sign in sign up
<<
Home , Shame

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by DSpace at VU

Child Development, November/December 2008, Volume 79, Number 6, Pages 1792 – 1801

Trumping by Blasts of Noise: , Self-Esteem, Shame, and Aggression in Young Adolescents

Sander Thomaes Brad J. Bushman VU University Amsterdam and Utrecht University University of Michigan and VU University Amsterdam

Hedy Stegge and Tjeert Olthof VU University Amsterdam

This experiment tested how self-views influence shame-induced aggression. One hundred and sixty-three young adolescents (M 5 12.2 years) completed measures of narcissism and self-esteem. They lost to an ostensible opponent on a competitive task. In the shame condition, they were told that their opponent was bad, and they saw their own name at the bottom of a ranking list. In the control condition, they were told nothing about their opponent and did not see a ranking list. Next, participants could blast their opponent with noise (aggression measure). As expected, narcissistic children were more aggressive than others, but only after they had been shamed. Low self-esteem did not lead to aggression. In fact, narcissism in combination with high self-esteem led to exceptionally high aggression.

Violent aggressive behavior in youth leaves scars on esteem would cause aggression. Even more problem- perpetrators, victims, and society at large. Among atic, a persuasive body of empirical evidence was the many factors that contribute to youth aggression, lacking. Although there have been a few exceptions the self-regard of perpetrators has been a theoreti- (Lochman & Dodge, 1994), most studies involving cally important but empirically controversial cause. elementary school-aged children found no concurrent For many years, the prevailing view has held that or prospective links between low self-esteem and aggressive youth have low self-esteem (e.g., Califor- aggression (Gresham, MacMillan, Bocian, Ward, & nia Task Force to Promote Self-Esteem and Personal Forness, 1998; Hymel, Rubin, Rowden, & LeMare, and Social Responsibility, 1990; Carr, 1999; Heide, 1990; Kupersmidt & Patterson, 1991). In fact, several 1997; Kaplan, 1975; Keith, 1984; Willock, 1987). studies found that aggressive children have inflated Applied and practical efforts have also focused on rather than deflated self-views (Brendgen, Vitaro, low self-esteem as a cause of violence. For example, Turgeon, Poulin, & Wanner, 2004; David & Kistner, following a series of incidents in which schoolchil- 2000; Zakriski & Coie, 1996) and that high self-esteem dren fired guns and killed their classmates at various is a risk factor for exacerbating aggression problems American schools, several organizations (including over time (Menon et al., 2007). the U.S. Department of Education) prepared lists of One possible explanation of these findings is that alleged warning signals to be used to identify youth low self-esteem may only cause aggression in adoles- who might be relatively likely to engage in such cents because they are more concerned than younger destructive violence, and nearly all the lists included children about maintaining desired self-images low self-esteem as a significant risk factor (e.g., Lord, (Harter, 2006; Rosenberg, 1986). Contrary to that re- 1999). asoning, however, the bulk of studies involving ado- Despite this apparent consensus, no compelling lescents also found no link between low self-esteem theoretical rationale existed to explain why low self- and aggression (East & Rook, 1992; Esposito, Kobak, & Little, 2005; Olweus, 1994; Prinstein, Boergers, & Vernberg, 2001; Salmivalli, Kaukiainen, Kaistaniemi, This research was supported by a Fulbright scholarship to S.T. We & Lagerspetz, 1999; but see Donnellan, Trzesniewski, are grateful to the children and staff of the Brighton School District Robins, Moffitt, & Caspi, 2005, for an exception). As it in Michigan for their kind cooperation. We also thank Marcel van Aken, Roy Baumeister, Keith Campbell, and Bram Orobio de Castro for their valuable comments on an earlier version of this article. Correspondence concerning this article should be addressed to Sander Thomaes, Department of Developmental Psychology, # 2008, Copyright the Author(s) Utrecht University, P.O. Box 80.140, 3584 CS, Utrecht, The Journal Compilation # 2008, Society for Research in Child Development, Inc. Netherlands. Electronic mail may be sent to [email protected]. All rights reserved. 0009-3920/2008/7906-0018 Trumping Shame by Blasts of Noise 1793 stands, there is little reason to adhere to the view that be reliably identified and distinguished from related low self-esteem causes aggressive behavior in youth. personality dimensions in older children and adoles- cents (C. T. Barry, Frick, & Killian, 2003; Frick, Bodin, & Barry, 2000). We recently developed the Childhood Narcissism and Aggression Narcissism Scale, a reliable and valid measure of In a comprehensive literature review, Baumeister, narcissistic traits in normal child and adolescent Smart, and Boden (1996) rejected the view that low populations (Thomaes, Stegge, Bushman, Olthof, & self-esteem causes aggression. They proposed in- Denissen, 2008). stead that violence most commonly occurs when inflated views of self and unstable beliefs in per- Shame-Induced Aggression sonal superiority are threatened. These conceptions of excessive self- are relevant to narcissism, Violence and aggression often occur when an a term that comes from the Greek myth about individual’s , reputation, or self-esteem has a handsome young man who falls in love with his been impugned or threatened. In late childhood ownreflectioninthewater.Initsextremeform, and adolescence, such situations are typically expe- narcissism is a personality disorder that involves rienced as shameful (Nishina & Juvonen, 2005; grandiose views of self, an inflated sense of entitle- Olthof,Ferguson,Bloemers,&Deij,2004;Reimer, ment, and exploitative attitudes toward others 1996). Shameful events often involve the public (American Psychiatric Association, 1994). Most cur- exposure of some failure or shortcoming (Olthof, rent psychological research focuses on ‘‘normal Schouten, Kuiper, Stegge, & Jennekens-Schinkel, narcissism,’’ operationalized as a trait on which 2000; Smith, Webster, Parrott, & Eyre, 2002). When people in the general population vary (Morf & shamed, people are painfully aware that others Rhodewalt, 2001; Raskin & Terry, 1988). Intuition might think they are flawed (Lewis, 1971; Tangney suggests that narcissism is equivalent to excessively & Dearing, 2002). Importantly, this awareness of high self-esteem, but research showed that narcis- others’ disapproval is easily internalized as a global sism and self-esteem are not strongly correlated condemnation of the self (e.g., ‘‘I am a bad and (Brown & Zeigler-Hill, 2004). Narcissists are self- worthless person’’). Over the course of late child- absorbed and arrogant, but they differ in of hood, such self-condemning negative self-appraisals self-assurance and sufficiency (reflecting higher become a more pronounced part of the shame levels of self-esteem) or self- and insufficiency experience (Ferguson, Stegge, & Damhuis, 1991). (reflecting lower levels of self-esteem; Cain, Pincus, & Gradually, shameful events come to constitute Ansell, 2007; Dickinson & Pincus, 2003; Rose, 2002; a more serious threat to self-esteem. Wink, 1991). How do people behave when they are shamed? The link between narcissism and aggression Shameful events may cause people to withdraw and has been firmly established in adults (e.g., Bushman hide from social contact (e.g., Lindsay-Hartz, De & Baumeister, 1998; Bushman, Bonacci, van Dijk, & Rivera, & Mascolo, 1995). Alternatively, shameful Baumeister, 2003; Donnellan et al., 2005; Stucke & events may cause people to lash out aggressively Sporer, 2002; Twenge & Campbell, 2003). Preliminary against others. Across the life span, shame-prone evidence suggests that narcissists with high self- individuals are predisposed to externalize , to esteem are particularly prone to behave aggressively feel , and to behave aggressively (Tangney, (Papps & O’Carroll, 1998; Wink, 1991). Unfortunately, Wagner, Hill-Barlow, Marschall, & Gramzow, 1996). few studies have examined the effects of narcissism Situationally induced shame produces similar reac- and self-esteem on aggression in youth. This lack of tions (Bennett, Sullivan, & Lewis, 2005; Thomaes, emphasis is surprising because childhood is the time Stegge, & Olthof, 2005). Shame-induced aggression when the foundation for lifelong aggressive or non- may serve an ego-protective function (Tangney & aggressive behavior styles is laid (e.g., Loeber & Hay, Dearing, 2002). By directing blame and anger on 1997). For both clinical and theoretical reasons, it is others, people can prevent their self-esteem from important to study the psychological mechanisms (further) damage. Aggression shifts away underlying aggressive inclinations at a time that they from painful awareness of a devalued self. Also, by unfold. One cause of the paucity of research on asserting the dominant aggressive stance, people can narcissistic aggression in youth has been the absence, reaffirm the self and ‘‘save face’’ in front of others. until recently, of a scale designed to measure early In summary, shame-induced aggression may ori- manifestations of narcissism. Recent developments ginate from the basic human motive to preserve indicate, however, that the construct of narcissism can self-esteem. 1794 Thomaes, Bushman, Stegge, and Olthof

Self-Views and Shame-Induced Aggression lescence is a time when children’s self-views are still relatively malleable before they become crystallized If the traditional view that low self-esteem causes in late adolescence and adulthood (Trzesniewski, aggression is true, one would predict that youth with Donnellan, & Robins, 2003). Thus, early adolescence low self-esteem should be particularly aggressive may be a critical age period to initiate aggression when shamed because shameful events make them interventions targeted at changing self-views. This feel even more inferior. This prediction, however, is experiment tested whether a link between self-views inconsistent with what we know about the motiva- and aggression has already been established by then. tions that surround self-esteem. Self-verification the- Aggressive behavior was examined in a situational ory holds that people generally try to maintain context of shame. We did not address how felt shame consistent self-appraisals and dislike changing their was involved in youth’s aggressive behavior. Several self-views (Swann & Read, 1981). From this perspec- authors have noted that although feelings of shame tive, youth with low self-esteem should be relatively may mediate shame-induced aggression, these feel- untouched by shameful events because their habitual ings are not necessarily consciously experienced (e.g., self-appraisals are already low. In contrast, youth Lewis, 1971; Robins, Tracy, & Shaver, 2001). For with inflated narcissistic self-views should be more conceptual clarity, we chose to focus on shame- vulnerable than others to shameful events because induced aggression on a situational level. they are highly motivated to protect their inflated self We used a shame manipulation based on the easy from being damaged. Indeed, vulnerability to shame task failure paradigm (e.g., Lewis, Alessandri, & has been described as a key component of narcissism Sullivan, 1992). People who fail an easy task are (Morrison, 1989; Tracy & Robins, 2004). These notions especially likely to experience shame. Participants in are consistent with the empirical findings in adult our experiment failed a competitive reaction time samples showing that narcissists are aggressive task. They were randomly assigned to shame or no- following ego threat. shame control conditions. Participants in the shame condition were told that their opponent was one of the slowest contestants tested so far, and they saw their Present Experiment own name at the bottom of a ranking list posted on the A between-subjects experimental design was used bogus FastKid! Web page (below their opponent’s to examine how self-views influence aggressive re- name). The Internet rankings highlighted public sponses to induced shame. Participants were 10 – 13 exposure, which should enhance feelings of shame years old. From several developmental perspectives, (Smith et al., 2002). Participants in the control condi- this is an ideal age period for the purposes of this tion were told nothing about their opponent’s abilities experiment. First, in early adolescence, shame is both and did not see the bogus Web page. A validation a frequent and an aversive emotional experience due study showed this shame manipulation to be highly to developmental increases in self-consciousness effective (Thomaes et al., 2005). The shame condition (Ryan & Kuczkowski, 1994; Simmons, Rosenberg, & elicits feelings of shame in young adolescents, Rosenberg, 1973), the ability to view the self from whereas the control condition does not. Next, partic- others’ perspective (Harter, 2006), and the ability ipants were given a chance to blast their opponent to make global negative evaluations of the self with loud noise through headphones (the aggression (Ferguson et al., 1991). Second, early adolescence is measure). We predicted that narcissistic youth would a time when children become increasingly concerned behave most aggressively, but only in the shame about maintaining worth and approval, and self- condition. We did not predict high levels of aggres- protective motives come to exert a stronger influence sion in youth with low self-esteem. On the contrary, on their behavior (Harter, 2006; Rosenberg, 1986). on the basis of prior research, we predicted that high Third, early adolescence is an ideal time to measure self-esteem would enhance narcissistic aggression childhood narcissism. Young children typically hold (Papps & O’Carroll, 1998; Wink, 1991). unrealistically positive self-views because they lack the abilities to differentiate their actual self from their ideal self and to base their self-views on social com- parisons (Harter, 2006). By age 10, most youth have Method overgrown age-normative overestimation of compe- Participants tence that likely is a prerequisite for the meaningful assessment of individual differences in childhood Participants were 163 young adolescents (54% boys) narcissism (Thomaes et al., 2008). Fourth, early ado- from two public middle schools serving middle- and Trumping Shame by Blasts of Noise 1795 upper-middle-class neighborhoods in southeastern bonus was the ability to send a written message to the Michigan. In the school district, 8% of students are opponent. The second-round bonus was the ability to eligible for reduced-price lunch programs. Partici- blast the opponent with loud white noise (sounds like pants ranged in age from 10 to 13 years (M 5 12.2, SD radio static) through headphones after winning a trial. 5 0.6). Almost all were Caucasians (96%). To partic- Through a rigged lottery, the opponent had the bonus ipate, adolescents received informed parental consent in the first round, whereas the participant had the (29% of parents consented) and gave their own assent bonus in the second round. Participants were given (98% of adolescents assented). Participants received samples of noise they could set for their opponent. a small (e.g., mechanical pens, markers) in The noise levels ranged from 55 (Level 1) to 100 dB exchange for their voluntary participation. (Level 10) in 5-dB increments. The maximum noise level, 100 dB, is about the same intensity as a smoke alarm. A nonaggressive no-noise setting (Level 0) was Self-View Questionnaire also included. A few weeks prior to the experiment (M 5 3 weeks, Participants were randomly assigned to shame or range 5 1 – 5 weeks), participants completed self- no-shame control conditions. In the shame condition, report measures of narcissism and self-esteem at their participants were told that they were lucky to com- school. Narcissism was measured using the reliable pete against one of the worst players thus far. The and valid Childhood Narcissism Scale (Thomaes experimenter then logged onto the fictitious FastKid! et al., 2008). This 10-item scale assesses grandiose Web site and showed participants their opponent’s views of self, inflated feelings of superiority and name at the bottom of the ranking list. The experi- , and exploitative interpersonal attitudes. menter said, ‘‘This means you should win easily!’’ Sample items include ‘‘Without me, our class would Participants were told that immediately after the first be much less fun,’’ ‘‘Kids like me deserve something round of the game, new rankings would appear on the extra,’’ and ‘‘I often succeed in getting .’’ very popular FastKid! Web site and that their own Items are rated along a 4-point scale ranging from name would be included in those rankings. After 0(not at all true)to3(completely true). Responses were competing with the opponent on the first five reaction summed, with higher scores indicating higher levels time trials, a message appeared on screen that said, of narcissism. In the present experiment, the alpha ‘‘Sorry (participant’s name), you lost!’’ The opponent coefficient for the scale was .76. then sent the participant a message that said, ‘‘Can’t Self-esteem was measured using the 6-item global wait to see the rankings!’’ Then, the new rankings self-worth subscale of the Self-Perception Profile for showed the participant’s name at the bottom of the Children (Harter, 1985). This scale assesses the extent list, beneath the opponent’s name. The control condi- to which participants are satisfied with themselves tion was similar to the shame condition, with three and the way they are leading their lives. Sample items exceptions. First, participants received no informa- include ‘‘Some kids like the kind of person they are’’ tion about how good their opponent was (and saw no and ‘‘Some kids are not very happy with the way they rankings on the Web site before the game). Second, do a lot of things.’’ Following others (e.g., Brendgen participants saw no rankings after the game. Third, et al., 2004), we used a 4-point scale response format the opponent’s message said, ‘‘Huh?! Is the first ranging from 0 (I am not like these kids at all)to3(Iam round finished already?’’ We used a losing control exactly like these kids). After reverse scoring negatively condition because we wanted to test the effects worded items, responses were summed, with higher of shame above and beyond the effects of mere scores indicating higher levels of self-esteem. The or from losing a game. alpha coefficient for the scale was .72. In the second round of the game, participants had the ‘‘noise bonus,’’ so they could blast their opponent with loud noise after winning a trial. Prior to each of Procedure the five trials of Round 2, participants set the noise Participants were tested individually in a quiet level their opponent would receive if the opponent room at their school. They were told that they would lost. After each trial, participants were informed be competing on an Internet reaction time game called whether they had won (i.e., Trials 1, 2, 4, and 5) or FastKid! against an opponent of the same sex and age lost (i.e., Trial 3) that trial. To obtain an aggression from a school in Columbus, OH. In reality, there was measure unconfounded by the (nonmanipulation) no opponent and the computer controlled all events. effect of losing Trial 3, the average level of noise set Participants were told that FastKid! consisted of two for the opponent across the first three trials was used 5-trial rounds, each with a bonus. The first-round to measure aggression. The alpha coefficient for the 1796 Thomaes, Bushman, Stegge, and Olthof aggression measure was .85. Finally, participants was 3.26, indicating that multicollinearity was not were thoroughly debriefed to remove lingering ef- a problem. fects of the manipulations and were given a small gift. The analysis revealed a main effect for narcissism, t(159) 5 2.01, p , .05, b 5 0.41, b 5 .16. This main effect, however, was qualified by a significant inter- action between narcissism and condition, t(156) 5 Results 2.03, p , .05, b 5 0.82, b 5 .21 (see Figure 1). As ex- Preliminary Analyses pected, narcissism was positively related to aggression when participants were shamed, t(81) 5 2.91, p , .01, Sex differences and age. Boys were significantly more b 5 0.89, b 5 .31. Narcissism was not related to ag- aggressive than girls, F(1, 161) 5 8.37, p , .01, d 5 0.45. gression when participants were not shamed, t(78) 5 Aggression was not affected by age, t(161) 5 À0.33, p . 0.04, p . .96, b 5 0.01, b 5 .01. .74, b 5 À0.07, b 5 À.03. Because there were also no Although self-esteem did not directly influence interactions involving sex or age, the data from boys aggression levels, on the basis of prior research, we and girls of different ages were combined for sub- anticipated that narcissism in combination with high sequent analyses. self-esteem would lead to exceptionally high levels Equivalence of experimental conditions. Narcissism of aggression in the shame condition (Papps & and self-esteem scores did not differ in the shame O’Carroll, 1998; Wink, 1991). As expected, there was and no-shame groups (ps . .30). Thus, random as- a significant Narcissism  Self-Esteem  Condition signment to conditions was effective. Narcissism and interaction, t(155) 5 2.00, p , .05, b 5 0.84, b 5 .19. To self-esteem were only weakly correlated (r 5 .09). interpret the three-way interaction, we examined the two-way interactions between narcissism and Primary Analyses self-esteem separately for the shame and no-shame control conditions. As expected, narcissism and self- Table 1 contains the descriptive statistics for the esteem interacted to influence aggression in the experiment. Data were analyzed using hierarchical shame condition, t(79) 5 2.65, p , .01, b 5 0.91, b 5 .28 multiple regression analysis. The dependent variable (see Figure 2; high values of narcissism and self- was aggressive behavior, defined as the average esteem were 1 SD above the mean and low values intensity of noise participants gave their ostensible were 1 SD below the mean; Aiken & West, 1991). opponent. The main effects for condition, narcissism, Figure 2 shows that narcissism and aggression were and self-esteem were entered in Step 1; the two-way interactions involving these variables were entered in Step 2; and the three-way interaction was entered in Step 3. Narcissism and self-esteem scores were cen- 10 tered to reduce multicollinearity (Aiken & West, 1991; Jaccard & Turrisi, 2003). A maximum variance infla- tion factor (VIF) greater than 10 indicates that multi- 8 collinearity may be unduly influencing the least squares estimates (e.g., Neter, Wasserman, & Kutner,

1990). The maximum VIF in the regression analysis 6

Aggression 4 Table 1 Means and Standard Deviations in the Shame and No-Shame Control Conditions 2 Shamed Shame (N 5 83) Control (N 5 80) Not Shamed Range M (SD) M (SD) 0 Narcissism 2 – 26 12.04 (4.32) 11.29 (4.92) 0 5 10 15 20 25 Self-esteem 3 – 18 13.60 (2.89) 13.51 (3.38) Narcissism Aggression 1 – 10 7.06 (2.51) 7.02 (2.66) Age (months) 131 – 166 146 (8) 146 (8) Figure 1. Relationship between narcissism and aggression for participants in the shame and the no-shame control conditions. Trumping Shame by Blasts of Noise 1797

10 10

8 8

6 6 Aggression 4 Aggression 4

2 Low Self-Esteem 2 Low Self-Esteem High Self-Esteem High Self-Esteem

0 0 Low High Low High Narcissism Narcissism

Figure 2. Relationship between narcissism and aggression for par- Figure 3. Relationship between narcissism and aggression for par- ticipants with high and low self-esteem in the shame condition. ticipants with high and low self-esteem in the no-shame control Note. High values of narcissism and self-esteem are 1 SD above the condition. mean; low values of narcissism and self-esteem are 1 SD below the Note. High values of narcissism and self-esteem are 1 SD above the mean. mean; low values of narcissism and self-esteem are 1 SD below the mean. strongly associated in shamed youth with high self- tic shame-induced aggression can likely be viewed as esteem, t(79) 5 4.00, p , .001, b 5 1.69, b 5 .59. In con- defensive effort to maintain self-worth. trast, narcissism and aggression were not associated in No support was found for the traditional view that shamedyouthwithlowself-esteem,t(79) 5 À0.29, p . low self-esteem underlies aggression. In fact, that .77, b 5 À0.14, b 5 À.05. As expected, narcissism and view was contradicted by the finding that high self- self-esteem did not interact to influence aggression in esteem (not low self-esteem) increased narcissistic the no-shame control condition, t(76) 5 0.30, p . .76, shame-induced aggression. One explanation for this b 5 0.07, b 5 .04. As shown in Figure 3, narcissism and finding, based on self-verification theory, is that nar- aggression were not associated in nonshamed youth cissistic youth with high self-esteem are more vulner- with high or low self-esteem, t(76) 5 0.24, p . .81, b 5 able to shameful events than are youth with low 0.09, b 5 .04 and t(76) 5 À0.17, p . .86, b 5 À0.06, b 5 self-esteem. Another explanation is that narcissistic À.03, respectively. youth with high and low self-esteem do not differ in their vulnerability to shameful events, but they do differ in the way they deal with those events. This latter explanation is consistent with research showing Discussion that narcissists’ interpersonal orientation tends to The present experiment examined how narcissism vary along with their level of self-esteem (e.g., Cain and self-esteem influence young adolescents’ shame- et al., 2007; Wink, 1991). Overt narcissists, who have induced aggressive behavior. As predicted, narcissis- high self-esteem, have been described as extraverts tic youth were more aggressive than others, but only marked by a dominant and aggressive interpersonal after they had been shamed. Narcissists seem highly orientation. Covert narcissists, who have much lower motivated to create and maintain a grandiose view of self-esteem (i.e., their self-absorption co-occurs with self. They tend to interpret social situations in terms of feelings of self-doubt and insufficiency), have been how they reflect on the self, and they engage in self- described as ‘‘worriers’’ marked by an anxious and regulatory strategies to protect self-esteem when they internalizing interpersonal orientation. They may feel need to (Morf & Rhodewalt, 2001). As shameful aggrieved and hurt when shamed, but in our exper- situations constitute a threat to , narcissis- iment, they were not aggressive. 1798 Thomaes, Bushman, Stegge, and Olthof

Researchers have argued that aggression is an letter from the principal endorsing our study was appealing behavioral alternative to shamed individ- accidentally not included with the consent form. Low uals because it serves an ego-protective function parental consent does not threaten the internal val- (Tangney & Dearing, 2002). Aggression provides idity of our experiment because participants were immediate relief from the of shame, which is randomly assigned to shame or no-shame groups. We a tempting benefit in the short run. In the long run, cannot, however, exclude the possibility that self- however, predispositions to behave aggressively selected biased sampling has influenced the magni- when shamed may have serious costs. Children who tude of the experimental effects observed. Future persistently deflect the painful feelings associated research may assess the extent to which our experi- with their flaws and shortcomings may become less mental findings generalize to other community and motivated to overcome them. Consequently, they may noncommunity samples. become less well adapted to the demands of their Experimental aggression research is sometimes social environment. Also, affectively aggressive chil- criticized for using laboratory paradigms that are dren are unpopular with peers (e.g., Price & Dodge, supposedly unrepresentative of ‘‘real-world’’ aggres- 1989; Prinstein & Cillessen, 2003). Thus, aggressive sion. The validity of laboratory aggression paradigms behaviors meant to discard shame in the short run (including the noise blast procedure we used) has may ironically increase children’s liability to be the been supported in two meta-analyses. One meta- target of victimization in the long run. analysis demonstrated high levels of convergence The present experiment contributes to the litera- across a wide range of laboratory aggression meas- ture in several ways. To our knowledge, it is the first ures (Carlson, Marcus-Newhall, & Miller, 1989). The study to examine the link between youth’s self-views other meta-analysis showed that ‘‘real’’ and labora- and their actual aggressive behavior in an in vivo tory measures of aggression are influenced in similar situational context. Other studies have relied on ways by situational variables (e.g., provocation) and measures of reported aggression that often include by individual difference variables (e.g., trait aggres- more diffusely defined antisocial acts such as lying or siveness; Anderson & Bushman, 1997). stealing (e.g., Donnellan et al., 2005). Our findings are In contrast to the adult literature, aggression in the consistent with those of previous work that found that context of self-esteem threats such as shame is rarely narcissism is linked to reported aggression and con- examined in the child literature. We believe this is duct problems in youth (C. T. Barry et al., 2003; T. D. unfortunate because such threats are common expe- Barry et al., 2007). Also, the present experiment high- riences (particularly in early adolescence) known to lights the importance of shame as an emotional elicit aggression in subsets of children (Dodge, Coie, & context for examining the link between self-views Lynam, 2006). In this experiment, we manipulated and aggression. Most important, this experiment a situation of self-attributable shortcoming. However, shows the value of differentiating among different shame and other self-esteem threats can also result forms of self-view (Salmivalli, 2001). Most previous from shortcomings pinpointed by others. In fact, peer research involving children has relied exclusively on harassment among schoolchildren typically involves measures of self-esteem, which by itself is an unreli- damaging others’ self-esteem or status (Galen & able predictor of aggression. This study indicates that Underwood, 1997; Nishina & Juvonen, 2005). Contin- narcissism is an important predictor of aggression, ued research on shame-induced aggression is needed especially in the context of shame. to obtain a more complete view on the emotional processes involved in children’s and adolescents’ aggression. Limitations and Future Research We have argued that early adolescence is an ideal Conclusion age period for the present experiment. However, our developmental focus limits the ability to make gen- Many efforts to reduce violence and aggression eralizations to other age periods. Future research in youth have relied on boosting self-esteem (e.g., should test whether the link between self-views and Kusche´ & Greenberg, 1994; Ringwalt, Graham, aggression is moderated by age. Such knowledge Paschall, Flewelling, & Browne, 1996). Although would help clinicians to identify critical age periods intuitively plausible, there are no clear theoretical or to target the self-views underlying children’s aggressive empirical reasons why boosting self-esteem should be behavior. effective in reducing aggression. In fact, the present The parental consent rate for our experiment was experiment suggests that practices aimed at boosting low (i.e., 29%). This may have occurred because the self-esteem may even bring substantial costs. If these Trumping Shame by Blasts of Noise 1799 practices cultivate the inflated views of self that are Carlson, M., Marcus-Newhall, A., & Miller, N. (1989). characteristic of narcissism, they are likely to increase Evidence for a general construct of aggression. Person- (rather than to decrease) the aggressive behavior of ality and Social Psychology Bulletin, 15, 377 – 389. youth at risk. Carr, A. (1999). The handbook of child and adolescent clinical psychology: A contextual approach. London: Routledge. David, C. F., & Kistner, J. A. (2000). Do positive self- perceptions have a ‘‘dark-side’’? Examination of the link References between perceptual bias and aggression. Journal of Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing Abnormal Child Psychology, 28, 327 – 337. and interpreting interactions. Newbury Park, CA: Sage. Dickinson, K. A., & Pincus, A. L. (2003). Interpersonal American Psychiatric Association. (1994). Diagnostic and analysis of grandiose and . Journal statistical manual of mental disorders (4th ed.). Washington, of Personality Disorders, 17, 188 – 207. DC: Author. Dodge, K. A., Coie, J. D., & Lynam, D. (2006). Aggres- Anderson, C. A., & Bushman, B. J. (1997). External validity sion and antisocial behavior in youth. In W. Damon & of ‘‘trivial’’ experiments: The case of laboratory aggres- R.Lerner(SeriesEds.)&N.Eisenberg(Vol.Ed.), sion. Review of General Psychology, 1, 19 – 41. Handbook of child psychology: Vol. 3. Social, emotional, Barry, C. T., Frick, P. J., & Killian, A. L. (2003). The relation and personality development (pp. 719 – 788). New York: of narcissism and self-esteem to conduct problems in Wiley. children: A preliminary investigation. Journal of Clinical Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., Child and Adolescent Psychology, 32, 139 – 152. Moffitt, T. E., & Caspi, A. (2005). Low self-esteem is Barry, T. D., Thompson, A., Barry, C. T., Lochman, J. E., related to aggression, antisocial behavior, and delin- Adler, K., & Hill, K. (2007). The importance of narcissism quency. Psychological Science, 16, 328 – 335. in predicting proactive and reactive aggression in mod- East, P. L., & Rook, K. S. (1992). Compensatory patterns of erately to highly aggressive children. Aggressive Behavior, support among children’s peer relationships: A test 33, 185 – 197. using school friends, nonschool friends, and siblings. Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation Developmental Psychology, 28, 163 – 172. of threatened to violence and aggression: The Esposito, A. J., Kobak, R., & Little, M. (2005). Aggression dark side of high self-esteem. Psychological Review, 103, and self-esteem: A diary study of children’s reactivity to 5 – 33. negative interpersonal events. Journal of Personality, 73, Bennett, D. S., Sullivan, M. W., & Lewis, M. (2005). Young 887 – 906. children’s adjustment as a function of maltreatment, Ferguson, T. J., Stegge, H., & Damhuis, I. (1991). Children’s shame, and anger. Child Maltreatment, 10, 311 – 323. understanding of and shame. Child Development, Brendgen, M., Vitaro, F., Turgeon, L., Poulin, F., & Wanner, 62, 827 – 839. B. (2004). Is there a dark side of positive illusions? Frick, P. J., Bodin, S. D., & Barry, C. T. (2000). Psychopathic Overestimation of social competence and subsequent traits and conduct problems in community and clinic- adjustment in aggressive and nonaggressive children. referred samples of children: Further development of the Journal of Abnormal Child Psychology, 32, 305 – 320. screening device. Psychological Assessment, Brown, R. P., & Zeigler-Hill, V. (2004). Narcissism and the 12, 382 – 393. non-equivalence of self-esteem measures: A matter of Galen, B. R., & Underwood, M. K. (1997). A developmental dominance? Journal of Research in Personality, 38, 585 – investigation of social aggression among children. Devel- 592. opmental Psychology, 33, 589 – 600. Bushman, B. J., & Baumeister, R. F. (1998). Threatened Gresham, F. M., MacMillan, D. L., Bocian, K. M., Ward, egotism, narcissism, self-esteem, and direct and dis- S. L., & Forness, S. R. (1998). Comorbidity of hyperac- placed aggression: Does self-love or self-hate lead to tivity-impulsivity-inattention and conduct problems: violence? Journal of Personality and Social Psychology, 75, Risk factors in social, affective, and academic domains. 219 – 229. Journal of Abnormal Child Psychology, 26, 293 – 406. Bushman, B. J., Bonacci, A. M., van Dijk, M., & Baumeister, Harter, S. (1985). The self-perception profile for children: R. F. (2003). Narcissism, sexual refusal, and aggression: Revision of the perceived competence scale for children. Testing a narcissistic reactance model of sexual . Denver, CO: University of Denver. Journal of Personality and Social Psychology, 84, 1027 – 1040. Harter, S. (2006). The self. In W. Damon & R. Lerner (Series Cain, N. M., Pincus, A. L., & Ansell, E. B. (2007). Eds.) & N. Eisenberg (Vol. Ed.), Handbook of child Narcissism at the crossroads: Phenotypic description of psychology: Vol. 3. Social, emotional, and personality devel- pathological narcissism across clinical theory, social/ opment (pp. 505 – 570). New York: Wiley. personality psychology, and psychiatric diagnosis. Clin- Heide, K. M. (1997). Juvenile homicide in America: How ical Psychology Review, 28, 638 – 656. can we stop the killing? Behavioral Sciences and the Law, California Task Force to Promote Self-Esteem and Personal 15, 203 – 220. and Social Responsibility. (1990). Toward a state of esteem. Hymel, S., Rubin, K. H., Rowden, L., & LeMare, L. (1990). Sacramento: California State Department of Education. Children’s peer relationships: Longitudinal prediction of 1800 Thomaes, Bushman, Stegge, and Olthof

internalizing and externalizing problems from middle to Olweus, D. (1994). at school: Long-term outcomes late childhood. Child Development, 61, 2004 – 2021. for the victims and an effective school-based interven- Jaccard, J., & Turrisi, R. (2003). Interaction effects in multiple tion program. In R. Huesmann (Ed.), Aggressive behavior: regression (2nd ed.). Newbury Park, CA: Sage. Current perspectives (pp. 97 – 130). New York: Plenum. Kaplan, H. B. (1975). Self-attitudes and deviant behavior. Papps, B., & O’Carroll, R. E. (1998). Extremes of self- Pacific Palisades, CA: Goodyear. esteem and narcissism and the experience and expres- Keith, C. R. (1984). The aggressive adolescent: Clinical per- sion of anger and aggression. Aggressive Behavior, 24, spectives. New York: Free Press. 421 – 438. Kupersmidt, J. B., & Patterson, C. J. (1991). Childhood peer Price, J. M., & Dodge, K. A. (1989). Reactive and proactive rejection, aggression, withdrawal, and perceived com- aggression in childhood: Relations to peer status and petence as predictors of self-reported behavior problems social-context dimensions. Journal of Abnormal Child in preadolescence. Journal of Abnormal Child Psychology, Psychology, 17, 455 – 471. 19, 427 – 449. Prinstein, M. J., Boergers, J., & Vernberg, E. M. (2001). Kusche´, C., & Greenberg, M. (1994). PATHS: Promoting Overt and in adolescents: Social- alternative thinking strategies. South Deerfield, MA: psychological adjustment of aggressors and victims. Developmental Research Programs. Journal of Clinical Child Psychology, 30, 479 – 491. Lewis, H. B. (1971). Shame and guilt in . New York: Prinstein, M. J., & Cillessen, A. H. N. (2003). Forms and International Universities Press. functions of adolescent peer aggression associated with Lewis, M., Alessandri, S. M., & Sullivan, M. W. (1992). high levels of peer status. Merrill-Palmer Quarterly, 49, Differences in shame and pride as a function of children’s 310 – 342. gender and task-difficulty. Child Development, 63, 630 – 638. Raskin, R., & Terry, H. (1988). A principal-components Lindsay-Hartz, J., De Rivera, J., & Mascolo, M. (1995). analysis of the narcissistic personality inventory and Differentiating shame and guilt and their effects on further evidence of its construct validation. Journal of motivation. In J. Tangney & K. Fischer (Eds.), Self- Personality and Social Psychology, 54, 890 – 902. conscious : Shame, guilt, , and pride Reimer, M. S. (1996). ‘‘Sinking into the ground’’: The (pp. 274 – 300). New York: Guilford. development and consequences of shame in adoles- Lochman, J. E., & Dodge, K. A. (1994). Social cognitive cence. Developmental Review, 16, 321 – 363. processes of severely violent, moderately aggressive, Ringwalt, C. L., Graham, L. A., Paschall, M. J., Flewelling, and nonaggressive boys. Journal of Consulting and Clinical R. L., & Browne, D. C. (1996). Supporting adolescents Psychology, 62, 366 – 374. with guidance and employment (SAGE). American Jour- Loeber, R., & Hay, D. (1997). Key issues in the develop- nal of Preventive Medicine, 12, 31 – 38. ment of aggression and violence from childhood to early Robins, R. W., Tracy, J. L., & Shaver, P. R. (2001). Shamed adulthood. Annual Review of Psychology, 48, 371 – 410. into self-love: Dynamics, roots, and functions of narcis- Lord, M. (1999, October 11). The violent kid profile: A sism. Psychological Inquiry, 12, 230 – 236. controversial new technique for beating violence. U.S. Rose, P. (2002). The happy and unhappy faces of narcis- News & World Report, 127, 56 – 57. sism. Personality and Individual Differences, 33, 379 – 391. Menon, M., Tobin, D. D., Corby, B. C., Menon, M., Hodges, Rosenberg, M. (1986). Self-concept from middle childhood E. V. E., & Perry, D. G. (2007). The developmental costs through adolescence. In J. Suls & A. Greenwald (Eds.), of high self-esteem. Child Development, 78, 1627 – 1639. Psychological perspective on the self (Vol. 3, pp. 107 – 135). Morf, C. C., & Rhodewalt, F. (2001). Unraveling the para- Hillsdale, NJ: Erlbaum. doxes of narcissism: A dynamic self-regulatory process- Ryan, R. M., & Kuczkowski, R. (1994). The imaginary ing model. Psychological Inquiry, 12, 177 – 196. audience, self-consciousness, and public individuation Morrison, A. P. (1989). Shame: The underside of narcissism. in adolescence. Journal of Personality, 62, 219 – 238. Hillsdale, NJ: Analytic. Salmivalli, C. (2001). good about oneself, being Neter, J., Wasserman, W., & Kutner, M. H. (1990). Applied bad to others? Remarks on self-esteem, , and linear statistical models (3rd ed.). Homewood, IL: Irwin. aggressive behavior. Aggression and Violent Behavior, 6, Nishina, A., & Juvonen, J. (2005). Daily reports of witness- 375 – 393. ing and experiencing peer harassment in middle school. Salmivalli, C., Kaukiainen, A., Kaistaniemi, L., & Lager- Child Development, 76, 435 – 450. spetz, K. (1999). Self-evaluated self-esteem, peer-eval- Olthof, T., Ferguson, T. J., Bloemers, E., & Deij, M. (2004). uated self-esteem, and defensive egotism as predictors Morality- and identity-related antecedents of children’s of adolescents’ participation in bullying situations. guilt and shame attributions in events involving phys- Personality and Social Psychology Bulletin, 25,1268– ical illness. and , 18, 383 – 404. 1278. Olthof, T., Schouten, A., Kuiper, H., Stegge, H., & Jennekens- Simmons, R. G., Rosenberg, F., & Rosenberg, M. (1973). Schinkel, A. (2000). Shame and guilt in children: Dif- Disturbance in the self-image at adolescence. American ferential situational antecedents and experiential Sociological Review, 38, 553 – 568. correlates. British Journal of Developmental Psychology, 18, Smith, R. H., Webster, J. M., Parrott, W. G., & Eyre, H. L. 51 – 64. (2002). The role of public exposure in moral and Trumping Shame by Blasts of Noise 1801

nonmoral shame and guilt. Journal of Personality and presented at the 12th European conference on Develop- Social Psychology, 83, 138 – 159. mental Psychology, La Laguna, Tenerife, Spain. Stucke, T. S., & Sporer, S. L. (2002). When a grandiose self- Tracy, J. L., & Robins, R. W. (2004). Keeping the self in the image is threatened: Narcissism and self-concept clarity self-conscious emotions: Further arguments for a theo- as predictors of negative emotions and aggression retical model. Psychological Inquiry, 15, 171 – 177. following ego-threat. Journal of Personality, 70, 509 – 532. Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. Swann, W. B., Jr., & Read, S. J. (1981). Self-verification (2003). Stability of self-esteem across the life-span. processes: How we sustain our self-conceptions. Journal Journal of Personality and Social Psychology, 84, 205 – 220. of Experimental Social Psychology, 17, 351 – 372. Twenge, J. M., & Campbell, W. K. (2003). ‘‘Isn’t it fun to get Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New the respect that we’re going to deserve?’’ Narcissism, York: Guilford. , and aggression. Personality and Social Tangney, J. P., Wagner, P. E., Hill-Barlow, D., Marschall, Psychology Bulletin, 29, 261 – 272. D. E., & Gramzow, R. (1996). Relation of shame and guilt Willock, B. (1987). The devalued (unloved, repugnant) self. to constructive versus destructive responses to anger A second facet of narcissistic vulnerability in the aggres- across the lifespan. Journal of Personality and Social sive, conduct-disordered child. Psychoanalytic Psychology, Psychology, 70, 797 – 809. 4, 219 – 240. Thomaes, S., Stegge, H., Bushman, B. J., Olthof, T., & Wink, P. (1991). Two faces of narcissism. Journal of Person- Denissen, J. (2008). Development and validation of the ality and Social Psychology, 61, 590 – 597. Childhood Narcissism Scale. Journal of Personality Assess- Zakriski, A. L., & Coie, J. D. (1996). A comparison of ment, 90, 382 – 391. aggressive-rejected and nonaggressive-rejected child- Thomaes, S., Stegge, H., & Olthof, T. (2005, August). ren’s interpretations of self-directed and other-directed ‘‘Shame feels angry’’: On humiliated fury in children. Poster rejection. Child Development, 67, 1048 – 1070.