Motivated Expectations of Positive Feedback in Social Interactions

ABSTRACT

People self-enhance in a variety of ways. For example, they generally expect to perform better than others, to be in control of events, and to have a brighter future. Might they also self-enhance by expecting to receive positive feedback in social interactions? Across five studies, we found that they did. People's desire for feedback correlated with how positive they expected it to be (Study 1), and their feedback expectations were more positive for themselves than for others (Study 2). People's positive feedback expectations also covaried with trait tendencies to self-enhance (i.e., self-esteem and narcissism; Study 3) and with a direct situational manipulation of self-enhancement motivation (Study 4). Finally, people expected to receive positive feedback but did not consistently expect to receive self-verifying feedback (Study 5). These findings are consistent with social expectations being driven in part by the self-enhancement motive.

Keywords:

• feedback
• self-enhancement
• self-motives
• social interaction

Acknowledgments

This research was supported by Economic and Social Research Council grant RES-000-23-0331. The authors would like to thank Marilyn Mbuthia and Vivien Ridley for their help with data collection.

Notes

1. In past research on seeking self-verifying feedback, participants are often asked to choose between feedback that confirms or actively disconfirms current self-views (Swann et al., 2003). Such a distinction differs from that used in Study 5, in which participants reported their expectations for feedback that confirms or fails to confirm a current self-view. Nevertheless, self-verification theory predicts that people are more motivated to obtain feedback that actively confirms a current self-view over either alternative (i.e., feedback that either disconfirms or neglects to confirm a current self-view). In line with this, several studies have shown that individuals choose to self-verify when, as here, they are given an alternative that neither confirms nor disconfirms self-views (e.g., taking part in a different experiment; Swann, Wenzlaff, & Tafarodi, 1992; making identity-irrelevant physical appearance choices; Swann et al., 2003). It would be valuable in future research to assess expectations for feedback that actively disconfirms a self-view (e.g., a friend provides positive feedback on a haircut that you personally do not like). Such a distinction would also allow for an equitable comparison of the relative strengths of the self-enhancement and self-verification motives in guiding feedback expectations.

2. The use of parametric statistics is not recommended when analyzing this type of ranked data, given that (a) the different ranks were not independent (e.g., having ranked one type of feedback as 1, participants had to rank all others differently), and (b) respondents were not obliged to assign a rank to all four types of feedback. To supplement the illustration provided by Figure 1, however, we conducted an ANOVA to examine the effects of feedback valence and focus at each rank level. A significant main effect of valence, indicating that positive feedback was more often assigned than negative, was found at rank 1, F(1, 56) = 33.59, p < .0005, and at rank 2, F(1, 56) = 8.01, p < .01, but not at rank 3, F(1, 56) = 2.65, p = .11. In addition, a significant main effect of valence, indicating that negative feedback was more often assigned than positive, was found at rank 4, F(1, 56) = 8.77, p < .01. In contrast, only at rank 2 did the main effect of focus approach marginal significance, F(1, 56) = 2.84, p = .10. These results, though they should be interpreted with caution, bear out the interpretation suggested by inspecting Figure 1. It is worth noting that the log-linear sampling distribution for the formally correct statistic is very difficult to estimate accurately, and that statistical simulations find log-linear analysis to provide only a modest increase in accuracy that rarely leads to altered statistical decisions (DeCarlo, Laczniak, Azevedo, & Ramaswami, 2000). Even if readers prefer not to view this analysis as demonstrative, they may still wish to regard it as illustrative.