Are there such things as “Narcissists” in social psychology? A taxometric analysis of the Narcissistic Personality Inventory

https://doi.org/10.1016/j.paid.2007.04.003Get rights and content

Abstract

Narcissism is typically viewed as a dimensional construct in social psychology. Direct evidence supporting this position is lacking, however, and recent research suggests that clinical measures of narcissism exhibit categorical properties. It is therefore unclear whether social psychological researchers should conceptualize narcissism as a category or continuum. To help remedy this, the latent structure of narcissism—measured by the Narcissistic Personality Inventory (NPI)—was examined using 3895 participants and three taxometric procedures. Results suggest that NPI scores are distributed dimensionally. There is no apparent shift from “normal” to “narcissist” observed across the NPI continuum. This is consistent with the prevailing view of narcissism in social psychology and suggests that narcissism is structured similar to other aspects of general personality. This also suggests a difference in how narcissism is structured in clinical versus social psychology (134 words).

Introduction

Narcissism has an extensive history as both a clinical and social psychological construct (see Campbell and Foster, 2007, Morf and Rhodewalt, 2001). As a clinical construct, narcissism grew to prominence with the writings of Freud (1914/1957). Kernberg, 1974, Kernberg, 1975, Kohut, 1977 continued this clinical tradition, which ultimately led to the inclusion of narcissistic personality disorder (NPD) in the DSM-III (American Psychiatric Association, 1980).

Freud’s (1931) essay on libidinal types and Murray’s (1938) work on “narcism” or egophilia mark classic examples of narcissism conceptualized as a component of normal personality. A recent surge in narcissism research by social psychologists was prompted by the development of the Narcissistic Personality Inventory (NPI; Raskin & Hall, 1979). The development of the NPI was based on the clinical criteria for NPD, with the idea being that non-disordered persons often exhibit cognitive/behavioral patterns similar to those with NPD. The NPI has now been used in well over 100 investigations, has proven to be a reliable and valid assessment of narcissism as it occurs in the general population (del Rosario and White, 2005, Raskin and Terry, 1988), and is the primary measure of narcissism in the social psychological literature.

The clinical and social psychology concepts of narcissism share many similarities, but differ on one important aspect. In clinical psychology, narcissism is specified as a personality disorder. Based upon a set of diagnostic criteria individuals either have NPD or they do not. Put differently, the structure of clinical narcissism is taxonic. In contrast, social psychologists generally view narcissism as a dimension. According to this view, there is no categorical property to the structure of narcissism. There exists no point along the narcissism continuum where one shifts from “normal” to “narcissist.”

The view that NPD, and personality disorders in general, have underlying taxa is controversial in the clinical domain. Many argue that personality disorders are extensions of normal personality continua and thus dimensional (e.g., Livesley et al., 1994, Markon et al., 2005, Widiger and Costa, 1994, Widiger et al., 2005). The taxonic orientation remains prominent, however, and has received empirical support (e.g., Haslam, 2003). Indeed, with specific respect to NPD, the current evidence favors a taxonic view (Fossati et al., 2005). That is, NPD appears to be a qualitatively distinct personality syndrome.

In social psychology, there has been far less debate over the dimensional/taxonic structure of personality traits. The general assumption is that personality is dimensional and the research generally supports this position. There are examples of proposed taxonic personality traits (e.g., self-monitoring, Gangestad & Snyder, 1985; type-A personality, Strube, 1989), but more often than not, personality exhibits dimensional properties (e.g., five-factor model of personality, Arnau et al., 1999, Green et al., 1999). Although the dimensional/taxonic properties of narcissism have not been investigated in the social psychology literature, the assumption seems to be that it too should be conceptualized as a dimension.

In summary, the taxonic view of narcissism has received empirical support in the clinical literature (Fossati et al., 2005). Little is known, however, about the structure of narcissism as it pertains to social psychology. The assumption then perhaps should be that narcissism is categorical, which would fit with the clinical evidence. This assumption is at odds, however, with how narcissism is viewed in social psychology. Indeed, to the extent that narcissism is categorical, changes to method, analysis, and theory may be necessitated. For example, some narcissism correlations may be driven by individuals with elevated NPI scores, but who are nevertheless members of a putative “non-narcissist” group. These findings may say very little about the functioning of “narcissists”. If the putative “narcissist” group is relatively small, it is even possible that some research on narcissism fails to capture any “narcissists” at all. This would be like studying gender differences in a female-only sample.

It is noteworthy that Fossati et al. (2005) investigated narcissism from a clinical perspective. They assessed narcissism using measures that capture NPD (e.g., Structured Clinical Interview for DSM; First, Spitzer, Gibbon, & Williams, 1995) administered to a clinical sample. It is unknown whether their results might generalize to narcissism as it is assessed by the NPI and in individuals selected from the general population, which is more representative of social psychological research. However, considering that (i) the content of the NPI was designed to reflect the diagnostic criteria for NPD (Raskin and Hall, 1979, Raskin and Terry, 1988) and (ii) scores on the NPI are strongly correlated with clinically valid measures of NPD that are administered to clinical samples (Prifitera & Ryan, 1984), there is reason to expect Fossati et al.’s (2005) results to generalize to the NPI. Consequently, there is a real need to formally test the structural properties of the NPI and specifically to determine whether social psychologists should conceptualize narcissism as a category or continuum.

Section snippets

Present study

The present study was conducted to remedy the lack of evidence concerning the taxonic/dimensional nature of narcissism as it is operationalized within the social psychology literature. To this end, the NPI was administered to 3895 participants. Responses were analyzed using three analytic procedures designed to distinguish taxa from dimensions—collectively referred to as taxometrics (Ruscio et al., 2006, Waller and Meehl, 1998).

Participants and procedure

A total of 3895 participants (75% female; 81% Caucasian; M age = 25.0, SD = 8.9; range = 17–70) completed an online version of the NPI (see Foster, Campbell, & Twenge, 2003 for details). All participants reported being American and English speaking and were part of a larger international study on narcissism (n = 5965) that was conducted between the years 2001 and 2003. A portion of these data was reported in Foster et al. (2003). Although it is likely that some of these participants were receiving

Results

We present the results of the taxometric procedures in three sections: evidence stemming from (i) visual inspection of plots, (ii) comparisons between research and simulated data, and (iii) consistency of putative taxon base rate estimates.1 Again, taxonicity is evidenced by (i) plots that visually reflect taxonic distributions (e.g., MAMBAC and MAXEIG plots that are peaked; L-Mode plots that

Discussion

Evidence from three independent taxometric procedures strongly supported the dimensional perspective of narcissism. Graphically, the shapes of the plots produced by the procedures were indicative of dimensionality. Comparing the research data to multiple sets of simulated taxonic and dimensional data showed that the research data more closely matched the simulated dimensional data. Finally, base rate estimates varied within and among the procedures and were unrealistic. Based upon these

References (36)

  • W.K. Campbell et al.

    The narcissistic self: Background, an extended agency model, and ongoing controversies

  • R.A. Emmons

    Narcissism: Theory and measurement

    Journal of Personality and Social Psychology

    (1987)
  • M. First et al.

    The structured clinical interview for DSM-III-R personality disorders (SCID-II): I. Description

    Journal of Personality Disorders

    (1995)
  • S. Freud

    On narcissism: An introduction

  • S. Freud

    Libidinal types

  • S. Gangestad et al.

    To carve nature at its joints: on the existence of discrete classes in personality

    Psychological Review

    (1985)
  • Green, B. A., Arnau, R. C., & Gleaves, D. H. (1999). Taxometric analysis of the five-factor model constructs. Poster...
  • O. Kernberg

    Barriers to falling and remaining in love

    Journal of the American Psychoanalytic Association

    (1974)
  • Cited by (178)

    View all citing articles on Scopus
    View full text