The robust self-esteem proxy: Impressions of self-esteem inform judgments of personality and social value

Abstract

People use impressions of an evaluative target’s self-esteem to infer their possession of socially desirable traits. But will people still use this self-esteem proxy when trait-relevant diagnostic information is available? We test this possibility in two experiments: participants learn that a target person has low or high self-esteem, and then receive diagnostic information about the target’s academic success or failure and positive or negative affectivity (Study 1), or watch a video of the target’s extraverted or introverted behavior (Study 2). In both experiments, participants’ impressions of the target’s traits accurately tracked diagnostic information, but impressions also revealed an independent self-esteem proxy effect. Evidently, the self-esteem proxy is robust and influences person perception even in the presence of vivid individuating information.

Key words:

• Self-esteem
• trait proxy
• implicit theories
• person perception

Notes

1. To date, only one study investigating implicit theories of self-esteem has included video recordings of targets (Zeigler-Hill et al., 2013b, Study 1). In their novel study, trained judges watched video recordings of participants with the aim of testing whether participants’ self-esteem was associated with judges’ impressions of personality and whether impressions of personality were associated with impressions of self-esteem. Because our goals diverge, we adopted a methodology that allowed us to contrast the robustness of self-esteem information to diagnostic information.

2. In Studies 1 and 2, the total percentage adds up to more than 100 because some individuals identified with multiple ethnic groups.

3. An independent sample of 192 undergraduate students (M_age = 20 years; SD = 4.62; 122 women, 69 men, 1 did not report; 100% Canadian born), read either the LSE or HSE description and rated how much they agreed that Alex had “HSE” using a 9-point scale (1 = strongly disagree; 9 = strongly agree). As intended, a Univariate Analysis of Variance (ANOVA) with condition as the between-subjects factor (Self-esteem condition: LSE vs. HSE) revealed that HSE Alex was perceived to have higher self-esteem (M = 7.17, SD = 1.27) than LSE Alex (M = 2.32, SD = 1.35), F(1, 191) = 657.98, p < .001, d = 3.70.

4. In both experiments, participants also indicated their own self-esteem using Rosenberg (1965) self-esteem scale in Study 1 and the single-item self-esteem scale (Robins, Hendin, & Trzesniewski, 2001) in Study 2. In both experiments, participants’ own self-esteem did not moderate how the other variables influenced their perceptions of the target. Similarly, participants’ own gender did not consistently influence results and thus, was not included in the analyses reported here. Study 1 also included the Ten-Item-Personality Inventory (Gosling, Rentfrow, & Swann, 2003), Personal Belief in a Just World scale and General Belief in a Just World scale (Dalbert, 1999). None of these additional scales moderated the reported effects.

5. Though we initially intended agreeableness to tap the warmth displayed in the extraverted videos, the TIPI agreeableness items did not directly tap the variable of interest. The two items, “sympathetic, warm” and reversed item “critical, quarrelsome” simply do not reflect the type of friendly and expressive behavior exhibited in the videos. Because the TIPI does not assess all facets of the Big Five traits (Gosling et al., 2003) and warmth does not perfectly align with agreeableness (DeYoung, Weisberg, Quilty, & Peterson, 2013), we opted to assign agreeableness as an “irrelevant” trait to the diagnostic videos.

6. An independent sample of 215 undergraduate students (M_age = 22.77 years; SD = 6.85; 116 women, 96 men, 3 did not identify with male or female gender identity; 84% Canadian born), rated the male and female targets as similarly attractive (Ms = 6.10 and 6.32, respectively; SDs = 1.51 and 1.52 respectively; where 1 = not at all attractive, and 9 = extremely attractive).

7. There were three main effects for target gender. Compared to when the target was a man, when the target was a woman, she was rated as less emotionally stable (M = 5.87 vs. 5.40), F(1, 128) = 3.81, p = .053, more agreeable (M = 5.58 vs. 6.07), F(1, 128) = 5.94, p = .016, and more physically attractive (M = 69.57 vs. 75.78), F(1, 128) = 5.45, p = .021, consistent with gender differences in social perception (attractiveness: Cameron et al., 2010; emotional stability and agreeableness: e.g., Koppensteiner & Grammer, 2011) and also consistent with the notion that gender stereotypes continue to influence personality impressions despite the presence of diagnostic information (e.g., Krueger & Rothbart, 1988).