Distinguishing Narcissism and Hostility: Similarities and Differences in Interpersonal Circumplex and Five-Factor Correlates

Home , Facet (psychology), Hostility

John M. Ruiz, Timothy W. Smith, and Frederick Rhodewalt Department of Psychology University of Utah

Narcissism and hostility are both characterized by dysfunctional social interactions, including tendencies to perceive slights, experience anger, and behave aggressively. The aim of this study was to examine the similarities and differences of narcissism and hostility, using 2 conceptual tools—the interpersonal circumplex and the Five-factor model. In a sample of 292 undergraduate men and women, composite measures of hostility (i.e., Buss–Perry Aggression Questionnaire [Buss & Perry, 1992] and Cook–Medley Hostility [Cook & Medley, 1954] total scores) were inversely correlated with affiliation and unrelated to dominance. In contrast, composite narcissism scores (i.e., Narcissistic Personality Inventory) were positively correlated with dominance and inversely correlated with affiliation. Examination of components of these traits revealed additional similarities and differences, as did associations with other dimensions of the Five-factor model. These findings suggest that the traits of narcissism and hostility are distinguishable by their interpersonal referents, as are their components.

Individual differences in the experience of interpersonal difficulties are a central focus in several aspects of personality psychology. Two widely studied personality characteristics, narcissism and hostility, are associated with dysfunctional social interactions. These constructs have been the focus of multiple lines of research in social (Emmons, 1987; Rhodewalt & Morf, 1995), clinical (Benjamin, 1996; Kernberg, 1984; Widiger & Trull, 1993), and health psychology (Miller, Smith, Turner, Guijarro, & Hallet, 1996; Siegman, 1994; Smith, 1992). 538 RUIZ, SMITH, RHODEWALT

Both narcissism and hostility reflect individual differences in the tendency to experience anger, behave aggressively, and perceive ambiguous social cues as threats or slights (Baumeister, Smart, & Boden, 1996; Christensen & Smith, 1993; Fox, 1974; Rhodewalt & Morf, 1995; Smith, 1992). Such conceptual similarities have not gone unnoticed (Raskin, Novacek, & Hogan, 1991; Rhodewalt & Morf, 1995). For example, in discussing personality traits conferring increased risk for coronary heart disease, Thoresen and Pattillo (1989) noted the similarities of Type A behavior pattern (TABP) and narcissism in terms of competitiveness, vigilance, and aggressive reactance to social threats. Moreover, Baumeister et al. (1996) suggested that inflated and fragile self-esteem, central in most descriptions of narcissism, underlies aggressive responses to threatened egotism. In this view, such aggressive behaviors serve as a common coping alternative to any downward revi- sion of the self-concept. Despite such similarities, these personality traits are rarely studied together. A more direct comparison may help to identify similarities and differences between measures intended to assess these related but distinct constructs, and suggest points of integration in the related literature. For example, hostility or aggression may serve an instrumental function associated with aspects of narcissism not uncovered with an examination that only compares composite measures of each trait. Thus, an interpersonal analysis can tell us something beyond the simple correlation of the global traits. A logical first step to distinguishing these constructs is comparison within a known taxonomy. Two conceptual tools—the Five-factor model (FFM) and the interpersonal circumplex—have been useful in developing a more integrated literature on personality traits (Kiesler, 1996; McCrae & Costa, 1987). The FFM provides a common conceptual framework for the organization and comparison of personality traits (McCrae & Costa, 1987; McCrae & John, 1992). By locating traits and scales in the conceptual space defined by the FFM, similarities and differences among personality scales and constructs can be identified. This process can foster integration and avoid redundancy in the otherwise haphazard proliferation of personality constructs and measures (McCrae & Costa, 1996). The integration of the FFM with the interpersonal circumplex (Trapnell & Wiggins, 1990) may be particularly useful in the development of an integrated literature on individual differences in interpersonal behavior. This approach integrates the FFM trait approach with the interpersonal circumplex by substituting the two principle circumplex dimensions of dominance versus submissiveness and friendliness versus hostility, respectively (Trapnell & Wiggins, 1990) for the interpersonally ori- ented traits in the FFM (i.e., Extraversion and Agreeableness). The dimensions of the interpersonal circumplex (Kiesler, 1996) are explicitly defined in terms of social behavior, and the related measures provide a well-validated (Trapnell & Wiggins, 1990; Wiggins & Broughton, 1991) framework for describing individual differences in social functioning (Gallo & Smith, 1998; Pincus & Gurtman, 1995). DISTINGUISHING NARCISSISM AND HOSTILITY 539

In this study we use this approach to describe similarities and differences in the traits of narcissism and hostility.

NARCISSISM, EMOTIONAL LABILITY, AND INTERPERSONAL HOSTILITY

Researchers in personality and clinical psychology have examined narcissism as a cause of emotional distress and dysfunctional social relations. The importance of social referents in characterizing narcissism is evident in both the personality and clinical literatures. For example, the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psychiatric Association, 1994) criteria for the diagnosis of narcissistic personality disorder (NPD) contain several maladaptive social behaviors and tendencies, volatile interpersonal experiences, and emotional lability. Kernberg’s (1975) conception of narcissism involves an egocentric de- pendence on positive and consistently reinforcing social relations. Emotional lability in response to negative social interactions is a hallmark trait of narcissism (American Psychiatric Association, 1994; Kernberg, 1975; Widiger & Trull, 1993). Specifically, individuals high in narcissism appear to be overly sensitive to criticism or threats to the self, often reacting with interpersonal hostility, including strong discounting of the source, feelings of rage, and potentially aggressive behaviors (Kernberg, 1975; Kohut, 1971; Rhodewalt, Madrian, & Cheney, 1998; Rhodewalt & Morf, 1998; Widiger & Trull, 1993). For example, Rhodewalt and Morf (1998) examined emotional responses of high and low narcissistic individuals to positive or negative feedback following task performance. Compared to participants low in narcissism, individuals high in narcissism demon- strated larger positive emotional responses to positive feedback and larger in- creases in anger following failure. Similar findings by Kernis (1994), Rhodewalt et al. (1998), and Rhodewalt and Morf (1995) suggested that the emotional reactions of individuals high in narcissism are strongly tied to interpersonal events, and self-relevant social information perceived as negative or ego threatening evokes higher levels of hostility, feelings of anger, and aggressive behavior in narcissistic individuals.

HOSTILITY AS AN INDIVIDUAL DIFFERENCE VARIABLE

In a largely separate literature, researchers in health psychology and behavioral medicine have examined hostility as a risk factor for physical illness (cf. Smith, 1992). From the investigation of the broadly defined TABP (Friedman & Rosenman, 1959), hostility has emerged as a multidimensional construct involving 540 RUIZ, SMITH, RHODEWALT affect, behavior, and cognition (Barefoot & Lipkus, 1994; Smith, 1992). Although some disagreement exists regarding precise definitions of these components, anger generally refers to negative affect ranging from feelings of irritation and annoyance to rage and may constitute an emotional state or an enduring trait. Aggression refers to attacking or destructive behavior. Hostility refers to a set of negative attitudes, beliefs, and appraisals of the worth, intent, and motives of others and often includes a desire to preemptively harm or see harm inflicted on others (Smith, 1992, 1994). These related facets, often collectively referred to with the umbrella term hostility represent separate although related dimensions of this individual difference variable. Furthermore, this set of traits confers increased risk of serious illness and pre- mature mortality (Miller et al., 1996). Previous research has documented consistent interpersonal correlates of trait hostility. For example, in several studies hostile persons displayed heightened anger and physiological reactivity to stressful interpersonal situations, relative to less hostile persons (Christensen & Smith, 1993; Smith & Gallo, 1999; Suarez & Wil- liams, 1989). In addition, hostile persons appraise the actions of others as inten- tionally aggressive and less friendly (Pope, Smith, & Rhodewalt, 1990; Smith, Sanders, & Alexander, 1990), process negative information about others more readily (Allred & Smith, 1991), behave in less friendly ways during interactions with family members (Smith et al., 1990), and experience discord and conflict in personal relationships (Miller et al., 1996; Newton & Kiecolt-Glaser, 1995; Smith, Pope, Sanders, Allred, & O’Keefe, 1988). The resulting interplay suggests that hostile individuals may not only respond to their environment in maladaptive ways, but also create increased frequency, intensity, and duration of interpersonal stress. This process could account, at least in part, for the health consequences of hostility (Smith, 1992). In summary, although generally studied independently, the personality traits of hostility and narcissism share the defining attributes of anger and aggressive response to perceived interpersonal slights, and a generally negative pattern of thoughts and beliefs about the motives and values of others. This suggests that a direct comparison of these traits and their respective components is needed to clarify their similarities and differences.

INTERPERSONAL CIRCUMPLEX AND FIVE-FACTOR APPROACHES TO CONSTRUCT VALIDATION

TheBigFiveorfive-factortraittaxonomyhasbeenusedextensivelytocompareand contrast personality constructs (Costa & McCrae, 1985, 1987; Piedmont, McCrae, & Costa, 1991; Wiggins & Trapnell, 1996). For example, in a prior study of narcissism, Bradlee and Emmons (1992) used the NEO Personality Inventory (NEO–PI; Costa & McCrae, 1985) to examine the five-factor correlates of a variant of the Nar- DISTINGUISHING NARCISSISM AND HOSTILITY 541 cissistic Personality Inventory (NPI; Raskin & Hall, 1979). The composite narcissism index was correlated with extroversion, antagonism, and low neuroticism; analyses of narcissism subscales revealed minor variations on this general pattern (see also Rhodewalt & Morf, 1995). Gallo and Smith (1998) recently used the interpersonal circumplex version of the FFM (Trapnell & Wiggins, 1990) to examine the construct validity of the Ag- gression Questionnaire (AQ; Buss & Perry, 1992). In the circumplex, all four subscales were associated with unfriendliness but differed in dominance. Hostility reflected unfriendly submission, verbal aggression reflected unfriendly dominance, and anger and physical aggression were unrelated to dominance. All four subscales were also positively correlated with neuroticism. Hence, the trait of dominance versus submission identified potentially important distinctions among components of this broader trait. In addition to their five-factor investigation of narcissism, Bradlee and Emmons (1992) attempted to explore interpersonal aspects of the NPI using the Personality Research Form (PRF; Jackson, 1974). Eight of the 22 PRF subscales were used to represent the interpersonal circumplex (Leary, 1957; Wiggins & Broughton, 1985). All NPI scales were positively related to agency with markers of affiliation distinguishing the various subscales. However, the PRF is not a validated measure of the interpersonal circumplex, leaving open the issue of a circumplex-based description of narcissism. Thus, the goals of this study are to examine the interpersonal circumplex and five-factor correlates of the total and subscale scores of the NPI and AQ using a well-validated measure of the interpersonal circumplex and five-factor traits (Trapnell & Wiggins, 1990). We predicted that both narcissism and hostility would be correlated with low affiliation (high hostility) and that the interpersonal dimension of dominance would distinguish these constructs and their components. Additionally, we predicted that aspects of narcissism associated with threatened egotism or entitlement (e.g., NPI subscale of Exploitiveness/Entitlement) would be most strongly associated with measures of hostility.

METHOD

Participants

Participants included 297 undergraduates (188 women and 109 men) from the Uni- versity of Utah. Participants were recruited from introductory psychology classes and received extra credit for their participation. The mean age was 23.5 years with a median and mode of 22 years. Participants were tested in groups of no more than 12 at a time. At the end of the session participants read the debriefing statement that addressed the purpose and basic goals of the study. 542 RUIZ, SMITH, RHODEWALT

Measures

Narcissism. The NPI (Raskin & Hall, 1979) is among the most widely used self-report measures of narcissism (Rhodewalt & Morf, 1995). It consists of 37 items in a true–false format adapted by Rhodewalt and Morf from the Emmons (1987) factor analysis. In constructing the NPI, Raskin and Hall (1979) drew from the Diagnostic and Statistical Manual of Mental Disorders (3rd ed.; American Psy- chiatric Association, 1980) criteria for NPD. Emmons’s (1984) factor analysis of the NPI revealed four separate factors he labeled Exploitiveness/Entitlement, Leadership/Authority, Superiority/Arrogance, and Self-Absorption/Self-Admira- tion, representing the principle criteria for diagnosis or characterization of narcissism. Emmons (1987) reported subscale correlations ranging from .16 to .44.

Hostility. The AQ (Buss & Perry, 1992) contains 29 self-report, Likert-type items assessing cognitive, affective, and behavioral components of this domain. Subscales include Hostility (i.e., hostile cognition), Anger (i.e., hostile affect), Physical Aggression, and Verbal Aggression. The factor structure of the AQ has been replicated in confirmatory analyses (Buss & Perry, 1992). Correlations among subscales range from .25 to .48. The Cook–Medley Hostility (Ho) scale (Cook & Medley, 1954) is a 50-item hostility measure derived from the Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1943). Originally designed to predict teacher rapport with students, the Ho scale is associated with experiential aspects of hostility including anger proneness, resentment, and mistrust as well as behaviorally assessed unfriendliness (Pope et al., 1990; Smith & Frohm, 1985; Smith et al., 1990).

Interpersonal circumplex and the FFM. The Revised Interpersonal Ad- jective Scales–Big Five (IASR–B5; Trapnell & Wiggins, 1990) is a 124-item adjective checklist with 64 items used to construct the circumplex and the remaining 60 items comprising the three remaining traits of the FFM (i.e., Neuroticism, Openness to Experience, Conscientiousness). Earlier versions relied on an 8-point rating format that was substituted here with a 5-point Likert rating format ranging from 1 (extremely accurate)to5(extremely inaccurate). To correct for this format change, standardized scores were calculated for this measure; means were calculated for the octant points and subtracted as constants from the individual participant’s octant scores(Wiggins,Trapnell,&Phillips,1988).Thesestandardizedoctantscoreswere thenusedtoderivetheprincipalcircumplexfactorsfordominanceandnurturance. The IASR–B5 also provides a measure of the other FFM traits—Neuroticism, Openness, and Conscientiousness—with 20 items (10 positive, 10 negative) each (Trapnell & Wiggins, 1990). The scores for dominance and nurturance are derived from a linear combination of the 64-item, eight-octant scales. The resulting factors DISTINGUISHING NARCISSISM AND HOSTILITY 543 for dominance and nurturance are related to Surgency/Extraversion and Agreeable- ness,respectively,andreplacethesetraitsoftheFFM(Trapnell&Wiggins,1990).

Data Analysis

There are a variety of methods available to examine the association of measures with the interpersonal circumplex dimensions (Gurtman, 1997; Trapnell & Wiggins, 1990). The method outlined by Wiggins and Broughton (1991) was em- ployed here. The first step in this procedure is to characterize the relation between a measure such as Verbal Aggression with the two principal axes of the circumplex. Two Pearson product–moment correlations are derived for each measure that serve as x (x = rvx) and y (y = rvy) coordinates and allow for locating the variable in the circumplex. The circumplex itself is based on an origin of r = 0 with r+1 delineat- ing the positive and negative poles of the individual axes. Vector lengths calculated using the multiple correlation for the variable of interest and ranging from a mini- mum of 0 projecting from the origin to a maximum of 1 characterize the strength of the interpersonal nature of the variable. The angle of displacement from the x axis is equal to q = tan–1(y/x). Additionally, the independent contributions of the remaining three factors of Neuroticism, Conscientiousness, and Openness were assessed using simultaneous multiple regression analyses.

RESULTS

Means, standard deviations, internal consistencies, and intercorrelations for the Cook–Medley, AQ, and NPI are displayed in Table 1. Consistent with expecta- tions, the composite measures of hostility were strongly correlated and showed a strong association with the overall measure of narcissism. Additionally, the NPI Exploitiveness/Entitlement subscale showed the strongest correlations with both the composite and facet measures of hostility.

Interpersonal Circumplex

The results for the projection of the AQ, the Cook–Medley Ho scale, and the NPI into circumplex space for the total sample are shown in Figures 1 and 2. All measures had significant associations with the dimensions of the two-factor interpersonal circumplex (i.e., significant multiple r). Composite score analysis for the hostility and narcissism measures is displayed in Figure 1. Both measures of hostility were significantly associated with the affiliation factor (see Table 2), but not with the dominance factor. The composite score for theNPIcorrelatedsignificantlywithbothdominanceandaffiliation(seeTable3). Comparison between Figures 1 and 2 demonstrates the value of examining facets as opposed to reliance on composite scores. The four AQ subscales were all 544

TABLE 1 Intercorrelations, Reliabilities, and Descriptive Statistics for the Cook–Medley Ho, AQ, and the NPI 12345 678 9 1011 1. Cook–Medley .62** .41** .38** .50** .60** .17** –.01 –.07 .21** .46** 2. AQ Total .79** .66** .81** .81** .20** .04 –.06 .22** .49** 3. Physical .36** .50** .44** .22** .10 .04 .25** .32** Aggression 4. Verbal .47** .39** .30** .20** .05 .27** .41** Aggression 5. Anger .59** .13** –.02 –.06 .12 .42** 6. Hostility .02 –.12* .20** .05 .41** 7. NPI Total .79** .67** .82** .59** 8. Leadership/ .40** .54** .24** Authority 9. Self-Absorption/ .36** .14* Self-Admiration 10. Superiority/ .50** Arrogance 11. Exploitiveness/ Entitlement α .88 .91 .82 .73 .72 .81 .83 .77 .65 .55 .67 M 17.61 61.22 17.29 13.36 13.15 17.43 14.69 4.89 4.61 3.64 1.55 SD 8.95 15.77 6.42 3.85 4.15 5.96 6.12 2.53 2.13 2.01 1.71

Note. n = 292. Ho = Hostility scale; AQ = Aggression Questionnaire; NPI = Narcissistic Personality Inventory. *p < .05. **p < .01. DISTINGUISHING NARCISSISM AND HOSTILITY 545

Dominance (PA)

(BC) (NO)

Hostility (DE) (LM) Friendliness

(FG) (JK)

(HI) Submissiveness

FIGURE 1 Projection of the composite scores of hostility and narcissism into interpersonal

circumplex space, for the total sample. Vector length for the Cook–Medley Ho measure is VLCM = .40. Vector length for Narcissistic Personality Inventory is VLNPI = .54. Vector length for Ag-

gression Questionnaire is VLAQ = .54. significantly and negatively associated with the affiliation factor to varying de- grees (see Table 2). Correlations with the dominance factor varied among the AQ subscales. Verbal aggression reflected unfriendly dominance, whereas both anger and hostility reflected unfriendly submission; physical aggression was unrelated. Correlations for the NPI are given in Table 3 and displayed in Figure 2. Leader- ship/Authority and Self-Absorption/Self-Admiration correlated with interpersonal dominance but not friendliness, Superiority/Arrogance reflected unfriendly dominance, and Exploitiveness/Entitlement reflected interpersonal hostility or unfriendliness with no association with the dominance factor. Again, the examination of subscale correlations with circumplex dimensions reveals important distinctions otherwise overlooked by the use of composite scores.

FFM. Each total scale and subscale was regressed simultaneously on the FFM traits. Results for the AQ are presented in Table 2 and for the NPI in Table 3.1

1Gender analyses revealed relatively few differences. However, for the AQ Hostility subscale, semipartial correlations revealed low dominance as the strongest predictor among men, whereas 546 RUIZ, SMITH, RHODEWALT

Dominance (PA)

(BC) (NO)

Hostility (DE) (LM) Friendliness

(FG) (JK)

(HI) Submissiveness

FIGURE 2 Projection of the Aggression Questionnaire (AQ) and Narcissistic Personality In- ventory (NPI) into interpersonal circumplex space, for the total sample. Vector lengths for the

AQ subscales are VLHO = .49, VLAN = .43, VLPA = .44, and VLVA = .44. Vector lengths for NPI subscales are VLLA = .60, VLSS = .36, VLSA = .43, and VLEE = .42.

Results are presented for the unstandardized (B) and standardized (b) regression coefficients, t-test values and significance levels, and zero-order (r) and semipartial (sr2) correlations. Additionally, multiple R, R2, and F values are given to provide the final, integrated interpersonal circumplex/five-factor framework solutions for each subscale. neuroticism was the best predictor of higher scores for women. For NPI composite scores, dominance, neuroticism, and openness emerged as significant predictors for men with dominance and low affiliation emerging for women. For the NPI Leadership/Authority subscale dominance, low affiliation, and low openness emerged as significant predictors for men, whereas dominance emerged as the only predictor for women. Finally, low affiliation emerged as the only significant predictor of higher Exploitiveness/Entitlement scores for men, whereas dominance, low affiliation, and neuroticism predicted higher scores for women. TABLE 2 Multiple Regression Analysis for the Cook–Medley, Aggression Questionnaire, and Five-Factor Model

Variable B β trsr2

Cook–Medleya Dominance 0.19 .02 0.34 –.06 .02 Affiliation –3.17 –.35 –5.73*** –.40*** –.32 Conscientiousness –0.27 –.02 –0.36 –.15** –.02 Neuroticism 1.88 .14 2.49** .23*** .15 Openness –0.91 –.05 –0.89 –.14** –.05 Aggression Questionnaire totalb Dominance 0.82 .05 0.96 –.08 .06 Affiliation –7.17 –.45 –8.35*** –.54*** –.44 Conscientiousness –0.64 –.03 –0.55 –.20*** –.03 Neuroticism 5.91 .26 5.02*** .37*** .28 Openness –2.87 –.09 –1.81 –.21*** –.11 Physical Aggressionc Dominance 0.68 .10 1.77 .03 .10 Affiliation –2.56 –.39 –6.60*** –.43*** –.36 Conscientiousness –0.45 –.05 –0.85 –.17** –.05 Neuroticism 0.14 .02 0.26 .10 .02 Openness –1.39 –.11 –1.94* –.20*** –.11 Verbal Aggressiond Dominance 0.94 .24 4.08*** .17** .23 Affiliation –1.47 –.38 –6.34*** –.39*** –.35 Conscientiousness 0.01 .02 0.30 –.05 .02 Neuroticism 0.84 .15 2.65** .18** .15 Openness 0.00 .00 0.05 –.05 .00 Angere Dominance 0.14 .03 0.58 –.12* .03 Affiliation –1.26 –.30 –5.36*** –.42*** –.30 Conscientiousness –0.26 –.04 –0.80 –.17** –.05 Neuroticism 2.24 .37 6.94*** .45*** .38 Openness –0.42 –.05 –0.97 –.14** –.60 Hostilityf Dominance –0.94 –.15 –2.83** –.28*** –.16 Affiliation –1.87 –.31 –5.64*** –.43*** –.31 Conscientiousness 0.00 .00 –0.06 –.19*** –.00 Neuroticism 2.69 .31 5.90*** .44*** .33 Openness –1.08 –.09 –1.77 –.21*** –.10

aMultiple R = .43; R2 = .18; F(5, 288) = 12.74, p < .001. bMultiple R = .60; R2 = .36; F(5, 291) = 32.18, p < .001. cMultiple R = .45; R2 = .20; F(5, 291) = 14.90, p < .001. dMultiple R = .46; R2 = 21; F(5, 291) = 15.72, p < .001. eMultiple R = .55; R2 = .30; F(5, 291) = 25.02, p < .001. fMultiple R = .57; R2 = .33; F(5, 291) = 28.29, p < .001. *p < .05. **p < .01. ***p < .001

547 548 RUIZ, SMITH, RHODEWALT

TABLE 3 Multiple Regression Analysis for the Narcissistic Personality Inventory and Five-Factor Model

Variable B β trsr2

NPI Totala Dominance 3.38 .56 9.78*** .51*** .50 Affiliation –0.96 –.15 –2.74** –.13* –.16 Conscientiousness –0.15 –.02 –0.32 .08 –.02 Neuroticism –0.38 –.04 –0.79 –.15** –.05 Openness –1.05 –.09 –1.63 –.02 –.10 Leadership/Authorityb Dominance 1.61 .61 11.86*** .60*** .57 Affiliation 0.00 .01 0.16 .04 .01 Conscientiousness 0.00 –.02 –0.45 .15** –.03 Neuroticism –0.15 –.04 –0.77 –.21*** –.05 Openness –0.39 –.08 –0.16 .05 –.09 Self-absorption/Self-admirationc Dominance 0.64 .29 4.86*** .36*** .27 Affiliation 0.00 –.01 –0.11 .08 –.01 Conscientiousness 0.16 .05 0.86 .16** .05 Neuroticism –0.49 –.16 –2.64** –.24*** –.15 Openness 0.15 .04 0.61 .11 .04 Superiority/Arroganced Dominance 0.86 .41 7.10*** .38*** .39 Affiliation –0.37 –.18 –2.98** –.17** –.17 Conscientiousness 0.00 –.03 –0.53 .02 –.03 Neuroticism –0.17 –.06 –1.02 –.12* –.06 Openness –0.51 –.13 –2.26* –.09 –.13 Exploitiveness/Entitlemente Dominance 0.27 .15 2.63** .05 .15 Affiliation –0.59 –.34 –5.69*** –.42*** –.32 Conscientiousness –0.13 –.05 –0.92 –.15** –.05 Neuroticism 0.46 .17 3.02** .23*** .18 Openness –0.31 –.09 –1.60 –.16** –.09

aMultiple R = .55; R2 = .30; F(5, 290) = 24.92, p < .001. bMultiple R = .61; R2 = .37; F(5, 290) = 34.38, p < .001. cMultiple R = .39; R2 = .15; F(5, 290) = 10.52, p < .001. dMultiple R = .46; R2 = .21; F(5, 290) = 15.10, p < .001. eMultiple R = .46; R2 = .21; F(5, 289) = 15.39, p <. 001. *p < .05. **p < .01. ***p < .001.

Cook–Medley Ho Scale

The overall F value was significant for the total sample. The FFM accounted for 18% of the variance with affiliation and neuroticism emerging as significant predictors. Higher Cook–Medley Ho scores were associated with higher neuroticism and lower affiliation. An examination of the semipartial correlation coefficients indicated that affiliation was the best predictor of Cook–Medley scores. DISTINGUISHING NARCISSISM AND HOSTILITY 549

Composite score. The overall F value for the composite AQ index was significant for the total sample. The FFM accounted for 36% of the variance with affiliation, conscientiousness, neuroticism, and openness emerging as significant predictors. Higher composite AQ scores were associated with lower affiliation, conscientiousness, and openness and higher neuroticism. An examination of the semipartial correlation coefficients indicated that affiliation was the best predictor of composite AQ scores.

Physical Aggression. The overall F value was significant for the total. The FFM accounted for 20% of the variance for the total sample, with affiliation and openness emerging as significant predictors. Higher Physical Aggression subscale scores were associated with lower affiliation and openness. An examination of the semipartial correlation coefficients revealed affiliation as the strongest predictor of Physical Aggression scores.

Verbal Aggression. The overall F value for Verbal Aggression was significant for the total sample. The FFM accounted for 21% of the variance. Dominance, affiliation, and neuroticism emerged as significant predictors. Higher Verbal Ag- gression scores were associated with higher dominance and neuroticism and lower affiliation. An examination of the semipartial correlation coefficients indicated that affiliation was the best predictor of Verbal Aggression scores.

Anger. The overall F value was again significant for the total sample. The FFM accounted for 30% of the variance in the total sample. Affiliation and neuroticism emerged as significant predictors. Higher Anger scores were associated with higher neuroticism and lower affiliation. An examination of the semipartial correlation coefficients revealed neuroticism as the strongest individual predictor of higher Anger scores.

Hostility. The overall F value was significant for the total sample. The FFM accounted for 33% of the variance. Dominance, affiliation, and neuroticism emerged as significant independent predictors. Higher Hostility scores were associated with lower dominance and affiliation and higher neuroticism. An examination of the semipartial correlation coefficients revealed that neuroticism was the best predictor of Hostility scores, with affiliation also emerging as a strong predictor.

NPI

Composite score. The overall F values were significant for the composite score for the total sample. The FFM accounted for 30% of the variance with domi- 550 RUIZ, SMITH, RHODEWALT nance and affiliation emerging as significant predictors. Higher composite NPI scores were associated with higher dominance and lower affiliation. An examination of the semipartial correlation coefficients revealed dominance as the best predictor of higher scores.

Leadership/Authority. The overall F values were significant for the composite score for the total sample. The FFM accounted for 37% of the variance with higher dominance emerging as the only significant predictor.

Self-Absorption/Self-Admiration. The overall F values were significant for the composite score for the total sample. The FFM accounted for 15% of the variance with higher dominance emerging as the only significant predictor.

Superiority/Arrogance. The overall F values were significant for the composite score for the total sample. The FFM accounted for 21% of the variance with dominance, affiliation, and openness emerging as significant predictors. Higher Superiority/Arrogance scores were associated with higher dominance and lower affiliation and openness. An examination of the semipartial correlation coefficients revealed dominance as the best predictor of higher scores.

Exploitiveness/Entitlement. The overall F value was significant for the total sample. The FFM accounted for 21% of the variance with dominance, affiliation, and neuroticism emerging as significant predictors. Higher Exploitiveness/ Entitlement scores were associated with higher dominance and neuroticism and lower affiliation. An examination of the semipartial correlation coefficients revealed affiliation as the strongest predictor of Exploitiveness/Entitlement scores.

DISCUSSION

As predicted, the results suggest that although both narcissism and hostility reflect a cold, unfriendly interpersonal style, they are clearly distinguishable. In the circumplex analyses, the composite measures of hostility and narcissism reflected low affiliation in the interpersonal circumplex, suggesting a basis for the common association between the personality traits. However, narcissism was also associated with interpersonal dominance, whereas the composite measures of hostility were not. In the FFM analyses, the composite hostility scales were associated with neuroticism, but narcissism was not. These findings are consistent with those of Gallo and Smith (1998) in describing trait hostility as a combination of low interpersonal affiliation and neuroticism. In contrast, the association of narcissism with dominance and hostility reflected the interpersonal arrogance commonly associated with this personality characteristic. DISTINGUISHING NARCISSISM AND HOSTILITY 551

An examination of the facet scales of the AQ and NPI revealed additional distinctions between these constructs and the related measures. For the AQ, the subscales of Physical Aggression and Verbal Aggression were most closely associated with low affiliation, whereas the Hostility and Anger subscales reflected a blend of both low affiliation and neuroticism. The AQ subscales were further distinguishable by their associations with dominance, positive in the case of Verbal Aggression and negative for both Anger and Hostility. Three of the four facets scales of the NPI were characterized by interpersonal dominance. The one excep- tion, Exploitiveness/Entitlement, was characterized by a combination of interpersonal hostility (i.e., low affiliation) and neuroticism, similar to the pattern associated with the measures of hostility. The NPI facets are further distinguishable by their association with the affiliation axis of the interpersonal circumplex, negative for both Exploitiveness/Entitlement and Superiority/Arrogance and unrelated to either the Leadership/Authority or Self-Absorption/Self-Admiration subscales. Hence, the potential overlap of narcissism and hostility varies consider- ably across components of these broad traits. These results emphasize the potential problems with the practice of combining distinct personality characteristics in composite scores. For example, the NPI Exploitiveness subscale was found to reflect hostility and neuroticism, whereas the NPI subscales assessing leadership and self-absorption are unrelated—or even inversely related—to these traits. Failure to recognize these distinctions could foster misleading conclusions about the association of narcissism with various aspects of emotional and social functioning. Furthermore, individuals could obtain a high narcissism score through quite different patterns of interpersonal behavior. Similarly, the various aspects of hostility share a major similarity (i.e., low affiliation), but differ in their association with dominance and negative emotionality. Such differences could suggest that types of hostility impact health to differing de- grees or through differing mechanisms (Gallo & Smith, 1998). Thus, the study of these multifaceted constructs must include examinations of their individual components (Carver, 1989). The combined approaches of the interpersonal circumplex and FFM provided a valuable framework in which to examine hostility and narcissism. As expected, all three measures were well represented in the two approaches, reflecting the importance of interpersonal and intrapersonal traits in their characterization. However, there are several important qualifications that must be addressed with respect to these findings. First, our data are limited to a young, predominantly White undergraduate sample, limiting generalizability with respect to age and race. Second, use of the NPI as a measure of narcissism, especially in this sample, precludes us from generalizing beyond normal personality variation to conclusions about diagnosable NPD. That is, just as findings of studies of variations in depressive symptoms within the normal range cannot be casually generalized to clinical affective disorder (Coyne, 1994; Smith & Rhodewalt, 1991), the study of variations in symptoms of 552 RUIZ, SMITH, RHODEWALT personality disorder within normal populations may or may not elucidate the clini- cally meaningful personality disorders that disrupt individuals’ lives. Although the personality taxonomy used in this study provides a useful method for the organization of personality traits, it is not without detractors (cf. Block, 1995; McAdams, 1992). Moreover, the item distribution of the IASR–B5 provides the affiliation and dominance factors (64 items) with potentially more precise assessment relativetothethreeremainingfactors(20itemseach).Thisinequalitycouldhaveled to an underestimate of the association involving these factors in the simultaneous regression analyses. Finally, although this version of the FFM taxonomy is particularly appropriate for this study given the largely interpersonal nature of the constructs of interest, other versions of the FFM (and related assessment proce- dures) such as the facet scales for each of the FFM traits provided by the NEO–PI–R (Costa & McCrae, 1992; McCrae & Costa, 1992) provide additional tools for com- paring and contrasting narcissism, hostility, and other traits. Finally, the interpersonal circumplex and FFM are descriptive tools and do not yet offer mechanistic explanations for the relations between narcissism, hostility, andsocialbehavior.Alogicalnextstepinfurtherdelineatingtheseconstructswould be to examine the common and unique underlying determinants or personality pro- cesses (Cantor, 1990) associated with narcissism, hostility, and their components. For example, in contrast to conventional wisdom, several researchers have suggested that such expressions are tied to threatened egotism rather than low self-esteem(Baumeisteretal.,1996;Kernis,Grannemann,&Barclay,1989).Nonetheless, thefindingsofthisstudyillustratethevalueofcomparingrelatedtraitsandmeasures through the use of well-established personality taxonomies and frameworks. Given the widespread proliferation of individual difference measures and related potential for rediscovering old traits under new labels, the interpersonal and five-factor approaches offer an important opportunity for organization, clarification, and integration across the literature.

ACKNOWLEDGMENTS

A previous version of this article was presented at the first annual meeting of the So- ciety for Personality and Social Psychology in Nashville, TN, February 2000. We thank Heidi A. Hamann, Bert N. Uchino, and two anonymous reviewers for their helpful comments on an earlier draft of this article.

REFERENCES

Allred, K. D., & Smith, T. W. (1991). Social cognition in cynical hostility. Cognitive Therapy and Re- search, 15, 399–412. American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. DISTINGUISHING NARCISSISM AND HOSTILITY 553

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Barefoot, J. C., & Lipkus, I. M. (1994). The assessment of anger and hostility. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp.43–66). Hillsdale, NJ: Lawrence Erlbaum Associ- ates, Inc. Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation of threatened egotism to violence and aggression: The dark side of high self-esteem. Psychological Review, 103, 5–33. Benjamin, L. S. (1996). Interpersonal diagnosis and treatment of personality disorders (2nd ed.). New York: Guilford. Block, J. (1995). A contrarian view of the five-factor approach to personality description. Psychological Bulletin, 117, 187–215. Bradlee, P. M., & Emmons, R. A. (1992). Locating narcissism within the interpersonal circumplex and the Five-factor model. Personality and Individual Differences, 13, 821–830. Buss, A. H., & Perry, M. (1992). The Aggression Questionnaire. Journal of Personality and Social Psy- chology, 63, 452–459. Cantor, N. (1990). From thought to behavior: “Having” and “doing” in the study of personality and cognition. American Psychologist, 45, 735–750. Carver, C. S. (1989). How should multifaceted personality constructs be tested? Issues illustrated by self-monitoring, attributional style, and hardiness. Journal of Personality and Social Psychology, 56, 577–585. Christensen, A. J., & Smith, T. W. (1993). Cynical hostility and cardiovascular reactivity during self- disclosure. Psychosomatic Medicine, 55, 193–202. Cook, W. W., & Medley, D. M. (1954). Proposed hostility and pharisaic-virtue scales for the MMPI. Journal of Applied Psychology, 38, 414–418. Costa, P. T., Jr., & McCrae, R. R. (1985). The NEO Personality Inventory manual. Odessa, FL: Psycho- logical Assessment Resources. Costa, P. T., Jr., & McCrae, R. R. (1987). Personality assessment in psychosomatic medicine. In T. M. Wise (Ed.), Advances in psychosomatic medicine (pp. 71–82). Basel, Switzerland: Krager. Costa, P. T., Jr., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO Personality Inventory. Psychological Assessment, 4, 5–13. Coyne, J. C. (1994). Self-reported distress: Analogue or ersatz depression? Psychological Bulletin, 116, 29–45. Emmons, R. A. (1984). Factor analysis and construct validity of the Narcissistic Personality Inventory. Psychological Reports, 48, 247–250. Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of Personality and Social Psy- chology, 52, 11–17. Fox, R. P. (1974). Narcissistic rage and the problem of combat aggression. Archives of General Psychia- try, 31, 807–811. Friedman, M., & Rosenman, R. H. (1959). Association of specific overt behavior pattern with blood and cardiovascular findings: Blood cholesterol level, blood clotting time, incidence of arcus senilis, and clinical coronary artery disease. Journal of the American Medical Association, 169, 1286–1296. Gallo, L. C., & Smith, T. W. (1998). Construct validation of health-relevant personality traits: Interper- sonal circumplex and Five-factor model analyses of the Aggression Questionnaire. International Journal of Behavioral Medicine, 5, 129–147. Gurtman, M. B. (1997). Studying personality traits: The circular way. In R. Plutchik & H. R. Conte (Eds.), Circumplex models of personality and emotions (pp. 81–102). Washington, DC: American Psychological Association. Hathaway, S. R., & McKinley, J. C. (1943). The Minnesota Multiphasic Personality Inventory. Minne- apolis: University of Minnesota Press. 554 RUIZ, SMITH, RHODEWALT

Jackson,D.N.(1974).PersonalityResearchFormmanual.Goshen,NY:ResearchPsychologistsPress. Kernberg, O. (1975). Borderline conditions and pathological narcissism. New York: Aronson. Kernberg, O. (1984). Severe personality disorders. New Haven, CT: Yale University Press. Kernis, M. H. (1994). Narcissism and reactions to interpersonal feedback. Journal of Research in Per- sonality, 28, 4–13. Kernis, M. H., Grannemann, B. D., & Barclay, L. C. (1989). Stability and level of self-esteem as predictors of anger arousal and hostility. Journal of Personality and Social Psychology, 56, 1013–1022. Kiesler, D. J. (1996). Contemporary interpersonal theory and research: Personality, psychopathology, and psychotherapy. New York: Wiley. Kohut, H. (1971). The analysis of the self. New York: International Universities Press. Leary, T. (1957). Interpersonal diagnosis of personality: A functional theory and methodology for personality evaluation. New York: Ronald. McAdams, D. P. (1992). The Five-factor model of personality: A critical appraisal. Journal of Personal- ity, 60, 329–361. McCrae, R. R., & Costa, P. T., Jr. (1987). Validation of the Five-factor model of personality across in- struments and observers. Journal of Personality and Social Psychology, 52, 81–90. McCrae, R. R., & Costa, P. T., Jr. (1992). Discriminant validity of NEO–PI–R facet scales. Educational and Psychological Measurement, 52, 229–237. McCrae, R. R., & Costa, P. T., Jr. (1996). Toward a new generation of personality theories: Theoretical contexts for the Five-factor model. In J. S. Wiggins (Ed.), The Five-factor model of personality (pp. 51–87). New York: Guilford. McCrae, R. R., & John, O. P. (1992). An introduction to the Five-factor model and its applications. Jour- nal of Personality, 60, 175–215. Miller, T. Q., Smith, T. W., Turner, C. W., Guijarro, M. L., & Hallet, A. J. (1996). A meta-analytic review of research on hostility and physical health. Psychological Bulletin, 119, 322–348. Newton, T. L., & Kiecolt-Glaser, J. K. (1995). Hostility and erosion of marital quality during early mar- riage. Journal of Behavioral Medicine, 18, 601–619. Piedmont, R. L., McCrae, R. R., & Costa, P. T., Jr. (1991). Adjective check list scales and the Five-factor model. Journal of Personality and Social Psychology, 60, 630–637. Pincus, A. L., & Gurtman, M. B. (1995). The three faces of interpersonal dependency: Structural analyses of self-report dependency measures. Journal of Personality and Social Psychology, 69, 744–758. Pope, M. K., Smith, T. W., & Rhodewalt, F. (1990). Cognitive, behavioral, and affective correlates of the Cook and Medley Hostility Scale. Journal of Personality Assessment, 54, 501–514. Raskin,R.N.,&Hall,C.S.(1979).ANarcissisticPersonalityInventory.PsychologicalReports,46,590. Raskin, R. N., Novacek, J., & Hogan, R. (1991). Narcissistic self-esteem management. Journal of Per- sonality and Social Psychology, 60, 911–918. Rhodewalt, F., Madrian, J. C., & Cheney, S. (1998). Narcissism, self-knowledge organization, and emotional reactivity: The effect of daily experiences on self-esteem and affect. Personality and Social Psychology Bulletin, 24, 75–87. Rhodewalt, F., & Morf, C. C. (1995). Self and interpersonal correlates of the Narcissistic Personality In- ventory: A review and new findings. Journal of Research in Personality, 29, 1–23. Rhodewalt, F., & Morf, C. C. (1998). On self-aggrandizement and anger: A temporal analysis of narcissism and affective reactions to success and failure. Journal of Personality and Social Psychology, 74, 672–685. Siegman, A. W. (1994). From Type A to hostility to anger: Reflections on the history of coronary-prone behavior. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp. 1–22). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Smith, T. W. (1992). Hostility and health: Current status of a psychosomatic hypothesis. Health Psy- chology, 11, 139–150. DISTINGUISHING NARCISSISM AND HOSTILITY 555

Smith, T. W. (1994). Concepts and methods in the study of anger, hostility, and health. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp. 23–42). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Smith, T. W., & Frohm, K. D. (1985). What’s so unhealthy about hostility? Construct validity and psychosocial correlates of the Cook and Medley Ho Scale. Health Psychology, 4, 503–520. Smith, T. W., & Gallo, L. (1999). Hostility and cardiovascular reactivity during marital interaction. Psy- chosomatic Medicine, 61, 436–445. Smith, T. W., Pope, M. K., Sanders, J. D., Allred, K. D., & O’Keefe, J. L. (1988). Cynical hostility at home and work: Psychosocial vulnerability across domains. Journal of Research in Personality, 22, 525–548. Smith, T. W., & Rhodewalt, F. T. (1991). Methodological challenges at the social/clinical interface. In C. R. Snyder & D. Forsyth (Eds.), Handbook of social and clinical psychology (pp. 739–756). New York: Pergamon. Smith, T. W., Sanders, J. D., & Alexander, J. F. (1990). What does the Cook and Medley Hostility Scale measure? Affect, behavior, and attributes in the marital context. Journal of Personality and Social Psychology, 58, 699–708. Suarez, E. C., & Williams, R. B., Jr. (1989). Situational determinants of cardiovascular and emotional reactivity in high and low hostile men. Psychosomatic Medicine, 51, 404–418. Thoresen, C. E., & Pattillo, J. R. (1989). Exploring the Type A behavior pattern in children and adoles- cents. In B. K. Houston & C. R. Snyder (Eds.), Type A behavior pattern: Research, theory, and inter- vention (pp. 98–145). New York: Wiley. Trapnell, P. D., & Wiggins, J. S. (1990). Extension of the interpersonal adjective scales to include the Big Five dimensions of personality. Journal of Personality and Social Psychology, 59, 781–790. Widiger, T. A., & Trull, T. J. (1993). Borderline and narcissistic personality disorders. In P. B. Sutker & H. E. Adams (Eds.), Handbook of psychopathology (2nd ed., pp. 371–394). New York: Plenum. Wiggins, J. S., & Broughton, R. (1985). The interpersonal circle: A structural model for the integration of personality research. In R. Hogan & W. H. Jones (Eds.), Perspectives in personality (pp. 1–47). Greenwich, CT: JAI. Wiggins, J. S., & Broughton, R. (1991). A geometric taxonomy of personality scales. European Journal of Personality, 5, 343–365. Wiggins, J. S., & Trapnell, P. D. (1996). A dyadic-interactional perspective on the Five-factor model. In J. S. Wiggins (Ed.), The Five-factor model of personality (pp. 88–162). New York: Guilford. Wiggins, J. S., Trapnell, P., & Phillips, N. (1988). Psychometric and geometric characteristics of the Revised Interpersonal Adjective Scales (IAS–R). Multivariate Behavioral Research, 23, 517–530.

John M. Ruiz Department of Psychology University of Utah 390 South 1530 East, Room 502 Salt Lake City, UT 84112 E-mail: [email protected]

Received April 2, 2000 Revised September 12, 2000