Self-evaluation motives 1 Running head: SELF-EVALUATION IN JUVENILE OFFENDERS Self-Evaluation in Naturalistic Context: The Case of Juvenile Offenders

The authors investigated how self-evaluation motives (self-enhancement, selfassessment, self-verification, self-improvement—and also self-diminishment and no information) shape self-knowledge preferences in male incarcerated juvenile offenders (IJOs). IJOs responded to questions on how much they would like to receive and actually received each of six types of feedback (positive, truthful, improving, consistent, negative, and no feedback) from each of six sources (teachers, parents, siblings, best friend, girlfriend, and behavioural specialists or psychologists). IJOs disliked negative feedback and the lack of feedback. They preferred truthful feedback to consistent feedback, and received truthful and positive feedback more frequently than improving feedback. Additionally, they received more negative or no feedback from parents than they would like. Finally, IJOs expressed a preference for receiving more improving feedback from their girlfriends than they did. The study highlights the interplay of self-evaluation motives in IJOs and opens up promising research and rehabilitation directions. Self-evaluation motives 3 Self-Evaluation in Naturalistic Context: The Case of Juvenile Offenders A fundamental epistemic concern is the pursuit of self-knowledge. This pursuit is guided by motives, to which we refer as self-evaluation motives. These motives influence not only the kind of information that individuals seek out, but also the way in which they judge, accept or reject, and remember this information. Prior research has identified four self-evaluation motives: self-enhancement, self-assessment, self-verification, and self-improvement (Sedikides & Strube, 1995; Sedikides & Gregg, 2003). The self-enhancement motive refers to strivings for increased positivity of the self-concept and protection of the self from negative feedback. This motive is thought to contribute to a relatively high level of self-esteem. The self-assessment motive refers to strivings for an accurate and objective image of the self. This motive is thought to reduce uncertainty about aspects of the self. The self-verification motive refers to preference for information that is consistent with existing self-conceptions. This motive is assumed to promote a sense of control and predictability. Lastly, the self-improvement motive reflects a genuine desire to improve traits or abilities. The quest for self-improvement is assumed to confer a sense of progress and growth. Substantial evidence exists to document the presence of each of these four motives as well as their functional importance for the self (Sedikides & Strube, 1997; Swann, Rentfrow, & Guinn, 2002; Taylor, Neter, & Wayment, 1995; Trope, 1986). Although an abundance of research exists on the ways in which self-evaluation motives guide self-knowledge preferences, this research has two major weaknesses: it is, almost exclusively, confined to the laboratory and to adult populations. The overarching objective of this investigation is to remedy these weaknesses. In recognition of the renewed emphasis on the study of self-evaluation across a Self-evaluation motives 4 wide range of situations (Bonanno, Field, Kovacevic, & Kaltman, 2002; Fleury, Sedikides, & Donovan, 2002; Tomich & Helgeson, 2004), we examine self-evaluation motives in a naturalistic setting (i.e., prison population). Furthermore, in recognising that little is known about self-evaluation motives among younger populations, we direct our attention to an adolescent population. We believe that motivation does matter in such a population: Whether individuals hold a learning or achievement goal has notable consequences for the positivity of their selfconcept (Dweck, Higgins, & Grant-Pillow, 2003; Dweck & Leggett, 1988; Ruble & Frey, 1991). Thus, research into the self-evaluation motives can provide insights into not only self-knowledge preferences but also consequences of such preferences for the self-system. In summary, we investigated how self-evaluation motives direct self-knowledge preferences, focusing on a particular male adolescent population: incarcerated juvenile offenders (IJOs). Self-Evaluation Motives in Incarcerated Juvenile Offenders Adolescence is a developmental period in which the search for self-knowledge is a primary concern (Roberts & Caspi, 2003). A rich history of research on identity status points to the relevance of exploration and commitment processes pivotal to identity formation during this period (Berzonsky & Adams, 1999; Erikson, 1982). Exploration, in particular, necessitates a search for self-knowledge in order for the individual to gain insight both into one’s self and one’s possible selves (i.e., the various personas and roles that the adolescent is considering for the future). By implication, the four self-evaluation motives should be particularly active during adolescence. Self-evaluation in juvenile delinquents is an especially important issue. Psychologists have long studied the link between self-esteem and delinquency (Emler, 2001; Emler & Hopkins, 1990; Emler & Reicher, 1995). However, findings have been contradictory: some studies report Deleted: Self-evaluation motives 5 relatively low self-esteem among delinquent youth (Mason, 2001; Rosenberg & Rosenberg, 1978), other studies report no relation between self-esteem and delinquency (Oyserman & Markus, 1990; Wells & Rankin, 1983), and still other studies report relatively high self-esteem among delinquent youth (Baumeister, Smart, & Boden, 1996; Jankowski, 1991). In a further variation on this theme, aggression and hostility have been linked to unstable high self-esteem: Individuals with high but frail self-esteem respond aggressively to perceived insult (Kernis, Grannemann, & Barclay, 1989; Kernis & Waschull, 1995). Hence, juvenile delinquents may have high, fragile self-esteem. Recently, Baumeister, Campbell, Krueger, and Vohs (2003) concluded that self-esteem alone provides insufficient insight into delinquent or aggressive behavior. Instead, it is important to consider additional variables. Our investigation approaches self-evaluation among delinquent adolescents from a distinctly different angle. Rather than focusing on the positivity of selfevaluation, we explored IJOs’ perceptions of self-related information. What type of feedback do IJOs receive and what type of feedback would they like to receive? Do IJOs report that they typically receive negative feedback (i.e., criticism) or positive feedback (i.e., praise) from others? Do IJOs wish to hear only positive feedback? More generally, we examined the presence and prevalence of the self-evaluation motives in IJOs. We asked two main questions. First, is each self-evaluation motive active? That is, are IJOs motivated to self-enhance, self-assess, self-verify, and self-improve? Second, what is the relative prevalence of the four self-evaluation motives? That is, are some self-evaluation motives more powerful in guiding the self-knowledge preferences of IJOs than other motives? The lack of relevant research and lack of a non-incarcerated comparison group led us to take an exploratory approach. Nevertheless, we expected that the therapeutic demands of the Youth Self-evaluation motives 6 Development Center environment coupled with the developmental tasks of adolescence would serve to activate multiple self-evaluation motives in IJOs. To consider this possibility, we present a more detailed discussion of the adult literature on the self-evaluation motives and link this literature to IJOs. Also, we articulate additional purposes of our research. Self-Enhancement Adults manifest several biases that serve to enhance the positivity of the self. For example, they spend more time processing positive than negative feedback (Baumeister & Cairns, 1992), remember favourable feedback better than unfavourable feedback (Green & Sedikides, 2004; Sedikides & Green, 2000), and take credit for their successes while denying responsibility for their failures (Campbell & Sedikides, 1999). Sedikides (1993) reported evidence that, when participants are in an information-gathering mode, self-enhancement concerns are more potent than either self-verification or self-assessment concerns. Self-enhancement serves to maintain a relatively high level of self-esteem (Baumeister, Tice & Hutton, 1989; Sedikides & Strube, 1997). Given that the self-enhancement motive is so pervasive in adults, and given the universal human need for high self-regard (Brown & Kobayashi, 2002; Kobayashi & Brown, 2003; Sedikides, Gaertner, & Toguchi, 2003; but see Heine, Lehman, Markus, & Kitayama 1999), we expected that this motive would influence strongly the self-knowledge preferences of IJOs. Self-Verification Research on adults is also supportive of the self-verification motive. Adults solicit selfconfirming rather than self-disconfirming feedback from interaction partners (Swann & Read, 1981a) and recall selectively information that is consistent with their self-concept (Swann & Read, 1981b), while attributing self-disconfirming behaviours to situational factors (Kulik, Self-evaluation motives 7 Sledge, & Mahler, 1986). In addition, young adolescents display a preference for feedback consistent with their self-perceptions (Cassidy, Ziv, Mehta, & Feeney, 2003), which is evidence for the presence of self-verification motives in younger populations. As stated previously, the primary function of self-verification is to afford controllability and predictability to the social world. For adolescents in general, developmental transitions (e.g., physical changes, joining other schools, shifting allegiances among friendship networks) and looming uncertainties about the future may activate the self-verification motive. The incarcerated youth, in particular, face additional threats to their sense of control. Their behavior within the Youth Development Center is strongly regulated by others. In our study, over half felt that they had little or no control over entering the Youth Development Center. Hence, IJOs may be especially motivated to seek information that is consistent with their self-concept in order to reintroduce a feeling of control in their lives. Self-Assessment The self-assessment motive is most typically observed in preferences for high (as opposed to low) diagnosticity achievement tasks (Strube & Roemmele, 1985; Trope, 1979, 1982). High diagnosticity tasks (e.g., standardised tests) provide individuals with accurate feedback regarding their relative standing on personality characteristics (i.e., skills, abilities, or attributes). Preference for such tasks indicates the desire for uncertainty reduction. Indeed, the self-assessment motive is activated in the face of uncertainty about one’s personality characteristics. Harter (1986, 2002) found that adolescents were more likely to report that they perceived characteristics within the self as opposites than were children. For example, an adolescent may report being shy in one situation, outgoing in another, and that these two “selves” are conflicting. Self-evaluation motives 8 Such perceptions were accompanied by feelings of confusion. Thus, the self-assessment motive may also be prevalent during adolescence, given that the motive follows from the desire to reduce feelings of confusion. For IJOs, self-assessment may be even more pressing. In reference to our study, the treatment objectives in the Youth Development Center were to encourage IJOs to acknowledge their crime, take responsibility for their actions, and understand the suffering that they caused to their victims. These goals presuppose an accurate perception of the self as an instigator of behaviour. Self-Improvement The self-improvement motive is distinct from the other motives in its orientation toward the future (Dweck, 1999; Pemberton & Sedikides, 2001; Sedikides, 1999). Individuals choose tasks (Taylor et al., 1995) or social comparison targets (Collins, 1996) with an eye toward personal improvement. For example, cancer patients make upward comparisons when selecting interaction partners among other cancer patients, a choice that is assumed to reflect an attempt to acquire more effective coping strategies (Van der Zee, Oldersma, Buunk, & Bos, 1998). The improvement motive is likely to be potent among incarcerated youth. The treatment objectives in the Youth Development Center were formulated to discourage further offending, and hence, to improve future behaviour. With incarceration serving as evidence of their failure to meet social goals, IJOs may have shifted their attentional orientation, reinvesting their hopes and aspirations in the future. Other Motives The special circumstances of IJOs raised additional issues. The participants in the current study entered treatment with negative backgrounds. The majority had been a witness to or a victim of multiple violent events (Shahinfar, Kupersmidt, & Matza, 2001). Many participants Self-evaluation motives 9 came from families with a history of criminal behaviour. Hence, participants had likely experienced either verbal or physical abuse. A negative social environment may in fact be common among many IJOs. Consistent feedback, for these offenders, may mean selfdiminishing feedback. To explore this possibility we assessed preferences for and receipt of negative feedback. Alternatively, IJOs may have experienced a lack of social feedback. Lack of parental monitoring and supervision is a predictor of delinquent behaviour (Jacobson & Crockett, 2000; Patterson & Stouthamer-Loeber, 1984), raising the possibility that low levels of parental supervision contributed to the delinquency of some of the participants. In the current study, about 70% of participants came from single-parent families, which may face greater difficulties in fully monitoring adolescents’ behaviour. In short, prior to their incarceration, a relatively high proportion of these adolescents may have not received enough guidance and feedback from adult sources. Hence, they may think that the absence of self-relevant information (i.e., the absence of constraints and discipline) is normal. Consequently, IJOs may actually prefer no feedback. By considering preferences for negative feedback as well as no feedback, we extended the scope of the current study beyond the traditional four self-evaluation motives. Source of Feedback Another novel contribution of this study rests in our explicit acknowledgement that the source of feedback matters. Pre-adolescent children report seeking different types of social support from various sources in their social network (Furman & Buhrmester, 1985). Adolescents may attempt to compensate for low support from one source by seeking more support from another (van Aken & Asendorpf, 1997). Given that peers and adults occupy different roles in adolescents’ lives, we might expect both desired and received feedback to vary across sources. Self-evaluation motives 10 This point is particularly relevant in the case of IJOs, who may receive quite different feedback from teachers, parents, siblings, best friends, romantic partners, and psychologists. For example, IJOs may receive self-verifying feedback from best friends, self-enhancing feedback from romantic partners, and self-improving feedback from teachers or psychologists. We also examined the possibility that type of feedback sought varies as a function of feedback source. For example, IJOs may report that positive feedback from peers is more desirable than positive feedback from adult figures. Preferred versus Received Feedback Additionally, we distinguished between preferred and received feedback. Social selfdiscrepancy theory emphasises that the discrepancy between actual and ideal social circumstances is of key importance in considering negative outcomes (Kupersmidt, Buchele, Voegler, & Sedikides, 1996; Kupersmidt, Sigda, Sedikides, & Voegler, 1999). In the current study, we explored discrepancies between actual feedback received on the one hand and preferred (ideal) feedback on the other. The discrepancy between the two is particularly germane in the case of IJOs. For example, they may prefer self-enhancing feedback, but receive instead self-diminishing feedback. Thus, we examined whether such discrepancies differed as a function of source and type of feedback. IJOs have clearly experienced negative outcomes. They are imprisoned for having committed crimes, and the majority report negative life experiences. These contextual factors prompted us to expect significant discrepancies between actual and ideal circumstances in these adolescents’ lives. Due to lack of informative literature, however, we took an exploratory approach regarding differences in discrepancy scores across sources. Self-evaluation motives 11 Overview We asked participants to indicate their preferences for various types of feedback from different people in their lives (e.g., best friends, romantic partners, teachers). We also asked participants about the feedback that they actually received from these differing sources. The questions focussed on role-specific feedback from each source. For example, “My teachers tell me how to get better at my schoolwork” assessed improving feedback from teachers. Of key interest was the presence and relative prevalence of the four major self-evaluation motives (selfenhancement, self-verification, self-assessment, self-improvement) as well as of two additional motives (self-diminishment, no information) in self-relevant feedback that IJOs prefer and receive. We operationalised self-enhancement in terms of positive feedback, self-assessment in terms of truthful feedback, self-improvement in terms of improving feedback, self-verification in terms of consistent feedback, self-diminishment in terms of negative feedback, and no information in terms of no feedback. We investigated the discrepancy between ideal and actual levels of feedback by subtracting participants’ level of preferred feedback from the level of received feedback across the six motives. Method Participants Participants were 110 male juvenile offenders between the ages of 13 and 17, incarcerated in a secure Youth Development Center environment in the southern United States. The current study includes only those adolescents who completed two questionnaires assessing self-evaluation motives. This limitation lowered the sample size to 45. (Details on the full sample are available in Shahinfar et al., 2001.) The considerable drop in sample size was primarily due to participants who completed only one or neither of the questionnaires (n = 43) because of Self-evaluation motives 12 scheduled releases from confinement that occurred over the course of the assessment procedure. The drop in sample size was also due to some participants skipping items within the questionnaires (n = 22). 1 Participants in the current study were on average 15.5 years of age (SD = 1.1). In terms of ethnicity, 71% of the participants were African-Americans and 29% were Whites. Procedure The Youth Development Center included measures of preferred and received feedback in a clinical assessment system whose primary purpose was treatment planning. The Youth Development Center staff and consultants supervised and trained adult assistants to conduct the clinical assessments. The assistants read aloud the questions and response choices to small groups of participants. Each participant marked his answer on his own questionnaire. The assistants administered questionnaires individually to participants who required greater assistance with reading. The assistants also allowed participants to complete the questionnaires individually, if they had missed portions of the group administration due to time constraints or scheduling conflicts. The assessment system was administered over several different sessions. Preferred feedback was assessed in a separate administration session than received feedback. Participants were told that the questionnaires would assist their psychologist in the formulation of their treatment plan. The North Carolina Department of Juvenile Justice and Delinquency Prevention (DJJDP) provided us with access to the data as anonymous archival records. Measures Preferred feedback. We determined participants’ feedback preferences by asking them how much they would like to receive each of six types of feedback from each of six sources. The Self-evaluation motives 13 feedback types included: positive, truthful, improving, consistent, negative, and no feedback. The sources of feedback included teachers, parents, siblings, best friend, girlfriend, and behavioural specialists or psychologists. The specific questions appear in Appendix A. The questions asked for role-specific feedback. Responses ranged from 1 (not at all) to 3 (a lot). To illustrate, the full set of questions for parents are reproduced below with the type of feedback indicated in parentheses. Participants were provided with the question but not the underlying type of feedback. I want my parent or the adults who have helped raised me to tell me ... I am a great son. (Self-enhancing) The truth about how good a son I am. (Self-assessing) How to be a better son. (Self-improving) I am the kind of son I think I am. (Self-verifying) I am a bad son. (Self-diminishing) Nothing about the kind of son I am. (No feedback) In a similar vein, questions regarding feedback from teachers focused on the participants’ schoolwork (e.g., “I would like for my teachers to tell me the truth about how good I am at my schoolwork”). Received feedback. We determined participants’ received feedback through a parallel series of questions comparable to the ideal feedback questions (Appendix B). Participants indicated how frequently they received each of six types of feedback (“My parents or the adults who have helped raised me tell me I am a great son”) from each of six sources. We tailored the questions to reflect role-specific feedback received. Responses ranged from 1 (not at all) to 3 (a lot). Self-evaluation motives 14 Feedback discrepancy. We created difference scores to assess the discrepancy between preferred and received feedback. Specifically, we subtracted participants’ received score from their preferred score. Values greater than zero indicated that participants would prefer more of a particular type of feedback than they received. Values less than zero indicated that participants received more of a particular type of feedback than they would prefer. Results Preliminary Analyses We began by examining ethnic differences in feedback type. Ethnicity did not interact with feedback type or feedback source for preferred feedback, received feedback, or feedback discrepancy. Also, the ethnicity main effect was not significant. Hence, we did not include ethnicity in subsequent analyses. We also examined the relation between age and preferred feedback, received feedback, and feedback discrepancy. Age was negatively related to consistent feedback received (r = -.23, p < .05), indicating that older participants reported receiving consistent feedback less frequently than did younger participants. Age was not related, however, to any other feedback type. We therefore excluded age from subsequent analyses. Main Analyses Preferred feedback. We analysed intraindividual differences in preferred feedback in a 6 (type) x 6 (source) within-participants repeated measures Multivariate Analysis of Variance (MANOVA). The main effect for source of information was marginal, F(5, 40) = 2.07, p < .09. We examined this trend with a post-hoc comparison of the absolute amount of feedback preferred from each source, using a Bonferroni adjustment with an alpha level of .05. Participants showed a significantly higher preference for feedback in general from girlfriends (M Self-evaluation motives 15 = 2.23), siblings (M = 2.23), and parents (M = 2.20) than from teachers (M = 2.16) and best friends (M = 2.03). The main effect for type of feedback was significant, F(5, 40) = 38.64, p < .001, indicating that participants preferred certain types of feedback more frequently than others. Post-hoc analyses revealed that participants showed a significantly greater preference for truthful (M = 2.62), positive (M = 2.57), self-improving (M = 2.45) and consistent (M = 2.42) feedback compared both to no feedback (M = 1.65) and negative feedback (M = 1.32). Additionally, preference for truthful feedback was significantly higher than preference for consistent feedback. Finally, negative feedback was significantly less preferred than no feedback. This main effect, however, was qualified by the interaction between type and source, F(25, 20) = 3.04, p < .01. In order to explore this interaction, we conducted several post-hoc comparisons. Within each type of feedback, or within each source, we used a Bonferroni adjustment for multiple comparisons with an alpha level of .05. Although this is less conservative than other possible adjustments, we chose to balance possible Type I errors with the relatively small sample size. Another factor that influenced our decision was the exploratory nature of our research, given that practically no empirical forays exist into the role of selfevaluation motives in either adolescents or juvenile offenders. We present the means in Table 1. In the top half of the table, we display means for preferred feedback, and in the lower half we provide means for received feedback. Within each row, we use different subscripts to note significant differences between types of feedback within each source. The interaction between type and source of preferred feedback primarily arises from differences in preferences of negative versus no feedback. Notably, this difference was significant only in preferred amounts of feedback from teachers and siblings. Self-evaluation motives 16 Within each feedback type (the columns in Table 1), participants generally did not report significant differences in preferred feedback across sources. That is, frequency of each type of feedback rarely differed by source. The only exception involved participants reporting that they would prefer to hear nothing (no feedback) from their teachers more frequently than from their parents. Received feedback. We analysed intraindividual differences in received feedback in a 6 (type) x 6 (source) within-participants repeated measures MANOVA. The main effect for source of information was not significant, F(5, 40) = .62, p < .69. The main effect for type of feedback was significant, F(5, 40) = 19.43, p < .001. Post-hoc analyses demonstrated that participants received truthful (M = 2.38), positive (M = 2.37), consistent (M = 2.27) and improving (M = 2.27) feedback significantly more often than they received either no feedback (M = 1.62) or negative feedback (M = 1.43). Participants reported receiving both truthful and positive feedback significantly more often than improving feedback. Participants received negative feedback less frequently than no feedback. This main effect, however, was qualified by the type X source interaction, F(25, 20) = 3.19, p < .01. In order to explore this interaction, we conducted several post-hoc comparisons. Again, we used a Bonferroni adjustment for multiple comparisons. In the lower half of Table 1 we present the means for received feedback. Participants reported that they received positive, truthful, improving, and consistent feedback more frequently than negative feedback or no feedback. We observed this pattern for received feedback from teachers, parents, and psychologists. Participants received improving feedback as infrequently as they received negative or no feedback from their best friends and from their romantic partners. In addition, Self-evaluation motives 17 participants received improving feedback more often from teachers and psychologists than from best friends or girlfriends. Feedback discrepancy. We analysed discrepancy scores in a 6 (type) x 6 (source) withinparticipants repeated measures MANOVA. The main effect for type was significant, F(5, 40) = 4.70, p < .01, although the main effect for source was not F(5, 40) = .85, p < .53. Post-hoc examination of the discrepancy scores revealed that participants would like to receive more improving, truthful, positive, and consistent feedback compared to negative feedback, which they would like to receive less often than they do. The source X type interaction was also significant, F(25, 20) = 2.07, p < .05. Analytical comparisons helped to illuminate these results (Table 2). We use different subscripts to denote significant comparisons within each row. Participants reported that, whereas they received negative and no feedback more than they would like from their parents, they would like to receive more truthful and improving feedback than they currently did from them. Participants also reported a greater discrepancy in the amount of improving feedback from their romantic partners as compared to positive feedback, negative feedback, or no feedback. This in part reflects the comparatively low levels of improving feedback that participants actually received from romantic partners. Only one effect within feedback type (columns of Table 2) reached significance. The discrepancy between preferred and received improving feedback from romantic partners was significantly higher than the discrepancy between preferred and received improving feedback from psychologists. Finally, we tested the magnitude of the discrepancy scores. Overall, about one-third of the scores differed significantly from zero (means displayed in bold-face in Table 2). This Self-evaluation motives 18 pattern of results suggests a substantial lack of fit between what participants receive and what they would like to receive. Inspection of the means reveals that participants would especially like to receive more improving feedback from people within their intimate social sphere (i.e., family and peers). Discussion Self-evaluation, the search for self-knowledge, is motivated. The motives underlie the selection of self-relevant information (i.e., feedback preferred) and the interpretation of it (i.e., feedback received). The current investigation was guided by consideration of ways in which the motives play out in a naturalistic context. In particular, the investigation examined possible developmental demands on adolescents in general and environmental demands specific to IJOs. These demands could influence the presence and relative prevalence of self-evaluation motives. Summary of Findings and Implications We assessed six self-evaluation motives. Four of them (self-enhancement, selfverification, self-assessment, and self-improvement) have been examined in the adult literature. We added two motives (self-diminishment and no information) in order to consider possible unique influences in the lives of incarcerated youth. We also considered sources of feedback (e.g., best friend vs. psychologist) as well as discrepancies between preferred and received