Measurement Invariance of the Pathological Narcissism Inventory and Multimethod Examination of Narcissistic Presentations in Community and Clinical Samples

Abstract

The Pathological Narcissism Inventory (PNI) is extensively used in recent empirical literature on pathological narcissism. However, most studies using the PNI are community-based, and no studies have used the PNI to investigate narcissistic presentations in personality disordered patients. This study investigates measurement invariance of the PNI in community participants and patients with personality disorders, and examines differences of narcissistic presentations in these samples through a multimethod approach. Results show that the PNI can be used reliably to measure and compare traits of pathological narcissism in community participants and patients with personality disorders. Personality disordered patients show higher traits reflecting vulnerable narcissism and overt manifestations of grandiose narcissism, compared with controls. Finally, network analysis indicates that traits of grandiose fantasies and entitlement rage have a central role in defining manifestations of PNI pathological narcissism, regardless of the presence of an underlying personality disorder. Research and clinical implications are discussed.

Keywords

pathological narcissism inventory measurement invariance network analysis personality disorders pathological narcissism

Pathological narcissism has suffered from a lack of consensual definition and exhaustive description of its grandiose and vulnerable manifestations for a long time (for a review, see Cain et al., 2008). As a consequence, empirical research resulted in a large number of studies whose findings conveyed only a partial and not completely appropriate understanding of pathological narcissistic functioning.

Pincus and colleagues (e.g., Cain et al., 2008; Pincus & Lukowitsky, 2010; Pincus et al., 2009) marked a turning point in the current understanding of pathological narcissism. By integrating clinical and empirical literature, the authors elaborated a comprehensive conceptualization of narcissism that clearly distinguished its adaptive and pathological aspects (Pincus & Lukowitsky, 2010). Accordingly, they developed a multidimensional self-report instrument for assessing pathological narcissism, namely the Pathological Narcissism Inventory (PNI; Pincus et al., 2009). Pincus and Lukowitsky (2010) indicated that narcissistic tendencies are common among individuals and involve needs for validation and affirmation. Individuals, however, may express their dispositional narcissistic tendencies through adaptive or pathological traits. Adaptive narcissism reflects the ability to cope with needs for validation and affirmation by maintaining a relatively positive self-image through adequate regulatory processes (Pincus & Lukowitsky, 2010). Conversely, pathological narcissism involves intense needs for validation and recognition in the context of impaired self-, field, and emotion regulation skills (Pincus & Lukowitsky, 2010). Furthermore, individuals may express pathological narcissism through grandiose and vulnerable manifestations. Grandiose manifestations include arrogance, a great sense of entitlement, fantasies of unlimited power, and exploitative attitudes, while feelings of emptiness, envy, hypersensitivity, and an avoidant interpersonal style typically describe vulnerable manifestations (Pincus et al., 2009).

Consistent with this taxonomy, the PNI comprises seven subscales (i.e., first-order factors) that are organized into two scales (i.e., higher-order factors), namely narcissistic grandiosity and narcissistic vulnerability (Wright et al., 2010). The PNI narcissistic grandiosity scale is described by dimensions of grandiose fantasy (i.e., compensatory fantasies of admiration and superiority over others), exploitativeness (i.e., manipulative interpersonal tendency), and self-sacrificing self-enhancement (i.e., proneness to engage in altruistic actions to sustain a grandiose self-image). The PNI vulnerability scale comprises four subscales: contingent self-esteem (i.e., proneness to dysregulate in the absence of external validation and admiration), entitlement rage (i.e., proneness to experience angry affects when others do not meet ones’ own expectations), hiding the self (i.e., unwillingness to show others one’s own flaws and failures), and the devaluing subscale (i.e., proneness to show disinterest in others who do not provide needed admiration, and to experience feelings of shame over needing recognition from disappointing others).

The PNI has a stable factorial structure and both grandiose and vulnerable pathological narcissism, as well as their specific traits, showed good reliability (e.g., Tritt et al., 2010; Wright et al., 2010). Empirical studies have also supported the criterion validity of both the PNI scales and subscales (e.g., Pincus et al., 2009; Thomas et al., 2012). Accordingly, the PNI scales and subscales have been used in empirical research (e.g., Di Sarno et al., 2020), even though some questions have been raised about the nature of the PNI grandiosity scale. More specifically, some authors (e.g., Miller et al., 2014) noted that the two PNI higher-order scales of narcissistic grandiosity and vulnerability covary strongly, and that the PNI grandiosity scale present associations with aspects of fragility (e.g., negative emotions) typically related to vulnerable narcissism. As pointed out by Edershile et al. (2019a), however, the associations of both the PNI scales of narcissistic grandiosity and vulnerability with measures of distress could be explained by their focus on pathological (i.e., maladaptive) manifestations of narcissism. Moreover, the authors demonstrated that the two PNI scales comprise unique aspects reflecting current conceptualizations of grandiose and vulnerable narcissism, as well as expected associations with psychological and behavioral correlates, when controlling for their shared variance.

Empirical Studies Using the PNI in Clinical Samples

The PNI has been translated into several languages (e.g., Italian, Chinese, and Serbian) over time, and it is one of the most widely used self-report measures of pathological narcissism all over the world. However, its use in clinical populations is still limited.

To have a clear picture of existing studies that have used the PNI in clinical samples, we conducted a systematic search using PsychInfo and PsychArticles databases, entering the keyword “Pathological Narcissism Inventory” and selecting English and peer-reviewed resources, published between 2009 and the date of our search (April 16, 2021). The search resulted in 668 resources, 55 of which were indexed in both PsychInfo and PsychArticles: Elimination of doubled resources led to a pool of 613 works (228 uniquely indexed in PsychInfo, 330 uniquely indexed in PsychArticles).¹ These works underwent a screening process to select only original studies that actually employed the PNI: We excluded original studies that did not administer the PNI, and other types of articles (e.g., reviews, meta-analyses, and theoretical contributions). The screening resulted in the elimination of 388 studies: The remaining 225 resources were original studies employing the PNI, its brief forms, or at least some of its subscales. These studies were subject to further in-depth exploration.

Inspection revealed that 198 studies out of 225 (88%) administered the PNI exclusively to nonclinical participants,² while only 27 studies (12%) involved clinical samples. These studies are displayed in Table 1.

Table 1.

Studies Administering the PNI to Clinical Samples.

Reference	Clinical participants	Specific PD diagnoses in the clinical sample^a	Nonclinical participants	PNI version	Mean differences in PNI
Boucher et al. (2018)	60 women evaluated at an Eating Disorders Unit	n/r	—	PNI	n/a
Brown et al. (2020)	700 MTurk participants (68.4% F) in psychiatric or psychotherapeutic treatment	n/r	527 undergraduates (71.2% F)	PNI	NG significantly higher in nonclinical; NV significantly higher in clinical
Dashineau et al. (2019)	288 participants (gender not reported) receiving treatment in mental health clinics in the past 2 years	n/r	—	PNI	n/a
Edershile et al. (2019a)	288 participants (67% F) recruited in mental health clinics	n/r	1635 undergraduates (66% F)	PNI	n/r
Ellison et al. (2013)	60 outpatients (85% F) in a university-based mental health center	22 (35.9%) with PD diagnosis:12 (19.7%) with BPD; 5 (8.2%) with NOS PD; 1 (1.6%) with NPD; 1 (1.6%) with APD; 1 (1.6%) with OCPD; 1 (1.6%) with HPD; 1 (1.6%) with STPD	—	PNI	n/a
Euler et al. (2018)	65 inpatients (76.9% F) from a specialized PD psychiatric unit	65 (100%) with BPD	—	PNI	n/a
Fossati et al. (2015)	103 outpatients (67% F) in a clinical psychology and psychotherapy hospital unit)	75 (72.8%) with PD diagnosis: 20 (19.4%) with NPD; 13 (12.6%) with BPD; 11 (10.7%) with HPD; 21 (20.4%) with mixed PDs; 10 (9.7%) n/r	394 community participants (55.3% F)	PNI	Clinical are higher on PNI total and NV
Fossati et al. (2017)	278 outpatients and inpatients (54% F) in a clinical psychology and psychotherapy hospital unit	n/r	—	PNI	n/a
Gamache et al. (2019)	106 outpatients (61.3% F) referring to a PD specialized clinic	106 (100%) with PD diagnosis: 30 (28.4%) with NPD; 24 (22.7%) with BPD-NPD; 19 (18.1%) with mixed or complex PDs; 13 (12.5%) with BPD; 9 (8.0%) with NOS PD; 6 (5.7%) with STPD; 4 (3.4%) with syndrome disorder diagnosis; 1 (1.1%) with HPD	—	B-PNI	n/a
Jaksic et al. (2017)	250 (61% F) outpatients in mental health hospitals	n/r	—	PNI	n/a
Kealy et al. (2012)	117 outpatients (73% F) in psychiatric treatment at mental health and substance use services	n/r	—	PNI	n/a
Kealy et al. (2016)	95 outpatients (68.4% F) in a public mental health clinic	n/r	—	PNI	n/a
Kealy et al. (2020)	99 outpatients (68.7% F) in a public mental health clinic	n/r	—	PNI	n/a
Marčinko et al. (2014)	234 psychiatric outpatients (42.7% F) in mental health units (	n/r	—	PNI	n/a
Miller et al. (2013)	98 (73.4% F) in psychiatric or psychotherapeutic treatment	31 (31.6%) with PD diagnosis: 7 (7.1%) with NPD; 5 (5.1%) with APD; 5 (5.1%) with OCPD (15%); 4 (4.1%) with BPD; 10 (10.2%) n/r	287 MTurk participants (66% F)	PNI	n/r
Miller et al. (2014)	98 (73.5% F) participants in psychiatric or psychotherapeutic treatment	n/r	214 undergraduates (67.1% F)	PNI	n/r
Morf et al. (2017)	214 inpatients (69.5% F) at a psychosomatic hospital	31 (14.5%) with PD diagnosis (specific PDs n/r)	1,837 undergraduates (67.5% F)	PNI	Clinical are higher in PNI total and subscales, except GF (no difference) and EXP (>in nonclinical)
Oltmanns and Widiger (2018)	241 MTurk participants (71% F) in current/past psychiatric or psychotherapeutic treatment	22 (9%) self-report a PD (specific PDs n/r)	—	PNI	n/a
Pincus et al. (2009)	26 outpatients (76.9% F) in a psychotherapy clinic	18 (68%) with PD diagnosis (one or more): 7 (26.9%) with BPD, 4 (15.4%) with NPD, 4 (15.4%) with HPD, 4 (15.4%) with OCPD, 2 (7.7%) with DPD, 2 (7.7%) with NOS PD, 1 (3.8%) with SZPD, 1 (7.7%) with APD	2 undergraduate samples (N₁ = 820, 73.9% F; N₂ = 399, 50.1% F)	PNI	n/r
Restek-Petrović et al. (2015)	48 patients (52.1% F) diagnosed with psychosis	n/r	—	PNI	n/a
Rogier and Velotti (2018)	74 patients (15.1% F) with Gambling Disorder	n/r	105 community participants (23.8% F)	PNI	Gamblers are higher on NG and NV, SSSE, HS, and DEV
Schoenleber et al. (2015)	60 outpatients (85.2% F) in a psychotherapy clinic	n/r	3 community and undergraduate samples (N₁ = 2862, 57.3% F; N₂ = 348, 73.9% F; N₃ = 581, 47.5% F)	PNI	n/r
Sherman et al. (2015)	98 participants (74% F) in psychiatric or psychotherapeutic treatment	41 (42%) with PD diagnosis (specific PDs n/r)	166 undergraduates (66% F); 287 MTurk participants (66% F)	PNI	n/r
Somma et al. (2018)	310 outpatients (53.5% F) in a clinical psychology and psychotherapy hospital unit	Total number of PD diagnoses n/r: 64 (20.6%) with NPD; 61 (19.7%) with PD with other specification; 36 (11.6%) with BPD; 15 (4.8%) with DPD; 14 (4.5%) with HPD	—	PNI	n/a
Stanton et al. (2019)	737 MTurk participants (67.7% F) in psychiatric or psychotherapeutic treatment	n/r	—	PNI	n/a
Stanton et al. (2018)	617 MTurk participants (68.6% F) in psychiatric or psychotherapeutic treatment	n/r	4911 MTurk participants (gender not available)	PNI	n/r
Tonna et al. (2018)	91 outpatients (65.9% F) referring to a PD hospital unit	n/r	—	PNI	n/a

Note. PNI = pathological narcissism inventory; PD = personality disorders; NG = narcissistic grandiosity; NV = narcissistic vulnerability; n/a = not applicable; n/r = not reported; MTurk = amazon mechanical turk; BPD = borderline personality disorder; NOS PD = not otherwise specified personality disorder; NPD = narcissistic personality disorder; APD = avoidant personality disorder; OCPD = obsessive–compulsive personality disorder; HPD = histrionic personality disorder; STPD = schizotypal personality disorder; (B-)PNI = (brief) pathological narcissism inventory; GF = grandiose fantasy subscale; EXP = exploitativeness subscale; DPD = dependent personality disorder; SZPD = schizoid personality disorder; SSSE = self-sacrificing self-enhancement subscale; HS = hiding the self subscale; DEV = devaluing subscale; F = females.

Percentages are computed with reference to the total number of clinical participants.

As shown in Table 1, only 4 out of 27 studies examined mean differences in PNI traits, comparing clinical and nonclinical participants. They indicate that people suffering from psychopathological conditions are higher than controls on PNI traits, especially vulnerable traits of pathological narcissism. These findings, however, need to be interpreted with extreme caution. In fact, no studies have investigated measurement invariance of the PNI across clinical and nonclinical samples so far (see Supplementary Table S1). Measurement invariance explores whether a measure is comparable across groups (MacCallum & Austin, 2000): It examines whether the factor structure is equal in such groups, and whether individuals belonging to different groups, and having the same position on a specific construct, show identical scores (Schmitt & Kuljanin, 2008). Therefore, measurement invariance is mandatory to studying mean differences across groups and to ensure that observed mean differences are due to genuine differences in the latent constructs, rather than to uneven relations between latent constructs and scores (Hirschfeld & Von Brachel, 2014; Jiang et al., 2017).

As shown in Supplementary Table S1, some studies (N = 6, 2.6%) examined measurement invariance of the PNI by focusing on gender (Jakšić et al., 2014; Schoenleber et al., 2015 for the PNI—Brief Form; Wright et al., 2010), cross-cultural (Dinić & Vujić, 2020; Wetzel et al., 2020 for the PNI—Brief Form), and age invariance (Somma et al., 2020).

Overall, these studies suggest that the PNI can be used reliably in both men and women, participants from different cultures, and people having different ages. However, we do not know whether the PNI scales reflect the same constructs in people with and without psychopathology.³

In summary, PNI-based empirical literature has been focused mainly on nonclinical samples, while only few studies involved clinical samples. As a consequence, further studies are needed to understand how the PNI works in clinical samples and what differences in PNI manifestations can be detected in people with and without psychopathological conditions.

Investigation of Pathological Narcissistic Manifestations: The Network Analysis Approach

As stated earlier, past studies explored differences in manifestations of PNI pathological narcissism between clinical participants and controls by comparing mean levels of PNI traits.

Recently, some authors (Borsboom, 2017) have pointed out that a different approach, namely network analysis, may be helpful to explore similarities and differences in psychopathological manifestations among different groups. Network analysis is increasingly gaining attention in the context of mental health research (for a review, see Contreras et al., 2019). This approach allows researchers to investigate the structure of psychopathologies by considering them as networks of symptoms and traits that interact with each other (e.g., Costantini et al., 2019). Specifically, network analysis may clarify which role symptoms and/or traits play in defining a specific phenomenon. Therefore, an inspection of traits-interconnections may provide valuable insights into how presentations of a psychopathological condition may differ across groups.

Di Pierro et al. (2019) used network analysis to investigate PNI presentations in a large community sample. The authors found that pathological narcissism was organized around core features related to both narcissistic grandiosity (i.e., grandiose fantasy) and vulnerability (i.e., contingent self-esteem and entitlement rage). Furthermore, the authors compared networks of participants with low versus high levels of personality organization to detect similarities and differences in narcissistic presentations. According to Kernberg’s object relations theory (Kernberg & Caligor, 2005), high levels of personality organization detail people showing identity integration, mature defense mechanisms, and preserved reality testing, while low levels of personality organization describe people showing difficulties in distinguishing inner and outer reality, employing primitive defenses (e.g., splitting, projection, and denial), and having poorly integrated identity. Results showed that grandiose fantasies, contingent self-esteem, and entitlement rage maintained a central role in narcissistic presentations across the levels of personality organization. Even the nature of traits-interconnections was very similar: A visual inspection revealed only slight differences in traits-interconnections suggesting more severe presentations in people with low levels of personality organization. Since low levels of personality organization are peculiar of people suffering from personality pathology (Di Pierro et al., 2016; Preti et al., 2015), the authors suggested that similar results might be found in patients with personality disorders. They also indicated, however, that such hypotheses needed to be tested empirically.

Another study (Dinić et al., 2021) has recently explored the structural organization of narcissism network, by administering four well-known measures of narcissism: the PNI, the Narcissism Personality Inventory (NPI; Raskin & Terry, 1988), the Grandiose Narcissism Scale (GNS; Foster et al., 2015), and the Five-Factor Narcissism Inventory–Short Form (FFNI; Sherman et al., 2015). Findings confirmed that core features of narcissistic presentations include both grandiose and vulnerable traits (Di Pierro et al., 2019). Furthermore, the authors showed that NPI entitlement and PNI grandiose fantasy have a “bridge” function in narcissism network, that is they are primarily responsible for the co-existence of both grandiose and vulnerable expressions of narcissism within individuals. As in Di Pierro et al. (2019), however, these findings were based on a community sample only.

Aims of the Study

This study investigates whether presentations of pathological narcissism, as assessed by the PNI, differ across community and clinical participants, namely personality disorder (PD) patients. In particular, our main aim was providing a comprehensive examination of the similarities and differences between PD patients and community participants in manifestations of pathological narcissism, both in terms of the average level of single traits (i.e., mean comparisons among groups) and in terms of relationships among traits (i.e., network analysis). The PNI has shown to assess reliably both the two phenotypic forms of pathological narcissism (Wright et al., 2010) and their specific manifestations (i.e., traits; Pincus et al., 2009). Empirical findings, however, suggest that PNI facet-level analyses can be particularly appropriate when a certain level of detail is desirable (e.g., Di Sarno et al., 2020). Thereby, we focused our investigation on the seven PNI traits, rather than considering the two PNI second-order scales.

The PNI measures pathological personality traits (i.e., pathological narcissistic traits) and it is well known that pathological narcissism may complicate treatment of co-occurring PDs (Caligor et al., 2015). Consequently, a better understanding of pathological narcissistic presentations in PD patients is particularly relevant from a clinical perspective. Therefore, we considered that a clinical sample of patients with proper personality pathology, as opposed to patients with any diagnosis, would be particularly suitable for our scope.

More specifically, the study has three main objectives. First, we tested measurement invariance across the two groups. No study has yet investigated whether the PNI can be reliably used to compare presentations of pathological narcissism in PD patients and community participants. Since the PNI is designed to be a dimensional measure of pathological narcissism, and dimensional trait approaches imply continuity between normal and pathological personality (Widiger, 1992) so that traits should be distributed to varying degrees in the general population, we expect the PNI to meet measurement invariance across the two groups.

Second, we tested differences in the levels of PNI traits between PD patients and community people. Past findings suggest that people with psychopathological conditions usually score higher in PNI traits of vulnerable narcissism than controls (Fossati et al., 2015; Karakoula & Triliva, 2016; Miller et al., 2017; Morf et al., 2017; Rogier & Velotti, 2018). Moreover, patients with narcissistic personality disorder usually seek help when they are in vulnerable narcissistic states (Ronningstam, 2014). Therefore, we hypothesize that PD patients would report higher PNI traits reflecting vulnerable manifestations than controls, while we expect no significant differences in traits defining grandiose narcissistic manifestations.

Third, we used network analysis to detect similarities and differences in PNI narcissistic presentations between PD patients and community people. We expect that PD patients and controls would show differences in the way PNI traits connect with each other, yet maintaining a stable core. In line with Di Pierro et al. (2019), we expect to find grandiose fantasies, contingent self-esteem, and entitlement rage to have a central role in defining manifestations of pathological narcissism in both PD patients and community participants. In other words, we expect that presentations of pathological narcissism would be organized around these traits, regardless of the existence of an underlying personality disorder. In line with clinical (Ronningstam, 2014) and empirical (Miller et al., 2017) observations, however, individuals suffering from PDs are expected to show stronger and larger interplays among traits of vulnerable narcissism. Finally, we expect that traits related to entitlement and grandiose fantasy would be primarily responsible for the co-existence of grandiose and vulnerable expressions of PNI pathological narcissism in community people, as Dinić et al. (2021) showed. Conversely, we cannot formulate specific hypotheses about bridge traits in PNI narcissistic presentations of PD patients because of the lack of past studies.

Method

Participants and Procedure

The study involved both PD patients and adult community participants. The community sample included 1311 participants (females: N = 847; 64.6%) with a mean age of 30.07 (range: 18–77; SD = 10.90). Community participants were involved in the study through posts and announcements on social networks. All participants signed informed consent prior to completing the online survey. Most participants were employed (45.1%; N = 591), 39.4% were students (N = 517), and 60 participants were unemployed (4.6%). Average years of education was 14.49 (range: 5–23; SD = 3.22). Due to technical problems, data on level of education and occupational status of 143 participants (10.9%) were not recorded. None of the participants reported being in psychotherapy or psychiatric treatment at the time of participation in the study.

The clinical PD sample (N = 153) was recruited among patients seeking treatment at mental health centers specialized in pathological addiction in the area of Milan (N = 56) and in outpatient psychiatric services of Parma Local Health Agency (N = 97). Inclusion criterion was endorsing at least one PD diagnosis at the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association [APA], 1994) Axis II disorders (First et al., 1997). Exclusion criteria were (a) being less than 18 years old, (b) cognitive impairment, (c) meeting criteria for a current psychotic disorder, and (d) reporting current use of illicit substances. The clinical sample included PD patients (females: N = 86; males: N = 66; other: N = 1) with a mean age of 40.82 years (range: 18–70; SD = 12.39). Most PD patients were employed (50.4%; N = 77), 36% participants were unemployed (N =55), and 13.1% were students (N = 20). Average years of education was 12.61 (range: 5–23; SD = 3.74). The majority of patients had one PD diagnosis only (N = 82; 53.6%). The most frequent diagnoses were Borderline Personality Disorder (N = 67; 43.8%), Narcissistic Personality Disorder (N = 53; 34.6%), and Obsessive-Compulsive Personality Disorder (N = 25; 16.3%). The majority of PD patients also had Axis I disorders (N = 110; 71.9%), as assessed by means of the Structured Clinical Interview for DSM-IV Axis I disorders (First et al., 1997). The most frequent diagnoses were Adjustment disorder (N = 34; 22.2%) and Mood disorders (N = 28; 18.3%). Some patients (N = 31; 20.3%) met criteria for Substance Use Disorders in the past year, but they had not been using illicit substances for the last 6 months. The clinical sample was not significantly different from community participants in the distribution of male and female, χ²(1) = 3.46, p = .063, but participants in the community sample were significantly younger than controls, t(1463) = −11.38, p < .001.

For the community sample and a portion of the clinical one (N = 56), PNI data were collected as part of a larger study that has been approved by the Ethical Committee of the University of Milano-Bicocca (protocol number: 0016178/13). PNI data of the remaining clinical sample (N = 97) were collected as part of another larger study which has been approved by the Ethical Committee of the University of Parma (protocol number 0037436).⁴

Measures

PNI (Pincus et al., 2009; Wright et al., 2010): The PNI is a 52-item self-report measure of pathological narcissistic traits. Items are rated on a 6-point Likert-type scale ranging from 0 (not at all like me) to 5 (very much like me). The PNI assesses seven first-order factors (subscales) and two second-order factors (scales). The PNI grandiosity scale comprises three subscales: grandiose fantasy, exploitativeness, and self-sacrificing self-enhancement. The PNI vulnerability scale includes four subscales: contingent self-esteem, hiding the self, devaluing, and entitlement rage. For the purpose of this study, we considered only the seven PNI subscales, which proved to be sufficiently reliable in our sample (range α = .68–.89).

Data Analytic Plan

First, we used confirmatory factor analyses (CFA) to test measurement invariance of the PNI across the two groups (PD patients and community participants). We examined configural, weak, and strong measurement invariance by performing a series of model comparisons (Hirschfeld & Von Brachel, 2014; Putnick & Bornstein, 2016). More specifically, we first fit a CFA model with seven correlated factors, reflecting the seven PNI subscales, in both groups (configural invariance); second, we constrained loadings to be equal across groups (weak invariance); third, we also constrained intercepts to be equal (strong invariance). To evaluate overall model fit, we considered values of comparative fit index (CFI) > .95, root mean square error of approximation (RMSEA) < .06 and standardized root mean square residual (SRMR) < .08 as indicating good fit (Hu & Bentler, 1999). To evaluate measurement invariance, we considered values of ΔCFI < .01, ΔRMSEA < .015, and ΔSRMR < .03 as indicating sufficient similarity between models for testing weak invariance, and values of ΔCFI < .01, ΔRMSEA < .015, and ΔSRMR < .01 for testing strong invariance (Chen, 2007; Sass, 2011).

Second, after establishing measurement invariance, we tested standardized differences of latent means between patients and controls (Sass, 2011). The CFA analysis was performed using R packages lavaan (Rosseel, 2012) and semTools (Jorgensen et al., 2021).

Third, we explored the connectivity between the seven PNI narcissistic traits using network analysis. In network models, nodes represent constructs of interest (in this case, narcissistic traits) and edges represent their pairwise relationships. We estimated regularized Gaussian Graphical Model networks for patients and controls (Costantini et al., 2019; Epskamp, Waldorp, et al., 2018) using Fused Graphical Lasso (Costantini et al., 2019; Danaher et al., 2014). In Gaussian Graphical Models, edges represent regularized estimates of partial correlations between any two nodes (e.g., Costantini et al., 2019). This method yields two networks, one for patients and one for controls, which are estimated jointly. Compared with separate network estimates, Fused Graphical Lasso improves edge estimates by exploiting similarities between samples. However, if the true networks are different and exploiting similarities does not improve model fit, this method becomes close to estimating networks independently, therefore allowing true differences to emerge (Costantini et al., 2019; Di Pierro et al., 2019; Fried et al., 2018). We also used the network comparison test (NCT), as implemented in the NetworkComparisonTest R package (van Borkulo et al., 2022), to investigate whether the network structure and global strength differed significantly between patients and controls. In particular, the network structure test examines whether the largest difference between two corresponding edges across groups (M) is significantly different from zero, whereas the global strength invariance test examines whether there is a significant difference in overall edge density (S).

We further analyzed the resulting networks in terms of fundamental properties of each node. In particular, we examined centrality indices that are more appropriate for psychopathology networks (Bringmann et al., 2019; Jones et al., 2021; Robinaugh et al., 2016).

Expected influence (EI) measures the overall connectedness of a node to the rest of the network. More specifically, EI accounts for the presence of both positive and negative edges in networks, by summing weights of the positive edges and subtracting those of the negative ones. For this reason, EI is particularly suited for psychopathology networks: In fact, negative edges can be clinically relevant, indicating potential suppressive effects among symptoms or maladaptive traits (Robinaugh et al., 2016). Bridge Expected Influence (bridge EI) is akin to EI, but focuses exclusively on the connectivity between different groups of nodes included in the same network (Jones et al., 2021). A single network can indeed include more than one group of nodes (e.g., network of general psychopathological symptoms, with groups of nodes for specific disorders). Bridge EI measures the connectedness of a node belonging to a group with nodes that are part of another group. Hence, nodes with high bridge EI can be considered primarily responsible for comorbidity between disorders (Jones et al., 2021). As to networks that include traits rather than symptoms, such as traits of pathological narcissism, bridge EI depends on connections between nodes of different groups of traits (i.e., vulnerable and grandiose ones). In particular, we calculated bridge EI to examine if any PNI trait of narcissistic grandiosity (i.e., grandiose fantasy [GF], exploitativeness [EXP], or self-sacrificing self-enhancement [SSSE]) is more strongly connected with PNI traits of narcissistic vulnerability (i.e., contingent self-esteem [CSE], hiding the self [HS], devaluing [DEV], and entitlement rage [ER]), and vice versa.

We also inspected the stability of EI and bridge EI indices using the Correlation Stability (CS) coefficient (Epskamp, Borsboom, & Fried, 2018). The CS-coefficient represents the proportion of participants that can be excluded from the sample in case-dropping bootstrap resamples, such that the resulting EI indices have a 95% probability to correlate ≥ 0.7 with the original EI index. As a rule of thumb, a CS-coefficient below 0.25 indicates insufficient stability and warns against interpreting the centrality indices. A CS-coefficient above 0.50 indicates good stability (Epskamp, Borsboom, & Fried, 2018). Since our analysis involved two groups, we computed the CS stability separately by group and considered the lowest value in the two groups as a conservative estimate of CS coefficient (Fried et al., 2018).

Results

Twenty seven missing values in the PNI (out of 76128 overall item responses collected) were imputed using the predictive mean matching algorithm, as implemented in the R package mice (van Buuren & Groothuis-Oudshoorn, 2011).

Measurement Invariance and Latent Mean Comparisons of PNI Traits Netween PDs and Controls

The baseline RMSEA was smaller than .158 both in the overall sample (baseline RMSEA = .120), and in separate models fit in the control (RMSEA = .118) and patient (RMSEA = .136) groups. Following recommendations by Kenny (2020; see also Snyder et al., 2021), we did not use CFI to assess absolute model fit, but only to compare nested models. The configural invariance model fit the data reasonably well according to the RMSEA and SRMR (Hu & Bentler, 1999; Kenny, 2020). Model comparisons reported in Table 2 supported strong measurement invariance. Factor loadings and intercepts are in Supplementary Material (Table S2).⁵

Table 2.

Measurement Invariance of PNI in Patients and Controls.

Model	χ²	df	p-value	RMSEA	SRMR	CFI	AIC	BIC
1. Configural inv.	8849.67	2506	<.001	0.059	0.063	.772	245,565.06	247,437.34
2. Weak inv.	8917.52	2551	<.001	0.058	0.064	.771	245,542.91	247,177.19
Δ2-1	67.85	45	.0154	<0.001	0.001	.001	–22.15	–260.15
3. Strong inv.	9165.90	2596	<.001	0.059	0.065	.764	245,701.30	247,097.57
Δ3-2	248.39	45	<.001	<0.001	0.001	.007	158.39	–79.613

Note. PNI = pathological narcissism inventory; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual; CFI = comparative fit index; AIC = Akaike information criterion; BIC = Bayesian information criterion.

Relying on the strong measurement invariance model, we estimated the standardized differences in the latent means of each of the seven first-order factors of the PNI model in PD patients and controls. The results showed that PD patients presented higher CSE (Δ = .604 SD, p < .001), DEV (Δ = .587 SD, p < .001), SSSE (Δ = .562 SD, p < .001), ER (Δ = .289 SD, p = .015), and HS (Δ = .172 SD, p = .050) than controls. No significant differences emerged for GF (Δ = −.017 SD, p = .878) and EXP (Δ = −.206 SD, p = .099).

Mean values and standard deviations of observed PNI variables and the PNI scale intercorrelations are reported in Table 3.

Table 3.

Mean Values, Standard Deviations, and PNI Scale Intercorrelations in the Community (Above the Diagonal) and Clinical Samples (Below the Diagonal).

		1	2	3	4	5	6	7	Community sample		Clinical sample
	PNI scale	1	2	3	4	5	6	7	M	SD	M	SD
1	EXP	_	.17***	.32***	.05*	.19***	.17***	.26***	2.12	.83	1.9	1.05
2	SSSE	.29***	_	.45***	.37***	.25***	.32***	.43***	2.79	.83	3.23	.96
3	GF	.31***	.41***	_	.53***	.43***	.41***	.55***	2.5	1.01	2.5	1.19
4	CSE	.12	.46***	.55***	_	.38***	.58***	.58***	2.08	.95	2.66	1.1
5	HS	.30***	.34***	.46***	.44***	_	.49***	.37***	2.67	.88	2.82	1.03
6	DEV	.26***	.36***	.47***	.62***	.65***	_	.53***	1.85	.95	2.34	1.12
7	ER	.36***	.47***	.62***	.62***	.40***	.59***	_	2.3	.87	2.54	1.12

Note. PNI = pathological narcissism inventory; M = mean; SD = standard deviation; EXP = exploitativeness subscale; SSSE = self-sacrificing self-enhancement subscale; GF = grandiose fantasy subscale; CSE = case of experiences of self-esteem dysregulation; HS = hiding the self subscale; DEV = devaluing subscale; ER = entitlement rage.

p < .05. ***p < .001.

Network of Pathological Narcissism in PD Patients and Controls

The Fused Graphical Lasso networks are reported in Figure 1 (exact values of edges are reported in Supplementary Table S4). At a visual inspection, most edges were similar in the two networks, with only a few exceptions. For instance, the connection between SSSE and EXP was present in PD patients but almost null in controls, and a small negative edge between ER and HS was present in patients but absent in controls. However, the network comparison test indicated that the two networks did not differ significantly in the structure (M = .145, p = .827) nor in the global strength (S = .089, p = .846).

Figure 1.

Networks Estimated With Fused Graphical Lasso.

The centrality of each node is visualized in Figure 2 (exact centrality values are reported in Supplementary Table S5). The CS-coefficients, in the control and patient networks, respectively, were .90 and .60 for EI, and .90 and .55 for bridge EI, thus indicating good to very large stability of all the coefficients. As shown in Figure 2, results indicated that ER, DEV, and GF were central in networks of both groups according to EI. Conversely, in both networks, node EXP was the least connected node, according to EI. Importantly, ER and GF had the highest values in bridge EI in both networks: ER presented the strongest connections to nodes reflecting grandiose narcissism, whereas GF had the strongest connections to nodes reflecting manifestations of vulnerable narcissism.

Figure 2.

Expected Influence (EI) and Bridge Expected Influence of Each Node in the Patient and Control Networks.

Discussion

Since 2009, the PNI has been used extensively in the empirical literature on pathological narcissism. The employment of this measure has undoubtedly increased understanding of both grandiose and vulnerable manifestations of pathological narcissism, and their psychological and behavioral correlates. However, information on the presentations of pathological narcissistic traits measured through the PNI in people with personality disorders is still scarce.

The Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; APA, 2013) Alternative model for personality disorders (AMPD; APA, 2013) considers pathological personality traits as extreme variants of “normal” personality, and some evidence supports continuity between healthy and pathological personality expressions (Krueger & Markon, 2006). According to this view, quantitative, but not qualitative, trait differences should be detected when comparing individuals with and without proper personality pathology. More specifically, people with PDs should report higher PNI traits than people without PDs. Few studies compared mean levels of PNI traits in community and clinical samples (Brown et al., 2020; Fossati et al., 2015; Morf et al., 2017; Rogier & Velotti, 2018), but none of these considered a clinical sample composed only of PD patients. Moreover, the legitimacy of comparing mean levels of PNI traits in community and clinical samples—including PD patients—has not been investigated yet. Only one study investigated measurement invariance of another popular self-report measure of narcissism (i.e., the FFNI) across clinical status (Miller et al., 2016). Results showed metric invariance of the FFNI (i.e., invariance of factor loadings), but they also suggest that means cannot be compared directly among community and clinical samples (i.e., invariance of factor loadings and intercepts was not met). It is of note, also, that the clinical sample in this study was poorly defined, since it included community participants currently receiving treatment for a psychological disorder, but no specific information on their diagnoses was provided. In line with the dimensional nature of the PNI, instead, our findings demonstrate for the first time strong measurement invariance of the PNI between patients with a diagnosis of PD and community participants, thus proving that scholars can employ this self-report instrument to measure and compare traits of pathological narcissism in such samples.

Our results showed that people with personality disorders scored higher in most PNI traits, particularly those related to narcissistic vulnerability. PD patients are thus more prone than community participants to devaluing others and to experiencing intense feelings of shame when others do not meet their needs for recognition. Moreover, PD patients experience more intense anger when their expectations are not met, are more prone to concealing their flaws or failures, and their self-esteem is strongly dependent on external sources of validation, compared with community participants. These results are in line with Miller et al. (2017) showing significant correlations between traits of vulnerable narcissism and traits related to several personality disorders—both measured according to the five factor model. As to traits of grandiose narcissism, PD patients showed higher levels of self-sacrificing self-enhancement, while they did not differ from community people in levels of grandiose fantasies and exploitativeness. According to Wright et al. (2010), PNI traits of grandiose fantasy capture covert aspects of narcissistic grandiosity, while self-sacrificing self-enhancement and exploitativeness represent overt manifestations. Therefore, our results indicate that PD patients are more likely than community participants to express overtly their pathological narcissistic grandiosity, but only by engaging in pseudo-altruistic acts, rather than through open interpersonal exploitation. Of note, network analysis also indicated that exploitativeness has a peripheral role in manifestations of PNI pathological narcissism, both in PD patients and community participants, according to EI and bridge EI. In a sense, our findings can be interpreted as indicative of those aspects that could perform as markers of narcissistic presentations in people with personality disorders. More specifically, important markers of PD patients’ narcissistic presentations could be the severity of emotional and self-esteem dysregulation, as well as the intensity of defensiveness and interpersonal avoidance, which are typical of narcissistic vulnerability, along with the strength of self-sacrificing strategies employed to sustain a grandiose self-image. Conversely, proneness to exploitation seems to be a less essential expression of PNI pathological narcissism in PD patients.

In line with our expectations, the interplay among traits of pathological narcissism was similar in the two groups. In other words, this study suggests that the presence of PDs does not affect the structural organization of pathological narcissistic presentations, beyond the mean levels of the traits. In line with Di Pierro and colleagues (2019), the present study shows that grandiose fantasy and entitlement rage are among the most central PNI narcissistic traits, both in terms of EI and bridge EI. According to Jones et al. (2021), nodes high in bridge expected influence can be seen as responsible for comorbid conditions in psychopathological networks. In this sense, the high bridge expected influence of grandiose fantasy and entitlement rage further suggests these traits might be primarily responsible for the coexistence of vulnerable and grandiose manifestations of PNI pathological narcissism within individuals. A third node, devaluing, was among the most central nodes according to EI, but among the most peripheral in terms of bridge EI. In other words, devaluing had very strong connections with other nodes connected to vulnerability, but very weak direct connections with traits of grandiosity. These findings extend those of Dinić et al. (2021) by showing that entitlement rage and grandiose fantasies have a “bridge” function both in community people and PD patients. More specifically, our study demonstrates that increased proneness to engaging in grandiose fantasies (GF) would also mirror increased manifestations of narcissistic vulnerability within individuals, as in the case of experiences of self-esteem dysregulation (CSE), shame over needing recognition from disappointing others (DEV), rage for unmet entitled expectations (ER), and tendency to hide one’s own flaws (HS). Moreover, increased proneness of individuals to experience anger when entitled expectations are unmet (ER) would also result in increased inclination to present manifestations of narcissistic grandiosity, such as GF and SSSE.

Previous studies suggested that grandiose and vulnerable manifestations of pathological narcissism may coexist within individuals and fluctuate over time (Edershile et al., 2019b; Gore & Widiger, 2016). Our study extends these findings by showing that both PD patients and community participants may present coexisting aspects of vulnerable and grandiose narcissism, and that traits of grandiose fantasy and entitlement rage are specifically responsible for this coexistence. A possible interpretation for the “bridge” function of these traits calls for the defensive role of grandiose manifestations against aspects of narcissistic vulnerability (Caligor & Stern, 2020). In this sense, we may interpret the bridge function of grandiose fantasies as evidence that individuals high in narcissistic vulnerability would engage in grandiose fantasies to protect themselves from vulnerable states. Similarly, the bridge function of entitlement rage may be seen as evidence that individuals who are highly sensitive to unmet entitled expectations may adopt a range of grandiose strategies (e.g., fantasies of grandiosity or pseudo-altruism) to avoid negative emotional states (i.e., resentment). This interpretation, however, needs further corroboration. In particular, existing studies (Edershile et al., 2019b; Gore & Widiger, 2016) investigated the coexistence of grandiose and vulnerable manifestations of pathological narcissism, as well as their fluctuation, without considering the specific role-played by their foundational expressions (i.e., traits). In this sense, further empirical studies are needed to deepen the “bridge” function of grandiose fantasies and entitlement rage, for example by examining the temporal dynamics characterizing the interplay of such traits by means of time series network analysis (Costantini et al., 2019; Epskamp, Waldorp, et al., 2018). Only then would the temporal relations between expressions of narcissistic vulnerability and manifestations of narcissistic grandiosity be clarified.

Di Pierro and colleagues (2019) found that contingent self-esteem also had a central role in the PNI network, something that did not emerge in this study. These seemingly discrepant results are explained by differences in node centrality measures. A strong negative connection of contingent self-esteem with exploitativeness has been found in both studies. However, Di Pierro et al. (2019) considered measures of centrality (e.g., strength centrality) that do not take into account negative edges in networks. Conversely, this study calculated node indices that are particularly suited for networks with both positive and negative edges (Jones et al., 2021), as is the case for PNI networks.

Overall, our findings are partly supportive of Miller et al.’s idea that self-centered antagonism is a core feature of pathological narcissism (Crowe et al., 2019; Miller et al., 2016). Indeed, self-centered antagonism includes a sense of self-importance and related aspects of entitlement (see also Krizan & Herlache, 2018), which also emerge as central nodes in both PD patients and community participants in our study. Furthermore, our findings suggest that self-importance may be qualified by intrapsychic processes (e.g., grandiose fantasies) in pathological narcissism, and not necessarily by overt manifestations. Yet, our results are at odds with the idea that exploitation of others (another aspect of self-centered antagonism according to Crowe et al., 2019) is a central aspect of pathological narcissism. According to Caligor and Stern (2020), moral functioning is compromised in patients with narcissistic pathology, but only those with an underlying low-level borderline personality organization would display frank antisocial features. Moreover, Gunderson and Ronningstam (2001) showed that, even though both narcissistic patients and antisocial ones are exploitative, “the antisocials were significantly more so, and were judged to more actively take advantage of or use other people” (p. 106). Therefore, it is plausible that traits of exploitativeness might have a more central role in defining manifestations of pathological narcissism only when considering patients with severe narcissistic conditions (i.e., those having a narcissistic personality disorder in the context of a low-level borderline personality organization) or with significant antisocial features. It is noteworthy that the antisocial personality disorder was underrepresented in our sample, since only 12 patients out of 153 (7.8%) had such diagnosis. Moreover, it is of note that only some patients had a narcissistic personality disorder in our clinical sample (N = 53; 34.6%) and that we did not evaluate the severity of PDs. In this sense, further empirical studies should investigate the role of exploitativeness in pathological narcissistic manifestations of PD patients with antisocial features and patients with severe narcissistic conditions.

Identification of central nodes in psychopathological networks, especially those having high bridge expected influence, can be particularly useful for clinical purposes since it may help clinicians identify targets for interventions (Jones et al., 2021). Given that grandiose fantasies and entitlement rage showed to have a bridge function in our study, they could potentially represent suitable targets for clinical interventions. Our study has a cross-sectional design that does not allow us to draw any definitive conclusion about the role of these traits in within-person change patterns of pathological narcissistic presentations. In a speculative way, however, our findings could suggest that clinicians might promote a significant change in PD patients who present manifestations of pathological narcissism by reducing proneness to engage in grandiose fantasies and sense of entitlement. By working on grandiose fantasy and entitlement rage, clinicians might expect to promote a significant decrease in other narcissistic traits reflecting both vulnerable and grandiose presentations. In addition, the negative link between contingent self-esteem and exploitativeness could also be clinically relevant. It suggests that when proneness to exploitative attitudes and behaviors decreases, a more unstable and fragile sense of self could emerge. Thus, these moments could be extremely painful for PD patients showing pathological narcissism. Albeit plausible, and grounded on empirical evidence (Jones et al., 2021), these clinical suggestions need to be tested empirically. Again, as reported above, longitudinal design studies would be particularly suitable to further examine these hypotheses, by detecting the effect of suggested clinical interventions on temporal changes in presentations of pathological narcissism in PD patients. It is also of note that centrality cannot be interpreted as an absolute measure of a node’s importance, being dependent on the presence and absence of nodes in the network that are relevant for the phenomenon at hand (Neal & Neal, 2021). In our study, we inspected traits of pathological narcissism captured by the PNI, which is currently one of the most widely used self-report measure of pathological narcissism. However, other popular measures of pathological narcissistic traits are available (e.g., FFNI) and showed to assess dimensions of pathological narcissism that overlap only partially with those captured by the PNI (e.g., Crowe et al., 2019). Therefore, future studies considering a broader set of narcissistic traits might further improve centrality estimates.

In summary, this study suggests that the structure of PNI pathological narcissistic presentations is similar in PD patients and community individuals, and that differences may be detected only when looking at mean levels of traits. The study also suggests that scholars should juxtapose visual inspection of networks and formal psychometric tests when inspecting group differences in trait-interconnections: Indeed, our findings demonstrated that, though different at the visual inspection, trait-interconnection differences were not statistically significant among PD patients and community participants.

Our findings need to be interpreted in light of some considerations on the composition of the two samples. Some of the PD patients showed comorbid clinical syndromes (i.e., mood and adjustment disorders). Although we cannot exclude an impact of these syndromes on our results, comorbidity of clinical syndromes in patients with PDs reflects a clinical reality. In the same vein, while most patients just endorsed one PD diagnosis, some suffered from more than one PD, and PD diagnoses were not distributed homogeneously in the clinical sample. Being diagnosed with more than one PD can be considered a severity index of patients’ clinical conditions (e.g., Preti et al., 2011). Moreover, some PDs rest on lower levels of personality organization than others, that is they reflect more severe impairments in dimensions such as identity and reality testing. Given the study’s aims, however, we did not investigate whether presentations of PNI pathological narcissism in PD patients may differ along with the severity of their clinical condition or level of personality organization. In our study, community participants were younger than PD patients, and empirical studies suggest that traits of both grandiose and vulnerable narcissism usually decrease with age (e.g., Wetzel et al., 2021). Albeit we did not find PD patients to be lower in PNI traits than community participants, we cannot exclude that the effect of age has not significantly impacted mean differences between the two samples in narcissistic traits. Further studies could deepen our results by investigating whether age differences may significantly affect the interplay among pathological narcissistic traits using the network analysis approach. Moreover, our findings should be interpreted with caution since the clinical and the community samples in our study were unbalanced in size, and this sample imbalance might have resulted in a lower precision of estimates for the clinical sample, compared with the control sample. Finally, the community sample of this study only included individuals with no current psychotherapy or psychiatric treatment. However, a formal screening for psychopathology (in its clinical and subclinical forms) was not performed, and we did not collect data on lifetime history of psychological and psychiatric treatments in community participants. Therefore, we were not able to control for the effects of these variables in our analyses. However, we may plausibly expect that the large sample size eventually mitigated the effects of current underdiagnosed (sub-)clinical conditions and of lifetime history of psychological or psychiatric treatments in the community sample.

Conclusion

This study shows that the PNI can be used reliably in both community samples and PD patients. These findings could favor the use of the PNI in clinical populations, especially in PD samples, thus increasing the number of empirical studies on pathological narcissism in clinical settings and leading to a deeper understanding of the way pathological narcissism may manifest in patients with PD.

By investigating similarities and differences in presentations of PNI pathological narcissism between PD patients and community participants, our study also contributes to a better understanding of the dimensional nature of pathological narcissism. Our results suggest that the presence of a PD diagnosis does not affect the interplay between PNI narcissistic traits. Albeit PD patients may express traits of vulnerable narcissism more strongly and engage more in grandiose pseudo-altruistic behaviors compared with community participants, the structure of PNI pathological narcissistic presentations is indeed similar in the two groups. In other words, in line with dimensional approaches to personality pathology (e.g., the AMPD), our study shows that presentations of pathological narcissism differ quantitatively, but not qualitatively, in people with and without a proper diagnosis of PD. Moreover, the study provides new insight into the specific role that grandiose fantasy and entitlement rage may have in manifestations of pathological narcissism, regardless of the presence of a PD diagnosis: These are the most influential traits, and they seem to be responsible for the co-existence of grandiose and vulnerable manifestations of PNI pathological narcissism within individuals.

Supplemental Material

sj-docx-1-asm-10.1177_10731911221101367 – Supplemental material for Measurement Invariance of the Pathological Narcissism Inventory and Multimethod Examination of Narcissistic Presentations in Community and Clinical Samples

Supplemental material, sj-docx-1-asm-10.1177_10731911221101367 for Measurement Invariance of the Pathological Narcissism Inventory and Multimethod Examination of Narcissistic Presentations in Community and Clinical Samples by Rossella Di Pierro, Giulio Costantini, Erika Fanti, Marco Di Sarno, Emanuele Preti, Fabio Madeddu, John F. Clarkin, Eve Caligor and Chiara De Panfilis in Assessment

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author (R.D.P.) upon reasonable request.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Rossella Di Pierro

Giulio Costantini

Supplemental Material

Supplemental material for this article is available online.

Notes

References

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Author.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). American Psychiatric Publishing.

Borsboom

(2017). A network theory of mental disorders. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 16(1), 5–13. https://doi.org/10.1002/wps.20375

Boucher

Côté

Gagnon-Girouard

M. P.

Ratté

Bégin

(2018). Eating pathology among patients with anorexia nervosa and bulimia nervosa: The role of narcissism and self-esteem. The Journal of Nervous and Mental Disease, 206(10), 776–782. https://doi.org/10.1097/NMD.0000000000000890

Bringmann

L. F.

Elmer

Epskamp

Krause

R. W.

Schoch

Wichers

Wigman

J. T.

Snippe

(2019). What do centrality measures measure in psychological networks? Journal of Abnormal Psychology, 128(8), 892–903. https://doi.org/10.1037/abn0000446

Brown

M. F. D.

Stanton

Watson

(2020). Replicable factor structure and correlates of an alternate version of the narcissistic personality inventory. Journal of Psychopathology and Behavioral Assessment, 42, 69–85. https://doi.org/10.1007/s10862-020-09790-y

Cain

N. M.

Pincus

A. L.

Ansell

E. B.

(2008). Narcissism at the crossroads: Phenotypic description of pathological narcissism across clinical theory, social/personality psychology, and psychiatric diagnosis. Clinical Psychology Review, 28(4), 638–656. https://doi.org/10.1016/j.cpr.2007.09.006

Caligor

Levy

K. N.

Yeomans

F. E.

(2015). Narcissistic personality disorder: Diagnostic and clinical challenges. American Journal of Psychiatry, 172(5), 415–422. https://doi.org/10.1176/appi.ajp.2014.14060723

Caligor

Stern

B. L.

(2020). Diagnosis, classification, and assessment of narcissistic personality disorder within the framework of object relations theory. Journal of Personality Disorders, 34(Suppl.), 104–121. https://doi.org/10.1521/pedi.2020.34.supp.104

10.

Chen

F. F.

(2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14(3), 464–504. https://doi.org/10.1080/10705510701301834

11.

Chen

F. F.

Sousa

K. H.

West

S. G.

(2005). Testing measurement invariance of second-order factor models. Structural Equation Modeling, 12(3), 471–492. https://doi.org/10.1207/s15328007sem1203_7

12.

Contreras

Nieto

Valiente

Espinosa

Vazquez

(2019). The study of psychopathology from the network analysis perspective: A systematic review. Psychotherapy and Psychosomatics, 88(2), 71–83. https://doi.org/10.1159/000497425

13.

Costantini

Richetin

Preti

Casini

Epskamp

Perugini

(2019). Stability and variability of personality networks. A tutorial on recent developments in network psychometrics. Personality and Individual Differences, 136, 68–78. https://doi.org/10.1016/j.paid.2017.06.011

14.

Crowe

M. L.

Lynam

D. R.

Campbell

W. K.

Miller

J. D.

(2019). Exploring the structure of narcissism: Toward an integrated solution. Journal of Personality, 87(6), 1151–1169. https://doi.org/10.1111/jopy.12464

15.

Danaher

Wang

Witten

D. M.

(2014). The joint graphical lasso for inverse covariance estimation across multiple classes. Journal of the Royal Statistical Society. Series B, Statistical Methodology, 76(2), 373–397. https://doi.org/10.1111/rssb.12033

16.

Dashineau

S. C.

Edershile

E. A.

Simms

L. J.

Wright

A. G. C.

(2019). Pathological narcissism and psychosocial functioning. Personality Disorders: Theory, Research, and Treatment, 10(5), 473–478. https://doi.org/10.1037/per0000347

17.

Dinić

B. M.

Sokolovska

Tomašević

(2021). The narcissism network and centrality of narcissism features. Current Psychology. Advance online publication. https://doi.org/10.1007/s12144-020-01250-w

18.

Dinić

B. M.

Vujić

(2020). The pathological narcissism inventory: Measurement invariance across Serbian and USA samples and further validation. European Journal of Psychological Assessment, 36, 670–680. https://doi.org/10.1027/1015-5759/a000537

19.

Di Pierro

Costantini

Benzi

I. M. A.

Madeddu

Preti

. (2019). Grandiose and entitled, but still fragile: A network analysis of pathological narcissistic traits. Personality and Individual Differences, 140, 15–20. https://doi.org/10.1016/j.paid.2018.04.003

20.

Di Pierro

Mattavelli

Gallucci

. (2016). Narcissistic traits and explicit self-esteem: The moderating role of implicit self-view. Frontiers in Psychology, 7, 1815. https://doi.org/10.3389/fpsyg.2016.01815

21.

Di Sarno

Zimmermann

Madeddu

Casini

Di Pierro

. (2020). Shame behind the corner? A daily diary investigation of pathological narcissism. Journal of Research in Personality, 85, Article 103924. https://doi.org/10.1016/j.jrp.2020.103924

22.

Edershile

E. A.

Simms

L. J.

Wright

A. G.

(2019a). A multivariate analysis of the pathological narcissism inventory’s nomological network. Assessment, 26(4), 619–629. https://doi.org/10.1177/1073191118766412

23.

Edershile

E. A.

Woods

W. C.

Sharpe

B. M.

Crowe

M. L.

Miller

J. D.

Wright

A. G.

(2019b). A day in the life of narcissus: Measuring narcissistic grandiosity and vulnerability in daily life. Psychological Assessment, 31(7), 913–924. https://doi.org/10.1037/pas0000717

24.

Ellison

W. D.

Levy

K. N.

Cain

N. M.

Ansell

E. B.

Pincus

A. L.

(2013). The impact of pathological narcissism on psychotherapy utilization, initial symptom severity, and early-treatment symptom change: A naturalistic investigation. Journal of Personality Assessment, 95(3), 291–300. https://doi.org/10.1080/00223891.2012.742904

25.

Epskamp

Borsboom

Fried

E. I.

(2018). Estimating psychological networks and their accuracy: A tutorial paper. Behavior Research Methods, 50(1), 195–212. https://doi.org/10.3758/s13428-017-0862-1

26.

Epskamp

Waldorp

L. J.

Mõttus

Borsboom

(2018). The gaussian graphical model in cross-sectional and time-series data. Multivariate Behavioral Research, 53(4), 453–480. https://doi.org/10.1080/00273171.2018.1454823

27.

Euler

Stöbi

Sowislo

Ritzler

Huber

C. G.

Lang

U. E.

Wrege

Walter

(2018). Grandiose and vulnerable narcissism in borderline personality disorder. Psychopathology, 51(2), 110–121. https://doi.org/10.1159/000486601

28.

First

M. B.

Gibbon

Spitzer

R. L.

Williams

J. B. W.

Benjamin

L. S.

(1997). Structured clinical interview for DSM-IV Axis II Personality Disorders (SCID-II). American Psychiatric Press.

29.

Fossati

Borroni

Grazioli

Dornetti

Marcassoli

Maffei

e Cheek

(2009). Tracking the hypersensitive dimension in narcissism: Reliability and validity of the Hypersensitive Narcissism Scale. Personality and Mental Health, 3, 235–247. https://doi.org/10.1002/pmh.92

30.

Fossati

Feeney

Pincus

Borroni

Maffei

(2015). The structure of pathological narcissism and its relationships with adult attachment styles: A study of Italian nonclinical and clinical adult participants. Psychoanalytic Psychology, 32(3), 403–431. https://doi.org/10.1037/a0037175

31.

Fossati

Somma

Borroni

Pincus

A. L.

Markon

K. E.

Krueger

R. F.

(2017). Profiling pathological narcissism according to DSM-5 domains and traits: A study on consecutively admitted Italian psychotherapy patients. Psychological Assessment, 29(11), 1400–1411. https://doi.org/10.1037/pas0000348

32.

Foster

J. D.

McCain

J. L.

Hibberts

M. F.

Brunell

A. B.

Johnson

R. B.

(2015). The grandiose narcissism scale: A global and facet-level measure of grandiose narcissism. Personality and Individual Differences, 73, 12–16. https://doi.org/10.1016/j.paid.2014.08.042

33.

Fried

E. I.

Eidhof

M. B.

Palic

Costantini

Huisman-van Dijk

H. M.

Bockting

C. L. H.

Engelhard

Armour

Nielsen

A. B. S.

Karstoft

K.-I.

(2018). Replicability and generalizability of Posttraumatic Stress Disorder (PTSD) networks: A cross-cultural multisite study of PTSD symptoms in four trauma patient samples. Clinical Psychological Science, 6(3), 335–351. https://doi.org/10.1177/2167702617745092

34.

Gamache

Savard

Leclerc

Côté

(2019). Introducing a short self-report for the assessment of level of personality functioning for personality disorders: The self and interpersonal functioning scale. Personality Disorders: Theory, Research, and Treatment, 10(5), 438–447. https://doi.org/10.1037/per0000335

35.

Gore

W. L.

Widiger

T. A.

(2016). Fluctuation between grandiose and vulnerable narcissism. Personality Disorders: Theory, Research, and Treatment, 7(4), 363–371. https://doi.org/10.1037/per0000181

36.

Gunderson

J. G.

Ronningstam

(2001). Differentiating narcissistic and antisocial personality disorders. Journal of Personality Disorders, 15(2), 103–109. https://doi.org/10.1521/pedi.15.2.103.19213

37.

Hirschfeld

Von Brachel

(2014). Improving multiple-group confirmatory factor analysis in R–A tutorial in measurement invariance with continuous and ordinal indicators. Practical Assessment, Research, and Evaluation, 19(1), Article 7. https://doi.org/10.7275/qazy-2946

38.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

39.

Jaksic

Marcinko

Skocic Hanzek

Rebernjak

Ogrodniczuk

J. S.

(2017). Experience of shame mediates the relationship between pathological narcissism and suicidal ideation in psychiatric outpatients. Journal of Clinical Psychology, 73(12), 1670–1681. https://doi.org/10.1002/jclp.22472

40.

Jakšić

Milas

Ivezić

Wertag

Jokić-Begić

Pincus

A. L.

(2014). The Pathological Narcissism Inventory (PNI) in transitional post-war croatia: Psychometric and cultural considerations. Journal of Psychopathology and Behavioral Assessment, 36(4), 640–652. https://doi.org/10.1007/s10862-014-9425-2

41.

Jiang

Mai

Yuan

K.-H.

(2017). Advances in measurement invariance and mean comparison of latent variables: Equivalence testing and a projection-based approach. Frontiers in Psychology, 8, 1823. https://doi.org/10.3389/fpsyg.2017.01823

42.

Jones

P. J.

McNally

R. J.

(2021). Bridge centrality: A network approach to understanding comorbidity. Multivariate Behavioral Research, 56, 353–367. https://doi.org/10.1080/00273171.2019.1614898

43.

Jorgensen

T. D.

Pornprasertmanit

Schoemann

A. M.

Rosseel

(2021). semTools: Useful tools for structural equation modeling (R package version 0.5-5). https://CRAN.R-project.org/package=semTools

44.

Karakoula

Triliva

(2016). Narcissistic vulnerability and addiction: Findings from a study of people in treatment. Journal of Drug Issues, 46(4), 396–410. https://doi.org/10.1177/0022042616659761

45.

Kealy

Laverdière

Pincus

A. L.

(2020). Pathological narcissism and symptoms of major depressive disorder among psychiatric outpatients: The mediating role of impaired emotional processing. The Journal of Nervous and Mental Disease, 208(2), 161–164. https://doi.org/10.1097/NMD.0000000000001114

46.

Kealy

Tsai

Ogrodniczuk

J. S.

(2012). Depressive tendencies and pathological narcissism among psychiatric outpatients. Psychiatry Research, 196(1), 157–159. https://doi.org/10.1016/j.psychres.2011.08.023

47.

Kealy

Tsai

Ogrodniczuk

J. S.

(2016). Pathological narcissism and somatic symptoms among men and women attending an outpatient mental health clinic. International Journal of Psychiatry in Clinical Practice, 20(3), 175–178. https://doi.org/10.1080/13651501.2016.1199811

48.

Kenny

(2020). Measuring model fit. http://www.davidakenny.net/cm/fit.htm

49.

Kernberg

O. F.

Caligor

(2005). A psychoanalytic theory of personality disorders. In Lenzenweger

M. F.

Clarkin

J. F.

(Eds.), Major theories of personality disorders (2nd ed., pp. 114–156). Guilford Press.

50.

Krizan

Herlache

A. D.

(2018). The narcissism spectrum model: A synthetic view of narcissistic personality. Personality and Social Psychology Review, 22(1), 3–31. https://doi.org/10.1177/1088868316685018

51.

Krueger

R. F.

Markon

K. E.

(2006). Understanding psychopathology: Melding behavior genetics, personality, and quantitative psychology to develop an empirically based model. Current Directions in Psychological Science, 15(3), 113–117. https://doi.org/10.1111/j.0963-7214.2006.00418.x

52.

MacCallum

R. C.

Austin

J. T.

(2000). Applications of structural equation modeling in psychological research. Annual Review of Psychology, 51, 201–226. https://doi.org/10.1146/annurev.psych.51.1.201

53.

Marčinko

Jakšić

Ivezić

Skočić

Surányi

Lončar

Franić

Jakovljević

(2014). Pathological narcissism and depressive symptoms in psychiatric outpatients: Mediating role of dysfunctional attitudes. Journal of Clinical Psychology, 70(4), 341–352. https://doi.org/10.1002/jclp.22033

54.

Miller

J. D.

Few

L. R.

Wilson

Gentile

Widiger

T. A.

MacKillop

Keith Campbell

(2013). The Five-Factor Narcissism Inventory (FFNI): A test of the convergent, discriminant, and incremental validity of FFNI scores in clinical and community samples. Psychological Assessment, 25(3), 748–758. https://doi.org/10.1037/a0032536

55.

Miller

J. D.

Lynam

D. R.

Hyatt

C. S.

Campbell

W. H.

(2017). Controversies in narcissism. Annual Reviews of Clinical Psychology, 13(1), 291–315. https://doi.org/10.1146/annurev-clinpsy-032816-045244

56.

Miller

J. D.

Lynam

D. R.

McCain

J. L.

Few

L. R.

Crego

Widiger

T. A.

Campbell

W. K.

(2016). Thinking structurally about narcissism: An examination of the Five-Factor Narcissism Inventory and its components. Journal of Personality Disorders, 30, 1–18. https://doi.org/10.1080/00273171.2019.1614898

57.

Miller

J. D.

McCain

Lynam

D. R.

Few

L. R.

Gentile

MacKillop

Campbell

W. K.

(2014). A comparison of the criterion validity of popular measures of narcissism and narcissistic personality disorder via the use of expert ratings. Psychological Assessment, 26(3), 958–969. https://doi.org/10.1037/a0036613

58.

Morf

C. C.

Schürch

Küfner

Siegrist

Vater

Back

Mestel

Schröder-Abé

(2017). Expanding the nomological net of the pathological narcissism inventory: German validation and extension in a clinical inpatient sample. Assessment, 24(4), 419–443. https://doi.org/10.1177/1073191115627010

59.

Neal

Z. P.

Neal

J. W.

(2021). Out of bounds? The boundary specification problem for centrality in psychological networks. Psychological Methods. Advance online publication. https://doi.org/10.1037/met0000426

60.

Ogrodniczuk

J. S.

Piper

W. E.

Joyce

A. S.

Steinberg

P. I.

Duggal

(2009). Interpersonal problems associated with narcissism among psychiatric outpatients. Journal of Psychiatric Research, 43, 837–842. https://doi.org/10.1016/j.jpsychires.2008.12.005

61.

Oltmanns

J. R.

Widiger

T. A.

(2018). Assessment of fluctuation between grandiose and vulnerable narcissism: Development and initial validation of the FLUX scales. Psychological Assessment, 30(12), 1612–1624. https://doi.org/10.1037/pas0000616

62.

Pincus

A. L.

Lukowitsky

M. R.

(2010). Pathological narcissism and narcissistic personality disorder. Annual Review of Clinical Psychology, 6, 421–446. https://doi.org/10.1146/annurev.clinpsy.121208.131215

63.

Pincus

A. L.

Ansell

E. B.

Pimentel

C. A.

Cain

N. M.

Wright

A. G. C.

Levy

K. N.

(2009). Initial construction and validation of the pathological narcissism inventory. Psychological Assessment, 21, 365–379. https://doi.org/10.1037/a0016530

64.

Preti

Prunas

De Panfilis

Marchesi

Madeddu

Clarkin

J. F.

(2015). The facets of identity: Personality pathology assessment through the inventory of personality organization. Personality Disorders: Theory, Research, and Treatment, 6(2), 129–140. https://doi.org/10.1037/per0000119

65.

Preti

Prunas

Ravera

Madeddu

(2011). Polydrug abuse and personality disorders in a sample of substance-abusing inpatients. Mental Health and Substance Use, 4(3), 256–266. https://doi.org/10.1080/17523281.2011.577751

66.

Putnick

D. L.

Bornstein

M. H.

(2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. https://doi.org/10.1016/j.dr.2016.06.004

67.

Raskin

R. N.

Terry

(1988). A principal-components analysis of the narcissistic personality inventory and further evidence of its construct validity. Journal of Personality and Social Psychology, 54, 890–902. https://dx-doi-org.web.bisu.edu.cn/10.1037/0022-3514.54.5.890

68.

Restek-Petrović

Bogović

Grah

Filipčić

Ivezić

(2015). Personality characteristics of psychotic patients as possible motivating factors for participating in group psychotherapy. Psychiatria Danubina, 27(Suppl. 1), 364–370.

69.

Robinaugh

D. J.

Millner

A. J.

McNally

R. J.

(2016). Identifying highly influential nodes in the complicated grief network. Journal of Abnormal Psychology, 125(6), 747–757. https://doi.org/10.1037/abn0000181

70.

Rogier

Velotti

(2018). Narcissistic implications in gambling disorder: The mediating role of emotion dysregulation. Journal of Gambling Studies, 34(4), 1241–1260. https://doi.org/10.1007/s10899-018-9759-x

71.

Ronningstam

(2014). Introduction to the special series on “narcissistic personality disorder-new perspectives on diagnosis and treatment.” Personality Disorders, 5(4), 419–421. https://doi.org/10.1037/per0000093

72.

Rosseel

(2012). Lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. https://www.jstatsoft.org/v48/i02/

73.

Sass

D. A.

(2011). Testing measurement invariance and comparing latent factor means within a confirmatory factor analysis framework. Journal of Psychoeducational Assessment, 29(4), 347–363. https://doi.org/10.1177/0734282911406661

74.

Schmitt

Kuljanin

(2008). Measurement invariance: Review of practice and implications. Human Resource Management Review, 18(4), 210–222. https://doi.org/10.1016/j.hrmr.2008.03.003

75.

Schoenleber

Roche

M. J.

Wetzel

Pincus

A. L.

Roberts

B. W.

(2015). Development of a brief version of the pathological narcissism inventory. Psychological Assessment, 27(4), 1520–1526. https://doi.org/10.1037/pas0000158

76.

Sherman

E. D.

Miller

J. D.

Few

L. R.

Campbell

W. K.

Widiger

T. A.

Crego

Lynam

D. R.

(2015). Development of a short form of the five-factor narcissism inventory: The FFNI-SF. Psychological Assessment, 27(3), 1110–1116. https://doi.org/10.1037/pas0000100

77.

Snyder

H. R.

Friedman

N. P.

Hankin

B. L.

(2021). Associations between task performance and self-report measures of cognitive control: Shared versus distinct abilities. Assessment, 28(4), 1080–1096. https://doi.org/10.1177/1073191120965694

78.

Somma

Boni

Arlotta

Nazzaro

Masci

Busso

Beretta

Manini

Borroni

Maffei

Fossati

(2018). Pathological narcissism measures as predictors of self-reported physical aggression among 310 consecutively-admitted Italian outpatients. Journal of Psychopathology, 24(4), 215–223.

79.

Somma

Pincus

A. L.

Fontana

Cianfanelli

Fossati

(2020). Measurement invariance of three versions of the pathological narcissism inventory across gender-matched Italian adolescent high school and young adult university students. Journal of Psychopathology and Behavioral Assessment, 42, 38–51. https://doi.org/10.1007/s10862-019-09758-7

80.

Stanton

McArtor

D. B.

Watson

(2019). Parsing the hypomanic personality: Explicating the nature of specific dimensions defining mania risk. Assessment, 26(3), 492–507. https://doi.org/10.1177/1073191117725170

81.

Stanton

Watson

Clark

L. A.

(2018). Belief in narcissistic insecurity: Perceptions of lay raters and their personality and psychopathology relations. Personality and Mental Health, 12(1), 73–81. https://doi.org/10.1002/pmh.1404

82.

Thomas

K. M.

Wright

A. G. C.

Lukowitsky

M. R.

Donnellan

M. B.

e Hopwood

C. J.

(2012). Evidence for the criterion validity and clinical utility of the pathological narcissism inventory. Assessment, 19, 135–145. http://doi.org/10.1177/1073191112436664

83.

Tonna

Paglia

Ottoni

Ossola

De Panfilis

Marchesi

(2018). Delusional disorder: The role of personality and emotions on delusional ideation. Comprehensive Psychiatry, 85, 78–83. https://doi.org/10.1016/j.comppsych.2018.07.002

84.

Tritt

S. M.

Ryder

A. G.

Ring

A. J.

Pincus

A. L.

(2010). Pathological narcissism and the depressive temperament. Journal of Affective Disorders, 122(3), 280–284. https://doi.org/10.1016/j.jad.2009.09.006

85.

van Borkulo

C. D.

van Bork

Boschloo

Kossakowski

J. J.

Tio

Schoevers

R. A.

Borsboom

Waldorp

L. J.

(2022). Comparing network structures on three aspects: A permutation test. Psychological Methods. Advance online publication. https://doi.org/10.1037/met0000476

86.

van Buuren

Groothuis-Oudshoorn

. (2011). Mice: Multivariate imputation by chained equations in R. Journal of Statistical Software, Foundation for Open Access Statistics, 45(3), 1–67. https://doi.org/10.18637/jss.v045.i03

87.

Wetzel

Grijalva

Robins

R. W.

Roberts

B. W.

(2020). You’re still so vain: Changes in narcissism from young adulthood to middle age. Journal of Personality and Social Psychology, 119(2), 479–496. http://doi.org/10.1037/pspp0000266

88.

Widiger

T. A.

(1992). Categorical versus dimensional classification: Implications from and for research. Journal of Personality Disorders, 6(4), 287–300.

89.

Wright

A. G. C.

Lukowitsky

M. R.

Pincus

A. L.

e Conroy

D. E.

(2010). The higher-order factor structure and gender invariance of the Pathological Narcissism Inventory. Assessment, 17, 467–483. http://doi.org/10.1177/1073191110373227