Don’t Believe the Hype: Shining a Light on the Dark Triad
Proceed with caution when reading studies on the Dark Triad.
Posted Jul 03, 2019
“Psychopaths are strong and narcissists are fast, but MACHs endure.”
“Narcissists prefer carbs while psychopaths and MACHs crave red meat, although psychopaths like it rarer.”
These are fictional headlines, but if you read in psychology or even the mainstream press then you have seen many just like these. They come from research studies aimed at understanding the Dark Triad—the three interrelated personality constellations of psychopathy, narcissism, and Machiavellianism.
Interest in this topic has grown remarkably quickly over the past decade, both among scholars and non-scholars. Who can blame people for their interest? As far as we can tell it is one of the few research areas that regularly publishes papers, in academic journals, with titles including “evil,” “good,” “malevolent,” and “dark.” The outcomes studied are legion, ranging from academic major and preference for plastic surgery to infidelity, mating success, preference for daytime vs. nighttime, geographical preferences (urban vs. rural), and aggression.
It’s a dastardly collection of characters straight out of the Game of Thrones, wrapped up in a dangerous sounding name. But we are not sure that you should believe all of the hype.
First, a little history. Psychopathy and narcissism have been in the psychological literature for more than 100 years. Psychopathy refers to a suite of traits that include callousness, dishonesty, grandiosity/egocentrism, shallow and deficient affective responses, irresponsibility, impulsivity, and (in some models) interpersonal boldness and charm. Narcissism refers to traits that overlap quite a bit with psychopathy and include grandiosity, entitlement, exploitativeness, vanity, and exhibitionism. Machiavellianism, although historically older, is a relative newcomer to empirical study. It has conceptual roots in the writings of Niccolò Machiavelli, who detailed how to acquire and maintain power.
Each had different origins and was studied within slightly different fields. Psychopathy has a long history in both clinical psychology and psychiatry. Narcissism can be found in both clinical and social psychology. Machiavellianism is studied mostly in social and industrial-organizational psychology.
In 2002, Paulhus and Williams published what, at the time, seemed a relatively innocuous paper in the Journal of Research in Personality titled “The dark triad of personality: Narcissism, Machiavellianism, and psychopathy” in which they examined the degree of relatedness among the three constructs and their personality correlates with the predominant measure of personality – the big five/five-factor model – as well as indices of intelligence in a moderately sized sample of undergraduates (N = 245).
They found that the three were interrelated, as expected, but had somewhat different basic personality trait profiles. For instance, while all three Dark Triad constructs were related to lower agreeableness, they differed in their relations to traits such as extraversion and conscientiousness. None showed a robust link to their measure of IQ.
This relatively brief empirical study spawned new literature on the “Dark Triad.” To date, this publication has been cited more than 2,500 times and more than 300 articles have been published with this term in the title.
Despite the enthusiasm with which this topic has been pursued by researchers and the spate of sensationalized headlines that have come out this research, basic questions remain about the “Dark Triad.” Some are conceptual. What does “dark” mean exactly in a scientific sense? Why are these three personality constellations included but other equally “dark” ones excluded – antisocial, borderline, and histrionic personality disorders, sadism, spitefulness, pettiness, venality, etc.? Why think in terms of a triad at all, rather than talk about them as different manifestations of a single underlying core of interpersonal antagonism (e.g., the degree to which one is callous, manipulative, immodest, non-compliant)? Still, other concerns are more methodological and empirical. We recently detailed these concerns in a paper published in Current Directions of Psychological Science (Miller, Vize, Crowe, & Lynam, in press) and review some of them here.
First, despite how they are studied, the dark triad constructs are not monolithic or unitary. They can’t be easily summarized with single scores as is commonly done in this literature.
All three Dark Triad constructs are multidimensional, made up of a variety of smaller, differentiable pieces of personality. For instance, psychopathy can be described via three or four basic “ingredients:” interpersonal antagonism (e.g., callousness; manipulativeness; grandiosity), disinhibition (e.g., impulsivity, irresponsibility), and boldness (e.g., emotional resilience, extraversion). Treating psychopathy as a single “thing” forecloses the ability to understand what pieces drive certain relations — e.g., relation between psychopathy and criminal behavior.
Similarly, two people can have the same total score on a dark triad construct but very different scores on the individual components which themselves have very different relations to dysfunction. Imagine that scores on each psychopathy component range from 0 to 10 for a maximum total score of 30. Now imagine that Bob has a score of 15 comprising scores of 10 on Boldness, two on antagonism, and three on disinhibition, whereas Sarah’s score of 15 comprises scores of 10 on antagonism and five on disinhibition. Those two people, despite their equivalent total psychopathy scores, are at very different risk for engaging in the behaviors we associate with psychopathy (e.g., violent and non-violent criminality). Sarah’s scores are much more concerning given that antagonism and disinhibition are strongly related to these problematic outcomes but boldness is not and is instead predominantly adaptive in nature.
One can see how the use of a total score, which can be a case of adding apples and oranges, fails to convey important information about dysfunction that would be communicated if the lower order scores were used (e.g., antagonism vs. disinhibition vs. boldness). Unfortunately, dark triad researchers have treated these constructs as unidimensional despite clear evidence to the contrary and often rely on measures that do not allow for a more accurate and nuanced examination. Moreover, treating all three dark triad components this way obscures the fact that they all include interpersonal antagonism as a core feature, which means that most dark triad research is research on interpersonal antagonism. But that doesn’t sound as sexy.
Second, existing measures of Machiavellianism (MACH) are reliable but not valid. That is, one tends to get the same score each time one takes a measure of Machiavellianism, but it is not a very good measure of the underlying Machiavellianism construct. In fact, measures of Machiavellianism capture something that bears only passing resemblance to the descriptions laid out in Machiavelli’s work and in expert ratings.
Most problematically, scores on existing measures of Machiavellianism yield a personality profile of antagonism, disinhibition and neuroticism that makes it virtually indistinguishable from psychopathy (Miller, Hyatt, Maples-Keller, Carter, & Lynam, 2017).
Impulsivity is supposed to distinguish Machiavellianism and psychopathy, with Machiavellian individuals showing average to good impulse control, which facilitates the long-term conning and manipulation thought characteristic of Machiavellianism – whereas psychopathic individuals are supposed to show poor impulse control.
But people who report high levels of MACH on existing scales also report poor impulse control. In fact, counter to key descriptions, Machiavellian individuals describe themselves as less competent, ambitious, hard-working, and self-controlled.
In a recent meta-analytic review of the entire dark triad literature (Vize, Lynam, Collison, & Miller, 2018), across a variety of different outcomes including impulsivity, the relations for measures of Machiavellianism never differed from those for psychopathy suggesting a problematic failure to measure Machiavellianism in a way that distinguishes it from psychopathy and aligns with theoretical descriptions.
Ultimately, while the three constructs may be differentiable theoretically, empirically this is really a literature on the “dark dyad” (i.e., narcissism and two slightly different measures of psychopathy). This is problematic because individuals studying these constructs still treat these scales as if they capture the construct intended despite all the evidence to the contrary.
A third problem is statistical: the use of an analytic approach in which the relation between one dark triad construct and an outcome is evaluated after controlling for the effects of the other two dark triad constructs. The greatest problem with such an approach, when variables are relatively highly correlated as they often are in the dark triad literature, is that one often finds relations that don’t appear when each construct is looked at by itself (i.e., when one does not remove the variance shared among the three dark triad constructs; see Sleep. Lynam, Hyatt, & Miller, 2017).
For an illustration of the problem, imagine studying the relations among sex, height, weight, muscle mass, testosterone, and maximum bench press. Based on world records and comparative physiology, we suspect that men will be able to bench press more than women. Now imagine that after controlling for differences in height, weight, muscle mass, and testosterone, women can bench press more than men. Which is the correct conclusion? We believe it is obviously the former: Men, on average, can bench press more than women.
In the Dark Triad literature, as one of many examples, one group of scholars found that Machiavellianism was significantly positively related to intelligence whereas psychopathy and narcissism were unrelated yielding a headline of “MACHs are smart but psychopaths and narcissists are not.” The problem is that when each Dark Triad construct is examined by itself, none of them are significantly related to intelligence, yielding a more accurate but less sexy conclusion of “None of the dark triad constructs are meaningfully related to intelligence.”
Greater attention to the difficulties of interpreting studies in which these types of statistical controls are used and what that means for the underlying constructs are desperately needed in this literature. For instance, when one examines only the unique part of each of the DT constructs, only psychopathy remains somewhat true to its original form – that is, it is still characterized by interpersonal antagonism (callousness, deceitfulness; noncompliance) and disinhibition, consistent with theoretical and empirical accounts of psychopathy.
Conversely, the unique part of narcissism is extremely high extraversion (warmth, gregariousness; assertiveness) and emotional well-being (e.g., low anxiety, depression). Most troubling is that the unique component of Machiavellianism is high negative affectivity – depression, anxiety, self-consciousness – and a distrustful, manipulative interpersonal style. The latter components align with expert ratings but the former do not – so one must ask what can be learned by studying the unique component of MACH or the rest of the dark triad.
The above are just a few of our concerns about this research literature. We hope the scientists studying these constructs work to improve it by endeavoring to use additional and putatively more relevant samples, to measure the constructs carefully and comprehensively, to think clearly about their underlying nature (and multidimensionality), and to prioritize these methodological, statistical, and theoretical issues over sexiness, provocativeness, and attention-grabbing headlines.
Until such time, we believe that readers should view such studies with a healthy dose of skepticism. These constructs matter and we should treat them with the seriousness and thoughtfulness they deserve.