Npi 16
Npi 16
Npi 16
www.elsevier.com/locate/jrp
Brief report
Abstract
Narcissism has received increased attention in the past few decades as a sub-clinical individ-
ual diVerence with important everyday consequences, such as self-enhancement in perceptions
of one’s own behavior and attributes. The most widespread measure used by non-clinical
researchers, the 40-item Narcissistic Personality Inventory or NPI-40, captures a range of
diVerent facets of the construct but its length may prohibit its use in settings where time pres-
sure and respondent fatigue are major concerns. In this article, we draw from the NPI-40 set of
items to create and validate a shorter, unidimensional measure, the NPI-16. In Wve studies, we
show that this short NPI closely parallels the NPI-40 in its relation to other personality mea-
sures and dependent variables. We conclude that the NPI-16 has notable face, internal, dis-
criminant, and predictive validity and that it can serve as an alternative measure of narcissism
when situations do not allow the use of longer inventories.
© 2005 Elsevier Inc. All rights reserved.
Keywords: Narcissism; Narcissist; NPI; Scale; Measure; Inventory; Validation; Hubris; Self-enhance-
ment; Individual diVerence
1. Introduction
*
Corresponding author.
E-mail address: [email protected] (D.R. Ames).
0092-6566/$ - see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jrp.2005.03.002
D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450 441
as well as a preoccupation with success and demands for admiration (see Morf &
Rhodewalt, 2001 for a recent review). In the past few decades, sub-clinical variance in
narcissism has received growing attention as a personality dimension, showing an
impressive ability to predict a wide range of dependent variables, ranging from emo-
tional reactivity to self-appraisals of performance.
Much current research on narcissism relies on either a 40-item Narcissistic Person-
ality Inventory (the NPI-40, Raskin & Terry, 1988) or a 37-item measure (Emmons,
1987). In this paper, we develop a shorter measure, the NPI-16, a new tool that could
expand scholarship on narcissism. Our hope is that the NPI-16 will facilitate research
where a longer measure would be impractical, as in certain Weld work and/or with
time- and attention-strapped respondents. We do not wish to obscure the subcompo-
nents of narcissism or to short-change its complexity; rather, our intent is to expand
researchers’ opportunities to further explore narcissism’s consequences.
Several studies have explored the factor structure of narcissism (e.g., Emmons,
1984, 1987; Raskin & Terry, 1988). However, nearly all recent empirical results
examining the eVects of narcissism have revolved around total NPI scores. Indeed,
in their work on narcissism and self-enhancement, Wallace and Baumeister (2002)
noted that total NPI scores were consistently more predictive than any of the sub-
scale scores. Thus, our approach was to create a short measure of narcissism as a
single construct, albeit one that represents the diVerent aspects of narcissism as
reXected in the original NPI.
We developed and tested a short measure of narcissism (the NPI-16) in Wve stud-
ies. In Study 1, we selected items from the larger NPI-40 and compared the short and
long measures as well as their relationships with the Big Five personality traits. In
Study 2, we considered convergent/discriminant validity and in Study 3, we examined
test–retest data. In Studies 4 and 5, we explored predictive validity.
2. Study 1
2.1. Method
2.2. Results
The NPI-16 score was computed as the mean across the 16 items, with narcis-
sism-consistent responses coded as 1 and narcissism-inconsistent responses coded
as 0. The NPI-16 had an ! of .72 while the full 40-item measure revealed an ! of
.841. The mean interitem correlations were .12 for the 40-item scale and .13 for the
16-item scale. For the 40-item scale, loadings on the Wrst unrotated factor ranged
from .06 to .59 with the Wrst factor capturing 15.0 percent of variance. For the NPI-
16, loadings on the Wrst unrotated factor ranged from .13 to .66 with the Wrst factor
capturing 19.9 percent of variance. The two measures were correlated at r D .90
(p < .001). The 16-item scale correlated with the remaining 24 items from the NPI-
40 at r D .71 (p < .001). The mean was .35 (.20) for the NPI-16 and .39 (.17) for the
NPI-40 (see Table 1).
1
To conWrm our expectations about scale length, we tested a number of alternative measures. Reliability
dropped substantially with shorter measures (e.g., an 8-item subset yielded an ! of .59 and showed more di-
vergent correlations with other personality measures). Longer measures showed improved reliability but
few gains in correlations.
D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450 443
Table 1
Mean values and gender diVerences in narcissism
16-item scale 40-item scale
M (SD) t (df) p M (SD) t (df) p
Study 1
All .35 (.20) .39 (.17)
Men .37 (.20) 1.83 (772) .07 .40 (.18) 1.61 (772) .10
Women .34 (.20) .38 (.17)
Study 2
All .40 (.19)
Men .41 (.19) 1.66 (165) .10 — — —
Women .36 (.17) —
Study 3
All .31 (.19)
Men .39 (.24) 3.00 (141) <.01 — — —
Women .29 (.16) —
Study 4
All .37 (.19) .40 (.17)
Men .41 (.17) 2.22 (173) .03 .44 (.16) 2.66 (173) <.01
Women .35 (.20) .38 (.17)
Study 5
All .39 (.18)
Men .43 (.21) 1.75 (40) .09 — — —
Women .33 (.11) —
Note. t and p values indicate tests of mean diVerences for men versus women.
As seen in Table 2, the NPI-16 and NPI-40 scales had extremely similar patterns
of correlations with the Big Five personality constructs. Both the 16- and 40-item
scales correlated positively with Openness to Experience, Conscientiousness, and
Extraversion and both correlated negatively with Agreeableness and Neuroticism.
None of the correlations between the two narcissism measures and the Big Five
dimensions diVered by more than .10. Consistent with prior work on self-esteem
and narcissism (Campbell, Rudich, & Sedikides, 2002), the scales showed similar
moderate and signiWcant correlations with self-esteem. As expected, neither the 16-
nor the 40-item measures showed a signiWcant correlation with the Just World
Scale.
This same pattern of parallel correlations emerged when the NPI-16 was used for
a randomly selected half of the dataset and the NPI-40 was used for the other half
(NPI-16 and -40 correlations were within .10 of one another and consistent in direc-
tion and signiWcance). Finally, as found in previous work (e.g., Foster, Campbell, &
Twenge, 2003; Watson, Grisham, Trotter, & Biderman, 1984), men had slightly
higher mean levels of narcissism across both measures (see Table 1).
In sum, the NPI-16 showed satisfactory internal consistency, correlated strongly
with the 40-item NPI, exhibited correlations to Big Five constructs and self-esteem
that were very similar to those of the longer measure, and showed the same sex diVer-
ences as the longer measure.
444 D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450
Table 2
Correlations with narcissism measures
16-item scale 40-item scale
Study 1
40-Item NPI .90¤¤ —
Remaining NPI items (24 non-included items) .71¤¤ —
Openness to experience (NEO measure) .24¤¤ .24¤¤
Conscientiousness (NEO measure) .14¤¤ .09¤
Extraversion (NEO measure) .36¤¤ .26¤¤
Agreeableness (NEO measure) ¡.23¤¤ ¡.25¤¤
Neuroticism (NEO measure) ¡.21¤¤ ¡.13¤¤
Self-esteem (Rosenberg measure) .30¤¤ .38¤¤
Belief in a just world .04 .06
Study 2
Openness to experience (BFI measure) .32¤¤ —
Extraversion (BFI measure) .41¤¤ —
Self-esteem (Robins et al. measure) .24¤¤ —
Self-monitoring .37¤¤ —
Dispositionism .06 —
Study 3
16-item scale retest (5 weeks later) .85¤¤ —
Study 4
40-item NPI .90¤¤ —
Self overall performance ranking ¡.25¤¤ ¡.24¤¤
Self eVort ranking ¡.17¤¤ ¡.20¤¤
Self creativity ranking ¡.28¤¤ ¡.21¤¤
Self power rating .29¤¤ .23¤¤
Self attractiveness rating .37¤¤ .31¤¤
Self cooperativeness ranking .05 ¡.04
Study 5
Actual performance .15
Self-esteem .34¤
Estimated relative accuracy .38¤ —
Estimated relative accuracy after controlling .36¤ —
for self-esteem and actual performance
Note. Study 1, n D 776; Study 2, n D 167; Study 3, n D 117; Study 4, n D 176; Study 5, n D 43.
¤
p < .05.
¤¤
p < .01.
3. Study 2
Study 2 sought additional convergent validity for the NPI-16 in a diVerent adult
population: master’s of business administration students (MBAs). We expected nar-
cissism would correlate positively with self-esteem, extraversion, openness to experi-
ence, and self-monitoring. For discriminant validity, we measured dispositionism,
expecting it to be unrelated to narcissism.
D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450 445
3.1. Method
3.1.1. Participants
One hundred sixty-seven full-time MBA students (44 women, 123 men) partici-
pated in Study 2 as part of a class exercise. Mean age was 28.31 (SD D 3.53).
3.1.2. Materials
Participants completed the NPI-16 as well as measures of self-esteem, extraver-
sion, openness, self-monitoring, and dispositionism. Self-esteem was measured using
the Robins, Hendin, and Trzesniewski (2001) single-item self-esteem scale. Partici-
pants indicated agreement with the statement “I have high self-esteem” on a scale
ranging from “Disagree strongly” (1) to “Agree strongly” (5). Extraversion and
Openness to Experience were measured with the related items from the Big Five
Inventory (John, Donahue, & Kentle, 1991) on a scale ranging from “Disagree
strongly” (1) to “Agree strongly” (5). Self-monitoring was measured with the Lennox
and Wolfe (1984) measure, with ratings on a six-point scale ranging from “Certainly
always false” (1) to “Certainly always true” (6). Dispositionism was measured using a
three-item scale (Chiu, Hong, & Dweck, 1997), including “The kind of person some-
one is something very basic about them and it can’t be changed very much.” Items
were rated on a Wve-point scale ranging from “Disagree strongly” (1) to “Agree
strongly” (5).
The NPI-16 had an ! of .68 and a mean of .40 (SD D .19). As expected, and as
shown in Table 2, narcissism was positively correlated with openness, extraversion,
self-esteem, and self-monitoring. Consistent with our expectations of divergent valid-
ity, narcissism was not signiWcantly correlated with dispositionism. As in Study 1,
men scored higher in narcissism than women (see Table 1). Study 2 thus provided
additional evidence of convergent validity for the NPI-16. The correlations we
observed with our 16-item narcissism measure were similar to those observed by
Emmons (1984) in his examination of a longer version of the Narcissistic Personality
Inventory.
4. Study 3
In Study 3, we examined the test–retest reliability for the NPI-16 over a 5-week
interval. We expected a high correlation between narcissism scores over this period.
4.1. Method
received extra course credit for their participation. The NPI-16 was included at the
end of each questionnaire packet.
The NPI-16 had an ! of .69 at Time 1 and an ! of .78 at Time 2. The Time 1 mean
for the NPI-16 was .31 (.19) while the Time 2 mean was .32 (.22). As shown in Table 2,
scores on the NPI-16 were stable over a 5-week period (r D .85, p < .01). As in Studies
1 and 2, men scored signiWcantly higher in narcissism than women (see Table 1).
5. Study 4
In Study 4, we sought evidence of the predictive validity of the NPI-16. Prior work
suggests that narcissists tend to give high estimates of their contributions to group tasks
(e.g., John & Robins, 1994) and their attractiveness (e.g., Gabriel, Critelli, & Ee, 1994).
Participants’ completed the NPI-40 and then participated in a group decision-making
task. After the task, they ranked their own contributions to the discussion and their own
attractiveness. We expected the NPI-16 to parallel the NPI-40 in predicting judgments of
one’s own performance, eVort, and creativity. Likewise, we expected both measures to
similarly predict participants’ ratings of their own power and attractiveness. We did not
expect either narcissism measure to predict rankings of cooperativeness (which we
included to test discriminant validity; cf. Campbell et al., 2002).
5.1. Method
5.1.1. Participants
One hundred and seventy-six undergraduate university students participated in
Study 4 as part of a research credit program for introductory-level courses in psy-
chology (105 women, 71 men). Mean age was 20.34 (SD D 3.19).
The NPI-16 had an ! of .69 while the NPI-40 had an ! of .83. The NPI-16 corre-
lated with the 40-item scale at r D .90 (p < .001).
Our expectations were conWrmed: narcissism was related to higher rankings for
overall performance, eVort, and creativity (Table 2). As predicted, narcissism was
also related to the tendency to rate oneself as more powerful and inXuential in the
discussion and as more attractive. As expected, narcissism was not signiWcantly
related to self-rankings of cooperativeness. As before, men showed higher levels of
narcissism (Table 1).
6. Study 5
6.1. Method
The NPI-16 had an ! of .65 and a mean of .39 (SD D .18). As shown in Table 2,
narcissism was not signiWcantly linked with actual performance, but was signiWcantly
correlated with participants’ estimates of their relative accuracy. Narcissism was also
positively correlated with self-esteem, though self-esteem was not signiWcantly corre-
lated with either actual or estimated relative accuracy (rs D ¡.05 and .06, ns, respec-
tively). In a multiple regression predicting self-estimates of accuracy with narcissism,
self-esteem, and actual percentile, only narcissism emerged as signiWcantly predictive.
The partial correlation between narcissism and estimated accuracy controlling for
self-esteem and actual accuracy was r D .36, p D .03. Consistent with the results of the
prior studies, men showed marginally higher levels of narcissism (Table 1).
448 D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450
In sum, as expected, Study 5 showed additional evidence for the predictive validity
of the NPI-16. Narcissism was signiWcantly and positively correlated with partici-
pants’ estimates of their own relative performance in a judgment task.
Further evidence for the validity of the NPI-16 comes from recent work by Ames
and Kammrath (2004) on self-estimated performance in interpersonal judgments. In
two studies, participants estimated their relative performance on social judgment
tasks (e.g., reading a partner’s emotions). The pattern of results parallel those
reported here in Study 5: estimated performance was largely unrelated to actual per-
formance, but estimated performance was strongly predicted by narcissism (mea-
sured using the NPI-16). The eVect of narcissism on self-estimates persisted after
controlling for self-esteem and actual performance.
After our validation work on the NPI-16, we learned of unpublished and in-press
work on a 15-item measure of narcissism (the NPI-15; Armor, 2002; Schütz, Marcus,
& Sellin, 2004; NPI-40 items 1, 4, 7, 9, 10, 12, 13, 18, 27, 30, 32, 33, 34, 36, and 40; nine
items in common with the NPI-16). These items were selected on the basis of loadings
on the Wrst unrotated principal component in Raskin and Terry’s analysis (1988).
This strategy produced a highly reliable scale, though one that is more concentrated
on certain narcissism components than the NPI-16. Seven of the 15 items fall on
Emmons’s (1987) leadership/authority factor. With Raskin and Terry’s (1988)
dimensions, six of the 16 fall on the authority factor. In contrast, our NPI-16 items
are, by design, more evenly distributed across components (e.g., four items from each
of the Emmons factors).
Having used the NPI-40 in Studies 1 and 4, we compared the NPI-16 with the
NPI-15. The NPI-15 revealed substantial alphas (.81 in Study 1; .78 in Study 4) and
was signiWcantly correlated with both the NPI-16 (.86 in Study 1; .84 in Study 4) and
the NPI-40 (.91 in both studies). The NPI-15, -16, and -40 all showed virtually the
same pattern of correlations with our other measures in both Studies 1 and 4. In only
one case was there is a diVerence in signiWcance: while the NPI-16 and -40 were sig-
niWcantly negatively related to self-rankings of creativity in Study 4, the NPI-15 cor-
relation was not signiWcant (r D ¡.11, p D .17). Our conclusion is that both the NPI-15
and the NPI-16 are meaningful short measures of narcissism with good predictive
validity. However, while the NPI-15 yields greater internal consistency by focusing
more on the authority dimension of narcissism, the NPI-16 spans more of the com-
ponents of narcissism identiWed in past research.
8. Conclusion
Acknowledgments
The authors thank Sam Gosling and Bill Swann for generous comments and sug-
gestions.
Appendix A
References
Ames, D. R., & Kammrath, L. K. (2004). Mind-reading and metacognition: Narcissism, not actual compe-
tence, predicts self-estimated ability. Journal of Nonverbal Behavior, 28, 187–209.
Armor, D., (2002). A 15-item short version of the NPI. Unpublished manuscript.
450 D.R. Ames et al. / Journal of Research in Personality 40 (2006) 440–450
Campbell, W. K., Rudich, E., & Sedikides, C. (2002). Narcissism, self-esteem, and the positivity of self-
views: Two portraits of self-love. Personality & Social Psychology Bulletin, 28, 358–368.
Chiu, C., Hong, Y., & Dweck, C. S. (1997). Lay dispositionism and implicit theories of personality. Journal
of Personality and Social Psychology, 73, 19–30.
Costa, P. T, Jr., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO Per-
sonality Inventory. Psychological Assessment, 4, 5–13.
Emmons, R. A. (1984). Factor analysis and construct validity of the Narcissistic Personality Inventory.
Journal of Personality Assessment, 48, 291–300.
Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of Personality and Social Psychol-
ogy, 52, 11–17.
Foster, J. D., Campbell, W. K., & Twenge, J. M. (2003). Individual diVerences in narcissism: InXated self-
views across the lifespan and around the world. Journal of Research in Personality, 37, 469–486.
Gabriel, M. T., Critelli, J. W., & Ee, J. S. (1994). Narcissistic illusions in self-evaluations of intelligence and
attractiveness. Journal of Personality, 62, 143–155.
John, O. P., & Robins, R. W. (1994). Accuracy and bias in self-perception: Individual diVerences in self-
enhancement and the role of narcissism. Journal of Personality and Social Psychology, 66, 206–219.
John, O. P., Donahue, E. M., & Kentle, R. (1991). The Big-Five inventory. Technical report, institute for per-
sonality and social research. Berkeley, CA: University of California.
Lennox, R. D., & Wolfe, R. N. (1984). Revision of the self-monitoring scale. Journal of Personality and
Social Psychology, 46, 1349–1364.
Morf, C. C., & Rhodewalt, F. (2001). Unraveling the paradoxes of narcissism: A dynamic self-regulatory
processing model. Psychological Inquiry, 12, 177–196.
Raskin, R., & Terry, H. (1988). A principal-components analysis of the Narcissistic Personality Inventory
and further evidence of its construct validity. Journal of Personality and Social Psychology, 54,
890–902.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct vali-
dation of a single-item measure and the Rosenberg self-esteem scale. Personality and Social Psychology
Bulletin, 27, 151–161.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, New Jersey: Princeton University
Press.
Rubin, Z., & Paplau, L. A. (1973). Belief in a just world and reactions to another’s lot: A study of partici-
pants in a national draft lottery. Journal of Social Issues, 29, 73–93.
Schütz, A., Marcus, B., & Sellin, I. (2004). Die Messung von Narzissmus als Persönlichkeitskon-strukt:
Psychometrische Eigenschaften einer Lang- und einer Kurzform des Deutschen NPI (Narcissistic Per-
sonality Inventory). Diagnostica, 202–218.
Wallace, H. M., & Baumeister, R. F. (2002). The performance of narcissists rises and falls with perceived
opportunity for glory. Journal of Personality and Social Psychology, 82, 819–834.
Watson, P. J., Grisham, S. O., Trotter, M. V., & Biderman, M. D. (1984). Narcissism and empathy: Validity
evidence for the Narcissistic Personality Inventory. Journal of Personality Assessment, 48, 301–305.