ArticlePDF Available

Abstract and Figures

The aim of this research was to explore the underlying structural organization and centrality of narcissism features by using network analysis. The research was conducted on a sample from the general population using subscales of four narcissism instruments (Narcissism Personality Inventory, Grandiose Narcissism Scale, Pathological Narcissism Inventory, Five-Factor Narcissism Inventory Short Form). The results revealed detection of four communities, one of which was interpreted as vulnerable narcissism, while the remaining three were interpreted as aspects of grandiose narcissism (grandiose exhibitionism, narcissistic antagonism, and authority). The results suggested that although features of grandiose narcissism (grandiose exhibitionism and leadership/authority) show consistent higher strength centrality across networks, entitlement/exploitativeness, followed by grandiose fantasy, has a central role in bringing together maladaptive grandiose and vulnerable aspects of narcissism. In addition, leadership/authority brings together various aspects of grandiose narcissism. The results support the role of entitlement features as the core of narcissism.
Content may be subject to copyright.
This is a post-peer-review, pre-copyedit version of an artcle published in Current Psychology.
The final authentcated version is available online at:
Please cite this paper as: Dinić, B. M., Sokolovska, V., & Tomašević, A. (2021). The narcissism
network and centrality of narcissism features. Current Psychology. Online First.
Full text is avaiable at:
6The narcissism network and centrality of narcissism features
7Bojana M. Dinić
&Valentina Sokolovska
&Aleksandar Tomašević
9Accepted: 3 December 2020
10 #Springer Science+Business Media, LLC, part of Springer Nature 2020
11 Abstract
12 The aim of this research was to explore the underlying structural organization and centrality of narcissism features by using
13 network analysis. The research was conducted on a sample from the general population using subscales of four narcissism
14 instruments (Narcissism Personality Inventory, Grandiose Narcissism Scale, Pathological Narcissism Inventory, Five-Factor
15 Narcissism Inventory Short Form). The results revealed detection of four communities, one of which was interpreted as vulner-
16 able narcissism, while the remaining three were interpreted as aspects of grandiose narcissism (grandiose exhibitionism, narcis-
17 sistic antagonism, and authority). The results suggested that although features of grandiose narcissism (grandiose exhibitionism
18 and leadership/authority) show consistent higher strength centrality across networks, entitlement/exploitativeness, followed by
19 grandiose fantasy, has a central role in bringing together maladaptive grandiose and vulnerable aspects of narcissism. In addition,
20 leadership/authority brings together various aspects of grandiose narcissism. The results support the role of entitlement features as
21 the core of narcissism.
22 Keywords Grandiose narcissism .Vulnerable narcissism .Narcissistic antagonism .Entitlement/exploitativeness .Leadership/
23 authority .Grandiose fantasies .Network analysis
25 Introduction
26 Narcissism is a complex trait in which two main dimensions
27 could be distinguished grandiose and vulnerable narcissism.
28 Grandiose narcissism refers to inflated self-image, sense of
29 superiority, and exhibitionism, which are associated with ap-
30 proach motivation and self-enhancement strategies (e.g.,
31 Krizan & Herlache, 2018; Miller, Lynam, Hyatt, &
32 Campbell, 2017). On the other hand, interest in vulnerable
33 narcissism comes from clinical research in which this form
34 of narcissism refers to experience of low self-esteem, negative
35 emotions, and hypersentivity to rejection and criticism, which
36 are associated with avoidance of interpersonal relationships
37and self-protective strategies (e.g., Krizan & Herlache, 2018;
38Miller et al., 2017).
39However, the complexity of narcissism is not limited to
40detection of its main dimensions, as each of these dimen-
41sions also has subdimensions which vary across different
42measurements and approaches. For example, the most com-
43monly used measure of narcissism Narcissistic
44Personality Inventory (NPI; Raskin & Terry, 1988)assesses
45only grandiose narcissism and originally it is supposed to
46capture seven subscales: authority, self-sufficiency, superi-
47ority, exhibitionism, exploitativeness, vanity, and entitle-
48ment. Although this measure provided good content and
49criterion validity (e.g., Miller, Lynam, & Campbell, 2016;
50Miller, Price, & Campbell, 2012b), the results regarding
51both the number of the subscales and its structure were in-
52consistent (for a review of proposed models, see Dinić&
53Vujić,2019a) while reliabilities of some subscales were
54reported to be low (e.g., Miller, Nicols, Clark, Daniels, &
55Grant, 2018). In order to overcome the shortcomings of
56NPI, Foster, McCain, Hibberts, Brunell, and Johnson
57(2015) developed a new measure of grandiose narcissism
58Grandiose Narcissism Inventory (GNI), based on the con-
59tent of the original NPI structure. GNI subscales showed
60good reliabilities and solely the relations with paired mea-
61sures, while NPI subscales showed more diverse relations
62with paired measures (Foster et al., 2015).
*Bojana M. Dinić
Valentina Sokolovska
Aleksandar Tomašević
Department of Psychology, Faculty of Philosophy, University of
Novi Sad, Dr Zorana Đinđića 2, Novi Sad 21000, Serbia
Department of Sociology, Faculty of Philosophy, University of Novi
Sad, Dr Zorana Đinđića 2, Novi Sad 21000, Serbia
Current Psychology
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
63 On the other hand, measurement of vulnerable narcissism
64 has been neglected for a long time. Although there are scales
65 which assess vulnerable narcissism (e.g., Hypersensitive
66 Narcissism Scale; Hendin & Cheek, 1997), they treat vulner-
67 able narcissism rather as a unidimensional construct.
68 Furthermore, there are controversies regarding whether NPI
69 measures normal or pathological narcissism (e.g., Rosenthal
70 &Hooley,2010; Pincus & Lukowitsky, 2010; Miller, Gentile,
71 & Campbell, 2012a). In order to develop a pathological mea-
72 sure of both grandiose and vulnerable narcissism, Pincus et al.
73 (2009) constructed Pathological Narcissism Inventory (PNI).
74 In the final structure of PNI, grandiose narcissism captures
75 exploitativeness, grandiose fantasy, and sacrificing self-en-
76 hancement, while vulnerable narcissism captures contingent
77 self-esteem, hiding the self, devaluating, and entitlement rage
78 (Wright, Lukowitsky, Pincus, & Conroy, 2010). In addition,
79 Glover, Miller, Lynam, Crego, & Widiger (2012) developed
80 the Five-Factor Narcissism Inventory which also measures
81 both grandiose and vulnerable narcissism. Grandiose narcis-
82 sism captures 11 subscales, including exhibitionism, authori-
83 tativeness, grandiose fantasies, manipulativeness,
84 exploaitativess, entitlement, while vulnerable narcissism cap-
85 tures four subscales reactive anger, shame, need for admira-
86 tion, and cynicism/distrust.
87 There are many proposals for the main features or coreof
88 both grandiose and vulnerable narcissism. Brown, Budzek,
89 and Tamborski (2009) proposed two independent features
90 grandiosity and entitlement, while Wright and Edershile
91 (2018) proposed only entitlement. These proposals are in line
92 with Narcissism Spectrum Model (NSM; Krizan & Herlache,
93 2018), in which entitled self-importance is the core feature and
94 narcissism expression can vary across grandiose and vulnera-
95 ble forms. Entitlement refers to the belief that one deserves
96 special treatment, while in NSM it also includes lack of em-
97 pathy and manipulation. The remaining two factors of NSM
98 are grandiosity/exhibitionism, which includes authoritative-
99 ness and exhibitionism,and vulnerability, which includes con-
100 tingent self-esteem and entitlement rage. Similarly, Miller
101 et al. (2017) propose a trifurcated model in which antagonism
102 serves as the narcissism core, but other characteristics modu-
103 late whether narcissism would be expressed as grandiose or
104 vulnerable. Thus, grandiose narcissism is modulated by
105 agentic aspect of extraversion and vulnerable narcissism by
106 negative affect of neuroticism.
107 The emerging consensus is that narcissism is best repre-
108 sented by a triad, but the core feature is somewhat different
109 across the approaches, although entitlement could be seen as a
110 part of a broader construct of antagonism. In factor analysis of
111 items of several narcissism measures, the core factor of antag-
112 onism actually includes entitlement (Crowe, Lynam,
113 Campbell, & Miller, 2019). However, there is disagreement
114 about centrality of narcissism features between the proposed
115 models and expertsviews. Namely, a study on expertsviews
116showed that they generally believe that grandiose features of
117narcissism are more central (e.g., grandiose presentation of
118self-concept, entitlement, exhibitionism, grandiose fantasies)
119than vulnerable features (e.g., Ackerman, Hands, Donnellan,
120Hopwood, & Witt, 2017).
121Previous studies used factor analytic approach in order to
122determine the central factor of narcissism (e.g., Krizan &
123Herlache, 2018: Crowe et al., 2019). The aim of this research
124is to explore the centrality of narcissism features by using a
125different approach, i.e., network analysis, which offers direct
126measurement of centrality of each variable based on interrela-
127tions between all other variables. Moreover, structural organi-
128zation of narcissism features was detected by community anal-
129ysis. Unlike latent structure models, community structure is
130detected from all indirect and direct relationships between the
131variables in question, focusing on the unique variance be-
132tween the variables and not on common shared variance
133(e.g., Constantini et al., 2015). Thus, there is no a priori as-
134sumption about the connection between the nodes and their
135community membership. Besides the community structure it-
136self, this analysis offers a better insight into the centrality of
137narcissism features, i.e., features which bring together the de-
138tected communities. Thus, the features that bring together the
139detected communities could serve as the core features of the
140network (bridge-centrality) and features that are highly con-
141nected with other nodes are the core features of the commu-
142nities themselves (strength centrality). Although there are sev-
143eral proposals of the main narcissism component (e.g., Krizan
144& Herlache, 2018; Miller et al., 2017), network analysis could
145be used to directly test the centrality of narcissism features
146considering the complexity of this trait. Previous network
147analysis research focused on only one measure of narcissism.
148For example, Briganti and Linkowski (2019) analyzed the
149network of NPI items and showed that the items referring to
150entitlement, authority, and superiority had the highest
151centrality. In addition, a study by Di Pierro, Constantini,
152Benzi, Madeddu, and Preti (2019) suggested that subscales
153of grandiose fantasies, entitlement rage, and contingent self-
154esteem had the central role in PNI. In the current study, we
155included various measures of narcissism which cover the main
156aspects of narcissism, in order to provide a better insight into
157central features of this multidimensional trait.
159Participants and Procedure
160The sample included 423 participants (50.1% females, 3 par-
161ticipants did not report sex) from the general population of
162Serbia, aged between 18 and 85 (M= 31.59, SD = 14.24).
163The majority of the sample was highly educated (40% under-
164graduate students and 25.5% higher school or faculty
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
165 graduates). Data were collected by trained undergraduate stu-
166 dents, each of whom was required to collect data from 6 par-
167 ticipants, in accordance with the given sex (3 males and 3
168 females) and age quotas (from 18 to 30 and over 30 years of
169 age). Regardless of the given quotas, it was a convenience
170 sample, thus the examiners recruited participants from their
171 community. A part of the data was also used in Citation
172 Blinded. The research was approved by the Institutional
173 Review Board.
174 Measures
175 Narcissism Personality Inventory (NPI: Raskin & Terry,
176 1988, for Serbian adaptation see Dinić&Vujić,2019a)as-
177 sesses grandiose narcissism and it consists of 40 items.
178 However, in this research we applied the most commonly
179 used scoring according to Ackerman et al. (2011)tobeable
180 to compare the results with previous studies. This scoring
181 includes 26 items measuring three subscales: leadership/au-
182 thority, grandiose exhibitionism, and entitlement/
183 exploitativeness. The participants rated each statement on a
184 6-point scale, ranging from 0 = not at all like me to 5 = very
185 much like me. Descriptives and reliabilities for all used mea-
186 sures are presented in Table 1.
187 Grandiose Narcissism Scale (GNS; Foster et al., 2014) con-
188 tains 35 items which measure seven subscales of grandiose
189 narcissism: authority, self-sufficiency, superiority, vanity,
190exhibitionism, entitlement, and exploitativeness. The partici-
191pants rated each statement on a 6-point scale, ranging from
1921= strongly disagree to 6 = strongly agree. Since this is the
193first use of Serbian adaptation of GNS, we calculated model fit
194for proposed 7-factor model and it was excellent
(539) = 2064.88, p< .001, CFI
= .976,
= .051, SRMR = .061). The pa-
197rameters were high, ranging from .51 to .86, while the corre-
198lations among the factors were in the range from .22 (between
199self-sufficiency and exhibitionism) to .81 (between exhibi-
200tionism and entitlement).
201Pathological Narcissism Inventory (PNI: Pincus et al.,
203cludes 52 items and measures seven subscales which assess
204pathological grandiose narcissism (exploitativeness, grandi-
205ose fantasy, and self-sacrificing self-enhancement) and vul-
206nerable narcissism (contingent self-esteem, hiding the self,
207devaluing, and entitlement rage). The participants rated each
208statement on a 6-point scale, ranging from 0 = not at all like
209me to 5 = very much like me.
210Five-Factor Narcissism Inventory Short Form (FFNI-SF;
211Sherman et al., 2015) contains 60 items and measures 15
212subscales that assess grandiose narcissism (indifference, exhi-
213bitionism, authoritativeness, grandiose fantasies, manipula-
214tiveness, exploaitativess, entitlement, lack of empathy, arro-
215gance, acclaim seeking, and thrill seeking) and vulnerable
216narcissism (reactive anger, shame, need for admiration, and
t1:1Table 1 Descriptives and reliability of narcissism subscales
t1:3Narcissism Personality Inventory Five Factor Narcissism Inventory
t1:4leadership/authority 24.63(12.62) .91 indifference 11.71(3.80) .73
t1:5grandiose exhibitionism 19.35(11.03) .86 exhibitionism 12.57(3.83) .75
t1:6entitlement/exploitativeness 8.34(4.64) .74 authoritativeness 11.93(3.82) .78
t1:7Grandiose Narcissism Scale grandiose fantasies 10.85(4.11) .75
t1:8authority 16.14(6.21) .89 manipulativeness 9.78(3.81) .83
t1:9self-sufficiency 21.02(5.24) .79 exploitativeness 7.15(3.50) .84
t1:10 superiority 13.89(5.41) .83 entitlement 9.21(3.56) .73
t1:11 vanity 20.00(5.97) .89 lack of empathy 7.98(3.34) .75
t1:12 exhibitionism 13.88(5.81) .88 arrogance 8.41(3.15) .58
t1:13 entitlement 14.10(4.99) .80 acclaim seeking 12.83(3.72) .79
t1:14 exploitativeness 10.24(5.08) .86 thrill seeking 9.69(3.81) .74
t1:15 Pathological Narcissism Inventory reactive anger 11.70(3.44) .64
t1:16 exploitativeness 10.89(5.50) .77 shame 12.55(3.75) .74
t1:17 grandiose fantasy 15.86(8.7) .87 need for admiration 10.50(3.23) .55
t1:18 self-sacrificing self-enhancement 15.60(6.05) .72 cynicism/distrust 11.72(3.23) .56
t1:19 contingent self-esteem 20.92(12.78) .90
t1:20 hiding the self 17.03(6.86) .72
t1:21 devaluing 12.47(7.20) .80
t1:22 entitlement rage 17.08(8.52) .83
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
217 cynicism/distrust). The participants rated each statement on a
218 5-point scale, ranging from 1 = strongly disagree to 5 = strong-
219 ly agree. Since this is the first use of Serbian adaptation of
220 FFNI, we calculated model fit for proposed 15-factor model
221 and it was good (WLSMχ
(1605) = 4811.96, p< .001,
222 CFI
= .931, TLI
= .923, RMSEA
= .054,
223 SRMR = .066). The parameters were in the range from .37
224 to .83, except for three reverse coded items which had load-
225 ings < .30 (items 19, 27, and 38). The correlations among the
226 factors were in the range from .31 (between indifference and
227 need for admiration) to .86 (between acclaim seeking and
228 grandiose fantasies).
229 Data Analysis
230 Network analysis was conducted on 32 subscales from 4 used
231 narcissism instruments. Statistical networks have two impor-
232 tant elements: (1) nodes, which represent observed variables
233 (e.g., subscales of narcissism), and (2) edges connecting the
234 nodes and representing some form of statistical relationship
235 which is estimated from data. In this study, the zero-order and
236 the partial correlations (with adaptive LASSO regularization)
237 were used as the method of edge estimation. Since all nodes
238 are from the same domain of narcissism, we included the
239 network based on zero-order correlations regarding some con-
240 cerns that residualization could affect the relations between
241 correlated variables (e.g., Vize, Collison, Miller, & Lynam,
242 2020). In contrast, the network based on regularized partial
243 correlations is sparser and more parsimonious (e.g., Epskamp
244 &Fried,2018). This network is recommended, for example,
245 in investigating organization of personality subtraits
246 (Constantini et al., 2015) and it was also used in previous
247 research investigating centrality of narcissism features
248 (Briganti & Linkowski, 2019; Di Pierro et al., 2019).
249 The structural organization of narcissism subscales was
250 explored by analyzing clustering of the nodes (local cluster-
251 ing) and the community structure of the network (global clus-
252 tering). Localized clustering of the nodes occurs when a
253 nodes neighbors tend to be directly connected to each other.
254 Anodethathasahighclusteringcoefficientcanbeseenasa
255 cohesive element in a well-connected group of neighboring
256 nodes. Degree of clustering is measured using a coefficient
257 proposed by Zhang and Horvath (2005). On the other hand,
258 communities are subsets of an entire set of network nodes
259 within which node-node edges are dense, but between which
260 the overall connectivity is weak (e.g., Newman, 2006).
261 Although different community detection methods and algo-
262 rithms have been proposed, the most prevalent one is the
263 walktrapalgorithm (e.g., Pons & Latapy, 2006).
264 Community detection was performed in R using igraph
265 package (Csárdi & Nepusz, 2006) and the algorithm is repeat-
266 ed 5000 times under different seed values. Communities were
267 estimated on the entire network without grouping of variables
268from the same scale. Modularity was used as a measure of
269how a community structure fits to the data of both networks.
270It indicates the probability that all nodes have the majority of
271their edges within their respective community. If a detected
272structure has low modularity, then it does not represent the
273structure of localized clustering of network edges.
274Modularity scores between 0.30 and 0.70 represent good fit
275for networks estimated from empirical data (Girvan &
276Newman, 2002).
277Two indices of centrality were calculated. The first is
278strength centrality which represents the sum of the edge
279weights per node translated to the sum of correlations of the
280node in respective networks. In reference to the entire net-
281work, nodes with high strength centrality are the nodes which
282are highly correlated with other nodes. In terms of the com-
283munity partition of the network, strength centrality serves as
284an indicator of node centrality within the community if the
285node shows moderate or high local clustering coefficient.
286This means that the node is highly correlated and that the
287majority of these correlations are with the nodes within the
288same community. Strength centrality is one of the most uti-
289lized centrality measures because its interpretation does not
290require strong assumptions (idea of distance or shortest path),
291which have been criticized in recent literature, along with
292demonstrated instability of the measures based on the con-
293cepts of betweenness and closeness due to sampling variabil-
294ity (Hallquist, Wright, & Molenaar, 2019; Bringmann et al.,
296The second centrality index is bridge-centrality which is
297based on a discovered community structure of the network.
298Nodes with high bridge-centrality play the role of a mediator
299or the middlemanconnecting nodes from different commu-
300nities. Bridge nodes bring all network nodes closer and create
301indirect paths between distant nodes and increase the cohesion
302of trait network conceived as a system of subscales. There are
303several different measures of how important a node is as a
304bridge between communities, but we focused on bridge ex-
305pected influence as a measure of centrality which indicates a
306nodes sum connectivity outside its own community while
307differentiating between positive and negative edges (Jones,
308Ma, & McNally, 2019). This measure is useful for networks
309that have both negative and positive edges (Robinaugh,
310Millner, & McNally, 2016) and it is based on strength central-
311ity. Bridge expected influence is not a measure estimated from
312the general topography of the network, but rather a measure
313indicating the existence of structural links between two or
314more network communities, which makes its interpretation
315restricted and straightforward. Bootstrap robustness tests (av-
316erages and confidence intervals based on non-parametric
317bootstrap and subsample stability based on case-drop
318boostrap) were performed for both types of centrality indices.
319Estimation of network, clustering, and centrality indices was
320performed using R package bootnet(Epskamp, Borsboom,
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
321 & Fried, 2018). Sample sizes greater than 250 are generally
322 sufficient for networks estimated from continuous variables
323 and moderate network size (e.g., 2535 nodes, see
324 Epskamp, 2016). Data and R code are available at https://
326 246b76dd19484c159e131e77d7562284.
327 Results
328 Results of the correlation analysis showed that there are only a
329 few negative zero-order correlations between the subscales
330 (red squares on Fig. 1), while the majority of the correlations
331 between the subscales were positive (blue squares). Most no-
332 table negative correlations were between need for admiration
333 and indifference, need for admiration and authoritativeness,
334 and shame and indifference, which is in line with the theoret-
335 ical assumptions (Glover, Miller, Lynam, Crego, & Widige,
336 2012).
337 The network based on zero-order correlations showed high
338 stability and robustness in terms of strength centrality (see Fig.
339 Cin Supplement). Entitlement/exploitativeness (15.06), gran-
340 diose exhibitionism (14.65), and leadership/authority from
341 NPI (14.58), as well as entitlement from GNS (14.22) had
342 the highest levels of strength centrality (see Table Ain
343 Supplement). However, low modularity was found for the
344 detected community structure (0.07). Although grandiose
345 and vulnerable narcissism features could be distinguished as
346 two communities (see Fig. Ain Supplement), low modularity
347 implies a weak community structure where nodes have many
348 strong edges with nodes from other communities. In general,
349 the networks based on zero-order correlations are less sparse
350 in comparison with the partial correlation networks. The pres-
351 ence of shared variance produces a multitude of zero-order
352 correlations which weaken the community structure of the
353 network as this type of variance distorts the structure of unique
354 associations between specific groups of nodes and results in
355 lower modularity score. Thus, bridge-centrality would not be
356 appropriate to interpret on this network.
357 Network based on regularized partial correlations showed
358 high stability and robustness in terms of both strength and
359 bridge expected influence centrality (see Fig. Bin
360 Supplement). Contingent self-esteem from PNI (1.46),
361 leadership/authority (1.33), and grandiose exhibitionism from
362 NPI (1.32) had the highest strength centrality (Fig. 4;see
363 Table Ain Supplement for numerical values). The community
364 structure of this network showed adequate modularity (0.34),
365 thus its interpretation is justifiable. Walktrap algorithm was
366 not affected by the value of random seed, giving identical
367 results in every iteration. Taken together, these results showed
368 that the detected community structure network could be con-
369 sidered as stableQ1 (Fig. 2).
370Walktrap community detection algorithm revealed the ex-
371istence of four communities. The first community consists of
372nodes from NPI, FFNI, and GNS, which refer to grandiose
373exhibitionism (Fig. 3). Based on a combination of high
374strength centrality and moderate/high clustering coefficient,
375the most central node in this community is grandiose exhibi-
376tionism from NPI (Fig. 4; see Table Ain Supplement for
377clustering coefficients). Thus, the majority of its edges are
378with nodes from the same community (high clustering) and
379the sum all of edge weights or partial correlations of this node
380is the highest within the community (strength centrality).
381Majority of the subscales from PNI, two subscales from
382GNS, and four from FFNI form the second community, refer-
383ring to vulnerable narcissism. The central node in this com-
384munity is contingent self-esteem. The third community is
385composed of nodes from FFNI and GNS and refers to
386exploitativeness, lack of empathy, and entitlement, thus it
387could serve as grandiose aspect of narcissistic antagonism.
388According to strength centrality, exploitativeness, entitlement,
389and supperiority from GNS and lack of empathy from FFNI
390have somewhat higher centrality than other nodes in this com-
391munity and among them entitlement has the highest clustering
392coefficient, thus this node could be seen as central. Finally, the
393fourth community is defined with a strong edge between ma-
394nipulativeness from FFNI and exploaitativess from PNI, as
395well as a triad consisting of authoritativeness, leadership/au-
396thority, and authority, among which autoritativeness stands
397out as the central node. Thus, this community mostly refers
398to authority but with elements of manipulation and exploita-
399tion. Finally, it could be seen that thrill seeking is very weakly
400connected to the rest of the community and it is connected to
401the rest of the nodes with weak edges only.
402According to the bridge expected influence centrality,
403entitlement/exploitativeness (0.81) and leadership/authority
404(0.6) from NPI and grandiose fantasy (0.61) from PNI are
405the most bridge-central nodes (Fig. 4). Entitlement/
406exploitativeness is the most central node and its bridge expect-
407ed influence is significantly higher compared to all other
408nodes, except for leadership/authority (see Fig. Ein
409Supplement). Entitlement/exploitativeness, followed by gran-
410diose fantasy, has the function in connecting the grandiose
411exhibitionism community and the vulnerable narcissism com-
412munity (the first and the second). On the other hand,
413leadership/authority has an important role in bridging together
414two grandiose narcissism communities, grandiose exhibition-
415ism and authority (the first and the fourth).
417The results of this study point to several main conclusions.
418First, community detection analysis showed that grandiose
419aspect of narcissism is presented as more complex, spreading
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
420 to three communities, while vulnerable aspect is detected as a
421 well-structured community. This result reflects the content of
422 the used instruments, as some of them measure only grandiose
423 narcissism (NPI and GNS) while other instruments measure
424 mainly grandiose narcissism by including more subscales that
425 constitute grandiose narcissism compared to vulnerable nar-
426 cissism (FFNI). Thus, although in PNI and FFNI vulnerable
427 narcissism is represented as a multifaceted dimension, it
428 seems that all features of vulnerable narcissism converge to
429 contingent self-esteem, which is in line with NSM (Krizan &
430 Herlache, 2018). This PNI subscale refers to fluctuating, un-
431 stable experience of self-esteem which, depending on the ex-
432 ternal sources, serves as the key feature of vulnerable
433narcissism. The centrality of contingent self-esteem is also
434recognized in a more recent network analysis of PNI, showing
435that contingent self-esteem and entitlement rage, along with
436grandiose fantasies, have the highest strength and closeness
437centrality (Di Pierro et al., 2019).
438Among PNI grandiose narcissism subscales, sacrificing
439self-enhancement was placed in the vulnerable narcissism
440community, while exploitativeness and grandiose fantasies
441belong to certain grandiose narcissism communities.
442Regarding some concerns that PNI measures only vulnerable
443narcissism (e.g., Miller et al., 2016), the results of this study
444showed that, among grandiose narcissism subscales, only
445sacrificing self-enhancement showed similarities with
Fig. 1 Zero-order correlation matrix plot. ACS = acclaim seeking,
ARR = arrogance, AU = authoritativeness, DI = cynicism/distrust,
ENT.FFNI = entitlement, EXH.FFNI = exhibitionism, EXP.FFNI =
exploitativeness, GF.FFNI = grandiose fantasies, IN = indifference,
LEM = lack of empathy, MAN = manipulativeness, NA = need for
admiration, RA = reactive anger, SH = shame, TS = thrill seeking,
CSE = contingent self-esteem, EXP.PNI = exploitativeness, SSSE = self-
sacrificing self-enhancement, HS = hiding the self, GF.PNI = grandiose
fantasy, DEV = devaluing, ER = entitlement rage, AUT = authority,
SUF = self-sufficiency, SUP = superiority, VAN = vanity, EXH.GNS =
exhibitionism, ENT.GNS = entitlement, EXP.GNS = exploitativeness,
LA = leadership/authority, GE = grandiose exhibitionism, EE =
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
446 vulnerable narcissism. This subscale refers to the use of pur-
447 portedly altruistic acts to support an inflated self-image.
448 Previous studies showed that it is positively related to Big
449 Five Agreeableness (e.g., Dinić&Vujić,2019b), thus it could
450 be seen as communion aspect of narcissism, which is however
451 manifested in submissive manner as well as vulnerable forms
452 of narcissism.
453 Second, the remaining three communities could be
454 interpreted as three aspects of grandiose narcissism: grandiose
455 exhibitionism, narcissistic antagonism, and authority. The dis-
456 tinction between the grandiose exhibitionism and authority
457 communities could be seen as distinction between more mal-
458 adaptive and more adaptive aspects of grandiose narcissism.
459 Previous studies showed that leadership/authority aspect of
460 narcissism is related to psychological adjustment and health,
461 while exhibitionism and entitlement/exploitativeness are
462 strongly related to psychopathological characteristics (e.g.,
463 Ackerman et al., 2011; Sedikides, Rudich, Gregg,
464 Kumashiro, & Rusbult, 2004; Stanton et al., 2017).
465 Results of this study showed that the central node in the
466 grandiose exhibitionism community is grandiose exhibition-
467 ism and the presence of entitlement/exploitativeness subscale
468highlights the maladaptive aspect of grandiose narcissism
469within this community. Namely, entitlement was character-
470ized as socially toxicaspect of narcissism due its relations
471with negative outcomes such are counterproductive school
472behavior and direct expression of hostile feelings and aggres-
473sion (see Ackerman et al., 2011). The centrality of overt
474(exhibitionism) manifestations of narcissistic grandiosity
475could be interpreted in light of Wrights(2016)notionthat
476the diagnosis of narcissistic personality disorder primarily fo-
477cuses on overt expressions of narcissistic grandiosity, com-
478pared to subtler forms, such are grandiose fantasies, although
479both forms are seen as central according to expertsrating
480(e.g., Ackerman et al., 2011).
481On the other hand, the authority community is the most
482diverse and captures both leadership/authority and
483manipulation/exploitation aspects. Previous studies on NPI
484showed that although leadership and manipulativeness corre-
485lated highly, leadership could be seen as adaptive and manip-
486ulativeness as maladaptive aspect of grandiose narcissism
487(Dinić&Vujić,2019a). The central node in this community
488is authoritativeness, which is conceptually close to leadership/
489authority which showed some positive outcomes as we
Fig. 2 Regularized partial correlation network
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
490 already mentioned (e.g., Ackerman et al., 2011). In addition,
491 authoritativeness was not related to Machiavellianism, psy-
492 chopathy, and externalizing behavior, showed significant neg-
493 ative correlation with depression, and among Five Factor traits
494 showed positive correlation only with Extraversion (Miller,
495 Gentle, & Campbell, 2012). Thus, we could conclude that this
496 community captures mainly adaptive features of narcissism.
497 The centrality of authority features is in line with a previous
498 study on NPI instrument, showing that authority, along with
499 entitlement and superiority, had the highest strength centrality
500 (Briganti & Linkowski, 2019). In addition, both exhibition
501 and authoritativeness were recognized as the main features
502 of grandiose narcissism in NSM (Krizan & Herlache, 2018).
503 The narcissistic antagonism community mainly comprises
504 FFNI antagonism subscales in grandiose narcissism along with
505 entitlement, exploitativeness, and superiority from GNS. The
506 detection of this community is in line with Miller et al. (2017)
507 trifurcated model in which the antagonism, as the core narcis-
508 sism element, is viewed as a broader constellation of antagonis-
509 tic features. According to Miller et al. (2017), antagonism is the
510 key component of both grandiose and vulnerable narcissism,
511but in the detected community it is more characterized by gran-
512diosity. This is in line with results of profile similarity that
513showed that antagonistic factor of narcissism is somewhat more
514similar to grandiose narcissism factor, compared to vulnerable
515narcissism factor (Crowe et al., 2019). The possible explanation
516could be that antagonism manifestations are more diverse in
517grandiose narcissism compared to vulnerable narcissism, which
518is reflected also by the fact that FFNI grandiose narcissism
519contains six subscales of antagonism, while vulnerable includes
520only two. There is no standout central node in this community
521but, based on a combination of strength centrality and clustering
522coefficient, entitlement from GNS could serve as the most cen-
523tral node, which is in line with NSM (Krizan & Herlache, 2018)
524and expertsrating of centrality features in narcissism (e.g.,
525Ackerman et al., 2017).
526Third, the two indices of centrality could offer an insight into
527the role of narcissism features. An important result is that there
528is no one central node, but rather several nodes, in line with
529heterogeneity of narcissism. In the network based on zero-order
530correlations, the features with the highest strength centrality
531include entitlement/exploitativeness, grandiose exhibitionism,
Fig. 3 Network communities in regularized partial correlation network. Legend: 1 = grandiose exhibitionism, 2 = vulnerable narcissism, 3 = grandiose
narcissistic antagonism, 4 = grandiose authority
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
Fig. 4 Sample (red) and bootstrap average (black) values and confidence
intervals (gray) of strength centrality (left) and bridge expected influence
(right) based on regularized partial correlation network. VAN = vanity,
TS = thrill seeking, SUP = superiority, SUF = self-sufficiency, SSSE =
self-sacrificing self-enhancement, SH = shame, RA = reactive anger,
NA = need for admiration, MAN = manipulativeness, LEM = lack of
empathy, LA = leadership/authority, IN = indifference, HS = hiding the
self, GF.PNI = grandiose fantasy, GF.FFNI = grandiose fantasies, GE =
grandiose exhibitionism, EXP.PNI = exploitativeness, EXP.GNS =
exploitativeness, EXP.FFNI = exploitativeness, EXH.GNS =
exhibitionism, EXH.FFNI = exhibitionism, ER = entitlement rage,
ENT.GNS = entitlement, ENT.FFNI = entitlement, EE = entitlement/
exploitativeness, DI = cynicism/distrust, DEV = devaluing, CSE =
contingent self-esteem, AUT = authority, AU = authoritativeness,
ARR = arrogance, ACS = acclaim seeking
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
532 leadership/authority, and entitlement. This result suggested that
533 entitlement features could be seen as the core element (Krizan
534 & Herlache, 2018), along with two grandiosity features,
535 reflecting the maladaptive and adaptive aspects of grandiosity
536 (e.g., Ackerman et al., 2011). However, the features of vulner-
537 able narcissism were missing, and these results are in line with
538 expertsfavoring grandiose features of narcissism as more cen-
539 tral (e.g., Ackerman et al., 2017). In the network based on
540 regularized partial correlation network, the centrality of the
541 same grandiose narcissism features (leadership/authority and
542 grandiose exhibitionism) is also recognized, along with contin-
543 gent self-esteem, highlighting the central aspect of vulnerable
544 narcissism. However, the features of entitlement or other antag-
545 onism aspects were missing. Thus, due to partialization, contin-
546 gent self-esteem emerged as central, meaning that this feature
547 has unique variance, not shared by the rest of the nodes in the
548 network. On the contrary, the missing central features of enti-
549 tlement could indicate that these features were highly connected
550 to other nodes, thus partialization affects their unique variance.
551 Regardless of these incongruent results, centrality indices on
552 both zero-order and partial correlations could offer a valuable
553 insight into the underlying structural organization of narcissism
554 features.
555 Furthermore, bridge-centrality highlighted the role of cer-
556 tain features in describing the narcissism core. Based on
557 bridge-centrality, entitlement/exploitativeness, leadership/au-
558 thority, and grandiose fantasy stand out as the central nodes
559 and their role is in connecting detected communities, i.e., dif-
560 ferent narcissism aspects. Entitlement/exploitativeness,
561 followed by grandiose fantasy, connects the grandiose exhibi-
562 tionism and vulnerable narcissism community. Although
563 overt (exhibitionism) manifestations of narcissism are seen
564 as strength-central within the grandiose exhibitionism com-
565 munity, high bridge-centrality of grandiose fantasy highlights
566 the role of more covert manifestation of narcissism in
567 connecting grandiose and vulnerable narcissism, which is in
568 line with expertsrating of central features of narcissism (e.g.,
569 Ackerman et al., 2017) and notion that vulnerable narcissism
570 comprises more subtle forms of narcissism (Wright, 2016).
571 Leadership/authority connects two grandiose narcissism com-
572 munities, grandiose exhibitionism and authority. Thus, the
573 results support the NSM structure (Krizan & Herlache,
574 2018) and the role of entitlement features as the common core
575 of grandiose and vulnerable narcissism. However, not all as-
576 pects of grandiose narcissism are connected through entitle-
577 ment features. Namely, authority/leadership aspect of grandi-
578 ose narcissism seems distant from the entitlement core,
579 supporting the view of this aspect of narcissism as more adap-
580 tive (e.g., Ackerman et al., 2011).
581 It should be noted that nodes from the narcissistic antago-
582 nism community were not listed among nodes with the highest
583 bridge-centrality. In this community, there are rather several
584 nodes with medium bridge-centrality, meaning that different
585nodes connect the narcissistic antagonism community with
586other communities. For example, the antagonism community
587is connected to the grandiose exhibitionism community main-
588ly through entitlement from GNS, and to the authority com-
589munity mainly through superiority.
590Fourth, the fact that subscales from NPI serve as the central
591features support the importance of this measure, despite some
592doubts about its validity (e.g., Rosenthal, Montoya, Ridings,
593Rieck, & Hooley, 2011). Previous research showed that NPI
594matched the expertsratings of narcissistic personality disor-
595der (e.g., Miller et al., 2014). The results of this research add to
596validity of NPI on the general and non-clinical sample, show-
597ing that some NPI subscales are very well-connected to other
598important features of narcissism in different communities.
599Exclusion of the mentioned NPI subscales would lead to a
600less cohesive narcissism network.
601Last, the results showed that thrill seeking has weak edges
602with the rest of the nodes in the narcissism network. Thrill
603seeking refers to a tendency to engage in risky behaviors for
604the sake of excitement. Previous studies showed mixed results
605on whether narcissism is more related to impulsivity (e.g.,
606Brunell & Buelow, 2018;Vazire&Funder,2006)orspecific
607impulsivity-related traits (e.g., sensation seeking, see Miller
608et al., 2009). Thus, future studies should further explore the
609status of thrill seeking in the narcissism network and model.
610There are some limitations of this study. First, it should be
611noted that there are other proposals of dimensionality of gran-
612diose narcissism, e.g., admiration and rivalry strategies (Back
613et al., 2013) or agentic and communal narcissism (e.g.,
614Gebauer, Sedikides, Verplanken, & Maio, 2012). Thus, the
615community structure can depend on specific instruments ap-
616plied in the analysis. Although there are other narcissism in-
617struments and models, in this study we focused on the widely
618used distinction between grandiose and vulnerable narcissism.
619Furthermore, previous results showed, for example, that ad-
620miration correlated highly with grandiosity (leadership/au-
621thority and grandiose exhibitionism) while rivalry correlated
622with both grandiosity (entitlement/exploitativeness) and vul-
623nerability (Back et al., 2013), thus it seems that inclusion of
624these strategies would not substantially change the explored
625narcissism network. However, grandiose narcissism is over-
626represented in narcissism assessment tools in general so future
627studies should include more aspects of vulnerable narcissism.
628Second, the stability of the networks was estimated using
629only bootstrap analysis, which does not resolve the issue
630concerning the influence of sampling variability and instabil-
631ity of the estimated network across different samples.
632Bootstrap estimated confidence intervals and significance test-
633ing should be treated with caution in regularized partial cor-
634relation network, which itself is a result of an automated mod-
635el selection procedure that limits the inference based on
636bootstrapped estimates. Therefore, the results of network anal-
637ysis are exploratory. However, the finding that there is no one
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
638 clear central node in the narcissism network, but rather several
639 ones, contributes to the robustness of the explored structure of
640 the narcissism network and since there is a clear community
641 structure with clear differentiation between the nodes of dif-
642 ferent communities, such network is more stable than a net-
643 work with one or two centrality nodes in the partial correla-
644 tional structure. Future study should consider Bayesian alter-
645 native (Bayesian Gaussian Graphical Models, see Williams &
646 Mulder, 2020) which allows for hypothesis testing based on
647 the assumptions about the structural features of the network.
648 However, application of these models into network approach
649 is still in the development phase.
650 To conclude, the results of the present study provide sup-
651 port for heterogeneity of grandiose narcissism, while vulner-
652 able narcissism could be seen as a more homogeneous dimen-
653 sion, at least in the used set of instruments. The central fea-
654 tures of vulnerable (contingent self-esteem) and both mal-
655 adaptive (grandiose exhibitionism) and adaptive (leadership/
656 authority, authoritativeness) aspects of narcissism were recog-
657 nized, along with entitlement as the central core of narcissistic
658 antagonism. Furthermore, entitlement/exploitativeness is the
659 main element that brings together maladaptive grandiose and
660 vulnerable aspects of narcissism, supporting the role of enti-
661 tlement features as the core of narcissism, while leadership/
662 authority brings together various aspects of grandiose narcis-
663 sism. The underlying organization of the explored narcissism
664 network reflects the complexity of narcissism and could offer
665 a promising support for development of the narcissism model.
666 Supplementary Information The online version contains supplementary
667 material available at
668 Funding This study was supported by the Ministry of Education,Science
669 and Technological Development of the Republic of Serbia [Grant
670 ON179006].
671 Data Availability
672 9484c159e131e77d7562284
673 Compliance with Ethical Standards
674 Conflict of Interest On behalf of all authors, the corresponding author
675 states that there is no conflict of interest.
676 Ethics Approval The research was approved by the by the Ethical
677 Committee of the Faculty of Philosophy, University of Novi Sad,
678 Serbia, which is the Second Instance Commission of the Ethical
679 Committee of the Serbian Psychological Society.
680Q2 References
681 Ackerman, R. A., Hands, A. J., Donnellan, M. B., Hopwood, C. J., &
682 Witt, E. A. (2017). Expertsviews regarding the conceptualization
683 of narcissism. Journal of Personality Disorders, 31(3), 346361.
685Ackerman, R. A., Witt, E. A., Donnellan, M. B., Trzesniewski, K. H.,
686Robins, R. W., & Kashy, D. A. (2011). What does the narcissistic
687personality inventory really measure? Assessment, 18(1), 6787.
689Back, M. D., Küfner, A. C. P., Dufner, M., Gerlach, T. M.,Rauthmann,J.
690F., & Denissen, J. J. A. (2013). Narcissistic admiration and rivalry:
691Disentangling the bright and dark sides of narcissism. Journal of
692Personality and Social Psychology, 105(6), 10131037. https://
694Bringmann, L. F., Elmer, T., Epskamp, S., Krause, R. W., Schoch, D.,
695Wichers, M., Wigman, J. T. W., & Snippe, E. (2019). What do
696centrality measures measure in psychological networks? Journal
697of Abnormal Psychology, 128(8), 892903.
699Brown, R. P., Budzek, K., & Tamborski, M. (2009). On the meaning and
700measure of narcissism. Personality and Social Psychology Bulletin,
70135(7), 951964.
702Brunell, A. B., & Buelow, M. T. (2018). Homogenous scales of narcis-
703sism: Using the psychological entitlement scale, interpersonal
704exploitativeness scale, and narcissistic grandiosity scale to study
705narcissism. Personality and Individual Differences, 123,182190.
707Costantini, G., Epskamp, S., Borsboom, D., Perugini, M., Mõttus, R.,
708Waldorp, L. J., & Cramer, A. O. J. (2015). State of the aRt person-
709ality research: A tutorial on network analysis of personality data in
710R. Journal of Research in Personality, 54,1329.
712Crowe, M. L., Lynam, D. R., Campbell, W. K., & Miller, J. D. (2019).
713Exploring the structure of narcissism: Toward an integrated solution.
714Journal of Personality. Online first., 87,11511169.
716Csárdi, G., & Nepusz, T. (2006). The igraph software package for com-
717plex network research. InterJournal, Complex Systems, 1695
718Retrieved from
720Di Pierro, R., Constantini, G., Benzi, I. M. A., Madeddu, F., & Preti, E.
721(2019). Grandiose and entitled, but still fragile: A network analysis
722of pathological narcissistic traits. Personality and Individual
723Differences, 140(1), 1520.
725Dinić, B. M., & Vujić, A. (2019a). Five-factor model best describes
726narcissistic personality inventory across different item response for-
727mats. Psychological Reports, 122(5), 19461966.
729Dinić,B.M.,&Vujić, A. (2019b). The pathological narcissism Inventory:
730Measurement invariance across Serbian and USA samples and fur-
731ther validation. European Journal of Personality Assessment.
732Online first.
733Epskamp, S. (2016). Regularized Gaussian psychological networks:
734Brief report on the performance of extended BIC model selection.
735Retrieved from
736Epskamp, S., Borsboom, D., & Fried, E. I. (2018). Estimating psycho-
737logical networks and their accuracy: A tutorial paper. Behavior
738Research Methods, 50(1), 195212.
740Epskamp, S., & Fried, E. I. (2018). A tutorial on regularized partial
741correlation networks. Psychological Methods, 23(4), 617634.
743Foster, J. D., McCain, J. L., Hibberts, M.F., Brunell, A. B., & Johnson, R.
744B. (2015). The grandiose narcissism scale: A global and facet-level
745measure of grandiose narcissism. Personality and Individual
746Differences, 73,1216.
747Gebauer, J. E., Sedikides, C., Verplanken, B., & Maio, G. R. (2012).
748Communal narcissism. Journal of Personality and Social
749Psychology, 103(5), 854878.
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
750 Girvan, M., & Newman, M. E. J. (2002). Community structure in social
751 and biological networks. Proceedings of the National Academy of
752 Sciences, 99(12), 78217826.
753 122653799.
754 Glover, N., Miller, J. D., Lynam, D. R., Crego, C., & Widige, T.
755 A. (2012). The five-factor narcissism inventory: A five-factor
756 measure of narcissistic personality traits. Journal of
757 Personality Assessment, 94(5), 500512.
758 1080/00223891.2012.670680.
759 Hallquist, M.N., Wright, A.G.C., & Molenaar, P.C.M. (2019).
760 Problems with centrality measures in psychopathology symp-
761 tom networks: Why network psychometrics cannot escape
762 psychometric theory. Multivariate Behavioral Research.
763 Online first. 125.
764 1640103.
765 Hendin, H. M., & Cheek, J. M. (1997). Assessing hypersensitive narcis-
766 sism: A reexamination of Murray's narcissism scale. Journal of
767 Research in Personality, 31(4), 588599.
768 jrpe.1997.2204.
769 Jones, P. J., Ma, R., & McNally, R. J. (2019). Bridge centrality: A net-
770 work approach to understanding comorbidity. Multivariate
771 Behavioral Research. Online first.
772 00273171.2019.1614898.
773 Krizan, Z., & Herlache, A. D. (2018). The narcissism Spectrum model: A
774 synthetic view of narcissistic personality. Personality and Social
775 Psychology Review, 22(1), 331.
776 1088868316685018.
777 Miller, B. K., Nicols, K. M., Clark, S., Daniels, A., & Grant, W. (2018).
778 Meta-analysis of coefficient alpha for scores on the narcissistic per-
779 sonality inventory. PLoS One, 13(12), e0208331.
780 1371/journal.pone.0208331.
781 Miller, J. D., Campbell, W. K., Young, D. L., Lakey, C. E., Reidy, D. E.,
782 Zeichner, A., & Goodie, A. S. (2009). Examining the relations
783 among narcissism impulsivity, and self-defeating behaviors.
784 Journal of Personality, 77(3), 761793.
785 1467-6494.2009.00564.x.
786 Miller, J. D., Gentile, B., & Campbell, W. K. (2012a). A test of the
787 construct validity of the five-factor narcissism inventory. Journal
788 of Personality Assessment, 95(4), 377387.
789 1080/00223891.2012.742903.
790 Miller, J. D., Lynam, D. R., & Campbell, W. K. (2016). Measures of
791 narcissism and their relations to DSM-5 pathological traits: A critical
792 reappraisal. Assessment, 23(1), 39.
793 1073191114522909.
794 Miller, J. D., Lynam, D. R., Hyatt, C. S., & Campbell, W. K. (2017).
795 Controversies in narcissism. Annual Review of Clinical Psychology,
796 13, 291315.
797 045244.
798 Miller, J. D., McCain, J., Lynam, D. R., Few, L. R., Gentile, B.,
799 MacKillop, J., & Campbell, W. K. (2014). A comparison of the
800 criterion validity of popular measures of narcissism and narcissistic
801 personality disorder via the use of expert ratings. Psychological
802 Assessment, 26(3), 958969.
803 Miller, J. D., Price, J., & Campbell, W. K. (2012b). Is the narcissistic
804 personality inventory still relevant? A test of independent grandios-
805 ity and entitlement scales in the assessment of narcissism.
806 Assessment, 19(1), 813.
807 1073191111429390.
808 Newman, M. E. J. (2006). Modularity and community structure in net-
809 works. Proceedings of the National Academy of Sciences, 103(23),
810 85778582.
811Newman, M. E. J. (2010). Networks: An introduction. Oxford, New
812York: Oxford University Press.
813Pincus, A. L., Ansell, E. B., Pimentel, C. A., Cain, N. M., Wright, A. G.
814C., & Levy, K. N. (2009). Initial construction and validation of the
815pathological narcissism inventory. Psychological Assessment,
81621(3), 365379.
817Pincus, A. L., & Lukowitsky, M. R. (2010). Pathological narcissism and
818narcissistic personality disorder. Annual Review of Clinical
819Psychology, 6, 421446.
821Pons, P., & Latapy, M. (2006). Computing communities in large net-
822works using random walks. Journal of Graph Algorithms and
823Applications, 10(2), 191218.
824Raskin, R., & Terry, H. (1988). A principal-components analysis
825of the narcissistic personality inventory and further evidence
826of its construct validity. Journal of Personality and Social
827Psychology, 54,890902.
829Robinaugh, D. J., Millner, A. J., & McNally, R. J. (2016). Identifying
830highly influential nodes in the complicated grief network. Journal of
831Abnormal Psychology, 125(6), 747757.
833Rosenthal, S. A., & Hooley, J. M. (2010). Narcissism assessment in
834socialpersonality research: Does the association between narcis-
835sism and psychological health result from a confound with self-
836esteem? Journal of Research in Personality, 44,453465. https://
838Rosenthal, S. A., Montoya, R. M., Ridings, L. E., Rieck, S. M., &
839Hooley, J.M. (2011). Further evidence of the narcissistic personality
840Inventorys validity problems: A meta-analytic investigation
841Response to Miller, maples, and Campbell (this issue). Journal of
842Research in Personality, 45(5), 408416.
844Sedikides, C., Rudich,E. A., Gregg, A.P., Kumashiro, M., & Rusbult, C.
845(2004). Are normal narcissists psychologically healthy?: Self-
846esteem matters. Journal of Personality and Social Psychology,
84787(3), 400416.
848Sherman, E. D., Miller, J. D., Few, L. R., Campbell, W. K., Widiger, T.
849A., Crego, C., & Lynam, D. R. (2015). Development of a short form
850of the five-factor narcissism inventory: The FFNI-SF. Psychological
851Assessment, 27(3), 11101116.
853Stanton, K., Daly, E., Stasik-OBrien, S. M., Ellickson-Larew, S., Clark,
854L. A., & Watson, D. (2017). An integrative analysis of the narcis-
855sistic personality inventory and the hypomanic personality scale:
856Implications for construct validity. Assessment, 24(6), 695711.
858Vazire, S., & Funder, D. C. (2006). Impulsivity and the self-defeating
859behavior of narcissists. Personality and Social Psychology Review,
86010(2), 154165.
861Vize, C. E., Collison, K. L., Miller, J. D., & Lynam, D. R. (2020). Using
862item-level analyses to better understand the consequences of
863partialing procedures: An example using the dark triad. Journal of
864Personality, 88(4), 719734.
865Williams, D. R., & Mulder, J. (2020). Bayesian hypothesis testing for
866Gaussian graphical models: Conditional independence and order
867constraints. Journal of Mathematical Psychology, 99, 102441.
869Wright, A. G., & Edershile, E. A. (2018). Issues resolved and unresolved
870in pathological narcissism. Current Opinion in Psychology, 21,74
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
872 Wright, A. G. C. (2016). On the measure and mismeasure of narcissism:
873 A response to measures of narcissim and their relations to DSM5
874 pathological traits: Acriticalreappraisal.Assessment, 23(1), 1017.
876 Wright, A. G. C., Lukowitsky, M. R., Pincus, A. L., & Conroy, D. E.
877 (2010). The higher order factor structure and gender invariance of
878 the pathological narcissism inventory. Assessment, 17(4), 467483.
880Zhang, B., & Horvath, S. (2005). A general framework forweighted gene
881coexpression network analysis. Statistical Applications in Genetics
882and Molecular Biology, 4(1).
884PublishersNoteSpringer Nature remains neutral with regard to jurisdic-
885tional claims in published maps and institutional affiliations.
Curr Psychol
JrnlID 12144_ArtID 1250_Proof# 1 - 05/12/2020
... On the other hand, grandiose and vulnerable narcissism positively correlate to psychological entitlement, more so than Machiavellianism and psychopathy (Miller et al., 2011;Turnipseed & Cohen, 2015). Recent network analyses also identified entitlement as the central feature of the narcissism network that connects grandiose and vulnerable narcissism (Dinić et al., 2021). Therefore, research suggests that entitled self-importance, as described by Krizan and Herlache (2018), is narcissism's core. ...
Full-text available
Narcissism as a psychological construct has had a contentious past both in its conceptualization and measurement. There is an emerging consensus that narcissism consists of grandiose and vulnerable subtypes, which share a common core. In the present research (N = 1002), we constructed a new measure of unified narcissism that reflects these contemporary understandings using items from the most widely used measures of grandiose and vulnerable narcissism: the Narcissistic Personality Inventory (NPI; Raskin & Terry, 1988,, and the Pathological Narcissism Inventory (PNI; Pincus et al., 2009, https://doi-org/10.1037/a0016530). We used classical test theory and item response theory approaches to devise a 29-item Unified Narcissism Scale. The scale showed good internal consistency, and convergent and discriminant validity, and showed evidence of measurement invariance between men and women. This research gave strong support for the structure, reliability, and validity of the unified measure, which offers a promising avenue for further enhancing our knowledge of narcissism.
... Regression analyses confirmed this pattern of associations, suggesting that vanity and self-sufficiency might protect against post-traumatic symptoms in some individuals, while entitlement would have a detrimental effect. Against this background, among all the grandiose narcissism dimensions, entitlement and exploitative-ness have been proposed to constitute the core of narcissism [32,63], and mainly reflect maladaptive narcissism and a general tendency toward antagonism [5][6][7][8]32,33]. ...
Full-text available
Background: Narcissism is characterized by entitlement, grandiose fantasies and the need for admiration. This personality trait has been associated with both traumatic experiences and emotional problems. Most studies have only focused on narcissism in the context of childhood trauma and negative emotional factors. However, dimensions of grandiose narcissism such as authority have been linked to adaptive outcomes. Furthermore, narcissism might not be linked only to negative childhood experiences; it may also be associated with the presence of post-traumatic symptoms. Therefore, the present study aimed to assess the associations between narcissism and the frequency and severity of post-traumatic symptoms and emotional factors (resilience capacity, emotional regulation, positive and negative affect, intolerance of uncertainty and perceived stress), as well as the possible mediational role of the latter in the relationship between narcissism and post-traumatic symptoms. Method: A total of 115 healthy young psychology undergraduates and their relatives, aged from 18 to 40 years, were asked to complete a set of questionnaires to evaluate the aforementioned variables. Results: The results showed that most of the grandiose narcissism dimensions were positively related to emotional adaptive outcomes, except exploitativeness and entitlement. The negative associations observed between the frequency and severity of post-traumatic symptoms and narcissism (self-sufficiency) were mediated by affect and resilience, which were in turn positively associated with the majority of the narcissism dimensions. Both positive affect and resilience were important factors mediating the association between grandiose narcissism and post-traumatic symptoms. Conclusions: Our findings reaffirm the need to assess not only desirable personality traits, but also ones that are not initially desirable, before pathologizing them. This consideration may be essential to achieve a personalized approach to the prevention of mental health problems, and promotion of positive emotions, in the general population.
... Matrica sklopa komponenti Petofaktorskog inventara makijavelizma Napomena: Prikazana su opterećenja ≥ .32 Kratka verzija Petofaktorskog inventara narcizma (Short Form of the Five-Factor Narcissism Inventory -FFNI-SF: Sherman et al., 2015, za adaptaciju na srpskom videtiDinić et al., 2021a) ...
Full-text available
Osnovni cilj ovog istraživanja je utvrđivanje relacija između crta Mračne tetrade i faktora aleksitimije, uzimajući u obzir multidimenzionalnu prirodu mračnih crta, kao i njihove antagonističke (maladaptivne) aspekte i aspekte delotvornosti ili agensne aspekte (adaptivnije aspekte). Dodatno, ispitana je medijatorska uloga faktora aleksitimije u odnosu između mračnih crta i distresa. Na uzorku od 355 ispitanika (71.3% ženskog pola) iz opšte populacije, primenjeni su sledeći instrumenti: Toronto skala aleksitimije, Skala depresije, anksioznosti i stresa, Levensonova skala psihopatije, Petofaktorski inventar makijavelizma, Kratka verzija Petofaktorskog inventara narcizma, Procena sadističke ličnosti. Rezultati pokazuju da mračne crte i faktori aleksitimije dele 56% zajedničke varijanse, te da su dominantni prediktori aleksitimije sekundarna psihopatija u pozitivnom smeru i makijavelistička delotvornost u negativnom smeru, a potom i narcistički antagonizam i primarna psihopatija, oba u pozitivnom smeru. Rezultati pokazuju da faktor mračnih crta koji okuplja agensne karakteristike na pozitivnom polu ostvaruje veći i negativni doprinos u objašnjenju aleksitimije i distresa, u odnosu na faktor koji okuplja primarno antagonističke karakteristike. Mračnim crtama je najviše objašnjen faktor aleksitimije koji se odnosi na probleme u identifikaciji emocija, i ujedno je ovaj faktor značajan medijator u predikciji distresa na osnovu sekundarne psihopatije, makijavelističke delotvornosti ili oba faktora mračnih crta. Naime, ovaj kognitivni deficit u razumevanju emocija doprinosi povećanju distresa u slučaju primarno antagonističkih mračnih crta, dok faktor delotvornosti preko manjih problema u identifikaciji emocija doprinosi redukciji distresa.
... These facets align with the grandiose and vulnerable aspects of the broad narcissism construct and were the characteristics which emerged homogenously across males and females and substantiated across samples. The two DSHS factors also corroborate the literature, which has contended that the phenotypic structure of narcissism is attributed to a core of entitlement which manifests in grandiosity and/or vulnerability (Ackerman et al., 2019;Dinić et al., 2021). ...
Full-text available
There has been an absence of consideration regarding measurement invariance across males and females in the widely available Dark Tetrad (DT) scales which measure psychopathy, Machiavellianism, narcissism and everyday sadism. This has resulted in criticisms of the measures, suggesting that the assessed constructs are not wholly relatable between the groups. This article documents the construction and validation of the Dark Side of Humanity Scale (DSHS), which measures dark personalities from an alternative viewpoint, determined by the constructs as they emerged from the male and female data, whilst aligning with theory and attaining invariance between sex. Across four samples (n = 2409), using a diverse range of statistical methods, including exploratory graph analysis, item response theory and confirmatory factor analysis, a divergence from the widely available DT measures emerged, whereby primary psychopathy and Machiavellianism were unified. This corroborated past research which had discussed the two constructs as being parallel. It further supported the DSHS with a shift away from the traditional DT conceptualisation. The resulting scale encompasses four factors which are sex invariant across samples and time. The first factor represents the successful psychopath, factor two addresses the grandiose form of entitlement, factor three taps into everyday sadism whilst the fourth factor pertains to narcissistic entitlement rage. Construct and external validity of the DSHS across two samples (n = 1338), as well as test-retest reliability (n = 413), was achieved. The DSHS provides an alternative approach to investigating the dark side of human nature, whilst also being sex invariant, thus making it highly suitable for use with mixed sex samples.
Childhood maltreatment and insecure attachment are associated with narcissism during adulthood. However, it is unclear how different types of maltreatment exert effects on specific features of narcissism and whether attachment security serves as a link between maltreatment and such features. We used network analyses to investigate the structure of narcissism and the complex interplay between different types of maltreatment, attachment security, and features of narcissism. A total of 718 participants were included in this study. The results revealed that (1) entitlement rage and contingent self-esteem were the most central features in the vulnerable narcissistic community, and leadership/authority was the most central feature in the grandiose narcissistic community; (2) childhood maltreatment was positively associated with features of vulnerable narcissism, whereas it had no or at most a weak negative correlation with features of grandiose narcissism, suggesting a different etiology for the two forms of narcissism; (3) emotional maltreatment had stronger associations with features of vulnerable narcissism compared with other types of maltreatment; and (4) attachment security served as a bridge node between emotional maltreatment and features of narcissism. Our findings suggest that therapists need to attend to maltreatment experienced by and the attachment security of individuals with vulnerable narcissism.
Full-text available
The aim of this research was to explore measurement invariance across samples from Serbia and the USA (Study 1) and to further validate the Serbian adaptation of the Pathological Narcissism Inventory – PNI (Study 2). The results supported the original seven-factor first-order structure as well as the hierarchical structure of the PNI with Narcissistic grandiosity and vulnerability as the second-order factors. Further, scalar invariance between the two versions of the PNI was achieved. Relations between Narcissistic grandiosity and vulnerability and other measures of grandiose and hypersensitive narcissism supported the validity of their scores. Among HEXACO traits, both Narcissistic grandiosity and vulnerability showed substantial negative correlations with Honesty-Humility. The main distinctions between the two aspects of narcissism lie in the positive relations with Neuroticism and negative relations with self-esteem, both of which are higher for Narcissistic vulnerability. The results support good psychometric properties of the PNI scores and add to the PNI’s cross-cultural validity.
Full-text available
Understanding patterns of symptom co-occurrence is one of the most difficult challenges in psychopathology research. Do symptoms co-occur because of a latent factor, or might they directly and causally influence one another? Motivated by such questions, there has been a surge of interest in network analyses that emphasize the putatively direct role symptoms play in influencing each other. In this critical paper, we highlight conceptual and statistical problems with using centrality measures in cross-sectional networks. In particular, common network analyses assume that there are no unmodeled latent variables that confound symptom co-occurrence. The traditions of clinical taxonomy and test development in psychometric theory, however, greatly increase the possibility that latent variables exist in symptom data. In simulations that include latent variables, we demonstrate that closeness and betweenness are vulnerable to spurious covariance among symptoms that connect subgraphs (e.g., diagnoses). We further show that strength is redundant with factor loading in several cases. Finally, if a symptom reflects multiple latent causes, centrality metrics reflect a weighted combination, undermining their interpretability in empirical data. Our results suggest that it is essential for network psychometric approaches to examine the evidence for latent variables prior to analyzing or interpreting patterns at the symptom level. Failing to do so risks identifying spurious relationships or failing to detect causally important effects. Altogether, we argue that centrality measures do not provide solid ground for understanding the structure of psychopathology when latent confounding exists.
Full-text available
Centrality indices are a popular tool to analyze structural aspects of psychological networks. As centrality indices were originally developed in the context of social networks, it is unclear to what extent these indices are suitable in a psychological network context. In this article we critically examine several issues with the use of the most popular centrality indices in psychological networks: degree, betweenness, and closeness centrality. We show that problems with centrality indices discussed in the social network literature also apply to the psychological networks. Assumptions underlying centrality indices, such as presence of a flow and shortest paths, may not correspond with a general theory of how psychological variables relate to one another. Furthermore, the assumptions of node distinctiveness and node exchangeability may not hold in psychological networks. We conclude that, for psychological networks, betweenness and closeness centrality seem especially unsuitable as measures of node importance. We therefore suggest three ways forward: (a) using centrality measures that are tailored to the psychological network context, (b) reconsidering existing measures of importance used in statistical models underlying psychological networks, and (c) discarding the concept of node centrality entirely. Foremost, we argue that one has to make explicit what one means when one states that a node is central, and what assumptions the centrality measure of choice entails, to make sure that there is a match between the process under study and the centrality measure that is used. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
Full-text available
Recently, researchers in clinical psychology have endeavored to create network models of the relationships between symptoms, both within and across mental disorders. Symptoms that connect two mental disorders are called "bridge symptoms." Unfortunately, no formal quantitative methods for identifying these bridge symptoms exist. Accordingly, we developed four network statistics to identify bridge symptoms: bridge strength, bridge betweenness, bridge closeness, and bridge expected influence. These statistics are nonspecific to the type of network estimated, making them potentially useful in individual-level psychometric networks, group-level psychometric networks, and networks outside the field of psychopathology such as social networks. We first tested the fidelity of our statistics in predicting bridge nodes in a series of simulations. Averaged across all conditions, the statistics achieved a sensitivity of 92.7% and a specificity of 84.9%. By simulating datasets of varying sample sizes, we tested the robustness of our statistics, confirming their suitability for network psychometrics. Furthermore, we simulated the contagion of one mental disorder to another, showing that deactivating bridge nodes prevents the spread of comorbidity (i.e., one disorder activating another). Eliminating nodes based on bridge statistics was more effective than eliminating nodes high on traditional centrality statistics in preventing comorbidity. Finally, we applied our algorithms to 18 group-level empirical comorbidity networks from published studies and discussed the implications of this analysis.
Full-text available
Objective Despite decades of work on narcissism there remain many active areas of exploration and debate including a clear and consensual description of its underlying components. Understanding narcissism's factor structure is necessary for precise measurement and investigation of specific psychological and behavioral processes. The aim of the current study was to explore the structure of narcissism by examining it at varying hierarchical levels. Method Participants recruited from Amazon's Mechanical Turk (N = 591) completed 303 narcissism items encompassing 46 narcissism scales and subscales. Criterion variables measuring the Five Factor Model, self‐esteem, aggression, and externalizing behavior were also collected. Results A series of factor analyses reveal the factor structure of narcissism at a range of specificities. No more than five meaningful factors (i.e., Grandiosity, Neuroticism, Antagonism, Distrustful Self‐reliance, Attention‐seeking) were identified and the most parsimonious model appears to be a three‐factor structure. Narcissism scales that effectively capture each of the identified factors are identified. Factors diverged in their associations with criterion variables. Conclusions A three‐factor model (i.e., Agentic Extraversion, Narcissistic Neuroticism, Self‐centered Antagonism) seems to be the most parsimonious conceptualization. Larger factor solutions are discussed, but future research will be necessary to determine the value of these increasingly narrow factors. This article is protected by copyright. All rights reserved.
Full-text available
The Narcissistic Personality Inventory (NPI) has greatly facilitated the scientific study of trait narcissism. However, there is great variability in the reported reliability of scores on the NPI. This study meta-analyzes coefficient alpha for scores on the NPI and its sub-scales (e.g. entitlement) with transformed alphas weighted by the inverse of the variance of alpha. Three coders evaluated 1213 individual studies for possible inclusion and determined that 1122 independent samples were suitable for coding on 12 different characteristics of the sample, scale, and study. A fourth author cross-coded 15 percent of these samples resulting in 85 percent overall agreement. In the independent samples, comprised of 195,038 self-reports, the expected population coefficient alpha for the NPI was .82. The population value for alpha on the various sub-scales ranged from .48 for narcissistic self-sufficiency to .76 for narcissistic leadership/authority. Because significant heterogeneity existed in coded study alphas for the overall NPI, moderator tests and an explanatory model were also conducted and reported. It was found that longer scales, the use of a Likert response scale as opposed to the original forced choice response format, higher mean scores and larger standard deviations on the scale, as well as the use of samples with a larger percentage of female respondents were all positively related to the expected population alpha for scores on the overall NPI. These results will likely aid researchers who are concerned with the reliability of scores on the NPI in their research on non-clinical subjects.
Full-text available
The objective of this research was to validate the Narcissistic Personality Inventory across different response formats, given that several factor structures were proposed, ranging from two to seven factors. The original forced-choice format of the Narcissistic Personality Inventory was given to 410 participants and a modified, i.e., Likert format was given to 423 participants from the general population, along with personality and other narcissism measures. The results showed that the five-factor model proposed by Ackerman et al. had the best model fit in both response formats and that a distinction between adaptive (Leadership, Vanity, and Superiority) and some aspects of maladaptive (Manipulativeness and Exhibitionism) narcissism factors could be established. However, the redundancy of items in certain factors could be problematic and further improvements of the Narcissistic Personality Inventory should include more indicators of some proposed factors, especially of Vanity.
Gaussian graphical models (GGM; partial correlation networks) have become increasingly popular in the social and behavioral sciences for studying conditional (in)dependencies between variables. In this work, we introduce exploratory and confirmatory Bayesian tests for partial correlations. For the former, we first extend the customary GGM formulation that focuses on conditional dependence to also consider the null hypothesis of conditional independence for each partial correlation. Here a novel testing strategy is introduced that can provide evidence for a null, negative, or positive effect. We then introduce a test for hypotheses with order constraints on partial correlations. This allows for testing theoretical and clinical expectations in GGMs. The novel matrix-F prior distribution is described that provides increased flexibility in specification compared to the Wishart prior. The methods are applied to PTSD symptoms. In several applications, we demonstrate how the exploratory and confirmatory approaches can work in tandem: hypotheses are formulated from an initial analysis and then tested in an independent dataset. The methodology is implemented in the R package BGGM.
Objective: Partialing procedures are frequently used in psychological research. The present study sought to further explore the consequences of partialing, focusing on the replicability of partialing-based results. Method: We used popular measures of the Dark Triad (DT; Machiavellianism, narcissism, and psychopathy) to explore the replicability of partialing procedures. We examined whether the residual content of popular DT scales is similar to the residual content of DT scales derived from separate samples based on relations with individual items from the IPIP-NEO-120, allowing for a finer-grained analysis of residual variable content. Results: Profiles were compared using three sample sizes (Small N = 156-157, Moderate N = 313-314, Large N = 627-628) randomly drawn from a large MTurk sample (N = 1,255). There was low convergence between original and residual DT scales within samples. Additionally, results showed that the content of residual Dirty Dozen scales was not similar across samples. Comparable results were found for short Dark Triad-Machiavellianism, but only in the moderate and small samples. Conclusion: The results indicate that there are important issues that arise when using partialing procedures, including replicability issues surrounding residual variables. Reasons for the observed results are discussed and further research examining the replicability of residual-based results is recommended.
We examined the extent to which the Psychological Entitlement Scale (PES), the Interpersonal Exploitativeness Scale (IES), and the Narcissistic Grandiosity Scale (NGS), when taken together, assess a broader construct or three distinct facets. In Study 1, a principal components analysis was conducted, demonstrating that the PES, IES, and NGS should be considered three separate traits rather than one overall construct. In Study 2, confirmatory factor analysis (CFA) revealed that the most efficient and best fitting model contained 8 items of the PES (dropping a reverse-scored item), the 6-item IES, and a revised 6-item model of the NGS. Study 3 replicated the CFA and examined the correlates of the PES, IES, and NGS with measures of narcissism and related measures such as empathic concern and self-esteem. Implications for future assessment of narcissism traits are discussed.