communic ation, the global score considered the following behaviors: Specific in requests and suggestions : the adolescent states their position clearly and concisely. Does not ask or suggest in a roundabout or confused way; Uses active listening skills : the adolesc ent maintains eye appropriate; Asks questions
25 point of view; : adolescent ma kes a Shows respect : the adolescent shows respect for the other r other forms of politeness; Paraphrases for clarity : adolescent repeats back all or part of Stays in control : if the adolescent begins to feel frustrated with the situation s/he does something (i.e. takes a deep breath) to stay in control and calm him/herself down. The skills examined in the present study were selected to match the specific social skills taught in the prevention program, Life Skills Training, a school based drug abuse and violence prevention program that the larger study was designed to evaluate (Graber et al. 2001). As indicated, these behaviors were the basis for a single score for communication skills in each vignette. Interrater agreement (the percent agreement within one point between coder and gold standard scores) for the competence scale was 99% ( = 87), 99% ( = 91), and 98% ( = 85) at 6 th grade; 98% ( = 77), 95% ( = 64), and 100% ( = 1.0 ) at 7 th grade; and 100% ( = 1.0 ), 98% ( = 76), and 100% ( = 1 0) at 8 th grade for the peer, parent, and teacher vignettes, respectively. Mean scor es for communication skills competence across the three vignettes at each grade were created as well. As with the global vignette scores, communication scores by grade were on a 1 5 point scale with higher scores indicating more skill in verbal communicati on. Social Competence: High Resolution Skills Quality of resolution skills was assessed separately for each of the three conflict resolution role play tasks. Behaviors
26 were initially coded into eight categories: Gave in ; Gave ineffective reason and/or re quest ; Gave effective reason/request ; Offers one suggestion for compromise ; Negotiated compromise ; Confederate ignores/accepts suggestion ; and Confederate suggestion accepted; and Confederate suggestion negotiated When occasional confederate errors were m ade, these errors were recorded and participant behaviors ability to demonstrate skill, otherwise they were coded as missing. Interrater agreement (the percent of exact m atches between coder and gold standard scores) for the resolution scale in the parent, teacher, and peer vignettes was 88% ( = 1.0 ), 78% ( = 82), and 86% ( = 85) at 6 th grade; 92% ( = 81), 89% ( = 83), and 90% ( = .81 ) at 7 th grade; and 90% ( = .89 ), 93% ( = 83), and 91% ( = .94 ) at 8 th grade, respectively. Because programs that promote effective conflict resolution focus on compromise as the desired outcome in conflict scenarios, only Offers one suggestion for compromise ; Negotiated compromis e ; Confederate suggestion negotiated were considered to be indicative of higher quality resolution skills (Graber et al. 2001); hence original codes for conflict were re coded as 1 for the compromise behaviors and 0 for all other codes for each vignette f or each grade. Resolution skills scores were summed within grade to create a 0 3 point scale, with higher scores indicating more instances of student initiated compromise during conflict resolution tasks at that assessment. As with the communication compet ence skills, the conflict resolution skills examined in the present study were selected to match the specific social skills taught in the prevention program that the larger study was designed to evaluate (Graber et al. 2001).
27 Self Report Measures Self Es teem. Self esteem was measured using five items from the scale (Strongly Disagree ) to 4 (Strongly Agree). Items were summed such that higher sc ores indicate high self esteem (6 th grade = .81; 7 th grade = .86; 8 th grade = .88) Anger Anger was measured with the seven the Buss and Perry (1992) A ggression Questionnaire. Students were asked to rate how sponse categories ranged from 0 (Really Not True for Me) to 4 (Really True for Me). Items were summed such that higher scores indicate greater anger (6 th grade = .74; 7 th grade = .72; 8 th grade = .68). Covariates analyses. Age was entered as a continuous variable, and race/ethnicity was dummy coded into three racial groups: African American/Black, Hispanic, and Other, with Caucasian as the omitted group. Analysis Plan Descriptive statistics including means, skewness, and kurtosis were examined to ensure that all variables were normally distributed. Correlations were also examined to assess the associations between the variables and to test for multicollinearity between the predictors. Ordinary least squared regressions were performed using social competence indicators (communication and resolution skills), self esteem, and anger
28 pred icting aggression both cross sectionally and longitudinally. Cross sectional analyses at each assessment point were examined to determine whether competence in communication and resolution skills, anger, and self esteem predicted concurrent aggression. Sp ecifically, the first block in regression models included covariates consisting of gender, dummy coded race/ethnicity variables, age, and school type, followed by 6 th grade social competence measures, self esteem, and anger in the second step. Centered int eraction terms were entered in the third step, as per the specifications of Baron and Kenny (1986) for testing moderation. Similar analyses were conducted for 7 th and 8 th grade time points. Following the concurrent analyses, longitudinal analyses were cond ucted to examine whether prior competence in communication and resolution skills, anger, and self esteem predicted change in aggression. To analyze change over time in aggression from 6 th to 7 th grade, similar models were used, but in this case the 6 th gra de aggression variable was entered on the second step in the model predicting 7 th grade aggression, following the covariates in the first step. As before, the main effects were tested by entering 6 th grade predictors (social competency measures, self estee m, and anger) in the third step, followed by centered interaction terms in the final block. Change in aggression from 7 th to 8 th grade was also examined, following the same procedure, with 7 th grade aggression entered in Block 2, predictors entered in Bloc k 3, and centered interaction terms in the final block.
29 Table 2 1. Attrition table 6 th grade a 7 th grade b 8 th grade b n Overall 453 278 245 n Aggression 453 278 245 n Anger 441 257 244 n Self esteem 431 246 241 M age at video data collect ion 11.8 13 14 % Female 50.4 51.1 51 % Black 46.7 44.2 44.9 % Hispanic/Latino 24.5 28.4 27.4 % White 15.6 16.2 17.1 a Includes control and intervention groups b Includes control group only
30 CHAPTER 3 RESULTS Means and standard deviations for core co nstructs can be found in Table 3 1 Results of repeated measures analysis of variance with covariates and test for quadratic effects show that the average number of incidents of aggression increased significantly from sixth grade (M = 12.46, SE = .66) to s eventh grade (M = 18.23, SE = .80), F (1, 207) = 4.84, p < .05, but not from seventh to eighth grade (M = 20.68, SE = .82), F (1, 208) = 1.39, p > .05. In sixth grade 40% of the participants reported never having engaged in an incidence of aggression in th e past month, while in seventh and eighth grades 23.2% and 18.5%, respectively, reported never having done this in the past month. These rates are consistent with those reported in a prior study examining the larger sample from which the observational stud y sample was drawn (Nichols et al., 2006). According to results of repeated measures controlling for covariates, rates of anger and self esteem did not increase significantly across grades. As can be seen in Table 3 1 participants showed moderate levels o f social competence across all time points. Correlations among the constructs are shown in Table 3 2 Scores for aggression, anger, and self esteem each demonstrated moderate to strong associations within construct across grades (e.g., r s for aggression ranged from .45 to .67, p < .01). In contrast, competence in communication scores in sixth, seventh, and eighth grade were weakly to moderately correlated ( r s ranging from .23 to .28, p < .01), as were competence in resolution skills scores ( r s ranging fr om .20 to .36, p < .01). As expected, anger was significantly positively correlated with aggression at each time point and self esteem had significant negative correlations with aggression at each time point. Social
31 competency skills (communication and res olution quality) demonstrated less consistent associations with aggression within time points, with no association between communication competence or resolution quality in sixth grade and negative significant associations in seventh and eighth grade for b oth social competence indicators. Few associations were found among anger, self esteem, and indicators of social competence at each grade (Table 3 2 ). Regressions Predicting Concurrent Aggression As indicated, regressions were used to analyze the relatio nships between social competence indicators, anger, and self esteem in sixth grade with concurrent levels of aggression. Three separate regressions were run at each grade. These analyses included overall communication competence and overall resolution qual ity along with anger and self esteem as predictors of concurrent aggression. As communication and resolution quality were not strongly correlated, both scores were included on the same step in a single model. For 6 th grade, as expected, anger and self es teem were significant predictors of aggression (Table 3 3 ). Specifically, having higher levels of anger in sixth grade was significantly associated with higher levels of aggression in sixth grade. Conversely, having lower levels of self esteem in sixth gra de was associated with higher levels of concurrent aggression. Sixth grade social competence in communication and resolution skills did not significantly predict levels of aggression in sixth grade. This is not surprising as preliminary analyses showed tha t sixth grade social competence and sixth grade aggression were not significantly correlated. The final R 2 of the sixth grade model suggest that approximately 29% of the variance in levels of aggression was explained by the model. The covariates (gender, r ace, and school type) were not significantly
32 associated with levels of aggression in the final model, with the exception of being black, which was associated with slightly higher levels of aggression. Additional regressions analyzed the effects of the inte ractions of sixth grade social competence in communication and resolution skills and sixth grade self esteem and anger on concurrent aggression. However, none of the interaction effects were significant. Following the same method described above, separat e regressions were conducted to examine concurrent aggression at seventh and eighth grades as well. As was found in the sixth grade analyses, in seventh grade anger and self esteem were significantly associated with aggression such that higher anger and lo wer self esteem predicted higher aggression (Table 3 4 ). For eighth grade, although higher anger was associated with increased aggression, self esteem was not a significant predictor (Table 3 5 ). Social competence indicators (communication and resolution s kills) did not significantly predict concurrent levels of aggression in seventh grade, but did show moderate associations at eighth grade. The final R 2 s suggest that for seventh grade, the model explained about 19% of the variance in aggression, while for eighth grade, approximately 24% of the variance in aggression was explained. Additional regressions were conducted to analyze the effects of interactions of communication competence, resolution skills, anger, and self esteem on concurrent aggression at s eventh and eighth grade. None of these interactions were significant. Regressions Predicting Change in Aggression from Sixth to Seventh Grade Another set of regressions analyzed the relationship between sixth grade social competence indicators (communica tion and resolution skills), anger, and self esteem with the change in aggression from sixth to seventh grade (Table 3 6 ). Longitudinal analyses were conducted because the constructs examined in this study have been
33 found to change during this developmenta l period. In particular, overall rates of aggression have been found to increase during the transition to and progression through adolescence. In fact, this pattern was observed in the present study, with aggression increasing from sixth to eighth grade, a lthough it did not reach significance. Thus, it is valuable to examine these changes in aggression over time from a longitudinal perspective. Results of the regressions predicting change in aggression from sixth to seventh grade showed that sixth grade agg ression was a significant predictor of seventh grade aggression in the final model as expected. None of the covariates were significant predictors of change in aggression from sixth to seventh grade. Social competence and resolution skills measures were no t significant predictors of change in aggression over time. Self esteem and anger at sixth grade, however, were significant predictors of change in aggression from sixth to seventh grade, such that high anger and low self esteem were significantly associat ed with change in aggression over time. As before, additional regressions were conducted to analyze the effects of interactions of communication competence, resolution skills, anger, and self esteem on changes in aggression. Again, none of the interactio ns was found to be significant in models assessing change in aggression from sixth to seventh grade. Regressions Predicting Change in Aggression from Seventh to Eighth Grade Regressions were then used to analyze the relationship between seventh grade soci al competence (communication and resolution skills), anger, and self esteem with the change in aggression from seventh to eighth grade (Table 3 7 ). Seventh grade seen in T able 3 1 the largest change in aggression occurred between sixth and
34 seventh grades. That said, seventh grade aggression was a significant predictor of eighth grade aggression, such that higher levels of aggression in seventh grade were significantly asso ciated with higher levels of aggression in eighth grade. None of the covariates were significant predictors of change in aggression from seventh to eighth grade. The self esteem and social competence measures, communication and resolution skills, were not significant predictors of change in aggression over time. Anger, however, was a significant predictor of change in aggression from seventh to eighth grades, such that high anger was significantly associated with change in aggression over time. As was done for analyses of change in aggression from sixth to seventh grades, regressions were also conducted to test interactions of resolution skills, communication competence, self esteem, or anger predicted change in aggression from seventh to eighth grades. No interactions were found to be significant predictors of change in aggression from seventh to eighth grade. Given the absence of effects of social competence variables in basic descriptive analyses and the regression models, these variables were re examined As indicated, the social competence indicators were based on the average performance across several vignettes and the overall levels of social skills indicated moderate competence in communication and resolution skills with relatively small size standar d deviations (Table 3 1 ). Hence, the overall means may not be as salient an indicator in contrast to identifying subgroups of youth who are consistently high or low in their social competence across vignettes. Thus, to examine the effects of social compet ence more fully, additional sensitivity analyses were conducted to explore subgroups of the competence indicators. However, findings from
35 these analyses did not differ substantially from findings for regressions using the continuous measures. Hence, the a nalyses of subgroups are not reported here. Additional information on these analyses can be found in the appendix.
36 Table 3 1. Means and standard deviations for primary variables Variable n M SD Aggression a 6 th grade Aggression 453 12.4 6 10.17 7 th grade Aggression 278 18.23 11.66 8 th grade Aggression 245 20.68 11.83 Anger a 6 th grade Anger 441 10.38 6.22 7 th grade Anger 257 12.61 5.9 8 th grade Anger 244 12.04 5.5 Self Esteem a 6 th grade Self Esteem 431 15.69 3.94 7 th grade Self Esteem 246 15.85 3.65 8 th grade Self Esteem 241 15.37 3.93 Social Competence Communication b 6 th grade Communication 449 3.08 0.44 7 th grade Communication 270 3.04 0.32 8 th grade Communication 249 3.09 0.28 Social Competence Resolution Sk ills c 6 th grade Resolution skills 446 0.82 0.8 7 th grade Resolution skills 269 0.97 0.87 8 th grade Resolution skills 245 1.4 0.95 a Lower score indicates less participation (0 = Never). b On a scale of 1 5 where 1 is not at all competent. c On a s cale of 0 3 where 0 is never high resolution skills.
37 Table 3 2 Correlations of primary variables Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. 6th grade Aggression 2. 7th grade Aggression .55 ** 3. 8th gra de Aggression .45 ** .67 ** 4. 6th grade Anger .39 ** .41 ** .38 ** 5. 7th grade Anger .28 ** .43 ** .33 ** .49 ** 6. 8th grade Anger .26 ** .29 ** .39 ** .45 ** .47 ** 7. 6th grade Self Esteem .21** .21 ** .23 ** .06 .09 .13 8. 7th grade Self Esteem .18 ** .25 ** .26 ** .13 .21 ** .17 .44 ** 9. 8th grade Self Esteem .12 .19 ** .16 .01 .13 .04 .34 ** .5 ** 10. 6th grade Communication .01 .04 .04 .03 .04 .06 .05 0 .07 11. 7th grade Communication .01 .09 .20 ** .04 .07 .01 .17 ** .13 .11 .28 ** 12. 8th grade Communication .02 .16 .14 .02 .02 .06 .12 .11 .11 .23 ** .28 ** 13. 6th grade Resolution Skills 0 .06 0 .016 .03 .03 .03 .04 .01 .24 ** .19 ** .07 14. 7th grade Resolution Skills .05 .06 .09 .02 .12 .07 .12 .07 .03 .15 .28 ** .14 .20 ** 15. 8th grade Resolution Skills .09 .14 .17 .07 .02 .06 .12 .04 .13 .29 ** .21 ** .25 ** .30 ** .36 ** p <.05. ** p <.01.
38 Table 3 3 Regression analysis summary for social competence indicators at sixth grade predicting concurrent aggression Step and Predictor Variable B SE B R2 R2 p Step 1: 0.06 0.06 .016* Gender 1.27 1.34 0.06 .344 Latino 1.09 2.03 0.05 .590 Black 3.96 1.98 0.20 .046* White 0.33 2.31 0.01 .886 School type 1.04 1.61 0.05 .519 Step 2: 0.29 0.23 .000** Overall Communication 0.96 1.42 0.04 .500 Overall High Resolution 0.02 0.80 0.00 .985 Anger 0. 70 0.10 0.44 .000** Self esteem 0.48 0.15 0.19 .002* p <.05. ** p <.01.
39 Table 3 4 Regression analysis summary for social competence indicators at seventh grade predicting concurrent aggression Step and Predictor Variable B SE B R2 R2 p Step 1: 0.005 0.005 0.971 Gender 0.186 1.743 0.008 0.915 Latino 1.373 2.681 0.054 0.609 Black 1.696 2.625 0.071 0.519 White 0.253 2.975 0.008 0.932 School type 0.004 2.178 0.000 0.999 Step 2: 0.189 0.184 .000** Overall Communication 2.994 2.683 0.083 0.266 Overall High Resolution 0.033 0.975 0.002 0.973 Anger 0.685 0.138 0.361 .000** Self esteem 0.471 0.232 0.151 0.044* p <.05. ** p <.01.
40 Table 3 5 Regression analysis s ummary for social competence indicators at eighth grade predicting concurrent aggression Step and Predictor Variable B SE B R2 R2 p Step 1: 0.019 0.019 0.493 Gender 0.876 1.576 0.037 0.579 Latino 0.018 2.500 0.001 0.994 Black 0.513 2.443 0.021 0.834 White 4.275 2.826 0.133 0.132 School type 1.679 1.846 0.067 0.364 Step 2: 0.241 0.221 .000** Overall Communication 5.723 2.617 0.136 0.03* Overall High Resolution 1.600 0.764 0.129 0.037* Anger 0.884 0.128 0.413 .000** Self esteem 0.286 0.180 0.096 0.113 p <.05. ** p <.01.
41 Table 3 6 Regression analysis s ummary for social competence indicators predicting change in aggression from sixth to seventh grades Step and Predictor Variable B SE B R2 R2 p Step 1: .31 .31 .000*** 6 th grade aggression .67 .07 .56 Step 2: .34 .02 .295 Gender .71 1.41 .03 .617 Latino 2.70 2.14 .11 .208 Black .51 2.11 .02 .808 White 1.05 2.46 .03 .671 School type 2.03 1.69 .08 .232 Step 3: .39 .05 .007** Overall Communication 2.02 1.65 .08 .223 Overall High Resolution 1.72 .96 .11 .074 Anger .34 .13 .18 .009* Self esteem .43 .19 .14 .022* p <.05. ** p <.01. *** p<.001
42 Table 3 7 Regression analysis summary for social competence indicators predicting change in aggression from seventh to eighth grades Step and Predictor Variable B SE B R2 R2 p Step 1: .51 .51 .000*** 7 th grade aggression .73 .05 .72 Step 2: .52 .01 .625 Gender 1.55 1.24 .07 .212 Latino 1.83 1.90 .07 .337 Black 1.42 1.86 .06 .447 White 3.03 2.11 .10 .152 School t ype .39 1.54 .01 .799 Step 3: .55 .03 .056 Overall Communication 3.78 2.06 .10 .068 Overall High Resolution .60 .74 .04 .424 Anger .18 .11 .09 .119 Self esteem .26 .18 .08 .154 p <.05. ** p <.01. *** p<.001
43 CHAPTER 4 DISCUSSION The purpose of the present study was to examine whether anger, self esteem, and social competence, as indicated by communication and conflict resolution skills, predict aggression both cross sectionally and over time, and whether anger and self esteem moderate the relationship between social competence and aggression. The present findings suggest that in nearly all analyses, anger and self esteem were each found to predict both concurrent aggression and change in aggression over ti me. However, according to the present findings, social competence does not predict aggression on its own, and does not interact with anger or self esteem to predict aggression. Self esteem and anger also did not interact with one another to predict aggress ion. Main effects in the expected directions were found for both anger and self esteem predicting aggression. In terms of anger, high anger was consistently associated with higher aggression. This is in line with the literature on anger and aggression, w hich has shown that trait anger, the type measured in the present study, is consistently associated with a variety of problem behaviors including aggression (e.g. Deffenbacher et al., 1996; Fives, Kong, Fuller, & DiGiuseppe, 2011; Mahon, 2000; Nichols et a l., 2006) Thus, the present findings help to strengthen the literature on trait anger. With regard to self esteem, the present findings show that low self esteem is associated with higher aggression in this sample. This direction of effects was expected i n the present study, although as described previously, there have been inconsistent findings with regard to self esteem and aggression in the literature. Inconsistencies in the self esteem and aggression literature may, at least partly, be the result of i nconsistent definitions of self esteem and, in turn, inconsistent methods of measuring self esteem.
44 For example, some researchers conceptualize self esteem as self evaluations that use a m ore global measure of self esteem (e.g. the Rosenberg Self Esteem Scale). Others, though, use the term self esteem to refer to self evaluations of domain specific behaviors or competences, and in turn focus their measures on these particular domains (Diama ntopoulou, Rydell, & Henricsson, 2008; Ostrowsky, 2010). Studies that use the global perspective tend to find relations between low self esteem and aggression, while those that use domain specific measures often find that high self esteem is related to agg ression (Diamantopoulou et al., 2008). In the present study, self esteem was conceptualized as a global self evaluation. Thus, in line with the findings on global self esteem, it was expected that adolescents with low self esteem would be higher in aggress ive behaviors than those with high self esteem, and overall, this is precisely what was found. This finding, that low self esteem was predictive of aggression in this sample, adds to the literature on adolescent self esteem and may help to clarify the role of self esteem in predicting adolescent aggression. The main findings on social competence were that social competence was not associated with aggression, anger, or self esteem and did not interact with the predictors. Thus, although social competence may serve as an important buffer for limiting adolescent aggression, the present study was unable to confirm this. To explain why the social competence findings were not as hypothesized, it is useful to look at the previous research on social competence an d the distinction between self reported and observed social competence measures. Social competence is commonly measured through the use of self report, and
45 report by parents, peers, or teachers, however there are many limitations to these methods. To sta rt with, these measures fail to take into account the importance of evaluating social competence within varying social contexts (Wright, Nichols, Graber, Brooks competency, response s in a variety of situations must be examined, as displaying social competence in one context does not necessarily mean one will display competence in others. Despite this, many studies of social competence rely on self report data. These instruments gener reporting of possible behavior may not provide the most accurate representation of social competence possible, and ma to engage in the desired behaviors. To avoid these limitations, social competence in the present study was assessed through direct observation, specifically by observing adolescen ts engaging in role plays of conflict situations. Standardized role play tasks were used to provide a more realistic measurement of social competence than self report or report by others can achieve. Observational methods to obtain social competence data, like the use of role play tasks, naturalistic observation of competence than other methods are able to, while also providing enough control to collect standard data (Foster, I nderbitzan, & Nangle, 1993). In addition, although they are not assessments of real world behavior, role play tasks responses on the spot (Foster et al., 1993; Wrig ht et al., 2004). Despite these benefits
46 and the findings that role play tasks are an effective way of studying social interaction, exceptions, see Borbely, Graber, Nichol s, Brooks Gunn, & Botvin, 2005; Wright et al., 2004). Thus, although the social competence findings in the present study were not as hypothesized, this study contributes to the social competence literature by demonstrating an uncommon approach to the measu rement of social competence in adolescence, through the use of observational role play measures. Limitations At the same time, this measurement approach had limitations. Most importantly, this study shows that more investigation is needed into the measurem ent of social competence and its role in predicting aggression. In particular, the present study found a notable lack of variability in the social competence indicators in this sample, with most participants displaying averag e social competence at all time points and no indication that competencies increased developmentally. Statistical methods for follow up, such as using a sensitivity analysis, cluster analysis, or latent class analysis, could be used to help clarify whether the social competence indicato rs used here can reasonably be analyzed by comparing high versus low social competence groups, and whether the observed measures of social competence used here map onto the same underlying competence factor. It is important to note that social competence c urrently does not have a standard set of measures, nor is there a standard definition in the literature; future research may consider other factors such as refusal skills and coping skills that may be important to social competence. In addition, alternate approaches to collecting observed social competence data should be considered, as the observed role play tasks used in the present study may
47 not have tapped behaviors most salient to aggression, or may not have differentiated among youth with communicatio n and conflict resolution skills that are most salient to avoiding aggressive situations in the middle school years. Still, the use of observed measures of social competence is a valuable approach to investigating social competence, as self report measures themselves and can therefore be difficult to disentangle from other concepts of self evaluation, such as self esteem. Alternate approaches to the measurement of social competence include a more naturalistic example, in a classroom setting, during a group project, or during school lunch. This approach may not be ideal, however, due to the lack of control in measurement. Another, more standardized, approach could be to develop a cooperative task, like a puzzle task, for participants to work on with a partner. This approach to observed social competence might elicit conflict more than the hypothetical role play scenario used in the present study was able to, and could the refore result in greater distribution of range in the social competence indicators. With regard to other limitations of the study aggression is self reported in the present study and hence, the validity of this measure may be questioned. However, to addr ess this concern, steps were taken during data collection to improve self report validity. Additionally, self report measures are believed to provide reliable reporting of aggressive behavior and illegal activities (Thornberry & Krohn, 2000), and typically result in higher rates of aggressive or delinquent behaviors than provided by arrest reports. Thus, self reported behaviors may allow a more encompassing measure of adolescent aggression than engagement in the legal system can provide. In addition,
48 the p resent study examined a limited range of aggressive behaviors. Specifically, this study focused on direct aggression, rather than relational or indirect aggression. Hence, it may be worthwhile to include measures of both direct and indirect aggression in f uture studies, as communication or conflict resolution skills may have associations with other forms of aggression that the present study would not be able to address. Implications Prior to this study, there has been very little longitudinal research focu sed on evaluations and the relationship of these factors with aggression. The knowledge to be gained from this study may help to strengthen preexisting prevention programs that develop and prom ote these factors in young, urban adolescents and may also help guide the development of future prevention programs targeting this demographic population. This is particularly important, as rates of direct aggression have been found to be higher in Black a nd Hispanic groups compared with youth from other ethnic groups, yet predictors of aggression in these samples remains understudied (Reingle et al., 2012). Thus, although results may not be generalizable to other adolescent populations, because urban minor ity samples are often underrepresented in research and rates of aggression are known to be higher in these groups, this study provides an important contribution to the literature. As noted, results from this study can be used to inform future prevention pr ograms targeting aggression. Many studies focus on youth who are at risk due to al., 2011 for review). Unfortunately, many of those factors cannot easily be changed throug h typical school based intervention programs focused on changing behavior at
49 the individual level. However interpersonal factors such as social competence, self esteem, and emotion regulation can be changed more easily (in comparison to structural changes in a community). Thus, determining which interpersonal factors are prevention programs and improving those that are currently in use. As previous research has clearly shown, the strongest predictor of future problem behaviors is an previous involvement in problem behaviors, looking at causes of low self esteem and poor anger regulation a nd potentially social competence in communication and conflict resolution, is very important in early adolescence, as this is a time when most adolescents have not begun to engage in many problem behaviors and levels of interpersonal violence are still rel atively low as compared to older adolescents.
50 APPENDIX SOCIAL COMPETENCE SUBGROUPS Truncation of range was a problem for both social competence variables. Specifically, little variability was found in either social competence variable, with enerally displaying moderate social competence at each time point, scoring an average of 3 on a scale of 1 5. As the present study aimed to investigate whether differences in social competence differentially predicted aggression, subgroups were created for both of the social competence measures to identify subgroups of youth who are consistently high or low in their social competence across vignettes. Thus, to examine the effects of social competence more fully, additional sensitivity analyses were conducte d to explore subgroups of the competence indicators. Unfortunately, the sample sizes of these subgroups were extremely small as a result of the truncation of range issues in the continuous variables, so the inclusion of these subgroup analyses in the thesi s seemed to suggest that there was greater variability in the continuous measures of social competence than was actually found in the present study (Figure A 1 shows a descriptive comparison of social competence groups). In light of these concerns, analysi s of the subgroups created from the communication and conflict resolution measures are excluded from the final analyses. More detailed discussion of the social competence subgroups and corresponding analyses is included here. Subgroup Analyses To examine the effects of social competence more fully, social competence in communication and resolution quality were examined for
51 consistency of competence. The social competence in communication measure was used to create two additional variables for each time poi nt: high competence and low competence (both dummy coded, split by one standard deviation above or below the mean, where 0 indicates never or rarely displayed high/low competence and 1 indicates consistently displayed high/low competence). The overall soci al competence in resolution quality measure was used similarly to create one additional dummy coded variable for each time point, where 0 indicates never or rarely displayed high quality resolution skills and 1 indicates consistently displayed high quality resolution skills in all three vignettes. Descriptive analyses of the social competence indicators showed the following: With regard to low competence in communication, 94.4% of participants never or rarely displayed low competence in sixth grade, 97.8% of participants never or rarely displayed low competence in seventh grade, 97.2% never or rarely displayed low competence in eighth grade. For high competence in communication, 14.5% consistently displayed high competence in sixth grade, 5.2% consistently displayed high competence in seventh grade, and 4.4% consistently displayed high competence in eighth grade. For high competence in resolution skills, 18.8% of participants consistently displayed high competence in sixth grade 26.8% of participants consist ently displayed high competence in seventh grade, and 48.2% of participants consistently displayed high competence in eighth grade. Descriptive comparisons between the competence groups on aggression showed no significant effects This can be seen in the
52 f ollowing figure (Figure A 1) where aggression appears to increase from sixth to eighth grade, but all groups display moderate aggression at each time point. The regression models tested for the continuous variables for communication and resolution were r e analyzed using either low communication and high resolution or high communication and high resolution in the models. Findings for main effects did not differ. However, u sing the high and low social competence measures as moderators in the models was fo und to differentially predict aggression, both in concurrent and change over time models. Sixth grade competence in communication appeared to show moderating effects predicting change in aggression from sixth to seventh grade. It must be noted, however, t hat these interactions were difficult to interpret due to the extremely small size of some of the groups; particularly the low communication competence high anger group, which may have driven the interaction. Similarly, seventh grade anger and concurrent s ocial competence predicted concurrent seventh grade aggression, such that high resolution skills appeared to buffer the effect of low communication competence and anger on aggression. Again, however, these interactions must be interpreted with caution, as low communication competence was unlikely to occur with high resolution skills, causing this group to drop out of the follow up analyses. Finally, in eighth grade, concurrent competence in communication moderated by concurrent resolution skills, predicted eighth grade aggression. Here too the interaction is difficult to interpret, as high communication competence rarely occurred without high conflict resolution skills, so the high communication competence not high
53 resolution skills group did not come out i n follow up analyses. Thus, none of the subgroup interactions are interpretable due to the small sample size of the groups. It is worth noting, however, that it does make conceptual sense that particular combinations of factors are unlikely to occur toge ther. For example, low communication competence was unlikely to be found in combination with high conflict resolution skills, and was also unlikely to be found in combination with high anger. With regard to the low communication competence and high confli ct resolution group, despite the fact that they seemed to be distinct indicators of overall social competence, as indicated by weak to moderate correlations between these variables at each time point, displaying high conflict resolution skills may require at least a moderate level of communication competence. In other words, if one is consistently unable to communicate effectively with others, it is likely that they will also have trouble resolving conflict through discussion. In the present study, communi cation and resolution were observed in the same vignette but were coded as independent constructs. Regarding the low communication and high anger group, given that both low communication competence and high anger were found to be less common individually, it is easy to see how this could result in few participants displaying both of these uncommon characteristics.
54 Figure A 1. Descriptive comparison of social competence subgroups on aggression
55 LIST OF REFERENCES Achenbach, T. M. & Edelbrock, C. S. (198 6). Burlington: University of Vermont. Baumeister, R. F., Bushman, B. J., & Campbell, W. K. (2000). Self esteem, narcissism, and aggression: Does violence result from low self esteem or from threatened egotism ? Current Directions in Psychological Science, 9 26 29. Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high self esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological Science in the Public Interest, 4, 1 44. Baumeister, R.F., Smart, L., & Boden, J.M. (1996). Relation of threatened egotism to violence and aggression: The dark side of high self esteem. Psychological Review 103 5 33. Bongers, I. L., Koot, H. M., van der Ende, J., & Verhu lst, F. C. (2003). The normative development of child and adolescent problem behavior. Journal of Abnormal Psychology, 112, 179 192. Borbely, C.J., Graber, J.A., Nichols, T., Brooks Gunn, J., & Botvin, G.J. (2005). Sixth n role plays with a peer, parent, and teacher. Journal of Youth and Adolescence, 34, 279 291. Botvin, G.J. & Griffin, K.W. (2004). Life skills training: Empirical findings and future directions. The Journal of Primary Prevention, 25 211 232. Botvin, G. J., Schinke, S. P., Epstein, J. E.,&Diaz, T. (1994). The effectiveness of culturally focused and generic skills training approaches to alcohol and drug abuse prevention among minority youth. Psychology of Addictive Behaviors, 8, 116 127. Buka, S.L. Stichick, T.L., Birdthistle, I., & Earls, F.J. (2001). Youth exposure to violence: Prevalence, risks, and consequences. American Journal of Orthopsychiatry, 71, 298 310. Bushman, B.J., & Baumeister, R.F. (1998). Threatened egotism, narcissism, self e steem, and direct and displaced aggression: Does self love or self hate lead to violence? Journal of Personality and Social Psychology 75 219 229. Crick, N. R., & Dodge, K. A. (1994). A review and reformation of social information processing mechanism Psychological Bulletin 115 74 101.
56 Caspi, A., & Moffitt, T. E. (1993). When do individual differences matter? A paradoxical theory of personality coherence. Psychological Inquiry 4 247 271. Crocker, J. & Luhtan en, R.K. (2003). Level of self esteem and contingencies of self worth: Unique effects on academic, social, and financial problems in college students. Personality and Social Psychology Bulletin 29 701 712. Deffenbacher, J. L., Oetting, E. R., Lynch, R. S., & Morris, C. D. (1996). The expression of anger and its consequences. Behaviour Research and Therapy 34 575 590. del Barrio, V., Aluja, A., & Spielberger, Ch. (2004). Anger assessment with the STAXI CA: Psychometric properties of a new instrumen t for children and adolescents. Personality and Individual Differences 37 227 244. Diamantopoulou, S., Rydell, A.M., & Henricsson, L. (2008). Can both low and high self esteem be related to aggression in children? Social Development 17 682 698. Dod Child Development 51 162 170. Dodge, K. A., Coie, J. D., & Lynam. D. (2006) Aggression and antisocial behavior in youth In Damon, W. (Series Ed.), Lerner, R. M. (Series Ed.), & Eisenberg, N. (Vol. Ed.). Handbook of child psychology: Vol. 3. Social, emotional, and personality development (6th ed.). New York: Wiley. Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., Mof tt, T. E., & Caspi, A. (2005). Low self esteem is related to aggression, antisocial behavior, and delinquency Psychological Science 16 328 335. Elliott D, Huizinga D, Menard S., (1989). Multiple Problem Youth: Delinquency, Substance Use, and Mental Health Problems New York: Springer. Evans, R. I. Hansen,W. B., & Mittlemark, M. B. (1977). Increasing the validity of self reports of smoking behavior in children. Journal of Applied Psychology, 62, 521 523. Fives, C.J., Kong, G., Fuller, J.R., & DiGiuseppe, R. (2011). Anger, aggression, and irratio nal beliefs in adolescents. Cognitive Theoretical Research 35 199 208. Foster, S. L., Inderbitzen, H. M., and Nangle, D. W. (1993). Assessing acceptance and social skills with peers in childhood: Current issues. Behavior Modification, 17 255 286.
57 Fraser, M. W., Galinsky, M. J., Smokowski, P. R., Day, S. H., Terzian, M. A., Rose, R. A., et al. (2005). Social information processing skills training to promote social competence and prevent aggressive behavior in the third grades. Journal of Consultin g and Clinical Psychology, 73 1045. Graber, J.A., Nichols, T., Lynne, S.D., Brooks Gunn, J., & Botvin G.J. (2006). A longitudinal examination of family, friend, and media influences on competent versus problem behaviors among urban minority youth. Appl ied Developmental Science 10 75 85. Harter, S. (1999). The construction of the self New York: Guilford. Joireman, J., Anderson, J., & Strathman, A. (2003). The aggression paradox: Understanding links among aggression, sensation seeking, and the consi deration of future consequences. Journal of Personality and Social Psychology 84 1287 1302. Keltikangas Jorvinen, L., & Kangas, P. (1988). Problem solving strategies in aggressive and nonaggressive children. Aggressive Behavior 14 255 264. Kim, S., Orpinas, P., Kamphaus, R., & Kelder, S.H. (2011). A multiple risk factors model of the development of aggression among early adolescents from urban disadvantaged neighborhoods. School Psychology Quarterly 26 215 230. Kotler, J.C. & McMahon, R.J. (2002) Differentiating anxious, aggressive, and socially competent preschool children: Validation of the social competence and behavior evaluation 30 (parent version). Behaviour Research and Therapy 40 947 959. LaRusso, M., & Selman, R. (2011). Early adoles cent health risk behaviors, conflict resolution strategies, and school climate. Journal of Applied Developmental Psychology, 32 354 362. Lemerise, E.A. & Arsenio, W.F. (2000). An integrated model of emotion proesses and cognition in social information processing. Child Development 71 107 118. Lochman, J. E., & Dodge, K. A. (1994). Social cognitive processes of severely violent, moderately aggressive, and nonaggressive boys. Journal of Consulting and Clinical Psychology 62 366 374. Lynne Landsman, S.D., Graber, J.A., Nichols, T.R., & Botvin, G.J. (2011). Trajectories of aggression, delinquency, and substance use across middle school among urban, minority adolescents. Aggressive Behavior 37 161 176. Mahon, N.E., Yarcheski, A., & Yarcheski, T.J. (2000). Positive and negative outcomes of anger in early adolescents. Research in Nursing and Health 23 17 24.
58 Masten, A. S., Coatsworth, D. J., Neemann, J., Gest, S. D., Tellegen, A., & Garmezy, N. (1995). The structure and coherence of competence fr om childhood through adolescence. Child Development, 66, 1635 1659. Masten, A. S., & Curtis, W. J. (2000). Integrating competence and psychopathology: Pathways toward a comprehensive science of adaptation in development. Development and Psychopathology, 12, 529 550. Moffitt, T. E. (1993). Adolescence limited and life course persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674 701. Monahan, K.C. & Steinberg, L. (2011). Accentuation of individual differences in social competence during the transition to adolescence. Journal of Research on Adolescence 21 576 585. Nichols, T. R., Graber, J. A., Brooks Gunn, J.,& Botvin, G. J. (2006). Sex differences in overt aggression and delinquency among urban minority middle scho ol students. Journal of Applied Developmental Psychology, 27, 78 91 Orobio de Castro, B., Merk, W., Koops, W., Veerman, J. W., & Bosch, J. D. (2005). Emotions in social information processing and their rela tions with reactive and proactive aggression in referred aggressive boys. Journal of Clinical Child and Adolescent Psychology, 34, 105 116. Ostrowsky, M.K. (2010). Are violent people more likely to have low self esteem or high self esteem? Aggression and Violent Behavior 15 69 75 Prinstein, M. J., & La Greca, A. M. (2004). Childhood peer rejection and aggression as year longitudinal study. Journal of Consulting and Clinical Psychology, 72, 103 112. Reing le, J.M., Maldonado Molina, M.M., Jennings, W.G., & Komro, K.A. (2012). Racial/Ethnic Differences in Trajectories of Aggression in a Longitudinal Sample of High Risk, Urban Youth. Journal of Adolescent Health, 51, 45 52. Rosenberg, M. (1965). Society an d the adolescent self image Princeton, NJ: Princeton University Press. Runions, K.C. & Keating, D.P. (2009). Anger and inhibitory control as moderators of children's hostile attributions and aggression Journal of Applied Developmental Psychology 31 370 378 Simmons, R. G., & Blyth, D. A. (1987). Moving into adolescence: The impact of pubertal change and school context. New York: Aldine.
59 Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends i n theory and research. (Vol. 1, pp. 24 49). New York: Academic Press. Spielberger, C. D. (1988). State Trait Anger Expression Inventory. Orlando, FL: Psychological Assessment Resources. Spielberger, C. D., Krasner, S., & Solomon, E. (1988). The experi ence, expression, and control of anger. In M. P. Janisse (Ed.), Health psychology: Individual differences and stress (pp. 89 108). New York: Springer Verlag. Sutherland, I., & Shepherd, J. P. (2002). A personality based model of adolescent violence. Bri tish Journal of Criminology, Delinquency and Deviant Social 42 433 441. Thornberry, T. P., & Krohn, M. D. (2000). The self report method for measuring delinquency and crime. Measurement and Analysis of Crime and Justice, 4, 33 83. Tourangeau, R., Smit h, 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 Webster Stratton, C., & Lindsay, D. W. (1999). Social competence and con duct problems in young children: Issues in assessment. Journal of Clinical Child Psychology 28 25 43. Wilkowski, B.M., & Robinson, M.D. (2008). The cognitive basis of trait anger and reactive aggression: An integrative analysis. Personality and Social Psychology Review, 12 3 21. Williams, J.R., Fredland, N., Han, H., Campbell, J.C., & Kub, J.E. (2009). Relational Aggression and Adverse Psychosocial and Physical Health Symptoms Among Urban Adolescents. Public Health Nursing, 26, 489 499. Wittman, M., Arce, E., & Santisteban, C. (2008). How impulsiveness, trait anger, and extracurricular activities might affect aggression in school children. Personality and Individual Differences 45 618 623. Wright, J.A., Nichols, T.R., Graber, J.A., Brooks Gunn divergent peer resistance skills and delinquency a year later. Journal of Adolescent Health, 35 380 391.
60 BIOGRAPHICAL SKETCH Allison Metz comp leted her undergraduate education at Clark University, majoring in psychology. She divided her time between psychology and sociology, graduating with a Bachelor of the Arts degree in the spring of 2008. After completing her B.A., Allison spent two years wo rking as a research a ssistant at National Health Promotion Associates, Inc. in White Plains, NY. In 2010, she left New York to pursue a Ph.D. in developmental psychology at the University of Florida.