Sex Roles

, Volume 63, Issue 7, pp 530–541

Male Narcissism and Attitudes Toward Heterosexual Women and Men, Lesbian Women, and Gay Men: Hostility toward Heterosexual Women Most of All

Authors

Original Article

DOI: 10.1007/s11199-010-9837-8

Cite this article as:
Keiller, S.W. Sex Roles (2010) 63: 530. doi:10.1007/s11199-010-9837-8

Abstract

The present study investigated links between heterosexual men’s narcissism and attitudes toward heterosexual and non-heterosexual women and men. Male narcissism was predicted to be associated with hostility toward heterosexual women more than toward other groups, indicating investment in patriarchal power more than in conservative gender ideology or nonspecific disdain toward all people. Hierarchical regression analyses of responses from 104 male undergraduates (95% Caucasian) from Ohio in the U.S. supported the hypothesis that men’s narcissism is related most robustly to hostility toward women, rather than to equivalent derogation of all groups. Moreover, men’s narcissism is not merely a maker of traditional gender ideology, but instead is associated with favorable attitudes toward lesbian women and is unrelated to attitudes toward gay men.

Keywords

NarcissismAmbivalent sexismBenevolent sexismAttitudes toward womenAttitudes toward menAttitudes toward homosexuality

Introduction

The goals of the present study were to show that heterosexual men’s narcissism is (1) linked to hostility beyond proximal targets (i.e., specific individuals in the context of interpersonal interactions and relationships) such that it is evident globally, in attitudes toward categories of people, and is (2) differentially related to attitudes toward various categories of people. It was predicted that among heterosexual men, narcissism is linked to hostility toward heterosexual women more robustly than to hostility toward other groups (heterosexual men, lesbian women, and gay men). Narcissists’ low empathy, feelings of entitlement, and perceptions of being deprived of “deserved” admiration and gratification make them prone to aggression and vengeance (Brown 2004; Bushman and Baumeister 1998; Bushman et al. 2003; Bushman et al. 2009). Identifying the group(s) most likely to be targeted by narcissistic hostility could guide legislation, law enforcement efforts, and public health policy decisions regarding the allocation of finite resources for violence prevention and intervention. In the present investigation, hierarchical regression analyses were used to disentangle narcissism’s link to antipathy toward heterosexual women from its link to attitudes toward heterosexual gender roles more broadly, and from its link to attitudes toward groups commonly perceived as violating gender role norms (lesbian women and gay men). Hypotheses were derived from an integration of three theoretical models: the Agency Model of Narcissism (Campbell et al. 2006), Ambivalent Sexism Theory (Glick and Fiske 1996, 1999), and Implicit Inversion Theory (Kite and Deaux 1987). The present study relied upon a sample of U.S. undergraduate students, as did nearly all the empirical research reviewed, with notable exceptions of Abele and Wojciske’s (2007) sample of Polish college students, Foster and Campbell’s (2007) U.S. community sample, and Glick and Fiske’s (1996, 1999, 2001), along with their colleagues (2000, 2004), multinational samples of both college students and the general population. Findings from U.S. student samples may have limited generalizability, although research using broader samples, published in this journal and elsewhere, suggests that narcissism and patriarchal norms exist in other cultures across the world; for example, Glick et al. (2000) found consistent patterns and correlates of sexist ideologies across samples from 19 nations.

Narcissists have callous, adversarial, and hostile approaches to interpersonal relationships, as suggested by the Diagnostic and Statistical Manual of Mental Disorders’ (DSM) diagnostic criteria for Narcissistic Personality Disorder (NPD), which include arrogance, grandiosity, excessive need for admiration from others, feelings of entitlement to special privileges, and deficits in empathy (American Psychiatric Association [APA] 2000). In dating and marital relationships, narcissists use disingenuous “game-playing” behaviors intended to keep the upper hand and to avoid emotional intimacy (Campbell et al. 2002a). Instead of offering nurturance, narcissists feel entitled to unreciprocated gratification and unwarranted admiration (Campbell 1999; Foster et al. 2006). Rather than seeking soulmates or confidants, narcissists are more likely to pursue casual sex (Foster et al. 2006), to avoid commitment, to be unfaithful, and to “trade up” to “trophy” partners who are more impressive in appearance or status, compared to previous partners (Campbell et al. 2002a, b).

Heterosexual Women as Primary Targets of Narcissistic Hostility

It was predicted that heterosexual women, more than other groups (heterosexual men, lesbian women, gay men), bear the brunt of narcissistic hostility, given that men tend to be more narcissistic than women, and given that most men, being heterosexual, seek and expect several key types of gratification from heterosexual women. On the widely-used and well-validated Narcissistic Personality Inventory (NPI; Raskin and Terry 1988), men score significantly higher than women (Twenge et al. 2008). Nearly 75% of people diagnosed with NPD are men (APA 2000). Traditional gender norms may contribute to higher levels of narcissism in men while discouraging narcissism in women. As posited by the agency model of narcissism (Campbell et al. 2006), narcissists are more agentic (i.e., individualistically-oriented, self-promoting) than communal (i.e., interdependently-oriented; Helgeson 1994). Although the agency model does not focus explicitly on gender, other research suggests that traditional masculine norms prescribe that men should be agentic, pursuing self-benefiting goals such as amassing status, power, and competence, while devaluing, as being “feminine,” communal goals, such as joining with others to build and nurture emotionally intimate relationships. (Abele and Wojciszke 2007; Bradlee and Emmons 1992; Campbell et al. 2002b; Gabriel et al. 1994). Similarly, many cultures’ predominant gender scripts for heterosexual interactions prescribe proactive, agentic roles for men but reactive, receptive, communal roles for women (Abele 2003), justifying patriarchal expectations that women should subserviently gratify men and affirm men’s self-views of superiority.

More specifically, heterosexual women are likely to be the group most targeted by narcissistic hostility because they have unparalleled potential for gratifying, or frustrating, heterosexual men’s narcissism. Heterosexual men look to heterosexual women more than other groups (heterosexual men, gay men or lesbian women) to provide sexual pleasure and to tend to housework and other domestic activities (Glick and Fiske 1996), and to serve as “trophy” partners to impress others ( Campbell 1999). For these reasons, heterosexual women are key resources, and often gatekeepers, in heterosexual men’s quests for gratification, patriarchal power, and status. The promise of obtaining such rewards motivates heterosexual men to seek interactions and relationships with women, but brings risks of conflict, frustration, and disappointment as well (Glick and Fiske 1996). Although narcissists initially charm their ways to obtaining gratification, their insensitivity soon strains relationships (Paulhus 1998), and their incessant demands for gratification may backfire, making partners feel burdened or unappreciated. Egocentrism and impaired empathy lead narcissists to externalize blame for relationship conflicts, with the likely result being that heterosexual women, more than other groups, are positioned to be primary targets of men’s narcissistic anger and resentment.

Additionally, narcissists’ inflated but fragile grandiosity makes them prone to see other people as insufficiently respectful and obliging, leading narcissists to retaliate vengefully (Baumeister et al. 2000, Bushman and Baumeister 1998, Twenge and Campbell 2003). Psychoanalyst Heinz Kohut (1972) conceptualized “narcissistic rage” as a defense against deep-rooted shame reactivated when grandiose views of the self are threatened by feeling disrespected or deprived of gratification. People who feel rejected often restore positive self-views by derogating those who reject them (Bourgeois and Leary 2001). Although narcissists may express hostility toward anyone, even displacing it onto conveniently available substitute targets (Twenge and Campbell 2003), narcissistic affect is likely directed most intensely toward objects for which expectations are highest. Male narcissism has been linked to hostility toward specific individual female targets in the context of direct interpersonal interactions. For example, in responding to a female confederate who withheld gratification, narcissistic men doled out harsher punishment than did other men. (Bushman et al. 2003). Building upon such findings, the present research examined whether male narcissism is linked more broadly to hostility toward heterosexual women in general, even in the absence of proximal events likely to elicit men’s hostility, and examined whether heterosexual men’s narcissism is differentially related to attitudes toward various categories of people.

Patriarchy and Hostility Toward Women

According to Glick and Fiske’s (1996) theory of Ambivalent Sexism, men often have mixed emotions toward women, feeling frustrated by women and expressing hostility toward them, but alternately or even simultaneously experiencing and expressing warm emotions (e.g., desire, affection) toward them. Hostility is obviously harmful to women, but even some of men’s seemingly positive attitudes pose dangers, more insidiously, when they are paternalistic and patronizing, promising men’s approval and support only to women who conform to patriarchal norms of female subservience and submission. Glick and Fiske (1996) labeled these warm but constricting attitudes “Benevolent Sexism.” Men’s benevolence often is conditional, given with expectations that women will sacrifice and selflessly provide communal benefits (e.g., nurturance), domestic services, and sexual gratification to men, and with the belief that women are deficient in agency (e.g., competence and status) and so should accept being dependent upon men to be their protectors and providers (Fiske et al. 2002). Dominance over women is framed as chivalry, flattery, and protectiveness in order to appease women’s objections to subordination. Glick and Fiske (1996) developed the Ambivalent Sexism Inventory (ASI), which has subscales to assess Hostility toward Women (also known as “Hostile Sexism”) and Benevolence toward Women (also known as “Benevolent Sexism”).

Glick and Fiske (1999) extended their theory of Ambivalent Sexism by developing the Ambivalence toward Men Inventory (AMI), which, like the ASI, measures hostile and benevolent attitudes, but toward men rather than women. The Hostility toward Men subscale measures beliefs that men lack attributes in stereotypically feminine domains (e.g., nurturance, domestic skills), but deserve higher status and power over women. The Benevolence toward Men subscale assesses a respondent’s agreement with a favorably-framed, although narrowly agentic, view of masculinity that reinforces patriarchal power. The AMI was designed to assess women’s attitudes toward men, but has been administered to men as well; its factor structure and predictive validity for men was comparable to that for women (Glick and Fiske 1999).

Although hostile and benevolent attitudes have opposite emotional tones, their functions are the same; each prescribes polarization of gender roles and justifies men’s dominance over women. Furthermore, attitudes (both hostile and benevolent) toward men and toward women are complementary, being two integrated sides of a traditional, patriarchal model of heterosexual relationships. Thus, all four attitudes prescribe communal, subservient roles for women and give men license to be agentic, enabling them to amass status and power over others. This traditional gender ideology prescribes that the ideal relationship is a heterosexual one between an agentic man and a communal woman. The complementary nature of all four attitudes is supported by findings that men’s hostile and benevolent attitudes toward women are correlated with their hostile and benevolent attitudes toward men (Glick and Fiske 2001; Glick et al. 2004).

Disentangling Narcissism’s Links to Gender Ideology and Patriarchy

The present study further analyzes the link between men’s narcissism and traditional gender-role orientations (e.g., Hurlbert et al. 1994), dissecting narcissism’s differential relationships with subcomponents of attitudes toward heterosexual relationships (hostile and benevolent attitudes toward women and hostile and benevolent attitudes toward men) and contrasting each to narcissism’s relationship with attitudes toward groups perceived as violating traditional gender norms (gay men and lesbian women; Kite and Deaux 1987; Madon 1997). Each group (heterosexual women, heterosexual men, lesbian women, gay men) represents a different combination of points along two dimensions: the group’s perceived conformity to traditional gender roles and the group’s potential for gratifying a heterosexual man. Male narcissism’s unique relationship to attitudes toward each group was examined in a series of correlational and hierarchical regression analyses aimed at teasing apart these two dimensions. These analyses were used to evaluate hypotheses that heterosexual men’s narcissism is differentially related to hostility toward each category of people.

It was predicted that although narcissists show adversarial approaches toward people in general, heterosexual male narcissism is associated with hostility aimed primarily at heterosexual women, rather than with diffuse hostility expressed equally toward all four groups of people, given that: men are more narcissistic than women; most men are heterosexual; narcissists use relationships for self-benefiting rather than for interdependent goals; gender norms prescribe that men are entitled to subservience and gratification from women; and people attack those whom they perceive as depriving them of gratification to which they feel entitled. In contrast, it was predicted that heterosexual men’s narcissism is related to flattering views of their own group, heterosexual men (as measured by the Benevolence toward Men subscale of the AMI), reflecting a self-serving bias of blaming heterosexual conflicts on women more than on men.

The link of heterosexual men’s narcissism with conservative attitudes toward gender roles was predicted to be limited to heterosexual relationships, rather than to extend to attitudes toward sexual orientation. Even though heterosexual men tend to show antipathy toward homosexuality (Kite and Whitley 1996, 1998), they likely expect less of and have had fewer meaningful relationships with gay men or lesbian women than with heterosexual women. No investigations of narcissism and attitudes toward homosexuality were found, but it was predicted that heterosexual male narcissism is unrelated to attitudes toward gay men, for they are the group least relevant to a heterosexual man’s pursuit of gratification. It was additionally predicted that male narcissism is associated with favorable attitudes, not hostility, toward lesbian women. Some men find lesbian women gratifying sexually, voyeuristically or in fantasy, which contributes to men’s more favorable attitudes toward them than toward gay men (Louderback and Whitley 1997; Whitley et al. 1999); it was predicted that narcissistic men are particularly likely to show egocentrically hedonistic views of lesbian women.

Predictions that male narcissism is associated with hostility toward heterosexual women more than toward other groups were evaluated against two competing possibilities: the Nonspecific Derogation of Any Group Hypothesis and the Narcissism as a Marker of Conservative Gender Ideology Hypothesis. The Nonspecific Derogation of Any Group Hypothesis would be supported by findings that men’s narcissism is associated equally with antipathy toward all groups, regardless of the group members’ sexual orientations or gender. Perhaps narcissistic men perceive threats from men as much as from women, or perceive threats in varied contexts, such as the work place or sports field, as much as in romantic relationships. Or, even if heterosexual relationships are the source of their deepest frustrations and disappointments, narcissistic heterosexual men may displace resentment and anger onto any available target. Narcissists experiencing rejection not only retaliate toward the individual causing narcissistic injury, but also displace aggression onto innocent bystanders who did not reject them (Twenge and Campbell 2003). Just as narcissistic hostility may be displaced onto any available individual target, it may be displaced non-specifically on the group level as well.

The Narcissism as a Marker of Conservative Gender Ideology Hypothesis would be supported by findings that male narcissism is associated with hostility toward any group perceived as threatening polarization of gender roles along a mutually-exclusive dichotomy of masculine versus feminine. Many people, particularly those invested in traditional gender ideology, view non-heterosexuals as gender inverts, believing that gay men resemble heterosexual women not only in feeling sexual attraction toward men, but also in possessing feminine interests, skills, and behaviors; further, many people believe that lesbian women are also “cross-gendered,” being masculine in all domains (Kite and Deaux 1987). Traditional, patriarchal beliefs are linked to unfavorable attitudes toward homosexuality in general (Whitley 2001) and specifically both toward gay men (Barron et al. 2008; Davies 2004; Parrott et al. 2002) and lesbian women (Wilkinson 2006, 2008). For a heterosexual man, subordinating outgroups, whether heterosexual women or gender norm “violators,” could function to unite with other men in asserting heterosexual male privilege and status. If narcissism is merely a marker of investment in conservative gender ideology, then higher narcissism would be associated not only with endorsement of traditional roles for women and men in heterosexual relationships (i.e., Hostility toward Women, Benevolence toward Women, Hostility toward Men, and Benevolence toward Men) but also with disapproval of homosexuality, evidenced by antipathy toward lesbian women and toward gay men.

Summary of Hypotheses

  1. H1:

    Heterosexual men’s narcissism is associated with patriarchal attitudes toward heterosexual relationships, as evidenced by positive correlations between narcissism and each of the ASI and AMI’s subscales (Hostility toward Women, Benevolence toward Women, Hostility toward Men, and Benevolence toward Men).

     
  2. H2:

    Hostility toward Women accounts for variance in narcissism beyond that accounted for by both Benevolence toward Women and Hostility toward Men.

     
  3. H3:

    Given that narcissists likely have flattering views of their own group, Benevolence toward Men accounts for narcissism variance beyond that accounted for by both Benevolence toward Women and Hostility toward Men.

     
  4. H4:

    Heterosexual men’s narcissism is correlated more strongly with hostility toward heterosexual women than toward gay men and lesbian women.

     
  5. H5:

    Heterosexual men’s narcissism is so robustly associated with antipathy toward heterosexual women that Hostility toward Women accounts for narcissism variance beyond that accounted for by men’s attitudes toward all the other groups (heterosexual men, gay men, and lesbian women) and by Benevolence toward Women.

     

Method

Participants

A total of 113 male undergraduates were recruited from General Psychology classes at regional campuses of a large university in the Midwest U.S.. Three men who self-reported a sexual orientation other than completely heterosexual (on a 5-point scale, with response choices of “Completely Gay or Lesbian,” “Mostly Gay or Lesbian,” “Bisexual,” “Mostly Heterosexual,” and “Completely Heterosexual”) were excluded from the analyses, as the hypotheses pertain to men’s views of women as sources of gratification or frustration in heterosexual relationships. Of the remaining 110 participants, six were removed from the sample for providing blatantly invalid responses, operationalized as inconsistent responses on many item-pairs with semantically matching content and correlations exceeding r = .60 in the sample overall. The final sample consisted of 104 men (M age = 21 years); a few (two to four) additional participants were excluded from testing of particular hypotheses only if they were missing data on the items comprising scales pertaining to the hypothesis. Most of the sample (95%) identified as Caucasian; the rest identified as African American (2%), Hispanic (1%), or “other” (3%). All participants were at least 18 years old; 33 % were 18 to 19, 45% were 20 to 24, 9% were 25 to 29, 5% were 30 to 34, 4% were 35–39, and 4% were 40 or older.

Procedure

After securing informed consent, a female research assistant provided survey packets and instructions to groups ranging from 15 to 45 participants. Both male and female participants were present at the sessions, but only data from male participants were analyzed in the present study. Participants were instructed to refrain from putting their names on any materials in order to maintain their anonymity. Participants received partial course credit for their time.

Measures

Narcissism

Narcissism was measured with the Narcissistic Personality Inventory-16 (NPI-16; Ames et al. 2006), a short form of the widely used 40-item Narcissistic Personality Inventory (NPI-40; Raskin and Terry 1988), which has been found to be a valid measure of narcissism in non-clinical populations (Morf and Rhodewalt 2001). The NPI-16 was utilized in order to minimize respondents’ fatigue and careless responding, as the measures analyzed in the present study were part of a longer packet of measures. Ames et al. (2006) demonstrated the validity of the NPI-16 in several ways. First, the NPI-16 was highly correlated (r = .90) with the NPI-40 and with measures to which the NPI-40 is related (i.e., measures of self-esteem and the Big Five personality traits). Further, NPI-16 scores were associated with grandiose self-evaluations in both group decision-making tasks and individual judgment tasks; NPI-16 scores were associated with overly favorable self-appraisals of performance (which were unrelated to actual quality of performance), effort, creativity, power and attractiveness. Finally, scores on the NPI-16 (M = .35) resembled those on the NPI-40 (M = .39), and men scored higher than women on both versions. The abbreviated length of the NPI-16 does not allow for scoring of the NPI-40’s subscales, but recent research has cast doubt on the validity of most of the subscales; only the full scale narcissism score and one of NPI-40’s seven subscales had adequate internal consistency (α > .70; del Rosario and White 2005). In the present study’s sample, participants provided generally consistent responses to the NPI, as suggested by a Cronbach’s alpha (α = .76) that was comparable to that reported in past research for the NPI-16 (α = .72; Ames et al. 2006) and for the NPI-40 (α = .80 to .82; del Rosario and White 2005). The NPI-16’s format is like that of the NPI-40; both consist of pairs of items from which respondents choose one item as being more descriptive of themselves. Narcissism scores were calculated as recommended by Ames et al. (2006); the sum of the number of times a participant endorsed the narcissistic item of each pair was divided by 16 (the number of items), resulting in a score (ranging from 0.00 to 1.00) indicating the percentage of items for which the participant chose the narcissistic self-description.

Attitudes Toward Women and Men

Sexist attitudes toward women and traditional female stereotypes were assessed with the Hostile Sexism and Benevolent Sexism subscales of the ASI (ASI; Glick and Fiske 1996), consisting of 22 items, half assessing Hostile Sexism (sample item: “Once a woman gets a man to commit to her, she usually tries to put him on a tight leash.”) and half assessing Benevolent Sexism (sample items: “Women should be cherished and protected by men.” and “Women, as compared to men, tend to have a more refined sense of culture and good taste.”). In this paper, Hostile Sexism is referred to as “Hostility toward Women” and Benevolent Sexism is referred to as “Benevolence toward Women” in keeping with Glick et al.’s (2004) suggestion that they are mirror images of the complementary attitudes toward men measured by the AMI. The Hostility toward Women subscale assesses directly-expressed antipathy toward women and unfavorable feminine stereotypes. The Benevolence toward Women subscale assesses agreement with positively-framed but confining and paternalistic views of women. Respondents indicate the extent of their agreement with each item on a six-point scale ranging from “Disagree Strongly” to “Agree Strongly.” Scores were calculated by reverse-scoring items as necessary, then computing the average of the participants’ responses to each item of the scale. In the present sample, Cronbach’s alphas for responses to items of the Hostile Sexism scale (α = .80) and of the BS scale (α = .75) were comparable to those reported in prior research: in the .70 to .90 range for Hostile Sexism and the .70 to .85 range for Benevolent Sexism (Glick and Fiske 1996; Glick et al. 2000).

Sexist attitudes toward men and heterosexual male stereotypes were assessed with the Hostility toward Men subscale (sample item: “Men pay lip service to equality, but can’t handle it.”) and the Benevolence toward Men subscale (sample item: “Men are less likely to fall apart in emergencies.”) of the AMI (Glick and Fiske 1999). Each AMI subscale consists of 10 items, and the item format and scoring parallels that of the ASI. In the present sample, Cronbach’s alphas for responses to the items of the Hostility toward Men scale (α = .73) and for the Benevolence toward Men scale (α = .83) were comparable to those reported in prior research: ranging from .81 to .86. for Hostility toward Men and from .79 to .83 for Benevolence toward Men (Glick and Fiske 1996; Glick et al. 2000).

Attitudes Toward Gay Men

The 5-item version of Herek’s (1988, 1998) Attitudes Toward Gay Men Scale, Revised Version (ATG) was used to assess respondents’ attitudes toward gay men. Herek (1988) reported that the 5-item version of the ATG correlated very highly (r = .96) with the longer 10-item version. Each item (e.g., “Male homosexuality is a perversion.”) is rated on a 7-point scale, ranging from “Strongly Disagree” to “Strongly Agree.” Higher scores indicate more unfavorable attitudes toward gay men. ATG scores were calculated by reverse scoring items when necessary, then summing participants’ responses to each item. In the present sample, the Cronbach’s alpha for responses to the items of the ATG (α = .87) was comparable to that of prior research (α = 87; Herek 1988).

Attitudes Toward Lesbian Women

The five items of Herek’s (1988) Attitudes Toward Gay Men (ATG) scale were modified in the present research, replacing the words “gay men” with the words “lesbian women,” the words “male homosexuality” with “female homosexuality,” and the word “men” with “women” in order to create a measure of attitudes toward lesbian women (ATL) parallel to the measure of attitudes toward gay men. In the full Attitudes Toward Lesbians and Gay Men Scale (ATLGM; Herek 1988), the items assessing attitudes toward gay men have different content than the items assessing attitudes toward lesbians. As Herek (1998) cautions:

“Researchers wishing to compare a subject population’s attitudes toward gay men with their attitudes toward lesbians are advised to use parallel forms of one of the subscales (the ATG items have usually been used for this purpose), with each item presented twice, once in reference to gay men and once in reference to lesbians” (1998, p. 392).

As with the ATG, respondents rated each item of the ATL using a 7-point scale, ranging from “Strongly Disagree” to “Strongly Agree.” Higher scores indicate unfavorable attitudes toward lesbian women. ATL scores were calculated by reverse scoring items when necessary, then summing participants’ responses to each item. In the present sample, the Cronbach’s alpha for responses to the items of the ATL (α = .85) was comparable to those found in various versions of the ATL and ATG as reported by Herek (1988, 1998), with typical Cronbach’s alphas in the mid to high .80s.

Results

Bivariate correlation coefficients, means, standard deviations for all variables are shown in Table 1. Of the AMI and ASI scales, the sample’s mean scores were highest on two scales: Hostility toward Women and Benevolence toward Women (measuring “benevolently,” yet restrictively sexist attitudes toward women), and these mean scores did not differ significantly from each other (using a Bonferroni-corrected alpha level of p < .05 divided by seven t-tests = p < .007), but were each significantly higher than the sample’s mean scores on the two remaining scales measuring attitudes toward heterosexual gender roles, specifically, sexist and stereotyped views of men’s gender roles (Hostility toward Men or Benevolence toward Men), as shown in the t-test results displayed in Table 2. Thus, of the four subscales of the ASI and AMI, men scored highest, on average, on both Hostility toward Women and Benevolence toward Women, then significantly lower on Benevolence toward Men, and significantly lower still on Hostility toward Men. Also notable, and consistent with past research (Kite and Whitley 1996, 1998), men showed considerably stronger antipathy toward gay men than toward lesbian women, as shown by the final t-test presented in Table 2.
Table 1

Bivariate correlations, means, and standard deviations of study variables.

Variable

NPI

HW

BW

HM

BM

ATG

ATL

NPI

.31 (.21)

.43***

.27**

.25**

.37***

.06

−.24*

HW

 

3.88 (.91)

.35***

.49***

.58***

.37***

.06

BW

  

3.63 (.90)

.33**

.63***

.26**

−.03

HM

   

2.88 (.73)

.61***

.30**

.10

BM

    

3.34 (.90)

.33**

.06

ATG

     

23.42 (8.68)

.69***

ATL

      

17.62 (7.63)

NPI = Narcissism. HW = Hostility toward Women. BW = Benevolence toward Women. HM = Hostility toward Men. BM = Benevolence toward Men. ATG = (Hostile) Attitudes toward Gay Men. ATL = (Hostile) Attitudes toward Lesbians. Higher scores on all variables represent higher levels of the construct. Means and standard deviations are displayed on the diagonal. Possible range of scores is 0.00 to 1.00 for NPI, 1 to 6 for HW, BW, HM, and BM, and 5 to 35 for ATG and ATL

*p < .05. **p < .01. ***p < .001. All two-tailed tests

Table 2

t-tests of mean differences of selected study variables.

Variables

M (SD)

Mean difference

t

p

Cohen’s d

HW

3.88 (.91)

.24537

2.417

.017

.238

BW

3.63 (.90)

    

HM

2.88 (.73)

−.46311

−6.335

.000

−.624

BM

3.34 (.90)

    

HW

3.88 (.91)

.99647

11.950

.000

1.177

HM

2.88 (.73)

    

HW

3.88 (.91)

.53806

6.472

.000

.640

BM

3.34 (.90)

    

BW

3.63 (.90)

.75044

8.066

.000

.790

HM

2.88 (.73)

    

BW

3.63 (.90)

.31642

4.220

.000

.416

BM

3.34 (.90)

    

ATG

23.42 (8.68)

5.798

9.096

.000

.892

ATL

17.63 (7.63)

    

HW = Hostility toward Women. BW = Benevolence toward Women. HM = Hostility toward Men. BM = Benevolence toward Men. ATG = (Hostile) Attitudes toward Gay Men. ATL = (Hostile) Attitudes toward Lesbians. Higher scores on all variables represent higher levels of the construct. Possible range of scores is 1 to 6 for HW, BW, HM, and BM, and is 5 to 35 for ATG and ATL

All two-tailed tests

H1, the prediction that heterosexual men’s narcissism is positively correlated with all four subscales of the ASI and AMI, was supported. Men’s narcissism showed nearly uniform moderate zero-order correlations (displayed in Table 1) with hostile and “benevolently” stereotyped attitudes toward women and men, suggesting that men’s narcissism was associated with beliefs that women and men should conform to traditional gender roles, and that heterosexual relationships should be patriarchal rather than egalitarian.

For the next hypothesis, and others tested using hierarchical regression analysis, the data met the assumptions of multiple regression; there were no indications of multicollinearlity, as tolerance statistics for all variables were >.38 and Variance Inflation Factors were all <2.6. H2, the prediction that Hostility toward Women accounts for variance in narcissism beyond that accounted for by Benevolence toward Women and Hostility toward Men, was supported by the hierarchical regression analysis results shown in Table 3. Hostility toward Women accounted for a statistically significant amount of variance (9.9%) in narcissism beyond that attributable to Benevolence toward Women and Hostility toward Men (10%). Further, the final beta weights displayed in Table 3 suggest that only Hostility toward Women accounted for a significant amount of unique variance in narcissism; Benevolence toward Women and Hostility toward Men were weaker correlates of narcissism and were redundant with Hostility toward Women.
Table 3

Hierarchical regression analysis testing hypothesis 2.

Predictor Variable

ra

semi-partial r

R2change (%)

R2

R2adj

Final β

Fchange

df1,df2

Step 1: Benevolence toward Women

.27**

.12

10.1**

.101

.083

.128

5.60**

2,100

 Hostility toward Men

.25**

.03

   

.029

  

Step 2: Hostility toward Women

.43***

.31

9.9**

.199

.175

.369**

12.19**

1,99

All variables were scored so that higher numbers represent higher amounts of the construct

aBivariate correlation with Narcissism

*p < .05. **p < .01. ***p < .001. All two-tailed tests

H3, the hypothesis that Benevolence toward Men accounts for variance in men’s narcissism beyond that accounted for by both Benevolence toward Women and Hostility toward Men, was supported by the results of the hierarchical regression displayed in Table 4.
Table 4

Hierarchical regression analysis testing hypothesis 3.

Predictor Variable

ra

semi-partial r

R2change (%)

R2

R2adj

Final β

Fchange

df1,df2

Step 1: Benevolence toward Women

.25**

.03

9.9**

.099

.081

.036

5.44**

2,99

 Hostility toward Men

.27**

.06

   

.072

  

Step 2: Benevolence toward Men

.37***

.21

4.1*

.140

.113

.302*

4.65*

1,98

All variables were scored so that higher numbers represent higher amounts of the construct

aBivariate correlation with Narcissism

*p < .05. **p < .01. ***p < .001. All two-tailed tests

Benevolence toward Men accounted for a statistically significant amount of variance (4%) in narcissism beyond that accounted for by Benevolence toward Women and Hostility toward Men (9.9%). Further, of the three predictor variables (Hostility toward Women had not been entered in this regression analysis), only Benevolence toward Men had a beta weight that was that was statistically significant, whereas Benevolence toward Women and Hostility toward Men were weaker correlates of narcissism and the variance they shared with narcissism was largely redundant the narcissism variance accounted for by Benevolence toward Men, as shown in Table 4.

H4, predicting that heterosexual men’s narcissism is correlated more strongly with antipathy toward heterosexual women than with antipathy toward non-heterosexual people, was supported. Male narcissism was not significantly related to attitudes toward gay men (r = .06, ns); this correlation coefficient was significantly smaller than that of narcissism with Hostility toward Women (r = .43, p < .001), r1- r2 = −0.37, Z(101) = −3.43, p < .01, and it was smaller than even narcissism’s weakest correlation with any of the four heterosexual attitudes variables, Hostility toward Men (r = .25, p < .01), r1- r2 = 0.198, Z(101) = −1.71, p < .05. In contrast, narcissism was negatively correlated with attitudes toward lesbian women (meaning that men who had the most favorable attitudes toward lesbian women tended to be the most narcissistic; r = −.24, p < .05), and that correlation was significantly different from narcissism’s correlation with hostility toward heterosexual women (r = .43, p < .001), r1-r2 = .67, Z(101) = −4.94, p < .01. Also statistically significant was the difference between narcissism’s correlation with attitudes toward lesbian women (r = −.24, p < .05) and its non-significant correlation with attitudes toward gay men (r = .06, ns), r1- r2 = 0.18, Z(101) = −2.43, p < .01. Thus, for men in the present sample, higher narcissism was related to discrepant attitudes toward different types of women (antipathy toward heterosexual women, favorable attitudes toward lesbian women) and was unrelated to attitudes toward gay men.

H5, the prediction that Hostility toward Women accounts for narcissism variance beyond that accounted for by men’s attitudes toward all the other groups (heterosexual men, gay men, and lesbian women) and by Benevolence toward Women, was supported by the hierarchical regression results displayed in Table 5. Hostility toward Women accounted for a significant amount of variance (5.3%) in men’s narcissism even beyond the considerable variance (24%) accounted for by the other five attitude variables. Moreover, of the final beta weights for the six predictors, only two were statistically significant; as predicted, narcissism was positively correlated with Hostility toward Women, but was negatively correlated with Attitudes toward Lesbian Women (i.e., again, higher narcissism tended to be associated with hostility toward heterosexual women but with favorable attitudes toward lesbian women).
Table 5

Hierarchical regression analysis testing hypothesis 5.

Predictor Variable

ra

semi-partial r

R2change (%)

R2

R2adj

Final β

Fchange

df1,df2

Step 1: Hostility toward Men.

.27**

.01

24.0***

.240

.199

.015

5.98***

5,95

 Benevolence toward Men

.37***

.12

   

.191

  

 Benevolence toward Women

.25**

−.03

   

−.032

  

 Attitudes Toward Gay Men

.07

.08

   

.128

  

 Attitudes Toward Lesbians

−.23*

−.25

   

−.356**

  

Step 2: Hostility toward Women

.44***

.23

5.3**

.292

.247

.304**

7.01**

1,94

All variables were scored so that higher numbers represent higher amounts of the construct

aBivariate correlation with Narcissism

*p < .05. **p < .01. ***p < .001. All two-tailed tests

Discussion

The present results align most closely with the prediction that men’s narcissism is associated with hostility targeting heterosexual women most of all; the results fail to support the competing hypotheses (the Nonspecific Derogation of Any Group Hypothesis and the Narcissism as a Marker of Conservative Gender Ideology Hypothesis). In the present sample, narcissism in heterosexual men was associated with patriarchal orientations toward heterosexual relationships, as measured by the ASI and AMI. When attitudes toward traditional gender roles were dissected into directly-expressed antipathy versus benevolently-sexist attitudes, and into attitudes toward women versus attitudes toward men, and then pitted against each other as predictors of men’s narcissism, Hostility toward Women emerged (as predicted) as the most robust predictor of narcissism. Consistently, in each hierarchical regression analysis, Hostility toward Women accounted for statistically significant amounts of unique variance in narcissism, whereas the other ASI and AMI subscales, in isolation, accounted for no unique variance in narcissism (see the beta weights of Tables 3 and 5). Only when Hostility toward Women was excluded did another ASI or AMI subscale emerge as a statistically significant predictor, by itself, of unique variance in narcissism; the only subscale to do so was, as predicted, Benevolence toward Men, as indicated by its beta weight (see Table 4) in the regression analysis that omitted Hostility toward Women. Further supporting the prediction that men’s narcissism is associated with hostility toward heterosexual women most of all were the results that higher narcissism was not linked to derogation of gay men or of lesbian women. Thus, in the present sample, heterosexual male narcissism’s association with attitudes toward traditional heterosexual gender roles was accounted for primarily by men’s resentment and anger toward heterosexual women.

The Nonspecific Derogation of Any Group Hypothesis was unsupported by the present results; men’s narcissism was associated with discrepant rather than equivalent evaluations of the target groups. The beta weights displayed in Tables 3, 4, and 5 suggest that Hostility Toward Men did not account for significant unique variance in narcissism. Moreover, narcissism was not associated with derogation of non-heterosexuals. In contrast, Hostility toward Women was a significant predictor of narcissism in every hierarchical regression analysis, even accounting for unique variance in narcissism when entered last after all the other ASI and AMI subscales and after attitudes toward gay men and lesbian women. Thus, narcissism was not a non-specific defense of inflating one’s self-evaluation by derogating any group, but instead was associated primarily with derogation of heterosexual women.

The Conservative Gender Ideology Hypothesis, too, was unsupported. Narcissism was not associated with antipathy toward non-heterosexuals, a group frequently perceived as gender-role violators (Kite and Deaux 1987; Madon 1997). Narcissism showed opposite correlations with attitudes toward each category of women (positively correlated with hostility toward heterosexual women but negatively correlated with hostility toward lesbian women). One interpretation of this finding is that, compared to other men, narcissistic men are more invested in patriarchal control over women, yet are less likely to perceive lesbian women as threats to heterosexual male status and power. Further, compared to other men, narcissistic men may feel more entitled to objectify a lesbian woman for one-sided gratification, at least voyeuristically or as fantasy objects, whereas less narcissistic men may feel more inhibited by guilt or empathic concerns about causing offense. Male narcissism was not, however, associated with greater acceptance of homosexuality in general. Male narcissists’ favorability toward lesbian women was not extended to gay men, as suggested by the lack of a significant correlation between narcissism and attitudes toward gay men. In total, the present results suggest that narcissistic hostility is associated with a group’s potential to provide or withhold gratification rather than ideology about a group’s sexual orientation or conformity to heterosexual gender roles. Future research could test these interpretations directly by assessing respondents’ perceptions of non-heterosexuals’ gender role conformity, and by assessing whether male narcissism is associated with egocentric and androcentric views of lesbian women as sexual objects.

One notable finding beyond the a priori hypotheses is that for all men in the sample, regardless of level of narcissism, investment in traditional gender roles was related differently to attitudes toward each target group of non-heterosexuals. Traditional gender roles were unrelated to men’s attitudes toward lesbian women; none of the four AMI and ASI subscales were correlated with attitudes toward lesbian women (see Table 1). In contrast, men’s investment in traditional gender roles was linked to harsher attitudes toward gay men, as indicated by the results that each of the ASI and AMI subscales was significantly positively correlated with attitudes toward gay men. One interpretation of this difference is that it adds support to the possibility that the heterosexual men in this sample who are strongly invested in traditional, patriarchal relationships consider gay men, but not lesbian women, to be threats to patriarchy. Although the present results may seem to contradict Wilkinson’s (2008) finding that people view lesbian women as a threat to the patriarchy, that author’s result was obtained from a sample that was nearly 70% female. Perhaps respondent gender is a moderating variable, such that female respondents view lesbian women as a threat to patriarchy (see also Wilkinson 2006), whereas (consistent with the present results) male respondents do not. Alternately, it may be that traditional men tend to find gay men aversive, but are no more or less likely (compared to other men) to find lesbian women aversive or appealing.

Limitations and Future Directions

The present study relied solely upon self-report measures, so its results may be distorted by response biases such as impression management or self-deception (particularly likely in narcissistic participants; Paulhus 1991). Although anonymous testing conditions minimize impression management distortions, self-deception biases remain, as all measures assessed self-perceptions of personality traits and attitudes. This study falls short of providing evidence that narcissism is associated with hostile behaviors toward women and other groups, and behaviors not infrequently differ from attitudes and self-reports. Future research could extend these findings by using behavioral measures, implicit measures of attitudes, or measures completed by informants other than the respondent, such as significant others or acquaintances. A related potential limitation is that the measures had been administered in a group format, which may have induced normative pressures more strongly than if participants had completed measures in solitary settings. No literature was found regarding the effects of group versus solo testing on the measures used in the present study, but the topic is worthy of future inquiry.

Another limitation of the present study is the homogeneous nature of its sample, which was predominantly Caucasian, young (M age = 21 years), and of a contemporary cohort that is more narcissistic than past generations (Twenge et al. 2008). Self-reported narcissism was lowest in older participants of a multi-national sample of over 3,000 respondents ranging from childhood to age 83 (Foster et al. 2003). On the other hand, the present sample underrepresented the high end of the narcissism continuum. Respondents in the present sample scored mostly in the low, subclinical end of the continuum of narcissism, and consequently are unlikely to meet diagnostic criteria for NPD. In the present sample of college undergraduates, NPD may have been underrepresented to the extent that narcissistic traits interfere with motivation or ability (e.g., delaying gratification) to succeed in college.

Although the present findings are compatible with the hypothesis that male narcissism causes hostility toward women, the cross-sectional and correlational nature of the present study limits inferences about causality. Given the very early origins of personality development, it is likely that narcissism precedes the development of attitudes toward potential romantic partners. Classic theories of narcissism posit that it is rooted in infancy or childhood (e.g., Sandler et al. 1991; Kohut and Wolf 1978). However, causality is likely reciprocal, multidirectional, and transactional. Narcissists may choose partners who are narcissistic (i.e., superficially impressive, success-oriented trophies who are shallow and self-promoting themselves) thus reinforcing and strengthening each other’s narcissistic and agentic approaches to relationships. Moreover, people who behave narcissistically in relationships are increasingly likely to elicit rejection over time, as partners come to see through their charming façades and initially-impressive bravado. People facing rejections may, in turn, react to and defend against relationship disappointments by becoming even more narcissistic. Other variables also likely play causal roles. For example, boys may learn hostile stereotypes of women (e.g., from media or from older males) first, which then cause them to have unrealistic expectations of female partners or cause them to behave defensively or selfishly toward women, which in turn elicits rejection from women, which exacerbates narcissistic defenses, and so on. Exposure to presentations of hostile stereotypes of women may encourage some boys and men to adopt cavalier or adversarial approaches toward real women. Future research could explore causal directions by manipulating a suspected cause (e.g., exposure to hostile stereotypes of women, or priming such stereotypes) of increased narcissism. Future research could alternately test causality between narcissism and hostility toward various groups by using a competitive aggression paradigm (Zeichner et al. 1999) and manipulating the “gratification potential” of the target, telling the participant that the supposed opponent is either a heterosexual women, a lesbian woman, or a man, while holding constant the level of insult to test whether participants are differentially aggressive toward the targets.

In sum, the present study suggests that heterosexual men’s narcissism is linked to an adversarial and angry stance toward heterosexual women more than toward other groups. Although narcissists may want to maintain feelings of superiority and power over all people, narcissistic heterosexual men are particularly invested in subordinating heterosexual women. This study’s results additionally suggest that neither lesbian women nor homosexuality in general are perceived as threats to a heterosexual man’s narcissism; to the contrary, narcissism is related to favorable attitudes toward lesbian women and is unrelated to antipathy toward gay men.

Acknowledgement

The author thanks David DeLong and two anonymous reviewers for providing helpful feedback on previous drafts of this paper; Lynzee Abel and Erica Salisbury for assistance with data collection; and Julie Cremeans-Smith, Lee Fox-Cardamone, Deborah A. Jones, Robin Lashley, and Brad Shepherd for permitting the recruitment of research participants from their classes.

Copyright information

© Springer Science+Business Media, LLC 2010