Elsevier Editorial System(tm) for

Personality and Individual Differences

Manuscript Draft

Manuscript Number: PAID-D-18-01485R1

Title: The Bright Side of Dark: Exploring the Positive Effect of

Grandiose Narcissism on Perceived Stress through Mental Toughness

Article Type: Full Length Article

Section/Category: Multiple Study articles (<10,000 words)

Keywords: subclinical narcissism; grandiose narcissism, vulnerable

narcissism; mediation; mental toughness; perceived stress; symptoms of
psychopathology.

Corresponding Author: Dr. KOSTAS A. PAPAGEORGIOU, PhD

Corresponding Author's Institution: Queen's University Belfast

First Author: KOSTAS A. PAPAGEORGIOU, PhD

Order of Authors: KOSTAS A. PAPAGEORGIOU, PhD; Foteini-Maria Gianniou;

Paul Wilson; Giovanni B Moneta; Delfina Bilello; Peter J Clough

Abstract: Previous research reported that Subclinical Narcissism (SN) may

increase Mental Toughness (MT) resulting in positive outcomes such as
lower psychopathy, higher school grades and lower symptoms of depression.
We conducted three studies (N = 364, 240 and 144 for studies 1, 2 and 3,
respectively) to test a mediation model, which suggests that SN may
increase MT predicting lower Perceived Stress (PS). The participants were
drawn from the general population in studies 1 and 2; and were
undergraduate students in study 3. SN exerted a negative indirect effect
on PS, through MT across all three studies: β = -.26, SE = .039, 95% CI
[-.338, -.187]); β = -.25, SE = .050, 95% CI [-.358, -.160]); β = -.31,
SE = .078, 95% CI [-.473, -.168]). The results were replicated in the
combined dataset. In study 3, we extended the sensitivity of the model
showing that, it is the Grandiose SN that decreases PS, through MT;
Vulnerable SN exhibited the reverse pattern. The findings indicate that
the model, from SN to MT, may predict positive outcomes in various
domains (e.g. in education and psychopathology) suggesting that inclusion
of SN in the dark triad of personality may need to be reconsidered.
Response to reviewers - WITHOUT author identities

Ms. Ref. No.: PAID-D-18-01485

Title: The Bright Side of Dark: Exploring the Positive Effect of Grandiose Narcissism on

Perceived Stress through Mental Toughness

Reviewer 1

Comment 1: In the introduction, the theoretical based review about the model is needed

included in 1.2 mental toughness and perceived stress.

Response: Thank you. We have merged sections 1.2 and 1.3 into one: “1.2 Mental

Toughness, Narcissism and Perceived Stress” as per your suggestion.

Comment 2: In the method, the study 1 2 3 were related to difference sample pool and need

combined as one study with 3 sample.

Response: Thank you. We would like to direct the reviewer’s attention to section 2.1.3,

where we described the combined dataset: “The datasets derived from the three studies were

combined to form a large dataset of 748 participants.”

Comment 3: In the statistic analysis, The SEM could be used to improve the mediator

Model.

Response: Thank you. We have thought about running SEM, however, the sample size of

study 3 is rather small for SEM to return reliable associations. As such, we would like to retain the

mediation analyses throughout this manuscript for consistency.

Reviewer 2
 2

Comment 1: I like this paper and I think it is potentially very important. My comments are

mostly for clarification or recommending specific additions, such as discussion of a variable or

addition of a statement about permissions.

Response: Thank you for your comments and we are pleased to read that you have liked

this manuscript.

Comment 2: The abstract is a bit lacking on information about the study. The first sentence

of the abstract is unclear; is it summarizing previous research or is it your hypothesis statement?

Sample size for each study, and some aspect of statistical analysis and obtained results should be

reported. The populations differ for the three studies, and that should also be addressed here (just

who the three samples represent).

Response: Thank you for this comment. We have revised the abstract and we present an

extended version in this revised manuscript as per your comments: “Previous research reported

that Subclinical Narcissism (SN) may increase Mental Toughness (MT) resulting in positive

outcomes such as lower psychopathy, higher school grades and lower symptoms of depression. We

conducted three studies (N = 364, 240 and 144 for studies 1, 2 and 3, respectively) to test a

mediation model, which suggests that SN may increase MT predicting lower Perceived Stress (PS).

The participants were drawn from the general population in studies 1 and 2; and were

undergraduate students in study 3. SN exerted a negative indirect effect on PS, through MT across

all three studies: β = -.26, SE = .039, 95% CI [-.338, -.187]); β = -.25, SE = .050, 95% CI [-.358,

-.160]); β = -.31, SE = .078, 95% CI [-.473, -.168]). The results were replicated in the combined

dataset. In study 3, we extended the sensitivity of the model showing that, it is the Grandiose SN

that decreases PS, through MT; Vulnerable SN exhibited the reverse pattern. The findings indicate

that the model, from SN to MT, may predict positive outcomes in various domains (e.g. in
 3

education and psychopathology) suggesting that inclusion of SN in the dark triad of personality

may need to be reconsidered.”

Comment 3: Participants: Study 3 is heavily female (86.1%). Studies 1 and 2 are more

balanced (56.9% and 64.60%, respectively). Study 3's mean age = 22.08 years (SD = 5.5, range =

18-52), and study 1's Mean Age = 24.31 years (SD = 9.16, range = 18-79); Study 2's mean age =

25.43 years (SD = 7.21, range = 18-59); This is young sample. Study 1: 33.47; Study 2: 32.64;

Study 3: 27.58 (approximately 84th percentile of age).

Response: Thank you. We have included the following sentences in the discussion of this

revised manuscript: “The samples’ mean age was rather young across all three studies, which may

limit the generalizability of the findings to older samples. Furthermore, Study 3 was based

completely on a student sample that consists mainly of female participants, which limits the

generalizability of the findings to the general population.”

Comment 4: Combined data set was 65% female. No SD for mean age in combined

sample.

Response: Thank you. We have revised the following sentence to read: “Their mean age

was 24.30 years (SD = 8.09; range = 18 – 79 years).”

Comment 5: Sample size was determined by reviewing previous research; did you conduct

a G*Power analysis or other power analytic to determine minimum effective sample size?

Response: Thank you. In addition to checking sufficient sample size using previous

literature, we also performed post-hoc power analyses to confirm the power of the current datasets.

We have added the following paragraph on section 2.1.3 of this revised manuscript: “We used the

R package Monte Carlo Power Analysis for Indirect Effects (Schoemann, Boulton, & Short, 2017)
 4

to estimate the power of the mediation effect for each study. This indicated that the analyses were

overpowered. We also estimated the minimum sample size that would be required for 80% power

for each dataset which indicated a sufficient sample would be 38 ≤ N ≤ 44 for the mediation effects

found in the paper. Specifically:

Study 1: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power =

44; Study 2: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power =

43; Study 3: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power =

38; Combined Data: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80%

power = 39.”

Comment 6: Measures: Please report reliability statistics for the SD3, FFNI, MTQ48, and

the Perceived Stress Scale. You only need to report reliability for the parts of the scale you use,

such as the SN scale from the SD3. Alpha reliability and test-retest estimates should suffice.

Response: Thank you. Cronbach’s Alphas have been reported for all variables in Table 1.

We had stated this in sections 2.4.1 and 3.1. We have added the following text in section 3.1 of

this revised manuscript: “Descriptive statistics and Cronbach’s alpha for SN, MT and PS are

presented in Table 1 for all three studies and the combined dataset. Cronbach’s alpha for SN

ranged from .67 (study 2) to .73 (study 3). Cronbach’s alphas for GN and VN were .92 and .82,

respectively. Cronbach’s alpha for MT ranged from .88 (study 1 and 2) to .92 (study 3).

Cronbach’s alpha for PS ranged from .84 (study 1) to .89 (study 2).”

Comment 7: You state that each of the above scales have "good validity" in the description.

Please provide examples of validity for each. For example, you reported that the MTQ48 has

verified construct and criterion validity. (Horsburgh et al., 2009). Briefly report the findings with
 5

estimates. You don't need to go into details on the factor structure, for example, but just enough so

that the reader knows what you mean by "good" reliability and "good" validity. Some consider r =

0.65 to be good test/retest reliability, for example, and others consider it to be insufficient,

depending on what is being measured and how well defined the construct is. Mental Toughness,

for example, is a lot messier of a construct than subclinical narcissism.

Response: Thank you. We have added the following paragraphs to present previously

reported alphas for the SD3 and FFNI-sf scales of narcissism, the MTQ48 and the PSS,

respectively:

“Jones and Paulhus (2014) reported Cronbach’s alphas for the scale of narcissism that

ranged from .68 to .78 in three independent studies.”

“In the original study Sherman et al. (2015) reported Cronbach’s alphas that ranged from

.70 for distrust to .89 for exploitativeness with a mean of .81.”

“A recent study that tested the psychometric properties of the MTQ48 reported an overall

omega value of .82 indicating high reliability (Vaughan, Hanna, & Breslin, 2017). Good

reliabilities were also shown in previous research for the MT subscales with alphas ranging

from .65 to .85 ( Perry, Clough, Crust, Earle, & Nicholls, 2013).”

“In the original study, Cohen et al., (1983) reported Cronbach’s alphas for the PSS that

ranged from .84 to .86 in three independent samples.”

Comment 8: Procedure: In our lab we have had mixed results with online surveys. This is

one reason why citing reliability statistics for the instruments is important and then reporting your

reliability findings from the study. Please address the steps you took to secure the online portions

of your study (to make sure the person signing up was actually the one who completed it) and
 6

ensure consistency in administration. For example, how long did it take on average fro people to

complete the survey? Did they have to do it all in one sitting or could they log off and come back?

Response: Thank you. We have amended the procedure section to read: “Questionnaire

completion was self-paced, and participants could proceed to the subsequent page only once they

had answered all items. Participants had to complete the survey in one session (i.e. they could not

save their responses and return to complete the questionnaires) and spent between 30 and 45

minutes to complete the survey.”

Please also see response to comment 6 above regarding reliabilities.

Comment 9: Also there is no statement that the study was approved by an IRB or human

subjects review board.

Response: Thank you. We have added the following sentence in the method section: “All

three studies were approved by the university’s ethics committee.”

Comment 10: 2.4 Statistical Analyses; 2.4.1 Descriptive Statistics and Covariates.

Some will be confused by your reporting of nonsignificant results. For example, you reported

"Participants' age did not correlate significantly with SN (r = -.05, p > .05; r = -.09, p > .05; r = -

.16, p > .05; for study 1, study 2 and study 3, respectively)." You reported p greater than .05, which

is appropriate. However, you reported p less than .05 for results you said were non-significant. For

example, "Participants' age did not correlate significantly with MT and PS (r = -.04, p < .05 and r =

.02, p < .05 for study 3)." Note the p less than .05 in this sentence. This occurs in other places.

Because "less than" and "greater than" signs are similar and many who read results have a tendency

to look for the "less than" sign, the greater than sign may be overlooked and thus misread. In the

cases where you used the incorrect sign, it will cause added confusion. I recommend first stating
 7

that you set alpha = .05 for each set of correlations, and then reporting p = ns (or

something similar) for your nonsignificant results.

Response: Thank you for noticing this mistake. We have corrected the direction of the

signs in the revised manuscript. As per your suggestion, we have also added the following

sentence: “The alpha value was set to .05 when testing for the possible effects of the variables age

and sex.”

Comment 11: You stated that you maintained gender as a covariate but I'm not sure what

you found with gender. Dark Triad traits are more common in males but your sample is more

heavily female. How does this affect the interpretation of your results? I think it should be

discussed.

Response: Please see response to Comment 14 below.

Comment 12: The MTQ48 is a proprietary instrument with a specific sales process behind

it. It is owned (I think) by AQR International. You put the items from a copyrighted instrument

into an online survey platform; please indicate that you had specific permission to do so from the

owners of the MTQ48 and any other copyright holders for the other instruments you used. In

general, a simple sales agreement (buying test forms) does not include permission to modify the

test (such as making it into an online test without the owner's knowledge.).

Response: The MTQ48 has been developed by one of the co-authors. He holds the rights to

use it and he can do so as he sees fit. More information on this may reveal information on the

identity of the authors that is (I believe) against the journal’s policy during the review process.

Comment 13: You said that partial correlations controlled for gender and age were reported

in table 2. Table 2 does not mention those variables as controlled and it only lists the correlations.
 8

This is important given that gender, as a nominal-scale variable, requires a correlation formula that

allows non-continuous variables (such as gender).

Response: Thank you. This is now specified at the bottom of the revised Table 2.

Comment 14: Dark triad traits appear to be more common in males; Jonason, P. K., Li, N.

P., Webster, G. W., & Schmitt, D. P. (2009). Does the relationship between SN, PS, and MT that

you propose hold true for males versus females? I can see why it would but also why it would not.

Males are rewarded for displays of subclinical GN in many first world countries (the USA,

Australia, England, etc.) and females are, at least, not rewarded, and in some cases are punished for

GN displays. I'd like to see you discuss these results in more detail with respect to the differences

in social roles and gender role.

Response: Thank you. We tested our model in the combined dataset in males and females

separately. In response to your question, we have added the following paragraph in section 3.3 of

this revised manuscript: “Finally, considering that all datasets included more female than male

participants; and the reported sex differences in dark traits (e.g. Jonason & Davis, 2018) we

tested our proposed mediation model for males and females, separately. We tested this in the

combined dataset only as this dataset in large enough to allow separating between males and

females and still be able to observe reliable associations. The indirect effect of SN on PS through

MT in the male-only sample (N = 258) was negative and significant: β = -.26, SE = .049, 95%

CI [-.364, -.171]). The indirect effect of SN on PS through MT in the female-only sample (N =

466) was negative and significant: β = -.27, SE = .038, 95% CI [-.361, -.208]). As such, the

results were very similar for both males and females and almost identical to the results obtained,

when using the combined dataset (β = -.27, SE = .028, 95% CI [-.327, -.216]).”
 9

We have also discussed the findings in the discussion of this revised manuscript:

“Considering previous findings on sex differences in the dark traits (e.g. Jonason & Davis,

2018), we tested our proposed mediation model in two sub-samples (derived from the combined

dataset) of male-only vs. female-only participants. The results in the two sub-samples were very

similar, which indicates that, while sex differences in SN are present, they do not appear to

influence proposed mediation model.”

Comment 15: Overall, I like this study and I want to see it published. You will have to

make some corrections, however, and possibly add some information.

Response: Thank you. We have tried to address all your comments in this revised

manuscript.
*Title page with author details

The Bright Side of Dark: Exploring the Positive Effect of Narcissism on Perceived Stress

through Mental Toughness

Kostas A. Papageorgiou, Foteini-Maria Gianniou, Paul Wilson, Giovanni B. Moneta, Delfina

Bilello, and Peter J. Clough

Author Note

Kostas A. Papageorgiou, School of Psychology, Queen’s University Belfast, UK; and

Tomsk State University, Tomsk, Russia; Foteini-Maria Gianniou, School of Psychology,

Queen’s University Belfast, UK; Paul Wilson, School of Psychology, Queen’s University

Belfast, UK; Giovanni B. Moneta, School of Social Sciences, London Metropolitan University,

UK; Delfina Bilello, School of Psychology, Queen’s University Belfast, UK; Peter J. Clough,

Department of Psychology, Huddersfield University, UK.

Correspondence concerning this article should be addressed to Kostas A. Papageorgiou,

School of Psychology, Queen’s University Belfast, United Kingdom; David Keir Building, 18-30

Malone Road, Belfast, BT9 5BN; email: [email protected]; phone: +44(0) 289097

5653

Conflict of interest

All authors declare no conflicts of interest.

This is a revised version of the manuscript: PAID-D-18-01485

Acknowledgements
We gratefully acknowledge all the participants that took part in these three studies. We

are particularly grateful to a number of undergraduate and MSc students that acted as research

assistants and contributed to data collection at London Metropolitan University and Queen’s

University Belfast. The studies were not preregistered.

Author Contributions
K.A.P. has written the manuscript and he was responsible for data collection and data

preparation; F.M.G., D.B., P.W. and G.B.M. were responsible for data collection and data

preparation for study 3; K.A.P., F.M.G. and G.B.M. have performed the statistical analyses;

P.J.C. and G.B.M. provided critical reviews and comments during the write-up of this

manuscript. P.J.C. is responsible for the conceptualisation of MT (as described in this

manuscript) and the creation of the MTQ48 to assess MT; K.A.P. was responsible for the overall

conceptualisation of the studies that are presented in this manuscript. All authors contributed to

the revisions of this manuscript.

*Manuscript without author identities
Click here to view linked References 1
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

The Bright Side of Dark: Exploring the Positive Effect of Grandiose Narcissism on

Perceived Stress through Mental Toughness

 2
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Abstract

Previous research reported that Subclinical Narcissism (SN) may increase Mental Toughness

(MT) resulting in positive outcomes such as lower psychopathy, higher school grades and lower

symptoms of depression. We conducted three studies (N = 364, 240 and 144 for studies 1, 2 and

3, respectively) to test a mediation model, which suggests that SN may increase MT predicting

lower Perceived Stress (PS). The participants were drawn from the general population in studies

1 and 2; and were undergraduate students in study 3. SN exerted a negative indirect effect on PS,

through MT across all three studies: β = -.26, SE = .039, 95% CI [-.338, -.187]); β = -.25, SE =

.050, 95% CI [-.358, -.160]); β = -.31, SE = .078, 95% CI [-.473, -.168]). The results were

replicated in the combined dataset. In study 3, we extended the sensitivity of the model showing

that, it is the Grandiose SN that decreases PS, through MT; Vulnerable SN exhibited the reverse

pattern. The findings indicate that the model, from SN to MT, may predict positive outcomes in

various domains (e.g. in education and psychopathology) suggesting that inclusion of SN in the

dark triad of personality may need to be reconsidered.

Keywords: subclinical narcissism; mental toughness; perceived stress; symptoms of

psychopathology.
 3
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

1.0 Introduction
Various personality factors have been associated with symptoms of psychopathology

(e.g. Kotov, Gamez, Schmidt, & Watson, 2010). These associations can be bidirectional:

symptoms of psychopathology can vary depending upon an individual’s personality traits, and

the appearance of personality traits may be influenced by the presence of symptoms of

psychopathology, such as stress (Widiger, 2011). Stress is a multifaceted construct and may be

defined as a state of imbalance, a disruption of homeostasis, between internal or external

demands and personal or social resources available to cope with such demands (Cohen,

Kamarck, & Mermelstein, 1983). Stress has become a major threat to public health, for instance

by putting college students at risk for a range of health conditions such as depression or anxiety

(Nguyen-Michel, Unger, Hamilton, & Spruijt-Metz, 2006) and cardiovascular disease (Steptoe &

Kivimäki, 2012). It is, however, often not the level of stress per se that negatively affects

individuals’ health but rather their appraisal of it (Tugade & Fredrickson, 2004).

Perceived Stress (PS) indicates the degree to which an individual appraises his or her own

life as stressful, when what is demanded exceeds the ability to cope with it (Cohen, et al., 1983).

Research has shown that high levels of PS is linked to psychopathology with a study reporting

that PS mediates the relationship between environmental stressors and individuals’ risk for

depression ( Lee, Joo, & Choi, 2013). Exploring factors that may guard against symptoms of PS,

such as personality traits, is of paramount importance because the negative consequences of

stress-related illness are large contributors to the global burden of disease (Walker, McGee, &

Druss, 2015). We conducted three studies to explore the degree to which the personality trait of

Subclinical Narcissism (SN) predicts indirectly lower symptoms of Perceived Stress (PS)

through Mental Toughness (MT).

 4
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

1.1 Subclinical Narcissism and Perceived Stress

Narcissus – a mythological figure derived from the Greek mythology – is the origin of

the term narcissism, a fixation with oneself and one's physical appearance and/or public

perception. Subclinical Narcissism (SN) is part of the Dark Triad (DT), a personality cluster that

is defined at the subclinical level and includes also the traits of Subclinical Psychopathy (SP) and

Machiavellianism (Paulhus &Williams, 2002). The personality trait of SN includes facets

retained from the clinical syndrome namely, grandiosity, entitlement, dominance, and superiority

(Paulhus & Williams, 2002).

Previous studies suggested that SN might be unique among the DT traits in that, it

encapsulates to a larger extent (in comparison to psychopathy and Machiavellianism), prosocial

and adaptive behaviours (e.g. Veselka et al., 2012). For example, Petrides et al. (2011) reported

that—unlike psychopathy and Machiavellianism—narcissism correlates positively with

emotional intelligence. Furthermore, Sedikides, Rudich, Gregg, Kumashiro and Rusbult (2004)

conducted five studies and reported positive associations between SN and daily and dispositional

subjective well-being as well as couple well-being; and negative associations between SN and

daily sadness, dispositional depression, daily and dispositional loneliness, daily anxiety, and

dispositional neuroticism. Importantly, self-esteem fully accounted for the relation between SN

and psychological health suggesting that SN is beneficial for psychological health only when

associated with high self-esteem (Sedikides et al., 2004).

Within SN, studies have identified two main types: Grandiose Narcissism (GN) and

Vulnerable Narcissism (VN; Miller et al., 2011). GN is characterised by exhibitionism, lack of

humility/modesty, and interpersonal dominance. VN is characterised by negative affect, distrust,

selfishness, and a need for attention and recognition (Dickinson & Pincus, 2003; Vize, Lynam,
 5
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Collison & Miller, 2016). Research suggested that GN is linked mainly to positive outcomes,

while VN predicts negative outcomes (Besser & Zeigler-Hill, 2011).

In regards to the association between types of SN and PS, a study has shown that GN

predicted higher life satisfaction and lower PS. These relationships were mediated by coping

flexibility suggesting that, apart from holding positive self-views, grandiose narcissists are also

flexible in coping with stress, which make them psychologically healthier than vulnerable

narcissists, who exhibited the opposite pattern of associations (Ng, Cheung, & Tam, 2014). This

finding has been confirmed by a recent study, which reported that participants scoring high on

VN also reported higher DT traits, lower Mental Toughness (MT), poor sleep quality, and higher

scores on PS (Annen, Nakkas, Sadeghi Bahmani, Gerber, Holsboer-Trachsler, & Brand, 2017).

The authors suggested that a more fine-grained exploration of the different types of SN and their

association with further behavior may be necessary to shed light on the role of this highly

complex personality trait on symptoms of psychopathology (Annen et al., 2017).

1.2 Mental Toughness, Narcissism and Perceived Stress

Several other personality factors have been identified as buffers against the negative

impact of PS as well as various symptoms of psychopathology (Lin, Mutz, Clough, &

Papageorgiou, 2017). These include grit, defined as a perseverance and passion for long-term

goals (Duckworth, Peterson, Matthews, & Kelly, 2007); buoyancy, a trait characterized by

positive, constructive and adaptive response to challenges and setbacks (Martin & Marsh, 2008);

motivation, the desire to engage in behaviors for sheer enjoyment, challenge, pleasure, or interest

(Lepper, Corpus, & Iyengar, 2005); resilience, the ability to adapt effectively in situations of

adversity or stress (McGeown, St Clair-Thompson, & Clough, 2016); and hardiness a

personality trait describing an individual with three closely related tendencies: challenge,
 6
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

commitment, and control (Kobasa, 1979). Mental Toughness (MT) has been suggested as a

construct that may subsume the aforementioned concepts (see Sadeghi Bahmani et al., 2016).

MT reflects an effective coping mechanism as reaction to stressors and it allows

individuals to proactively seek out opportunities for personal growth (Lin, Clough, Welch, &

Papageorgiou, 2017; Mutz, Clough, & Papageorgiou, 2017; Papageorgiou, Mutz, Lin, & Clough,

2018). While MT shares some conceptual similarities with hardiness, it clearly differs in its

additional emphasis on confidence in one’s abilities and interpersonal relations. Individuals who

score high on MT are not only able to remain committed when confronting with stress, they are

also confident about successfully completing the task and are assertive in social situations. MT

is also distinct from grit, described by Duckworth and colleagues (2007) as perseverance and

passion for long-term goals. While individuals, who score high on grit, may work strenuously

toward goals despite self-doubt, mentally tough individuals believe that they are truly

worthwhile people and maintain the self-confidence to achieve goals. Furthermore, they are able

to control their emotion effectively in the face of setbacks and challenges. Clough et al., (2002)

characterised MT as a composite of four interrelated but independent components: (1) control

(life and emotion): the tendency to feel and act as if one is influential and keep anxieties in

check; (2) commitment: the tendency to be deeply involved in pursuing goals despite difficulties

that arise; (3) challenge: the tendency to see potential threats as opportunities for self-

development and to continue to strive in changing environments; and (4) confidence (in abilities

and interpersonal): the belief that one is a truly worthwhile person in spite of setbacks, and the

ability to push oneself forward in social settings.

In a longitudinal study, Gerber, Brand, et al. (2013) explored the relationships between

MT, PS, depressive symptoms, and life satisfaction. Levels of PS were assessed to provide an
 7
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

estimate of adverse life experiences; depressive symptoms and life satisfaction were assessed to

estimate overall levels of adjustment. In a sample of 865 students at vocational schools, both PS

and depressive symptoms correlated negatively with MT. Moreover, MT was positively

associated with life satisfaction. The researchers also found that well-adjusted individuals (low

levels of stress, few depressive symptoms, and high life satisfaction) scored high on MT,

whereas maladjusted individuals (high levels of stress, depressive symptoms, and little life

satisfaction) tended to have lower levels of MT. Resilient (moderate levels of stress at baseline,

decreased depressive symptoms and increased life satisfaction at follow-up) and deteriorated

(increasing levels of stress, increasing depressive symptoms, and decreasing life satisfaction)

individuals did not differ at baseline but showed an increase/decline of MT over time (resilient

and deteriorating individuals, respectively).

In line with these findings, Gerber, Kalak, et al. (2013) showed that MT was associated

with lower PS and fewer depressive symptoms in a sample of 284 high school students and in a

sample of 140 undergraduate students. They also showed that MT moderates the relationship

between high PS and depressive symptoms. More specifically, high levels of MT were

associated with lower depressive symptoms, when PS levels were high.

Finally, Gucciardi and Jones (2012) showed small to moderate negative correlations

between MT and PS, anxiety, and depression in a sample of 226 cricketers. This finding was

replicated by another study that reported negative correlations between MT and PS in a sample

of 217 international students (Jin & Wang, 2016).

Studies on the relationship between MT and SN have consistently reported positive and

moderate in size correlations (Onley et al., 2013; Sabouri et al., 2015; Papageorgiou, Wong, &
 8
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Clough, 2017; Papageorgiou et al., 2018). Papageorgiou et al. (2017) showed that, despite being

part of the dark triad, SN exerted a significant negative indirect effect on both Subclinical

Psychopathy (SP) and Machiavellianism through MT. Recently, Papageorgiou, Denovan and

Dagnall (2018) reported that SN contributes to significantly lower symptoms of depression,

through MT. Finally, a study used longitudinal data to show that SN at time 1 increases MT at

time 2 contributing indirectly to higher school grades in adolescents (Papageorgiou et al., 2018).

The authors concluded that the path model, from SN to MT to a further outcome, may predict

positive effects across various contexts including in the domain of education (e.g. higher school

grades) and psychopathology (e.g. lower symptoms of SP and symptoms of depression).

1.3 The Present Studies

The present investigation1 had a fourfold aim: (I) To test and replicate the correlations

between SN, MT and PS in three studies and in the combined dataset containing the data from all

three studies; (II) To test the correlations between GN and VN with MT and PS; (III) To test and

replicate in three studies a mediation model suggesting that SN increases MT resulting in

significantly lower levels of PS; (IV) To extend the sensitivity of the aforementioned mediation

model showing that, it is the GN that decreases PS, through MT, while VN exhibits the reverse

pattern. We hypothesized that: (1) SN will correlate positively with MT across all three studies;

(2) MT will correlate negatively with PS across all three studies and in the combined dataset; (3)

SN will correlate negatively with PS across all three studies and in the combined dataset; (4) SN

will exert a significant negative indirect effect on PS through MT across all three studies and in

the combined dataset; (5) GN will correlate negatively with PS (study 3); (6) VN will correlate

1
The studies were not pre-registered.
 9
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

positively with PS (study 3); (7) GN will exert a significant negative indirect effect on PS

through MT; while VN will exert a significant positive indirect effect on PS through MT in study

3. Progressive extension of the tested model provided an increasingly sophisticated

understanding of variable relationships.

2.0 Method
2.1 Sample
2.1.1 Study 1 and Study 2. Participants (N = 364 for study 1 and N = 240 for study 2)

were recruited online through advertisements on social networks (e.g. Facebook) as well as

through word of mouth. For study 1 participants’ mean age was 24.31 years (SD = 9.16, range =

18-79); 56.9% of the participants were females. For study 2 participants’ mean age was 25.43

years (SD = 7.21, range = 18-59); 64.60% of the participants were females. Participants received

detailed information regarding the aims of the studies and the voluntary basis of their

participation and they signed informed consent forms. The participants did not receive

compensation for taking part in the studies. We wanted to test the same hypotheses using

identical measures in two independent samples to be able to replicate the findings of study 1.

2.1.2 Study 3. Participants (N=144) were undergraduate students recruited through

advertisements in the class and around the university. They enrolled to participate in the study

through the university’s SONA system. Participants’ mean age was 22.08 years (SD = 5.5,

range = 18-52); 86.1% of the participants were females. Participants received detailed

information regarding the aim of the studies and they were asked to sign an informed consent

forms. Once they completed the study, participants received their credits (SONA points) as part

of a course fulfillment. They did not receive any other compensation for taking part in the study.

All three studies were reviewed and approved by the university’s ethics committee.
 10
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

2.1.3 Combined Dataset. The datasets derived from the three studies were combined to

form a large dataset of 748 participants. Sixty-five percent of the participants were females.

Their mean age was 24.30 years (SD = 8.09; range = 18 – 79 years). The sample size was

determined by reviewing previous research in this area of research and post-hoc power analyses.

We used the R package Monte Carlo Power Analysis for Indirect Effects (Schoemann, Boulton,

& Short, 2017) to estimate the power of the mediation effect for each study. This indicated that

the analyses were overpowered. We also estimated the minimum sample size that would be

required for 80% power for each dataset which indicated a sufficient sample would be

38 ≤ N ≤ 44 for the mediation effects found in the paper. Specifically:

Study 1: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power =

44; Study 2: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power = 43;

Study 3: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power = 38;

Combined Data: Power = 1.0 (rounded to 3 dp), minimum sample-size required for 80% power =

39.

2.2 Measures
2.2.1 Subclinical Narcissism (SN). The SD3 (Jones & Paulhus, 2014) assesses

subclinical narcissism, subclinical psychopathy and Machiavellianism and includes 27 items, 9

for each scale with responses given on a 5-point Likert scale, with 1 = strongly disagree and 5 =

strongly agree. Jones and Paulhus (2014) reported Cronbach’s alphas for the scale of narcissism

that ranged from .68 to .78 in three independent studies. We used only the 9 items that assess

subclinical narcissism. Example items include: “People see me as a natural leader”. The score

for the subscale represents the average score of the 9 corresponding items.
 11
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

The FFNI (Glover, Miller, Lynam, Crego, & Widiger, 2012) is a 148-item self-report

inventory designed to assess the basic elements of narcissism from the perspective of a 5-factor

model. The FFNI allows discriminating between the vulnerable (i.e., cynicism/distrust, need for

admiration, reactive anger, and shame) and grandiose (i.e., acclaim seeking, arrogance,

authoritativeness, entitlement, exhibitionism, exploitativeness, grandiose fantasies, indifference,

lack of empathy, manipulativeness, and thrill seeking) variants of narcissism. As such, it offered

the opportunity to test a model with higher sensitivity (i.e. to examine whether it is the grandiose

or vulnerable narcissism that predicts higher MT leading to lower symptoms of stress). In the

present study, we used the 60-item short form of the FFNI (Sherman et al., 2015). The average

completion time is 20 minutes and each item is scored on a 1 to 5 scale, where 1 = the statement

is false or that you strongly disagree and 5 = the statement is definitely true or you strongly

agree. Grandiose Narcissism (GN) derives by estimating the sum of the following subscales:

Indifference, Exhibitionism, Authoritativeness, Grandiose Fantasies, Manipulativeness,

Exploitativeness, Entitlement, Lack of Empathy, Arrogance, Acclaim Seeking, and Thrill Seeking.

Vulnerable Narcissism (VN) derives by estimating the sum of the following subscales: Reactive

Anger, Shame, Need for Admiration, and Distrust. Studies have shown that the grandiose and

vulnerable composites behave almost identically across the short- and long-form versions so that

the FFNI–Short Form (FFNI-SF) offers a well-articulated assessment of the basic traits

comprising GN and VN. In the original study Sherman et al. (2015) reported Cronbach’s alphas

that range from .70 for distrust to .89 for exploitativeness with a mean of .81.

2.2.2 Mental Toughness (MT). The Mental Toughness Questionnaire 48 (MTQ48) is the

most frequently used measure of MT as conceptualized by Clough et al. (2002). The MTQ48

has an average completion time of 10 minutes and responses to its 48 items are given on a 5-
 12
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

point Likert scale anchored at 1 = strongly disagree and 5 = strongly agree. A mean MT score

can be calculated by summing up individual items of the questionnaire and dividing by the total

number of items (48). Twenty-two items are reverse coded. Example items include "I can

usually adapt myself to challenges that come my way" and "I don’t usually give up under

pressure". A recent study that tested the psychometric properties of the MTQ48 reported an

overall omega value of .82 indicating high reliability (Vaughan, Hanna, & Breslin, 2017). Good

reliabilities were also shown in previous research for the MT subscales with alphas ranging from

.65 to .85 ( Perry, Clough, Crust, Earle, & Nicholls, 2013).

2.2.3 Perceived Stress (PS). The Perceived Stress Scale (PSS; Cohen, Kamarck, &

Mermelstein, 1983) was used to measure the degree to which respondents perceived their life as

unpredictable, uncontrollable, and overloading during the past month. The instrument has an

average completion time of less than 3 minutes, and responses to its 10 items are given on a 5-

point Likert scale (0 = never, 1 = almost never, 2 = sometimes, 3 = fairly often and 4 = very

often). Four items are reverse coded, and an overall perceived stress score can be obtained by

calculating the mean value of the scores that were reported for individual items of the

questionnaire. Example items include "In the last month, how often have you been upset because

of something that happened unexpectedly?" and "In the last month, how often have you felt

nervous and stressed?” In the original study, Cohen et al., (1983) reported Cronbach’s alphas

for the PSS that ranged from .84 to .86 in three independent samples.

2.3 Procedure
2.3.1 Study 1 and Study 2. All questionnaires were combined to form a single document

and they were made available online via SurveyMonkey (www.surveymonkey.com). Each

participant received a message containing a link to the online questionnaire and password access
 13
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

as well as a unique participant code. Questionnaire completion was self-paced, and participants

could proceed to the subsequent page only once they had answered all items. Participants had to

complete the survey in one session (i.e. they could not save their responses and return to

complete the questionnaires) and spent between 30 and 45 minutes to complete the survey. Upon

completion of the study the participants were given an online written debrief.

2.3.2 Study 3. All questionnaires were combined to form a single document; participants were

booking slots and they were coming in the class at the university to complete the paper and

pencil copies of the booklets. Each participant received his/her own booklet containing detailed

instructions for completing the study. Questionnaire completion was self-paced and participants

spent between 30 and 45 minutes to complete the booklet. Upon completion of the study the

participants were given a written debrief.

2.4 Statistical Analyses

The statistical analyses described below were conducted separately for study 1, study 2

and study 3. The results of the correlation and mediation analyses were also replicated in the

combined dataset that included all participants from the three studies.

2.4.1 Descriptive Statistics and Covariates. Kurtosis and skewness, were calculated to

test for normality in the distribution (skewness < 1.0). All variables were approximately

normally distributed. Cronbach’s alphas were calculated to estimate the measures’ internal

consistency (see Table 1 below for details).

The alpha value was set to .05 when testing for the possible effects of the variables age

and sex. Participants’ age did not correlate significantly with SN (r = -.05, p > .05; r = -.09, p >

.05; r = -.16, p > .05; for study 1, study 2 and study 3, respectively). Participants’ age did not
 14
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

correlate significantly with MT and PS (r = -.04, p > .05 and r = .02, p > .05 for study 3).

Participants’ age correlated positively with MT (r = .16, p < .01 and r = .13, p < .05 for study 1

and study 2, respectively) and negatively with PS (r = -.12, p < .05 and r = -.19, p < .01 for study

1 and study 2, respectively). Participants’ gender predicted MT and PS in study 1 (F = 6.038; p

= .003; η2 = .032; F = 4.207; p = .01; η2 = .02, respectively); SN, MT and PS in study 2 (F =

3.262; p = .04; η2 = .027; F = 3.290; p = .039; η2 = .027; F =11.674; p < .001; η2 = .09,

respectively); SN and GN in study 3 (F = 3.765; p = .02; η2 = .06; F = 6.90; p = .001; η2 = .10,

respectively); and SN and PS in the combined dataset (F = 1.536; p = .033; η2 = .06; F = 1.600;

p = .01; η2 = .08, respectively). For consistency, age and gender were used as a covariate in all

subsequent analysis.

2.4.3 Correlations and Mediation Analyses. Partial correlations (controlling for age)

were carried out to explore the amount of variance that is shared between SN, MT and PS. To

test the proposed mediation models, hierarchical regression analyses were performed using the

PROCESS macro for SPSS (Version 2.13; Hayes, 2012) in study 1, study 2, study 3 and in the

combined dataset (data derived from all three studies). PROCESS utilizes an ordinary least

squares path analytical framework to estimate direct, indirect, and total effects of mediation

models. The direct effect provides an estimate of the effect of the independent variable (IV) on

the dependent variable (DV). The indirect effect of the IV on the DV via a potential mediator

(M) can be estimated from bias-corrected bootstrap 95% confidence intervals. Confidence

intervals that do not contain zero give an indication of a significant mediation effect (Hayes,

2013). The total effect provides an estimate of the combined direct and indirect effects. In the

present study we used 5,000 bootstrap resamples as suggested by Preacher and Hayes (2008).
 15
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

The bootstrapping approach to estimating indirect effects is advantageous over traditional

procedures, as it does not rely on assumptions about the distribution of the indirect effect.

3.0 Results
3.1. Descriptive Statistics

Descriptive statistics and Cronbach’s alpha for SN, MT and PS are presented in Table 1

for all three studies and the combined dataset. Cronbach’s alpha for SN ranged from .67 (study

2) to .73 (study 3). Cronbach’s alphas for GN and VN were .92 and .82, respectively.

Cronbach’s alpha for MT ranged from .88 (study 1 and 2) to .92 (study 3). Cronbach’s alpha for

PS ranged from .84 (study 1) to .89 (study 2).

Table 1 should be placed here

3.2. Correlations between Subclinical Narcissism, Mental Toughness and Perceived Stress

Partial correlations (controlling for age and gender) between SN, MT and PS are

presented in Table 2. SN assessed using the SD3 scale, correlated positively with MT (r = .42, p

< .001, 95% CI [.32, .51]; r = .42, p < .001, 95% CI [.30, .53]; r = .42, p < .001, 95% CI [.25,

.56]; r = .44, p < .001, 95% CI [.37, .51] for study 1, 2, 3 and the combined dataset,

respectively).

SN assessed using the SD3 scale, correlated negatively with PS (r = -.18, p < .01, 95% CI

[-.31, -.04]; r = -.20, p < .001, 95% CI [-.33, -.06]; r = -.17, p < .001, 95% CI [-.25, -.08] for

study 1, 2 and the combined dataset, respectively). SN assessed using the SD3 scale, did not

correlate significantly with PS in study 3 (r = -.13, p = .17, 95% CI [-.29, .03]), although the

direction of the correlation was again negative.

 16
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

MT correlated negatively with PS (r = -.60, p < .01, 95% CI [-.66, -.52]; r = -.58, p <

.001, 95% CI [-.67, -.47]; r = -.66, p < .001, 95% CI [-.76, -.53]; r = -.57, p < .001, 95% CI [-.62,

-.52] for study 1, 2, 3 and the combined dataset, respectively).

Table 2 should be placed here

Partial correlations (controlling for age) between grandiose and vulnerable narcissism

(GN and VN, respectively) with SN (assessed using the SD3), MT and PS are presented in Table

3. Firstly, GN and VN did not correlate significantly (r = .008, p = .92, 95% CI [-.17, .19]). GN

correlated positively with SN (assessed using the SD3; r = .70, p < .001, 95% CI [.60, .78]) and

MT (r = .37, p < .001, 95% CI [.22, .50]). GN correlated negatively with PS but the correlation

was not significant (r = -.07, p = .47, 95% CI [-.22, .11]). VN correlated negatively with SN

(assessed using the SD3) but the correlation was not significant (r = -.08, p = .41, 95% CI [-.24,

.11]). VN correlated negatively with MT (r = -.64, p < .001, 95% CI [-.76, -.50]) and it correlated

positively with PS (r = .52, p < .001, 95% CI [.35, -.67]).

Table 3 should be placed here

3.3 Indirect Effects of Subclinical Narcissism on Perceived Stress through Mental

Toughness

Figure 1 to 4 below illustrates our proposed mediation model for study 1 to 3 and the

combined dataset, respectively. The mediation model suggests that SN (assessed using the SD3)

increases MT resulting in significantly lower PS. In line with this hypothesis, the indirect effect

of SN on PS through MT was negative and statistically significant for study 1, 2, 3 and the

combined dataset, respectively: (b = -2.80, SE = 0.46, 95% CI [-3.703, -1.902]; β = -.26, SE =

 17
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

.039, 95% CI [-.338, -.187]); (b = -3.38, SE = 0.67, 95% CI [-4.771, -2.119]); β = -.25, SE =

.050, 95% CI [-.358, -.160]); (b = -3.835, SE = 0.97, 95% CI [-5.859, -2.020]; β = -.31, SE =

.078, 95% CI [-.473, -.168]); (b = -3.269, SE = 0.358, 95% CI [-3.997, -2.604]; β = -.27, SE =

.028, 95% CI [-.327, -.216]).

Figure 1 should be placed here

Figure 2 should be placed here

Figure 3 should be placed here

Figure 4 should be placed here

Additional analysis was performed in the combined dataset containing only participants

(N = 353, mean age = 24.80, SD = 8.26, range = 18 – 79) that score below the 25 percentile

(scores ranged from 0 to 13) and above the 75 percentile (scores ranged from 24 to 40) in PS.

This analysis was performed to examine whether the proposed mediation model is particularly

relevant in explaining variation in PS in the quantitative extremes of the distribution (i.e.

individuals who score very low or very high on PS). Indeed, the indirect effect of SN on PS

through MT was negative and significantly stronger (b = -5.57, SE = 0.65, 95% CI [-6.929, -

4.378]; β = -.36, SE = .042, 95% CI [-.448, -.278]) in comparison to the indirect effect (b = -

3.269, SE = 0.358, 95% CI [-3.997, -2.604]; β = -.27, SE = .028, 95% CI [-.327, -.216])

observed, when testing the same model in the combined dataset containing all participants. The

results of this additional analysis are illustrated in Figure 5 below.

Figure 5 should be placed here

 18
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Finally, considering that all datasets included more female than male participants; and the

reported sex differences in dark traits (e.g. Jonason & Davis, 2018) we tested our proposed

mediation model for males and females, separately. We tested this in the combined dataset only

as this dataset in large enough to allow separating between males and females and still be able to

observe reliable associations. The indirect effect of SN on PS through MT in the male-only

sample (N = 258) was negative and significant: β = -.26, SE = .049, 95% CI [-.364, -.171]). The

indirect effect of SN on PS through MT in the female-only sample (N = 466) was negative and

significant: β = -.27, SE = .038, 95% CI [-.361, -.208]). As such, the results were very similar for

both males and females and almost identical to the results obtained, when using the combined

dataset (β = -.27, SE = .028, 95% CI [-.327, -.216]).

3.4 Indirect Effects of Grandiose and Vulnerable Narcissism on Perceived Stress through

Mental Toughness (Study 3)

Figures 6 below illustrates our proposed mediation model including GN and VN

(predictors), MT (mediator) and PS (outcome). The indirect effect of GN on PS through MT was

negative and statistically significant (b = -3.66, SE = 0.85, 95% CI [-5.438, -2.070]; β = -.26, SE

= .060, 95% CI [-.380, -.150]). The indirect effect of VN on PS through MT was positive and

statistically significant (b = 5.16, SE = 0.98, 95% CI [3.37, 7.24]; β = .45, SE = .084, 95% CI [-

.289, .621]). The omnibus indirect effect was negative and statistically significant (b = -6.42, SE

= 1.42, 95% CI [-9.31, -3.78]; β = -.36, SE = .080, 95% CI [-.530, -.216]).

Figure 6 should be placed here

4.0 Discussion
 19
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

It has been now more than 15 years since Paulhus and Williams (2002) coined the term

dark triad (DT) of personality to refer to the traits of SN, Subclinical Psychopathy (SP) and

Machiavellianism (see Furnham, Richards, & Paulhus, 2013). Accumulating evidence however

(see for example, Sedikides et al., 2004; Zeigler-Hill & Besser, 2011; Petrides et al., 2011;

Veselka et al., 2012; Onley et al., 2013; Ng et al., 2014; Sabouri et al., 2015; Papageorgiou et al.,

2017; Papageorgiou et al., 2018) suggests that when studying SN in relation to—not only the

other two dark traits—but also other personality traits, school performance and symptoms of

psychopathology, its inclusion in the malevolent side of human personality may need to be

reconsidered.

For example, previous cross-sectional and longitudinal research has shown that

Subclinical Narcissism (SN) may increase Mental Toughness (MT) resulting in positive

outcomes in the domains of psychopathology (e.g. lower scores on subclinical psychopathy;

Papageorgiou et al., 2017;) and education (e.g. higher school grades; Papageorgiou et al. 2018).

In a series of three studies we have shown that the path model, from SN to higher MT to a

positive outcome, is a reliable predictor of lower Perceived Stress (PS).

Specifically, confirming hypotheses 1 to 3, we replicated previous findings showing that

SN correlates positively with MT (hypothesis 1; see also Onley et al., 2013; Sabouri et al., 2015);

MT correlates negatively with PS (hypothesis 2; see also Gerber, Brand, et al., 2013; Gerber,

Kalak, et al., 2013); and that SN correlates negatively with PS in studies 1 and 2 and in the

combined dataset (hypothesis 3, see also Ng et al., 2014). With the only exception being the non-

significant (yet, as expected, negative) correlation between SN and PS in study 3—probably

because of the small sample size—all results were replicated with the direction and size of the

correlations being almost identical across all three studies and in the combined dataset.
 20
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Importantly, we showed that SN, assessed with the SD3 (Jones & Paulhus, 2014), may

increases MT resulting in significantly lower PS in three independent samples (Hypothesis 4).

The results were very stable (similar associations in size and direction) and they were replicated

in the combined dataset including data from all three studies. Additional analyses demonstrated

that the proposed mediation model is particularly relevant in explaining variation in the

quantitative extremes of the distribution on PS, hence differentiating reliably between individuals

that score very low and very high on PS. Considering previous findings on sex differences in the

dark traits (e.g. Jonason & Davis, 2018), we tested our proposed mediation model in two sub-

samples (derived from the combined dataset) of male-only vs. female-only participants. The

results in the two sub-samples were very similar, which indicates that, while sex differences in

SN are present, they do not appear to influence proposed mediation model.

Previous investigations (e.g. Papageorgiou et al., 2017 and Papageorgiou et al., 2018)

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 assessing the

antisocial aspects of narcissism. Maples, Lamkin and Miller (2014) also suggested that the SD3

narcissism scale measures primarily the grandiose aspects of this construct; while other short

measures of this trait, such as the Dirty Dozen (Jonason & Webster, 2010) might capture both

vulnerable and grandiose features of narcissism. To account for this bias, we included an

additional valid measure of SN, the 60-item short form of the FFNI (Sherman et al., 2015),

which differentiates between Grandiose SN and Vulnerable SN. The inclusion of this measures

offered the opportunity to gain a better insight into the role of SN in the proposed mediation

model.
 21
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Confirming our hypotheses, we showed that indeed, VN exhibited a moderate to strong

positive correlation with PS (hypothesis 6). VN also exerted a positive indirect effect on PS

through MT, while GN exerted a negative indirect effect on PS through MT (hypothesis 7).

Hypothesis 5 was only partially supported because GN correlated negatively with PS but the

correlation was not significant probably because of the small sample size of study 3. However,

GN did correlate positively and strongly with SN assessed using the SD3, confirming

suggestions from previous research (e.g. Maples, et al., 2014) that the SD3 assesses mainly the

grandiose, prosocial, aspects of SN.

This investigation shares well-reported limitations with other research in the domains of

personality and psychopathology, namely, the self-reported nature of the data that may be

influenced by common-method variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003); and

social desirability, particularly in the context of the assessment of a “dark” trait (narcissism).

The samples’ mean age was rather young across all three studies, which may limit the

generalizability of the findings to older samples. Furthermore, Study 3 was based completely on

a student sample that consists mainly of female participants, which limits the generalizability of

the findings to the general population. However, the present investigation establishes previous

findings and, most importantly, it provides new and counterintuitive insights into the role of a

seemingly “dark” trait in significantly reducing symptoms of psychopathology, through MT.

Considering the current (and previous) findings we suggest that simply including the SN

into the DT as a trait that links to poor and toxic psychosocial outcomes may need to be

revisited. This suggestion finds partial support also in a recent large meta-analysis and critical

review of the literature on the DT traits that, when controlling for SP and Machiavellianism,

failed to report any statistically significant correlation between SN and various measures of
 22
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

negative psychosocial outcome (with the exception of a weak positive correlation between SN

and interpersonal difficulties) such as, antisocial tactics, aggression, sex-related issues and

morality problems (Muris, Merckelbach, Otgaar, & Meijer, 2017). The authors argued however

that the way the DT traits are currently measured is too simple to capture the malevolent sides of

personality (Muris et al., 2017).

The findings of the current investigation support the view that SN is a particularly

complex personality trait involving both positive (grandiose) and negative (vulnerable) aspects

and exploring its relation to prosocial traits, such as MT, can be particularly helpful when trying

to identify and promote its adaptive tendencies. Studying the proposed path model from SN to

higher MT, while taking under consideration the distinction between GN as opposed to VN may

prove to be particularly useful in explaining variation in symptoms of psychopathology as well

as in other domains that span a number of human behaviors.

 23
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

References

Annen, H., Nakkas, C., Bahmani, D. S., Gerber, M., Holsboer-Trachsler, E., & Brand, S. (2017).

Vulnerable narcissism as key link between dark triad traits, mental toughness, sleep

quality and stress. European Psychiatry, 41, S261.

Besser, A., & Zeigler-Hill, V. (2011). Pathological forms of narcissism and perceived stress

during the transition to the university: The mediating role of humor styles. International

Journal of Stress Management, 18(3), 197.

Clough, P., Earle, K., & Sewell, D. (2002). Mental toughness: The concept and its measurement.

In I. M. Cockerill (Ed.), Solutions in sport psychology (pp. 32-43). Boston, MA: Cengage

Learning.

Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived

stress. Journal of health and social behavior, 385-396.

Dickinson, K. A., & Pincus, A. L. (2003). Interpersonal analysis of grandiose and vulnerable

narcissism. Journal of personality disorders, 17(3), 188-207.

Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: perseverance and

passion for long-term goals. Journal of personality and social psychology, 92(6), 1087.

Gerber, M., Brand, S., Feldmeth, A. K., Lang, C., Elliot, C., Holsboer-Trachsler, E., & Pühse, U.

(2013). Adolescents with high mental toughness adapt better to perceived stress: A

longitudinal study with Swiss vocational students. Personality and Individual Differences,

54(7), 808-814.

Gerber, M., Kalak, N., Lemola, S., Clough, P., Perry, J. L., Pühse, U., . . . Brand, S. (2013). Are

adolescents with high mental toughness levels more resilient against stress? Stress and

Health, 29(2), 164-171.

 24
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Glover, N., Miller, J. D., Lynam, D. R., Crego, C., & Widiger, T. A. (2012). The five-factor

narcissism inventory: A five-factor measure of narcissistic personality traits. Journal of

personality assessment, 94(5), 500-512.

Gucciardi, D. F., & Jones, M. I. (2012). Beyond optimal performance: Mental toughness profiles and

developmental success in adolescent cricketers. Journal of Sport and exercise

Psychology, 34(1), 16-36.

Hayes, A. F. (2013). Methodology in the social sciences. Introduction to mediation, moderation,

and conditional process analysis: A regression-based approach. New York, NY, US:

Guilford Press.

Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation,

moderation, and conditional process modeling.

Jin, L., & Wang, C. D. (2016). International students’ attachment and psychological well-being: The

mediation role of mental toughness. Counselling Psychology Quarterly. Advance online

publication. http://dx.doi.org/10.1080/09515070.2016.1211510

Jonason, P. K., & Webster, G. D. (2010). The dirty dozen: A concise measure of the dark

triad. Psychological assessment, 22(2), 420.

Jonason, P. K., & Davis, M. D. (2018). A gender role view of the Dark Triad traits. Personality

and Individual Differences, 125, 102-105.

Jones, D. N., & Paulhus, D. L. (2014). Introducing the short dark triad (SD3) a brief measure of dark

personality traits. Assessment, 21(1), 28-41.

Kobasa, S. C. (1979). Stressful life events, personality, and health: an inquiry into hardiness.

Journal of Personality and Social Psychology, 37(1), 1-11.

 25
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking “big” personality traits to

anxiety, depressive, and substance use disorders: a meta-analysis. Psychological

bulletin, 136(5), 768.

Lee, J. S., Joo, E. J., & Choi, K. S. (2013). Perceived stress and self‐ esteem mediate the effects

of work‐ related stress on depression. Stress and Health, 29(1), 75-81.

Lepper, M. R., Corpus, J. H., & Iyengar, S. S. (2005). Intrinsic and extrinsic motivational

orientations in the classroom: Age differences and academic correlates. Journal of

Educational Psychology, 97(2), 184-196.

Lin, Y., Clough, P. J., Welch, J., & Papageorgiou, K. A. (2017). Individual differences in mental

toughness associate with academic performance and income. Personality and Individual

Differences, 113, 178-183.

Lin, Y., Mutz, J., Clough, P. J., & Papageorgiou, K. A. (2017). Mental toughness and individual

differences in learning, educational and work performance, psychological well-being, and

personality: A systematic review. Frontiers in psychology, 8, 1345.

Maples, J. L., Lamkin, J., & Miller, J. D. (2014). A test of two brief measures of the dark triad:

The dirty dozen and short dark triad. Psychological assessment, 26(1), 326.

Martin, A. J., & Marsh, H. W. (2008). Academic buoyancy: Towards an understanding of

students' everyday academic resilience. Journal of school psychology, 46(1), 53-83.

McGeown, S. P., St Clair-Thompson, H., & Clough, P. (2016). The study of non-cognitive

attributes in education: Proposing the mental toughness framework. Educational

Review, 68(1), 96-113.

 26
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Miller, J. D., Hoffman, B. J., Gaughan, E. T., Gentile, B., Maples, J., & Keith Campbell, W.

(2011). Grandiose and vulnerable narcissism: A nomological network analysis. Journal

of personality, 79(5), 1013-1042.

Muris, P., Merckelbach, H., Otgaar, H., & Meijer, E. (2017). The malevolent side of human

nature: A meta-analysis and critical review of the literature on the dark triad (narcissism,

Machiavellianism, and psychopathy). Perspectives on Psychological Science, 12(2), 183-

204.

Mutz, J., Clough, P., & Papageorgiou, K. A. (2017). Do individual differences in emotion

regulation mediate the relationship between mental toughness and symptoms of

depression? Journal of Individual Differences.

Ng, H. K., Cheung, R. Y. H., & Tam, K. P. (2014). Unraveling the link between narcissism and

psychological health: new evidence from coping flexibility. Personality and Individual

Differences, 70, 7-10.

Nguyen‐ Michel, S. T., Unger, J. B., Hamilton, J., & Spruijt‐ Metz, D. (2006). Associations

between physical activity and perceived stress/hassles in college students. Stress and

Health: Journal of the International Society for the Investigation of Stress, 22(3), 179-

188.

Onley, M., Veselka, L., Schermer, J. A., & Vernon, P. A. (2013). Survival of the Scheming: A

Genetically Informed Link Between the Dark Triad and Mental Toughness. Twin

Research and Human Genetics, 16(06), 1087-1095.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method

biases in behavioral research: A critical review of the literature and recommended

remedies. Journal of applied psychology, 88(5), 879.

 27
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Papageorgiou, K. A., Denovan, A., Dagnall, N. (2018). The Positive Effect of Narcissism on

Depressive Symptoms through Mental Toughness: Narcissism may be a Dark Trait but it

does help with seeing the World Less Grey. European Psychiatry

Papageorgiou, K. A., Malanchini, M., Denovan, A., Clough, P. J., Shakeshaft, N., Schofield, K.,

& Kovas, Y. (2018). Longitudinal associations between narcissism, mental toughness and

school achievement. Personality and Individual Differences, 131, 105-110.

Papageorgiou, K.A., Mutz, J., Lin, Y., & Clough, P.J. (2018). Mental Toughness: A Personality

Trait That Is Relevant across Achievement Contexts and Mental Health Outcomes In

Zeigler-Hill, V., & Shackelford, T.K., The SAGE Handbook of Personality and

Individual Differences (pp. 587-604). Los Angeles, LA: Sage

Papageorgiou, KA., Wong, B., & Clough, P. (2017). Beyond Good and Evil: Exploring the

Mediating Role of Mental Toughness on the Dark Triad of Personality Traits.

Personality and Individual Differences, 119, 19-23.

Paulhus, D. L., & Williams, K. M. (2002). The dark triad of personality: Narcissism,

Machiavellianism, and psychopathy. Journal of research in personality, 36(6), 556-563.

Perry, J. L., Clough, P. J., Crust, L., Earle, K., & Nicholls, A. R. (2013). Factorial validity of the

Mental Toughness Questionnaire-48. Personality and Individual Differences, 54(5), 587-

592. doi:10.1016/j.paid.2012.11.020

Petrides, K. V., Vernon, P. A., Schermer, J. A., & Veselka, L. (2011). Trait emotional

intelligence and the dark triad traits of personality. Twin Research and Human

Genetics, 14(01), 35-41.

 28
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and

comparing indirect effects in multiple mediator models. Behavior research

methods, 40(3), 879-891.

Sabouri, S., Gerber, M., Bahmani, D. S., Lemola, S., Clough, P. J., Kalak, N., ... & Brand, S.

(2016). examining Dark Triad traits in relation to mental toughness and physical activity

in young adults. Neuropsychiatric disease and treatment, 12, 229.

Sadeghi Bahmani, D., Hatzinger, M., Gerber, M., Lemola, S., Clough, P. J., Perren, S., ... &

Brand, S. (2016). The origins of mental toughness–prosocial behavior and low

internalizing and externalizing problems at age 5 predict higher mental toughness scores

at age 14. Frontiers in Psychology. Advance online publication. doi:

10.3389/fpsyg.2016.01221

Schoemann, A. M., Boulton, A. J., & Short, S. D. (2017). Determining Power and Sample Size

for Simple and Complex Mediation Models. Social Psychological and Personality

Science, 8(4), 379–386.

Sedikides, C., Rudich, E. A., Gregg, A. P., Kumashiro, M., & Rusbult, C. (2004). Are normal

narcissists psychologically healthy?: Self-esteem matters. Journal of personality and

social psychology, 87(3), 400.

Sherman, E. D., Miller, J. D., Few, L. R., Campbell, W. K., Widiger, T. A., Crego, C., & Lynam,

D. R. (2015). Development of a Short Form of the Five-Factor Narcissism Inventory: The

FFNI-SF. Psychological Assessment, 27(3), 1110.

St Clair-Thompson, H., Bugler, M., Robinson, J., Clough, P., McGeown, S. P., & Perry, J.

(2015). Mental toughness in education: Exploring relationships with attainment,

attendance, behaviour and peer relationships. Educational Psychology 35(7), 886-907.

 29
POSITIVE EFFECT OF NARCISSISM ON STRESS THROUGH MENTAL TOUGHNESS

Steptoe, A., & Kivimäki, M. (2012). Stress and cardiovascular disease. Nature Reviews

Cardiology, 9(6), 360.

Tugade, M. M., & Fredrickson, B. L. (2004). Resilient individuals use positive emotions to

bounce back from negative emotional experiences. Journal of personality and social

psychology, 86(2), 320.

Vaughan, R., Hanna, D., & Breslin, G. (2017). Psychometric properties of the Mental Toughness

Questionnaire 48 (MTQ48) in elite, amateur and nonathletes. Sport, Exercise, and

Performance Psychology, 7(2), 128-140.

Veselka, L., Schermer, J. A., & Vernon, P. A. (2012). The Dark Triad and an expanded

framework of personality. Personality and Individual Differences, 53, 417–425.

Vize, C. E., Lynam, D. R., Collison, K. L., & Miller, J. D. (2016). Differences among dark triad

components: A meta-analytic investigation.

Walker, E. R., McGee, R. E., & Druss, B. G. (2015). Mortality in mental disorders and global

disease burden implications: a systematic review and meta-analysis. JAMA

psychiatry, 72(4), 334-341.

Widiger, T. A. (2011). Personality and psychopathology. World Psychiatry, 10(2), 103-106.

Table 1

Table 1. Descriptive Statistics for Subclinical Narcissism, Mental Toughness and Perceived Stress

Variable Name Mean Variance Median Range Kurtosis Skewness Cronbach’s α

STUDY 1

Subclinical Narcissism 2.93 (0.57) 0.32 2.88 3.22 -0.05 - 0.01 .70

Mental Toughness 3.44 (0.37) 0.14 3.45 2.19 0.24 - 0.32 .88

Perceived Stress 20.22 (6.16) 38.00 20.00 34.00 0.01 0.018 .84

STUDY 2

Subclinical Narcissism 2.89 (0.56) 0.32 2.88 3.22 -0.27 0.10 .67

Mental Toughness 3.42 (0.39) 0.15 3.42 2.35 0.58 0.07 .88

Perceived Stress 16.57 (7.76) 60.30 15.00 40.00 -0.10 0.43 .89

STUDY 3

Subclinical Narcissism (SD3) 2.53 (0.57) 0.33 2.55 2.78 -0.25 0.14 .73

Grandiose Narcissism 2.38 (0.52) 0.28 2.36 2.50 0.34 0.69 .92

Vulnerable Narcissism 3.19 (0.61) 0.37 3.25 2.88 -0.08 -0.27 .82

Mental Toughness 3.21 (0.41) 0.17 3.20 2.38 0.44 -0.28 .92

Perceived Stress 19.95 (7.31) 53.54 19.00 38.00 0.20 -0.07 .87

COMBINED DATASET

Subclinical Narcissism 2.85 (0.59) 0.34 2.89 4.00 -0.18 0.01 .71

Mental Toughness 3.40 (0.39) 0.15 3.42 3.00 0.44 -0.22 .91

Perceived Stress 18.97 (7.12) 50.69 19.00 40.00 -0.08 0.09 .87

Note: N =364; N = 240; N = 144; N = 748 for Study 1, Study 2, Study 3 and Combined Dataset, respectively. Standard
deviations are given in parentheses
Table 2

Table 2. Correlations between Subclinical Narcissism, Mental Toughness and Perceived Stress
Variable Name Subclinical Narcissism Mental Toughness Perceived Stress

Study 1

Subclinical Narcissism .42*** -20***

Mental Toughness -.60***

Study 2

Subclinical Narcissism .42*** -.18**

Mental Toughness -.58***

Study 3

Subclinical Narcissism (SD3) .41*** -.15

Mental Toughness -.67***

Combined Dataset

Subclinical Narcissism .44*** -.17***

Mental Toughness -.57***

Note: N = 364; N = 240; N = 144; N = 748 for Study 1, Study 2, Study 3 and Combined Dataset, respectively;
**p ≤ .01; ***p ≤ .001; covariates include age and sex.
Table 3

Table 3. Correlations
Variable Name between Grandiose and Vulnerable Narcissism
Grandiose Narcissismwith Narcissism
Mental(SD3) Mental
Perceived
Toughness and Perceived Stress (Study 3)
Narcissism (SD3) Toughness Stress

Grandiose Narcissism .70*** .37*** -.07

Vulnerable Narcissism .008 -.08 -.64*** .53***

Note: N = 144; ***p ≤ .001

Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 6

 Subclinical Narcissism predicted lower perceived stress through mental toughness in three

independent studies

 Grandiose narcissism predicted lower perceived stress through mental toughness

 Vulnerable narcissism predicted higher perceived stress through mental toughness

 The proposed mediation model is particularly relevant in explaining variation in the

quantitative extremes of the distribution on perceived stress

 Grandiose narcissism could be an important resilience factors against symptoms of

psychopathology
Figure 6

The Bright Side of Dark: Exploring the Positive Effect of Narcissism on Perceived Stress

through Mental Toughness

Kostas A. Papageorgiou, Foteini-Maria Gianniou, Paul Wilson, Giovanni B. Moneta, Delfina

Bilello, and Peter J. Clough

Author Note

Kostas A. Papageorgiou, School of Psychology, Queen’s University Belfast, UK; and

Tomsk State University, Tomsk, Russia; Foteini-Maria Gianniou, School of Psychology,

Queen’s University Belfast, UK; Paul Wilson, School of Psychology, Queen’s University

Belfast, UK; Giovanni B. Moneta, School of Social Sciences, London Metropolitan University,

UK; Delfina Bilello, School of Psychology, Queen’s University Belfast, UK; Peter J. Clough,

Department of Psychology, Huddersfield University, UK.

Correspondence concerning this article should be addressed to Kostas A. Papageorgiou,

School of Psychology, Queen’s University Belfast, United Kingdom; David Keir Building, 18-30

Malone Road, Belfast, BT9 5BN; email: [email protected]; phone: +44(0) 289097

5653

Conflict of interest

All authors declare no conflicts of interest.

This is a revised version of the manuscript: PAID-D-18-01485

Ms. Ref. No.: PAID-D-18-01485

Title: The Bright Side of Dark: Exploring the Positive Effect of Grandiose Narcissism on
Perceived Stress through Mental Toughness

Dear Professor Ronald R. Holden,

Thank you for your email dated 28th September 2018 and for the helpful comments from
the two reviewers. We were glad to read that the manuscript captured the reviewers’ interest. We
have carefully addressed the points made by the two reviewers in the document titled “Responses
to Reviewers” and in the revised manuscript, which we hope you will find sufficient to consider
for publication. The changes are presented in the document for each comment separately.

Sincerely,

Dr Kostas A. Papageorgiou