See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/244479376

A Test of Two Brief Measures of Grandiose Narcissism: The Narcissistic

Personality Inventory-13 and the Narcissistic Personality Inventory-16

Article in Psychological Assessment · July 2013

DOI: 10.1037/a0033192 · Source: PubMed

CITATIONS READS
302 13,533

6 authors, including:

Brittany Christine Gentile Joshua D Miller

Genentech University of Georgia
30 PUBLICATIONS 4,169 CITATIONS 419 PUBLICATIONS 26,608 CITATIONS

SEE PROFILE SEE PROFILE

Brian James Hoffman Dennis E. Reidy

University of Georgia Georgia State University
50 PUBLICATIONS 7,746 CITATIONS 78 PUBLICATIONS 3,150 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by W. Keith Campbell on 13 February 2015.

The user has requested enhancement of the downloaded file.

 Psychological Assessment © 2013 American Psychological Association
2013, Vol. 25, No. 3, 000 1040-3590/13/$12.00 DOI: 10.1037/a0033192

A Test of Two Brief Measures of Grandiose Narcissism: The Narcissistic

Personality Inventory–13 and the Narcissistic Personality Inventory–16
Brittany Gentile, Joshua D. Miller, Brian J. Hoffman, Dennis E. Reidy, Amos Zeichner,
and W. Keith Campbell
University of Georgia

The most widely used measure of trait narcissism is the Narcissistic Personality Inventory (NPI), which
can provide both total and subscale scores. However, with a length of 40 items, this measure may not be
ideal in settings in which time or participant attention may limit the types of measures that can be
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

administered. In response, Ames, Rose, and Anderson (2006) created the NPI-16, which provides a
This document is copyrighted by the American Psychological Association or one of its allied publishers.

shorter, unidimensional measure of the construct. In the present research, we examine the reliability and
validity of the NPI-16 in conjunction with a new short measure of narcissism, the NPI-13, which provides
both a total score and 3 subscale scores (Leadership/Authority; Grandiose Exhibitionism; Entitlement/
Exploitativeness). Across 2 studies, we demonstrate that both short measures manifest good convergent
and discriminant validity and adequate overall reliability. The NPI-13 may be favored over the NPI-16
because it allows for the extraction of 3 subscales, consistent with the use of its parent measure.

Keywords: narcissism, grandiose, measurement, personality, assessment

Supplemental materials: http://dx.doi.org/10.1037/a0033192.supp

Narcissism is a construct of increasing interest to psychologists comprehensive test of the reliability and validity of the NPI-16.
from a variety of disciplines including clinical, social-personality, Although researchers have traditionally focused primarily on the
and industrial-organizational psychology. Trait narcissism is NPI-40 total scores, there has been a recent push to use subscales
thought to exist on a continuum (Foster & Campbell, 2007) and is as well, as they manifest divergent patterns of relations with
marked by a grandiose sense of self, feelings of entitlement, and a important external criteria such as self-esteem and psychological
dominant and antagonistic interpersonal style. Growing evidence distress. The second goal of this study was to create and test
suggests that narcissism is a heterogeneous construct composed of another brief measure of narcissism, the NPI-13, that would pro-
grandiose and vulnerable dimensions (Dickinson & Pincus, 2003; vide a total score and three subscale scores, in line with recent
Fossati et al., 2005; Miller & Campbell, 2008; Miller, Hoffman, et work on the factor structure of the NPI-40 (Ackerman et al., 2011).
al., 2011; Russ, Shedler, Bradley, & Westen, 2008; Wink, 1991);
however, the vast majority of research on the topic has focused on
the former. Factor Structure of the NPI-40
By far, the most widely used measure of grandiose narcissism is Explorations of the underlying factor structure of the NPI-40
the Narcissistic Personality Inventory (NPI; Raskin & Terry, have yielded a variety of solutions. Originally, Emmons (1984,
1988), often the 40-item forced-choice version (hereafter referred 1987) found evidence for four-factors (i.e., leadership/authority,
to as the NPI-40), which can be broken down further into subscales superiority/arrogance, self-absorption/self-admiration, and exploit-
(e.g., Ackerman et al., 2011; Emmons, 1984, 1987; Raskin & ativeness/entitlement) using both principal components analysis
Terry, 1988). For the sake of increased efficiency, Ames, Rose, (PCA) and exploratory factor analysis (EFA). Using similar meth-
and Anderson (2006) created a 16-item short form of the NPI-40. ods, Raskin and Terry (1988) found a seven-factor solution (i.e.,
The NPI-16 generates a global narcissism score that has been exploitativeness, exhibitionism, entitlement, superiority, self-
shown to manifest a pattern of results consistent with those pro- sufficiency, authority, and vanity). Amid concerns regarding the
duced by the NPI-40. The first goal of this study was to provide a stability of these factor solutions, researchers have revisited this
issue and found converging evidence for a more parsimonious
solution. Using PCA and confirmatory factor analysis (CFA),
Kubarych, Deary, and Austin (2004) found evidence for two- and
three-factor solutions pertaining to power, exhibitionism, and (in
Brittany Gentile, Joshua D. Miller, Brian J. Hoffman, Dennis E. Reidy, the case of the three-factor solution) being a special person. Corry,
Amos Zeichner, and W. Keith Campbell, Department of Psychology,
Merritt, Mrug, and Pamp (2008) also found evidence for a two-
University of Georgia.
Correspondence concerning this article should be addressed to Brittany
factor solution; the factors were titled leadership/authority and
Gentile, Department of Psychology, Psychology Building, University of exhibitionism/entitlement. Combining these methods in, perhaps,
Georgia, Athens, GA 30602-3013, or to Joshua D. Miller, Department of the most rigorous analysis, Ackerman and colleagues (2011) found
Psychology, Psychology Building, University of Georgia, Athens, GA support for a three-factor solution containing similar factors (i.e.,
30602-3013. E-mail: [email protected] or [email protected] leadership/authority, grandiose exhibitionism, entitlement/exploit-

1
 2 GENTILE ET AL.

ativeness). In sum, there is converging evidence that the NPI-40 Study 1

contains two to three distinguishable factors.

Differences Among the NPI-40 Subscales Method

Scholars have increasingly called for the use of NPI-40 subscale Participants. Sample 1 was composed of 361 undergraduates
scores in place of, or in addition to, total scores, given some (225 women, 135 men, and one who did not identify as either)
evidence that they manifest divergent relations with certain im- from the University of Georgia with a mean age of 19.16 (SD ⫽
portant external criteria. For instance, among the three NPI-40 1.42). Of these, 86.1% were White. Data from this sample were
subscales, the entitlement/exploitativeness factor is thought by previously published in Miller et al. (2010).2
some to be the most indicative of narcissistic personality pathology Sample 2 was composed of 238 undergraduates (143 women, 95
as it is related to lower self-esteem and extraversion (Ackerman et men) from the University of Georgia with a mean age of 19.13
al., 2011; Brown, Buzdek, & Tamborski, 2009), as well as higher (SD ⫽ 1.26). Of these, 79.8% were White. Data from this sample
mood variability (Emmons, 1987) and neuroticism (Ackerman et were previously published in Miller, Hoffman, et al. (2011).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

al., 2011; Emmons, 1984). In addition, this NPI-40 subscale is Sample 3 was composed of 86 undergraduate men from the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

related to both grandiose and vulnerable narcissism (Ackerman et University of Georgia with a mean age of 19.67 (SD ⫽ 1.26). Of
al., 2011; Miller, Gentile, Wilson, & Campbell, 2013; Miller, these 86.2% were White. Data from this sample were previously
Price, Gentile, Lynam, & Campbell, 2012) and narcissistic per- published in Reidy, Zeichner, Foster, and Martinez (2008).
sonality disorder (NPD; Emmons, 1987). In contrast, the leader- Sample 4 and Sample 5 were composed of adults who partici-
ship/authority and grandiose exhibitionism subscales appear to be pated via Amazon’s Mechanical Turk (MTurk) website. MTurk
more specific markers of grandiose narcissism, as they are asso- allows for the collection of data from individuals using an online
ciated with higher self-esteem (Ackerman et al., 2011; Brown et approach and results in more diverse samples than American
al., 2009) and extraversion and lower neuroticism (Ackerman et undergraduate samples (see Buhrmester, Kwang, & Gosling, 2011,
al., 2011). It is important to note that the three NPI-40 subscales for a review). Sample 4 participants were 306 adults (mean age ⫽
also manifest a number of converging relations with important 29.66, SD ⫽ 10.16), the majority of which came from India
external criteria. For instance, all three are correlated with alter- (45.1%) and the United States (41.5%), and the rest came from
native measures of exploitativeness, entitlement, antagonism, and various other countries.3 Of these, 42.4% were women, 44.1%
aspects of psychopathy (Ackerman et al., 2011). Nonetheless, were White, and 61.8% reported that English was their first lan-
given that these subscales differ in certain important ways, it guage. Sample 5 participants were 277 adults (mean age ⫽ 31.34,
would be advantageous for any brief measure of narcissism to SD ⫽ 10.98) living in the United States. Of these, 64.6% were
maintain a factor structure of this nature. The NPI-13 was created women, 73.9% were White, and 97.5% said English was their first
with this goal in mind, so that both total and subscale scores could language. Data from these samples were previously published in
be used in future research. Miller, Gentile, Wilson, and Campbell (2013) and Miller, Price,
Gentile, et al. (2012), respectively.
The Present Research Sample 6 was composed of 48 White clinical outpatients (60.4%
In the studies presented here, we created a new short measure of women, mean age ⫽ 31.65, SD ⫽ 10.47) who were recruited via
grandiose narcissism, the NPI-13, and tested its validity in con- advertisements placed in an outpatient psychology clinic and local
junction with that of the NPI-16. In Study 1, we used archival data newspapers. Potential participants were screened for eligibility
from a variety of samples to examine the relations between the based on three inclusion criteria: aged 18 – 60, currently seeing a
NPI-13, NPI-16, NPI-40, and a number of alternative measures of psychologist or psychiatrist, and absence of psychotic symptoms.
narcissism. We also examined these three versions of the NPI in A comparison of the sample to established norms on the General
relation to a number of constructs considered important to narcis- Severity Index (GSI) of the Symptom Checklist-90-Revised (SCL-
sism’s nomological network. These included self, parental, and 90; Derogatis, 1975) indicated that they were comparable in terms
“thin-slice” ratings1 of five-factor model (FFM) personality traits, of psychological symptoms to a psychiatric outpatient sample.
as well as “near neighbor” personality disorders (e.g., antisocial, Data from this sample were previously published in Miller,
histrionic, psychopathy), pathological personality traits from the Gaughan, Pryor, Kamen, and Campbell (2009).
new Diagnostic and Statistical Manual of Mental Disorders (5th
ed.; DSM–5; www.dsm5.org) trait model of personality disorder, 1
A thin-slice is a brief (e.g., 60 s) video-recorded clip of an individual’s
and affect. We likewise examined the three measures in relation to behavior that is then coded by blind-raters for various personality traits.
several outcomes including lifetime histories of externalizing be- The clip can involve a person answering questions about themselves,
haviors (e.g., antisocial behavior; substance use), as well as ag- performing an activity, or interacting with others in a group setting. The
gression manifested in a laboratory paradigm. Finally, we tested purpose of thin-slice ratings is to assess how much information regarding
an individual’s personality can be gleaned from a first impression.
the factor structure of the NPI-13 and NPI-16. Based on these 2
Data from all samples were screened for excessive missing data or
results, we examined the differential pattern of relations between random responding (e.g., high numbers of consecutive answers of the same
the three factors extracted from the NPI-13 and constructs in the number such as “1”).
3
nomological network of narcissism. In Study 2, participants com- A minority of participants (i.e., 2.5% or less) stated they were from the
following countries: Argentina, Austria, Bangladesh, Brazil, Bulgaria,
pleted separate versions of the NPI-13, NPI-16, and NPI-40 to Canada, China, Cuba, Germany, Greece, Israel, Italy, Mexico, Netherlands,
again examine the relations between the three versions and assess Pakistan, Philippines, Poland, Romania, Serbia, Singapore, Slovakia, Swe-
the time savings associated with the brief measures. den, United Kingdom, and Ukraine.
 TEST OF TWO BRIEF MEASURES 3

All participants were originally administered the NPI-40, from Hypersensitive Narcissism Scale (HSNS). The HSNS (Hen-
which the NPI-16 and NPI-13 were derived. Table 1 lists the din & Cheek, 1997) is a 10-item self-report measure of vulnerable
criterion measures that were available in each sample. narcissism. Alphas ranged from .66 (Sample 1) to .81 (Sample 5).
Scale construction of the NPI-13. The goal in creating the Psychological Entitlement Scale (PES). The PES (Campbell,
NPI-13 was to develop a brief measure of narcissism while main- Bonacci, Shelton, Exline, & Bushman, 2004) is a nine-item self-
taining the three-factor structure explicated by Ackerman et al. report measure of the extent to which individuals believe that they
(2011). To that end, we first selected the NPI-40 items with the are more deserving than others. Items are scored on a 1 (strong
highest factor loadings on the leadership/authority (LA), grandiose disagreement) to 7 (strong agreement) scale. Alphas were .86
exhibitionism (GE), and entitlement/exploitativeness (EE) factors (sample 2) and .88 (samples 4 and 5).
from the Ackerman et al. (2011) analyses. Since there were only Rosenberg Self-Esteem Scale (RSE). The RSE (Rosenberg,
four items for the entitlement/exploitativeness factor, all four items 1965) is a 10-item measure of global self-esteem. Alphas ranged
were included. from .88 (Sample 4) to .91 (Sample 5).
Next, we examined whether these items were also categorized as Personality measures.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

being among the most prototypical of narcissism via expert ratings, Revised NEO Personality Inventory (NEO-PI-R). The NEO-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

item-response theory (IRT) analyses, and exploratory factor anal- PI-R (Costa & McCrae, 1992) is a 240-item self-report measure of
yses described in Rosenthal and Hooley (2010). We gave prefer- the five-factor model (FFM), which includes the domains of Neu-
ence to items that were identified by all three techniques as being roticism, Extraversion, Openness to Experience, Agreeableness,
indicative of the prototypical description of narcissism but in- and Conscientiousness. Alphas for the domains ranged from .87 to
cluded items identified by only one or two of these metrics so long .92, .89 to .91, and .86 to .94 for Samples 1, 2, and 6, respectively.
as they corresponded to those selected from Ackerman et al. Parental reports of FFM personality. Parental ratings of
(2011). personality were collected from participants in Sample 1. A
In all, we selected a final pool of 13 items (4, 12, 13, 14, 15, 19, packet containing several questionnaires was sent to the homes
20, 24, 25, 27, 29, 32, 36; ␣ ⫽ .73). These included four LA (␣ ⫽ of participants’ parents. The parent(s) completed an informant
.66, mean interitem r ⫽ .32) items, five GE items (␣ ⫽ .65, mean version of the NEO Five-Factor Inventory (NEO-FFI; Costa &
interitem r ⫽ .26), and four EE items (␣ ⫽ .51, mean interitem McCrae, 1992), a 60-item measure of the FFM domains. Alphas
r ⫽ .21).4 Although the GE subscale initially had four items, an for these domains ranged from .63 (Openness) to .90 (Consci-
additional item was chosen to improve its internal consistency. entiousness).
Given the item constraints on the EE subscale, it was not possible Thin-slice ratings. Using the protocol described by Oltmanns,
to boost reliability by adding additional items. However, lower Friedman, Fiedler, and Turkheimer (2004), each participant in
reliability in this subscale is not uncommon and does not appear to Sample 2 was individually videotaped for 60 s while answering the
limit its correlations with important external criteria (Ackerman et question: “What do you enjoy doing?” Each clip was rated by an
al., 2011). average of 11 raters who were doctoral students in a clinical
Narcissism measures. psychology program. The graduate students rated the clips on the
Narcissistic Personality Inventory (NPI-40). The NPI-40 following constructs (using one item per construct) using a 5-point
(Raskin & Terry, 1988) is a 40-item self-report measure of trait Likert scale: Neuroticism, Extraversion, Openness to Experience,
narcissism. The reliability across samples (N ⫽ 1,316) was .87. Agreeableness, Conscientiousness, physical attractiveness, likabil-
Narcissistic Personality Inventory–16 (NPI-16). The NPI-16 ity, and narcissism. The five personality domain descriptions were
(Ames et al., 2006) is a 16-item self-report measure of trait consistent with FFM definitions (e.g., Costa & McCrae, 1992). No
narcissism derived from the NPI-40. The reliability across samples descriptors were given for physical attractiveness. Likability was
(N ⫽ 1,316) was .75. gauged with the question “How likable do you find this individual
Pathological Narcissism Inventory (PNI). The PNI (Pincus et (would you want to get to know him/her better)?” For narcissism,
al., 2009) is a 52-item self-report measure of both vulnerable and raters were given several descriptors (i.e., self-centered, grandiose,
grandiose narcissism traits. The PNI contains four vulnerable and overly confident) to go with the “narcissistic” label. Intraclass
narcissism subscales (i.e., Contingent Self-Esteem, Hiding the correlations (ICCs) indicated that interrater reliability was high,
Self, Devaluing, and Entitlement Rage) and three grandiose nar- ranging from .77 (likability) to .92 (physical attractiveness), with
cissism subscales (i.e., Self-Sacrificing Self-Enhancement, Gran- a median of .86. Composites were created by taking the mean of all
diose Fantasies, and Exploitativeness). Alphas in Sample 2 ranged available ratings.
from .74 to .94, and alphas in Sample 5 ranged from .82 to .95. Interpersonal Adjective Scales (IAS). The IAS (Wiggins,
Narcissistic Grandiosity Scale (NGS). The NGS (Rosenthal, 1995) contains 64 adjectives, scored on a 1 to 8 scale, that provide
Hooley, & Steshenko, 2007) is a measure of grandiose narcissism, scores on the interpersonal circumplex (IPC). The scale includes
which requires participants to rate themselves on 16 adjectives eight octant scores and scores on the two primary axes of domi-
such as “superior” and “omnipotent” on a 1 (not at all) to 7
(extremely) scale. The reliability was the same in both Sample 4 4
With the exception of Sample 6, which exhibited low alpha coeffi-
and Sample 5 (␣ ⫽ .96). Scores from the NGS are significantly cients on the LA (␣ ⫽ .19) and EE (␣ ⫽ .33) subscales, the reliabilities
correlated with other measures of grandiose narcissism and traits were similar across samples (LA ␣s ranged .55–.72, GE ␣s ranged .52–.68,
and EE ␣s ranged .41–.62). Despite the lower internal consistencies in
associated with narcissism such as agreeableness and extraversion Sample 6 for two of the subscales, these subscales generally manifested
(e.g., Miller, Price, & Campbell, 2012; Miller, Price, Gentile, et al., similar patterns of correlations with external criteria to those found in the
2012). nonclinical samples.
 4 GENTILE ET AL.

Table 1
External Criterion Measures Available in Each Sample

Measure Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6

N 361 238 86 306 277 48

Narcissism/self-esteem
PNI X X
NGS X X
HSNS X X X X
PES X X X
RSE X X X X
Personality
NEO-PI-R X X X
FFM parental report X
FFM thin-slice ratings X
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

IAS X
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Personality disorders
SCID-II-PQ X (NPD) X X (NPD) X (NPD)
SCID-II-PQ Interview X
PID5 X
Psychopathy
SRP-III X
Symptoms/affect
BSI X X
PANAS-X X X
Behavioral outcomes
CAB X
RCAP X
Note. PNI ⫽ Pathological Narcissism Inventory; NGS ⫽ Narcissistic Grandiosity Scale; HSNS ⫽ Hypersensitive Narcissism Scale; PES ⫽ Psychological
Entitlement Scale; RSE ⫽ Rosenberg Self-Esteem Scale; NEO-PI-R ⫽ Revised NEO Personality Inventory; FFM ⫽ five-factor model; IAS ⫽ Interpersonal
Adjective Scales; SCID-II-PQ ⫽ Structured Clinical Interview for DSM–IV Personality Disorders—Personality Questionnaire; PID5 ⫽ Personality
Inventory for DSM–5; SRP-III ⫽ Self-Report Psychopathy Scale, Version III; BSI ⫽ Brief Symptom Inventory; PANAS-X ⫽ Positive and Negative Affect
Scale—Expanded Form; CAB ⫽ Crime and Analogous Behavior Scale; RCAP ⫽ Response Choice Aggression Paradigm; NPD ⫽ narcissistic personality
disorder; DSM–IV ⫽ Diagnostic and Statistical Manual of Mental Disorders (4th ed.); DSM–5 ⫽ Diagnostic and Statistical Manual of Mental Disorders
(5th ed.).

nance and nurturance. The alphas for the octants ranged from .79 personality disorder scores (Hopwood, Thomas, Markon,
(Unassuming-Ingenuous) to .91 (Cold-hearted). Wright, & Krueger, 2012).
Personality disorders. Psychopathy measures: Self-Report Psychopathy Scale–III
Structured Clinical Interview for DSM–IV Personality Disor- (SRP-III). The SRP-III (Paulhus, Neumann, & Hare, in press)
ders—Personality Questionnaire (SCID-II-PQ). The SCID- is a 64-item self-report measure of psychopathy that has four
II-PQ (First, Gibbon, Spitzer, Williams, & Benjamin, 1997) is a subscales. Factor 1 psychopathy is measured by the Interper-
119-item self-report measure that assesses the diagnostic criteria sonal Manipulation (SRP-IPM; ␣ ⫽ .86) and Callous Affect
for the Diagnostic and Statistical Manual of Mental Disorders (4th (SRP-CA; ␣ ⫽ .80) scales, whereas Factor 2 psychopathy is
ed., text rev.; DSM–IV–TR; American Psychiatric Association, measured by the Erratic Lifestyle (SRP-ELS; ␣ ⫽ .81) and
1994) personality disorders. In Sample 2, the full scale was used, Antisocial Behaviors (SRP-ASB; ␣ ⫽ .78) scales. The SRP-III
and it manifested reliabilities ranging from .44 (obsessive-
scales demonstrate substantial correlations with alternative
compulsive) to .89 (antisocial). In Samples 1, 4, and 5, only the
measures of psychopathy (Few, Miller, & Lynam, 2013; Seib-
NPD subscale was used, with alphas ranging from .65 to .82. In
ert, Miller, Few, Zeichner, & Lynam, 2011) and have a well-
Sample 6, a SCID-II-PQ semistructured interview was used to
validated factor structure (e.g., Neal & Sellbom,
assess NPD (␣ ⫽ .76; interrater reliability ICC ⫽ .77).
2012).
Personality Inventory for DSM–5 (PID5). The PID5
(Krueger, Derringer, Markon, Watson, & Skodol, 2012) is a Symptom and affect measures.
220-item self-report measure that was created to assess the 25 Brief Symptom Inventory (BSI). The BSI (Derogatis &
personality traits proposed for use as part of a new alternative Melisaratos, 1983) is a 53-item measure of psychological symp-
diagnostic model for personality disorders in the DSM–5 (to be toms experienced in the past week. It includes specific symptom
included in Section 3 in order to stimulate further research on scales and a global severity index (GSI). We report only on the
this approach). Items are scored on a 0 (Very false or Often GSI (Samples 1 and 2: ␣s ⫽ .97).
False) to 3 (Very True or Often True) scale. Alphas across Positive and Negative Affect Schedule—Expanded Form
facets ranged from .68 to .94. The PID5 scales manifest good (PANAS-X). The PANAS-X (Watson & Clark, 1994) is a
structural validity (Wright et al., 2012) and strong correlations 60-item self-report measure of affect. Here, we report on the
with DSM–IV–TR (American Psychiatric Association, 2000) subscales of positive (10 items; Sample 1: ␣ ⫽ .87; Sample 2:
 TEST OF TWO BRIEF MEASURES 5

␣ ⫽ .84) and negative affect (10 items; Sample 1: ␣ ⫽ .83; relations, 95% confidence intervals were computed and used to
Sample 2: ␣ ⫽ .85). compare correlations across the three NPI measures.5
Behavioral outcome measures. First, we examined the relations between the three NPI scales
Crime and Analogous Behavior Scale (CAB). The CAB and other measures of narcissism and self-esteem (see Table 2).
(Miller & Lynam, 2003) is a self-report inventory that assesses All three scales manifested nearly identical correlations with the
externalizing behaviors such as substance use, antisocial behavior, PNI, HSNS, PES, and both interview and self-report measures of
gambling, and intimate partner violence. An alcohol use variable NPD. In general, the three NPI scales were significantly positively
was created by averaging five standardized variables (i.e., use of correlated with grandiose scales from the PNI and NGS, psycho-
alcohol, age of first use, current pattern of use, ever binge drinking, logical entitlement, and both self-report and interview-based
number of binge drinking episodes during the past month). A symptom of NPD. All three were also positively related to self-
lifetime substance use variety count was created by giving partic- esteem scores, although the correlation manifested by the NPI-13
ipants a “1” for every substance they endorsed using (five items; (r ⫽ .15) was significantly smaller than the corresponding corre-
e.g., marijuana). A lifetime antisocial behavior count was created lation manifested by the NPI-40 (r ⫽ .28).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

by giving participants a “1” for every relevant act they endorsed Relations between NPI scales and personality measures.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

(nine items; e.g., stealing). A lifetime gambling (GAMB) count Next, we examined the relations between the three NPI scales and
was created by giving participants a “1” for every relevant act they both self- and other-rated FFM traits. The three NPI scales man-
endorsed (six items; e.g., played card or other games for money).
ifested similar correlations across the NEO-PI-R domains (see
Last, a lifetime intimate partner violence (IPV) count was created
Table 3).6 For instance, all were negatively related to Agreeable-
using this same approach (six items; e.g., slapped my partner). The
ness and Neuroticism, and positively related to Extraversion. The
antisocial and IPV variables were log-transformed prior to use.
only significant difference was between the NPI-13 and the
Response Choice Aggression Paradigm (RCAP). Participants
NPI-40 correlation with the domain of Extraversion (NPI-13: r ⫽
in Sample 3 completed the RCAP (Zeichner, Frey, Parrott, &
.25, NPI-40: r ⫽ .41). With regard to parental-reports and thin-
Butryn, 1999), ostensibly a reaction time competition with another
slice ratings of FFM domains, there were no significant differences
person (see Reidy et al., 2008, for details). Participants were told
that they could administer shocks ranging from “1 to 10” to their across the three NPI measures (see Table 3). There were also no
opponent (no opponent was actually involved; participants re- differences in the relations manifested by the three NPI scores with
ceived shocks from and gave shocks to a computer). After each the thin-slice ratings of likability, attractiveness, and narcissism; in
trial, participants received feedback on whether they had won or general, the NPI measures were positively related to thin-slice
lost and were allowed to administer shocks to the other person, ratings of all three. Overall, the self-, parental-report, and thin-slice
regardless of the outcome. They, in turn, could receive shocks ratings produced similar relationships between the FFM domains
from the “other person” (actually sent at random intervals by a and narcissism across the three scales. Of note however, the low
computer). Shocks were administered via two electrodes attached Agreeableness manifested on the self- and parental-reports was not
to two fingers of the participant’s nondominant hand. corroborated by the thin-slice ratings, nor was the high Extraver-
We combined three aggression scores (i.e., shock intensity, sion manifested by the self-report and thin-slice ratings corrobo-
shock duration, and shock frequency) to create an aggression rated by the parental-reports. Last, no significant differences were
composite (␣ ⫽ .80). Shock intensity was the average intensity of found between the NPI scales and the IPC octants. In general, all
shocks for trials in which the participant administered a shock. three NPI scores manifested their largest correlations with octants
Shock duration was the average duration of shocks for trials during associated with interpersonal dominance and coldness.
which the participant administered a shock. Shock frequency was Relations between NPI scales and psychological functioning
the number of trials during which the participant chose to admin- measures. We next examined the relations between the three
ister a shock. NPI scales and measures of personality pathology (PID5, SCID-
II-PQ) and psychopathy (SRP-III). No significant differences
were found between the three NPI scales for any of the DSM–5
Results personality disorder (PD) traits, as measured by the PID5,
Relations among NPI scales. The NPI-13 was highly corre- although the correlations with the NPI-13 were generally higher
lated with both the NPI-16 (r ⫽ .83, p ⬍ .001) and the NPI-40 than those with either the NPI-16 or NPI-40 (see Table 4). In
(r ⫽ .87, p ⬍ .001). Likewise, the NPI-16 was highly correlated general, all three produced substantial correlations with the
with the NPI-40 (r ⫽ .90, p ⬍ .001). The high correlation between Antagonism subscales such as Grandiosity and Attention Seek-
the NPI-13 and NPI-16 was partially due to their degree of over- ing, as well as traits from Disinhibition (e.g., Irresponsibility),
lap; seven of the items on the NPI-13 also appear on the NPI-16.
Relations between NPI scales and non-NPI narcissism 5
It should be noted that there is very little overlap between the corre-
measures. Across analyses, we tested whether there were sig- lations presented here and those presented in the previously published
nificant differences between the NPI-13, NPI-16, and NPI-40 with articles from which the data were derived. Of the three versions of the NPI,
measures of external criteria relevant to the nomological network only the NPI-40 was used in the previous studies, and in most cases,
of narcissism. In some cases, measures were unique to a single subscale scores were used in lieu of total scores and then combined into
grandiose narcissism composites.
sample (e.g., SRP-III), whereas in other cases, measures were 6
The relations between the three NPI scales and the 30 FFM facets were
repeated across samples (e.g., NEO-PI-R). In the latter case, we examined and no significant differences were found. The results of these
combined the correlations using a weighted average. For all cor- analyses are provided in the supplemental material.
 6 GENTILE ET AL.

Table 2
Correlations Between Non-NPI Narcissism Measures and the Three NPI Scales

NPI-13 NPI-16 NPI-40

95% CI 95% CI 95% CI
Measure N k Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

PNI
CSE 515 2 .00 ⫺.08 .09 ⫺.02 ⫺.11 .06 ⫺.08 ⫺.17 .01
HS 515 2 ⫺.03 ⫺.12 .05 ⫺.02 ⫺.10 .07 ⫺.06 ⫺.15 .02
D 515 2 .15 .06 .23 .10 .01 .18 .04 ⫺.04 .13
ER 515 2 .32 .24 .39 .29 .21 .37 .25 .17 .33
SSSE 515 2 .07 ⫺.01 .16 .07 ⫺.02 .16 .11 .02 .19
GF 515 2 .25 .17 .33 .24 .16 .32 .26 .18 .34
E 515 2 .39 .31 .46 .45 .38 .52 .51 .45 .58
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

HSNS 1,181 4 .16 .11 .22 .09 .03 .14 .05 ⫺.01 .11
This document is copyrighted by the American Psychological Association or one of its allied publishers.

PES 820 3 .45 .40 .51 .47 .41 .52 .46 .41 .52
NGS 582 2 .62 .56 .66 .58 .53 .64 .65 .60 .69
SCID-II-PQ NPD 1,175 4 .54 .50 .58 .52 .48 .56 .52 .47 .56
SCID-II-PQ NPD Interview 48 1 .61 .39 .76 .56 .33 .73 .54 .30 .71
RSE 1,153 4 .15a .09 .21 .22ab .16 .27 .28b .23 .34
Note. Correlations within each row with different subscripts are significantly different. PNI ⫽ Pathological Narcissism Inventory; CSE ⫽ Contingent
Self-Esteem; HS ⫽ Hiding the Self; D ⫽ Devaluing; ER ⫽ Entitlement Rage; SSSE ⫽ Self-Sacrificing Self-Enhancement; GF ⫽ Grandiose Fantasies;
E ⫽ Exploitativeness; HSNS ⫽ Hypersensitive Narcissism Scale; PES ⫽ Psychological Entitlement Scale; NGS ⫽ Narcissistic Grandiosity Scale; NPD ⫽
narcissistic personality disorder; RSE ⫽ Rosenberg Self-Esteem Scale; NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence interval.

Negative Affectivity (e.g., Hostility), and Psychoticism (e.g., NPI-16: ricc ⫽ .96; NPI-13 vs. NPI-40: ricc ⫽ .95; NPI-16 vs.
Unusual Beliefs/Perceptions). NPI-40: ricc ⫽ .98).
The three NPI scales also manifested similar correlations Factor structure of the NPI-16 and NPI-13.
with the DSM–IV–TR PDs (see Table 5). Specifically, all three NPI-16. Because the NPI-16 was not created with a specific a
evinced moderate positive relations with paranoid, antisocial, priori factor structure, we split the total sample into two random
and histrionic PDs, and moderate negative relations with halves and used EFA (Sample 1) and CFA (Sample 2) to examine
avoidant and dependent PDs. With regard to psychopathy, the its factor structure on the raw data. First, an EFA using principal
three NPI scales were similarly and strongly positively corre- axis factoring with direct oblimin rotation was performed on the
lated with all four of the SRP-III subscales; the strongest NPI-16 items. Five factors had eigenvalues of 1 or greater; the first
correlations were with the two “Factor 1” psychopathy scales seven eigenvalues were 3.52, 1.49, 1.20, 1.12, 1.00, .92, and .89.
(Interpersonal Manipulation; Callous Affect). In terms of over- An examination of the scree plot suggested that a two-factor
all psychological distress, as measured by the BSI’s Global solution was appropriate. Parallel Analysis (PA) and the Minimum
Severity Index, all three scales manifested similarly small neg- Average Partial (MAP; Velicer, 1976) methods were also utilized
ative correlations that did not differ from one another (see Table to identify the optimal number of factors. The results from the PA
5). With regard to positive and negative affect (see Table 5), all suggested that up to six factors could be extracted, although the
three NPI scales were unrelated to negative affect and posi-
real eigenvalues for Factors 4 – 6 were very similar to those gen-
tively related to positive affect.
erated by the random data (mean difference between real eigen-
Relations between NPI scales and behavioral outcome
values and random data for Factors 4 – 6: .03). Finally, the results
measures. We next examined the relations between the NPI
from the MAP test suggested that two factors should be extracted.
scales and several outcome measures (see Table 5). With regard
On the basis of the EFA results, we next used confirmatory
to externalizing behaviors, the three NPI scales manifested
factor analysis to test five competing models underlying the struc-
similar correlations with no significant differences among the
ture of the NPI-13 and NPI-16 (see Table 6).7 The fit of all CFA
correlations. In general, the three NPI scales produced small
positive correlations with histories of antisocial behavior, gam- models was evaluated in accordance with four fit indices: (a) the
bling, and alcohol use. All three scales were also equally chi-square goodness-of-fit test (␹2), (b) the comparative fit index
correlated with aggression manifested in a well-validated be- (CFI; Bentler, 1990), (c) the root-mean-square error of approxi-
havioral paradigm. mation (RMSEA; Steiger, 1990), and (d) the standardized root-
Intraclass correlations between the NPI scales. To examine mean-square residual (SRMR; Bentler, 1995). Researchers gener-
the similarity of the relations produced by the three NPI scales ally agree that CFI values greater than .90 indicate adequate fit,
across all the aforementioned criteria, we calculated second- while values at or above .95 indicate good fit; RMSEA values less
order intraclass correlations of the bivariate correlations man-
ifested by the three NPI scores across all the external criteria 7
For the NPI-16, we did not test the six-factor solution as the sixth factor
reported in Tables 2 through 5. The pattern of correlations was composed of a single item. Thus, this model could not be tested in
manifested by the NPI scales were nearly identical (NPI-13 vs. CFA and would not be practically useful.
 TEST OF TWO BRIEF MEASURES 7

Table 3
Correlations Between the Five-Factor Model and Interpersonal Circumplex Personality Measures and the Three NPI Scales

NPI-13 NPI-16 NPI-40

95% CI 95% CI 95% CI
Measure N k Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

FFM self-report
Neuroticism 636 3 ⫺.16 ⫺.24 ⫺.09 ⫺.23 ⫺.30 ⫺.16 ⫺.28 ⫺.35 ⫺.21
Extraversion 636 3 .25a .17 .32 .29ab .22 .36 .41b .34 .47
Openness 636 3 .04 ⫺.04 .11 .05 ⫺.02 .13 .09 .01 .17
Agreeableness 636 3 ⫺.53 ⫺.58 ⫺.47 ⫺.51 ⫺.57 ⫺.45 ⫺.50 ⫺.56 ⫺.44
Conscientiousness 636 3 .03 ⫺.05 .11 .07 .00 .15 .11 .03 .19
FFM parental report
Neuroticism 143 1 ⫺.16 ⫺.32 .00 ⫺.22 ⫺.37 ⫺.06 ⫺.17 ⫺.33 ⫺.01
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Extraversion 143 1 .05 ⫺.12 .21 .15 ⫺.01 .31 .12 ⫺.05 .28
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Openness 143 1 .20 .04 .35 .06 ⫺.11 .22 .07 ⫺.10 .23
Agreeableness 143 1 ⫺.21 ⫺.36 ⫺.05 ⫺.18 ⫺.33 ⫺.02 ⫺.21 ⫺.36 ⫺.05
Conscientiousness 143 1 ⫺.08 ⫺.24 .09 ⫺.01 ⫺.17 .15 ⫺.06 ⫺.22 .11
FFM thin-slice ratings
Neuroticism 230 1 ⫺.27 ⫺.39 ⫺.15 ⫺.19 ⫺.31 ⫺.06 ⫺.29 ⫺.40 ⫺.17
Extraversion 230 1 .38 .26 .49 .28 .16 .40 .40 .29 .50
Openness 230 1 .10 ⫺.03 .23 .10 ⫺.03 .23 .12 ⫺.01 .25
Agreeableness 230 1 ⫺.11 ⫺.24 .02 ⫺.09 ⫺.22 .04 ⫺.09 ⫺.22 .04
Conscientiousness 230 1 ⫺.16 ⫺.28 ⫺.03 ⫺.19 ⫺.31 ⫺.06 ⫺.16 ⫺.28 ⫺.03
Attractiveness 230 1 .21 .08 .33 .12 ⫺.01 .25 .23 .10 .35
Likebility 230 1 .17 .04 .29 .11 ⫺.02 .24 .20 .07 .32
Narcissism 230 1 .33 .21 .44 .28 .16 .40 .35 .23 .46
IAS
Assured-Dominant 277 1 .41 .31 .50 .47 .37 .56 .53 .44 .61
Arrogant-Calculating 277 1 .46 .36 .55 .48 .38 .57 .53 .44 .61
Cold-Hearted 277 1 .48 .38 .57 .43 .33 .52 .42 .32 .51
Aloof-Introverted 276 1 .05 ⫺.07 .17 ⫺.03 ⫺.15 .09 ⫺.10 ⫺.22 .02
Unassured-Submissive 277 1 ⫺.23 ⫺.34 ⫺.12 ⫺.32 ⫺.42 ⫺.21 ⫺.38 ⫺.48 ⫺.27
Unassuming-Ingenuous 277 1 ⫺.25 ⫺.36 ⫺.14 ⫺.28 ⫺.39 ⫺.17 ⫺.34 ⫺.44 ⫺.23
Warm-Agreeable 277 1 ⫺.37 ⫺.47 ⫺.26 ⫺.34 ⫺.44 ⫺.23 ⫺.30 ⫺.40 ⫺.19
Gregarious-Extraverted 277 1 .12 .00 .23 .17 .05 .28 .28 .17 .39
Note. Correlations within each row with different subscripts are significantly different. NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence
interval; FFM ⫽ five-factor model; IAS ⫽ Interpersonal Adjective Scales.

than .08 and SRMR values less than .10 indicate acceptable fit, 6). In addition, restricting LA to load on a common factor with GE
whereas RMSEA values below .06 and SRMR values below .08 resulted in a significant decrement in fit. Accordingly, we adopted
indicate good fit (e.g., Hu & Bentler, 1999; Lance & Vandenberg, the three-factor model for subsequent analyses. The standardized
2002). Finally, we compared parameter-nested competing models factor loadings for this solution are presented in Table 7.
using ⌬␹2. Each of the five competing structures provided a Relations between NPI-13 subscales and non-NPI narcissism
reasonable approximation of the NPI-16 data, but only the five- measures. The three NPI-13 subscales manifested small to mod-
factor model met the criteria for an acceptable fit in terms of erate intercorrelations (LA vs. GE: r ⫽ .35; LA vs. EE: r ⫽ .34;
RMSEA, SRMR, and CFI. Clearly, the successively increasing GE vs. EE: r ⫽ .20) and, as expected, manifested evidence of both
complexity of the models contributed to the findings. That is, the convergence and divergence with the external criteria, including
five-factor model is practically problematic because it is not par- the alternative measures of narcissism (see Table 8).8 In general,
simonious, and multiple factors include three or fewer items,
only the NPI-13 EE subscale manifested significant correlations
which is below recommended standards for CFA. Ultimately, none
of the models provided a particularly close fit to the data.
NPI-13. We next tested the structure of the NPI-13 using 8
The NPI-40 subscales were similarly intercorrelated (LA vs. GE: r ⫽
confirmatory factor analysis (EFA was not used as was done with .49; LA vs. EE: r ⫽ .31; GE vs. EE: r ⫽ .21). Comparing the relations
the NPI-16 because the NPI-13 was designed based on an a priori manifested by the NPI-13 and NPI-40 LA and GE subscales with the
external criteria reveled significant differences on only two variables:
three-factor structure). CFA was used to compare the a priori self-esteem (NPI-13 LA r ⫽ .17 vs. NPI-40 LA r ⫽ .33) and self-reported
three-factor model to one-and two-factor models. The two-factor extraversion (NPI-13 GE r ⫽ .31 vs. NPI-40 GE r ⫽ .45). Comparisons
model combined the LA and GE subscales, as previous research were not made between the NPI-13 and NPI-40 EE subscales, as they
has suggested these are more uniformly grandiose than the EE contain the same items. The NPI-13 and NPI-40 subscales were highly
correlated (LA r ⫽ .86 and GE r ⫽ .88), and the profiles of correlations
subscale (Ackerman et al., 2011; Brown et al., 2009). Although the manifested by each with the external criteria were almost identical (NPI-13
one- and two-factor models provided a reasonable approximation LA vs. NPI-40 LA: ricc ⫽ .94; NPI-13 GE vs. NPI-40 GE: ricc ⫽ .97). The
to the data, the three-factor model provided a closer fit (see Table full results of these analyses are provided in the supplemental material.
 8 GENTILE ET AL.

Table 4
Correlations Between the PID5 Traits and the Three NPI Scales

NPI-13 NPI-16 NPI-40

95% CI 95% CI 95% CI
Measure Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

Negative affectivity
Emotional lability .27 .16 .37 .16 .05 .27 .17 .06 .28
Anxiousness .06 ⫺.05 .17 ⫺.09 ⫺.20 .02 ⫺.10 ⫺.21 .01
Separation insecurity .33 .23 .43 .23 .12 .33 .25 .14 .35
Perseveration .24 .13 .34 .12 .01 .23 .10 ⫺.01 .21
Submissiveness .11 .00 .22 .02 ⫺.09 .13 .03 ⫺.08 .14
Hostility .40 .30 .49 .31 .20 .41 .29 .18 .39
Restricted affect .19 .08 .30 .17 .06 .28 .15 .04 .26
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Depressivity .18 .07 .29 .07 ⫺.04 .18 .04 ⫺.07 .15
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Suspiciousness .30 .19 .40 .19 .08 .30 .21 .10 .31
Detachment
Withdrawal .09 ⫺.02 .20 .01 ⫺.10 .12 ⫺.01 ⫺.12 .10
Anhedonia .07 ⫺.04 .18 ⫺.04 ⫺.15 .07 ⫺.08 ⫺.19 .03
Intimacy avoidance .19 .08 .30 .13 .02 .24 .11 .00 .22
Antagonism
Manipulativeness .50 .41 .58 .44 .34 .53 .49 .40 .57
Deceitfulness .46 .37 .54 .37 .27 .46 .39 .29 .48
Grandiosity .57 .49 .64 .51 .42 .59 .57 .49 .64
Attention seeking .59 .51 .66 .54 .46 .61 .57 .49 .64
Callousness .45 .36 .54 .35 .25 .44 .35 .25 .44
Disinhibition
Irresponsibility .40 .30 .49 .31 .20 .41 .28 .17 .38
Impulsivity .36 .26 .45 .26 .15 .36 .26 .15 .36
Rigid perfectionism .29 .18 .39 .19 .08 .30 .20 .09 .31
Distractibility .22 .11 .32 .10 ⫺.01 .21 .06 ⫺.05 .17
Risk taking .25 .14 .35 .26 .15 .36 .36 .26 .45
Psychoticism
Unusual beliefs/perceptions .40 .30 .49 .36 .26 .45 .37 .27 .46
Eccentricity .18 .07 .29 .11 .00 .22 .08 ⫺.03 .19
Cognitive/perceptual dysregulation .41 .31 .50 .31 .20 .41 .31 .20 .41
Note. N ⫽ 306. k ⫽ 1. Correlations within each row with different subscripts are significantly different. PID5 ⫽ Personality Inventory for DSM–5; NPI ⫽
Narcissistic Personality Inventory; CI ⫽ confidence interval; DSM–5 ⫽ Diagnostic and Statistical Manual of Mental Disorders (5th ed.).

with measures/subscales of vulnerable narcissism (PNI: Contin- vs. NPD: ricc ⫽ .75; GE vs. NPD: ricc ⫽ .68; EE vs. NPD: ricc ⫽
gent Self-Esteem, Hiding the Self, Devaluing, and Entitlement .59).
Rage; HSNS), whereas all three subscales manifested significant In terms of the interpersonal circumplex, derived from the IAS,
positive correlations with measures/subscales of grandiose narcis- the three subscales differed such that LA was most strongly related
sism (PNI: Grandiose Fantasies, Exploitativeness; Narcissistic to interpersonal dominance and EE was most strongly related to
Grandiosity Scale), psychological entitlement, and NPD. The interpersonal coldness; GE manifested correlations that typically
NPI-13 subscales also manifested divergent relations with self- fell between those manifested by LA and EE.
esteem, with NPI-13 LA and GE manifesting positive correlations Relations between NPI-13 subscales and psychological func-
and NPI-13 EE manifesting a negative correlation. tioning measures. With regard to the DSM–5 personality disor-
Relations between NPI-13 subscales and personality der traits, the NPI-13 subscales manifested similar correlations
measures. In terms of their relations to personality traits, the with facets measuring Antagonism and with unusual beliefs or
NPI-13 LA and GE subscales manifested a pattern of correlations perceptions from the Psychoticism domain (see Table 10). The
with FFM ratings (self-, parental-report, thin slices) that is char- subscales differed with regard to their correlations with many of
acteristic of grandiose narcissism (i.e., negative correlations with the facets from the Negative Affectivity (e.g., hostility, anxious-
Agreeableness and Neuroticism, positive correlation with Extra- ness, depressivity), Detachment (e.g., anhedonia), and Disinhibi-
version), whereas the EE subscale manifested a pattern of corre- tion (e.g., irresponsibility, distractibility) domains such that the
lations that falls between that of vulnerable and grandiose narcis- NPI-13 EE subscale manifested stronger positive correlations.
sism, as its primary FFM correlate is Agreeableness (see Table 9).9
We also calculated similarity scores between the three NPI-13
9
subscales with the meta-analytically derived correlational profile The relations between the three NPI subscales and the self-report FFM
facets were examined, and significant differences were found in the ma-
of NPD (Samuel & Widiger, 2008) using intraclass correlations. jority of facets (Neuroticism: 5, Extraversion: 6, Openness: 2, Agreeable-
All three manifested significant intraclass correlations with the ness: 6, and Conscientiousness: 4). The results of these analyses are
meta-analytic profile of NPD from the perspective of the FFM (LA provided in the supplemental material.
 TEST OF TWO BRIEF MEASURES 9

Table 5
Correlations Between Psychopathology, Psychopathy, and Behavioral Outcome Measures and the Three NPI Scales

NPI-13 NPI-16 NPI-40

95% CI 95% CI 95% CI
Measure N k Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

SCID-II-PQ
Paranoid 238 1 .24 .12 .36 .24 .12 .36 .15 .02 .27
Schizoid 238 1 .06 ⫺.07 .19 .00 ⫺.13 .13 ⫺.05 ⫺.18 .08
Schizotypal 238 1 ⫺.01 ⫺.14 .12 ⫺.02 ⫺.15 .11 ⫺.07 ⫺.20 .06
Antisocial 238 1 .26 .14 .37 .29 .17 .40 .24 .12 .36
Borderline 238 1 .02 ⫺.11 .15 .09 ⫺.04 .21 .00 ⫺.13 .13
Histrionic 238 1 .45 .34 .55 .48 .38 .57 .52 .42 .61
Avoidant 238 1 ⫺.31 ⫺.42 ⫺.19 ⫺.29 ⫺.40 ⫺.17 ⫺.42 ⫺.52 ⫺.31
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Dependent 238 1 ⫺.20 ⫺.32 ⫺.07 ⫺.15 ⫺.27 ⫺.02 ⫺.21 ⫺.33 ⫺.09
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Obsessive-compulsive 238 1 .01 ⫺.12 .14 .03 ⫺.10 .16 .00 ⫺.13 .13
SRP-III
SRP-IPM 361 1 .53 .45 .60 .50 .42 .57 .50 .42 .57
SRP-CA 361 1 .42 .33 .50 .39 .30 .47 .39 .30 .47
SRP-ELS 361 1 .33 .23 .42 .28 .18 .37 .34 .25 .43
SRP-ASB 361 1 .27 .17 .36 .26 .16 .35 .23 .13 .33
BSI
GSI 598 2 ⫺.04 ⫺.12 .04 ⫺.05 ⫺.13 .03 ⫺.11 ⫺.19 ⫺.03
PANAS-X
Negative affect 596 2 .04 ⫺.04 .12 .00 ⫺.08 .08 ⫺.04 ⫺.12 .04
Positive affect 596 2 .20 .12 .27 .24 .16 .31 .29 .22 .37
CAB
Alcohol use 361 1 .16 .06 .26 .12 .02 .22 .18 .08 .28
Substance use 361 1 .06 ⫺.04 .16 .04 ⫺.06 .14 .04 ⫺.06 .14
Antisocial behavior 361 1 .23 .13 .33 .21 .11 .31 .21 .11 .31
Gambling 361 1 .23 .13 .33 .23 .13 .33 .23 .13 .33
Intimate partner violence 361 1 .11 .01 .21 .08 ⫺.02 .18 .07 ⫺.03 .17
RCAP
Composite 86 1 .45 .26 .60 .47 .29 .62 .40 .21 .56
Note. Correlations within each row with different subscripts are significantly different. NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence
interval; SCID-II-PQ ⫽ Structured Clinical Interview for DSM–IV Personality Disorders—Personality Questionnaire; SRP-III ⫽ Self-Report Psychopathy
Scale, Version III; IPM ⫽ Interpersonal Manipulation; CA ⫽ Callous Affect; ELS ⫽ Erratic Lifestyle; ASB ⫽ Antisocial Behaviors; BSI ⫽ Brief Symptom
Inventory; PANAS-X ⫽ Positive and Negative Affect Scale—Expanded Form; CAB ⫽ Crime and Analogous Behavior Scale; RCAP ⫽ Response Choice
Aggression Paradigm; DSM–IV ⫽ Diagnostic and Statistical Manual of Mental Disorders (4th ed.); GSI ⫽ General Severity Index.

With regard to DSM–IV–TR PDs, all three NPI-13 subscales man- NPI-13 subscales with the external criteria reported in Tables 8
ifested positive correlations with antisocial and histrionic PDs (see through 11 were examined using intraclass correlations: LA versus
Table 11). The LA and GE subscales were also negatively related GE: ricc ⫽ .86; LA versus EE: ricc ⫽ .66; GE versus EE: ricc ⫽
to avoidant and dependent PDs, whereas the EE subscale was .50.10 LA and GE manifested similar patterns of correlations with
unrelated to both. Last, both the LA and EE subscales were an array of constructs, whereas EE manifested more moderate
positively correlated with paranoid PD. In terms of psychopathy, similarity (although still substantial) with LA and GE.
the correlations between the NPI-13 subscales and the SRP-III
subscales were positive and generally similar, with the GE sub-
Study 2
scale generally producing the smallest correlations.
In terms of general psychological distress, only the NPI-13 EE In discussing the creation of short forms, Smith, McCarthy, and
subscale was positively related to the GSI, although the correlation Anderson (2000) argue that, in order to test whether one’s “short
was small. Similarly, the EE subscale manifested a small positive form has adequate overlapping variance with the full form” (p.
correlation with negative affect and was unrelated to positive 105), the measures must be given separately (vs. deriving the long
affect, whereas the LA and GE subscales showed the inverse and short form scores from the same test administration, as was
pattern. done in Study 1). In Study 2, we address this issue by adminis-
Relations between the NPI-13 subscales and outcome tering all three separate versions of the NPI in order to test the
measures. The three NPI-13 subscales manifested similar posi- degree to which they overlap. In addition, this approach allowed us
tive correlations with alcohol use and gambling, as well as aggres-
sion measured in the laboratory paradigm; there were no differ- 10
The similarities of the profiles of correlations manifested by the three
ences in these correlations across the three NPI-13 subscales. NPI-40 subscales with the external criteria were similar to those of the
Intraclass correlations between the NPI-13 subscales. The NPI-13 (LA vs. GE: ricc ⫽ .85; LA vs. EE: ricc ⫽ .42; GE vs. EE: ricc ⫽
similarities of the profiles of correlations manifested by the three .52), with the exception that the LA versus EE correlation was smaller.
 10 GENTILE ET AL.

Table 6
Model Fit Statistics for CFA Models of the NPI-16 and the NPI-13

Model fit statistics

Model df ␹2 RMSEA SRMSR CFI

NPI-16
1. 1 Factor 104 554.61 .085 .064 .830
2. 2 Factor 103 495.20 .077 .070 .852
3. 3 Factor 102 472.29 .079 .073 .860
4. 4 Factor 98 386.49 .068 .064 .891
5. 5 Factor 94 302.54 .058 .058 .921

NPI-13
1. 1 Factor 65 821.06 .103 .075 .805
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2. 2 Factor 64 702.90 .095 .069 .836

This document is copyrighted by the American Psychological Association or one of its allied publishers.

3. 3 Factor 62 333.41 .059 .056 .931

Model 3 vs. Model 2 ⌬2 ⌬369.49ⴱ
Note. For NPI-16, N ⫽ 657; for NPI-13, N ⫽ 1,316. CFA ⫽ confirmatory factor analysis; NPI ⫽ Narcissistic
Personality Inventory; df ⫽ degrees of freedom; RMSEA ⫽ root-mean-square error of approximation; SRMSR ⫽
standardized root-mean-square residual; CFI ⫽ comparative fit index.
ⴱ
p ⬍ .001.

to examine the time savings associated with the two short forms .84 (NPI-16), and .93 (NPI-40). With the exception of the EE
when compared to the full 40-item measure. subscale, the subscales on both the NPI-13 (LA: ␣ ⫽ .73, GE: ␣ ⫽
.75, EE: ␣ ⫽ .52) and NPI-40 (LA: ␣ ⫽ .87, GE: ␣ ⫽ .82, EE: ␣ ⫽
Method .46) manifested adequate reliability.
Response time. Using Qualtrics metadata, participants were
Participants. The sample was composed of 150 adults (mean timed (in seconds) to assess how long it took them to complete
age ⫽ 33.30, SD ⫽ 12.21) living in the United States. Of these each version of the NPI in order to test whether there were time
48.7% were women, 72.0% were White, and 96.7% reported that savings associated with the brief measures.
English was their first language. The data were collected via
MTurk, and participants were paid $1.00 in exchange for their
Results
participation. Institutional review board approval was obtained for
this study. Relations among NPI scales and subscales. The NPI-13 was
Measures. highly correlated with both the NPI-16 (r ⫽ .84, p ⬍ .001) and
NPI. Participants were administered the NPI-13, NPI-16, and NPI-40 (r ⫽ .88, p ⬍ .001), as was the NPI-16 with the NPI-40
NPI-40 in randomized order with short filler measures placed in (r ⫽ .93, p ⬍ .001). The relations between the subscales of the
between so that no two versions of the NPI were presented back- NPI-13 and NPI-40 are shown in Table 12. The subscales of the
to-back. The three versions had alpha coefficients of .82 (NPI-13), NPI-13 were highly correlated with the corresponding subscales of

Table 7
Standardized Factor Loadings for the Three-Factor NPI-13 Model

Item Narcissistic response LA GE EE

12 I like having authority over other people. .60

27 I have a strong will to power. .60
32 People always seem to recognize my authority. .61
36 I am a born leader. .49
4 I know that I am a good person because everybody keeps telling me so. .35
15 I like to show off my body. .60
19 I like to look at my body. .69
20 I will usually show off if I get the chance. .36
29 I like to look at myself in the mirror. .63
13 I find it easy to manipulate people. .55
14 I insist upon getting the respect that is due me. .35
24 I expect a great deal from other people. .41
25 I will never be satisfied until I get all that I deserve. .50
Note. N ⫽ 1,316. All factor loadings are significant at p ⬍ .01. NPI ⫽ Narcissistic Personality Inventory;
LA ⫽ Leadership/Authority; GE ⫽ Grandiose Exhibitionism; EE ⫽ Entitlement/Exploitativeness. Adapted from
“A Principal-Components Analysis of the Narcissistic Personality Inventory and Further Evidence of Its
Construct Validity,” by R. Raskin & H. Terry, 1988, Journal of Personality and Social Psychology, p. 894.
Copyright by the American Psychological Association.
 TEST OF TWO BRIEF MEASURES 11

Table 8
Correlations Between the Non-NPI Narcissism Measures and the NPI-13 Subscales

LA GE EE
95% CI 95% CI 95% CI
Measure k Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

PNI
CSE 2 ⫺.04a ⫺.13 .04 ⫺.10a ⫺.18 ⫺.01 .17b .09 .26
HS 2 .03a ⫺.05 .12 ⫺.15b ⫺.24 ⫺.07 .07a ⫺.02 .15
D 2 .11a .03 .20 ⫺.04a ⫺.13 .05 .30b .22 .37
ER 2 .25ab .16 .33 .10a .01 .19 .39b .32 .46
SSSE 2 .09 .01 .18 .07 ⫺.02 .15 ⫺.02 ⫺.11 .07
GF 2 .27a .19 .35 .09b .00 .17 .21ab .12 .29
E 2 .37a .30 .44 .20b .11 .28 .31ab .23 .38
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

HSNS 4 .06a .01 .12 .05a .00 .11 .28b .23 .34
This document is copyrighted by the American Psychological Association or one of its allied publishers.

PES 3 .36 .30 .42 .31 .24 .37 .35 .29 .41
NGS 2 .53a .47 .59 .45ab .38 .51 .38b .31 .45
SCID-II-PQ NPD 4 .41 .36 .46 .35 .30 .40 .45 .40 .49
SCID-II-PQ NPD Interview 1 .63 .42 .78 .23 ⫺.06 .48 .36 .08 .58
RSE 4 .17a .11 .22 .21a .15 .26 ⫺.10b ⫺.16 ⫺.04
Note. PNI Ns ⫽ 514 –515, HSNS Ns ⫽ 1180 –1181, PES Ns ⫽ 819 – 820, NGS Ns ⫽ 581–582, SCID-II-PQ NPD Ns ⫽ 1174 –1175, SCID-II-PQ NPD
Interview N ⫽ 48, RSE Ns ⫽ 1152–1153. Correlations within each row with different subscripts are significantly different. PNI ⫽ Pathological Narcissism
Inventory; CSE ⫽ Contingent Self-Esteem; HS ⫽ Hiding the Self; D ⫽ Devaluing; ER ⫽ Entitlement Rage; SSSE ⫽ Self-Sacrificing Self-Enhancement;
GF ⫽ Grandiose Fantasies; E ⫽ Exploitativeness; HSNS ⫽ Hypersensitive Narcissism Scale; PES ⫽ Psychological Entitlement Scale; NGS ⫽ Narcissistic
Grandiosity Scale; NPD ⫽ narcissistic personality disorder; RSE ⫽ Rosenberg Self-Esteem Scale; LA ⫽ Leadership/Authority; GE ⫽ Grandiose
Exhibitionism; EE ⫽ Entitlement/Exploitativeness; NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence interval.

the NPI-40 (NPI-13 LA vs. NPI-40 LA: r ⫽ .82, p ⬍ .001; NPI-13 In Study 1, we sought to avoid this “sin” by conducting a
GE vs. NPI-40 GE: r ⫽ .84, p ⬍ .001; NPI-13 EE vs. NPI-40 EE: comprehensive test of the validity of the two brief forms of the
r ⫽ .86, p ⬍ .001). Likewise, within and between each version the NPI—the NPI-16 (Ames et al., 2006) and the newly created
subscales were highly correlated, although the lowest correlations NPI-13— by comparing their patterns of correlations with a vast
were found with the EE subscale. array of important criteria from narcissism’s nomological network
Time savings. We next examined whether there was a time with those manifested by the NPI-40. The results suggested that
savings associated with the brief measures of the NPI. A within- both the NPI-13 and NPI-16 total scores result in patterns of
subjects analysis of variance with a Greenhouse-Geisser correction convergent and discriminant validity that are nearly identical to
indicated that the response time significantly differed across ver- those manifested by the NPI-40. For instance, both brief forms
sions of the NPI, F(1.26, 187.85) ⫽ 53.90, p ⬍ .001. Paired manifested strong correlations with symptoms of NPD derived
comparisons using a Bonferroni correction indicated that while from interviews and self-reports, alternative measures of narcis-
both the NPI-13 (M ⫽ 64.47, SD ⫽ 63.90) and NPI-16 (M ⫽ 78.65 sism and narcissism-related traits such as entitlement and grandi-
SD ⫽ 91.40) took significantly less time to complete (both p ⬍ osity, and “near neighbor” disorders such as psychopathy and
.001) than the NPI-40 (M ⫽ 209.41, SD ⫽ 212.85; d [NPI-13 vs. histrionic personality disorder. Both brief forms were also associ-
NPI-40] ⫽ ⫺0.92; d [NPI-16 vs. NPI-40] ⫽ ⫺0.80), there was no ated with the general personality traits rated as prototypical of
time savings associated with the NPI-13 over the NPI-16 (d ⫽ NPD and grandiose narcissism by researchers and clinicians in-
⫺0.18). cluding extraversion and disagreeableness (Lynam & Widiger,
2001; Samuel & Widiger, 2004; Thomas, Wright, Lukowitsky,
General Discussion Donnellan, & Hopwood, 2012). Similarly, both brief forms were
Brief personality measures can be invaluable because of their also strongly positively correlated with the two pathological
efficiency. The creation of reliable and valid brief forms can even traits— grandiosity and attention seeking—that are central to the
drive new advances in research as scholars from other content alternative diagnostic approach for NPD that will be included in
areas and disciplines may decide to include brief measures in their DSM–5 (Miller, Gentile, & Campbell, 2012). From an interper-
own research endeavors if the costs of inclusion are minimal. sonal circumplex perspective, both NPI brief forms were related to
There are a number of difficulties and dangers, however, in cre- dominance and coldness, as was expected. Finally, both brief
ating these measures, several of which were documented by Smith forms were positively associated with externalizing behaviors in-
and colleagues (2000). One of the general “sins” that Smith and cluding aggressive behavior manifested in an experimental para-
colleagues (2000) describe in the development of brief forms is “to digm involving electric shocks. In short, both the NPI-13 and
assume that because the new measure is shorter, less validity NPI-16 appear to do an excellent job of replicating the results
evidence is required” (p. 103). These authors go on to state, “a found with the longer NPI-40.
short-form developer must meet the same standards of validity as In Study 2, we individually administered each of the three
are required for any test” (p. 103). versions of the NPI to a new sample of participants in order to
 12 GENTILE ET AL.

Table 9
Correlations Between the Five-Factor Model and Interpersonal Circumplex Personality Measures and the NPI-13 Subscales

LA GE EE
95% CI 95% CI 95% CI
Measure Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

FFM self-report
Neuroticism ⫺.21a ⫺.28 ⫺.13 ⫺.17a ⫺.25 ⫺.09 .07b .00 .15
Extraversion .25a .17 .32 .31a .24 .38 ⫺.12b ⫺.19 ⫺.04
Openness ⫺.01ab ⫺.09 .07 .13a .05 .21 ⫺.08b ⫺.15 .00
Agreeableness ⫺.46a ⫺.52 ⫺.40 ⫺.23b ⫺.31 ⫺.16 ⫺.49a ⫺.55 ⫺.43
Conscientiousness .11a .04 .19 .02ab ⫺.05 .10 ⫺.09b ⫺.17 ⫺.02
FFM parental report
Neuroticism ⫺.13 ⫺.29 .03 ⫺.12 ⫺.28 .05 ⫺.07 ⫺.23 .10
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Extraversion .07 ⫺.10 .23 .13 ⫺.03 .29 ⫺.17 ⫺.33 ⫺.01
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Openness .07 ⫺.10 .23 .19 .03 .34 .17 .01 .33
Agreeableness ⫺.14 ⫺.30 .02 ⫺.17 ⫺.33 ⫺.01 ⫺.13 ⫺.29 .03
Conscientiousness .02 ⫺.14 .18 ⫺.15 ⫺.31 .01 .02 ⫺.14 .18
FFM thin-slice ratings
Neuroticism ⫺.25 ⫺.37 ⫺.12 ⫺.26 ⫺.38 ⫺.13 ⫺.06 ⫺.19 .07
Extraversion .31 .19 .42 .36 .24 .47 .12 ⫺.01 .25
Openness .05 ⫺.08 .18 .08 ⫺.05 .21 .10 ⫺.03 .23
Agreeableness ⫺.14 ⫺.26 ⫺.01 ⫺.01 ⫺.14 .12 ⫺.11 ⫺.24 .02
Conscientiousness ⫺.13 ⫺.26 .00 ⫺.14 ⫺.26 ⫺.01 ⫺.07 ⫺.20 .06
Attractiveness .10 ⫺.03 .23 .23 .10 .35 .12 ⫺.01 .25
Likebility .11 ⫺.02 .24 .22 .09 .34 .03 ⫺.10 .16
Narcissism .32 .20 .43 .26 .13 .38 .13 .00 .26
IAS
Assured-Dominant .46a .36 .55 .24b .13 .35 .20b .08 .31
Arrogant-Calculating .40 .30 .49 .31 .20 .41 .32 .21 .42
Cold-Hearted .33ab .22 .43 .25a .14 .36 .51b .42 .59
Aloof-Introverted ⫺.05a ⫺.17 .07 ⫺.10a ⫺.22 .02 .29b .18 .39
Unassured-Submissive ⫺.30a ⫺.40 ⫺.19 ⫺.22a ⫺.33 ⫺.10 .03b ⫺.09 .15
Unassuming-Ingenuous ⫺.27 ⫺.38 ⫺.16 ⫺.18 ⫺.29 ⫺.06 ⫺.11 ⫺.22 .01
Warm-Agreeable ⫺.25ab ⫺.36 ⫺.14 ⫺.16a ⫺.27 ⫺.04 ⫺.44b ⫺.53 ⫺.34
Gregarious-Extraverted .17a .05 .28 .23a .12 .34 ⫺.18b ⫺.29 ⫺.06
Note. FFM self-report Ns ⫽ 635– 636. k ⫽ 3. FFM parental report N ⫽ 143, FFM thin-slices Ns ⫽ 229 –230, IAS Ns ⫽ 276 –277. k ⫽ 1. Correlations
within each row with different subscripts are significantly different. NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence interval; FFM ⫽ five-factor
model; IAS ⫽ Interpersonal Adjective Scales; LA ⫽ Leadership/Authority; GE ⫽ Grandiose Exhibitionism; EE ⫽ Entitlement/Exploitativeness.

ensure that the substantial overlapping variance found in Study sidering that the measurement of narcissism as a unidimensional
1 was not due to the use of NPI scores derived from a single construct has been challenged in favor of using subscales (Brown
administration. As with Study 1, we again found substantial et al., 2009; Miller, Price, & Campbell, 2012). The three subscales
convergence between participants’ scores on the three mea- manifested some degree of divergence in their relations with the
sures, bolstering our claim that the NPI-13 and NPI-16 are brief, non-NPI narcissism measures and with the variables in the nomo-
yet accurate, measures of trait narcissism that can be used in logical network of narcissism. The LA and GE subscales generated
place of the NPI-40. Furthermore, comparisons of response patterns of correlations that were most similar to one another and
time across the three measures support the notion that the most closely resembled the personality profile of grandiose nar-
NPI-13 and NPI-16 offer a time savings over the NPI-40, cissism (Miller, Hoffman, et al., 2011). The EE subscale, which
although the two brief measures did not differ from one another was primarily associated with antagonism, is not as uniquely
in the amount of time necessary to complete them. The main associated with either grandiose or vulnerable narcissism. Instead,
benefit of the NPI-13 over the NPI-16 is that it preserves the EE appears to be a blend of both dimensions. Although the EE
three-factor structure believed to underlie the NPI-40 (Acker-
subscale is associated with greater vulnerability than the other
man et al., 2011) and, thus, allows for the use of subscales in the
subscales, its relations with these markers (e.g., positive correla-
assessment of narcissism.
tions with neuroticism and psychological distress; negative corre-
lation with self-esteem and extraversion) are substantially smaller
NPI-13 Subscales than those found with explicit measures of vulnerable narcissism
In Study 1, we generated and validated three subscales for the (e.g., Miller et al., 2010; Miller, Hoffman, et al., 2011). Overall,
NPI-13 (i.e., LA, GE, and EE) based on those derived by Acker- our results are consistent with research showing that of the three
man et al. (2011). These subscales were similar to those previously subscales, EE is most closely associated with certain maladaptive
identified as being central to the narcissism construct (Corry et al., traits and outcomes such as psychological distress, negative affec-
2008; Kubarych et al., 2004). This is particularly important con- tivity, impulse-control problems, and submissiveness (Ackerman
 TEST OF TWO BRIEF MEASURES 13

Table 10
Correlations Between the PID5 Traits and the NPI-13 Subscales

LA GE EE
95% CI 95% CI 95% CI
Measure Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

Negative affectivity
Emotional lability .15 .04 .26 .17 .06 .28 .28 .17 .38
Anxiousness ⫺.03a ⫺.14 .08 ⫺.02a ⫺.13 .09 .21b .10 .31
Separation insecurity .16 .05 .27 .29 .18 .39 .29 .18 .39
Perseveration .10a ⫺.01 .21 .13ab .02 .24 .32b .22 .42
Submissiveness .11 .00 .22 .03 ⫺.08 .14 .12 .01 .23
Hostility .22a .11 .32 .23a .12 .33 .47b .38 .55
Restricted affect .15 .04 .26 .11 .00 .22 .19 .08 .30
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Depressivity .06a ⫺.05 .17 .07a ⫺.04 .18 .32b .22 .42
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Suspiciousness .17 .06 .28 .18 .07 .29 .33 .23 .43
Detachment
Withdrawal .09ab ⫺.02 .20 ⫺.07a ⫺.18 .04 .24b .13 .34
Anhedonia ⫺.02a ⫺.13 .09 ⫺.06a ⫺.17 .05 .28b .17 .38
Intimacy avoidance .07a ⫺.04 .18 .08ab ⫺.03 .19 .30b .19 .40
Antagonism
Manipulativeness .38 .28 .47 .33 .23 .43 .42 .32 .51
Deceitfulness .31 .20 .41 .28 .17 .38 .47 .38 .55
Grandiosity .40 .30 .49 .45 .36 .54 .41 .31 .50
Attention seeking .42 .32 .51 .48 .39 .56 .39 .29 .48
Callousness .26a .15 .36 .26a .15 .36 .50b .41 .58
Disinhibition
Irresponsibility .18a .07 .29 .26a .15 .36 .48b .39 .56
Impulsivity .17a .06 .28 .25ab .14 .35 .39b .29 .48
Rigid perfectionism .19 .08 .30 .19 .08 .30 .26 .15 .36
Distractibility .09a ⫺.02 .20 .08a ⫺.03 .19 .36b .26 .45
Risk taking .20 .09 .31 .21 .10 .31 .15 .04 .26
Psychoticism
Unusual beliefs/perceptions .28 .17 .38 .25 .14 .35 .38 .28 .47
Eccentricity .09ab ⫺.02 .20 .04a ⫺.07 .15 .30b .19 .40
Cognitive/perceptual dysregulation .25 .14 .35 .26 .15 .36 .41 .31 .50
Note. Ns ⫽ 305–306. k ⫽ 1. Correlations within each row with different subscripts are significantly different. PID5 ⫽ Personality Inventory for DSM–5;
NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence interval; LA ⫽ Leadership/Authority; GE ⫽ Grandiose Exhibitionism; EE ⫽ Entitlement/
Exploitativeness; DSM–5 ⫽ Diagnostic and Statistical Manual of Mental Disorders (5th ed.).

et al., 2011). It is important to note, however, that EE is not more Rosenthal & Hooley, 2010). For instance, some critics argue that
strongly associated with many other critical criteria such as inter- the NPI-40 is a measure of “normal” or “adaptive narcissism” but
view or self-reports of DSM–IV–TR NPD symptoms, DSM–5 NPD not pathological narcissism (Pincus & Lukowitsky, 2010), despite
traits, or interpersonal antagonism from the FFM. the fact that it correlates as highly with measures of NPD as do
In Study 2, we demonstrated that the corresponding subscales of other measures of narcissism such as the PNI (Pincus et al., 2009).
the NPI-13 and NPI-40 were highly related to one another and that Others argue that the NPI-40 includes content that is not central to
noncorresponding subscales were similarly intercorrelated be- narcissism such as extraversion (Ackerman, Donnellan, & Robins,
tween and within measures. Thus, the subscales of the NPI-13 2012), despite ratings by narcissism scholars, clinicians, and non-
appear to closely approximate those of the NPI-40, suggesting that experts that suggest otherwise (Lynam & Widiger, 2001; Samuel
they may be used effectively in place of those of the NPI-40 when & Widiger, 2004; Thomas et al., 2012).
time is limited. We have suggested previously that the problems with the
NPI-40 have been overstated (Miller & Campbell, 2011; Miller,
Criticism of the NPI-40 Maples, & Campbell, 2011; Miller, Price, & Campbell, 2012). For
In testing and discussing the validity of these two brief forms of instance, it is difficult to conceive of the NPI-40 as a measure of
the NPI-40, it is important to acknowledge that the NPI-40 has adaptive narcissism when it manifests substantial correlations with
come under substantial scrutiny and criticism over the past several interview and self-reported symptoms of NPD, as well as other
years. These criticisms have involved a variety of issues including personality disorders such as antisocial PD, histrionic PD, and
the NPI-40’s relative focus on the assessment of grandiose rather psychopathy; general personality traits such as immodesty, deceit-
than vulnerable narcissism, adaptivity versus maladaptivity, the fulness, noncompliance, and dominance; pathological personality
reliability and replicability of its factor structure, and its relations traits such as manipulativeness, grandiosity, attention seeking,
with self-esteem and psychological functioning (e.g., Ackerman et callousness, hostility, and irresponsibility; and externalizing be-
al., 2011; Brown et al., 2009; Cain, Pincus, & Ansell, 2008; haviors such as antisocial behavior, gambling, and aggression.
 14 GENTILE ET AL.

Table 11
Correlations Between Psychopathology, Psychopathy, and Behavioral Outcome Measures and the NPI-13 Subscales

LA GE EE
95% CI 95% CI 95% CI
Measure k Rxy Lower Upper Rxy Lower Upper Rxy Lower Upper

SCID-II-PQ
Paranoid 1 .21ab .09 .33 .02a ⫺.11 .15 .35b .23 .46
Schizoid 1 .04 ⫺.09 .17 .02 ⫺.11 .15 .07 ⫺.06 .20
Schizotypal 1 ⫺.02 ⫺.15 .11 .00 ⫺.13 .13 ⫺.01 ⫺.14 .12
Antisocial 1 .21 .09 .33 .16 .03 .28 .21 .09 .33
Borderline 1 .00 ⫺.13 .13 ⫺.09 ⫺.21 .04 .17 .04 .29
Histrionic 1 .36 .24 .47 .40 .29 .50 .21 .09 .33
Avoidant 1 ⫺.28ab ⫺.39 ⫺.16 ⫺.31a ⫺.42 ⫺.19 ⫺.05b ⫺.18 .08
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Dependent 1 ⫺.21 ⫺.33 ⫺.09 ⫺.17 ⫺.29 ⫺.04 ⫺.03 ⫺.16 .10
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Obsessive-compulsive 1 .08 ⫺.05 .21 ⫺.09 ⫺.21 .04 .04 ⫺.09 .17
SRP-III
SRP-IPM 1 .42 .33 .50 .31 .21 .40 .47 .39 .55
SRP-CA 1 .41a .32 .49 .13b .03 .23 .44a .35 .52
SRP-ELS 1 .27 .17 .36 .20 .10 .30 .25 .15 .34
SRP-ASB 1 .17a .07 .27 .09a ⫺.01 .19 .39b .30 .47
BSI
GSI 2 ⫺.07a ⫺.15 .01 ⫺.11a ⫺.19 ⫺.03 .16b .08 .23
PANAS-X
Negative affect 2 ⫺.02 ⫺.10 .06 ⫺.01 ⫺.09 .07 .14 .06 .22
Positive affect 2 .20a .12 .28 .20a .12 .28 ⫺.03b ⫺.11 .05
CAB
Alcohol use 1 .10 .00 .20 .17 .07 .27 .08 ⫺.02 .18
Substance use 1 .00 ⫺.10 .10 .09 ⫺.01 .19 .03 ⫺.07 .13
Antisocial behavior 1 .15 .05 .25 .15 .05 .25 .22 .12 .32
Gambling 1 .24 .14 .33 .13 .03 .23 .13 .03 .23
Intimate partner violence 1 .06 ⫺.04 .16 .13 .03 .23 .04 ⫺.06 .14
RCAP
Composite 1 .25 .04 .44 .28 .07 .46 .47 .29 .62
Note. SCID-II-PQ Ns ⫽ 237–238, SRP-III Ns ⫽ 361, BSI Ns ⫽ 597–598, PANAS-X Ns ⫽ 595–596, CAB Ns ⫽ 361, RCAP Ns ⫽ 86. Correlations within
each row with different subscripts are significantly different. NPI ⫽ Narcissistic Personality Inventory; CI ⫽ confidence interval; LA ⫽ Leadership/
Authority; GE ⫽ Grandiose Exhibitionism; EE ⫽ Entitlement/Exploitativeness; SCID-II-PQ ⫽ Structured Clinical Interview for DSM–IV Personality
Disorders—Personality Questionnaire; SRP-III ⫽ Self-Report Psychopathy Scale, Version III; IPM ⫽ Interpersonal Manipulation; CA ⫽ Callous Affect;
ELS ⫽ Erratic Lifestyle; ASB ⫽ Antisocial Behaviors; BSI ⫽ Brief Symptom Inventory; PANAS-X ⫽ Positive and Negative Affect Scale—Expanded
Form; CAB ⫽ Crime and Analogous Behavior Scale; RCAP ⫽ Response Choice Aggression Paradigm; DSM–IV ⫽ Diagnostic and Statistical Manual of
Mental Disorders (4th ed); GSI ⫽ General Severity Index.

Similarly, we disagree with those who would suggest that the scales are associated with both adaptive (e.g., low neuroticism, low
NPI-40 subscales of LA and/or GE are irrelevant to the study of distress; high extraversion, positive affect, and self-esteem) and
narcissism or only assess “adaptive aspects of personality and maladaptive (e.g., symptoms of NPD, interpersonal antagonism
therefore should be excluded from inventories designed to measure from the perspective of the FFM and the DSM–5 trait model,
maladaptive personality features” (Ackerman et al., 2011, p. 82). psychopathy, and externalizing behaviors) features and outcomes.
The current data suggest that both the NPI-13 LA and GE sub- In addition, these two subscales create personality profiles that are

Table 12
Correlations Between the NPI-13 and NPI-40 Subscales

NPI-13 NPI-40
Subscale LA GE EE LA GE EE

NPI-13 LA —
NPI-13 GE .48 —
NPI-13 EE .48 .37 —
NPI-40 LA .82 .51 .46 —
NPI-40 GE .60 .84 .42 .61 —
NPI-40 EE .53 .44 .86 .52 .47 —
Note. The bold correlations are the correlations between the same subscales across the NPI-13 and NPI-40. All correlations are significant at p ⬍ .001.
NPI ⫽ Narcissistic Personality Inventory; LA ⫽ Leadership/Authority; GE ⫽ Grandiose Exhibitionism; EE ⫽ Entitlement/Exploitativeness.
 TEST OF TWO BRIEF MEASURES 15

generally consistent with the meta-analytically derived profile towards internal consistency . . . is not necessarily the path towards
found for DSM–IV–TR NPD. predictive validity” (p. 1006).
Despite our disagreements with some of the aforementioned Despite these limitations, both the NPI-13 and NPI-16 appear to
critiques of the NPI-40, we took previous criticisms of the NPI-40 be promising brief measures of grandiose narcissism. The empir-
into account in creating the NPI-13 by choosing items that ical evidence presented here suggests that both measures are
Rosenthal and Hooley (2010)— critics of the NPI-40 —reported comparable to the NPI-40 in terms of convergent and discriminant
were most relevant to NPD as judged by expert ratings, IRT validity, while demonstrating adequate overall reliability. Consis-
analyses, or the results of exploratory factor analyses. The choice tent with suggestions by Smith and colleagues (2000) regarding
to include items that these researchers deemed most relevant and the development of brief forms, the two NPI brief forms save time
central to narcissism likely explains why the NPI-13 appears to without a resultant loss in the validity of the scale scores. Further-
assess a slightly more pathological and “darker” variant of narcis- more, in addition to its truncated length, the NPI-13 exhibits a
sism, as seen in the smaller positive correlations with self-esteem relatively clear factor structure that suggests it can be useful for a
and extraversion. This change is also likely due to the fact that the more fine-grained assessment of narcissism. In sum, we believe
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

EE items comprise 31% of the NPI-13 items versus 10% in the that both the NPI-13 and NPI-16 provide valid assessments of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

NPI-40; as such, the NPI-13 total score weights traits associated grandiose narcissism but that the NPI-13 might be more advanta-
with entitlement and exploitativeness to a much more substantial geous due the availability of the three subscales in addition to a
degree than is the case in the NPI-40. These changes, while total score. This feature allows researchers to continue to test the
relatively small in terms of differences in the external correlates of manner in which these narcissism traits converge and diverge from
these measures, may help mitigate concerns that have been voiced one another with regard to their relations with other central fea-
regarding the NPI-40. tures of narcissism’s nomological network.

Limitations and Conclusions References

Although the current study provides a great deal of data on the Ackerman, R. A., Donnellan, M. B., & Robins, R. W. (2012). An item
performance of these three versions of the NPI in relation to a wide response theory analysis of the Narcissistic Personality Inventory. Jour-
nal of Personality Assessment, 94, 141–155. doi:10.1080/00223891
array of relevant criterion variables and measures, and across a
.2011.645934
variety of sample populations (i.e., college students, adults via the Ackerman, R. A., Witt, E. A., Donnellan, M. B., Trzesniewski, K. H.,
Internet, and clinical patients), this study is not all-inclusive and Robins, R. W., & Kashy, D. A. (2011). What does the narcissistic
presents several limitations. First, additional research is needed to personality inventory really measure? Assessment, 18, 67– 87. doi:
test the reliability and validity of the two NPI brief forms in 10.1177/1073191110382845
clinical and community samples, particularly ones that are ethni- American Psychiatric Association. (2000). Diagnostic and statistical man-
cally diverse, as our own clinical sample was limited to White ual of mental disorders (4th ed., text rev). Washington, DC: Author.
participants, and in relation to additional criteria assessed using Ames, D., Rose, P., & Anderson, C. (2006). The NPI-16 as a short measure
diverse methodologies. Second, while we combined data where of narcissism. Journal of Research in Personality, 40, 440 – 450. doi:
10.1016/j.jrp.2005.03.002
possible, a number of analyses were conducted on only one sample
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psy-
and, having a smaller sample size, may thus be less generalizable. chological Bulletin, 107, 238 –246. doi:10.1037/0033-2909.107.2.238
Third, all of the data presented here were taken at one time point Bentler, P. M. (1995). EQS structural equations program manual. Encino,
and, thus, preclude analysis of the stability of the three NPI CA: Multivariate Software.
versions over time. Additional data are needed in the form of Brown, R. P., Budzek, K., & Tamborski, M. (2009). On the meaning and
longitudinal research to examine the test–retest reliability of the measure of narcissism. Personality and Social Psychology Bulletin, 35,
two short-forms in particular. Lastly, previous research has sug- 951–964. doi:10.1177/0146167209335461
gested age and gender differences on the NPI-40 (Foster, Camp- Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechan-
bell, & Twenge, 2003). Although not examined in this article, ical Turk: A new source of inexpensive, yet high-quality, data? Per-
spectives on Psychological Science, 6, 3–5. doi:10.1177/
future research should investigate the impact of these effects
1745691610393980
across versions of the NPI. Cain, N. M., Pincus, A. L., & Ansell, E. B. (2008). Narcissism at the
An important limitation in terms of scale usage is the lower crossroads: Phenotypic description of pathological narcissism across
internal consistency noted for the NPI-13 EE subscale, which is clinical theory, social/personality psychology, and psychiatric diagnosis.
typically the case with this scale (e.g., Ackerman et al., 2012, Clinical Psychology Review, 28, 638 – 656. doi:10.1016/j.cpr.2007.09
2011). If future researchers wish to study the role of these traits .006
using the NPI-13 (or NPI-40, since this subscale is the same across Campbell, W. K., Bonacci, A. M., Shelton, J., Exline, J. J., & Bushman,
both measures), they may consider bolstering it with an additional B. J. (2004). Psychological entitlement: Interpersonal consequences and
measure of entitlement such as the PES. Like Ackerman et al. validation of a self-report measure. Journal of Personality Assessment,
83, 29 – 45. doi:10.1207/s15327752jpa8301_04
(2011), we would note that the NPI-13/NPI-40 EE subscale still
Corry, N., Merritt, R. D., Mrug, S., & Pamp, B. (2008). The factor structure
manifests evidence of good criterion validity, which is consistent of the Narcissistic Personality Inventory. Journal of Personality Assess-
with the recent data that suggest that internal consistency may not ment, 90, 593– 600. doi:10.1080/00223890802388590
be as relevant to scale validity as was once thought (McCrae, Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory
Kurtz, Yamagata, & Terracciano, 2011). Addressing a similar (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional
point, Thalmayer, Saucier, and Eigenhuis (2011) noted, “the path manual. Lutz, FL: Psychological Assessment Resources.
 16 GENTILE ET AL.

Derogatis, L. R. (1975). The SCL-90-R. Baltimore, MD: Clinical Psycho- Miller, J. D., & Campbell, W. K. (2008). Comparing clinical and social-
metric Research. personality conceptualizations of narcissism. Journal of Personality, 76,
Derogatis, L. R., & Melisaratos, N. (1983). The Brief Symptom Inventory: 449 – 476. doi:10.1111/j.1467-6494.2008.00492.x
An introductory report. Psychological Medicine, 13, 595– 605. doi: Miller, J. D., & Campbell, W. K. (2011). Addressing criticisms of the
10.1017/S0033291700048017 Narcissistic Personality Inventory (NPI). In W. K. Campbell & J. D.
Dickinson, K. A., & Pincus, A. L. (2003). Interpersonal analysis of gran- Miller (Eds.), Handbook of narcissism and narcissistic personality dis-
diose and vulnerable narcissism. Journal of Personality Disorders, 17, order: Theoretical approaches, empirical findings, and treatments (pp.
188 –207. doi:10.1521/pedi.17.3.188.22146 146 –152). Hoboken, NJ: Wiley.
Emmons, R. A. (1984). Factor analysis and construct validity of the Miller, J. D., Dir, A., Gentile, B., Wilson, L., Pryor, L. R., & Campbell,
Narcissistic Personality Inventory. Journal of Personality Assessment, W. K. (2010). Searching for a vulnerable dark triad: Comparing factor 2
48, 291–300. doi:10.1207/s15327752jpa4803_11 psychopathy, vulnerable narcissism, and borderline personality disorder.
Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of Journal of Personality, 78, 1529 –1564. doi:10.1111/j.1467-6494.2010
Personality and Social Psychology, 52, 11–17. doi:10.1037/0022-3514 .00660.x
.52.1.11 Miller, J. D., Gaughan, E. T., Pryor, L. R., Kamen, C., & Campbell, W. K.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Few, L. R., Miller, J. D., & Lynam, D. R. (2013). An examination of the (2009). Is research using the narcissistic personality inventory relevant
This document is copyrighted by the American Psychological Association or one of its allied publishers.

factor structure of the Elemental Psychopathy Assessment. Personality for understanding narcissistic personality disorder? Journal of Research
Disorders: Theory, Research, and Treatment. Advance online publica- in Personality, 43, 482– 488. doi:10.1016/j.jrp.2009.02.001
tion. Miller, J. D., Gentile, B., & Campbell, W. K. (2012). A test of the construct
First, M. B., Gibbon, M., Spitzer, R. L., Williams, J. B. W., & Benjamin, validity of the Five-Factor Narcissism Inventory. Journal of Personality
L. S. (1997). Structured Clinical Interview for DSM–IV Axis II Person- Assessment. Advance online publication.
ality Disorders (SCID-II). Washington, DC: American Psychiatric Press. Miller, J. D., Gentile, B., Wilson, L., & Campbell, W. K. (2013). Grandiose
Fossati, A., Beauchaine, T. P., Grazioli, F., Carretta, I., Cortinovis, F., & and vulnerable narcissism and the DSM–5 pathological personality trait
Maffei, C. (2005). A latent structure analysis of Diagnostic and Statis- model. Journal of Personality Assessment, 95, 284 –290.
tical Manual of Mental Disorders, fourth edition, narcissistic personality Miller, J. D., Hoffman, B. J., Gaughan, E. T., Gentile, B., Maples, J., &
disorder criteria. Comprehensive Psychiatry, 46, 361–367. doi:10.1016/ Campbell, W. K. (2011). Grandiose and vulnerable narcissism: A no-
j.comppsych.2004.11.006 mological network analysis. Journal of Personality, 79, 1013–1042.
Foster, J. D., & Campbell, W. K. (2007). Are there such things as “nar- doi:10.1111/j.1467-6494.2010.00711.x
cissists” in social psychology? A taxometric analysis of the Narcissistic Miller, J. D., & Lynam, D. R. (2003). Psychopathy and the five-factor
Personality Inventory. Personality and Individual Differences, 43, model of personality: A replication and extension. Journal of Person-
1321–1332. doi:10.1016/j.paid.2007.04.003 ality Assessment, 81, 168 –178. doi:10.1207/S15327752JPA8102_08
Foster, J. D., Campbell, W. K., & Twenge, J. M. (2003). Individual Miller, J. D., Maples, J., & Campbell, W. (2011). Comparing the construct
differences in narcissism: Inflated self-views across the lifespan and validity of scales derived from the Narcissistic Personality Inventory: A
around the world. Journal of Research in Personality, 37, 469 – 486. reply to Rosenthal and Hooley (2010). Journal of Research in Person-
doi:10.1016/S0092-6566(03)00026-6 ality, 45, 401– 407. doi:10.1016/j.jrp.2010.12.004
Hendin, H. M., & Cheek, J. M. (1997). Assessing hypersensitive narcis- Miller, J. D., Price, J., & Campbell, W. K. (2012). Is the narcissistic
sism: A reexamination of Murray’s Narcissism Scale. Journal of Re- personality inventory still relevant? A test of independent grandiosity
search in Personality, 31, 588 –599. doi:10.1006/jrpe.1997.2204 and entitlement scales in the assessment of narcissism. Assessment, 19,
Hopwood, C. J., Thomas, K. M., Markon, K. E., Wright, A. G., & Krueger, 8 –13. doi:10.1177/1073191111429390
R. F. (2012). DSM–5 personality traits and DSM–IV personality disor- Miller, J. D., Price, J., Gentile, B., Lynam, D. R., & Campbell, W. K.
ders. Journal of Abnormal Psychology, 121, 424 – 432. doi:10.1037/ (2012). Grandiose and vulnerable narcissism from the perspective of the
a0026656 interpersonal circumplex. Personality and Individual Differences. doi:
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance 10.1016/j.paid.2012.04.026
structure analysis: Conventional criteria versus new alternatives. Struc- Neal, T. M., & Sellbom, M. (2012). Examining the factor structure of the
tural Equation Modeling, 6, 1–55. doi:10.1080/10705519909540118 Hare Self-Report Psychopathy Scale. Journal of Personality Assessment,
Krueger, R. F., Derringer, J., Markon, K. E., Watson, D., & Skodol, A. E. 94, 244 –253. doi:10.1080/00223891.2011.648294
(2012). Initial construction of a maladaptive personality trait model and Oltmanns, T. F., Friedman, J. N. W., Fiedler, E. R., & Turkheimer, E.
inventory for DSM–5. Psychological Medicine, 42, 1879 –1890. doi: (2004). Perceptions of people with personality disorders based on thin
10.1017/S0033291711002674 slices of behaviors. Journal of Research in Personality, 38, 216 –229.
Kubarych, T. S., Deary, I. J., & Austin, E. J. (2004). The Narcissistic doi:10.1016/S0092-6566(03)00066-7
Personality Inventory: Factor structure in a non-clinical sample. Person- Paulhus, D. L., Neumann, C., & Hare, R. D. (in press). Manual for the
ality and Individual Differences, 36, 857– 872. doi:10.1016/S0191- Self-Report Psychopathy Scale. Toronto, Ontario, Canada: Multi-Health
8869(03)00158-2 Systems.
Lance, C. E., & Vandenberg, R. J. (2002). Confirmatory factor analysis. In Pincus, A. L., Ansell, E. B., Pimentel, C. A., Cain, N. M., Wright, A., &
F. Drasgow & N. Schmitt (Eds.), Measuring and analyzing behavior in Levy, K. N. (2009). Initial construction and validation of the Patholog-
organizations (pp. 221–254). San Francisco, CA: Jossey-Bass. ical Narcissism Inventory. Psychological Assessment, 21, 365–379. doi:
Lynam, D. R., & Widiger, T. A. (2001). Using the five-factor model to 10.1037/a0016530
represent the DSM–IV personality disorders: An expert consensus ap- Pincus, A. L., & Lukowitsky, M. R. (2010). Pathological narcissism and
proach. Journal of Abnormal Psychology, 110, 401– 412. doi:10.1037/ narcissistic personality disorder. Annual Review of Clinical Psychology,
0021-843X.110.3.401 6, 421– 446. doi:10.1146/annurev.clinpsy.121208.131215
McCrae, R. R., Kurtz, J. E., Yamagata, S., & Terracciano, A. (2011). Raskin, R., & Terry, H. (1988). A principal-components analysis of the
Internal consistency, retest reliability, and their implications for person- Narcissistic Personality Inventory and further evidence of its construct
ality scale validity. Personality and Social Psychology Review, 15, validity. Journal of Personality and Social Psychology, 54, 890 –902.
28 –50. doi:10.1177/1088868310366253 doi:10.1037/0022-3514.54.5.890
 TEST OF TWO BRIEF MEASURES 17

Reidy, D. E., Zeichner, A., Foster, J. D., & Martinez, M. A. (2008). Effects Steiger, J. H. (1990). Structural model evaluation and modification: An
of narcissistic entitlement and exploitativeness on human physical ag- interval estimation approach. Multivariate Behavioral Research, 25,
gression. Personality and Individual Differences, 44, 865– 875. doi: 173–180. doi:10.1207/s15327906mbr2502_4
10.1016/j.paid.2007.10.015 Thalmayer, A. G., Saucier, G., & Eigenhuis, A. (2011). Comparative
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, validity of brief to medium-length Big Five and Big Six personality
NJ: Princeton University Press. questionnaires. Psychological Assessment, 23, 995–1009. doi:10.1037/
Rosenthal, S. A., & Hooley, J. M. (2010). Narcissism assessment in a0024165
social–personality research: Does the association between narcissism Thomas, K. M., Wright, A. C., Lukowitsky, M. R., Donnellan, M. B., &
and psychological health result from a confound with self-esteem? Hopwood, C. J. (2012). Evidence for the criterion validity and clinical
Journal of Research in Personality, 44, 453– 465. doi:10.1016/j.jrp.2010 utility of the Pathological Narcissism Inventory. Assessment, 19, 135–
.05.008 145. doi:10.1177/1073191112436664
Rosenthal, S. A., Hooley, J. M., & Steshenko, Y. (2007). Distinguishing Velicer, W. F. (1976). Determining the number of components from the
grandiosity from self-esteem: Development of the Narcissistic Grandi- matrix of partial correlations. Psychometrika, 41, 321–327. doi:10.1007/
osity Scale. Unpublished manuscript. BF02293557
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Russ, E., Shedler, J., Bradley, R., & Westen, D. (2008). Refining the Watson, D., & Clark, L. A. (1994). The PANAS-X: Manual for the Positive
construct of narcissistic personality disorder: Diagnostic criteria and and Negative Affect Schedule—Expanded Form. Unpublished manu-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

subtypes. The American Journal of Psychiatry, 165, 1473–1481. doi: script, University of Iowa, University Heights.
10.1176/appi.ajp.2008.07030376 Wiggins, J. S. (1995). Interpersonal Adjective Scales: Professional man-
Samuel, D. B., & Widiger, T. A. (2004). Clinicians’ personality descrip- ual. Odessa, FL: Psychological Assessment Resources.
tions of prototypic personality disorders. Journal of Personality Disor- Wink, P. (1991). Two faces of narcissism. Journal of Personality and
ders, 18, 286 –308. doi:10.1521/pedi.18.3.286.35446 Social Psychology, 61, 590 –597. doi:10.1037/0022-3514.61.4.590
Samuel, D. B., & Widiger, T. A. (2008). A meta-analytic review of the Wright, A. G., Thomas, K. M., Hopwood, C. J., Markon, K. E., Pincus,
relationships between the five-factor model and DSM–IV–TR personality A. L., & Krueger, R. F. (2012). The hierarchical structure of DSM–5
disorders: A facet level analysis. Clinical Psychology Review, 28, 1326 – pathological personality traits. Journal of Abnormal Psychology, 121,
1342. doi:10.1016/j.cpr.2008.07.002 951–957. doi:10.1037/a0027669
Seibert, L. A., Miller, J. D., Few, L. R., Zeichner, A., & Lynam, D. R. Zeichner, A., Frey, F. C., Parrott, D. J., & Butryn, M. F. (1999). Measure-
(2011). An examination of the structure of self-report psychopathy ment of laboratory aggression: A new response-choice paradigm. Psy-
measures and their relations with general traits and externalizing behav- chological Reports, 85, 1229 –1237. doi:10.2466/pr0.1999.85.3f.1229
iors. Personality Disorders: Theory, Research, and Treatment, 2, 193–
208. doi:10.1037/a0019232
Smith, G. T., McCarthy, D. M., & Anderson, K. G. (2000). On the sins of Received September 29, 2012
short-form development. Psychological Assessment, 12, 102–111. doi: Revision received March 4, 2013
10.1037/1040-3590.12.1.102 Accepted April 12, 2013 䡲

View publication stats