The One Trait Behind Narcissism, Psychopathy and Manipulation at Work

Three colleagues. The first takes credit for my idea in the meeting, then acts wounded when I mention it afterwards. The second agrees to the plan in the room and works against it the moment the room empties. The third barely reacts when a decision they made costs someone else their job. Different people, different files in HR's head.

Narcissist, Machiavellian and psychopath. And here's what that filing costs. Three files means three vocabularies and three sets of advice when most people never get past working out which one applies. They spend months trying to name the problem and by the time they have, the situation has usually moved on without them.

In 2018, three psychologists asked an inconvenient question. What if these are not different files? What if they're the same disposition, wearing three different masks? They Here's the assumption almost every workplace training still runs on, that difficult people come in fixed types and the task is to identify which one is in front of you.

Is she a narcissist or just confident? Is he a Machiavellian or just political? Each label has been given its own questionnaire, its own case studies, its own confident advice column. Workplace psychology has built three decades of separate vocabulary this way and the cost is real. If these are genuinely separate conditions, they would have to be screened separately, one at a time.

However, there's a pattern the research kept running into. People who score high on narcissism also tend to score high on Machiavellianism. People high in psychopathy tend to score high in both. Not occasionally, but reliably, across study after study, far more often than chance would produce. A fourth label was already on its way.

In 2013, Erin Buckles and colleagues found that some people would relish the chance to hurt someone who had done nothing wrong to them. And that this held after narcissism, Machiavellianism and psychopathy had been accounted for. Everyday sadism became the fourth trait. The dark triad became the dark tetrad and the overlap problem stayed exactly where it was.

That kept happening for a reason nobody had properly tested. Then in 2018, somebody finally did. They would have to get through the働 back. They The 2018 research paper by Moshigan, Hilbig and Zettler was published in the Psychological Review. Their argument borrows from a different corner of psychology.

In cognitive research, there's the G-factor for general intelligence. It's not a test anyone takes directly. It's the common thread that may explain why someone good at logic puzzles also tends to be good at maths. Moshigan and colleagues propose the same structure exists for human darkness. They called it D.

The definition is precise enough to test. D is the general tendency to maximise one's own utility, meaning gains, advantage, satisfaction, while disregarding, accepting or actively promoting harm to others, accompanied by beliefs that serve as justifications. The middle component is where the explanatory work happens, because it's not one position, it's a spectrum and people sit at different points along it.

At one end, the cost is simply disregarded. It does not register or it does not weigh, and the person reads as cold rather than cruel. That end sits closest to psychopathy. Further along, the cost is registered accurately and accepted as the price of getting what you want, and the person reads as calculating.

That is closest to Machiavellianism. At the far end, the cost stops being a price and becomes part of the payoff, and the person reads as cruel. That is where sadism sits. Those are the three masks, and it's worth noticing what has just happened to the labels. Three of the four went onto that spectrum.

Narcissism didn't, and that's not an oversight. The version of narcissism that matters here is organised around standing rather than around cost, so it could turn up anywhere along the spectrum. This is the first sign that the labels were never the right unit. Same disposition, recognisably different people, depending on where along that spectrum they sit.

And the third piece, the justifying belief, is what lets any of them look in the mirror and still feel reasonable. It's a competitive jungle out there. They're not as sharp as me. The story changes person to person. Its function and consequence do not. Self- Now, what this framework is not, because that boundary is where most coverage of it goes wrong.

Once a model like this exists, it becomes tempting to go looking for famous examples, and business books do exactly that. Elizabeth Holmes, Robert Maxwell, Al Dunlap and Fred Goodwin, the same four names confidently diagnosed. I'm not going to characterise any of them, and the reason is more interesting than the caution.

Their positions in the public record are not remotely equivalent. Holmes was convicted in 2022 of defrauding Theranos investors, and acquitted of every charge relating to patients. Maxwell died in 1991 before any proceedings. The pension misappropriation was established afterwards. Dunlap settled accounting fraud charges without admitting or denying them, and Goodwin was never found guilty of anything at all.

Four genuinely different legal positions, which collapse into a single storey remarkably easily once somebody wants them to. That collapse is the thing to notice. None of these people has ever sat a de facto test. A retrospective read of press cuttings is not a psychological assessment, whatever it's called.

Measurement here is contested, even when done properly. A widely cited instrument for scoring corporate psychopathy was examined in print by Robert Hare, who developed the clinical checklist it has been adapted from, and judged to misspecify the construct it claimed to measure. If a purpose-built workplace instrument can attract that verdict, a reading assembled from newspaper archives sits on weaker ground still.

So what's the point of these stories? They're not a diagnosis of anyone. They illustrate the pattern. Utility maximised. Costs to others disregarded or accepted. And a story that made it feel justified. That's a fair use of the public record. A clinical label is not. A clinical label is not. So what do the masks look like up close?

The list first, because the number matters later. The 2018 paper tested nine traits. Narcissism, Machiavellianism, Psychopathy and Sadism, which are the dark tetra you've heard of. Then spitefulness, psychological entitlement, egoism, moral disengagement and self-interest. I'm taking the four, then spitefulness and entitlement.

The last three behave much as their names suggest. The tells are subtler than the stereotypes. Narcissism isn't simply confidence. Plenty of good leaders have narcissistic tendencies. Research distinguishes its two faces, admiration, which is largely self-promoting, and rivalry, which is defensive and derogating of others.

The rivalry mask generally carries the dark relevant signal, and it surfaces when a colleague's success threatens someone's own standing. Enthusiastic credit claiming when things go well, a conspicuous absence the moment they don't. Machiavellianism isn't cynicism in general. It's the colleague who registers the cost their actions impose on other people accurately, and simply decides that cost is worth paying.

Psychopathy here is not the criminal stereotype. The scales this literature runs on were built on ordinary non-institutionalised samples, university students rather than prisoners. What they pick up at the workplace relevant end is callousness. The cost registers accurately, and simply does not carry the weight that would normally stop someone.

That's the third colleague from the cold open, and it's the disregard end of the spectrum. Everyday sadism rarely announces itself as cruelty. It tends to hide inside a defensible surface, the manager who lingers a beat too long on a mistake, past the point the correction required. The tell is not what was said.

It's the surplus, the satisfaction a situation doesn't explain. Spikefulness may be the strangest of the nine, the one trait where someone will accept a real cost to themselves just to make sure someone else pays a bigger one. Not a rational trait at all, which is exactly why it can be such a reliable tell.

And entitlement supplies a story for all the others. It's the belief, often dressed as a fairness complaint, that being merely treated as an equal is itself a kind of injury. Nine traits built by separate research groups for separate purposes over decades, and they correlate far more than separate traits ever should.

Thank Naming a pattern is one thing. Establishing it as a genuine stable feature of a person rather than five questionnaires asking the same thing five different ways is another thing entirely. Here's the strongest single piece of evidence. Zetler, Moshigan and Hilbig followed more than 1200 adults across four years.

These showed higher rancor to stability than any of the nine specific traits it's built from. It predicted those traits four years later and changing the traits can be traced back to changing D. That's considerably harder to explain away as coincidence than a same day correlation. A second detail worth knowing, because D was defined as a fluid construct it should be recoverable from any sufficiently broad set of dark trait indicators and in practice it generally is.

The published item sets come in 70, 35 and 16 item versions and the short one behaves much like the long one. Whatever this is, it doesn't appear to be fragile. A 2025 study extends a picture from the individual to the society around them. Countries and even individual US states with more corruption and more inequality score higher on average D.

That's an association, not a demonstrated cause. However, it fits a straightforward reading. Environments may modulate the same disposition up or down. That's an association, Here's an honest complication, and it's one that serious critics keep returning to. Long before D existed, personality researchers had two candidates for the dark core.

Low honesty humility, from the Hexaco model, and low agreeableness, from the Big Five. If either already explained the overlap, D isn't a discovery, it's a rebrand. The D team tested that directly, and the answer is more interesting than a clean win. D and low honesty humility overlap heavily, but come apart in places, with each predicting something the other doesn't.

The agreeableness version produced a published commentary, a published reply from the model's authors, and further papers on both sides, as recently as 2025. That exchange is still evolving. There's an even sharper problem. A 2015 study tried to fit narcissism, Machiavellianism, and psychopathy as three equal partners in the original dark triad, and the model wouldn't hold.

All of the shared variants traced back to psychopathy alone. That study ran on a correctional sample rather than a workplace one, which limits how far the finding travels, but it still raises an uncomfortable question. Is this genuinely nine traits sharing one route, or several measures that were already redundant with each other before the unifying theory arrived?

Nobody has settled that question yet, neither the researchers nor their critics. The four-year stability evidence appears to rule out pure statistical noise. It doesn't rule out how many real, independent things are sitting underneath. So, where do people who score high on this tend to end up? A study published in 2026 pooled six samples, more than 8,000 people, across Germany, the United States and Denmark, using both self-reported interests and official job registry data.

The clearest finding, people high in D are reliably less drawn to social roles, such as teaching, caring and helping, and to creative and artistic work. That held whether researchers asked what people wanted, or checked what job they actually had. What it does not cleanly support is a popular idea that dark personalities systematically climb into leadership.

The link to enterprising roles was positive in the German data, statistically insignificant in the US data, and mixed in the Danish records. These are correlations at a single point in time. They indicate where people are, not how they got there. An older and much larger meta-analysis supports a narrower claim.

Machiavellianism and psychopathy tend to predict lower job performance, and all three of the original dark triad traits predict more counterproductive behaviour at work, with the size of the effect depending on how much authority the person holds, and on the surrounding culture. So the accurate version is not that dark personalities run the company.

Wherever they land, the behaviour that follows tends to cost the organisation more than it shows up on anyone's objectives. In my own experience, I have observed that the prevalence of these dark traits may be as much related to the type of environment as it is to the role position in that environment.

In other words, certain vocations may be more heavily represented in certain dark traits, and the extent of these dark traits is amplified with the role seniority. For example, the shark-infested waters within law firms, the highly Machiavellian environment of politics, and the types of individuals that function successfully as surgeons or military operators.

This is already subject to various studies, but more research is required. One more finding is worth adding, because it cuts against how we should talk about this. In 2024, Furnham and Capello asked what the four dark traits shared most strongly. Out of the four work-related traits they measured, the answer wasn't cruelty or coldness or even self-interest.

It was competitiveness, the need for the outcome to be ranked and for someone else to come second. That was a self-report study of 447 people, so I would hold it lightly. It's a lead worth following rather than a settled conclusion. It does suggest something about where this shows up. Competition does not stop at the office door.

The disposition that turns a project review into a scoreboard may turn a family dinner into one. People are often startled to find the colleague they've described for months sounds exactly like a relative they've described for years. So while this episode is framed around work, none of it is confined to work.

Ready and none of it is destiny which matters more than it might sound the same disposition can look very different depending on the environment around it a team with clear metrics real oversight and genuine consequence for crossing a line tends to constrain this behavior a team with none of that gives it room to surface two colleagues with a genuinely similar disposition may therefore produce very different outcomes at the same company the series companion article on the bad apple effect covers that mechanism in full i'm showing the link to the article on screen now the short version knowing someone scores high on this indicates what to watch for it doesn't indicate what they'll actually do So what should you actually do with this?

I want to start with what not to do because it's the part where the most damage gets done. Do not administer a personality test. Not formally, not informally, not as a team building exercise. The item sets behind D were validated in anonymous research conditions where there was no incentive for anyone to fake their answers.

But the same question is in front of a job applicant and that protection disappears. When people are instructed to fake responses to dark trait measures, their scores move about as much as they do in the same situation with the big five measures. These questionnaires are not resistant to being gamed.

They're vulnerable, which in a selection context is a bad outcome. There's also a straightforward signal from the people who built the instrument. The published D item sets are distributed for non-commercial use only and anyone downloading them has to agree to that. Whatever else that is, it's not an endorsement of use in hiring and a broad value-laden measure may carry real discrimination risk if used to filter people who've not actually done anything yet.

So here's the line to hold precisely. This evidence is strong enough to take a documented, repeated pattern seriously once it's actually been seen. It was never intended to justify testing someone for a pattern before they've shown you anything at all. So Here are three questions you can ask instead.

These require no clinical qualification and they are not arbitrary. Each question maps onto one of the three components that define D. I shall be clear about what these are. They are reasoned judgement from practice, derived from the definition, not a validated screening tool. Nothing here diagnoses anyone.

Question 1. Does this person's account of events consistently minimise the cost their choices impose on someone else? Not once under pressure, but consistently across situations where the cost was obvious. Question 2. Does their justification under scrutiny reflect a worldview where ruthlessness is simply realism, or where they are quietly exempt from standards everyone else has to meet?

Question 3. Do they appear to take some disproportionate satisfaction in a colleague's setback that the situation itself does not explain? All three describe things you can actually document, and documentation is the part almost everyone gets wrong. There's a difference between a record and an accusation, and that may decide almost everything that happens next.

This is a practitioner judgment rather than research, and I'll say so plainly because no study has tested it. An accusation is a conclusion about a person. He is manipulative. She's a narcissist. Each of these tends to invite a counter-conclusion about you. And here's the uncomfortable part. Someone high on that disposition is very likely to be better at managing that kind of exchange than you are.

That's close to the definition of the trait. Bring a conclusion, and you're competing on their ground. A record is different. A record is behaviour, date, who else was present, what was said, and what happened next. Nothing more. No adjectives. No inference. No diagnosis. Which is a practical reason the three questions are phrased the way they are.

For example, So, the 11th of March, in the project review, this commitment was made in front of these four people. The 19th of March, that commitment was described to the sponsor as never having been given. That's a record. It doesn't call anyone anything, and it doesn't need to, because it stands up to scrutiny in a way that an accusation doesn't.

Then the harder question, and the one most people skip, what is a record for? There are two answers, and their thresholds differ. The first is a governance matter. Other people are affected. There may be a policy breach, or something affecting safeguarding, or regulatory obligation. That record goes to whoever owns the risk, and it needs to be dry, dated, and specific, because it will be read by people who were not there.

The second is a behavioural boundary matter. You're the only one materially affected. Nobody is going to intervene, and, honestly, nobody should have to. That record is for you. Its purpose is to stop the same event being re-litigated in your head at 2am. Confusing the two is a common mistake, I see.

A boundary matter taken to HR usually gets a person labelled as difficult. A governance matter treated as a personal boundary leaves other people unprotected while you quietly absorb it. After all, Now for the part most videos on this subject will not tell you. There's no validated protocol for handling a high D colleague.

None. No trial, no controlled study, no evidence base establishing that any approach reliably works. The research offers reasonably strong evidence about what this disposition predicts and almost nothing about what to do once it's been found. So when a consultant offers a five-step system for handling the narcissist in your team, be aware that they're not reporting research.

They're selling a framework. It may be a good framework, but it's worth understanding the complete lack of an evidence base, which puts the efficacy of the framework into question. Beyond that, I can tell you what my research in combination with 25 plus years in and around senior teams has taught me.

And, to be clear, this is purely my judgment, also not evidence. Two things seem to help more reliably than anything else. Reduce the person's discretion over what affects you. Decisions in writing. Commitments made in front of others. Few arrangements that exist privately between the two of you. And increase the number of people who can see the interactions.

Neither change is the person. Both change your exposure. And on the evidence we have, that's the honest goal. It tends to hold up well outside work, also, in your personal life. So the three colleagues from the start. Not three separate problems requiring three playbooks. And not three labels to sort people into.

Very possibly one disposition. Showing up at three different points on the same spectrum. Each held in place by whatever story lets that person still feel reasonable. Nobody needs diagnosing. What is needed is a description of what happened. A decision about which destination it belongs to. And appropriate actions in response.

Yeah. Next week, I'm addressing the trait everybody thinks they can spot, grandiose narcissism, and why the people promoted on it so often underperform once they get there. It's a harder episode than it sounds because most of us mislabel confidence as narcissism, and the research says that mistake is expensive in both directions.

Look out for episode 5, coming next week. If you want the full evidence behind today's episode, including where this model's own critics land some real punches, you can read the companion article on my website. It carries every reference, the verification statement of each one, and six questions I frequently get asked.

The link is the first thing in the description below, and on the screen now. If this was useful, please like and subscribe. It helps the channel reach the people it's actually for, and it'll ensure you're notified about the next episode in the series. I'm Dr Nick Kecker. Thank you for watching. I'll see you again next week.

Thank