Clever and crude but not kind: Narcissism, self-esteem, and the self-reference effect

Abstract

According to the agency model of narcissism (Campbell, Brunell, & Finkel, 2006) narcissists view themselves as high on agentic traits but low on communal traits. To test if this self-view extends to recall, two experiments examined the extent to which narcissism was associated with self-ratings and recall of agentic and communal traits that varied in valence. Across both experiments a trait description task was followed by a surprise recall task for the trait words and then completion of the Narcissistic Personality Inventory (NPI; Raskin & Terry, 1988). Within the self-reference condition narcissism was related to higher selection in the trait description task and to higher recall of positive-agentic (e.g., clever) traits. This general pattern of results occurred for narcissism even while controlling for the related personality variables of self-esteem, agency, and communion. In contrast to narcissism, within the self-referent group self-esteem predicted higher recall for positive-communal traits (e.g., kind) but lower recall for negative-communal traits, a finding consistent with mnemic neglect. Overall, results supported the agency model of narcissism and extended this model to suggest that narcissists rate themselves more highly not only on positive-agentic traits but also on negative-communal traits.

Keywords:

• Self-reference effect
• Narcissism
• Self-esteem
• Mnemic neglect
• Narcissistic Personality Inventory (NPI)

Notes

¹ As a matter of convenience and convention we use the term “narcissists” to refer to individuals with above-average scores on the NPI, and “non-narcissists” to refer to those with below-average NPI scores.

² Independent-samples t-tests confirmed that there were no differences between the Wayne State University and Francis Marion University participants on NPI score, number of correct items recalled, or the number of recalled items within each of the four trait-types (all ps>.55). Hence all subsequent analyses collapsed across university.

³ Notably, the failure of narcissism to predict positive-agentic ratings contrasts with that found in Experiment 1. Therefore we further investigated the relationship between NPI scores and positive-agentic ratings first by visual inspection of a scatterplot. The scatterplot and subsequent regression analyses revealed an emergence of narcissism as a predictor of positive-agentic ratings for NPI scores ≥13 (≥−.67 standard deviations from the mean). For this subset of 88 self-referent group participants with NPI scores ≥ 13, narcissism was a reliable predictor of positive-agentic ratings, β=.21, t=2.24, p<.05. For all other variables the pattern of results for these 88 participants did not differ from the overall self-referent sample.