Professional Documents
Culture Documents
Misperceiving Grandiose Narcissism As Self Esteem
Misperceiving Grandiose Narcissism As Self Esteem
Misperceiving Grandiose Narcissism As Self Esteem
DOI: 10.1111/jopy.12436
ORIGINAL ARTICLE
1
Department of Psychology, MacEwan
University, Edmonton, Alberta, Canada
Abstract
2
Department of Psychology, Wilfrid Laurier Objective: We examine why people form positive first impressions of grandiose
University, Waterloo, Ontario, Canada narcissists, even though they can identify others’ narcissism. We test whether this
occurs because narcissists are perceived to have especially high self‐esteem, which
Correspondence
Miranda Giacomin, Department of is socially valued.
Psychology, MacEwan University, 10700 Method: Across four studies, undergraduate perceivers viewed photographs of tar-
104 Ave, Edmonton Alberta T5J 4S2,
gets (for whom narcissism and self‐esteem were known) and rated perceptions of
Canada.
Email: miranda.giacomin@macewan.ca their narcissism and self‐esteem, as well as how much they liked them.
Funding information Results: Perceivers rated more narcissistic targets to be higher in self‐esteem (even
Social Sciences and Humanities Research compared to targets with equally high self‐esteem) and liked them more. Perceptions
Council of Canada
of self‐esteem, moreover, mediated the effect of target narcissism on liking (Study
1). This effect disappeared when targets’ narcissism was made salient, suggesting
that trait narcissism is not inherently attractive (Study 2). Finally, path models re-
vealed a negative effect of perceptions of narcissism on liking that was suppressed by
a positive effect of perceptions of self‐esteem on liking (Study 3a), even for ratings
of people’s online dating profiles (Study 3b).
Conclusions: Positive initial impressions of narcissists may be driven by inflated
perceptions that they have high self‐esteem.
KEYWORDS
impression formation, narcissism, person perception, self‐esteem
1 | IN T RO D U C T ION but over time like them less, as they become more hostile in
groups, neglectful in relationships, and exploitative as lead-
People can accurately perceive grandiose narcissism in oth- ers (Judge, LePine, & Rich, 2006; Paulhus, 1998; Wurst et
ers based on minimal information (e.g., Giacomin & Rule, al., 2017). We investigate why people initially seem to over-
2018; Holtzman, 2011; Holtzman & Strube, 2013; Naumann, look others’ narcissism and form positive, yet misguided,
Vazire, Rentfrow, & Gosling, 2009; Vazire, Naumann, first impressions of narcissists. Across four studies, we ex-
Rentfrow, & Gosling, 2008). They use cues associated with amine whether grandiose narcissists make positive first im-
grandiose narcissists’ physical appearance—such as stylish, pressions, in part, because perceivers focus predominantly
expensive clothing and neater, more organized appearance on perceptions that narcissists have high self‐esteem rather
(Back, Schmukle, & Egloff, 2010; Vazire et al., 2008)—to than their evident narcissism. Our studies, moreover, are rel-
accurately perceive narcissism in others, yet still form pos- atively novel in examining how targets’ personality along two
itive first impressions of grandiose narcissists.1 People ini- conceptually distinct dimensions (i.e., narcissism and self‐es-
tially like narcissists more than less narcissistic individuals, teem) jointly determine the first impressions they make.
self‐esteem. We also expect perceivers to like narcissists forced‐choice items, with one option more narcissistic than
more, and for this effect to be mediated by perceptions of the other. Narcissistic choices were summed to create a nar-
self‐esteem rather than narcissism. Such results would sug- cissism score.
gest perceivers rely on perceptions of self‐esteem when
forming impressions of narcissists and neglect perceptions Self‐esteem
of narcissism. Targets completed the 10‐item Rosenberg Self‐Esteem
We also explore the possibility that people like narcis- Scale (RSES; Rosenberg, 1965; α = 0.78), indicating how
sists more, despite recognizing their narcissism, because much they agree with each item (1 = strongly disagree,
they misunderstand the nature of narcissism. Narcissists 9 = strongly agree). Items were averaged, with higher scores
are more egoistic and less communal than less narcissis- indicating higher self‐esteem.2
tic individuals (Campbell, Rudich, & Sedikides, 2002).
Accordingly, we had perceivers rate targets on egoistic and Interview
communal qualities. Perceivers may like narcissists more, Targets completed a videotaped interview, answering a series
despite recognizing their narcissism, if they fail to recog- of questions, which were transcribed verbatim (for details,
nize that they are egoistic and less communal. Alternatively, see the supplemental materials or Giacomin & Jordan, 2016).
however, they may accurately perceive that narcissists are The interview began with simple get‐to‐know‐you questions
egoistic and less communal, but nevertheless like them (e.g., “What is your academic major?”) to make participants
anyway. This latter pattern of results would further suggest comfortable and then got more personal (e.g., “Describe your
that perceivers rely on perceptions of self‐esteem and ne- best attribute”).
glect perceptions of narcissism when forming first impres-
sions of others.
2.1.2 | Target selection
We selected five high self‐esteem/high narcissism targets
2.1 | Method
(three females; Mage = 21.60, SD = 1.95), five high self‐es-
teem/low narcissism targets (five females; Mage = 20.80,
2.1.1 | Targets
SD = 0.84), and five low self‐esteem/low narcissism tar-
Data from 47 undergraduates were collected to serve as gets (three females; Mage = 23.80, SD = 6.10), based on
target stimuli for subsequent studies. They participated tertile splits of the initial sample (see Table 1). We al-
in a study described to them as investigating a variety of lowed targets to vary naturalistically along all dimensions
attitudes in exchange for partial course credit or $11.00 other than self‐esteem and narcissism, though this created
(CDN). Two photographs (one smiling, one with neutral a gender imbalance across target types. To alleviate con-
expression) were taken of each target above the waist cerns that our results are due to target gender, we replicated
against a gray background. Thirty‐three undergraduates the analyses excluding male targets, and results remained
(14 males; Mage = 22.21, SD = 3.91; 72% Caucasian, 16% identical. Target gender also did not moderate any of the
Middle Eastern, 9% Asian, 3% African American) con- results.
sented to have their photographs shown to future partici-
pants. There were no differences between targets who did
and did not consent on narcissism, self‐esteem, or demo-
2.1.3 | Perceivers
graphic information. We needed 116 perceivers to achieve 80% power to de-
tect a medium effect size in a mixed‐model ANOVA
Narcissism (Faul, Erdfelder, Lang, & Buchner, 2007). Undergraduates
Targets completed the Narcissistic Personality Inventory (N = 128; 92 females; Mage = 18.84, SD = 5.70) participated
(NPI; Raskin & Terry, 1988; α = 0.78), consisting of 40 for partial course credit.
TABLE 1 Study 1: Targets’ self‐reported self‐esteem and narcissism as a function of target type
Target type
TABLE 3 Study 1: Ratings of the primary dependent variables as a function of target type and information condition
Target Type
3.2 | Results
3 | ST U DY 2
Table 4 presents descriptive statistics and correlations for
In Study 2, we further test whether people like narcissists study variables at the perceiver level. Each dependent vari-
because they focus predominantly on perceptions of high able was submitted to a 3 (target type: high self‐esteem/high
self‐esteem and neglect perceptions of narcissism. If this is narcissism, high self‐esteem/low narcissism, low self‐es-
so, then making targets’ narcissism salient should attenuate teem/low narcissism) × 2 (self‐esteem information: speci-
the extent to which narcissists are well liked. Accordingly, fied, unspecified) × 2 (narcissism information: specified,
we manipulated whether perceivers were explicitly told unspecified) mixed‐model ANOVA (Table 5). We decom-
about targets’ self‐esteem and narcissism (see also Zeigler‐ pose significant interactions by comparing first across target
Hill & Myers, 2009). Perceivers viewed the same targets as type and then within target type across information condi-
in Study 1. We expected to replicate Study 1 when partici- tions. We again use a modified Bonferroni correction (Holm,
pants were not told about targets’ self‐esteem and narcis- 1979).
sism. In addition, we expected participants to continue to
like targets explicitly labeled as having high self‐esteem.
However, we expected perceivers to like narcissistic targets
3.2.1 | Perceptions of self‐esteem
less when made explicitly aware of targets’ narcissism be- For self‐esteem ratings, there were main effects of narcissism
cause in this case perceivers cannot easily disregard targets’ information, F(1, 187) = 7.99, p = 0.005, 𝜂p2 = 0.04, and tar-
narcissism. get type, F(2, 374) = 335.10, p < 0.001, 𝜂p2 = 0.64, and a
GIACOMIN and JORDAN
| 833
TABLE 4 Study 2: Descriptive statistics and correlations Across all information conditions, perceivers rated high
between primary dependent variables self‐esteem/high narcissism targets higher in narcissism than
Perceptions
high self‐esteem/low narcissism targets and low self‐esteem/
low narcissism targets (all Fs ≥ 17.50, ps < 0.001), replicat-
1 2 3 4 5 ing Study 1. Perceivers also rated high self‐esteem/low nar-
1. Self‐esteem 0.14†
0.56**
0. 41 **
0.03 cissism targets higher in narcissism than low self‐esteem/low
2. Narcissism –0.12 –0.29 **
0.69** narcissism targets when they were told about targets’ self‐es-
3. Liking 0.56 **
–0.16* teem but not narcissism, F(1, 47) = 35.05, p < 0.001, 𝜂p2 =
4. Communal –0.31**
0.43, or both self‐esteem and narcissism, F(1, 47) = 5.43,
traits p = 0.024, 𝜂p2 = 0.10, but not when they were given no infor-
5. Egoistic mation, F(1, 47) = 3.35, p = 0.073, 𝜂p2 = 0.07, or told only
traits about targets’ narcissism, F(1, 46) = 1.28, p = 0.263, 𝜂p2 =
M 5.72 3.77 5.32 6.31 3.52 0.03. Thus, participants tended to track the information given
to them about targets’ narcissism.
SD 0.81 1.15 1.07 0.79 1.24
† * ** In addition, perceivers rated high self‐esteem/high narcis-
p < 0.06. p < 0.05. p < 0.01.
sism targets higher in narcissism when told about their nar-
cissism relative to when they were not, F(1, 190) = 93.05,
p < 0.001, 𝜂p2 = 0.33. Thus, participants attended to narcis-
significant Target Type × Self‐Esteem Information interac- sism information. Perceivers may also underestimate the
tion, F(2, 374) = 81.60, p < 0.001, 𝜂p2 = 0.30 (Table 5). narcissism of highly narcissistic individuals when they lack
Replicating Study 1, perceivers rated high self‐esteem/ direct information about their narcissism. Conversely, per-
high narcissism targets higher in self‐esteem than high self‐ ceivers may interpret a label of “high” narcissism to indicate
esteem/low narcissism targets and low self‐esteem/low nar- a higher degree of narcissism than suggested by targets’ self‐
cissism targets, and rated high self‐esteem/low narcissism reported narcissism.
targets as higher in self‐esteem than low self‐esteem/low nar- Likewise, perceivers rated high self‐esteem/low narcis-
cissism targets, when perceivers were explicitly told targets’ sism targets lower in narcissism when they were told of their
self‐esteem level and when they were not (all Fs ≥ 12.71, narcissism relative to when they were not, F(1, 190) = 4.31,
ps ≤ 0.001). Perceivers were not oblivious to information p = 0.039, 𝜂p2 = 0.02. This suggests that perceivers may over-
about targets’ self‐esteem, however: They rated high self‐es- estimate the narcissism of high self‐esteem individuals who
teem/low narcissism targets higher in self‐esteem when they are low in narcissism when they lack direct information about
were explicitly told these targets had high self‐esteem rela- their narcissism. This pattern primarily occurred, however,
tive to when they were not, F(1, 190) = 23.73, p < 0.001, when participants were also informed of targets’ self‐esteem.
𝜂p2 = 0.11. They also rated low self‐esteem/low narcissism That is, when perceivers were told about targets’ self‐esteem,
targets lower in self‐esteem when they were told these indi- they rated high self‐esteem/low narcissism targets lower in
viduals had low self‐esteem relative to when they were not, narcissism if they were also told about their narcissism rela-
F(1,189) = 73.21, p < 0.001, 𝜂p2 = 0.28. Information about tive to when they were not, F(1, 95) = 9.41, p = 0.003, 𝜂p2 =
self‐esteem did not affect perceptions of high self‐esteem/ 0.09. In contrast, information about targets’ narcissism did
high narcissism targets’ self‐esteem, F(1,190) = 1.46, p = not affect perceptions of low self‐esteem/low narcissism tar-
0.229, 𝜂p2 = 0.01. gets, F(1, 189) = 0.06, p = 0.806, 𝜂p2 = 0.00.
Lastly, perceivers rated high self‐esteem targets (high self‐
esteem/high narcissism and high self‐esteem/low narcissism
3.2.2 | Perceptions of narcissism
targets) higher in narcissism when they were told about their
For narcissism ratings, there were main effects of target self‐esteem, relative to when they were not, but only when
type, F(2, 374) = 258.42, p < 0.001, 𝜂p2 = 0.58, and nar- targets’ narcissism was unspecified, (Fs ≥ 9.87, ps ≤ 0.002).
cissism information, F(1, 187) = 12.56, p < 0.001, 𝜂p2 = When given information about targets’ high self‐esteem, par-
0.06. There were also significant two‐way interactions ticipants may assume they are also more narcissistic, unless
between target type and self‐esteem information, F(2, specifically told otherwise.
374) = 5.01, p = 0.007, 𝜂p2 = 0.03, target type and narcis-
sism information,F(2, 374) = 76.59, p < 0.001, 𝜂p2 = 0.29,
and self‐esteem information and narcissism information,
3.2.3 | Liking
F(1, 187) = 7.94, p = 0.005, 𝜂p2 = 0.04, as well as a signifi- There was a main effect of target type on liking, F(2, 374) =
cant three‐way interaction, F(2, 374) = 4.24, p = 0.015, 𝜂p2 12.86, p < 0.001, 𝜂p2 = 0.06, that was qualified by an inter-
= 0.02 (Table 5). action between target type and narcissism information, F(2,
834
| GIACOMIN and JORDAN
TABLE 5 Study 2: Ratings of the dependent variables as a function of target type, self‐esteem information, and narcissism information
Target type
374) = 6.74, p = 0.001, 𝜂p2 = 0.04 (Table 5). When perceiv- about targets’ narcissism or not, they liked high self‐esteem/
ers were not told about targets’ narcissism, they liked high low narcissism targets more than low self‐esteem/low narcis-
self‐esteem/high narcissism targets more than high self‐es- sism targets (Fs ≥ 13.10, ps < 0.001). When perceivers are
teem/low narcissism targets, F(1, 96) = 6.11, p = 0.015, 𝜂p2 = explicitly told that narcissists are high in narcissism, they like
0.06, and low self‐esteem/low narcissism targets, F(1, 94) = them less, suggesting that perceptions of higher narcissism do
26.18, p < 0.001, 𝜂p2 = 0.22. This pattern was reversed, how- not necessarily make individuals more attractive.
ever, when perceivers were told about targets’ narcissism:
Perceivers liked high self‐esteem/high narcissism targets less
3.2.4 | Perceptions of communal and
than high self‐esteem/low narcissism targets, F(1, 94) = 4.96,
egoistic attributes
p = 0.028, 𝜂p2 = 0.05, and to the same degree as low self‐es-
teem/low narcissism targets, F(1, 94) = 0.36, p = 0.548, 𝜂p2 For ratings of communal attributes, there were main effects
= 0.00. Lastly, regardless of whether participants were told of target type, F(2, 374) = 69.60, p < 0.001, 𝜂p2 = 0.27, and
GIACOMIN and JORDAN
| 835
narcissism information, F(1, 187) = 4.99, p = 0.027, 𝜂p2 = 2 suggests that people do not view narcissism as inherently
0.03, which were qualified by a Target Type × Narcissism attractive or socially desirable when it is pointed out to them.
Information interaction, F(2, 374) = 10.52, p < 0.001, 𝜂p2 =
0.05 (Table 5). Whether perceivers were told about targets’
narcissism or not, they rated high self‐esteem/high narcissism
4 | STUDIES 3A AND 3B
targets as less communal than high self‐esteem/low narcis- Our focus on specific target types in Studies 1 and 2 allows
sism targets and low self‐esteem/low narcissism targets (all us to stringently test differences in perceptions of targets who
Fs ≥ 26.14, ps < 0.001). They also rated high self‐esteem/ are equally high in self‐esteem but differ in narcissism, but it
high narcissism targets as even less communal when explic- also restricts the variability in target narcissism and self‐es-
itly told about their narcissism compared to when they were teem. In Studies 3a and 3b, we present perceivers with targets
not explicitly told, F(1, 190) = 16.08, p < 0.001, 𝜂p2 = 0.08. representing a broader range of self‐esteem and narcissism
For ratings of egoistic attributes, there were significant using two new samples of targets. In Study 3a, we use a new
main effects of target type, F(2, 374) = 142.63, p < 0.001, sample of undergraduate targets, and in Study 3b, we use
𝜂p2 = 0.43, and narcissism information, F(1, 187) = 26.64, male Tinder dating profile photographs to further generalize
p = 0.014, 𝜂p2 = 0.032, which were qualified by a significant our findings to a meaningful context. Tinder is a mobile app
Target Type × Narcissism Information interaction, F(2, 374) that helps people to meet and potentially form romantic re-
= 22.52, p < 0.001, 𝜂p2 = 0.11 (Table 5).5 Whether narcis- lationships. People view photos of other users and “swipe”
sism was specified or not, high narcissism/high self‐esteem right to “like” a potential match or “swipe” left to ignore the
targets were rated as more egoistic than high self‐esteem/low potential match. Thus, people make decisions about who
narcissism targets and low self‐esteem/low narcissism targets they are interested in pursuing romantically largely based on
(all Fs ≥ 26.39, ps < 0.001). High self‐esteem/low narcissism physical appearance.
targets were also rated as more egoistic than low self‐esteem/ In both samples, we had perceivers rate targets’ self‐es-
low narcissism targets (all Fs ≥ 9.32, ps < 0.003). High self‐ teem and narcissism and report their liking of targets. Doing
esteem/high narcissism targets were also rated as more egois- so within a broader range of target narcissism and self‐esteem
tic when perceivers were told about their narcissism relative to helps ensure our results are not an artifact of examining only
when they were not, F(1, 190) = 24.68, p < 0.001, 𝜂p2 = 0.12. specific types of targets, and it allows us to test a fully spec-
Narcissism information, in contrast, did not affect ratings of ified path model of our proposed mediation. Consistent with
egoistic attributes for non‐narcissistic targets (all Fs ≤ 1.54, ps earlier findings, we expect perceivers to accurately perceive
≥ 0.216). Participants, as in Study 1, had a clear idea of what it targets’ self‐esteem and narcissism and to like narcissistic tar-
means to be narcissistic in terms of egoistic attributes. gets more than less narcissistic targets. We expected percep-
tions of self‐esteem to mediate the effect of target narcissism
on liking. In contrast, we expected perceptions of narcissism
3.3 | Discussion to reduce liking of targets. Such results would suggest that
liking of narcissists at zero acquaintance is driven primar-
Replicating Study 1, when participants were not told about
ily by perceptions that they have high self‐esteem rather than
targets’ narcissism, they accurately perceived narcissists to
perceptions of their narcissism, even for actual online dating
be more narcissistic, less communal, and more egoistic than
profile photos.
non‐narcissistic targets, including those with equally high
self‐esteem. They also, however, overestimated narcissists’
self‐esteem and liked them more than high self‐esteem tar- 4.1 | Method
gets with lower narcissism. This pattern of liking, however,
was completely reversed when participants were explicitly 4.1.1 | Targets
informed of targets’ narcissism. Perceivers continued to view For Study 3a, we took smiling photographs of 34 new under-
narcissists as more narcissistic, less communal, and more graduate targets (nine males) who had completed the NPI (α
egoistic than targets with high self‐esteem but low narcissism = 0.84) and RSES (α = 0.90).6 In Study 3b, we included only
(tendencies that were, in fact, accentuated), but they liked male undergraduate targets (N = 20; Mage = 19.38, SD = 1.87)
them less. This reversal is consistent with our contention that who completed the NPI (α = 0.84) and the RSES (α = 0.86),
perceivers typically like narcissists at zero acquaintance be- and provided researchers with their Tinder profile photograph.
cause they focus on exaggerated perceptions of these targets’
high self‐esteem and neglect accurate perceptions of their nar-
cissism. It is possible that an explicit label of high narcissism 4.1.2 | Perceivers
made participants less likely to report liking targets because In Study 3a, undergraduates (N = 252; 175 females, one un-
narcissism may be socially undesirable. Nevertheless, Study disclosed; Mage = 18.55, SD = 1.32) participated for partial
836
| GIACOMIN and JORDAN
course credit, with one participant excluded for invariant re- positively to perceived narcissism. In Study 3a, target self‐
sponding across items. In Study 3b, female undergraduates esteem related positively to perceived self‐esteem, though
who self‐identified as heterosexual (N = 225; Mage = 19.03, this was not significant in Study 3b. Interestingly, target self‐
SD = 2.68) participated for partial course credit.7 Sample size esteem related negatively to perceived narcissism; that is,
recommendations made by Westfall, Kenny, and Judd (2014) perceivers viewed high self‐esteem targets, overall, as less
suggest we should have 80% power to detect medium effect narcissistic. Target narcissism, however, related positively to
sizes. perceived self‐esteem, demonstrating that more narcissistic
targets are perceived to have higher self‐esteem even when
controlling their self‐reported self‐esteem. Next, we exam-
4.1.3 | Procedure and materials
ined direct effects of target narcissism and self‐esteem on
In Study 3a, perceivers viewed photographs of targets and liking. Target narcissism related positively to liking. Target
then rated their self‐esteem and narcissism, as well as how self‐esteem, somewhat surprisingly, related negatively to
much they liked them (α = 0.93), as in earlier studies. In liking in Study 3a but was positively related to liking in
Study 3b, perceivers viewed targets’ Tinder profile photo, Study 3b. Our final model predicted liking from target self‐
rated their self‐esteem (i.e., “This person is high in self‐es- esteem and narcissism as well as perceived self‐esteem and
teem”), rated their narcissism (i.e., “This person is narcis- narcissism (Figures 1 and 2).8 Perceived narcissism related
sistic”), and indicated whether they would want to meet negatively to liking, whereas perceived self‐esteem related
this person (i.e., “I would like to meet this person”) on
a scale ranging from 1 (strongly disagree) to 7 (strongly
agree).
TABLE 6 Studies 3a and 3b: Descriptive statistics and
correlations between primary dependent variables
4.2 | Results Perceptions
Table 6 presents descriptive statistics and correlations for
1 2 3
study variables at the perceiver level for Studies 3a and 3b.
1. Self‐esteem –0.002 0.44**
We conducted a series of cross‐classified analyses to test our
**
predictions. All scores were standardized to facilitate inter- 2. Narcissism 0.29 –0.24**
pretation. We examined direct effects of target narcissism and 3. Liking 0.44** 0.12**
self‐esteem on liking and indirect effects mediated through Study 3a M (SD) 5.57 (0.83) 3.69 (1.27) 5.33 (1.51)
perceptions of self‐esteem and narcissism (see Figure 1 for Study 3b M (SD) 4.55 (1.36) 3.58 (1.32) 3.64 (1.57)
Study 3a; Figure 2 for Study 3b). Note. Zero‐order correlations for Study 3a appear on top of the diagonal, and
First, we examined the relations between target self‐es- Study 3b correlations appear below the diagonal. Standard deviations are in
teem and narcissism and perceived self‐esteem and narcis- parentheses.
**
sism. As expected, in both samples, target narcissism related p < 0.01.
F I G U R E 1 Study 3a: Multilevel analyses investigating perceived self‐esteem and perceived narcissism as mediators between targets’ self‐
reported narcissism and self‐esteem and perceivers’ ratings of how much they like each target. Standardized betas and 95% confidence intervals (CI;
[lower limit, upper limit]) are presented.
GIACOMIN and JORDAN
| 837
F I G U R E 2 Study 3b: Multilevel analyses investigating perceived self‐esteem and perceived narcissism as mediators between targets’
self‐reported narcissism and self‐esteem and perceivers’ ratings of how much they would like to meet each target. Standardized betas and 95%
confidence intervals (CI; [lower limit, upper limit]) are presented
positively to liking, even with target self‐esteem and narcis- perceived to have higher self‐esteem, even when their self‐
sism controlled. reported self‐esteem was controlled. In addition, perceivers
We examined a series of multilevel mediation models liked more narcissistic targets more overall. This total effect
using Monte Carlo simulation with 20,000 resamples. First, occurred despite a negative effect of perceived narcissism on
we tested whether perceived narcissism mediates the relation liking because it was overshadowed by a positive effect of
between target narcissism and liking. The indirect effect (IE) perceived self‐esteem on liking. These two mediation paths
was negative, Study 3a: IE = –0.012, CI [–0.016, –0.008], provide strong support for our contention that narcissists
and Study 3b: IE = –0.005, CI [–0.008, –0.003], suggesting are generally well liked at zero acquaintance because peo-
that perceivers like narcissistic targets less because they view ple focus predominantly on their perceptions that narcissists
them as more narcissistic. Next, we tested whether perceived have high self‐esteem and neglect their perceptions that they
self‐esteem mediates the relation between target narcissism are narcissistic. Indeed, to the extent that they perceive them
and liking. The indirect effect was positive, Study 3a: IE = to be more narcissistic, they like them less. Moreover, this
0.017, CI [0.009, 0.025], and Study 3b: IE = 0.045, CI [0.031, process occurred when heterosexual female perceivers rated
0.059], suggesting that perceivers like narcissistic targets male online dating profile photos.
more because they perceive them to have higher self‐esteem. We also examined the processes that link targets’ self‐es-
We also examined whether perceived self‐esteem medi- teem to liking. In Study 3a, targets’ self‐esteem was posi-
ates the relation between target self‐esteem and liking. The tively associated with perceptions of self‐esteem, which were
indirect effect was significant only in Study 3a, Study3a: IE in turn positively associated with liking. Targets’ self‐esteem
= 0.010, CI [0.002, 0.018], and Study 3b: IE = 0.010, CI was negatively associated with perceptions of narcissism,
[–0.007, 0.025], indicating that perceivers like high self‐es- which were in turn negatively associated with liking. Thus,
teem targets more because they perceive them to have higher higher self‐esteem targets were better liked because they
self‐esteem. Finally, we tested whether perceived narcissism were perceived to be higher in self‐esteem and less narcis-
mediates the link between target self‐esteem and liking. The sistic. This overall pattern of results suggests that perceptions
indirect effect was significant, Study 3a: IE = 0.011, CI of self‐esteem and narcissism are influential in leading high
[0.007, 0.014], and Study 3b: IE = 0.004, CI [0.003, 0.007], self‐esteem individuals to be better liked at zero acquain-
indicating that participants like high self‐esteem targets more tance. Surprisingly, however, perceivers liked higher self‐
because they perceive them to be less narcissistic. esteem targets less overall, even when perceived narcissism
and self‐esteem were taken into account in Study 3a. It is
not clear why this is so. Presumably, observable cues asso-
4.3 | Discussion
ciated with self‐esteem, but not perceptions of self‐esteem
Studies 3a and 3b replicate and extend our findings, notably or narcissism, made high self‐esteem targets less likable. In
to a psychologically meaningful context (i.e., online dating). Study 3b, however, target self‐esteem was positively related
Perceivers relatively accurately perceived targets’ narcis- to liking overall. Further research is needed to establish how
sism and self‐esteem (though not as well through online self‐esteem relates to perceptions of narcissism and liking at
dating profiles). More narcissistic targets, moreover, were zero acquaintance.
838
| GIACOMIN and JORDAN
TABLE 7 Supplemental Study: Perceptions of self‐esteem, narcissism, and liking as a function of target type
Target type
narcissistic targets more than targets who were high in self‐ ENDNOTES
esteem but low in narcissism, F(2, 150) = 28.71, p < 0.001, 1
We focus here on grandiose narcissism rather than vulnerable narcissism
𝜂p2 = 0.28. These findings suggest that our central results are (Campbell & Foster, 2007). We refer to narcissists as shorthand for “in-
not a consequence of perceivers’ concurrently judging targets’ dividuals high in narcissism.”
self‐esteem and narcissism, and reporting their own liking. 2
Targets completed additional measures relevant to research questions ad-
dressed in Giacomin and Jordan (2016), which examined whether per-
ceptions of targets’ self‐esteem relate to targets’ implicit self‐esteem, and
7 | IM P LICAT ION S A N D did not consider how targets’ narcissism affects perceptions of self‐es-
CO NC LU S IO N S teem or liking.
3
Although it does not qualify the effects of primary interest, the interac-
People appear to like narcissists during initial encounters be- tion indicates that participants in the photograph‐and‐transcript condition
cause they overestimate their self‐esteem and neglect their rated targets as higher in self‐esteem than those in the photograph‐only
condition in the two high self‐esteem target conditions (all Fs ≥ 8.14, ps
narcissism. Highly positive first impressions of narcissists are
≤ 0.005), but not the low self‐esteem target condition, F(1, 126) = 0.95,
potentially problematic because narcissists tend to become p = 0.331, 𝜂p2 = 0.01.
disliked over time (Campbell, 1999; Paulhus, 1998). This 4
The interaction on perceptions of narcissism indicates that participants in
dynamic may have important implications in a variety of do- the photograph‐and‐transcript condition rated the high self‐esteem/low
mains, such as politics and leadership. People report greater narcissism targets, F(1, 127) = 4.14, p = 0.044, 𝜂p2 = 0.03—but not other
willingness to vote for presidential candidates who they targets (all Fs ≤ 0.92, ps ≥ 0.679)—as higher in narcissism than did
perceive to have higher self‐esteem (Zeigler‐Hill & Myers, participants in the photograph‐only condition. This interaction, however,
2009). It may be that people perceive narcissists more through does not qualify the effects of target type, nor are the simple effects sig-
the lens of high self‐esteem than narcissism, even when mak- nificant with the modified Bonferroni correction.
5
ing consequential decisions, like choosing whom to vote for. There was also a Target Type × Self‐Esteem Information interaction,
Understanding why narcissists make positive first impres- F(2, 374) = 4.33, p = 0.014, 𝜂p2 = 0.02 (Table 5). Although none of
the simple effects were significant using the modified Bonferroni
sions may also suggest ways to temper those impressions. This
correction, perceivers rated high self‐esteem/high narcissism targets
may be desirable because people tend to dislike narcissists over
as marginally more egoistic when told about their self‐esteem rela-
time. Study 2 suggests that making narcissism more salient to tive to when they were not, F(1, 190) = 3.75, p = 0.054, 𝜂p2 = 0.02.
perceivers might reverse the tendency to like them more than Ratings of high self‐esteem/low narcissism and low self‐esteem/
less narcissistic individuals. Focusing people on the possibility low narcissism targets were not affected by self‐esteem information
that others may be narcissistic in early dating encounters or (all Fs ≤ 2.67, ps ≥ 0.104). In addition, there was an interaction
when hiring might lead them to form different preferences to- between self‐esteem information and narcissism information that
ward others. Understanding the dynamics of how impressions did not qualify any of the findings, F(1, 187) = 4.79, p = 0.030,
𝜂p2 = 0.03. When perceivers were not told about targets’ self‐es-
of others relate to those individuals’ narcissism may be import-
teem, they perceived targets to be more egoistic when told about
ant for ultimately allowing people to form first impressions
their narcissism relative to when they were not, F(1, 93) = 10.83,
that they are more satisfied with over the long term. Similar p < 0.001, 𝜂p2 = 0.40. However, when perceivers were told about tar-
dynamics may govern first impressions of many individuals gets’ self‐esteem, narcissism information did not affect perceptions
with inconsistent personal qualities, such as being likable but of egotism, F(1, 95) = 0.11, p = 0.741, 𝜂p2 = 0.00).
untrustworthy, charming but manipulative, or competent but 6
Targets completed additional measures relevant to research questions not
disagreeable. Some personal qualities may be associated with addressed here (see Giacomin & Jordan, 2016).
clearer cues of likability than others and may suggest higher 7
We limited Study 3b to heterosexual female perceivers and male targets to
self‐esteem while masking less desirable qualities. avoid potential gender differences in perceptions of personality in online
dating contexts.
8
In Study 3a, targets also completed the Big Five Aspects Scale (DeYoung,
ACKNOWLEDGMENTS
Quilty, & Peterson, 2007). The results of Study 3a are identical when
The author(s) disclosed receipt of the following financial sup- target extraversion is controlled.
port for the research, authorship, and/or publication of this ar-
ticle: Preparation of this manuscript was supported by Social
Sciences and Humanities Research Council of Canada. ORCID
Miranda Giacomin http://orcid.org/0000-0001-5310-3765
CONFLICT OF INTERESTS
R E F E R E NC E S
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this Back, M. D., Schmukle, S. C., & Egloff, B. (2008). How extraverted
article. is honey.bunny77@ hotmail.de? Inferring personality from e‐mail
GIACOMIN and JORDAN
| 841
addresses. Journal of Research in Personality, 42, 1116–1122. Giacomin, M., & Jordan, C. H. (2016). How implicit self‐esteem affects
https://doi.org/10.1016/j.jrp.2008.02.001 perceptions of self‐esteem at zero and non‐zero acquaintance. Self
Back, M. J., Schmukle, S. C., & Egloff, B. (2010). Why are narcissists & Identity, 15, 615–637. https://doi.org/10.1080/15298868.2016.1
so charming at first sight? Decoding the narcissism‐popularity link 183516
at zero‐acquaintance. Journal of Personality and Social Psychology, Giacomin, M., & Rule, N. O. (2018). Eyebrows cue grandiose narcis-
98, 132–145. https://doi.org/10.1037/a0016338 sism. Journal of Personality. Advance Online Publication. https://
Bosson, J. K., Lakey, C. E., Campbell, W. K., Zeigler‐Hill, V., Jordan, doi.org/10.1111/jopy.12396
C. H., & Kernis, M. H. (2008). Untangling the links between nar- Gibson, J. J. (1979). The ecological approach to visual perception.
cissism and self‐esteem: A theoretical and empirical review. Social Boston, MA: Houghton Mifflin.
and Personality Psychology Compass, 2, 1415–1439. https://doi. Hirschmüller, S., Schmukle, S. C., Krause, S., Back, M. D., & Egloff,
org/10.1111/j.1751-9004.2008.00089.x B. (2018). Accuracy of self‐esteem judgments at zero acquain-
Brown, R. P., & Zeigler‐Hill, V. (2004). Narcissism and the non‐equiv- tance. Journal of Personality, 86, 308–319. https://doi.org/10.1111/
alence of self‐esteem measures: A matter of dominance? Journal jopy.12316
of Research in Personality, 38, 585–592. https://doi.org/10.1016/j. Holm, S. (1979). A simple sequentially rejective multiple test proce-
jrp.2003.11.002 dure. Scandinavian Journal of Statistics, 6, 65–70. Retrieved from
Brummelman, E., Thomaes, S., & Sedikides, C. (2016). Separating https://www.jstor.org/stable/4615733
narcissism from self‐esteem. Current Directions in Psychological Holtzman, N. S. (2011). Facing a psychopath: Detecting the Dark
Science, 25, 8–13. https://doi.org/10.1177/0963721415619737 Triad from emotionally‐neutral faces, using prototypes from the
Brunswik, E. (1956). Perception and the representative design of exper- Personality Faceaurus. Journal of Research in Personality, 45, 648–
iments. Berkeley, CA: University of California Press 654. https://doi.org/10.1016/j.jrp.2011.09.002
Buffardi, L. E., & Campbell, W. K. (2008). Narcissism and social net- Holtzman, N. S., & Strube, M. J. (2013). People with dark per-
working web sites. Personality and Social Psychology Bulletin, 34, sonalities tend to create a physically attractive veneer. Social
1303–1314. https://doi.org/10.1177/0146167208320061 Psychological and Personality Science, 4, 461–467. https://doi.
Cameron, J. J., MacGregor, J. C. D., & Kwang, T. (2013). Self‐esteem org/10.1177/1948550612461284
as an interpersonal signal. In V. Zeigler‐Hill (Ed.), Current issues Judd, C. M., Kenny, D. A., & McClelland, G. H. (2001). Estimating
in self‐esteem research (pp. 145–162). New York, NY: Psychology and testing mediation and moderation in within‐subject de-
Press. signs. Psychological Methods, 6, 115–134. https://doi.
Cameron, J. J., Stinson, D. A., Hoplock, L., Hole, C., & Schellenberg, org/10.1037/1082-989X.6.2.115
J. (2016). The robust self‐esteem proxy: Impressions of self‐esteem Judge, T. A., LePine, J. A., & Rich, B. L. (2006). Loving yourself abun-
inform judgments of personality and social value. Self and Identity, dantly: Relationship of the narcissistic personality to self and other
1–18, https://doi.org/10.1080/15298868.2016.1175373 perceptions of workplace deviance, leadership, and task and con-
Campbell, K. W. (1999). Narcissism and romantic attraction. Journal textual performance. Journal of Applied Psychology, 91, 762–776.
of Personality and Social Psychology, 77, 1254–1270. https://doi. https://doi.org/10.1037/0021-9010.91.4.762
org/10.1037/0022-3514.77.6.1254 Küfner, A. C., Nestler, S., & Back, M. D. (2013). The two pathways to
Campbell, W. K., & Foster, J. D. (2007). The narcissistic self: being an (un‐)popular narcissist. Journal of Personality, 81, 184–
Background, an extended agency model, and ongoing controversies. 195. https://doi.org/10.1111/j.1467-6494.2012.00795.x
In C. Sedikides & S. Spencer (Eds.), Frontiers in social psychology: Leckelt, M., Küfner, A. C., Nestler, S., & Back, M. D. (2015). Behavioral
The self (pp. 115–138). Philadelphia, PA: Psychology Press. processes underlying the decline of narcissists’ popularity over
Campbell, W. K., Rudich, E. A., & Sedikides, C. (2002). Narcissism, time. Journal of Personality and Social Psychology, 109, 856–871.
self‐esteem, and the positivity of self‐views: Two portraits of self‐ https://doi.org/10.1037/pspp0000057
love. Personality and Social Psychology Bulletin, 28, 358–368. Naumann, L. P., Vazire, S., Rentfrow, P. J., & Gosling, S. D. (2009).
https://doi.org/10.1177/0146167202286007 Personality judgments based on physical appearance. Personality
Clifton, A., Turkheimer, E., & Oltmanns, T. F. (2009). Personality disor- and Social Psychology Bulletin, 35, 1661–1671. https://doi.
der in social networks: Network position as a marker of interpersonal org/10.1177/0146167209346309
dysfunction. Social Networks, 31, 26–32. https://doi.org/10.1016/j. Nestler, S., & Back, M. D. (2013). Applications and extensions of the
socnet.2008.08.003 lens model to understand interpersonal judgments at zero acquain-
DeYoung, C. G., Quilty, L. C., & Peterson, J. B. (2007). Between tance. Current Directions in Psychological Science, 22, 374–379.
facets and domains: 10 aspects of the Big Five. Journal of https://doi.org/10.1177/0963721413486148
Personality and Social Psychology, 93, 880–896. https://doi. Oltmanns, T. F., Friedman, J. N., Fiedler, E. R., & Turkheimer, E.
org/10.1037/0022-3514.93.5.880 (2004). Perceptions of people with personality disorders based on
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G* Power thin slices of behavior. Journal of Research in Personality, 38, 216–
3: A flexible statistical power analysis program for the social, be- 229. https://doi.org/10.1016/S0092-6566(03)00066-7
havioral, and biomedical sciences. Behavior Research Methods, 39, Paulhus, D. L. (1998). Interpersonal and intrapsychic adaptive-
175–191. https://doi.org/10.3758/BF03193146 ness of trait self‐enhancement: A mixed blessing. Journal of
Funder, D. C (1995). On the accuracy of personality judgment: A re- Personality and Social Psychology, 74, 1197–1208. https://doi.
alistic approach. Psychological Review, 102, 652–670. https://doi. org/10.1037/0022-3514.74.5.1197
org/10.1037/0033-295X.102.4.652 Raskin, R., & Terry, H. (1988). A principal‐components analysis of the
Funder, D. C (2012). Accurate personality judgment. Current Narcissistic Personality Inventory and further evidence of its con-
Directions in Psychological Science, 21, 177–182. https://doi. struct validity. Journal of Personality and Social Psychology, 54,
org/10.1177/0963721412445309 890–902. https://doi.org/10.1037/0022-3514.54.5.890
842
| GIACOMIN and JORDAN
Rosenberg, M. (1965). Society and the adolescent self‐image. Princeton, Zeigler‐Hill, V., & Myers, E. M. (2009). Is high self‐esteem a path to the
NJ: Princeton University Press. White House? The implicit theory of self‐esteem and the willing-
Uleman, J. S., Adil Saribay, S., & Gonzalez, C. M. (2008). Spontaneous ness to vote for presidential candidates. Personality and Individual
inferences, implicit impressions, and implicit theories. Annual Differences, 46, 14–19. https://doi.org/10.1016/j.paid.2008.08.018
Review of Psychology, 59, 329–360. https://doi.org/10.1146/an- Zeigler‐Hill, V., & Myers, E. M. (2011). An implicit theory of self‐
nurev.psych.59.103006.093707 esteem: The consequences of perceived self‐esteem for romantic
Vazire, S., Naumann, L. P., Rentfrow, P. J., & Gosling, S. D. (2008). desirability. Evolutionary Psychology, 9, 147–180. https://doi.
Portrait of a narcissist: Manifestations of narcissism in physical org/10.1177/147470491100900202
appearance. Journal of Research in Personality, 42, 1439–1447.
https://doi.org/10.1016/j.jrp.2008.06.007
Westfall, J., Kenny, D. A., & Judd, C. M. (2014). Statistical power and
optimal design in experiments in which samples of participants re- SUPPORTING INFORMATION
spond to samples of stimuli. Journal of Experimental Psychology:
Additional supporting information may be found online in
General, 143, 2020–2045. https://doi.org/10.1037/xge0000014
Wurst, S. N., Gerlach, T. M., Dufner, M., Rauthmann, J. F., Grosz, M. P.,
the Supporting Information section at the end of the article.
Küfner, A. C., & Back, M. D. (2017). Narcissism and romantic rela- Appendix S1
tionships: The differential impact of narcissistic admiration and ri- Appendix S2
valry. Journal of Personality and Social Psychology, 112, 280–306. Table S1
https://doi.org/10.1037/pspp0000113
Zeigler‐Hill, V., & Besser, A. (2014). Self‐esteem and evaluations of tar-
gets with ostensibly different levels of self‐worth. Self and Identity, How to cite this article: Giacomin M, Jordan CH.
13, 146–161. https://doi.org/10.1080/15298868.2013.770194 Misperceiving grandiose narcissism as self‐esteem:
Zeigler‐Hill, V., Besser, A., Myers, E. M., Southard, A. C., & Malkin, Why narcissists are well liked at zero acquaintance.
M. L. (2013). The status‐signaling property of self‐esteem: The role
Journal of Personality. 2019;87:827–842. https://doi.
of self‐reported self‐esteem and perceived self‐esteem in person-
ality judgments. Journal of Personality, 81, 209–220. https://doi.
org/10.1111/jopy.12436
org/10.1111/j.1467-6494.2012.00790.x
This document is a scanned copy of a printed document. No warranty is given about the
accuracy of the copy. Users should refer to the original published version of the material.