A Test of the Construct Validity of the Five-Factor Narcissism Inventory

Abstract

The Five-Factor Narcissism Inventory (FFNI) is a new self-report measure that was developed to assess traits associated with grandiose and vulnerable narcissism from a Five-factor model (FFM) perspective. In a sample of undergraduates (N = 283), the relations among the FFNI scales, grandiose and vulnerable dimensions, and an array of relevant criteria were examined including self- and informant reports of the Big Five domains, measures of the Dark Triad, ratings of the interpersonal circumplex, externalizing and internalizing behaviors and symptoms, and romantic and attachment styles. The FFNI grandiose and vulnerable dimensions demonstrated good convergent and criterion validity. The FFNI grandiose and vulnerable dimensions manifested converging (e.g., disagreeableness, low love/communion, psychopathy, Machiavellianism, Ludus/Manic love styles) and diverging (e.g., neuroticism, extraversion, dominance, externalizing, internalizing, attachment anxiety) relations in a manner largely consistent with predictions. The FFNI joins the Pathological Narcissism Inventory as a measure that can simultaneously assess both grandiose and vulnerable dimensions of narcissism.

Notes

An exploratory factor analysis was conducted using principal axis factoring with an oblique rotation to test whether the four narcissism scales could be used to form a grandiose composite. All indexes supported the extraction of just one factor (i.e., eigenvalues over 1, Minimum Average Partial [MAP] and parallel analyses). The first factor accounted for 50.62% of the variance.

A reviewer noted that the FFNI scales manifested more significant relations with the self-report Big Five domains than with the informant-report domains. We think this is due to three issues. First, the analyses using the informant reports used a smaller sample and thus had lower power (offset to some degree by using a lower p value). Second, the correlations between the self-report FFNI scales and the self-report Big Five dimensions might be inflated due to the role of common method variance, whereas the correlations using the informant reports could have been attenuated due to the use of two different methodologies. Third, the FFNI scales manifested a tighter pattern of convergent and discriminant validity correlations with the informant-report Big Five domains. For the informant reports, the Big Five domains were primarily correlated only with the FFNI facets derived from that domain. This was not as much the case for the correlations between the FFNI scales and the self-report Big Five domains as the FFNI scales often correlated with their “home” domains as well as other Big Five domains.

The relation between the FFNI grandiose fantasies and the FFM domain of openness requires further examination. Although this FFNI subscale was derived from the openness to fantasy facet, the grandiose fantasies subscale has manifested nonsignificant relations with this facet (Glover et al., 2012) and domain (this study) to date. Glover et al. (2012) suggested it “should perhaps be conceptualized more accurately as a maladaptive variant of low FFM modesty (i.e., as another expression of narcissistic arrogance)” (p. 508).