DARK TRIAD AMONG AT-RISK YOUTHS 1
Running-head: DARK TRIAD AMONG AT-RISK YOUTHS
Dirty Dozen: A concise measure of Dark Triad traits among at-risk youths
Pedro Pechorro *
University of Minho, Braga, Portugal
Peter K. Jonason
Western Sydney University, Sydney, Australia
Lusófona University of Humanities and Technologies, Lisbon, Portugal
William James Centre for Research, ISPA – Instituto Universitário, Lisbon, Portugal
* Address corresponding to Pedro Pechorro. School of Psychology, University of
Minho, Portugal. Postal address: Campus de Gualtar, 4710-057 Braga, Portugal. Phone:
+351 253601397. Email: firstname.lastname@example.org
Conflict of Interest: The authors declare that they have no conflict of interest.
This study was partially funded by the Portuguese Foundation for Science and
Technology (FCT; Grant SFRH/BPD/86666/2012).
DARK TRIAD AMONG AT-RISK YOUTHS 2
The Dark Triad is a term used to describe a constellation of three socially undesirable
personality traits: psychopathy, narcissism, and Machiavellianism. The aim of the present study
was to examine the psychometric properties of the Dirty Dozen scale, a brief measure of the
Dark Triad, among a sample of 412 at-risk youths from Portugal (M = 13.19 years; SD = 1.41;
age range = 12-17 years), composed of male (n = 200; M = 13.32 years; SD = 1.41 years) and
female youths (n = 212; M = 13.08 years; SD = 1.41 years). As expected, our translation had a
three-factor structure that was invariant across the sexes. Adequate psychometric properties
were demonstrated in terms of internal consistency (i.e., Cronbach´s alpha and Omega
coefficient), convergent validity (i.e., with measures of self-reported delinquency, and sensation
seeking), discriminant validity (i.e., with measures of self-esteem, and self-control), criterion-
related validity (i.e., with drug use, and risky sex), and know-groups validity (boys versus girls).
We also found that sex differences in self-reported delinquency were partially mediated by sex
differences in the Dark Triad traits. Our findings suggest the Dirty Dozen is a valid and useful
measure in the study of at-risk for delinquency adolescents.
Keywords: Assessment; Dirty Dozen; Dark Triad; Youth; Validation
DARK TRIAD AMONG AT-RISK YOUTHS 3
Recently there has been an explosion of interest in the Dark Triad traits (i.e.,
psychopathy, narcissism, and Machiavellianism) which are three related, socially undesirable
personality traits (Paulhus & Williams, 2002). Psychopathy manifests in emotional callousness
and remorselessness, narcissism is related with egotism and a grandiose view of the self, and
Machiavellianism is characterized by manipulation of other people and an absence of
internalized morality. This interest in Dark Triad traits compliments most work in personality
psychology on the Big Five traits (Furnham, Richards, & Paulhus, 2013). One reason for this
increased in interest has been the development of concise measures of the traits, one of which is
the Dirty Dozen measure (Jonason & Webster, 2010). This 12-item measure has considerable
psychometric evidence to support its utility in English (Maples, Lamkin, & Miller, 2014), Polish
(Czarna, Jonason, Dufner, & Kossowska, 2016), German (Küfner, Dufner, & Back, 2015),
Japanese (Tamura, Oshio, Tanaka, Masui, & Jonason, 2015), Turkish (Özsoya, Rauthmann,
Jonason, & Ardıç, 2017), French-Canadian (Savard, Simard, & Jonason, 2017), Italian (Chiorri,
Garofalo, & Velotti, in press; Schimmenti et al., in press), and Spanish (Pineda, Sandín, &
Muris, in press) speaking samples.
There is a wider range of variables of interest when studying the Dark Triad traits and
cause to examine the utility of this scale in other countries (Jonason, Li, & Czarna, 2013). For
example, the Dark Triad traits are correlated with “risky” sexual behavior (Jonason, Li,
Webster, & Schmitt, 2009), drug use (Gott & Hetzel-Riggin, 2018), impulsivity (Jones &
Paulhus, 2011), limited self-control (Jonason & Tost, 2010), and petty theft (Lyons & Jonason,
2015), all of which may manifest themselves in adolescent samples (Chabrol, Van Leeuwen,
Rodgers, & Séjourné, 2009; Klimstra, Sijtsema, Henrichs, & Cima, 2014; Lau & Marsee, 2013)
and have relevance to various psychosocial outcomes.
The existing literature suggests that Dark Triad traits, in general, consistently tend to be
more prevalent in men than in women (e.g., Carter, Campbell, Muncer, & Carter, 2015; Czarna
et al., 2016; Furnham & Trickey, 2011). Life history models of the Dark Triad traits (Jonason,
Koenig, & Tost, 2010) suggest that men may be better characterized by and benefited more over
evolutionary time from engaging in selfish and antisocial approaches to life. In contrast, while
DARK TRIAD AMONG AT-RISK YOUTHS 4
women may reap some rewards from engaging in Dark Triad style behavior, they also are more
likely to suffer more costs than men are from both social and reproductive sources. As such,
several studies have shown that the Dark Triad traits facilitate (i.e., mediate) sex differences in
important outcomes like interest in casual sex (see Jonason et al., 2009). If we apply this to
delinquency behaviors in youths, we expect that boys will be better characterized by the Dark
Triad traits and delinquency traits than girls do and that the former traits will mediate sex
differences in the latter. In addition, to better ensure that these sex differences and mediation
effects are trustworthy, we replicate them after testing for measurement invariance (Millsap &
Research on the Dark Triad traits, however, has focused almost exclusively on
participants in college or adult community members. As far as we can tell, there have been six
investigations of these traits in individuals under the age of 18 years (see below). Most of the
studies assessing Dark Triad traits among school and community samples of youths found
unique positive associations between each of these traits and various forms of delinquent
behavior (Chabrol et al., 2009; Wright et al., 2017), aggression (Kerig & Stellwagen, 2010;
Muris, Meesters, & Timmermans, 2013), emotional/behavioral dysregulation (Lau & Marsee,
2013), and risk-taking (Malesza & Ostaszewski, 2016). Regarding the psychometric properties
of the Dirty Dozen measure, we are only aware of one study using a school sample of Dutch
youths (Klimstra et al., 2014) that provided strong evidence for the validity, reliability, and
measurement invariance of data collected with this measure, and also showed the unique
associations of each Dark Triad trait with particular forms of aggression.
The current study examines the psychometric properties and utility of the Dark Triad
Dirty Dozen scale (Jonason & Webster, 2010) among a sample of at-risk male and female
Portuguese youths. This is the first study examining the Dark Triad traits among Portuguese
youths and one of the few examining such traits in people under 18 years of age. We predicted
that the Dirty Dozen would (1) confirm the presumed three-factor structure and demonstrate
cross-sex invariance; (2) demonstrate adequate internal consistency; (3) show convergent
validity with measures of youth delinquency and sensation seeking, and discriminant validity
DARK TRIAD AMONG AT-RISK YOUTHS 5
with measures of self-esteem and self-control; (4) show criterion-related validity with outcomes
like drug use and risky unprotected sex; (5) show in terms of known-groups validity that boys
will score higher than girls; and (6) demonstrate that sex differences in self-reported
delinquency are mediated by individual differences in Dark Triad traits.
We used a convenience sample of volunteers who were recruited from public schools
managed by the Portuguese Ministry of Education located in disadvantaged and impoverished
zones with high rates of criminality in the greater Lisbon area. The population that resides in
these zones tends to suffer from social exclusion, and has limited access to resources and
economic opportunities. A school sample — aged 12 to 17 — of 412 participants (Mage = 13.19
years; SDage = 1.41), was composed of male participants (n = 200; Mage = 13.32 years; SDage =
1.41 years) and female participants (n = 212; Mage = 13.08 years; SDage = 1.41 years).
The participants were white Europeans (74.9%) and members of ethnic minorities
(25.1%; e.g., black Africans, mixed race from South-America) from an urban background. Most
had a low socioeconomic status (88.6%) and had completed an average of six years of education
(M = 5.78, SD = .94, range = 4–10 years). About 18.3% of the participants reported getting into
problems with the law in the last 12 months and three participants reported having been
previously incarcerated into juvenile detention centers.
The Dirty Dozen (DD; Jonason & Webster, 2010) is a brief measure of the Dark Triad
traits and is composed of subscales of Machiavellianism (e.g., “I have used deceit or lied to get
my way.”), Psychopathy (e.g., “I tend to lack remorse.”), and Narcissism (e.g., “I tend to want
others to admire me.”) dimensions. Participants completed the 12 items by denoting their
agreement (Strongly disagree = 1; Strongly agree = 5). Internal consistency values for the
current Portuguese adaptation of this measure are reported in Table 5.
We used the Self-Report Delinquency (SRD; Elliott, Ageton, & Huizinga, 1985), which
was created for the original National Youth Survey (NYS), a longitudinal study of delinquent
DARK TRIAD AMONG AT-RISK YOUTHS 6
behavior among American youth. The scale was developed with the intention of including items
(e.g., “Been involved in gang fights.”) that were representative of the full range of acts for
which juveniles could be arrested and involved a recall period of one year. The scale can be
scored by adding the 24 item scores on a 9-point ordinal scale (Never = 1; Two-Three times a
day = 9). Higher scores indicate higher levels of juvenile delinquency. The Portuguese version
of the scale (Pechorro, Lima, Simões, & DeLisi, in press) was used with a 5-point ordinal scale.
Item scores were summed to create an index of self-reported delinquency (Cronbach’s α = .91).
The Add Health Self-Report Delinquency (AHSRD) was designed for the National
Longitudinal Study of Adolescent Health (Add Health), a prospective study of American
adolescents in the seventh through the twelfth grade (Udry, 2003). The scale can be scored by
adding the 10 item scores of the Non-violent factor (e.g., “Take something from a store without
paying for it.”) and the seven item scores of the Violent factor (e.g., “Pull a knife or a gun on
someone.”) on a 4-point ordinal scale (None = 0; Five or more times = 3). Higher scores
indicate higher levels of juvenile delinquency. The Portuguese version of the scale was used
(Pechorro, Moreira, Basto-Pereira, Oliveira, & Ray, in press). Item scores were summed to
create an index of self-reported delinquency (α = .90).
The Brief Sensation Seeking Scale (BSSS; Hoyle, Stephenson, Palmgreen, Lorch, &
Donohew, 2002) was used as a measure of individual differences in the four basic facets of
sensation seeking, namely: Experience Seeking, Boredom Susceptibility, Thrill and Adventure
Seeking, and Disinhibition. The scale can be scored by adding the eight item scores (e.g., “I like
wild parties.”) on a 5-point ordinal scale (Strongly disagree = 1; Strongly agree = 5). Higher
scores indicate higher levels of sensation-seeking. The Portuguese version of the scale was used
(Pechorro, Castro, Hoyle, & Simões, 2018). Item scores were summed to create an index of
sensation-seeking (α = .83).
The Brief Self-Control Scale (BSCS; Tangney, Baumeister, & Boone, 2004) is a short
self-report measure of general self-control. The scale can be scored (after reverse scoring the
appropriate items) by adding the item scores (e.g., “I am good at resisting temptation.”) on a 5-
point ordinal scale (Not at all like me = 0; Very much like me = 4). Higher scores indicate higher
DARK TRIAD AMONG AT-RISK YOUTHS 7
levels of self-control. The Portuguese version of the scale was used (Pechorro, Pontes, DeLisi,
Alberto, & Simões, in press). Item scores were summed to create an index of self-control (α =
The Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1989) was used to assess juvenile
self-esteem. The scale can be scored by adding the 10 item scores (e.g., “I take a positive
attitude toward myself.”) on a 4-point ordinal scale (Strongly Disagree = 0; Strongly Agree = 3)
after reverse scoring the appropriate items. Higher scores indicate higher levels of self-esteem.
The Portuguese version of the scale was used (Pechorro, Marôco, Poiares, & Vieira, 2011). Item
scores were summed to create an index of self-esteem (α = .77).
An ad hoc questionnaire was designed to describe the participants’ socio-demographic
characteristics, including variables such as age, sex, ethnicity, years of education completed,
socioeconomic status, tobacco use, alcohol use, drug use, and having risky unprotected sex
(these last variables coded as five point ordinal variables from 0 = Almost never/Never, to 4 =
Almost always/Always). Socioeconomic status was measured by taking into consideration both
the parental level of education and the parental profession, appropriate to the Portuguese reality
(Simões, 2000). Three levels of socioeconomic status were considered: low, middle, and high.
The standard procedure of back-translation was used (van de Vijver, 2016). The
translation from English into Portuguese was completed by the first and third authors of this
article, taking into consideration a previous translation of the Dirty Dozen for use with
adolescents (Muris et al., 2013) and also previous translation procedures of Dark Triad traits
measures (e.g., Malesza, Ostaszewski, Büchner, & Kaczmarek, in press). The back-translation
into English being completed by a professional native English speaker translator. Discrepancies
were revised until a consensus was reached and no semantic differences were detected between
the English version and the Portuguese version of the Dirty Dozen. We strived to use an easy
but correct grammar. Then this version of the translation was piloted tested on a small group of
youths not included in the present investigation to make sure they could properly understand the
DARK TRIAD AMONG AT-RISK YOUTHS 8
meaning of the items. This way we made sure that the wording of the items was adequate to the
comprehension and reading level of the participants.
The General Directorate of Education of the Portuguese Ministry of Education (DGE-
ME) granted the permission to assess the participants coming from three public schools of the
greater Lisbon area. The participants were informed about the nature of the study and asked to
voluntarily collaborate. The participation rate was around 86%. Approximately equal
proportions of participants were selected from each of the three schools, and the participants’
sociodemographic characterization was considered quite similar. Some youths did not agree to
collaborate or were not able to collaborate (e.g., refused to participate, did not have
authorization from their legal tutors, had low reading ability). The measures were administered
in small group settings. Parental authorization to participate was mandatory and was previously
obtained. Informed consent was obtained from all participants.
Data were analyzed using SPSS v25 (IBM Corp., 2017) and EQS 6.3 (Bentler & Wu,
2015). The factor structure was assessed with Conﬁrmatory Factor Analysis (CFA) performed
in EQS 6.3 with the robust estimation methods. CFA’s were performed on the ordinal items and
items with standardized loadings below .30 were excluded. Goodness-of-ﬁt indices were
calculated to assess the different models tested. A Satorra-Bentler chi-square (S-Bχ²)/degrees of
freedom value < 5 was considered acceptable, a value ≤2 was considered good, and a value of =
1 was considered very good (Blunch, 2016; Maroco, 2014). A CFI ≥ .90 and RMSEA ≤ .08
indicated adequate fit whereas a CFI ≥ .95 and RMSEA ≤ .06 indicated good model fit. IFI
values of ≥ .90 were considered acceptable, and the model with the lowest AIC value should be
selected. No modification indices were used. Measurement invariance across participant sex
was investigated by testing configural, metric, and scalar invariance. Prior research suggests that
changes in CFI equal to or less than .01, and changes in RMSEA of equal to or less than .015
provides evidence in support of invariance (Chen, 2007; Cheung & Rensvold, 2002).
ANOVAs were used to examine differences between the male and female groups,
including the partial Eta squared (ηp2) effect size. Chi-square tests were used to compare
DARK TRIAD AMONG AT-RISK YOUTHS 9
nominal variables, including the phi (Φ) effect size. Cronbach’s alpha (α) and omega (ω)
coefficients (considered satisfactory if above .70), mean inter-item correlations (MIIC;
considered good if within the .15-.50 range), and corrected item-total correlation ranges
(CITCR; considered adequate if above .30) were used to assess reliability (Blunch, 2016; Dunn,
Baguley, & Brunsden, 2014). Correlations and partial correlations were used to evaluate
convergent, discriminant validity, and criterion-related validity (Finch, Immekus, & French,
2016). Correlations were considered low if below .20, moderate if between .20 and .50, and
high if above .50 (Ferguson, 2009). Fisher’s z (2-tailed) test was used to test differences
between two independent correlation coefficients. We also examined the intercorrelations
among the total score and the factor scores in order to further examine its construct validity.
The approach to mediation developed by Baron and Kenny (1986) was followed to test
whether sex differences in delinquency might be accounted for (i.e., mediated) by sex
differences in the Dark Triad traits given that there was reasonably normally distributed data
and a sufﬁciently large sample size (Fritz, Kenny, & MacKinnon, 2016). The PROCESS Macro
(Hayes, 2012, 2013) software was used, with bootstrapping being done with 5,000 samples.
In the initial stage of data analysis (Table 1), the socio-demographic variables were
examined taking into account the sex of the participants. No differences were found between
males and females regarding age [F(1, 410) = 2.74; p = .10; ηp2 = .01], years of education (F(1,
410) = 3.44; p = .06; ηp2 = .01)], ethnicity [χ2(1)= 0.00, p = 1, Φ = .00]; nationality [χ2(1)= 3.16,
p = .08, Φ = .09], and SES [χ2(2)= 3.26, p = .25, Φ = .09].
[Insert Table 1 about here]
Regarding the psychometric properties of our translation of the Dirty Dozen into
Portuguese using confirmatory factor analysis we found that a three-dimensional model fit the
data best (see Table 2; Figure 1) and that there was measurement invariance across the sexes
DARK TRIAD AMONG AT-RISK YOUTHS 10
with the ΔS-B χ2(df) being non-significant (see Table 3). We also report the correlations
between the traits in men, women, and overall (see Table 4).
[Insert Tables 2, 3, 4, and Figure 1 about here]
In terms of Cronbach’s α, Ω coefficients, mean inter-item correlations, and corrected
item-total correlations (see Table 5), we found sufficient-to-good internal consistency (Schmitt,
1996). However, the Machiavellianism and Psychopathy dimensions failed to meet the .30
threshold for the corrected item-total correlations, indicating that the associations between the
items of each of these dimensions were weaker than expected.
[Insert Table 5 about here]
Next, we examined the nomological network surrounding the Dirty Dozen measure in
Portuguese youth. Table 6 contains the convergent validity with measures of self-reported
delinquency and sensation-seeking, and the discriminant validity with self-esteem and self-
control. Using Fisher’s z (2-tailed) tests for independent correlations, we compared the
correlations in boys and girls displayed in Table 6 and found evidence that some correlations
were significantly different across sex (e.g., the correlations between narcissism and
delinquency stronger in males than in females). Using partial correlations to control for the
remaining dimensions of the Dark Triad trait (i.e., controlling for shared variance) among the
total sample we found some unique associations (e.g., the three dark triad traits showed mostly
similar statistically significant associations with delinquency).
[Insert Table 6 about here]
We also examined the criterion-related validity. Table 7 presents the correlations with
tobacco, alcohol, cannabis use, cocaine/heroin use, and unprotected sex. Fisher’s z (2-tailed)
tests for independent correlations indicated that some correlations were significantly different
across sex (e.g., the correlation between narcissism and substance use was stronger in males
than in females). Using partial correlations to control for the remaining dimensions of the Dirty
Dozen among the total sample, we found some unique associations (e.g., Machiavellianism was
the only dimension associated with alcohol use).
[Insert Table 7 about here]
DARK TRIAD AMONG AT-RISK YOUTHS 11
In terms of known-groups validity boys obtained higher scores on Machiavellianism
[t(353.71) = -3.40, p < .01, Cohen’s d for effect size = 0.36; Mboys = 5.83, SDboys = 1.88; Mgirls =
5.27, SDgirls = 1.35], psychopathy [t(361.23) = -2.13, p < .05, d = 0.22; Mboys = 6.92, SDboys =
3.01; Mgirls = 6.35, SDgirls = 2.26], and narcissism scores [t(339.35) = -3.82, p ≤.001, d = 0.42;
Mboys = 6.68, SDboys = 3.05; Mgirls = 5.68, SDgirls = 2.01].
In terms of the mediation analysis we tested each Dark Triad trait independently, and
we found three instances of partial mediation. The relationship between sex and delinquency
[Total Effect = 2.96, SE = 0.48, t = 6.14, p < .001, 95% CI (2.01, 3.91)] was partially mediated
by Machiavellianism as the direct effect between sex and delinquency remained significant after
Machiavellianism was added [Direct Effect = 2.28, SE = 0.45, t = 5.12, p < .001, 95% CI (1.41,
3.16)], and the indirect effect of Machiavellianism were significant [Indirect Effect = 0.68, Boot
SE = 0.25, 95% CI (.28, 1.29), p<.05]. It was partially mediated by psychopathy as the direct
effect between sex and delinquency remained significant after psychopathy was added [Direct
Effect = 2.72, SE = 0.47, t = 5.76, p < .001, 95% CI (1.79, 3.65)], and the indirect effect of
psychopathy were significant [Indirect Effect = 0.24, Boot SE = 0.13, 95% CI (.04, .53), p<.05].
Finally, it was also partially mediated by narcissism as the direct effect between sex and
delinquency remained significant after narcissism was added [Direct Effect = 2.35, SE = 0.46, t
= 5.07, p < .001, 95% CI (1.43, 3.26)], and the indirect effect of Narcissism were significant
[Indirect Effect = 0.61, Boot SE = 0.24, 95% CI (.26, 1.24); p<.05].
In this study we aimed to assess the psychometric properties of the Dirty Dozen among
a Portuguese sample of at-risk male and female youths. The three-dimensional model that fit the
data best was like the original English version (Jonason & Webster, 2010). Consistent with
previous research our findings suggest that the three Dark Triad traits are separable entities in
adolescents like they are in adults (Klimstra et al., 2014). Metric and scalar measurement
invariance was also demonstrated in both sexes. This is quite relevant because scale validity is
DARK TRIAD AMONG AT-RISK YOUTHS 12
ultimately the most important feature of scale construction and validation (Hamby, Ickes, &
Babcock, 2016). This leads us to conclude that our first hypothesis was confirmed.
We found acceptable internal consistency values in terms of Cronbach’s alpha for the
Portuguese translated version of the Dirty Dozen that was equivalent in each sex, but again our
values were lower the ones reported by Klimstra et al. (2014). The Omega coefficients, which
are better true reliability estimators than alpha, presented considerably higher values (Dunn et
al., 2014). The intercorrelations between the Dark Triad traits, however, were not as high as the
ones reported by Klimstra et al. (2014) among youths from the Netherlands, possibly caused by
sampling error as per our at-risk Portuguese sample. Nevertheless, our results provide overall
evidence for the psychometric validity and reliability in boys and girls. This means that not only
does the scale return similar properties as translations in other languages, it paves the way for
potential cross-cultural studies and presents of picture of universality in the nature of these traits
in measurement properties. As such, our second hypothesis was also confirmed.
We also made a wide-scale nomological network assessment to evaluate the validity of
the translation of the Dirty Dozen. We found good validity for the translation through the
assessment of the nomological network surrounding the Portuguese translation among youths.
Focusing on the partial-correlations among the total sample the convergent validity with self-
reported delinquency and sensation-seeking was generally demonstrated. The three dimensions
of the Dirty Dozen showed mostly similar positive associations with self-reported delinquency,
with narcissism showing slightly lower values. This is consistent with the recent finding that the
nomological networks of psychopathy and Machiavellianism overlap substantially (Vize,
Lynam, Collison, & Miller, 2018). The narcissism and Machiavellianism dimensions showed
the strongest associations with sensation-seeking consistent with some previous research
(Crysel, Crosier, & Webster, 2013) even after controlling for the remaining dimensions of the
Dirty Dozen. The discriminant validity with self-esteem and self-control was generally
demonstrated. In line with previous research, negative correlations were found between
psychopathy and self-esteem (Pechorro et al., 2014) and between Machiavellianism and self-
control (Jonason & Tost, 2010). Criterion-related validity with tobacco, alcohol, cannabis use,
DARK TRIAD AMONG AT-RISK YOUTHS 13
cocaine/heroin use, and unprotected sex was also demonstrated as expected (e.g., Flexon,
Meldrum, Young, & Lehmann, 2016; Jonason, Koenig & Tost, 2010). After controlling for the
remaining dimensions of the Dirty Dozen, significant associations were found between
narcissism and tobacco, cannabis, and cocaine/heroin use; between Machiavellianism and
alcohol use; and between psychopathy and risky unprotected sex. Thus, our third and fourth
hypotheses were supported.
The sex of the participant played in important role in understanding the Dark Triad
traits. Sex differences in Dark Triad traits were consistent with most previous studies among
adults and youths, with males tending to score higher than females (e.g., Carter et al., 2015;
Czarna et al., 2016; Furnham & Trickey, 2011). It is worth mentioning that Cohen’s effect size
d was lower for psychopathy than for Machiavellianism and narcissism, which might be
attributed to the at-risk characteristics of our sample, that is, a higher effect size for psychopathy
could be expected if our sample was a forensic sample. Our findings regarding these sex
differences are supported by the measurement invariance previously reported that excludes
measurement error. There was also clear evidence that boys tended to present the strongest
correlations, namely of narcissism with tobacco, alcohol, cannabis, and cocaine/heroin; of
psychopathy with alcohol, cannabis, and unprotected sex; and of Machiavellianism with
alcohol. The correlations of narcissism and Machiavellianism with delinquency were stronger in
boys, while the correlations of psychopathy with sensation-seeking, self-esteem, and self-
control were stronger in girls. Therefore, our fifth hypothesis was also confirmed.
Finally, sex differences in delinquency were mediated by a latent Dark Triad factor,
suggesting that the psychological mechanisms behind sex differences in the former are a partial
function of individual differences in the latter. This highlights the need of different theoretical
models for males and females and potential sex-specific targeted interventions based on the
assessment of Dark Triad traits. This leads us to conclude that the last hypothesis was also
Limitations and future research
DARK TRIAD AMONG AT-RISK YOUTHS 14
Despite the use of data from adolescents in a non-English speaking sample and the
broad cross-section of potential correlates, our study was characterized by several limitations.
First, this convenience sample could still be described as mostly white and coming from a
democratic and industrialized country (Henrich, Heine, & Norenzayan, 2010). Second, internal
consistency estimates mostly passed the standard (i.e., .70) threshold (Dunn et al., 2014); a few
only passed the more liberal threshold (i.e., .50) as set out for basic research (Schmitt, 1996),
but on average we had adequate internal consistency in our scales. Third, the cross-sectional
nature of the current study did not allow for an examination of the stability of the features
assessed over time (i.e., temporal stability). Fourth, additional psychometric procedures should
be done in the future (e.g., cross-validation using other samples, convergent validity with other
Dark Triad measures). Fifth, the fact that the measures used in the current study where
presented in the self-report format may be problematic in terms of shared methods variance (i.e.,
variance may be attributable to the measurement method rather than to the constructs that the
measures are assumed to represent). Despite these limitations, our ﬁndings provides support for
use of the Dirty Dozen with at-risk youths, along with its use across different samples and
Overall, findings suggest that the Portuguese version of the Dirty Dozen provides a
good overall assessment of the Dark Triad constructs. However, further psychometric research
is needed (e.g., cross-validation, test–retest reliability) to arrive at more concrete conclusions.
Nonetheless, the present study corroborates prior research (e.g., Klimstra et al., 2014),
suggesting that it is, indeed, possible to assess the Dark Triad traits among youth populations
using a concise, 12-item measure which has favorable psychometric properties and predictive
DARK TRIAD AMONG AT-RISK YOUTHS 15
Ethical approval: “All procedures performed in studies involving human participants were in
accordance with the ethical standards of the institutional and/or national research committee and
with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.”
On behalf of all authors, the corresponding author states that there is no conflict of interest.
DARK TRIAD AMONG AT-RISK YOUTHS 16
Baron, R., & Kenny, D. (1986). Moderator-mediator variables distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173-1182. doi: 10.1037/0022-35188.8.131.523
Bentler, P., & Wu, E. (2015). Supplement to EQS 6.3 for Windows user’s guide. Temple City,
CA: Multivariate Software.
Blunch, N. (2016). Introduction to Structural Equation Modeling Using IBM SPSS Statistics
and EQS. Thousand Oaks, CA: SAGE Publications.
Carter, G. L., Campbell, A. C., Muncer, S., & Carter, K. A. (2015). A Mokken analysis of the
Dark Triad ‘Dirty Dozen’: Sex and age differences in scale structures, and issues with
individual items. Personality and Individual Differences, 83, 185–191. doi:
Chabrol, H., Van Leeuwen, N., Rodgers, R., & Séjourné, N. (2009). Contributions of
psychopathic, narcissistic, Machiavellian, and sadistic personality traits to juvenile
delinquency. Personality and Individual Differences, 47, 734–739. doi:
Chiorri, C., Garofalo, C., & Velotti, P. (in press). Does the Dark Triad manifest similarly in men
and women?: Measurement invariance of the Dirty Dozen across sex. Current
Psychology. Advance online publication. doi: 10.1007/s12144-017-9641-5
Crysel, L.C., Crosier, B.S., & Webster, G.D. (2013). The Dark Triad and risk behavior.
Personality and Individual Differences, 54, 35-40. doi: 10.1016/j.paid.2012.07.029
Czarna, A.Z., Jonason, P.K., Dufner, M., & Kossowska, M. (2016). The Dirty Dozen scale:
Validation of a Polish version and extension of the nomological net. Frontiers in
Psychology, 7, 445. doi: 10.3389/fpsyg.2016.00445
Dunn, T.J., Baguley, T., & Brunsden, V. (2014). From alpha to omega: A practical solution to
the pervasive problem of internal consistency estimation. British Journal of Psychology,
105, 399-412. doi: 10.1111/bjop.12046
DARK TRIAD AMONG AT-RISK YOUTHS 17
Elliott, D., Ageton, S., & Huizinga, D. (1985). Explaining delinquency and drug use. Beverly
Hills, CA: Sage.
Ferguson, C. (2009). An effect size primer: A guide for clinicians and researchers. Professional
Psychology: Research and Practice, 40, 532–538. doi: 10.1037/a0015808
Finch, W., Immekus, J., & French, B. (2016). Applied Psychometrics Using SPSS and AMOS.
Charlotte, NC: Information Age Publishing Inc.
Flexon, J., Meldrum, R., Young, J., & Lehmann, P. (2016). Low self-control and the Dark
Triad: Disentangling the predictive power of personality traits on young adult substance
use, offending and victimization. Journal of Criminal Justice, 46, 159-169. doi:
Fritz, M., Kenny, D., & MacKinnon, D. (2016). The combined effects of measurement error and
omitting confounders in the single-mediator model. Multivariate Behavioral Research,
51, 681–697. doi: 10.1080/00273171.2016.1224154
Furnham, A., & Trickey, G. (2011). Sex differences in the dark side traits. Personality and
Individual Differences, 50, 517–522. doi: 10.1016/j.paid.2010.11.021
Furnham, A., Richards, S.C, & Paulhus, D.L. (2013). The Dark Triad of personality: A 10-year
review. Social and Personality Compass, 7, 199-216. doi: 10.1111/spc3.12018
Gott, A., & Hetzel-Riggin, M. (2018). What did you expect? Substance use expectancies
mediate the relationships between dark triad traits, substance use, and substance
preference. Psychological Reports. doi: 10.1177/0033294118755094
Hamby, T., Ickes, W. & Babcock, M. (2016). Evidence for context switching in the effects of
average item length and item-length variability on internal consistency. Journal of
Personality Assessment, 98, 491-502. doi: 10.1080/00223891.2016.1147044
Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable
mediation, moderation, and conditional process modeling [White paper]. Retrieved
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis:
A regression-based approach. New York, NY: Guilford Press.
DARK TRIAD AMONG AT-RISK YOUTHS 18
Henrich, J., Heine, S., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral
and Brain Sciences, 33, 61-135. doi: 10.1017/S0140525X0999152X
Hoyle, R., Stephenson, M., Palmgreen, P., Lorch, E., & Donohew, R. (2002). Reliability and
validity of a brief measure of sensation seeking. Personality and Individual
Differences, 32, 401–414. doi: 10.1016/S0191-8869(01)00032-0
IBM SPSS. (2017). IBM SPSS statistics base 25. Chicago, IL: SPSS.
Jonason, P.K., & Webster, G.D. (2010). The Dirty Dozen: A concise measure of the Dark Triad.
Psychological Assessment, 22, 420–432. doi: 10.1037/a0019265
Jonason, P.K., & Tost, J. (2010). I just cannot control myself: The Dark Triad and self-control.
Personality and Individual Differences, 49, 611–615. doi: 10.1016/j.paid.2010.05.031
Jonason, P.K., Koenig, B.L., & Tost, J. (2010). Living a fast life: The Dark Triad and life
history theory. Human Nature, 21, 428–442. doi: 10.1007/s12110-010-9102-4
Jonason, P.K., Li, N.P., & Czarna, A.Z. (2013). Quick and dirty: Some psychosocial costs
associated with the Dark Triad in three countries. Evolutionary Psychology, 11, 172–
185. doi: 10.1177/147470491301100116
Jonason, P.K., Li, N.P., Webster, G.W., & Schmitt, D.P. (2009). The Dark Triad: Facilitating
short-term mating in men. European Journal of Personality, 23, 5–18. doi:
Jones, D.N., & Paulhus, D.L. (2011). Differentiating the Dark Triad within the interpersonal
circumplex. In L.M. Horowitz & S. Strack (Eds.), Handbook of interpersonal
psychology (pp. 249–269). New York, NY: Guilford.
Jones, D.N., & Paulhus, D.L. (2011). The role of impulsivity in the Dark Triad of personality.
Personality and Individual Differences, 51, 679–682. doi: 10.1016/j.paid.2011.04.011
Jones, D.N., & Paulhus, D.L. (2014). Introducing the Short Dark Triad (SD3). A brief measure
of dark personality traits. Assessment, 21, 28–41. doi: 10.1177/1073191113514105
Kerig, P.K., & Stellwagen, K.K. (2010) Roles of callous-unemotional traits, narcissism, and
Machiavellianism in childhood aggression. Journal of Psychopathology and
Behavioral Assessment, 32, 343–352. doi: 10.1007/s10862-009-9168-7
DARK TRIAD AMONG AT-RISK YOUTHS 19
Klimstra, T., Sijtsema, J., Henrichs, J., & Cima, M. (2014). The dark triad of personality in
adolescence: Psychometric properties of a concise measure and associations with
adolescent adjustment from a multi-informant perspective. Journal of Research in
Personality, 53, 84–92. doi: 10.1016/j.jrp.2014.09.001
Küfner, A., Dufner, M., & Back, M. (2015). Das Dreckige Dutzend und die Niederträchtigen
Neun – Kurzskalen zur Erfassung von Narzissmus, Machiavellismus und Psychopathie
[The Dirty Dozen and the Naughty Nine – short scales for the assessment of
narcissism, Machiavellianism, and psychopathy]. Diagnostica, 61, 76–91. doi:
Lau, K.S.L., & Marsee, M.A. (2013). Exploring narcissism, psychopathy, and Machiavellianism
in youth: Examination of associations with antisocial behavior and aggression. Journal
of Child and Family Studies, 22, 355-367. doi: 10.1007/s10826-012-9586-0
Lyons, M., & Jonason, P.K. (2015). Dark Triad, tramps, and thieves: Psychopathy predicts a
diverse range of theft-related attitudes and behaviors. Journal of Individual
Differences, 36, 215-220. doi: 10.1027/1614-0001/a000177
Malesza, M., & Ostaszewski, P. (2016). The utility of the Dark Triad model in the prediction of
the self-reported and behavioral risk-taking behaviors among adolescents. Personality
and Individual Differences, 90, 7-11. doi: 10.1016/j.paid.2015.10.026
Malesza, M., Ostaszewski, P., Büchner, S., & Kaczmarek, M. C. (in press). The adaptation of
the Short Dark Triad personality measure-Psychometric Properties of a German
Sample. Current Psychology. Advance online publication. doi: 10.1007/s12144-017-
Maples, J., Lamkin, J., & Miller, J. (2014). A test of two brief measures of the Dark Triad: The
Dirty Dozen and Short Dark Triad. Psychological Assessment, 26, 326–331. doi:
Maroco, J. (2014). Análise de Equações Estruturais: Fundamentos teóricos, software &
aplicações [Structural equations analysis: Theoretical foundations, software and
applications]. Pero Pinheiro: ReportNumber, Ltd.
DARK TRIAD AMONG AT-RISK YOUTHS 20
Millsap, R., & Olivera-Aguilar, M. (2012). Investigating measurement invariance using
confirmatory factor analysis. In R. Hoyle (Ed.), Handbook of structural equation
modeling (pp. 380-392). New York, NY: Guilford Press.
Muris, P., Meesters, C., & Timmermans, A. (2013). Some youths have a gloomy side:
Correlates of the Dark Triad personality traits in non-clinical adolescents. Child
Psychiatry and Human Development, 44, 658–665. doi: 10.1007/s10578-013-0359-9
Özsoya, E., Rauthmann, J., Jonason, P.K., & Ardıç, K. (2017). Reliability and validity of the
Turkish versions of Dark Triad Dirty Dozen (DTDD-T), Short Dark Triad (SD3-T),
and Single Item Narcissism Scale (SINS-T). Personality and Individual Differences,
117, 11–14. doi: 10.1016/j.paid.2017.05.019
Paulhus, D. L., & Williams, K. M. (2002). The dark side of normal personality: Self-report and
behavioral correlates. Unpublished manuscript, University of British Columbia.
Pechorro, P., Marôco, J., Poiares, C., & Vieira, R. (2011). Validação da Escala de Auto-Estima
de Rosenberg com adolescentes portugueses em contexto forense e escolar [Validation
of the Rosenberg Self-Esteem Scale with Portuguese adolescents in forensic and school
contexts]. Arquivos de Medicina, 25, 174-179.
Pechorro, P., Castro, A., Hoyle, R., & Simões, M. (2018). The Brief Sensation Seeking Scale:
Latent structure, reliability, and validity among at-risk for delinquency youths. Journal
of Forensic Psychology: Research and Practice, 18, 99-113. doi:
Pechorro, P., Lima, K., Simões, M., & DeLisi, M. (in press). Validity and reliability of the Self-
Report Delinquency among a sample of at-risk youths. Journal of Forensic Psychiatry
and Psychology. Advance online publication. doi: 10.1080/14789949.2018.1439991
Pechorro, P., Pontes, C., DeLisi, M., Alberto, I., & Simões, M. (in press). Escala Breve de
Autocontrolo: Validação e invariância numa amostra de jovens portugueses [Brief Self-
Control Scale: Validation and invariance in a sample of Portuguese youths]. Advance
online publication. Revista Iberoamericana de Diagnóstico y Evaluación Psicológica.
DARK TRIAD AMONG AT-RISK YOUTHS 21
Pechorro, P., Moreira, K., Basto-Pereira, M., Oliveira, J.P., & Ray, J. (in press). The Self-
Report Delinquency scale from the National Longitudinal Study of Adolescent to Adult
Health among at-risk for delinquency youths. Advance online publication. Violence &
Pechorro, P., Gonçalves, R., Maroco, J., Gama, A., Neves, S., & Nunes, C. (2014). Juvenile
delinquency and psychopathic traits: An empirical study with Portuguese adolescents.
International Journal of Offender Therapy and Comparative Criminology, 58, 174-
189. doi: 10.1177/0306624X12465584
Pineda, D., Sandín, B., & Muris, P. (in press). Psychometrics properties of the Spanish version
of two Dark Triad scales: The Dirty Dozen and the Short Dark Triad. Current
Psychology. Advance online publication. doi: 10.1007/s12144-018-9888-5
Rosenberg, M. (1989). Society and the adolescent self-image (Rev. ed.). Middletown, CT:
Wesleyan University Press.
Savard, C., Simard, C., & Jonason, P.K. (2017). Psychometric properties of the French-
Canadian version of the Dark Triad Dirty Dozen. Personality and Individual
Differences, 119, 122-128. doi: 10.1016/j.paid.2017.06.044
Schmitt, N. (1996). Uses and abuses of coefficient Alpha. Psychological Assessment, 8, 350-
353. doi: 10.1037/1040-35184.108.40.2060
Schimmenti, A., Jonason, P. K., Passanisi, A., La Marca, L., Di Dio, N., & Gervasi, A. M. (in
press). Exploring the dark side of personality: Emotional awareness, empathy, and the
Dark Triad traits in an Italian sample. Current Psychology. Advance online publication.
Simões, M.R. (2000). Investigações no âmbito da aferição nacional do teste das Matrizes
Progressivas Coloridas de Raven [Investigations in the context of the national
standardization of the Raven Coloured Progressive Matrices test]. Lisboa: Fundação
DARK TRIAD AMONG AT-RISK YOUTHS 22
Tamura, A., Oshio, A., Tanaka, K., Masui, K., & Jonason, P.K. (2015). Development,
reliability, and validity of the Japanese version of the Dark Triad Dirty Dozen (DTDD-
J). The Japanese Journal of Personality, 24, 26–37. doi: 10.2132/personality.24.26
Tangney, J., Baumeister, R., & Boone, A. (2004). High self-control predicts good adjustment,
less pathology, better grades, and interpersonal success. Journal of Personality, 72,
271–324. doi: 10.1111/j.0022-3506.2004.00263.x
Udry, J. (2003). The national longitudinal study of Adolescent Health (Add Health), waves I and
II, 1994–1996; wave III, 2001–2002. Chapel Hill, NC: Carolina Population Center,
University of North Carolina at Chapel Hill.
van de Vijver, F. (2016). Test adaptations. In F. Leong, D. Bartram, F. Cheung, K. Geisinger, &
D. Iliescu (Eds.), The ITC international handbook of testing and assessment (pp. 364–
376). New York, NY: Oxford University Press.
Vize, C. E., Lynam, D. R., Collison, K. L., & Miller, J. D. (2018). Differences among Dark
Triad components: A meta-analytic investigation. Personality Disorders: Theory,
Research, and Treatment, 9, 101–111. doi: 10.1037/per0000222
Wright, J. P., Morgan, M. A., Almeida, P. R., Almosaed, N. F., Moghrabi, S. S., & Bashatah, F.
S. (2017). Malevolent forces, self-control, the Dark Triad, and crime. Youth Violence
and Juvenile Justice, 15, 191-215. doi: 10.1177/1541204016667995
DARK TRIAD AMONG AT-RISK YOUTHS 23
Age in years (SD)
Education in years (SD)
Note. SES = socioeconomic status
DARK TRIAD AMONG AT-RISK YOUTHS 24
Goodness of fit indexes for the different models of the Dirty Dozen
3-factor, 2nd order
3-factor, 2nd order
3-factor, 2nd order
Note. S-Bχ2(df) = Satorra-Bentler chi-square (degrees of freedom); IFI = Incremental Fit Index;
CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation; C.I. =
confidence interval; AIC = Akaike Information Criteria
DARK TRIAD AMONG AT-RISK YOUTHS 25
Goodness of fit statistics for measurement invariance of the Dirty Dozen
Configural (no constraints)
Weak (metric) invariance
Strong (scalar) invariance
Note. S-Bχ2(df) = Satorra-Bentler chi-square (degrees of freedom); CFI = Comparative Fit
Index; RMSEA = Root Mean Square Error of Approximation; C.I. = confidence interval
DARK TRIAD AMONG AT-RISK YOUTHS 26
Pearson correlation matrixes of the Dirty Dozen
Note. No values were statistically significant in the sexes
*** p < .001
DARK TRIAD AMONG AT-RISK YOUTHS 27
Internal consistency of the Dirty Dozen
Note. α = Cronbach´s alpha; ω = Omega coefficient; MIIC = Mean inter-item correlation;
CITCR = Corrected item-total correlation range
DARK TRIAD AMONG AT-RISK YOUTHS 28
Convergent and discriminant validity of the Dirty Dozen
Note. SRD = Self-Report Delinquency; AHSRD = Add Health Self-Report Delinquency; BSSS
= Brief Sensation Seeking Scale; RSES = Rosenberg Self-Esteem Scale; BSCS = Brief Self-
Bolded values were statistically significant in the sexes
Partial correlations controlling for the remaining dimensions of the Dirty Dozen among the total
sample are given in parentheses
* p < .05; ** p < .01; *** p < .001
DARK TRIAD AMONG AT-RISK YOUTHS 29
Criterion-related validity of the Dirty Dozen
Bolded values were statistically significant in the sexes
Partial correlations controlling for the remaining dimensions of the Dirty Dozen among the total
sample are given in parentheses
* p < .05; ** p < .01; *** p < .001
DARK TRIAD AMONG AT-RISK YOUTHS 30
Figure 1. Three-factor structure with standardized item loadings for the total sample.
DARK TRIAD AMONG AT-RISK YOUTHS 31
Appendix: Portuguese translation of the Dirty Dozen
1. Tenho tendência a levar as outras pessoas a fazerem o que eu quero.
(I tend to manipulate others to get my way)
2. Já enganei ou menti para obter o que eu queria.
(I have used deceit or lied to get my way)
3. Já elogiei (engraxei) pessoas para obter o que eu queria.
(I have used ﬂattery to get my way)
4. Tenho tendência a usar as outras pessoas em meu benefício pessoal.
(I tend to exploit others towards my own end)
5. Tenho tendência a não sentir remorsos ou arrependimento.
(I tend to lack remorse)
6. Tenho tendência a não me preocupar com o que é certo ou errado.
(I tend to be unconcerned with the morality of my actions)
7. Tenho tendência a ser uma pessoa insensível ou fria.
(I tend to be callous or insensitive)
8. Tenho tendência a não me importar com as regras e normas sociais.
(I tend to be cynical)
9. Tenho tendência a querer que as outras pessoas sintam admiração por mim.
(I tend to want others to admire me)
10. Tenho tendência a querer que as outras pessoas me prestem atenção.
(I tend to want others to pay attention to me)
11. Tenho tendência a querer ter prestígio ou estatuto social alto.
(I tend to seek prestige or status)
12. Tenho tendência a esperar que os outros me façam favores especiais.
(I tend to expect special favors from others)
Note. Items 1-4 assess Machiavellianism; Items 5-8 assess psychopathy; Items 9-12 assess