Essay
Journal of Social and Clinical Psychology, Vol. 33, No. 1, 2014, pp. 25-50
25
© 2014 Guilford Publications, Inc.
Address correspondence to Peggy S. Keller, Ph.D., 012i Kastle Hall, Department of Psychology, University of Kentucky, Lexington, KY 40506; E-mail: peggy.keller@uky.edu
NARCISSISM IN ROMANTIC RELATIONSHIPS KELLER ET AL.
NARCISSISM IN ROMANTIC RELATIONSHIPS: A DYADIC PERSPECTIVE
PEGGY S. KELLER University of Kentucky
SARAI BLINCOE Longwood University
LAUREN R. GILBERT, C. NATHAN DEWALL, ERIC A. HAAK, AND THOMAS WIDIGER University of Kentucky
Previous literature has associated narcissism with interpersonal difficulties. How- ever, there is a lack of studies investigating the impact of narcissism within ro- mantic dyads. The current study extended the literature by examining relations between narcissism and conflict behaviors in both members of a romantic dyad. Participants in the study were 190 college-student couples, who completed ques- tionnaire measures of their narcissism, conflict within their relationship, and a behavioral measure of aggression toward their partners. The results indicated a tendency towards homogamy for narcissism. A person’s narcissism was related to their use of aggression and the partner’s use of aggression. Several interactions between male and female narcissism were observed. For total NPI scores, the combination of high male and female narcissism was related to greater aggres- sion. Grandiose Exhibitionism was only related to aggressive behavior when part- ner Grandiose Exhibitionism was lower; when partner Grandiose Exhibitionism was higher, aggressive behavior was generally higher regardless of a person’s own Grandiose Exhibitionism.
26 KELLER ET AL.
Although there is debate among researchers about the potential beneficial or detrimental effects of narcissism for mental health and distress (Cain, Pincus, & Ansell, 2008; Campbell, 2001), researchers generally concur that narcissism is associated with interpersonal difficulties (e.g., Campbell & Foster, 2002). However, the majority of studies on narcissism within interpersonal relationships have focused on the individual. The present investigation takes a dy- adic perspective, answering questions about associations between partners’ narcissism and relations between narcissism and conflict behaviors. Narcissistic personality disorder is relatively rare in the general population, but narcissism falls along a continuum and nar- cissistic traits in non-clinical samples are consistently related to in- terpersonal problems (Miller & Campbell, 2008). We use the term narcissist to indicate persons at the higher end of the spectrum rela- tive to others.
Narcissists do not make warm, caring partners. Narcissism is characterized by feelings of superiority and entitlement, expecta- tions for special treatment, exaggerations of abilities and personal qualities, demands for attention and admiration, exploitation of others, contempt, and lack of empathy (American Psychiatric As- sociation, 2000). Narcissists idealize their romantic partners at first, but when imperfections become apparent they develop feelings of contempt (Masterson, 1988). Narcissists believe that they have greater alternatives for romantic partners than non-narcissists, and this leads them to be less committed to their relationships (Camp- bell & Foster, 2002). Individuals higher in narcissism report greater vindictiveness, domineering and controlling approaches to others, and intrusiveness in their relationships (e.g., Ogrodniczuk, Piper, Joyce, Steinberg, & Duggal, 2009).
Narcissism is also associated with aggression and hostility. Ac- cording to the dynamic self-regulatory processing model (Morf, Horvath, & Torchetti, 2010), narcissism is a personality process in which self-regulatory strategies are designed to build and maintain a grandiose self-view. Interpersonal interactions are adversarial and narcissists may respond to threats to their inflated self-view with hostility, rage, or physical aggression. For example, narcissists pro- vide negative evaluations of experimenters giving them negative performance feedback (Stucke & Sporer, 2002). They are angrier than non-narcissists after social rejection, and they direct their an- ger at those who reject them as well as innocent parties (Twenge & Campbell, 2003). However, narcissists do not need to be provoked in
NARCISSISM IN ROMANTIC RELATIONSHIPS 27
order to become aggressive. Stronger relations between narcissism and proactive aggression than between narcissism and reactive ag- gression have been reported (Reidy, Foster, & Zeichner, 2010). The little research that has examined narcissism and romantic relation- ships has found that narcissists are more likely to engage in marital violence and sexual coercion (Baumeister, Smart, & Boden, 1996).
Despite the clear potential for narcissism to create problems in romantic relationships, the role of narcissism in romantic relation- ships is understudied. Questions remain regarding how narcissism relates to the use of destructive behavior with romantic partners and the use of destructive behavior by the partners of narcissists. Given that narcissists may be cruel, aggressive, and generally annoying and exasperating, partners may react to narcissists with a certain degree of aggression and irritation of their own. Narcissists do re- port more negative interpersonal interactions (Rhodewalt, Madrian, & Cheney, 1998), and report being victimized by others to a greater degree than non-narcissists (McCullough, Emmons, Kilpatrick, & Mooney, 2003). However, narcissists may simply perceive greater victimization. On the other hand, the more times someone interacts with a narcissist, the less they like the narcissist (Paulhus, 1998). In- dependent and blind coders of audiotaped conversations reported that narcissists are less likeable than non-narcissists (Morf, 1994). Wink (1991) found that the spouses of overt narcissists described them as aggressive, outspoken, egotistical, self-centered, intolerant, arrogant, demanding, and argumentative.
The present investigation builds on this prior research by taking a dyadic perspective on the role of narcissism in romantic relation- ships. The key to a dyadic perspective is to measure both partners in a relationship. First, we examine relations between self and part- ner narcissism. Assortative mating is well-established (Thiessen & Gregg, 1980). Individuals tend to be attracted to and choose partners who are similar to them (homogamy) on a variety of characteristics, including psychiatric conditions (Merikangas, 1982; Merikangas & Spiker, 1982). Previous research has found some evidence of assor- tative mating among narcissists (Campbell, Foster, & Finkel, 2002). Theorists propose that homogamy is the result of (1) opportunities to meet potential mates across daily life lead to meeting partners with similar interests and characteristics; (2) involvement of family and friends in introducing romantic partners, as these third parties are themselves similar to the partners; and (3) personal preferences for similar people (Kalmijn, 1994). Narcissists may choose activities
28 KELLER ET AL.
where they are likely to meet other narcissists, such as performance- related hobbies or competitive occupations. Their friends and fam- ily may introduce them to each other, thinking that they would be a good match. They may also prefer each other because of similar values or interests. We hypothesize a replication of previous find- ings that homogamy for narcissism will occur.
We also examine relations between narcissism and self or partner conflict behaviors. This research question is based on prior research demonstrating interpersonal difficulties and aggression in the con- text of high narcissism, but extends this prior research to identify actor and partner effects. Our second hypothesis is that narcissists will engage in greater aggression and stonewalling, and less co- operation in their romantic relationships. Our third hypothesis is that narcissists’ partners will also engage in greater aggression and stonewalling, and less cooperation during conflict. We also exam- ine statistical interactions between male and female narcissism. Our fourth hypothesis is that the combination of male and female narcis- sism will be especially likely to lead to aggression and poor conflict behavior. The examination of interactions between male and female narcissism will provide important information about the potential dyadic nature of narcissism.
Narcissism is a multi-dimensional construct, including grandi- ose and vulnerable dimensions (Cain, Pincus, & Ansell, 2008). The Narcissistic Personality Inventory (NPI; Raskin & Hall, 1979), the most widely used measure of narcissism, is considered an index of grandiose narcissism (Miller & Campbell, 2008). A total score for narcissism is most frequently used, but there are potential sub-scale measures which may permit greater precision in the study of the correlates of narcissism (Maxwell, Donnellan, Hopwood, & Ack- erman, 2011). Recently, Ackerman and colleagues (2011) have rec- ommended the use of three subscales from the NPI: Leadership/ Authority, Grandiose Exhibitionism, and Entitlement/Exploitative- ness. The Leadership/Authority subscale is proposed to measure healthy or normal narcissism, while the other subscales are consid- ered measures of pathological narcissism. When specific aspects of narcissism have been examined, exploitativeness emerges as an especially deleterious characteristic (Dickinson & Pincus, 2003; Mc- Cullough et al., 2003). The current study employs these three sub- scales in order to determine whether distinctions between normal and pathological narcissism apply to the context of conflict and ag- gression in romantic relationships.
NARCISSISM IN ROMANTIC RELATIONSHIPS 29
METHOD
PARTICIPANTS
Participants were 190 college-student couples participating in a larger study of romantic relationships. To be eligible, one partner had to be currently enrolled in an introductory psychology course, and couples had to have been together for at least three months. The majority of participants were European American (84.1% of women and 89.6% of men). Participants were between the ages of 18 and 24 years (women: M = 19.07, SD = 1.30; men: M = 19.63, SD = 1.36). On average, couples had been together between 6 and 18 months.
PROCEDURE
This study was conducted with the approval of the institution’s in- ternal review board (IRB) and informed consent was obtained. Cou- ples attended a laboratory session during the afternoon. Both part- ners came to the session, but completed questionnaires separately. One-week later, a subset of participants (N = 74 couples) completed a competitive reaction time task (Taylor, 1967) ostensibly against their partner. This task is a well-established, reliable, and valid measure of aggression (e.g., Giancola & Zeichner, 1995). Participants were told that they and their romantic partners would have to press a button as fast as possible on each of 25 trials, and that whoever was slower would receive a blast of white noise through a pair of head- phones. At the beginning of each trial, participants set the loudness of the noise their partners would receive if the partners lost the trial, from 60 dB (level 1) to 105 dB (level 10; about the same volume as a smoke alarm). Participants were also able to select 0 dB (level 0). The duration of the white noise was also set by the participants pre- trial. Of the 25 trials, the participants won a randomly determined subset of 12 trials. Over the course of the task, participants experi- enced increasing levels of aggression from their partner after losing trials. A computer recorded all events in the task. Participants were debriefed at the end of the laboratory session.
30 KELLER ET AL.
MEASURES
Narcissism. Participants completed the Narcissistic Personality Inventory (Raskin & Hall, 1979). This questionnaire includes 40 forced-choice items. Items were scored based on the factor structure identified by Ackerman and colleagues (2011). The Leadership/Au- thority subscale consists of 12 items (e.g., I have a natural talent for influencing people vs. I am not good at influencing people). The Grandiose Exhibitionism subscale consists of 10 items (e.g., I like to look at my body vs. My body is nothing special). The Entitlement/ Exploitation subscale consists of 4 items (e.g., I find it easy to ma- nipulate people vs. I don’t like it when I find myself manipulating people.). Reliability coefficients (Cronbach’s alpha) in the present study were: .79 for male Leadership/Authority, .76 for female Lead- ership/Authority, .73 male Grandiose Exhibitionism, .75 for female Grandiose Exhibitionism, .29 for male Entitlement/Exploitation, and .27 for female Entitlement/Exploitation. Due to low reliabil- ity, the Entitlement/Exploitation scale was removed from analyses. The total NPI score, which includes Entitlement/Exploitation items, was also used in analyses, α = .86 and .82, for males and females, respectively.
Conflict Behaviors. Each partner completed the Conflict Properties Scales (CPS; Kerig, 1996) measures of Cooperation (6 items; e.g., Lis- ten to the other’s point of view.), Stonewalling (6 items; e.g., Bicker without really getting anywhere.), Verbal Aggression (8 items; e.g., Raise voice, yell, shout.), and Physical Aggression (7 items; e.g., Slap partner.). Items were rated on a scale from 0 (never) to 3 (often) in terms of how frequently participants engaged in each behavior during conflicts. Participants completed the CPS once with regard to their own behavior and once with regard to their partners’ be- havior. Self and partner reports were significantly correlated (r’s ranging from .43 to .89, all p’s < .001) and so were averaged to sim- plify analyses. Reliability coefficients were as follows: .77 for male Cooperation, .84 for female Cooperation, .67 for male Stonewall- ing, .75 for female Stonewalling, .86 for male Verbal Aggression, .86 for female Verbal Aggression, .67 for male Physical Aggression and .79 for female Physical Aggression. Additional analyses were conducted using the separate reports of conflict behaviors; model coefficients and their significance levels were highly similar to those based on the averaged scores and therefore are not presented.
NARCISSISM IN ROMANTIC RELATIONSHIPS 31
Competitive Reaction Time Task. The following measures were de- rived from the Competitive Reaction Time Task: (1) Unprovoked Aggression = Noise intensity set by participants before the initial trial; (2) Provoked Aggression = Average noise intensity set by par- ticipants in the final block of trials, in which their partners have blasted them with the highest levels of noise; (3) Extreme Aggres- sion = Number of trials for which participants set the noise intensity to levels 9 or 10 (the two highest intensity levels); and (4) Total Ag- gression = An index of aggression across the entire task that takes into account intensity and duration. The Total Aggression index was computed as follows: (1) intensity score was standardized for each trial (25 standardized scores per person); (2) the same thing was done for the duration score; (3) for each trial, the standardized intensity and standardized duration were multiplied together (25 products for each participant); and (4) the products were averaged across the 25 trials to provide the Total Aggression score for each participant.
Depression. Each participant completed the Center for Epidemio- logic Studies Depression Scale (CESD; Radloff, 1977). The CESD is a 20-item measure that is widely used to assess depressive symptoms. Cronbach’s alpha was .82 for women and .83 for men. Depression scores were included as control variables in all models. Optimally, the estimation of predicted associations would control for self-es- teem (Campbell, Rudlich, & Sedikides, 2002). There was no mea- sure of self-esteem available in the current study, but depression is often negatively correlated with self-esteem (Kernis, Grannemann, & Mathis, 1991) and therefore was used as an alternative control. Further, depression is associated with relationship difficulties (At- kins, Dimidjian, Bedics, & Christensen, 2009).
DATA ANALYSES
Means, standard deviations, and correlations among variables are presented in Table 1. Data analyses for hypothesis 1 involved ex- amination of bivariate correlations between male and female nar- cissism. For the remaining hypotheses, analyses were conducted with structural equation modeling in MPLUS version 6 (Muthen & Muthen, 2011). Maximum Likelihood with Robust Standard Errors (the MLM option in MPLUS) was used to estimate model param-
32 KELLER ET AL.
TA B
LE 1
. M
ea n s,
S ta
n d ar
d D
ev ia
ti o n s,
a n d C
o rr
el at
io n s
V ar
ia b le
1 2
3 4
5 6
7 8
9 1 0
1 1
1 2
1 3
1 4
1 . R
el at
io n sh
ip L
en gt
h
2 . M
al e
D ep
re ss
io n
.0 8
3 . Fe
m al
e D
ep re
ss io
n –.
0 1
.2 3
4 . M
al e
G ra
n d io
se E
xh ib
it io
n –.
0 6
–. 0 3
.1 6
5 . M
al e
Le ad
er sh
ip /A
u th
o ri
ty –.
0 4
–. 3 0 *
.1 4
.4 8 **
*
6 . Fe
m al
e G
ra n d io
se E
xh ib
it io
n –.
0 1
–. 1 4
–. 1 8
.2 6 **
* .1
8 *
7 . Fe
m al
e Le
ad er
sh ip
/A u th
o ri
ty –.
1 0
–. 1 7
–. 0 1
.2 3 **
.2 3 **
.3 2 **
*
8 . M
al e
C o o p er
at io
n .0
5 –.
0 7
–. 2 0
–. 0 3
–. 0 3
.0 0
.0 0
9 . M
al e
St o n ew
al li n g
.0 7
.2 5 *
.3 4 **
.2 1 **
.0 6
.0 9
.0 4
–. 2 4 **
1 0 . M
al e
V er
b al
A gg
re ss
io n
.1 0
.0 7
.1 0
.1 9 **
.1 5 *
.1 5 *
.0 0
–. 3 8 **
* .7
4 **
*
1 1 . M
al e
P hy
si ca
l A gg
re ss
io n
.1 1
.0 7
.1 6
.2 3 **
.1 2
.0 6
.0 8
–. 2 7 **
* .4
3 **
* .5
0 **
*
1 2 . M
al e
U n p ro
vo ke
d A
gg re
ss .
–. 0 1
.1 2
–. 0 4
.1 3
–. 0 4
.0 1
.1 8
–. 0 9
–. 0 4
–. 0 5
.1 8
1 3 . M
al e
P ro
vo ke
d A
gg re
ss io
n .1
9 –.
2 2
–. 1 8
.2 2
.0 5
.1 0
.2 6 *
–. 1 4
.0 3
.1 0
.2 6 *
.5 7
** *
1 4 . M
al e
Ex tr
em e
A gg
re ss
io n
.1 9
–. 2 2
–. 0 8
.0 3
.0 4
–. 0 5
.0 7
–. 1 8
.0 0
.0 6
.1 0
.6 9
** *
.6 3
** *
1 5 . M
al e
To ta
l A gg
re ss
io n
.1 7
.1 0
–. 0 7
.2 4
.0 2
.1 5
.3 2 **
–. 0 8
–. 0 2
.0 4
.2 2
.7 4
** *
.9 0
** *
.7 7
** *
1 6 . Fe
m al
e C
o o p er
at io
n –.
0 1
–. 1 8
–. 2 9 *
–. 0 8
–. 1 0
–. 0 6
–. 0 5
.6 6 **
* –.
3 0 **
* –.
3 8 **
* –.
2 9 **
* .0
1 –.
0 5
–. 2
0
1 7 . Fe
m al
e St
o n
ew al
li n g
.1 4
.2 7 *
.3 2 **
.1 5 *
.0 7
.0 9
.0 4
–. 3 0 **
* .7
6 **
* .6
5 **
* .3
6 **
* –.
0 6
.1 9
.0 8
1 8 . Fe
m al
e V
er b al
A gg
re ss
io n
.1 2
.1 3
.1 3
.1 5 *
.1 3
.1 5 *
.0 7
–. 3 9 **
* .7
4 **
* .8
7 **
* .4
7 **
* –.
0 4
.1 6
.1 0
1 9 . Fe
m al
e P hy
si ca
l A gg
re ss
io n
.1 4 *
.0 8
.3 5 **
.2 1 **
.1 3
.1 4
.1 5 *
–. 2 7 **
* .4
3 **
* .4
3 **
* .5
8 **
* –.
0 6
.0 6
.0 3
2 0 . Fe
m al
e U
n p
ro vo
ke d A
gg re
ss io
n .1
6 –.
1 3
–. 1 6
.0 7
.2 3
.2 1
.2 9 *
.0 6
.1 0
.1 6
–. 0 1
.2 8
* .3
8 **
.4 2
**
2 1 . Fe
m al
e P ro
vo ke
d A
gg re
ss io
n .0
9 –.
0 2
–. 0 5
.0 8
.1 9
.2 1
.2 9 *
–. 1 1
.0 4
.0 8
.0 0
.0 7
.2 6
* .1
3
2 2 . Fe
m al
e Ex
tr em
e A
gg re
ss io
n .1
0 –.
0 3
.0 3
.0 2
.1 2
.1 7
.2 5 *
.0 2
.0 4
.1 0
.1 1
.1 0
.2 3
.1 8
2 3 . Fe
m al
e To
ta l A
gg re
ss io
n .1
5 –.
1 9
–. 0 6
.0 8
.2 2
.2 2
.3 3 *
–. 0 9
.1 0
.1 4
.0 3
.1 6
.3 2
* .2
6 *
M 1 .6
4 2 7 .4
6 2 9 .2
6 4 .0
1 6 .5
9 3 .7
2 5 .2
3 2 .5
6 0 .8
1 1 .3
5 0 .0
9 0
.1 9
0 .4
3 8
.3 5
SD 1 .4
2 6 .6
8 5 .6
9 2 .5
2 2 .8
5 2 .4
8 2 .8
2 0 .3
4 0 .3
9 0 .5
1 0 .1
4 1
.9 6
1 1
.8 5
8 .4
0
N o te
. R
el at
io n sh
ip l en
gt h w
as c
o d ed
s u ch
t h at
0 =
3 -6
m o
n th
s, 1
= 6
-1 2 m
o n th
s, 2
= 1
2 -1
8 m
o n th
s, 3
= 1
8 -2
4 m
o n th
s, a
n d 4
= 2
4 o
r m
o re
m o n
th s.
* p
< . 0
5 , **
p <
. 0
1 ,
** *p
< . 0 0 1
NARCISSISM IN ROMANTIC RELATIONSHIPS 33
eters, because several variables were skewed and MLM is designed to handle the analysis of non-normal data. The very minimal miss- ing data (<1% of cases) was treated with mean replacement because MLM cannot be conducted with missing data. Narcissism scales were treated as single indicators in all models. Two types of models were fit: (1) models with the total NPI score and (2) models with the subscale scores for Leadership/Authority and Grandiose Ex- hibitionism. Both subscales were included in each model, so that associations can be interpreted as controlling for the other subscale score. Correlations among narcissism measures and among conflict behavior’s residuals were estimated. Models were fully saturated and no fit indices are reported (fully saturated models are a perfect fit for the data). Although each model included tests of multiple hypotheses, results are presented for each hypothesis separately to ease interpretation. Example models are provided in figures. All effects control for self and partner depression (Watson, Sawrie, Greene, & Arredondo, 2002).
Models were organized as actor-partner interdependence models (APIM; Cook & Kenny, 2005), such that male and female narcissism (Total NPI score, Grandiose Exhibitionism or Leadership/Author- ity) were included as predictors of male and female conflict behav- iors (either cooperation, stonewalling, verbal aggression, physical aggression, or a behavioral aggression measure). Actor-partner interdependence model allows for the analysis of dyadic data by treating the couple as the unit of analysis (each case represents a couple, with variables distinguishing whether scores are for the male or female partner). This approach avoids violations of the as- sumption of independent observations. It also maintains the unique features in each partner’s scores. Actor effects are those where a person’s narcissism predicts his or her own conflict behavior. Part- ner effects are those where a person’s narcissism predicts his or her partner’s conflict behavior. We use the term effects to be consistent with standard terminology in APIM, but data are correlational and cross-sectional findings do not imply causal effects. In APIM, actor effects are calculated controlling for partner effects and estimates of partner effects control for actor effects. This partials out shared variance between the dyad in order to meet assumptions of inde- pendence required by structural equation modeling. In terms of re- search questions, whether narcissism is related to conflict behavior is tested via actor effects; whether narcissism is related to conflict behaviors of partners is tested via partner effects.
34 KELLER ET AL.
Male and female actor effects were compared by constraining them to be equal and observing the change in model chi square. Male and female partner effects were compared in a similar way. Because model chi square values derived from MLM estimation cannot be compared with the typical difference chi square test, the Satorra-Bentler scaled chi square test for use with MLM and other alternative estimation procedures was employed (Satorra & Bentler, 1999). A significant result of the scaled chi square test indicates that the actor (or partner) effects are significantly different from each other. These comparisons were conducted in cases where at least one actor (or partner) effect was statistically significant (whether the other was significant or not).
Interaction terms between male and female narcissism were also included in each model, as predictors of conflict behavior. Male and female narcissism scores were centered before each cross-product was computed. Significant interactions were probed according to criteria put forth by Aiken and West (1991) by submitting MLM- derived estimates to an online utility provided by Preacher and col- leagues (Preacher, Curran, & Bauer, 2006).
RESULTS
HYPOTHESIS 1: INDIVIDUAL NARCISSISM WILL BE ASSOCIATED WITH PARTNER NARCISSISM
Male Leadership/Authority was significantly associated with fe- male Leadership/Authority, r(188) = .23, p < .01. Male Grandiose Ex- hibitionism was related to female Grandiose Exhibitionism, r(188) = .26, p < .001. The full scores of the NPI for male and female partners were also correlated, r(188) = .36, p < .001. Further, male Grandiose Exhibitionism was related to female Leadership/Authority, r(188) = .26, p < .001, and male Leadership/Authority was related to female Grandiose Exhibitionism, r(188) = .18, p < .05. These findings indi- cate significant homogamy for narcissism.
Dependent measures t tests were conducted to examine whether narcissism was higher in one gender. Male and female Grandiose Exhibitionism was similar, but males reported greater Leadership/ Authority (M = 6.59) than females (M = 5.23), t(189) = 5.34, p < .001. The NPI Total Score was also significantly greater for males than females, t(189) = 5.33, p < .001.
NARCISSISM IN ROMANTIC RELATIONSHIPS 35
HYPOTHESIS 2: GREATER NARCISSISM WILL BE RELATED TO SELF CONFLICT BEHAVIORS
See Figure 1 for an example model. See Table 2 for model coeffi - cients. Female NPI score was positively associated with female use of Cooperation, B = .023, p < .01, Stonewalling, B = .02, p < .05, Ver- bal Aggression, .004, p < .05, Unprovoked Aggression, B = .079, p < .001, Provoked Aggression, B = .452, p < .01, Extreme Aggression, B = .275, p < .05, and Total Aggression, B = .094, p < .05. Male NPI score was only marginally related to male use of Verbal Aggression, B = .015, p < .10. Statistical comparisons of actor effects are shown in Table 3. Two comparisons were statistically signifi cant and one was marginally signifi cant. In all cases, the female actor effect was stronger than the male actor effect, indicating that female narcis- sism is more strongly linked to female relationship behavior than male narcissism is related to male relationship behavior.
Female Grandiose/Exhibitionism was related to female use of greater Cooperation, B = .030, p < .05, Unprovoked Aggression, B = .120, p < .01, and marginally related to female Extreme Aggres- sion, B = .462, p < .06. Male Grandiose/Exhibitionism was related to male Stonewalling, B = .034, p < .05, Verbal Aggression, B = .040, p < .05, and Provoked Aggression, B = 1.253, p < .01. One comparison
FIGURE 1. Example Model for Subscales of NPI and Prediction of Verbal Aggression.
36 KELLER ET AL.
of actor effects was marginally significant. Male Grandiose Exhibi- tionism was more strongly linked to male Stonewalling than female Grandiose Exhibitionism was linked to female Stonewalling, Δχ2(1) = 3.652, p < .06.
Female Leadership/Authority was related to female greater Stonewalling, B = .004, p < .05, and Verbal Aggression, B = .005, p < .01, but lower Unprovoked Aggression, B = -.012, p < .01. Male Leadership/Authority was related to male greater Stonewalling, B = .006, p < .001, Verbal Aggression, B = .006, p < .001, Extreme Ag- gression, B = .182, p < .001, and Total Aggression, B = .024, p < .001, but lower Provoked Aggression, B = -.099, p < .01. No comparisons between female and male actor effects were significant for Leader- ship/Authority.
HYPOTHESIS 3: GREATER NARCISSISM WILL BE RELATED TO PARTNERS’ NEGATIVE CONFLICT BEHAVIORS
Results are shown in Table 2. Female NPI scores were related to male Stonewalling, B = .021, p < .01, Verbal Aggression, B = .02, p < .05, Unprovoked Aggression, B = .074, p < .01, Provoked Aggres- sion, B = .406, p < .05, and Total Aggression, B = .099, p < .001. Male NPI scores were related to greater female use of Physical Aggres- sion, B = .277, p < .001, and Extreme Aggression, B = .163, p < .05. Three comparisons of partner effects were significant, and one was marginally significant (See Table 3). The female partner effect was stronger than the male partner effect for Verbal Aggression, Un- provoked Aggression, and marginally for Total Aggression. These findings indicate that female narcissism is more strongly related to certain male relationship behaviors than male narcissism is related to those female relationship behaviors, The male partner effect was stronger than the female partner effect for Physical Aggression, sug- gesting that women are more likely to be physically aggressive to- ward their narcissistic partners than men are.
Female Grandiose Exhibitionism was unrelated to male relation- ship behavior. Male Grandiose Exhibitionism was significantly re- lated to lesser female Cooperation, B = -.034, p < .01, marginally greater female Stonewalling, B = .034, p < .07, and greater female Physical Aggression, B = .028, p < .05. As shown in Table 3, the part- ner effect of male Grandiose Exhibitionism on female Cooperation was stronger than the partner effect of female Grandiose Exhibi-
NARCISSISM IN ROMANTIC RELATIONSHIPS 37
TABLE 2. Narcissism as a Predictor of Conflict Behavior: Model Results
Total NPI Grandiose Exhibitionism Leadership/Authority
Actor (Self) Effects Male Female Male Female Male Female
Cooperation –.002 .023** –.012 .030* –.002 .003
Stonewalling .000 .020* .034* –.012 .006*** .004*
Verbal Aggression .015+ .004* .040* .006 .006*** .005**
Physical Aggression .004 –.103 .464 –.004 –.113 –.002
Unprovoked Aggression –.017 .079*** .084 .120** .000 –.012**
Provoked Aggression .115 .452** 1.253** .570 –.099** –.053
Extreme Aggression .103 .275* .084 .468+ .182*** .013
Total Aggression .012 .094* .161* .078 .024*** –.015
Partner Effects
Cooperation .001 –.007 –.034** .016 -.003 .003
Stonewalling .004 .021** .034+ –.014 .010*** .003**
Verbal Aggression .001 .020** .026 .017 .009*** .004+
Physical Aggression .277*** .005 .028* –.163 .001 –.120
Unprovoked Aggression .009 .074** –.006 –.005 –.004 –.004
Provoked Aggression .042 .406* .321 .112 –.093** –.073*
Extreme Aggression .163* .094 –.007 –.088 .005 –.013
Total Aggression –.027 .099*** –.025 .065 .028*** .005
Interactions Between Male and Female Narcissism Predicting…
Female Cooperation .029* .012* .001
Female Stonewalling .042 –.018** –.006
Female Verbal Aggression .002 –.019* –.010*
Female Physical Aggression 2.170*** –.014** .003
Female Unprovoked Aggression .000 –.044** .002
Female Provoked Aggression –.007 –.153 –.222*
Female Extreme Aggression .053 .120 .224*
Female Total Aggression .004 –.049+ .006
Male Cooperation –.034** .009+ .005
Male Stonewalling .037* –.009+ .000
Male Verbal Aggression .027 –.006 –.012*
Male Physical Aggression .016 –.149 .071
Male Unprovoked Aggression .000 –.056* .055**
Male Provoked Aggression .019 –.195 .125
Male Extreme Aggression –.020 –.159+ .018
Male Total Aggression –.002 –.049* .035+
Note. Findings are from a total of 16 models in which self narcissism, partner narcissism, and the interaction between the two were included as predictors of self and partner relationship behavior, controlling for self and partner depression. +p < .10; *p < .05; **p < .01; ***p < .001. Coefficients are unstandardized.
38 KELLER ET AL.
TABLE 3. Results of Parameter Comparisons of Actor and Partner Effects
Male Actor vs. Female Actor Effects Predicting… Male Actor Female Actor Delta Chi Square
Total NPI Score
Cooperation –.002 .023* 6.246*
Stonewalling .000 .02* 3.156+
Verbal Aggression .015 .004* 1.921
Unprovoked Aggression –.017 .079*** 9.669**
Provoked Aggression .115 .452** 1.996
Extreme Aggression .103 .275* 1.017
Total Aggression .012 .094* 3.188+
Grandiose Exhibitionism
Cooperation –.012 .030* 1.248
Stonewalling .034* –.012 3.652+
Verbal Aggression .040* .006 1.302
Unprovoked Aggression .084 .120** 0.309
Provoked Aggression 1.253** .570 1.110
Extreme Aggression .084 .462+ 1.248
Leadership/Authority
Stonewalling .034* –.012 1.053
Verbal Aggression .006 .005** 0.191
Unprovoked Aggression .000 –.012** 1.457
Provoked Aggression –.099** –.053 0.945
Extreme Aggression .182*** .013 2.411
Total Aggression .024*** –.015 0.768
Male Partner vs. Female Partner Effects Predicting… Male Partner Female Partner Delta Chi Square
Total NPI Score
Stonewalling .004 .021** 2.325
Verbal Aggression .001 .020** 5.186*
Physical Aggression .277*** .005 17.767***
Unprovoked Aggression .009 .074** 4.856*
Provoked Aggression .042 .406* 2.197
Extreme Aggression .163* .094 0.207
Total Aggression –.027 .099*** 3.412+
Grandiose Exhibitionism
Cooperation –.034** .016 5.731*
Stonewalling .034+ –.014 3.441+
Physical Aggression .028+ –.163 1.062
Leadership/Authority
Stonewalling .010*** .003** 1.444
Verbal Aggression .009*** .004+ 1.434
Provoked Aggression –.093** –.073* 0.371
Total Aggression .028*** .005 1.532
Note. Parameters were compared by placing an equality constraint on the coefficients and observing the increase in model chi square. The increase is denoted delta chi square, it follows a chi square distribution with one df. The Delta Chi Square is based on the Satorra-Bentler scaled chi square difference test for use with MLM estimation. +p < .10; *p < .05; **p < .01; ***p < .001. Coefficients are unstandardized.
NARCISSISM IN ROMANTIC RELATIONSHIPS 39
tionism on male Physical Aggression. The partner effect of male Grandiose Exhibitionism on female Stonewalling was marginally stronger than the partner effect of female Grandiose Exhibitionism on male Stonewalling. This again suggests that women are more likely to respond negatively to their narcissistic partners than men are.
Female Leadership/Authority was related to greater male Stone- walling, B = .003, p < .01, and marginally greater male Verbal Ag- gression, B = .004, p < .06, but lesser male Provoked Aggression, B = -.073, p < .05. Male Leadership/Authority was related to greater female Stonewalling, B = .004, p < .05, Verbal Aggression, B = .009, p < .001, and Total Aggression, B = .028, p < .001, but less female Pro- voked Aggression, B = -.093, p < .01. One partner effect comparison was significant (see Table 3). No comparisons of male and female partner effects were significant.
HYPOTHESIS 4: MALE AND FEMALE NARCISSISM WILL INTERACT
See Table 2 for findings. Several significant or marginally signifi- cant interactions were observed. The interaction between male and female NPI scores significantly predicted Female Cooperation, B = .029, p < .05, Female Physical Aggression, B = 2.170, p < .001, Male Cooperation, B = -.034, p < .01, and male Stonewalling, B = .037, p < .05, and marginally predicted female Stonewalling, B = .042, p < .10. Several of these interactions were very similar in form and an example plot is shown in Figure 2A. Narcissism is only related to a person’s behavior (e.g., greater female Cooperation, female Stone- walling, female Physical Aggression, and male Stonewalling) when partner narcissism was high. In contrast, narcissism was related to less Cooperation, Stonewalling, and Physical Aggression when the partner had low narcissism. The one exception was that high narcis- sism in both partners was related to less male Cooperation, but high narcissism in males combined with low narcissism in females was associated with greater male Cooperation (Figure 2B).
The interaction between male and female Grandiose Exhibition- ism significantly predicted female Cooperation, B = .012, p < .05, female Stonewalling, B = -.018, p < .01, female Verbal Aggression, B = -.019, p < .05, female Physical Aggression, B = -.014, p < .01, female Unprovoked Aggression, B = -.044, p < .01, and marginally
40 KELLER ET AL.
predicted female Total Aggression, B = -.049, p < .10. The interaction between male and female Grandiose Exhibitionism also significant- ly predicted male Unprovoked Aggression, B = -.056, p < .05, and Total Aggression, B = -.049, p < .05, and marginally predicted male Cooperation, B = .009, p < .10, male Stonewalling, B = -.009, p < .10, and male Extreme Aggression, B = -.159, p < .10. Again, the majority of these interactions followed a nearly identical pattern (see Figure 2C). Grandiose Exhibitionism was related to a person’s behavior (greater female Unprovoked Aggression, female Total Aggression, female Verbal Aggression, female Stonewalling, male Stonewalling, male Total Aggression, male Unprovoked Aggression, and male Ex- treme Aggression) only when the partner’s Grandiose Exhibition- ism was lower. When partner Grandiose Exhibitionism was higher, no significant association between self Grandiose Exhibitionism and self relationship behavior was observed; a person’s aggressive behavior was generally moderate or high regardless of their own Grandiose Exhibitionism. For male Cooperation the pattern of find- ings matched the other interactions, such that male Grandiose Ex- hibitionism was related to lesser Cooperation when female Grandi- ose Exhibitionism was lower (not shown). For female Cooperation, there was no association between female Grandiose Exhibitionism and female Cooperation when male Grandiose Exhibitionism was lower or higher (not shown).
The interaction between male and female Leadership/Authority significantly predicted female Verbal Aggression, B = -.010, p < .05, female Provoked Aggression, B = -.222, p < .05, and female Extreme Aggression, B = .224, p < .05. The interaction between male and female Leaderships/Authority also significantly predicted male Verbal Aggression, B = -.012, p < .05, and male Unprovoked Ag- gression, B = .055, p < .01, and marginally predicted male Total Ag- gression, B = .035, p < .10. The pattern of interaction was less clear for Leadership/Authority. In three cases (male Unprovoked Ag- gression, male Total Aggression, and female Extreme Aggression), a person’s Leadership/Authority was related to their greater aggres- sive behavior only when their partner was also higher in Leader- ship/Authority (see Figure 2D). When partner Leadership/Author- ity was lower, a person’s Leadership/Authority was related to less aggression. The opposite pattern was observed for female Provoked Aggression, female Verbal Aggression, and male Verbal Aggression: A person’s Leadership/Authority was linked to greater aggressive behavior only when partner Leadership/Authority was lower, but
NARCISSISM IN ROMANTIC RELATIONSHIPS 41
was linked to lower aggressive behavior when partner Leadership/ Authority was higher (see Figure 2E).
DISCUSSION
The present study investigated narcissism in romantic relation- ships. Our first hypothesis, that male and female narcissism would be positively related, was supported. There are several reasons why homogamy occurs, including personal preferences, exposure to others in daily life, and meeting people through friends and fam- ily (Kalmijn, 1994). In terms of personal preferences, people may choose to be with others who share the same values, beliefs, and characteristics that they do. Kalmijn (1994) refers to this as cultural status and it is typically adaptive in that it will likely lead to less conflict and a more stable relationship (Schramm, Marshall, Har- ris, & Lee, 2012). However, homogamy is observed in mental ill- ness (Merikangas & Spiker, 1982), which seems to increase the odds of a less stable relationship. It is possible that this phenomenon is the result of contagion or is secondary to assortment, in which per- sons choose partners who are similar in personality characteristics and those personality characteristics represent a vulnerability to the same psychopathology (Maes et al., 1998). Alternatively, people may prefer to be with persons who are more attractive than they are. Kalmijn (1994) refers to this as economic status. The more attrac- tive people will receive many opportunities to find a mate, and can therefore choose the most attractive ones. In terms of narcissism, it is possible that less narcissistic individuals have more opportunities to be with others who are also low in narcissism, and that narcissists are simply stuck with each other. Narcissistic homogamy has im- plications for understanding the impact of narcissism on romantic relationships. Persons high in narcissism may be interacting with partners who are also high in narcissism, rather than an innocent party as is sometimes assumed by researchers. We revisit this point in our discussion of hypothesis four.
It was also found that a person’s narcissism is related to his or her relationship behavior. Many of these associations represent aver- age effects that were qualified by interactions. In general, persons higher in narcissism, regardless of whether that was measured by the total NPI score, Grandiose/Exhibitionism, or Leadership/Au- thority, engaged in more aggressive behavior toward their partners.
42 KELLER ET AL.
A
B
C
FIGURE 2. Interactions Between Male and Female Narcissism.
NARCISSISM IN ROMANTIC RELATIONSHIPS 43
FIGURE 2. continued
F
E
D
44 KELLER ET AL.
However, there were a few exceptions: female Leadership/Author- ity was related to less female Unprovoked Aggression in the reac- tion time task and greater female self-reported Cooperation. Male Leadership/Authority was related to less male Provoked Aggres- sion. Because findings were especially strong for Grandiose/Exhi- bitionism and the total NPI score (which also includes Exploitative- ness), they are largely consistent with the current understanding of narcissism as deleterious for interpersonal relationships (Campbell & Foster, 2002) and the particular maladaptiveness of grandiose or exploitative narcissism (Cain, Pincus, & Ansell, 2008). Persons who have symptoms of grandiose narcissism may engage in greater ag- gression with their partners as a way to maintain those feelings of grandiosity (Morf et al., 2001).
The ambiguous findings for Leadership/Authority are some- what consistent with the concept of adaptive forms of narcissism (Sedikides, Rudich, Gregg, Kumashiro, & Rusbult, 2004). For some aspects of romantic relationships, Leadership/Authority may be beneficial because it represents confidence, high self-esteem, and extraversion. However, for other aspects of romantic relationships, including male and female Stonewalling, Verbal Aggression, and Extreme Aggression, Leadership/Authority may be detrimental. These associations were found even after controlling for Grandiose Exhibitionism, and so cannot be explained by associations between Grandiose Exhibitionism and Leadership/Authority. Further, in relation to partner effects (hypothesis three), male Leadership/ Authority was associated with greater female physical aggression and total aggression during the reaction time task. To our knowl- edge, there has been no investigation of Leadership/Authority in the context of romantic relationships. Our findings suggest that the adaptiveness of Leadership/Authority may depend on the form of aggression under consideration (e.g., behavioral aggression vs. self- reported aggression; aggression vs. cooperation).
Several partner effects were also observed. Female total Narcis- sism scores were associated with greater male aggression on several counts (Stonewalling, Verbal Aggression, and Behavioral Aggres- sion). Female Grandiose Exhibitionism was not related to male be- havior, on average, but female Leadership/Authority was related to greater male Stonewalling and Verbal Aggression but lesser male Provoked Aggression. Male total Narcissism scores were associated with female Physical Aggression and Extreme Aggression. Male Grandiose Exhibitionism was related to female Physical Aggression
NARCISSISM IN ROMANTIC RELATIONSHIPS 45
and Stonewalling, and male Leadership/Authority was related to greater female Verbal Aggression, Stonewalling, Provoked Aggres- sion, and Total Aggression, Taken together, these findings indicate that narcissists may indeed be more likely to be victimized by their partners. The pattern of findings for Leadership/Authority was similar to that of total Narcissism and Grandiose Exhibitionism, but there was one exception (for male Provoked Aggression). These findings support the hypothesis that narcissistic persons are obnox- ious and frustrating to their partners. Prior research does suggest that interacting with a narcissist can be frustrating. The more times someone interacts with a narcissist, the less they like the narcissist (Paulhus, 1998). Observers rate narcissists as less likeable than non- narcissists (Morf, 1994). Spouses of overt narcissists find them ag- gressive, outspoken, egotistical, self-centered, intolerant, arrogant, demanding, and argumentative (Wink, 1991).
Several interactions between male and female narcissism were also observed. Taken together, these findings indicate that the ways in which narcissism is related to relationship behavior depends on the aspect of narcissism being investigated, gender, and the level of narcissism in the partner. In several cases, the hypothesis that high narcissism in both partners would be especially detrimental to relationships was supported. This was the case for most inter- actions involving the total Narcissism score and some interactions involving Leadership/Authority. The implication of homogamy in narcissism is that the aggressive behavior of narcissists may be partially due to the narcissism of their partner; when coupled with individuals lower in narcissism, their behavior may be less aggres- sive. Narcissists’ arrogant and superior behavior may be especially provoking to partners who are equally arrogant and superior, espe- cially because such narcissistic partners are likely to perceive such behavior as a threat to their own superiority (Morf et al., 2010). Such findings were most consistent for the total NPI score, perhaps be- cause the total NPI score included items assessing exploitativeness. Exploitative forms of narcissism have been strongly implicated in relationship problems in prior research (Dickinson & Pincus, 2003; McCullough et al., 2003).
One interesting finding for the total NPI score was that the combi- nation of high NPI scores in both partners was related to greater fe- male Cooperation. Higher NPI scores in males combined with low NPI scores in females was related to greater male Cooperation. This suggests that higher narcissism may be related to both increased
46 KELLER ET AL.
aggression but also greater cooperation. A possible explanation for this finding is that there are both adaptive and maladaptive features of narcissism that are captured by the NPI (Ackerman et al., 2011).
A different pattern emerged for Grandiose Exhibitionism and some interactions involving Leadership/Authority. When partner Grandiose Exhibitionism was higher, most participants engaged in moderate to high levels of aggressive behavior regardless of their own Grandiose Exhibitionism. Grandiose Exhibitionism may be particularly provoking to both other narcissists and non-narcis- sists. One’s own Grandiose Exhibitionism is related to one’s own aggressive behavior only when the partner is lower on Grandiose Exhibitionism. The one exception for this was female Cooperation. Interestingly, the combination of higher male and female Grandi- ose Exhibitionism was related to greater female Cooperation. It is unclear why findings were different for female Cooperation. One possibility is that women who are higher in Grandiose Exhibition- ism are more likely to rate themselves higher on positive attributes (like Cooperation) due to their feelings of superiority. They may be especially likely to do so when their partner is higher in Grandiose Exhibitionism as a way of blaming their partner rather than them- selves for problems in the relationship. Future research using be- havioral measures of positive behavior in romantic relationships is needed to confirm this speculation.
The findings for interactions between male and female Leader- ship/Authority were more mixed. In three cases, the combination of high male and female Leadership/Authority was related to greater aggression (male and female, but only for behavioral measures of aggression). Such findings are consistent with high narcissism be- ing detrimental for relationships regardless of the form that it takes. However, in the other three cases, the combination of higher Lead- ership/Authority in both partners was either unrelated to aggres- sion or was related to less female Provoked aggression. It is pos- sible that Leadership/Authority reflects a greater tendency to be dominant in relationships (Emmons, 1984) and when both partners are dominant it creates greater conflict. The opposite pattern was observed for female Provoked Aggression. Higher male and female Leadership/Authority was related to less female Provoked Ag- gression, but higher female Leadership/Authority and lower male Leadership/Authority was related to greater female Provoked Ag- gression. Additional research is needed to replicate these findings
NARCISSISM IN ROMANTIC RELATIONSHIPS 47
and determine why different forms of female aggression are related to narcissism in different ways.
Findings should be interpreted in light of study limitations. The most serious limitation is that data are cross-sectional. Therefore, causal interpretations are not possible, and explanations for find- ings are speculative. This limitation also makes it impossible to examine potential changes in narcissism over time among dyads. It is possible that dynamics within the relationship may influence individual’s NPI scores. Longitudinal research is needed to exam- ine these potential relationships. It is also not apparent whether findings would generalize to individuals suffering from narcissistic personality disorder or to adults, as participants were drawn from a population of college students. Many couples in the study had been together for only a short time, and almost none were engaged or married. A restricted range in the length of romantic relation- ships may have affected associations. Further, relations may differ depending on marriage or cohabitation. Finally, the sample of 74 couples who completed the competitive reaction time task was rela- tively small and tests involving these variables were under-pow- ered.
Despite these limitations, the present study addresses important gaps in the understanding of narcissism in romantic relationships. Findings emphasize the value of a dyadic perspective, and suggest that the narcissism of one’s partner may have an important role in one’s own aggressive behavior and that this may be especially the case when one is also high in narcissism.
REFERENCES
Ackerman, R. A., Witt, E. A., Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., & Kashy, D. A. (2011). What does the narcissistic personality inventory really measure? Assessment, 18, 67–87.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interac- tions. Newbury Park, CA: Sage.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders, fourth edition, text revision (DSM-IV-TR). Washington, DC: American Psychiatric Association.
Atkins, D. C., Dimidjian, S., Bedics, J. D., & Christensen, A. (2009). Couple discord and depression in couple therapy and in depressed individuals during de- pression treatment. Journal of Consulting and Clinical Psychology, 77, 1089 –1099.
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.
48 KELLER ET AL.
Cain, N. M., Pincus, A. L., & Ansell, E. B. (2008). Narcissism at the crossroads: Phe- notypic description of pathological narcissism across clinical theory, social/ personality psychology, and psychiatric diagnosis. Clinical Psychology Review, 28, 638–656.
Campbell, W. K. (1999). Narcissism and romantic attraction. Journal of Personality and Social Psychology, 77, 1254–1270.
Campbell, W. K. (2001) Is narcissism really so bad? Psychological Inquiry, 12, 214–216. Campbell, W. K., & Foster, C. A. (2002) Narcissism and commitment in romantic
relationships: An investment model analysis. Personality and Social Psychology Bulletin, 28, 484–494.
Campbell, W. K., Foster, C. A., & Finkel, E. J. (2002). Does self-love lead to love of others? A story of narcissistic game playing. Journal of Personality and Social Psychology, 83, 340–354.
Campbell, W. K., Rudlich, E. A., & Sedikides, C. (2002). Narcissism, self-esteem, and the positivity of self-views: Two portraits of self-love. Personality and Social Psychology Bulletin, 28, 358–368.
Carnelley, K. B., Pietromonaco, P. R., & Kenneth, J. (1994). Depression, working models of others, and relationship functioning. Journal of Personality and Social Psychology, 66, 127–140.
Cook, W. L., & Kenny, D. A. (2005). The Actor-Partner Interdependence Model: A model of directional effects in developmental studies. International Journal of Behavioral Development, 29, 101–109.
Dickinson, K. A., & Pincus, A. L. (2003). Interpersonal analysis of grandiose and vulnerable narcissism. Journal of Personality Disorders, 17, 188–207.
Emmons, R. A. (1984). Factor analysis and construct validity of the Narcissistic Per- sonality Inventory. Journal of Personality Assessment, 48, 291–300.
Foster, J. D., & Campbell, W. K. (2005). Narcissism and resistance to doubts about romantic partners. Journal of Research in Personality, 39, 550–557.
Giancola, P. R., & Zeichner, A. (1995). Construct validity of a competitive reaction‐ time aggression paradigm. Aggressive Behavior, 21(3), 199-204.
Kalmijn, M. (1994). Assortative mating by cultural and economic occupational sta- tus. American Journal of Sociology, 100, 422–452.
Kerig, P. K. (1996). Assessing the links between interparental conflict and child ad- justment: The conflicts and problem-solving scales. Journal of Family Psychol- ogy, 10, 454–473.
Kernis, M. H., Grannemann, B. D., & Mathis, L. C. (1991). Stability of self-esteem as a moderator of the relation between level of self-esteem and depression. Journal of Personality and Social Psychology, 61, 80–84.
Liben, L. S., & Bigler, R. S. (2002). Extending the study of gender differentiation. Monographs of the Society for Research in Child Development, 67, 175–183.
Maes, H.H.M., Neale, M. C., Kendler, K. S., Hewitt, J. K., Silberg, J. L., Foley, D. L. et al. (1998). Assortative mating for major psychiatric diagnoses in two popula- tion based samples. Psychological Medicine, 28, 1389–1401.
Masterson, J. F. (1988). The search for the real self: Unmasking the personality disorders of our age. New York: Taylor & Francis.
Maxwell, K., Donnellan, M. B., Hopwood, C. J., & Ackerman, R. A. (2011). The two faces of Narcissus? An empirical comparison of the narcissistic personality inventory and the pathological narcissism inventory. Personality and Individu- al Differences, 50, 577–582.
NARCISSISM IN ROMANTIC RELATIONSHIPS 49
McCullouch, M. E., Emmons, R. A., Kilpatrick, S. D., & Mooney, C. N. (2003). Nar- cissists as “victims”: The role of narcissism in the perception of transgres- sions. Personality and Social Psychology Bulletin, 29, 885–893.
Merikangas, K. R. (1982) Assortative mating for psychiatric disorders and psycho- logical traits. Archives of General Psychiatry, 39, 1173–1180.
Merikangas, K. R., & Spiker, D. G. (1982). Assortative mating among in-patients with primary affective disorder. Psychological Medicine, 12, 753–764.
Miller, J. D., & Campbell, W. K. (2008). Comparing clinical and social-personality conceptualizations of narcissism. Journal of Personality, 76, 449–476.
Morf, C. C. (1994). Interpersonal consequences of narcissists’ continual effort to maintain and bolster self-esteem (Doctoral dissertation, University of Utah, Salt Lake City, 1994). Dissertation Abstracts International, 55(6–B), 2430.
Morf, C. C., Horvath, S., & Torchetti, L. (2010). Narcissistic self-enhancement: Tales of (successful?) self-portrayal. In M. D. Alicke & C. Sedikides (Eds.), Handbook of self-enhancement and self-protection (pp. 399–424). New York: Guilford.
Muthén, L. K., & Muthén, B. O. (2011). Mplus User’s Guide, 5th ed. Los Angeles: Muthén & Muthén.
Ogrodniczuk, J. S., Piper, W. E., Joyce, A. S., Steinberg, P. I., & Duggal, S. (2009). Interpersonal problems associated with narcissism among psychiatric outpa- tients. Journal of Psychiatric Research, 43, 837–842.
Paulhus, D. L. (1998). Interpersonal and intrapsychic adaptiveness of trait self-en- hancement: A mixed blessing? Journal of Personality and Social Psychology, 74, 1197–1208.
Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and la- tent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.
Raskin, R., & Hall, C. S. (1979). A narcissistic personality inventory. Psychological Reports, 45, 590.
Reidy, D. E., Foster, J. D., & Zeichner, A. (2010). Narcissism and unprovoked aggres- sion. Aggressive Behavior, 43, 414–422.
Rhodewalt, F., Madrian, J. C., & Cheney, S. (1998). Narcissism, self-knowledge or- ganization, and emotional reactivity: The effect of daily experiences on self- esteem and affect. Personality and Social Psychology Bulletin, 24, 75–87.
Satorra, A., & Bentler, P. M. (1999). A scaled difference chi-square test statistic for moment structure analysis. Unpublished Manuscript. http://statistics.ucla. edu/preprints/uclastat-preprint-1999:19
Schramm, D. G., Marshall, J. P., Harris, V. W., & Lee, T. R. (2012). Religiosity, ho- mogamy, and marital adjustment: An examination of newlyweds in first mar- riages and remarriages. Journal of Family Issues, 33, 246–268.
Sedikides, C., Rudich, E. A., Gregg, A. P., Kumashiro, M., & Rusbult, C. (2004). Are normal narcissists socially healthy? Self-esteem matters. Journal of Personality and Social Psychology, 87, 400–416.
Smolewska, K., & Dion, K. L. (2005). Narcissism and adult attachment: A multivari- ate approach. Self and Identity, 4, 59–68.
Stucke, T. S., & Sporer, S. L. (2002). When a grandiose self-image is threatened: Nar- cissism and self-concept clarity as predictors of negative emotions and ag- gression following ego-threat. Journal of Personality, 70, 509–032.
50 KELLER ET AL.
Taylor, S. P. (1967). Aggressive behavior and physiological arousal as a function of provocation and the tendency to inhibit aggression. Journal of Personality, 35(2), 297–310.
Thiessen, D. D., & Gregg, B. (1980). Human assortative mating and genetic equilib- rium: An evolutionary perspective. Ethology and Sociobiology, 1, 111–140.
Twenge, J. M., & Campbell, W. K. (2003). “Isn’t it fun to get the respect that we’re going to deserve?” Narcissism, social rejection, and aggression. Personality and Social Psychology Bulletin, 29, 261–272.
Watson, P. J., Sawrie, S. M., Greene, R. L., & Arredondo, R. (2002). Narcissism and depression: MMPI-2 evidence for the continuum hypothesis in clinical sam- ples. Journal of Personality Assessment, 79, 85–109.
Wink, P. (1991). Two faces of narcissism. Journal of Personality and Social Psychology, 61, 590–597.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.