Paulhus and Williams (2002) devised the Dark Triad model consisting of three socially maladaptive personality traits. Psychopathy is characterized by high impulsivity and risk taking as well as low neuroticism and empathy (Paulhus & Williams, 2002). Machiavellianism involves manipulation, deceitfulness, and a cynical view of human nature (Christie & Geis, 1970). Narcissism describes a sense of grandiosity, entitlement, and superiority (Raskin & Hall, 1979). Given their maladaptive nature, research has focused on adverse outcomes and correlates; however, some research suggests that the dark traits may be related to some prosocial behaviour (Palmer & Tackett, 2018).

A behaviour is considered prosocial if the act is done with the intention to help others (e.g., giving someone a compliment; Vinacke, 1980). Relatedly, altruism is prosocial behaviour that involves helping another person in a way that could potentially incur a cost to the individual helping (Oda et al., 2014). Studies that investigate altruistic behaviours have shown that people are generally more generous than not (e.g., people usually tend to donate at least a little money in the ultimatum game; Forgas & Tan, 2013); however, it could be argued that behaviours, such as donating money, cannot be altruistic unless the helper incurs a cost to the donor because of the behaviour (Oda et al., 2014).

Altruism has been found to be correlated with personality traits of extraversion and agreeableness (Oda et al., 2014), but relationships between dark personality traits and altruism have not been firmly established. Research has previously investigated the Dark Triad traits relative to prosocial behaviour (e.g., Wertag & Bratko, 2019); however, the few studies examining altruism specifically have generated mixed results (e.g., Palmer & Tackett, 2018; Sakai et al., 2019). Altruism is commonly misinterpreted as simply any prosocial behaviour. For example, Palmer and Tackett (2018) state that individuals high on the Dark Triad may engage in altruistic behaviours if they perceive some benefit to themselves; however, those behaviours cannot be considered altruistic if they incur a benefit. Some studies have found a negative relationship between altruistic behaviours and Machiavellianism (Palmer & Tackett, 2018) and psychopathy (Palmer & Tackett, 2018; Sakai et al., 2019; Vize et al., 2018), while others have reported small positive relationships between altruistic behaviours and Machiavellianism (Jonason et al., 2010) and narcissism (Chen et al., 2021; Palmer & Tackett, 2018). Still other results have suggested no significant relationships with altruism and dark traits (Jonason et al., 2010). Evidently, the findings thus far are inconclusive. One possible weakness has been that previous studies have failed to investigate different types of altruism, such as relational altruism or altruism towards different groups. Perhaps individuals scoring high on the Dark Triad traits are more likely to perform altruistic acts to family or friends rather than strangers. The present study attempts to address the research question with respect to the target of the altruistic acts.


The aim of this study was to evaluate the relationships between the Dark Triad and altruism using two different measures of altruism, allowing us to explore different interpretations of altruism (outward actions and emotional support) and altruistic behaviour in general as well as specifically towards family, friends, and strangers. Based on findings that individuals high in narcissism engage in some altruistic behaviours (e.g., narcissism and reciprocal altruism; Palmer & Tackett, 2018) and the finding that individuals scoring higher in narcissism often engage in impression management (e.g., Mehdizadeh, 2010), suggesting that individuals engage in altruistic acts in order to be perceived as a good person, it was hypothesized that narcissism will positively correlate with altruistic behaviours, more towards family and friends than strangers. Psychopathy was hypothesized to negatively correlate with altruism, based on the findings that individuals scoring high in psychopathy are less likely to preform costly helping behaviours (Sakai et al., 2019) and the fact that psychopathy is characterized by a lack of concern for others (Paulhus & Williams, 2002). In addition, because of the conceptualization that individuals scoring higher on Machiavellianism usually act in their own self-interest (Christie & Geis, 1970), we predicted that Machiavellianism would have a negative correlation with altruism as a true altruistic act is one that results in a disadvantage to the helper (Oda et al., 2014).



Participants were 286 adult volunteers (42 men and 244 women) from the North American community. The participants had a mean age of 38.64 (SD = 15.68; range 16-75 years).


Psychopathy. Participants completed the Self-Report Psychopathy Scale (SRP-III-R12; Paulhus et al., 2016) measuring subclinical psychopathy with 62 items, responded to on a five-point Likert scale (example item, “I’m a rebellious person”) ranging from 1 (strongly disagree) to 5 (strongly agree). Paulhus et al. (2016) reported that the scale has a high internal consistency estimate (α = .79) and has strong convergent validity.

Machiavellianism. Machiavellianism was measured using the Mach-IV (Christie & Geis, 1970), consisting of 20 self-report items (example item, “The best way to handle people is to tell them what they want to hear”) responded to on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Fehr et al. (1992) reported that The Mach-IV has acceptable internal consistency values (ranging from .70 to .76).

Narcissism. Participants also completed the Narcissistic Personality Inventory (NPI; Raskin & Hall, 1979) assessing subclinical narcissism. The NPI has 40 forced choice items, representing a narcissistic response and a non-narcissistic response, of which one option is selected (example item, “When people compliment me I sometimes get embarrassed” versus “I know that I am good because everybody keeps telling me so”). Paulhus and Williams (2002) reported that this scale has a high internal consistency estimate (α = .84). In the present study, the dark variables were scored automatically and therefore coefficient α values could not be computed with this data set.

Altruism. Individuals also completed two measures of altruism. First was the Self-Report Altruism Scale (SRAS; Rushton et al., 1981) consisting of 20 items responded to using a 1 (never) to 5 (very often) response key (sample item, “I have donated goods or clothes to a charity”). Items are composed of altruistic behaviours that one might also describe as a “good deed” (e.g. donating to charity, lending items to neighbours, or giving a stranger a ride). The coefficient α for the 20 items was .87.

Participants also completed the 45 behavioural items from the Compassionate Altruism Scale (CAS; O’Connor et al., 2015). This scale contains items asking about good deeds, but also includes items that relate more to altruism as social support, such as showing affection, helping others with their problems, or visiting. The CAS was created by taking each item from the Social Support Behaviors (SS-B) scale (Vaux et al., 1987) and changing the instructions from receiving the support from others to giving the support to others. For each behaviour, such as, “Paid for lunch when they were broke”, respondents indicate on a 1 (never) to 5 (very often) response key how often they have engaged in the behaviour, first for their family, then for their friends, and then for strangers. The coefficient α values are .97 for both the family and friend scales and .98 for the stranger scale. For both altruism scales, the scale scores are linear aggregates.


Participants provided consent and completed the scales on-line. Individuals were recruited through newspaper advertisements and referrals from other participants. Initially, participants completed the dark scales as part of another study (Vernon et al., 2008). One year later, some of those individuals completed the altruism scales. Those who completed both measures represent the present sample. The study received institutional ethical approval.



Table 1 lists the descriptive statistics and inter-scale correlations between the Dark Triad traits and the altruism measures. The altruism scales had positive correlations among them, with a strong positive correlation between the scores for acting altruistically with friends and family. The only significant zero-order correlation across dimensions was a negative relationship between SRAS and Machiavellianism.

Table 1

Descriptive statistics and inter-scale correlations

ScaleM (SD)123456
1. Self-Report Altruism61.40 (11.36)
2. Altruism towards Family171.76 (27.97).34**
3. Altruism towards Friend157.42 (28.56).32**.67**
4. Altruism towards Stranger90.64 (32.48).28*.33**.50**
5. Narcissism0.39 (0.16).
6. Machiavellianism2.47 (0.38)–.16*–.10–.01–.01.07
7. Psychopathy2.02 (0.34)–.07–.**.41**

[i] Note. *p < .01, **p < .001, two-tailed.

Table 2 contains results from a regression analysis, evaluating whether any of the Dark Triad traits were significant predictors of altruism. Each altruism scale was considered separately. Age and narcissism positively predicted SRAS scores. Being a woman was the only significant predictor of altruism towards family members. Being younger was the sole significant predictor of altruism towards friends. None of the variables significantly predicted altruism towards strangers.

Table 2

Regression results predicting each altruism scale response

Altruism Scale
Altruism towards
Altruism towards
Altruism towards
Adjusted R2.

[i] Note. a1 – men, 2 – women; *p < .01, **p < .001.


This study investigated the relationships between the Dark Triad and altruism. The hypotheses were partially supported; while narcissism was not significantly correlated with altruism, when controlling for the other dark traits, it was a significant predictor of general altruism as measured by the SRAS (Rushton et al., 1981). This finding furthers the research by Palmer and Tackett (2018) as it suggests that narcissism is related to altruism in general and not just reciprocal altruism. Items on the SRAS focus on charitable and prosocial acts, in contrast with the compassionate altruistic behaviours of the CAS. Narcissists may engage in outward prosocial behaviour because the benefits of being thought of as altruistic outweigh the cost of the act. Future research may want to examine how public recognition for altruistic acts influences the correlation with narcissism.

Machiavellianism had a small negative correlation with general altruism, as was predicted; however, when controlling for age, gender, and the other dark traits, Machiavellianism was not a significant predictor of any type of altruism. In addition, psychopathy scale scores were not significantly correlated with, or predictive of, any type of altruism.

Future research investigating altruism and the dark traits may wish to focus on motivational factors behind the altruistic acts. For example, altruistic acts may be reciprocated (e.g., Palmer & Tackett, 2018); therefore, although the individual may incur a cost in the short term, they are likely to benefit from their altruistic behaviours in the long term (Osiński, 2009). Additionally, overt acts may be more common in individuals with dark traits, especially narcissism, because the behaviour may impress an audience (e.g. Bereczkei et al., 2010). Relatedly, Jones and Mueller (2021) found that high scorers on Machiavellianism, compared to those who score high on psychopathy, are more strategic when they engage in antisocial behaviours. Perhaps a similar pattern of relationships may apply for prosocial behaviours; individuals scoring high on Machiavellianism may engage in altruistic acts in some contexts (e.g., relating to a romantic partner) compared to others, whereas high scorers on psychopathy might not show such a pattern.

The correlations between general altruism, as measured by the SRAS, and the three CAS subscales were small to moderate, suggesting that the two scales do not measure the same construct. At face value, the CAS contains some items that are similar to the SRAS (helping in practical ways such as financially, or with physical tasks), but other items on the CAS are related to emotional support such as spending time talking about problems, or providing encouragement. The present study suggests that these two scales measure related, but distinct, concepts of altruism given the moderate correlation between them, as well as the different correlations with the Dark Triad traits.


Limitations of this study include the small number of men in the study. Future research needs to include a more gender-balanced sample. Future research may also want to examine more extreme cases of altruism. For example, Konrath et al. (2016) found that individuals scoring higher in narcissism, who also chose to volunteer, reported that they considered altruistic motives as less important. How costly acts of altruism relate to dark traits requires future study.

It is possible that impression management played a role in the relationship between narcissism and altruism. Gulliford and colleagues (2019) suggested that gratitude could have an impression management component. Specifically, they reported a significant relationship between gratitude and both self-monitoring and social intelligence (dimensions associated with impression management). Perhaps narcissism predicts altruism because those scoring high in this trait perform good deeds to boost others’ impressions of them as a result of the gratitude the recipient experiences from the act. For example, Kowalski et al. (2018) reported positive correlations between narcissism and both social desirability responding and self-monitoring. Future research investigating the narcissism-altruism relationship may want to examine the reasoning behind why individuals engage in altruistic behaviours. Finally, the forced-choice version of the NPI was used, which may present a limitation to this study. Research suggests that using a Likert-scale format can result in higher internal consistency for narcissism (Miller et al., 2018). Future research should try to replicate this work using a narcissism measure with a Likert format, as well as possibly including other “dark” personality dimensions.


The results of this study suggest that endorsement of the Dark Triad does not necessarily indicate a complete lack of altruistic behaviours, but that the relationships are close to being independent. Overall, this study has helped in advancing the conceptualization of the Dark Triad traits.