1. Introduction
Various personality factors are associated with symptoms of psychopathology [Reference Kotov, Gamez, Schmidt and Watson1]. Symptoms of psychopathology vary depending upon an individual’s personality traits, and personality traits may influence the presence of symptoms of psychopathology [Reference Widiger2]. Exploring factors that potentially guard against psychopathology, such as indicators of depression (DS) is of paramount importance because of the prevalence of depression (there are over 298 million sufferers worldwide) [Reference Ferrari, Charlson, Norman, Patten, Freedman and Murray3]. Depression has been associated with an increased risk of mortality in general community populations, as well as in patient populations with chronic illnesses, such as coronary heart disease, cancer, diabetes and stroke [Reference Cuijpers, Vogelzangs, Twisk, Kleiboer, Li and Penninx4]. The researchers conducted three studies to explore the degree to which Subclinical Narcissism (SN) predicted indirectly lower DS, through Mental Toughness (MT) and the Big Five (BF) personality trait of Openness to Experience (OE).
2. Subclinical narcissism, mental toughness and psychopathology
Subclinical Narcissism (SN) is part of the Dark Triad (DT), a personality cluster defined at the subclinical level. This includes the traits of Subclinical Psychopathy (SP) and Machiavellianism [Reference Paulhus and Williams5]. SN includes facets retained from the clinical syndrome, namely grandiosity, entitlement, dominance and superiority [Reference Paulhus and Williams5]. SN includes two main types: Grandiose Narcissism (GN) and Vulnerable Narcissism (VN) [Reference Miller, Hoffman, Gaughan, Gentile, Maples and Keith Campbell6]. GN is characterised by exhibitionism, lack of humility/modesty and interpersonal dominance. VN encompasses negative affect, distrust, selfishness, and a need for attention and recognition [Reference Dickinson and Pincus7, Reference Vize, Lynam, Collison and Miller8]. Research suggests that GN links mainly to positive outcomes, whereas VN predicts negative outcomes [Reference Besser and Zeigler-Hill9]. For example, a recent study observed that participants scoring high on VN also reported higher DT traits, lower Mental Toughness (MT), poor sleep quality, and higher scores on perceived stress [Reference Annen, Nakkas, Sadeghi Bahmani, Gerber, Holsboer-Trachsler and Brand10].
Previous studies suggest that SN might be unique among the DT traits in that, it encapsulates to a large extent prosocial and adaptive behaviours [Reference Veselka, Schermer and Vernon11]. For example, a series of five studies indicated positive associations between SN and daily and dispositional subjective well-being, and couple well-being. Negative associations were found between SN and daily sadness, dispositional depression, daily and dispositional loneliness, daily anxiety and dispositional neuroticism [Reference Sedikides, Rudich, Gregg, Kumashiro and Rusbult12].
More recently, several studies [Reference Onley, Veselka, Schermer and Vernon13, Reference Sabouri, Gerber, Bahmani, Lemola, Clough and Kalak14, Reference Papageorgiou, Wong and Clough15, Reference Papageorgiou, Malanchini, Denovan, Clough, Shakeshaft and Schofield16] have reported moderate positive correlations between SN and MT. MT reflects an effective coping mechanism as reaction to stressors. Specifically, MT allows individuals to proactively seek out opportunities for personal growth [Reference St Clair-Thompson, Bugler, Robinson, Clough, McGeown and Perry17]. Many studies have shown that MT is a personality trait that acts as a resilient factor against psychopathology (see [Reference Lin, Mutz, Clough and Papageorgiou18] for a review). Congruently, both cross-sectional and longitudinal research has consistently linked high MT to significantly lower DS [Reference Gerber, Brand, Feldmeth, Lang, Elliot and Holsboer-Trachsler19–Reference Mutz, Clough and Papageorgiou21].
In terms of the factors that may influence MT, studies using both cross-sectional and longitudinal designs have shown that SN may increase MT. This results in various positive outcomes, such as lower psychopathy and higher school grades [Reference Papageorgiou, Wong and Clough15, Reference Papageorgiou, Malanchini, Denovan, Clough, Shakeshaft and Schofield16]. Furthermore, the authors concluded that the path model from SN to MT to a further outcome might predict positive effects across various contexts, including psychopathology.
3. Subclinical narcissism, mental toughness and openness to experience
Both SN and MT have been associated with the Big Five trait of Openness to Experience (OE) [Reference Furnham, Hughes and Marshall22, Reference Horsburgh, Schermer, Veselka and Vernon23]. In fact, a study reported that the correlation between SN and OE was the strongest out of all correlations between SN and the other four BF traits (Conscientiousness, Extraversion, Neuroticism and Agreeableness) [Reference Furnham, Hughes and Marshall22]. Another study reported that OE correlated with MT and in particular, with the MT facets of interpersonal confidence and challenge [Reference Horsburgh, Schermer, Veselka and Vernon23]. Confidence is the strongest correlate (out of all four facets of MT) between MT and SN [Reference Papageorgiou, Wong and Clough15]. The link between Challenge and OE may be explained conceptually by the fact that individuals that score high on Challenge perceive change and new experiences as an opportunity for growth rather than as a threat [Reference Clough, Earle, Sewell and Cockerill24].
Collectively, these findings suggest that individuals scoring high on SN and MT may be particularly open to experiences. Particularly, that they possess the inclination and confidence to seek out new opportunities for personal growth.
4. The present investigation
The present investigation explored a statistical model testing the notion that SN increases MT and OE contributing indirectly to lower DS. Extensions to the model further examined whether the grandiose, as opposed to the vulnerable aspect of SN, increased MT leading to positive outcomes. It was hypothesised that [Reference Kotov, Gamez, Schmidt and Watson1] SN would exert a significant negative indirect effect on DS through MT across studies. The researchers posited also that [Reference Widiger2] SN would exert a significant negative indirect effect on DS through MT and OE in Study 2 and 3. Furthermore, it was predicted [Reference Ferrari, Charlson, Norman, Patten, Freedman and Murray3] that GN would exert a significant negative indirect effect on DS through MT and OE (Study 3). Finally, it was anticipated [Reference Cuijpers, Vogelzangs, Twisk, Kleiboer, Li and Penninx4] that VN would exert a significant positive indirect effect on DS through MT and OE (Study 3).
5. Method
5.1. Sample
Three studies tested hypotheses 1–4. The first examined whether SN exerted a significant negative indirect effect on DS through MT. The second replicated the findings of the first study with an independent sample, and developed the model by testing whether SN exerted a significant negative indirect effect on DS through MT and OE. The third replicated the findings of Studies 1 and 2, and further extended the model by assessing the differential effect of GN (as opposed to VN) on DS through MT and OE. Progressive extension of the tested model provided an increasingly sophisticated understanding of variable relationships.
5.1.1. Study 1
Participants (N = 364) were recruited online through advertisements on social networks (e.g., Facebook) and word of mouth. Participant mean age was 24.31 years (SD = 9.16, range = 18–79); 56.9% were females. Inspection of univariate outliers identified that three z-scores exceeded 3.5 standard deviations (SDs). These were removed [Reference Bajorski25]. Multivariate outliers exist when data points exceed Cook’s distance of 1.0 [Reference Montgomery, Peck and Vining26]. No responses surpassed Cook’s distance. Skewness values were within the recommended range of −2.0 to +2.0 SDs from the mean [Reference Byrne27]. Specifically, MT = −0.32; Challenge = −0.27; Commitment = −0.14; Control = −0.42; Confidence = −0.40; Narcissism = −0.01; Depression = 1.33. Participants did not receive compensation for taking part.
5.1.2. Study 2
Participants (N = 364) were also recruited online through advertisements on social networks (e.g., Facebook) and word of mouth. Participant mean age was 25.30 years (SD = 7.22, range = 18–59); 64.75% females. Examination of outliers resulted in the removal of five data points. No issues existed with skewness, MT = 0.12; Challenge = −0.58; Commitment = −0.02; Control = 0.15; Confidence = −0.12; Narcissism = −0.09; Openness = −0.20; Depression = 1.19. Participants received no compensation for taking part.
5.1.3. Study 3
Participants (N = 144) were undergraduate students recruited through advertisements in the class and university. They enrolled to participate through the Sona system, which manages research participation. Specifically, it allows students to browse ongoing research and enrol to studies. Participant mean age was 22.08 years (SD = 5.50, range = 18–52); 86.1% females. Twenty-three participants did not disclose their age. Assessment of univariate outliers supported removal of four data points. Cook’s distance values were below 1.0. Skewness results fell between -2 and +2 SDs (Table 1). Once participants completed the study, they received credits as part of course fulfilment.
5.2. Measures
5.2.1. Study 1
Used the 9-item narcissism scale of the Short Dark Triad questionnaire (SD3) [Reference Jones and Paulhus28] to assess Subclinical Narcissism (Cronbach’s alpha = 0.70). The Mental Toughness Questionnaire 48 (MTQ48) [Reference Clough, Earle, Sewell and Cockerill24] measured the four dimensions of MT (Control, Confidence, Challenge and Commitment, respectively) and total MT (Cronbach’s alpha = 0.88, 0.53, 0.80, 0.53, and 0.80 for total MT). The Patient Health Questionnaire 9 (PHQ-9) [Reference Kroenke, Spitzer and Williams31] assessed the DSM-IV symptoms of major depressive disorder (Cronbach’s alpha = 0.76).
5.2.2. Study 2
Again, employed the SD3 [Reference Jones and Paulhus28] (Cronbach’s alpha = 0.67), MTQ48 [Reference Clough, Earle, Sewell and Cockerill24] (Cronbach’s alpha = 0.88, 0.50, 0.81, 0.58, and 0.74 for total MT), and the PHQ-9 [Reference Kroenke, Spitzer and Williams31] (Cronbach’s alpha = 0.86). These were presented alongside the 10-item scale of the Big Five Inventory (BFI) [Reference John and Srivastava30], which measured OE (Cronbach’s alpha = 0.72).
5.2.3. Study 3
Once more used the SD3 [Reference Jones and Paulhus28] (Cronbach’s alpha = 0.73), MTQ48 [Reference Clough, Earle, Sewell and Cockerill24] (Cronbach’s alpha = 0.92, 0.72, 0.84, 0.74, and 0.82 for total MT) and the 10-item BFI [Reference John and Srivastava30] (Cronbach’s alpha = 0.76). These core scales were presented together with the Five-Factor Narcissism Inventory – Short Form (FFNI-SF) [Reference Sherman, Miller, Few, Campbell, Widiger and Crego29] (Cronbach’s alpha = 0.92 and 0.82 for the scales of Grandiose and Vulnerable Narcissism), and The Beck Depression Inventory (BDI) [Reference Beck32] (Cronbach’s alpha = 0.90). Detailed information on the measures appears in the supplementary material section.
5.3. Procedure
Combined measures formed a single document. In Study 1 and 2, the online platform SurveyMonkey (www.surveymonkey.com) hosted this in electronic form. Participants accessed the questionnaire via a message containing a link, a password and a unique participant code. For Study 3, participants completed paper and pencil copies of the booklets within the university. Questionnaire completion was self-paced.
5.4. Analysis
Inspection of descriptive statistics and intercorrelation preceded tests of mediation. Mediation analysis used Process (Model 4) [Reference Hayes and Andrew42] with bootstrapping (1000 resamples) to generate indirect effect estimates with 95% bias-corrected confidence intervals.
In Study 1, the mediation model examined DS as the outcome variable, SN as the predictor and MT as a mediator. Study 2 assessed OE in addition to MT as mediators of the SN-DS relationship. Mediation analysis in Study 3 built upon Study 2 by testing whether facets of SN (GN and VN) had an indirect effect on DS through MT and OE. Specifically, Model 3 assessed GN as a predictor (controlling for VN). Mediator variables were MT and OE, with DS the outcome. Analysis of Model 4 tested VN as a predictor (controlling for GN). For comparison purposes, analysis considered also total SN in Model 5. Given MT is multidimensional [Reference Perry, Clough, Crust, Earle and Nicholls34], subfactors of Challenge, Commitment, Control and Confidence were examined as mediators in each study in addition to total MT.
For assessing mediation, various effects and statistical weights exist. The total effect (c weight) of a predictor on an outcome comprises an indirect effect (a*b weight) and a direct effect controlling for the influence of a mediator (c’ weight). Weight a relates to the effect of the predictor on the mediator. Weight b is the effect of the mediator on the outcome while excluding the effect of the predictor. An indirect effect represents a combination of the regression weight of the predictor on the mediator and the regression weight of the mediator on the outcome.
6. Results
6.1. Study 1
Inspection of descriptive statistics and intercorrelations revealed significant relationships between SN (M = 2.93, SD = 0.57) and MT (M = 3.44, SD = 0.37) (r = 0.40), SN and DS (M = 5.43, SD = 4.52) (r = −0.16), and MT and DS (r = −0.53). Analysis of Model 1 (Fig. 1) found that SN positively predicted MT and negatively predicted DS (a weight). MT possessed a negative relationship with DS (b weight) and a mediating effect as indicated by a significant a*b weight. The direct effect of SN was non-significant (c’ weight), supporting presence of mediation in addition to a meaningful indirect effect. Since the model was cross-sectional, reversing the paths between DS and SN revealed a considerably weaker indirect effect (a*b weight = −0.02 vs. −1.77). Similarly, a model assessing SN as mediator between MT-DS revealed a non-significant indirect effect, a*b = 0.31, 95% CI [−0.16, 0.84], supporting the direction in Model 1.
Further scrutiny of MT subfactors revealed significant indirect effects of SN on DS through Challenge, a*b = 0.40, 95% CI [0.18, 0.74], Commitment, a*b = −0.29, 95% CI [−0.64, −0.08], Control, a*b = −0.32, 95% CI [−0.63, −0.11], and Confidence, a*b = −1.86, 95% CI [−2.60, −1.22]. All subfactors were in the expected direction apart from Challenge, which evidenced a positive mediating effect on DS. Study 1 data available via figshare: https://doi.org/10.6084/m9.figshare.6887123.
6.2. Study 2
Analysis followed the same steps as Study 1. Descriptive statistics and intercorrelations revealed significant relationships between SN (M = 2.88, SD = 0.55) and MT (M = 3.42, SD = 0.38) (r = 0.44), SN and OE (M = 3.46, SD = 0.51) (r = 0.27), MT and DS (M = 7.13, SD = 5.43) (r = −0.51), and SN and DS (r = −0.13). No significant association existed between OE and DS.
Model 1: Mediation effects of Mental Toughness in the relation between narcissism and depression. Note. R2 for final model = 28.89%; results are based on 1000 bootstrap samples; values in square brackets indicate bias-corrected 95% confidence intervals; **p < 0.01, ***p < 0.001.
Analysis of Model 2 (Fig. 2) indicated that SN positively predicted MT and OE (a weights). MT negatively predicted DS, whereas OE positively predicted DS (b weights). SN had a significant indirect effect on DS through MT and OE (a*b weights). Additionally, a non-significant c’ path supported mediation. Reversing paths between SN and DS revealed considerably weaker indirect effects (MT a*b weight = −0.02 vs. −2.59; OE a*b weight = 0.01 vs. 0.52). Assessing SN as mediator of MT-DS revealed a non-significant indirect effect, a*b = 0.43, 95% CI [−0.24, 1.14]. Testing MT as a mediator of OE-DS (controlling for SN) reported a non-significant total effect, c = 0.05, 95% CI [−0.08, 0.19], supporting variable direction in Model 2. OE possessed a non-significant correlation with DS but a significant mediating effect. Possibly, this was due to shared variance with SN and MT. The correlation between OE and DS reached significance when controlling for the effects of SN and MT, partial r = 0.22, p = 0.001.
Scrutiny of MT subfactors revealed significant indirect effects of SN on DS via Control, a*b = −3.48, 95% CI [−5.86, −1.11], and Confidence, a*b = −4.24, 95% CI [−5.98, −2.51]. As with Model 2, OE exerted a significant mediating effect, a*b = 0.41, 95% CI [0.15, 0.80]. This is likely due to shared variance given the correlation between OE and DS was significant when controlling for SN and MT subfactors, partial r = 0.17, p = 0.008. Study 2 data available through figshare: https://doi.org/10.6084/m9.figshare.6887132.
6.3. Study 3
Consideration of zero-order correlations revealed several significant associations (Table 1). Grandiose Narcissism (GN), SN and MT correlated positively. OE was positively associated with SN and MT. DS correlated negatively with GN, SN and MT, and positively with Vulnerable Narcissism (VN). There was no significant relationship between DS and OE.
Model 3 (Fig. 3) revealed GN positively predicted MT (not OE) and VN negatively predicted MT (not OE) (a weights). MT and OE revealed negative and positive relationships with DS respectively. GN had a significant indirect effect on DS through MT (a*b weight). OE did not have a mediating effect.
Model 4 (Fig. 3) reported identical a and b weights. VN had a significant indirect effect on DS via MT (a*b weight). OE did not have a mediating effect. Note the negative relationship between MT and DS in Model 3 and 4, and the non-significant direct effects of GN and VN (c’ weights). These results, alongside significant indirect effects, support mediation. Reversing relationships reported a comparatively weaker indirect effect through MT with both GN (a*b weight = −0.01 vs. −3.26) and VN (a*b weight = 0.02 vs. 4.66). Similarly, testing MT as a mediator of OE-DS (controlling for GN and VN) reported a non-significant total effect, c = 0.79, 95% CI [−1.10, 2.70].
Model 2: Mediation effects of Mental Toughness and openness in the relation between narcissism and depression. Note. R2 for final model = 30.03%; results are based on 1000 bootstrap samples; values in square brackets indicate bias-corrected 95% confidence intervals; *p < 0.05, **p < 0.01, ***p < 0.001.
Models 3 and 4: Mediation effects of Mental Toughness and openness in the relation between grandiosity, vulnerability and depression. Note. R2 for Model 3 = 44.75%, R2 for Model 4 = 44.75%; results are based on 1000 bootstrap samples; values in square brackets indicate bias-corrected 95% confidence intervals; *p < 0.05, ***p < 0.001.
Analysis of MT subfactors for Model 3 revealed significant indirect effects of GN on DS through Control, a*b = −1.38, 95% CI [−2.87, −0.33], and Confidence, a*b = −1.53, 95% CI [−3.33, −0.38]. For Model 4, significant indirect effects of VN on DS occurred via Control, a*b = 2.62, 95% CI [0.39, −5.40], and Confidence, a*b = 1.78, 95% CI [0.44, 3.48]. Similar to Models 3 and 4, OE did not exert a mediating effect.
In comparison with Study 1 and 2, Model 5 (Fig. 4) indicated SN positively predicted MT and OE (a weights). MT evinced a negative relationship with DS, and OE positively predicted DS (b weights). MT had a negative mediating effect, whereas OE exhibited a positive mediating effect (a*b weights). Weaker mediation effects were evident in the context of reversed relationships for both MT (a*b weight = −0.02 vs. −4.58) and OE (a*b weight = 0.01 vs. 0.47). Assessing SN as mediator of MT-DS (controlling for OE) revealed a non-significant indirect effect, a*b = 0.68, 95% CI [−0.39, 2.08].
Model 5: Mediation effects of Mental Toughness and openness in the relation between narcissism and depression. Note. R2 for final model = 43.40%; results are based on 1000 bootstrap samples; values in square brackets indicate bias-corrected 95% confidence intervals; *p < 0.05, **p < 0.01, ***p < 0.001.
Lastly, given OE demonstrated a non-significant correlation with DS but a significant mediating effect, controlling for MT and SN revealed a significant association, partial r = 0.19, p = 0.028. This indicates that MT and SN inflated the link between OE and DS. The results of these further analyses are consistent with Study 1 and Study 2. Data for Study 3 available through figshare: https://doi.org/10.6084/m9.figshare.6887135.
7. Discussion
The path model, from Subclinical Narcissism (SN) to higher Mental Toughness (MT) to a positive outcome, is a reliable and strong predictor of lower symptoms of depression (DS). Explicitly, the results demonstrate that SN, assessed with the SD3 [Reference Jones and Paulhus28], increases MT resulting in significantly lower DS in three independent samples. Importantly, the results were stable across studies explaining almost 30% of the variation in DS. Assessment of MT subfactors replicated these results in relation to Control and Confidence. Considering the conceptual link between MT Challenge and the Big Five trait of Openness to Experience (OE), the authors extended the proposed mediation model showing (in Study 2) that SN may decrease DS through OE. Study 3, however, did not replicate this finding. This was possibly due to small sample size.
Previous investigations [Reference Papageorgiou, Wong and Clough15, Reference Papageorgiou, Malanchini, Denovan, Clough, Shakeshaft and Schofield16] have acknowledged that, scores for SN obtained using the SD3 might be biased towards assessing narcissism as a prosocial trait, linked to healthy self-esteem, rather than evaluating the antisocial aspects of narcissism (see also [Reference Maples, Lamkin and Miller43]). To account for this bias, the researchers included an additional valid measure of SN, the short form of the FFNI [Reference Sherman, Miller, Few, Campbell, Widiger and Crego29], which differentiates between Grandiose SN and Vulnerable SN. This illustrated that GN may predict lower psychiatric symptoms (DS) through MT. Whereas VN exerted a positive indirect effect on DS through MT explaining more than 40% of the variation in DS.
This investigation shares well-reported limitations with other research in the domains of personality and psychopathology. Namely, self-report data may be influenced by common-method variance [Reference Podsakoff, MacKenzie, Lee and Podsakoff44], and social desirability, particularly in the context of the assessment of a “dark” trait (narcissism). Another potential limitation of this investigation is that the researchers did not employ survey validity checks [Reference Becker, Dvorsky, Holdaway and Luebbe45] to ensure that measures were reliably completed. This does not appear to have significantly influenced the results because Study 3 (completed via pen and paper) produced similar findings (correlations and mediation models) to Study 1 and 2.
Additionally, the studies were cross-sectional, which precludes definitive conclusions concerning the causal order of the variables. Mitigating this concern, a test of reverse relationships supported the hypothesised variable order, and previous longitudinal work is consistent with the notion that Mental Toughness mediates the effects of narcissism (cf. [Reference Papageorgiou, Wong and Clough15]).
The present investigation provides new and counterintuitive insights into the role of a seemingly “dark” trait in reducing indirectly and significantly psychiatric symptoms in the general population. Current (and previous) findings, suggest that simply including the SN into the DT, as a trait that links to poor and toxic psychosocial outcomes, requires revision. This suggestion finds partial support in a recent large meta-analysis and critical review of the literature on the DT traits: the study failed to report statistically significant correlations between SN and various measures of negative psychosocial outcome, such as, antisocial tactics, aggression, sex-related issues and morality problems (with the exception of a weak positive correlation between SN and interpersonal difficulties), when controlling for SP and Machiavellianism [Reference Muris, Merckelbach, Otgaar and Meijer46].
8. Conclusion
The present investigation has direct theoretical and indirect applied implications. The findings support the view that SN is a complex personality trait involving both positive (grandiose) and negative (vulnerable) aspects. Exploring its relation to prosocial traits, such as MT, can be particularly helpful when trying to identify and promote SN’s adaptive tendencies. Studying the proposed path model from SN to higher MT, while considering other personality traits (e.g. Openness to Experience) and the distinction between GN and VN, may explain and predict variation in psychiatric symptoms. Considering the malleability of personality traits, joint intervention programmes could promote the adaptive—rather than maladaptive—aspects of SN and train MT in an attempt to reduce DS and possibly other psychiatric symptoms.
Data deposition
Data is accessible through figshare:
https://doi.org/10.6084/m9.figshare.6887123(Study1); https://doi.org/10.6084/m9.figshare.6887132(Study2); https://doi.org/10.6084/m9.figshare.6887135(Study3).
Comments
No Comments have been published for this article.