Feedback Seeking in Early Adolescence: Self-Enhancement or Self-Verification?

Abstract

The authors examined whether early adolescents (N = 90) solicit self-enhancing feedback (i.e., positive feedback) or self-verifying feedback (i.e., feedback congruent with self-views, even when these views are negative). Sixth, seventh, and eighth graders first completed a self-perception measure and then selected whether to receive positive or negative feedback from an unknown peer in different domains of self. Results were consistent with self-verification theory; adolescents who perceived themselves as having both strengths and weaknesses were more likely to seek negative feedback regarding a self-perceived weakness compared to a self-perceived strength. The authors found similar support for self-verification processes when they considered the entire sample regardless of perceived strengths and weaknesses; hierarchical linear modeling (HLM) examined the predictive power of ratings of self-perceived ability, certainty, and importance on feedback seeking for all participants and provided additional evidence of self-verification strivings in adolescence.

Feedback from peers is an important aspect of establishing a stable identity during early adolescence (Finkenauer, Engels, Meeus, & Oosterwegel, 2002). People prefer certain types of feedback and actively solicit evaluations from others. Two theories, self-enhancement and self-verification, make different predictions about the type of feedback adolescents seek (Kwang & Swann, in press; Leary, 2007; Swann, 1990; Swann, Rentfrow, & Guinn, 2003). According to self-enhancement theory, people are motivated to increase the positivity of their self-views and consequently prefer favorable feedback. Alternatively, self-verification theory posits that people are motivated to maintain stable self-views and seek feedback that is congruent with those views, even when they are negative (Swann, 1983, 1987; Swann, Pelham, & Krull, 1989).

Developmental investigations of self-enhancement and self-verification may help explain the emergence and maintenance of certain types of adjustment problems such as depression and eating disturbances during adolescence (Joiner, 1999; Swann, in press; Swann, Wenzlaff, Krull, & Pelham, 1992). Self-enhancement may be positively associated with mental health, although the benefits depend on situational factors (Kwan, Kuang, & Zhao, 2008; Taylor, Lerner, Sherman, Sage, & McDowell, 2003). Preference for self-enhancing feedback has been found to predict adjustment (Sedikides, Gregg, & Hart, 2007). Alternatively, seeking negative self-view verification may result in increased feelings of depression (Joiner, 1995; Swann, Wenzlaff, & Tafarodi, 1992) or body dissatisfaction (Joiner, 1999). People with negative self-views may surround themselves with others who view them in an unfavorable light, which in turn creates a cycle of persistent low self-esteem and depression (North & Swann, 2008).

Except for work by Cassidy and her colleagues, self-enhancement and self-verification during adolescence have not received significant research attention (Cassidy, Aikins, & Chernoff, 2003; Cassidy, Ziv, Mehta, & Feeney, 2003). Cassidy et al. studied a variety of ages and found both self-enhancement and self-verification strivings, which led them to call for future research to better understand how these processes operate. Given that adolescence is a period in which self-esteem tends to decrease (Robins, Trzesniewski, Tracy, Gosling, & Potter, 2002; Shapka & Keating, 2005; Twenge & Campbell, 2001), and because identity formation begins in early adolescence (e.g., role examination; Finkenauer et al., 2002), examining self-motives in this age group is important and warrants additional attention. This article first reviews the literature on feedback seeking in adulthood and then discusses the extant work examining this process in adolescence.

Feedback Seeking in Adulthood

Swann et al. (1989) examined the interplay of self-enhancement and self-verification processes by having participants indicate the type of feedback they wanted to receive. Participants examined a list with questions that tapped five dimensions of self (intelligence and physical appearance as well as artistic, social, and athletic abilities). For each domain, some questions were intended to elicit positive feedback (e.g., “What is this person’s greatest intellectual strength?”), whereas others were intended to elicit negative feedback (e.g., “What about this person makes you think she would have problems in academia?”; Swann et al., 1989, p. 785). When possible, participants chose to receive feedback in areas that they considered to be their strengths. This finding demonstrated both self-enhancement and self-verification because they wanted positive feedback for areas in which they already believed themselves to be skilled. When forced to select feedback regarding both self-perceived weaknesses and strengths, however, the way in which participants viewed their abilities in each area was critical: Regardless of whether they had high or low global self-worth, participants tended to ask for negative feedback in domains that were self-perceived weaknesses and positive feedback in domains that were self-perceived strengths. These findings were consistent with self-verification theory and counter to self-enhancement theory.

Certain factors may influence whether self-enhancement or self-verification is the more likely outcome. People are more likely to verify self-views for which they have a great deal of psychological investment as indexed by certainty and importance but are more likely to seek enhancing appraisals for domains in which they lack psychological investment. That is, self-verification is the more likely process when adults are certain of their self-views and consider those self-views to be important (Pelham, 1991; Swann, in press; Swann & Ely, 1984; Swann & Pelham, 2002).

The Development of Feedback Seeking

Swann and colleagues (e.g., Swann, 1990; Swann, Hixon, Stein-Seroussi, & Gilbert, 1990) propose that self-verification is more complex and involves more sophisticated cognitive operations than self-enhancement. Although both self-verification and self-enhancement require that feedback be categorized as positive or negative, self-verification also requires comparing that social feedback to self-views. Thus, Swann and colleagues hypothesize that self-enhancement precedes self-verification because the skills necessary for self-verification are not seen until later in development. For example, children first engage in social comparisons in middle childhood. As they develop, their self-concept becomes increasingly dependent on how they relate to their peers as well as how their behavior relates to that of their peers (Harter, 1999; Ruble, Boggiano, Feldman, & Loebl, 1980). Children at this age also begin to describe themselves in terms of traits by generalizing across specific behaviors (Harter, 1999). Both of these developmental shifts in self-concept may facilitate the evaluation of one’s traits in light of social feedback. Furthermore, with the transition to adolescence, perspective-taking abilities become more advanced, and individuals begin to consider the self and other from the viewpoint of a third, impartial party, in what Selman referred to as the third-party or mutual role-taking stage (Selman, 1976; Selman & Byrne, 1974). The manner in which feedback is viewed might change as perspective-taking abilities develop such that once individuals reach the third-party stage they may recognize that peers can have opinions of them that would differ from that of an impartial observer, and thus, possibly be more likely to reject this feedback.

However, the extant developmental research on feedback seeking does not support Swann’s supposition that self-verification should emerge later in development as a result of being the more complex process. Using a hypothetical peer selection paradigm, Cassidy, Aikins et al. (2003) presented seventh graders with positive and negative evaluations in a particular domain (e.g., intellectual ability). The majority of participants chose to interact with the peer who viewed them positively, consistent with self-enhancement. However, contingency table analyses also suggested some limited support for self-verification. Whereas all participants with positive self-views in that domain selected the positive evaluator, only 75% of those with negative views in that domain chose the positive evaluator. Stronger evidence of self-verification was found in a sample of third graders assessed using a developmentally appropriate puppet selection method; as was the case with seventh graders, all third graders with positive self-views in a particular domain selected the positive evaluator for that domain. However, the majority who viewed themselves negatively in a particular domain (61%) chose the peer who also viewed them negatively in that domain, consistent with self-verification theory.

Cassidy, Ziv et al. (2003) had seventh graders examine a list of three positive questions and three negative questions for each domain of self and select the questions for which they most wanted feedback. On average, participants desired positive feedback in areas that they perceived to be their strengths and negative feedback in areas that they perceived as their weaknesses, as predicted by self-verification theory.

Current Research

The current study builds on the work of Cassidy and colleagues (Cassidy, Aikins et al., 2003; Cassidy, Ziv et al., 2003) to examine self-enhancement and self-verification motives in sixth through eighth graders. Our investigation was guided by two research questions:

Research Question 1: Are early adolescents more likely to seek self-enhancing or self-verifying feedback?

Research Question 2: Does psychological investment (i.e., ratings of self-view certainty and importance of self-views) influence the feedback-seeking process?

Research with adults suggests that psychological investment is associated with greater self-verification strivings. To our knowledge, this investigation is the first to examine psychological investment in the feedback-seeking process during early adolescence.

To examine feedback seeking, we first asked participants to evaluate their abilities along different dimensions of self. Participants also rated how certain they were of their abilities within each domain as well as their importance. Finally, participants selected the type of feedback they wished to receive from an unknown peer. In line with self-verification theory, we hypothesized that adolescents would prefer feedback that was consistent with their self-views, even if that feedback was negative. Furthermore, we hypothesized that self-verification strivings would be most pronounced in the areas in which adolescents were psychologically invested (i.e., higher certainty and importance). We did not expect gender differences, but did expect age differences such that self-verification strivings would be strongest in eighth grade.

Method

Participants

Ninety early adolescents participated (51 female; X̄ = 12.9 years, SD = 1.7 years). The data from two additional adolescents were excluded from the analyses because they failed to complete the questionnaires. Parents identified the ethnicity of adolescent participants as Caucasian (n = 54), Hispanic (n = 5), Asian American (n = 3), African American (n = 2), Native American (n = 1), and Other (n = 10). The parents of 15 adolescents did not indicate ethnicity. We recruited the majority of participants (n = 85) from private schools in an urban area of the Southwest. The remaining participants (n = 5) were recruited from a local chapter of the Boys and Girls Club.

Measures

Adolescent Self-Attribute Questionnaire (ASAQ)

We modified the short version of the Self-Attribute Questionnaire (Pelham & Swann, 1989) for use with adolescents. The ASAQ assesses self-perceived intellectual/academic ability, social competence, artistic/musical ability, athletic competence, and physical attractiveness as well as the degree of psychological investment in self-views as indexed by ratings of certainty and importance. In each of the five domains of self, participants used a Likert-type scale to do the following: assess their ability in comparison to other adolescents ranging from 1 (a lot below average) to 5 (a lot above average), indicate how certain they were of their ability from 1 (very uncertain) to 5 (very certain), and rate how important each of these abilities was to them from 1 (very unimportant) to 5 (very important). Specifically, participants were instructed to “rate how you compare to other teenagers your age,” “rate how certain you are about your abilities,” and “rate how important each of these abilities is to you” for intellectual/academic ability, social competence, artistic/musical ability, athletic competence, and physical attractiveness.

Although we developed the ASAQ for the current study, the Self-Attribute Questionnaire upon which it is based has been used to assess feedback seeking in adults (Swann et al., 1989). To assess the validity of our revised measure, we created a composite self-view score by summing participants’ self-ratings across the different domains of self (higher scores indicated that participants viewed themselves more positively) and compared this score with the well-established Self-Perception Profile for Adolescents (Harter, 1988). The composite score of self-perceived ability as assessed by the ASAQ was positively correlated with global self-worth as assessed by the Self-Perception Profile (r = .45, p < .01).

Adolescent Feedback-Seeking Questionnaire (AFSQ)

We revised the Feedback-Seeking Questionnaire (Swann et al., 1989) to include language and examples appropriate for use with adolescent participants. Like the ASAQ, the AFSQ was developed for the current study but based on a measure frequently used to assess self-verification strivings in adults (i.e., the Feedback-Seeking Questionnaire; Swann et al., 1989). The AFSQ includes six questions that correspond to each dimension of the ASAQ (i.e., intellectual/academic ability, social competence, artistic/musical ability, athletic competence, and physical attractiveness). For each dimension, half of the questions elicit positive feedback (e.g., “What is my greatest athletic talent?”) and the other half negative feedback (e.g., “What is my greatest athletic weakness?”). Sample items on the AFSQ include “What about me makes you think I would be comfortable interacting with other people?,” “What academic subjects (like Math, Reading, and Science) would you expect to be difficult for me? Why?,” and “What is my greatest artistic or musical talent?” Participants selected two questions from each area for which they would most like to receive peer feedback. For each domain of self, participants received a score of 0, 1, or 2 based on the number of positive questions they selected for the area.

Filler items

To make the feedback task more realistic and interesting to the participants, we included questions about their interests and background characteristics (Cassidy, Ziv et al., 2003).

Procedure

We tested participants at their school or an after-school activity setting. A research assistant introduced the study as a project examining middle school life. The research assistant told students that, as part of the research project, some participants would have the opportunity to interact with students from a different school at a later date.

After introducing the study, the research assistant distributed a packet of questionnaires that included the ASAQ, AFSQ, and filler items. The research assistant instructed participants to complete the questionnaires while walking around the room to answer questions.

Data Analyses

We conducted two sets of analyses to examine whether adolescents sought self-enhancing or self-verifying feedback. Our first set of analyses were limited to participants who had both a clearly identified strength and weakness, and our second set of analyses included the entire sample.

Following Swann et al. (1989), we first compared the type of feedback adolescents selected for self-perceived strengths and weaknesses. We considered participants as having a self-perceived strength if they assigned themselves a 4 or 5 on one of the dimensions of the ASAQ and a self-perceived weakness if they assigned themselves a rating of 1 or 2 on the ASAQ. We limited this particular analysis to participants with a discrepancy in self-ratings of at least three points (e.g., 1-5, 1-4, 2-5) between their strength and weakness (n = 30); however, our hierarchical liner modeling (HLM) analyses, discussed below, included all participants in the sample (N = 90). Because some adolescents view themselves very favorably across all domains and even view their “worst” characteristic in positive terms, we followed Swann et al. and separately examined the data of those who perceived themselves as having a clear strength and a clear weakness. Swann and colleagues believed examining those with extremely positive and negative self-views would provide a clearer test of feedback seeking (Cassidy, Ziv, et al., 2003; Swann et al., 1989). If adolescents had multiple strengths or weaknesses, we used the areas with most disparate ratings. In the event that the ratings in various domains were identical, we selected the areas randomly. Because the data were not normally distributed, we used a Wilcoxon signed-rank test to examine whether adolescents were more likely to seek negative feedback for a self-perceived weakness than for a self-perceived strength across five domains: athletic ability, social ability, artistic/musical ability, intelligence, and physical attractiveness.

For each domain, participants evaluated not only their ability but also the certainty they felt about that ability as well as the amount of importance they placed upon the ability. Our second major analysis was to use a HLM to assess whether participant characteristics such as sex and grade level as well as domain characteristics such as ability, certainty, and importance predicted feedback-seeking behavior. This HLM analysis is analogous to a repeated-measures ANOVA (analysis of variance; for example, each participant has multiple dependent variables, or in the parlance of multilevel designs, domain characteristics [Level 1] are nested within participants [Level 2]); however, an HLM design is advantageous because it allows us to account for variability between participants for a more powerful test of the relationships between the independent and dependent variables of interest. In addition, we employed an ordinal HLM because the dependent measure (i.e., the number of positive questions that participants selected on the AFSQ [0, 1, or 2]) had a clear, ordered structure. For the sake of brevity, we report only the final significant model using the notation established by Raudenbush and Bryk (2002). All multilevel modeling was conducted on HLM 6.0.

Results

Table 1 presents descriptive statistics and correlations.

Table 1.

Correlations and Descriptive Statistics

	Mean (SD)	1	2	3	4
1. Ability	18.03 (2.59)	—	.66**	.41**	.25*
2. Certainty	18.33 (3.65)		—	.34**	.29**
3. Importance	19.49 (2.76)			—	.20
4. Positive feedback	7.15 (1.92)				—

^*p < .05.

^**p < .01.

Feedback Seeking: Self-Perceived Strengths and Weakness

A Wilcoxon signed-rank test found that adolescents were more likely to seek negative feedback for a self-perceived weakness than for a self-perceived strength, Z = −2.78, p < .01. Because the test for the converse (i.e., adolescents seek more positive feedback for a self-perceived strength) is statistically redundant, we did not execute the analysis but note that the significance of the first implies significance of the second.

Feedback Seeking: All Domains of Self

An ordinal HLM that examined whether ratings of self-perceived ability, certainty, and importance predicted feedback seeking found a significant effect of self-perceived ability on feedback seeking such that participants sought positive feedback on domains that they judged themselves as having high ability, β = −.55, t(417) = −4.23, p < .001 (see Equation 1).

$$\text{Formula: feedback seeking}={\gamma }_{00}+{\beta }_1\left(\text{Ability}\right)+{\mu }_0+\delta$$ (1)

Additionally, these results suggest that adolescents with high self-ratings are more likely to select a greater number of positive questions than adolescents with low self-ratings. Unsurprisingly, there was a significant variability across participants (τ = 1.03, χ² = 183.28, p < .001); nevertheless, participants did not vary significantly with regard to perceived ability and feedback seeking. There were no additional significant Level 1 effects of certainty or importance ratings, nor were there interactions of these variables with self-perceived ability; like-wise, there were no significant Level 2 effects of gender or grade (all p’s > .30).

Discussion

The current study examined whether adolescents were more likely to seek self-enhancing or self-verifying feedback. Consistent with past adult research (Swann et al., 1989), early adolescents desired self-verifying feedback. Adolescents sought more negative feedback regarding a self-perceived weakness than a self-perceived strength. As a stronger test of self-verification theory, we examined the entire sample and found a positive association between overall self-views and feedback seeking; participants with positive self-views were more likely to seek positive feedback from an unknown peer than those with negative self-views. Contrary to our predictions, feedback seeking did not differ as a function of investment in self-views (i.e., the extent to which adolescents were certain of their self-views and considered them to be important). Although we expected self-verification would be more likely with age, there were no significant differences based on grade.

The findings of the current study support the theory that self-verification motives operate across early adolescence. Cassidy et al. found mixed support for self-verification in seventh graders. When presented with positive and negative evaluations in a particular domain, the majority of participants chose to interact with the positive evaluator regardless of their domain-specific self-views. This finding suggests self-enhancement theory (Cassidy, Aikins et al., 2003); however, adolescents might simply prefer positive peers because they believe them to be more pleasant interaction partners. When presented with a feedback-seeking task similar to the current study, adolescents behaved in line with self-verification theory and preferred positive feedback in areas of perceived strength (Cassidy, Ziv et al., 2003). The current study provides additional support for the operation of self-verification motives throughout early adolescence.

All results must be interpreted considering the study’s methodological limitations. The sample was relatively small; however, significant effects still emerged, suggesting that our effect size was large. Unfortunately, no adequate or agreed upon measure of effect size exists for HLM analyses. Furthermore, we were unable to assess feedback seeking in the context of interpersonal interactions. In addition to using questionnaires, previous research with adults has examined how self-enhancement and self-verification motives guide interactions with both known and unknown partners (Swann, De La Ronde, & Hixon, 1994; Swann & Ely, 1984).

Despite these limitations, this study has many strengths. This study is one of the first to examine self-enhancement and self-verification in early adolescence. We also assessed the degree to which participants were invested in their self-views. Psychological investment is a concept that has not been examined in any studies of feedback seeking in developing populations. The current study adds to past research by examining how perceived certainty and importance influence the feedback-seeking process. The correlational analyses presented in Table 1 suggest that the summative certainty score was positively associated with the amount of positive feedback sought; that is, those who were certain of their self-views across the domains of self were more likely to select positive questions throughout AFSQ. However, no significant effects of certainty and importance emerged in the HLM analyses. The interactions between certainty and self-perceived ability and importance and self-perceived ability were also not significant, which was counter to what we expected based on previous findings with adults, who are more likely to verify self-views in which they have a great deal of psychological investment (Swann, in press; Swann & Ely, 1984). Certainty and importance might have a more limited effect on the feedback-seeking process during adolescence because teenagers are still working toward establishing their own identity (Finkenauer et al., 2002). Thus, self-views may be particularly salient for this group and, in turn, render self-verification particularly likely. By late adolescence or adulthood, people may become more certain of their self-views and better able to weigh the importance of different domains of self after they have explored different alternatives and made identity commitments. Adolescents may be becoming aware that they are less certain of their abilities in some areas whereas other areas are less important for their self-conceptions. Until this point, psychological investment may not influence feedback seeking; however, future research is needed to examine certainty and importance as moderators of the feedback-seeking process across adolescence and emerging adulthood.

Examining self-enhancement and self-verification across adolescence has important implications for health and adjustment. Adolescence is a period during which certain psychopathologies such as depression and eating disorders commonly emerge (Angold, Costello, & Worthman, 1998; Attie & Brooks-Gunn, 1989; Ma, Lee, & Stafford, 2005). Self-enhancement is associated with positive mental health (Sedikides et al., 2007), whereas seeking verification for negative self-views may partially explain the emergence and maintenance of these forms of psychopathology (Joiner, 1995, 1999). Our results suggest that adolescents self-verify, and seeking verification for negative self-views may contribute to depression and eating disturbances.

The current study built on past research to examine self-enhancement and self-verification strivings in early adolescence. Although these processes are known to operate in adulthood, limited developmental research has examined feedback seeking. Still, additional research is needed to better understand the developmental interplay of self-enhancement and self-verification. Longitudinal investigations may be especially informative to determine when in development self-verification first becomes evident. Swann and colleagues (1990) hypothesized that self-enhancement precedes self-verification, but the current results suggest that self-verification strivings already guide feedback seeking during early adolescence. More research should also directly examine how self-enhancement and self-verification influence adjustment during adolescence and specifically how individual differences in these processes relate to well-being. Such work may hold important implications for the treatment of adolescent depression and eating disorders.

Bios

Lisa H. Rosen is an assistant professor of psychology at Texas Woman’s University.

Her research focuses on self-esteem, identity, and peer relations.

Connor P. Principe is an assistant professor of psychology at Pacific University. His current areas of interest include the affective antecedents and consequences of experience with faces, behavioral outcomes of differential treatment based on attractiveness stereotypes, and the processing of uncanny facial stimuli in adults and children.

Judith H. Langlois is the Charles and Sarah Seay Regents’ professor of developmental psychology at the University of Texas at Austin. Her research interests include the effects of individual characteristics (physical appearance, gender, age) on the differential socialization of men and women and on the development of social behavior, the origin of social stereotypes, and the development of social competence.