- Email: [email protected]
The associations of adolescent invulnerability and narcissism with problem behaviors
Personality and Individual Differences 47 (2009) 577–582
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The associations of adolescent invulnerability and narcissism with problem behaviors Christopher T. Barry *, Jessica D. Pickard, Lisa L. Ansel Department of Psychology, The University of Southern Mississippi, 118 College Dr., Box 5025, Hattiesburg, MS 39406, United States
a r t i c l e
i n f o
Article history: Received 20 February 2009 Received in revised form 11 May 2009 Accepted 13 May 2009 Available online 16 June 2009 Keywords: Invulnerability Narcissism Delinquency
a b s t r a c t This study examined the correlations between invulnerability, narcissism, self-esteem, delinquency, and aggression in a sample of at-risk adolescents. Participants were 213, 16–18 year-olds (169 male, 44 female). As expected, narcissism and invulnerability were related to delinquency and aggression. However, maladaptive narcissism predicted unique variance in delinquency and relational aggression. A negative effect for self-esteem emerged for predicting delinquency when controlling for narcissism and perceived invulnerability. The distributions of narcissism and invulnerability indicated that these constructs may not, as a rule, be elevated in such youth. Some preliminary implications for the role of these variables in adolescent problem behaviors are presented. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Elkind’s (1967) theory of adolescent egocentrism describes adolescence as a time of perceived invulnerability to risk and danger. This perceived invulnerability is thought to be related to the adolescent’s tendency to view him/herself as unique—known as the ‘‘personal fable”—which includes being uniquely immune to the consequences that might befall others (see Millstein & HalpernFelsher, 2002a). However, some research has disputed the notion that perceived invulnerability is confined to adolescence or that adolescence is a period of particularly strong beliefs in one’s invulnerability (e.g., Millstein & Halpern-Felsher, 2002a). As with invulnerability, narcissism—including grandiosity and concern for one’s social status (Raskin, Novacek, & Hogan, 1991)—could be considered descriptive of adolescence partly based on the concept of imaginary audience, wherein one perceives that others are closely observing him/her (Elkind, 1967). Another view is that narcissism comprises personality characteristics that develop prior to adolescence from early experiences with caretakers and that persist into adulthood (e.g., Kohut, 1971). Narcissism and invulnerability were positively correlated in one study (Aalsma, Lapsley, & Flannery, 2006), but their association has not been extensively investigated. The current study empirically investigated the relations between invulnerability, narcissism, and problem behaviors in at-risk adolescents, a sample selected because of the expected wide variability on the constructs of interest compared to what might occur in a community or detained sample. The distributions of invulnerability and narcissism were * Corresponding author. Tel.: +1 601 266 5374; fax: +1 601 266 5580. E-mail address: [email protected] (C.T. Barry). 0191-8869/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2009.05.022
also examined. Some theories might suggest that the distributions are negatively skewed among adolescents (Elkind, 1967; Waddell, 2006); however, a study with invulnerability found no evidence for this position (Millstein & Halpern-Felsher, 2002a), and another study found meaningful individual differences at low, moderate, and high levels of adolescent narcissism (Lapsley & Aalsma, 2006). Therefore, the central question of the present study was, ‘‘Are there identifiable and meaningful individual differences in invulnerability and narcissism during adolescence?” Research suggests that perceptions of invulnerability are not endemic to adolescents as a group (see Steinberg, 2007); instead, individual differences and the cognitive processes involved in perceived invulnerability may help explain how invulnerability might be relevant for delinquency and aggression for some adolescents. Perceived invulnerability has been associated with delinquency (Greene, Krcmar, Walters, Rubin, & Hale, 2000) and risk behaviors presumably because some adolescents’ sense of uniqueness and desire for independence make such behaviors appealing (Aalsma et al., 2006). However, some adolescents, although still developing their personalities and abilities to evaluate risks, may demonstrate sound judgment regarding the likelihood of negative consequences following risky behaviors (Millstein & Halpern-Felsher, 2002a). Recent approaches toward defining and measuring invulnerability have considered invulnerability multidimensional—including domains of general, danger, and interpersonal invulnerability—and as indicative of the adolescent developmental process of separation–individuation (see Aalsma et al., 2006). Such a conceptualization may shed light on whether or not particular thoughts of invulnerability translate to delinquent/aggressive actions, yet this possibility has not been extensively examined. A general sense of invulnerability may be inadequate for explaining why adolescents
578
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
might engage in risk-taking behavior (Goldberg, Halpern-Felsher, & Millstein, 2002). Instead, perceived invulnerability to danger may elucidate a link between invulnerability and delinquent activity (e.g., drug use, property crime) because of a lack of inhibition against acts that would typically fit societal notions of dangerous or problematic behavior. Interpersonal invulnerability (e.g., feeling immune one’s feelings being hurt) has been the target of much less research than danger invulnerability and does not directly indicate beliefs that would be related to delinquency or aggression. A tenuous link is that portraying oneself as immune to criticism may actually be associated with heightened sensitivity to criticism and aggression in the face of such incidents, not unlike previous findings regarding narcissism and responses to negative feedback from others (e.g., Bushman & Baumeister, 1998; Thomaes, Stegge, Bushman, & Olthof, 2008). Overall, based on previous research (see Millstein & Halpern-Felsher, 2002b) as well as conceptual differences in facets of invulnerability, the most likely area in which invulnerability is tied to adolescent problem behaviors is in perceived invulnerability to danger which is perhaps indicative of one’s beliefs of power over the probability of facing harm after risky actions. Beliefs of power over, and admiration by, others exemplified by narcissism may also be related to delinquency and aggression. Indeed, decades of research has indicated that at least some facets of narcissism in adults are related to aggression toward others and other antagonistic social behaviors (e.g., Bushman & Baumeister, 1998; Raskin & Terry, 1988), presumably because of the narcissist’s attempt to restore his/her desired self-image (Thomaes et al., 2008) or obtain admiration from others (Raskin et al., 1991). Initial adolescent research has also shown a link between narcissism and delinquency and aggression (e.g., Barry, Grafeman, Adler, & Pickard, 2007; Thomaes et al., 2008; Washburn, McMahon, King, Reinecke, & Silver, 2004). Furthermore, there appear to be meaningful differences between different domains of narcissism, with socalled maladaptive narcissism (Barry, Frick, Adler, & Grafeman, 2007; Washburn et al., 2004) being particularly associated with behavioral problems. On the other hand, some ‘‘adaptive” aspects of narcissism may be positively regarded and encouraged in adolescence, particularly in more individualistic societies (see Twenge & Foster, 2008). Adaptive narcissism (e.g., leadership, self-sufficiency) has been related to self-confidence and assertiveness (e.g., Emmons, 1984), whereas maladaptive narcissism (i.e., entitlement, exhibitionism, and exploitativeness) has been associated with behavioral and social maladjustment in adults (see Emmons, 1984; Raskin & Terry, 1988). Initial evidence suggests that this distinction between adaptive and maladaptive narcissism is meaningful in youth (e.g., Barry, Frick, & Killian, 2003; Barry, Frick et al., 2007; Washburn et al., 2004). Previous research has not directly considered the relation between narcissism and perceived invulnerability. However, other elements of self-perception such as grandiosity/beliefs of power or self-esteem may offer some insights into this issue. Although narcissism and invulnerability would be presumably associated with high, perhaps exaggerated, self-esteem, the correlation between narcissism and self-esteem has generally been non-significant in early adolescents (e.g., Barry et al., 2003; Thomaes et al., 2008) or positive in middle to older adolescents (e.g., Barry, Grafeman et al., 2007). In addition, adaptive narcissism has been consistently positively correlated with self-esteem, yet maladaptive narcissism has demonstrated negative or no relations with selfesteem in youth (Barry et al., 2003; Barry, Grafeman et al., 2007). The importance of self-esteem is not necessarily in its convergence with, or divergence from, narcissism or invulnerability; rather, self-esteem may help explain the connection between these constructs and behavioral problems. Research (Barry et al., 2003; Washburn et al., 2004) suggests that a combination of narcissism
and low self-esteem may be most indicative of youth behavioral problems; however, with older adolescent samples such as the sample in this study, a combination of high narcissism and high self-esteem may be associated with relatively high levels of aggression (Bushman et al., 2009). The associations between self-esteem and dimensions of invulnerability, as well as the interaction between self-esteem and invulnerability in the prediction of problem behaviors, have not yet been investigated. It was hypothesized that overall narcissism and maladaptive narcissism would be related to delinquency, overt (i.e., direct physical or verbal acts toward a victim), and relational (i.e., covert gossip, rumors, manipulation of others’ social status), aggression (Hypothesis 1). Danger invulnerability was expected to be correlated with delinquency and aggression (Hypothesis 2). All dimensions of invulnerability as well as adaptive and overall narcissism were hypothesized to be positively related to self-esteem (Hypothesis 3). It was also hypothesized (Hypothesis 4) that the combination of high self-esteem and high narcissism would correspond to the highest levels of aggression and delinquency based on findings with older adolescents (Bushman et al., 2009). Another, more theoretical issue involved the distributions of invulnerability and narcissism as individual difference variables in adolescents. The distributions were not expected to be skewed based on empirical evidence regarding individual differences in adolescent invulnerability and narcissism to date (Hypothesis 5). If individual differences are identifiable, their meaning will rest in their ability to predict variance in adolescent behavioral problems.
2. Method 2.1. Participants Participants were 213 adolescents (169 males, 44 females), ages 16–18 (mean = 16.77 years, sd = .69 years) enrolled in a voluntary 22-week military-style intervention program for youth who have dropped out of school. Participation in this study did not affect the intervention provided in the program, which provides academic remediation and life skills training in a structured environment. The sample was 67% Caucasian, 32% African American, and 1% ‘‘other,” all residing in the southeastern United States. Information regarding parents’/guardians’ occupation was obtained to calculate a socioeconomic index score (SEI) (Hauser & Featherman, 1977), with the sample having a mean SEI of 38.1 (sd = 21.89, range 0–92). The SEI was not correlated with aggression, delinquency, invulnerability, or narcissism and was therefore not used as a control variable in subsequent analyses. 2.2. Materials 2.2.1. Adolescent invulnerability scale (AIS; Duggan, Lapsley, & Norman, 2000) The AIS contains 21 statements to which respondents indicate their level of agreement. Although psychometric evidence is limited, the AIS has demonstrated good construct validity in adolescent community samples (Duggan et al., 2000). The AIS consists of six items assessing danger invulnerability (e.g., ‘‘I’m unlikely to get hurt if I did a dangerous thing.”), six items assessing interpersonal invulnerability (e.g., ‘‘It is just impossible for people to hurt my feelings.”), and nine items assessing general invulnerability (e.g., ‘‘Nothing bad will happen if I go to a place by myself.”). The possible range of scores for both danger and interpersonal invulnerability is 0–24, whereas responses can range from 0 to 36 for general invulnerability. The internal consistencies were a = .73 (danger invulnerability), a = .76 (interpersonal invulnerability), and a = .82 (general invulnerability).
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
2.2.2. Narcissistic personality inventory for children (NPIC; Barry et al., 2003) The NPIC contains 40 items developed directly from the 40 adult NPI items (Raskin & Terry, 1988). The items are in a forcedchoice format, such that respondents choose one statement from a pair. Respondents then rate the selected statement as being ‘‘really true” or ‘‘sort of true” for them. Thus, the NPIC provides a 4-point response scale for each item. The reliability and validity of the NPIC have been supported (Barry et al., 2003; Barry, Frick et al., 2007). One item was removed from analyses due to a low (i.e., r = .01) item-total correlation, resulting in a possible range of 0–117. In the present study, the internal consistency coefficient for the overall scale was a = .85. Based on correlates of the analogous adult scales, the items drawn from the NPI Self-sufficiency and Authority scales were summed to form an adaptive narcissism composite (see Barry et al., 2003) with an internal consistency of a = .72 for the current sample. A maladaptive narcissism composite (internal consistency = .69) was formed using the items from the Exhibitionism, Exploitativeness, and Entitlement scales (Raskin & Terry, 1988). Scores could range from 0 to 42 and 0 to 51 for adaptive and maladaptive narcissism, respectively. 2.2.3. Self-report of delinquency (SRD; Elliott & Ageton, 1980) The SRD assesses self-report of 34 illegal juvenile acts, asking respondents whether or not they have ever engaged in each act at any time in their lives. The SRD was developed from offenses reported in the Uniform Crime Report with a juvenile base rate greater than 1%. A participant’s delinquency score represented the number of separate acts in which he/she reported having engaged, with a possible range of 0–34. This score had a coefficient alpha of .92 for this sample. 2.2.4. Peer conflict scale (PCS; Marsee et al., 2007) The PCS, a 40-item self-report questionnaire, with 20 items assessing relational (e.g., ‘‘If others make me mad, I tell their secrets.”) and 20 items for overt (e.g., ‘‘If others make me mad, I hurt them.”) aggression. Responses are made on a 0–3 scale ranging from ‘‘not at all true” to ‘‘definitely true,” with scores potentially ranging from 0 to 60 for both overt and relational aggression. A large factor analysis of the PCS with adolescents has supported the conceptual differences of the PCS content (Marsee et al., 2007). The internal consistency coefficient for this sample was .88 for both the relational and the overt aggression scales. 2.2.5. Rosenberg self-esteem scale (RSES; Rosenberg, 1965) The RSES is a measure of global self-esteem, with 10 items on a four-point Likert-type scale ranging from ‘‘strongly disagree” to ‘‘strongly agree,” resulting in a possible range of 0–30. The RSES has been used widely with adolescents and adults and has repeatedly demonstrated good reliability and validity. The internal consistency of the RSES in this sample was .81. 2.3. Procedure The Institutional Review Board of the university with which the authors are affiliated approved this study. The director of the intervention program, who serves as guardian ad litem for the youths during their enrollment, was fully informed of this study and gave informed consent for the youths to participate. All adolescents attending the program were invited to participate, with approximately 90% of those individuals agreeing to participate following a detailed assent procedure. Participants completed questionnaires for this study and a larger research project in a classroom setting, with researchers monitoring the sessions to ensure that the participants had adequate understanding of all instruments. The items
579
were read aloud to participants, with the items also being provided to participants on paper. 3. Results Descriptive statistics are presented in Table 1 and indicate that the distributions of most variables had little skew and adequate variability, supporting the hypothesis. The distribution of relational aggression was positively skewed and somewhat leptokurtic indicating that most participants reported low rates of such behaviors with little variability. Correlations among the study variables are shown in Table 2. As shown in Table 2, narcissism, including maladaptive and adaptive narcissism, was significantly correlated with each of the invulnerability scales, with the exception of maladaptive narcissism not being significantly associated with interpersonal invulnerability. Interpersonal invulnerability was positively correlated with self-esteem, r = .23, p < .001 as were overall narcissism, r = .36, p < .001, adaptive narcissism, r = .39, p < .001, and maladaptive narcissism, r = .19, p < .01. The correlation between self-esteem and adaptive narcissism was significantly higher than the correlation between self-esteem and maladaptive narcissism, t(210) = 3.57, p < .01. Correlational analyses were suggestive of gender differences on overt aggression and invulnerability, with males tending to report higher scores on each of the AIS scales1 (see Table 2). Delinquency was positively correlated with overall narcissism as well as adaptive, r = .18, p < .01, and maladaptive narcissism, r = .27, p < .001. Danger invulnerability was significantly correlated with delinquency, r = .28, p < .001, overt aggression, r = .22, p < .01. General invulnerability was also correlated with overt aggression, as were overall, adaptive, and maladaptive narcissism (see Table 2). Similarly, relational aggression, was correlated with overall narcissism, r = .36, p < .001, adaptive, r = .24, p < .001, and maladaptive narcissism, r = .37, p < .001. Self-esteem was not significantly correlated with delinquency, overt aggression, or relational aggression. These results partially support Hypotheses 1–3. In regression analyses, each of the study variables (i.e., general, danger, and interpersonal invulnerability; adaptive and maladaptive narcissism; self-esteem) was entered as a predictor. The model predicting overt aggression included gender as a control variable in light of the significant relation between gender and overt aggression.1 Results of these analyses are shown in Table 3. Other than adaptive narcissism and interpersonal invulnerability, each of the other variables predicted unique variance in delinquency, R2 = .16, p < .001. However, the effects for general invulnerability, b = .21, p < .01, and self-esteem, b = .19, p < .05, were negative, even though neither was significantly correlated with delinquency. Danger invulnerability and maladaptive narcissism predicted unique variance in delinquency (see Table 3). For overt aggression, R2 = .15, p < .001, there was a gender main effect, b = .20, p < .01, with males tending to report higher overt aggression. Maladaptive narcissism also predicted unique variance in relational aggression, b = .33, p < .001, R2 = .18, p < .001. There was a significant negative effect for interpersonal invulnerability in this model, b = .18, p < .05. To test Hypothesis 4, concerning the moderating effect of selfesteem, narcissism and self-esteem were entered simultaneously
1 Post hoc revealed that males scored higher on average than females on all three facets of invulnerability and overt aggression. In addition, for females, adaptive narcissism was a unique predictor of variance in overt and relational aggression when controlling for maladaptive narcissism and the invulnerability scales. Analyses treating gender as a moderator in reduced regression models revealed a significant interaction for predicting relational aggression, b = .15, p < .05, such that adaptive narcissism was associated with relatively higher relational aggression only for females. This finding may be worth further examination in subsequent research.
580
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
Table 1 Descriptive statistics. Variable
Mean (sd)
Min.
Max.
Skewness
Kurtosis
Total narcissism Adaptive narcissism Maladaptive narcissism General invulnerability Danger invulnerability Interpersonal invulnerability Delinquency Overt aggression Relational aggression Self-esteem
54.71 (14.74) 21.03 (6.72) 21.94 (6.78) 12.55 (6.25) 8.27 (4.34) 13.27 (4.73) 12.65 (7.74) 13.06 (9.07) 6.30 (6.96) 18.77 (3.78)
15 1 6 0 0 1 0 0 0 6
101 42 42 36 22 24 32 39 38 25
.20 .02 .33 .28 .30 .09 .28 .60 1.85 .50
.70 .33 .33 .68 .19 .01 .81 .24 4.24 .01
Table 2 Correlations among study variables. 1. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. * ** ***
Overall narcissism Maladaptive narcissism Adaptive narcissism General invulnerability Danger invulnerability Interpersonal invulnerability Self-esteem Delinquency Relational aggression Overt aggression Gender
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
.85***
.85*** .61***
.25*** .32*** .23**
.18* .27*** .17* .62***
.15* .11 .29*** .43*** .23**
.36*** .19** .39*** .03 .02 .23**
.25*** .27*** .18** .13 .28*** .10 .10
.36*** .37*** .24*** .09 .08 .12 .02 .19**
.26*** .27*** .24*** .20** .22** .12 .00 .49*** .49***
.08 .11 .13 .25*** .17* .14 .13 .08 .01 .24***
p < .05. p < .01. p < .001.
Table 3 Regression analyses using invulnerability and narcissism as predictors of delinquency and aggression.
Gender General invulnerability Danger invulnerability Interpersonal invulnerability Maladaptive narcissism Adaptive narcissism Self-esteem R2
Delinquency
Overt aggression
Relational aggression
.18* .28** .11 .23** .07 .20* .16***
.20** .02 .12 .04 .16 .14 .12 .15***
.05 .02 .18* .33*** .11 .09 .18***
Note: standardized beta-weights are shown. * p < .05. ** p < .01. *** p < .001.
as predictors with delinquency as the criterion. In the subsequent step, the interaction term for narcissism and self-esteem was added. There was a significant main effect for narcissism, b = .33, p < .001, and a significant negative main effect for self-esteem, b = .21, p < .001, R2 = .10, p < .001, but no significant selfesteem by narcissism interaction. Likewise, there was no significant narcissism by self-esteem interaction in the prediction of overt or relational aggression. There was a main effect for narcissism for both overt, b = .31, p < .001, R2 = .08, and relational aggression b = .44, p < .001, R2 = .16, as well as a significant negative main effect for self-esteem in predicting relational aggression, b = .18, p < .05. Repeating this approach using danger invulnerability, instead of narcissism, yielded no significant interactions for predicting delinquency or either form of aggression. Only main effects for danger invulnerability for predicting delinquency, b = .28, p < .001, and overt aggression, b = .22, p < .01, were found.
4. Discussion This study replicated previous research (e.g., Barry, Grafeman et al., 2007; Washburn et al., 2004) regarding the association between narcissism, especially maladaptive narcissism, and adolescent delinquency and aggression. A sense of invulnerability to danger was also associated with problem behaviors, and the distributions of invulnerability and narcissism were relatively normal in shape. Nevertheless, the cross-sectional nature of this study makes it impossible to determine the typical developmental trajectories of these variables. That is, narcissism and perceptions of invulnerability may be individual difference variables that develop in childhood or adolescence and remain relatively stable into adulthood. On the other hand, it is possible that high levels of these characteristics are less prevalent later in life such that their distributions become positively skewed. Clearly, longitudinal studies are necessary to address such questions.
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
The current study was also an initial attempt to evaluate the relation between adolescent invulnerability and narcissism. As expected, they were significantly correlated, although the effects were small to moderate in magnitude. In essence, narcissism and invulnerability appear to have behavioral relevance for adolescents, and these theoretically similar constructs have relatively limited redundancy. However, there is a need for further understanding of the mechanisms involved in the development of narcissism and perceived invulnerability as well as the factors that might explain their link to problem behaviors. For example, the association between maladaptive narcissism and relational aggression may be particularly influenced by aspects of maladaptive narcissism such as a sense of entitlement and tendency to exploit others which might serve to justify relational aggression as means to desired ends. Thus, beliefs of deserving positive responses from the outside world may be an important issue to consider and upon which to intervene. Although the etiology of maladaptive narcissism is unclear, it could be an individual difference variable indicative of greater persistence in problem behaviors as opposed to behavior problems that are more contextually influenced and relatively limited to adolescence (see Moffitt, 2006). Similarly, the content of the danger invulnerability scale of the AIS (e.g., ‘‘I could probably drink and drive without getting into an accident.”) is suggestive of thought processes that could promote, or justify, delinquent activity. An obvious next step in this line of research is to evaluate the cognitive interpretations of individuals with high levels of narcissism and invulnerability in social and dangerous situations as well as the roles of narcissism and invulnerability in the justification of antisocial behavior. Nevertheless, as the developmental psychopathology and the personality literatures attest, it would be premature and unsound to paint narcissism and invulnerability in adolescence as broadly maladaptive. In the current study, when controlling for other dimensions of invulnerability, narcissism, and self-esteem, general invulnerability had a significant negative effect in predicting delinquency. This finding may indicate some unique aspects of general invulnerability that are not indicative of behavioral problems. A similar pattern was found for interpersonal invulnerability in the prediction of relational aggression, suggesting that adolescents who feel impervious to the consequences of others’ actions toward them may feel less compelled to participate in aggression that targets others’ social status or reputation. However, individuals who are more sensitive to social criticism may be more likely to engage in relational aggression. Indeed, narcissistic adults are particularly likely to respond to criticism with aggression (e.g., Bushman & Baumeister, 1998). Youth with high levels of narcissism would appear to be at-risk for responding in a hostile manner to social criticism or rejection based on the emerging research. The interpretation of, and coping with, negative social feedback by these individuals is a necessary next avenue of exploration. The role of self-esteem in adolescent behavioral problems remains unclear. Self-esteem did not moderate the associations between narcissism and problem behaviors. In addition, despite the positive relation between narcissism and self-esteem and interpersonal invulnerability, low self-esteem was associated with delinquency when controlling for invulnerability and narcissism, replicating similar findings (Barry, Grafeman et al., 2007). More research is needed to understand the relation between self-esteem and adolescent behavior problems and the influence of narcissism in this relation. This study had a number of limitations. Shared source variance may account for some of the observed relations because of the reliance on self-reports. However, the participants were likely old enough to provide accurate accounts of their behavior, and selfreport may be the optimal way to evaluate narcissism, invulnerability, and self-esteem. Future studies should, nevertheless, inves-
581
tigate the issues addressed in this study, using other sources. Importantly, this sample consisted of adolescents who had dropped out of school and were attending a voluntary residential program, limiting the study’s generalizability. These participants were part of a larger project on risk and protective factors related to adolescent problem behaviors and were selected because they presumably would exhibit suitable variability on problem behaviors as well as other variables of interest. This sample was suitable for the exploratory nature of this study, yet this research should be replicated in broader adolescent samples. Furthermore, the retrospective nature of the self-reported delinquency measure makes it impossible to determine any causal mechanisms relative to present narcissism or feelings of invulnerability. As noted above, the intervening variables that may explain the relations between narcissism, invulnerability, and behavioral problems should be explored in future research. Contextual variables such as parenting and peer delinquency (see Steinberg, 2007) may offer additional insight as to conditions under which narcissism and invulnerability may be particularly related to aggression and delinquency. In reality, some degree of invulnerability and narcissism may be normative through adolescence and early adulthood, yet heightened danger invulnerability or a tendency to exploit others may be risk factors to which attention is needed. Therefore, awareness of these perceptions, their development, and their behavioral manifestations could help inform efforts to prevent or curtail problematic behaviors among adolescents. References Aalsma, M. C., Lapsley, D. K., & Flannery, D. (2006). Personal fables, narcissism, and adolescent adjustment. Psychology in the Schools, 43, 481–491. Barry, C. T., Frick, P. J., Adler, K. K., & Grafeman, S. J. (2007). The predictive utility of narcissism among children and adolescents: Evidence for a distinction between adaptive and maladaptive narcissism. Journal of Child and Family Studies, 16, 508–521. Barry, C. T., Frick, P. J., & Killian, A. L. (2003). The relation of narcissism and selfesteem to conduct problems in children: A preliminary investigation. Journal of Clinical Child and Adolescent Psychology, 32, 139–152. Barry, C. T., Grafeman, S. J., Adler, K. K., & Pickard, J. D. (2007). The relations among narcissism, self-esteem, and delinquency in a sample of at-risk adolescents. Journal of Adolescence, 30, 933–942. Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcissism, selfesteem and direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of Personality and Social Psychology, 75, 219–229. Bushman, B. J., Baumeister, R. F., Thomaes, S., Ryu, E., Begeer, S., & West, S. G. (2009). Looking again, and harder, for a link between low self-esteem and aggression. Journal of Personality, 77, 427–446. Duggan, P.M., Lapsley, D.K. & Norman, K. (2000). Adolescent invulnerability and personal uniqueness: Scale development and initial construct validation. In Paper presented at the biennial meeting of the Society for Research on Adolescence, Chicago. Elkind, D. (1967). Egocentrism in adolescence. Child Development, 38, 1025– 1034. Elliott, D. S., & Ageton, S. S. (1980). Reconciling race and class differences in selfreported and official estimates of delinquency. American Sociological Review, 45, 95–110. Emmons, R. A. (1984). Factor analysis and construct validation of the narcissistic personality inventory. Journal of Personality Assessment, 48, 291–300. Goldberg, J., Halpern-Felsher, B., & Millstein, S. (2002). Beyond invulnerability: The importance of benefits in adolescents’ decision to drink alcohol. Health Psychology, 21, 477–484. Greene, K., Krcmar, M., Walters, L., Rubin, D., & Hale, J. (2000). Targeting adolescent risk-taking behaviors: The contribution of egocentrism and sensation-seeking. Journal of Adolescence, 23, 439–461. Hauser, R. M., & Featherman, D. L. (1977). The process of stratification. New York: Academic Press. Kohut, H. (1971). The analysis of the self. New York: International Universities Press. Lapsley, D. K., & Aalsma, M. C. (2006). An empirical typology of narcissism and mental health in late adolescence. Journal of Adolescence, 29, 53–71. Marsee, M.A., Barry, C.T., Frick, P.J., Kimonis, E.R., Munoz, L.C., & Aucoin, K.J. (2007). Factor structure of the Peer Conflict Scale in at-risk youth. In Paper presented at the 2007 Southeastern Psychological Association Conference in New Orleans, LA. Millstein, S. G., & Halpern-Felsher, B. L. (2002a). Judgments about risk and perceived invulnerability in adolescents and young adults. Journal of Research on Adolescence, 12, 399–422. Millstein, S. G., & Halpern-Felsher, B. L. (2002b). Perceptions of risk and vulnerability. Journal of Adolescent Health, 31, 10–27.
582
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
Moffitt, T. E. (2006). Life-course-persistent versus adolescence-limited antisocial behavior (2nd ed.. In D. Cicchetti & D. J. Cohen (Eds.). Developmental psychopathology (Vol. 3, pp. 570–598). Hoboken, NJ: John Wiley & Sons, Inc. Raskin, R., Novacek, J., & Hogan, R. (1991). Narcissism, self-esteem, and defensive self-enhancement. Journal of Personality, 59, 19–38. Raskin, R. N., & Terry, H. (1988). A principal-components analysis of the narcissistic personality inventory and further evidence of its construct validity. Journal of Personality and Social Psychology, 54, 890–902. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Steinberg, L. (2007). Risk taking in adolescence. New perspectives from brain and behavioral science. Current Directions in Psychological Science, 16, 55–59.
Thomaes, S., Stegge, H., Bushman, B. J., & Olthof, T. (2008). Trumping shame by blasts of noise: Narcissism, self-esteem, shame, and aggression in young adolescents. Child Development, 79, 1792–1801. Twenge, J., & Foster, J. (2008). Mapping the scale of the narcissism epidemic: Increases in narcissism 2002–2007 within ethnic groups. Journal of Research in Personality, 42, 1619–1622. Waddell, M. (2006). Narcissism – an adolescent disorder? Journal of Child Psychotherapy, 32, 21–34. Washburn, J. J., McMahon, S. D., King, C. A., Reinecke, M. A., & Silver, C. (2004). Narcissistic features in young adolescents: Relations to aggression and internalizing symptoms. Journal of Youth and Adolescence, 33, 247–260.
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
The associations of adolescent invulnerability and narcissism with problem behaviors Christopher T. Barry *, Jessica D. Pickard, Lisa L. Ansel Department of Psychology, The University of Southern Mississippi, 118 College Dr., Box 5025, Hattiesburg, MS 39406, United States
a r t i c l e
i n f o
Article history: Received 20 February 2009 Received in revised form 11 May 2009 Accepted 13 May 2009 Available online 16 June 2009 Keywords: Invulnerability Narcissism Delinquency
a b s t r a c t This study examined the correlations between invulnerability, narcissism, self-esteem, delinquency, and aggression in a sample of at-risk adolescents. Participants were 213, 16–18 year-olds (169 male, 44 female). As expected, narcissism and invulnerability were related to delinquency and aggression. However, maladaptive narcissism predicted unique variance in delinquency and relational aggression. A negative effect for self-esteem emerged for predicting delinquency when controlling for narcissism and perceived invulnerability. The distributions of narcissism and invulnerability indicated that these constructs may not, as a rule, be elevated in such youth. Some preliminary implications for the role of these variables in adolescent problem behaviors are presented. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Elkind’s (1967) theory of adolescent egocentrism describes adolescence as a time of perceived invulnerability to risk and danger. This perceived invulnerability is thought to be related to the adolescent’s tendency to view him/herself as unique—known as the ‘‘personal fable”—which includes being uniquely immune to the consequences that might befall others (see Millstein & HalpernFelsher, 2002a). However, some research has disputed the notion that perceived invulnerability is confined to adolescence or that adolescence is a period of particularly strong beliefs in one’s invulnerability (e.g., Millstein & Halpern-Felsher, 2002a). As with invulnerability, narcissism—including grandiosity and concern for one’s social status (Raskin, Novacek, & Hogan, 1991)—could be considered descriptive of adolescence partly based on the concept of imaginary audience, wherein one perceives that others are closely observing him/her (Elkind, 1967). Another view is that narcissism comprises personality characteristics that develop prior to adolescence from early experiences with caretakers and that persist into adulthood (e.g., Kohut, 1971). Narcissism and invulnerability were positively correlated in one study (Aalsma, Lapsley, & Flannery, 2006), but their association has not been extensively investigated. The current study empirically investigated the relations between invulnerability, narcissism, and problem behaviors in at-risk adolescents, a sample selected because of the expected wide variability on the constructs of interest compared to what might occur in a community or detained sample. The distributions of invulnerability and narcissism were * Corresponding author. Tel.: +1 601 266 5374; fax: +1 601 266 5580. E-mail address: [email protected] (C.T. Barry). 0191-8869/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2009.05.022
also examined. Some theories might suggest that the distributions are negatively skewed among adolescents (Elkind, 1967; Waddell, 2006); however, a study with invulnerability found no evidence for this position (Millstein & Halpern-Felsher, 2002a), and another study found meaningful individual differences at low, moderate, and high levels of adolescent narcissism (Lapsley & Aalsma, 2006). Therefore, the central question of the present study was, ‘‘Are there identifiable and meaningful individual differences in invulnerability and narcissism during adolescence?” Research suggests that perceptions of invulnerability are not endemic to adolescents as a group (see Steinberg, 2007); instead, individual differences and the cognitive processes involved in perceived invulnerability may help explain how invulnerability might be relevant for delinquency and aggression for some adolescents. Perceived invulnerability has been associated with delinquency (Greene, Krcmar, Walters, Rubin, & Hale, 2000) and risk behaviors presumably because some adolescents’ sense of uniqueness and desire for independence make such behaviors appealing (Aalsma et al., 2006). However, some adolescents, although still developing their personalities and abilities to evaluate risks, may demonstrate sound judgment regarding the likelihood of negative consequences following risky behaviors (Millstein & Halpern-Felsher, 2002a). Recent approaches toward defining and measuring invulnerability have considered invulnerability multidimensional—including domains of general, danger, and interpersonal invulnerability—and as indicative of the adolescent developmental process of separation–individuation (see Aalsma et al., 2006). Such a conceptualization may shed light on whether or not particular thoughts of invulnerability translate to delinquent/aggressive actions, yet this possibility has not been extensively examined. A general sense of invulnerability may be inadequate for explaining why adolescents
578
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
might engage in risk-taking behavior (Goldberg, Halpern-Felsher, & Millstein, 2002). Instead, perceived invulnerability to danger may elucidate a link between invulnerability and delinquent activity (e.g., drug use, property crime) because of a lack of inhibition against acts that would typically fit societal notions of dangerous or problematic behavior. Interpersonal invulnerability (e.g., feeling immune one’s feelings being hurt) has been the target of much less research than danger invulnerability and does not directly indicate beliefs that would be related to delinquency or aggression. A tenuous link is that portraying oneself as immune to criticism may actually be associated with heightened sensitivity to criticism and aggression in the face of such incidents, not unlike previous findings regarding narcissism and responses to negative feedback from others (e.g., Bushman & Baumeister, 1998; Thomaes, Stegge, Bushman, & Olthof, 2008). Overall, based on previous research (see Millstein & Halpern-Felsher, 2002b) as well as conceptual differences in facets of invulnerability, the most likely area in which invulnerability is tied to adolescent problem behaviors is in perceived invulnerability to danger which is perhaps indicative of one’s beliefs of power over the probability of facing harm after risky actions. Beliefs of power over, and admiration by, others exemplified by narcissism may also be related to delinquency and aggression. Indeed, decades of research has indicated that at least some facets of narcissism in adults are related to aggression toward others and other antagonistic social behaviors (e.g., Bushman & Baumeister, 1998; Raskin & Terry, 1988), presumably because of the narcissist’s attempt to restore his/her desired self-image (Thomaes et al., 2008) or obtain admiration from others (Raskin et al., 1991). Initial adolescent research has also shown a link between narcissism and delinquency and aggression (e.g., Barry, Grafeman, Adler, & Pickard, 2007; Thomaes et al., 2008; Washburn, McMahon, King, Reinecke, & Silver, 2004). Furthermore, there appear to be meaningful differences between different domains of narcissism, with socalled maladaptive narcissism (Barry, Frick, Adler, & Grafeman, 2007; Washburn et al., 2004) being particularly associated with behavioral problems. On the other hand, some ‘‘adaptive” aspects of narcissism may be positively regarded and encouraged in adolescence, particularly in more individualistic societies (see Twenge & Foster, 2008). Adaptive narcissism (e.g., leadership, self-sufficiency) has been related to self-confidence and assertiveness (e.g., Emmons, 1984), whereas maladaptive narcissism (i.e., entitlement, exhibitionism, and exploitativeness) has been associated with behavioral and social maladjustment in adults (see Emmons, 1984; Raskin & Terry, 1988). Initial evidence suggests that this distinction between adaptive and maladaptive narcissism is meaningful in youth (e.g., Barry, Frick, & Killian, 2003; Barry, Frick et al., 2007; Washburn et al., 2004). Previous research has not directly considered the relation between narcissism and perceived invulnerability. However, other elements of self-perception such as grandiosity/beliefs of power or self-esteem may offer some insights into this issue. Although narcissism and invulnerability would be presumably associated with high, perhaps exaggerated, self-esteem, the correlation between narcissism and self-esteem has generally been non-significant in early adolescents (e.g., Barry et al., 2003; Thomaes et al., 2008) or positive in middle to older adolescents (e.g., Barry, Grafeman et al., 2007). In addition, adaptive narcissism has been consistently positively correlated with self-esteem, yet maladaptive narcissism has demonstrated negative or no relations with selfesteem in youth (Barry et al., 2003; Barry, Grafeman et al., 2007). The importance of self-esteem is not necessarily in its convergence with, or divergence from, narcissism or invulnerability; rather, self-esteem may help explain the connection between these constructs and behavioral problems. Research (Barry et al., 2003; Washburn et al., 2004) suggests that a combination of narcissism
and low self-esteem may be most indicative of youth behavioral problems; however, with older adolescent samples such as the sample in this study, a combination of high narcissism and high self-esteem may be associated with relatively high levels of aggression (Bushman et al., 2009). The associations between self-esteem and dimensions of invulnerability, as well as the interaction between self-esteem and invulnerability in the prediction of problem behaviors, have not yet been investigated. It was hypothesized that overall narcissism and maladaptive narcissism would be related to delinquency, overt (i.e., direct physical or verbal acts toward a victim), and relational (i.e., covert gossip, rumors, manipulation of others’ social status), aggression (Hypothesis 1). Danger invulnerability was expected to be correlated with delinquency and aggression (Hypothesis 2). All dimensions of invulnerability as well as adaptive and overall narcissism were hypothesized to be positively related to self-esteem (Hypothesis 3). It was also hypothesized (Hypothesis 4) that the combination of high self-esteem and high narcissism would correspond to the highest levels of aggression and delinquency based on findings with older adolescents (Bushman et al., 2009). Another, more theoretical issue involved the distributions of invulnerability and narcissism as individual difference variables in adolescents. The distributions were not expected to be skewed based on empirical evidence regarding individual differences in adolescent invulnerability and narcissism to date (Hypothesis 5). If individual differences are identifiable, their meaning will rest in their ability to predict variance in adolescent behavioral problems.
2. Method 2.1. Participants Participants were 213 adolescents (169 males, 44 females), ages 16–18 (mean = 16.77 years, sd = .69 years) enrolled in a voluntary 22-week military-style intervention program for youth who have dropped out of school. Participation in this study did not affect the intervention provided in the program, which provides academic remediation and life skills training in a structured environment. The sample was 67% Caucasian, 32% African American, and 1% ‘‘other,” all residing in the southeastern United States. Information regarding parents’/guardians’ occupation was obtained to calculate a socioeconomic index score (SEI) (Hauser & Featherman, 1977), with the sample having a mean SEI of 38.1 (sd = 21.89, range 0–92). The SEI was not correlated with aggression, delinquency, invulnerability, or narcissism and was therefore not used as a control variable in subsequent analyses. 2.2. Materials 2.2.1. Adolescent invulnerability scale (AIS; Duggan, Lapsley, & Norman, 2000) The AIS contains 21 statements to which respondents indicate their level of agreement. Although psychometric evidence is limited, the AIS has demonstrated good construct validity in adolescent community samples (Duggan et al., 2000). The AIS consists of six items assessing danger invulnerability (e.g., ‘‘I’m unlikely to get hurt if I did a dangerous thing.”), six items assessing interpersonal invulnerability (e.g., ‘‘It is just impossible for people to hurt my feelings.”), and nine items assessing general invulnerability (e.g., ‘‘Nothing bad will happen if I go to a place by myself.”). The possible range of scores for both danger and interpersonal invulnerability is 0–24, whereas responses can range from 0 to 36 for general invulnerability. The internal consistencies were a = .73 (danger invulnerability), a = .76 (interpersonal invulnerability), and a = .82 (general invulnerability).
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
2.2.2. Narcissistic personality inventory for children (NPIC; Barry et al., 2003) The NPIC contains 40 items developed directly from the 40 adult NPI items (Raskin & Terry, 1988). The items are in a forcedchoice format, such that respondents choose one statement from a pair. Respondents then rate the selected statement as being ‘‘really true” or ‘‘sort of true” for them. Thus, the NPIC provides a 4-point response scale for each item. The reliability and validity of the NPIC have been supported (Barry et al., 2003; Barry, Frick et al., 2007). One item was removed from analyses due to a low (i.e., r = .01) item-total correlation, resulting in a possible range of 0–117. In the present study, the internal consistency coefficient for the overall scale was a = .85. Based on correlates of the analogous adult scales, the items drawn from the NPI Self-sufficiency and Authority scales were summed to form an adaptive narcissism composite (see Barry et al., 2003) with an internal consistency of a = .72 for the current sample. A maladaptive narcissism composite (internal consistency = .69) was formed using the items from the Exhibitionism, Exploitativeness, and Entitlement scales (Raskin & Terry, 1988). Scores could range from 0 to 42 and 0 to 51 for adaptive and maladaptive narcissism, respectively. 2.2.3. Self-report of delinquency (SRD; Elliott & Ageton, 1980) The SRD assesses self-report of 34 illegal juvenile acts, asking respondents whether or not they have ever engaged in each act at any time in their lives. The SRD was developed from offenses reported in the Uniform Crime Report with a juvenile base rate greater than 1%. A participant’s delinquency score represented the number of separate acts in which he/she reported having engaged, with a possible range of 0–34. This score had a coefficient alpha of .92 for this sample. 2.2.4. Peer conflict scale (PCS; Marsee et al., 2007) The PCS, a 40-item self-report questionnaire, with 20 items assessing relational (e.g., ‘‘If others make me mad, I tell their secrets.”) and 20 items for overt (e.g., ‘‘If others make me mad, I hurt them.”) aggression. Responses are made on a 0–3 scale ranging from ‘‘not at all true” to ‘‘definitely true,” with scores potentially ranging from 0 to 60 for both overt and relational aggression. A large factor analysis of the PCS with adolescents has supported the conceptual differences of the PCS content (Marsee et al., 2007). The internal consistency coefficient for this sample was .88 for both the relational and the overt aggression scales. 2.2.5. Rosenberg self-esteem scale (RSES; Rosenberg, 1965) The RSES is a measure of global self-esteem, with 10 items on a four-point Likert-type scale ranging from ‘‘strongly disagree” to ‘‘strongly agree,” resulting in a possible range of 0–30. The RSES has been used widely with adolescents and adults and has repeatedly demonstrated good reliability and validity. The internal consistency of the RSES in this sample was .81. 2.3. Procedure The Institutional Review Board of the university with which the authors are affiliated approved this study. The director of the intervention program, who serves as guardian ad litem for the youths during their enrollment, was fully informed of this study and gave informed consent for the youths to participate. All adolescents attending the program were invited to participate, with approximately 90% of those individuals agreeing to participate following a detailed assent procedure. Participants completed questionnaires for this study and a larger research project in a classroom setting, with researchers monitoring the sessions to ensure that the participants had adequate understanding of all instruments. The items
579
were read aloud to participants, with the items also being provided to participants on paper. 3. Results Descriptive statistics are presented in Table 1 and indicate that the distributions of most variables had little skew and adequate variability, supporting the hypothesis. The distribution of relational aggression was positively skewed and somewhat leptokurtic indicating that most participants reported low rates of such behaviors with little variability. Correlations among the study variables are shown in Table 2. As shown in Table 2, narcissism, including maladaptive and adaptive narcissism, was significantly correlated with each of the invulnerability scales, with the exception of maladaptive narcissism not being significantly associated with interpersonal invulnerability. Interpersonal invulnerability was positively correlated with self-esteem, r = .23, p < .001 as were overall narcissism, r = .36, p < .001, adaptive narcissism, r = .39, p < .001, and maladaptive narcissism, r = .19, p < .01. The correlation between self-esteem and adaptive narcissism was significantly higher than the correlation between self-esteem and maladaptive narcissism, t(210) = 3.57, p < .01. Correlational analyses were suggestive of gender differences on overt aggression and invulnerability, with males tending to report higher scores on each of the AIS scales1 (see Table 2). Delinquency was positively correlated with overall narcissism as well as adaptive, r = .18, p < .01, and maladaptive narcissism, r = .27, p < .001. Danger invulnerability was significantly correlated with delinquency, r = .28, p < .001, overt aggression, r = .22, p < .01. General invulnerability was also correlated with overt aggression, as were overall, adaptive, and maladaptive narcissism (see Table 2). Similarly, relational aggression, was correlated with overall narcissism, r = .36, p < .001, adaptive, r = .24, p < .001, and maladaptive narcissism, r = .37, p < .001. Self-esteem was not significantly correlated with delinquency, overt aggression, or relational aggression. These results partially support Hypotheses 1–3. In regression analyses, each of the study variables (i.e., general, danger, and interpersonal invulnerability; adaptive and maladaptive narcissism; self-esteem) was entered as a predictor. The model predicting overt aggression included gender as a control variable in light of the significant relation between gender and overt aggression.1 Results of these analyses are shown in Table 3. Other than adaptive narcissism and interpersonal invulnerability, each of the other variables predicted unique variance in delinquency, R2 = .16, p < .001. However, the effects for general invulnerability, b = .21, p < .01, and self-esteem, b = .19, p < .05, were negative, even though neither was significantly correlated with delinquency. Danger invulnerability and maladaptive narcissism predicted unique variance in delinquency (see Table 3). For overt aggression, R2 = .15, p < .001, there was a gender main effect, b = .20, p < .01, with males tending to report higher overt aggression. Maladaptive narcissism also predicted unique variance in relational aggression, b = .33, p < .001, R2 = .18, p < .001. There was a significant negative effect for interpersonal invulnerability in this model, b = .18, p < .05. To test Hypothesis 4, concerning the moderating effect of selfesteem, narcissism and self-esteem were entered simultaneously
1 Post hoc revealed that males scored higher on average than females on all three facets of invulnerability and overt aggression. In addition, for females, adaptive narcissism was a unique predictor of variance in overt and relational aggression when controlling for maladaptive narcissism and the invulnerability scales. Analyses treating gender as a moderator in reduced regression models revealed a significant interaction for predicting relational aggression, b = .15, p < .05, such that adaptive narcissism was associated with relatively higher relational aggression only for females. This finding may be worth further examination in subsequent research.
580
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
Table 1 Descriptive statistics. Variable
Mean (sd)
Min.
Max.
Skewness
Kurtosis
Total narcissism Adaptive narcissism Maladaptive narcissism General invulnerability Danger invulnerability Interpersonal invulnerability Delinquency Overt aggression Relational aggression Self-esteem
54.71 (14.74) 21.03 (6.72) 21.94 (6.78) 12.55 (6.25) 8.27 (4.34) 13.27 (4.73) 12.65 (7.74) 13.06 (9.07) 6.30 (6.96) 18.77 (3.78)
15 1 6 0 0 1 0 0 0 6
101 42 42 36 22 24 32 39 38 25
.20 .02 .33 .28 .30 .09 .28 .60 1.85 .50
.70 .33 .33 .68 .19 .01 .81 .24 4.24 .01
Table 2 Correlations among study variables. 1. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. * ** ***
Overall narcissism Maladaptive narcissism Adaptive narcissism General invulnerability Danger invulnerability Interpersonal invulnerability Self-esteem Delinquency Relational aggression Overt aggression Gender
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
.85***
.85*** .61***
.25*** .32*** .23**
.18* .27*** .17* .62***
.15* .11 .29*** .43*** .23**
.36*** .19** .39*** .03 .02 .23**
.25*** .27*** .18** .13 .28*** .10 .10
.36*** .37*** .24*** .09 .08 .12 .02 .19**
.26*** .27*** .24*** .20** .22** .12 .00 .49*** .49***
.08 .11 .13 .25*** .17* .14 .13 .08 .01 .24***
p < .05. p < .01. p < .001.
Table 3 Regression analyses using invulnerability and narcissism as predictors of delinquency and aggression.
Gender General invulnerability Danger invulnerability Interpersonal invulnerability Maladaptive narcissism Adaptive narcissism Self-esteem R2
Delinquency
Overt aggression
Relational aggression
.18* .28** .11 .23** .07 .20* .16***
.20** .02 .12 .04 .16 .14 .12 .15***
.05 .02 .18* .33*** .11 .09 .18***
Note: standardized beta-weights are shown. * p < .05. ** p < .01. *** p < .001.
as predictors with delinquency as the criterion. In the subsequent step, the interaction term for narcissism and self-esteem was added. There was a significant main effect for narcissism, b = .33, p < .001, and a significant negative main effect for self-esteem, b = .21, p < .001, R2 = .10, p < .001, but no significant selfesteem by narcissism interaction. Likewise, there was no significant narcissism by self-esteem interaction in the prediction of overt or relational aggression. There was a main effect for narcissism for both overt, b = .31, p < .001, R2 = .08, and relational aggression b = .44, p < .001, R2 = .16, as well as a significant negative main effect for self-esteem in predicting relational aggression, b = .18, p < .05. Repeating this approach using danger invulnerability, instead of narcissism, yielded no significant interactions for predicting delinquency or either form of aggression. Only main effects for danger invulnerability for predicting delinquency, b = .28, p < .001, and overt aggression, b = .22, p < .01, were found.
4. Discussion This study replicated previous research (e.g., Barry, Grafeman et al., 2007; Washburn et al., 2004) regarding the association between narcissism, especially maladaptive narcissism, and adolescent delinquency and aggression. A sense of invulnerability to danger was also associated with problem behaviors, and the distributions of invulnerability and narcissism were relatively normal in shape. Nevertheless, the cross-sectional nature of this study makes it impossible to determine the typical developmental trajectories of these variables. That is, narcissism and perceptions of invulnerability may be individual difference variables that develop in childhood or adolescence and remain relatively stable into adulthood. On the other hand, it is possible that high levels of these characteristics are less prevalent later in life such that their distributions become positively skewed. Clearly, longitudinal studies are necessary to address such questions.
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
The current study was also an initial attempt to evaluate the relation between adolescent invulnerability and narcissism. As expected, they were significantly correlated, although the effects were small to moderate in magnitude. In essence, narcissism and invulnerability appear to have behavioral relevance for adolescents, and these theoretically similar constructs have relatively limited redundancy. However, there is a need for further understanding of the mechanisms involved in the development of narcissism and perceived invulnerability as well as the factors that might explain their link to problem behaviors. For example, the association between maladaptive narcissism and relational aggression may be particularly influenced by aspects of maladaptive narcissism such as a sense of entitlement and tendency to exploit others which might serve to justify relational aggression as means to desired ends. Thus, beliefs of deserving positive responses from the outside world may be an important issue to consider and upon which to intervene. Although the etiology of maladaptive narcissism is unclear, it could be an individual difference variable indicative of greater persistence in problem behaviors as opposed to behavior problems that are more contextually influenced and relatively limited to adolescence (see Moffitt, 2006). Similarly, the content of the danger invulnerability scale of the AIS (e.g., ‘‘I could probably drink and drive without getting into an accident.”) is suggestive of thought processes that could promote, or justify, delinquent activity. An obvious next step in this line of research is to evaluate the cognitive interpretations of individuals with high levels of narcissism and invulnerability in social and dangerous situations as well as the roles of narcissism and invulnerability in the justification of antisocial behavior. Nevertheless, as the developmental psychopathology and the personality literatures attest, it would be premature and unsound to paint narcissism and invulnerability in adolescence as broadly maladaptive. In the current study, when controlling for other dimensions of invulnerability, narcissism, and self-esteem, general invulnerability had a significant negative effect in predicting delinquency. This finding may indicate some unique aspects of general invulnerability that are not indicative of behavioral problems. A similar pattern was found for interpersonal invulnerability in the prediction of relational aggression, suggesting that adolescents who feel impervious to the consequences of others’ actions toward them may feel less compelled to participate in aggression that targets others’ social status or reputation. However, individuals who are more sensitive to social criticism may be more likely to engage in relational aggression. Indeed, narcissistic adults are particularly likely to respond to criticism with aggression (e.g., Bushman & Baumeister, 1998). Youth with high levels of narcissism would appear to be at-risk for responding in a hostile manner to social criticism or rejection based on the emerging research. The interpretation of, and coping with, negative social feedback by these individuals is a necessary next avenue of exploration. The role of self-esteem in adolescent behavioral problems remains unclear. Self-esteem did not moderate the associations between narcissism and problem behaviors. In addition, despite the positive relation between narcissism and self-esteem and interpersonal invulnerability, low self-esteem was associated with delinquency when controlling for invulnerability and narcissism, replicating similar findings (Barry, Grafeman et al., 2007). More research is needed to understand the relation between self-esteem and adolescent behavior problems and the influence of narcissism in this relation. This study had a number of limitations. Shared source variance may account for some of the observed relations because of the reliance on self-reports. However, the participants were likely old enough to provide accurate accounts of their behavior, and selfreport may be the optimal way to evaluate narcissism, invulnerability, and self-esteem. Future studies should, nevertheless, inves-
581
tigate the issues addressed in this study, using other sources. Importantly, this sample consisted of adolescents who had dropped out of school and were attending a voluntary residential program, limiting the study’s generalizability. These participants were part of a larger project on risk and protective factors related to adolescent problem behaviors and were selected because they presumably would exhibit suitable variability on problem behaviors as well as other variables of interest. This sample was suitable for the exploratory nature of this study, yet this research should be replicated in broader adolescent samples. Furthermore, the retrospective nature of the self-reported delinquency measure makes it impossible to determine any causal mechanisms relative to present narcissism or feelings of invulnerability. As noted above, the intervening variables that may explain the relations between narcissism, invulnerability, and behavioral problems should be explored in future research. Contextual variables such as parenting and peer delinquency (see Steinberg, 2007) may offer additional insight as to conditions under which narcissism and invulnerability may be particularly related to aggression and delinquency. In reality, some degree of invulnerability and narcissism may be normative through adolescence and early adulthood, yet heightened danger invulnerability or a tendency to exploit others may be risk factors to which attention is needed. Therefore, awareness of these perceptions, their development, and their behavioral manifestations could help inform efforts to prevent or curtail problematic behaviors among adolescents. References Aalsma, M. C., Lapsley, D. K., & Flannery, D. (2006). Personal fables, narcissism, and adolescent adjustment. Psychology in the Schools, 43, 481–491. Barry, C. T., Frick, P. J., Adler, K. K., & Grafeman, S. J. (2007). The predictive utility of narcissism among children and adolescents: Evidence for a distinction between adaptive and maladaptive narcissism. Journal of Child and Family Studies, 16, 508–521. Barry, C. T., Frick, P. J., & Killian, A. L. (2003). The relation of narcissism and selfesteem to conduct problems in children: A preliminary investigation. Journal of Clinical Child and Adolescent Psychology, 32, 139–152. Barry, C. T., Grafeman, S. J., Adler, K. K., & Pickard, J. D. (2007). The relations among narcissism, self-esteem, and delinquency in a sample of at-risk adolescents. Journal of Adolescence, 30, 933–942. Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcissism, selfesteem and direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of Personality and Social Psychology, 75, 219–229. Bushman, B. J., Baumeister, R. F., Thomaes, S., Ryu, E., Begeer, S., & West, S. G. (2009). Looking again, and harder, for a link between low self-esteem and aggression. Journal of Personality, 77, 427–446. Duggan, P.M., Lapsley, D.K. & Norman, K. (2000). Adolescent invulnerability and personal uniqueness: Scale development and initial construct validation. In Paper presented at the biennial meeting of the Society for Research on Adolescence, Chicago. Elkind, D. (1967). Egocentrism in adolescence. Child Development, 38, 1025– 1034. Elliott, D. S., & Ageton, S. S. (1980). Reconciling race and class differences in selfreported and official estimates of delinquency. American Sociological Review, 45, 95–110. Emmons, R. A. (1984). Factor analysis and construct validation of the narcissistic personality inventory. Journal of Personality Assessment, 48, 291–300. Goldberg, J., Halpern-Felsher, B., & Millstein, S. (2002). Beyond invulnerability: The importance of benefits in adolescents’ decision to drink alcohol. Health Psychology, 21, 477–484. Greene, K., Krcmar, M., Walters, L., Rubin, D., & Hale, J. (2000). Targeting adolescent risk-taking behaviors: The contribution of egocentrism and sensation-seeking. Journal of Adolescence, 23, 439–461. Hauser, R. M., & Featherman, D. L. (1977). The process of stratification. New York: Academic Press. Kohut, H. (1971). The analysis of the self. New York: International Universities Press. Lapsley, D. K., & Aalsma, M. C. (2006). An empirical typology of narcissism and mental health in late adolescence. Journal of Adolescence, 29, 53–71. Marsee, M.A., Barry, C.T., Frick, P.J., Kimonis, E.R., Munoz, L.C., & Aucoin, K.J. (2007). Factor structure of the Peer Conflict Scale in at-risk youth. In Paper presented at the 2007 Southeastern Psychological Association Conference in New Orleans, LA. Millstein, S. G., & Halpern-Felsher, B. L. (2002a). Judgments about risk and perceived invulnerability in adolescents and young adults. Journal of Research on Adolescence, 12, 399–422. Millstein, S. G., & Halpern-Felsher, B. L. (2002b). Perceptions of risk and vulnerability. Journal of Adolescent Health, 31, 10–27.
582
C.T. Barry et al. / Personality and Individual Differences 47 (2009) 577–582
Moffitt, T. E. (2006). Life-course-persistent versus adolescence-limited antisocial behavior (2nd ed.. In D. Cicchetti & D. J. Cohen (Eds.). Developmental psychopathology (Vol. 3, pp. 570–598). Hoboken, NJ: John Wiley & Sons, Inc. Raskin, R., Novacek, J., & Hogan, R. (1991). Narcissism, self-esteem, and defensive self-enhancement. Journal of Personality, 59, 19–38. Raskin, R. N., & Terry, H. (1988). A principal-components analysis of the narcissistic personality inventory and further evidence of its construct validity. Journal of Personality and Social Psychology, 54, 890–902. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Steinberg, L. (2007). Risk taking in adolescence. New perspectives from brain and behavioral science. Current Directions in Psychological Science, 16, 55–59.
Thomaes, S., Stegge, H., Bushman, B. J., & Olthof, T. (2008). Trumping shame by blasts of noise: Narcissism, self-esteem, shame, and aggression in young adolescents. Child Development, 79, 1792–1801. Twenge, J., & Foster, J. (2008). Mapping the scale of the narcissism epidemic: Increases in narcissism 2002–2007 within ethnic groups. Journal of Research in Personality, 42, 1619–1622. Waddell, M. (2006). Narcissism – an adolescent disorder? Journal of Child Psychotherapy, 32, 21–34. Washburn, J. J., McMahon, S. D., King, C. A., Reinecke, M. A., & Silver, C. (2004). Narcissistic features in young adolescents: Relations to aggression and internalizing symptoms. Journal of Youth and Adolescence, 33, 247–260.