The Narcissistic Personality Inventory: factor structure in a non-clinical sample

Abstract

The Narcissistic Personality Inventory (NPI) is the most widely used measure of the construct of narcissism. Four- and seven-factor solutions have been reported for the instrument. In the present study, 338 undergraduates completed the NPI along with a battery of personality questionnaires that include the NEO-FFI. Exploratory principal components analysis indicated that the NPI had a two- or three-factor structure. Confirmatory factor analyses were undertaken for one-, two- and three-factor models of the instrument. Fit indices were poor, typical of models with many item–level variables. The fits can be improved by allowing plausible correlated error terms in instances where items have very similar content. As a whole, the NPI is measuring a general narcissism construct, with two or three separable, correlated factors relating to ‘power’, ‘exhibitionism’, and being a ‘special person’. A psychometrically improved version of the NPI could be developed based on these factors. In the present study confirmatory factor analysis provided some insights not available from exploratory factor analysis, but was still largely exploratory in nature. NEO correlations with the overall factor were 0.36 for both extraversion and low agreeableness, with additional highly significant correlations for low neuroticism and high openness to experience. NEO correlations for the lower-order factors were similar.

Introduction

The research literature on narcissism contains numerous complaints that the term is overused (Kernberg, 1975, Pulver, 1986, Westen, 1990). In some variants of the myth from which the construct takes its name (Bulfinch, 1959), there are two protagonists, Narcissus and Echo. Narcissus is grandiose, arrogant, and ruthlessly defends his self-image of perfection; Echo is interested in others only as a means of regulating her fragile self-esteem. The traditional interpretation of these versions of the myth (Jorstad, 1995) is that Narcissus and Echo represent two personality types which look as if they could not be more different, but are really just two sides of the same coin: the failure to love oneself leading through one path or the other to the inability to love others.

In some of the psychological literature on narcissism (Wink, 1991) there is a distinction between “overt” narcissism (represented by Narcissus) and “covert” narcissism (represented by Echo). The DSM-IV Narcissistic Personality Disorder corresponds to overt narcissism; covert narcissism is probably closest to Borderline Personality Disorder. The best empirical support for overt and covert variants of narcissism is Wink's (1991) principal components analysis of six MMPI Narcissism Scales, which found two orthogonal components: one was associated with extraversion, aggression, and self-assurance, the other with introversion, anxiety and defensiveness, but both were associated with conceit, self-indulgence, and disregard of others.

In other versions of the myth (Graves, 1990), Narcissus is called “Antheus”, a surname of Dionysus, the god to whom Nietzsche (1973a) traced the will to power. Nietzsche (1973b) proposed that all psychology be viewed as the morphology and development of will to power. Nietzsche's theory had a profound influence on psychoanalysis, which is where most of the theorizing on narcissism has come from. Some of the literature on narcissism (e.g. Frommm, 1965, Kohut, 1986) traces virtually all human initiative, creativity and goodness to healthy narcissism, and virtually all psychopathology to unhealthy narcissism. The upshot seems to be that narcissism has to do with relationships between the part (individual) and the whole, which inevitably lead to issues of love (what am I willing to do for others?) and power (how do I get what I want from others?).

Until 1979, research on the clinically important construct of narcissism was hampered by the lack of an appropriate measurement instrument (Emmons, 1984). Raskin and Hall (1979) developed the Narcissistic Personality Inventory (NPI), which has since become the most widely used measure of narcissism. Raskin and Terry (1988) refined the instrument but argue that the NPI item pool does not account for all the dimensions central to narcissism, and urge continued efforts to improve the scale. The present study extends these efforts, arguing for an overall construct of narcissism and three interesting factors that call for further development.