Abstract
The question this article answers: why does the same organisational response, whether a warning, a sanction, a piece of feedback or an offer of support, change the behaviour of one difficult colleague and leave another untouched, and what should an organisation do given that it cannot and should not work out which is which?
Psychopathy, as studied in working populations, is a set of traits on which everybody sits somewhere, long argued to take a cold, low-anxiety primary form and an anxious, impulsive secondary form that look alike from outside [NEEDS SOURCE: Karpman (1941), volume and pages unconfirmed against a primary index]. This article asks what that predicts about workplace responsiveness.
Its central workplace source is a 2025 meta-analysis, meaning a study that pools the results of earlier studies rather than running a new one, of 166 samples and 49,350 psychopathy scores, which reports that psychopathy is associated with lower task performance, less helpful discretionary behaviour and more misconduct at work, and that these effects were mostly more pronounced for the secondary form (Roth & Klehe, 2025). Two organisational studies supply the responsiveness argument directly: of the features tested, clear rules in particular weakened the link between leaders’ primary psychopathy and their self-serving and abusive behaviour (Laurijssen, Wisse, Sanders & Sleebos, 2024), and employees high in the primary form reported more positive mood and engagement under abusive supervisors (Hurst, Simon, Jung & Pirouz, 2019).
The mechanism evidence is drawn from questionnaire work on sensitivity to punishment and reward (Newman, MacCoon, Vaughn & Sadeh, 2005; Ross et al., 2007), laboratory work on the dependence of fear on attention (Newman, Curtin, Bertsch & Baskin-Sommers, 2010), and a pooled analysis of the size of the difficulty such people have in pulling back once set on a goal, pooled across 94 samples (Smith & Lilienfeld, 2015). A chapter sets out where that evidence disagrees with itself, including a 2025 study of 446 adults that does not fit the neat account (Castagna & Kinrade, 2025) and the unresolved dispute over whether the secondary form is a distinct thing at all (Miller, Gaughan & Pryor, 2008).
A late chapter sets out the people most often mislabelled in this territory, and the practical section argues that responsiveness is observable where the underlying pattern is not, and that the responses best supported by the evidence are those that reach both forms. It is candid that almost none of those responses have been tested as an intervention inside an organisation.
Keywords: psychopathy; primary psychopathy; secondary psychopathy; punishment sensitivity; counterproductive work behaviour; abusive supervision; organisational rules; evidence appraisal; organisational psychology.
PART I. WHAT ACTUALLY HAPPENS
Chapter 1: Two Warnings, One Outcome
1.1 Two Files on the Same Desk
The scenarios below are composite. They draw on recurring patterns in research and advisory practice, and they describe no real, identifiable person or organisation.
A regional head of lending has two relationship managers on her team who have each been overriding the same credit control, the one that requires a second opinion before a facility is extended beyond the client’s approved limit, and on the same Tuesday she gives each of them the same written warning, in the same words, in the same room; the first, whom I will call the anxious one, goes visibly pale, apologises twice before she has finished speaking, and for five weeks is the most scrupulous person on the floor. Then the quarter closes badly, his numbers are exposed, and in a single afternoon he overrides the control twice and shouts at a junior analyst who asks him whether he meant to.
The second, whom I will call the calm one, listens without interrupting, thanks her for being clear, and asks a sensible question about the escalation route. Nothing changes. He overrides the control the following month, more discreetly, with a file note that would survive an audit, and when she raises it again, he is exactly as courteous as before, which leaves her with the distinct and uncomfortable sense that she is the only person in the conversation who is under any pressure at all.
Three months later the bank changes the system itself, so that an override above the limit cannot be booked without a second named signatory, and the calm one’s overrides stop within the month. He never mentions the change. The anxious one’s continue, in bursts, every time a deadline lands on him.
1.2 What the Two Have in Common
From the regional head’s chair, the two cases look almost identical, and that is the point of starting with them. The same control, the same breach, the same warning, the same file, and the same conclusion forming in her head that she is dealing with two people of the same kind who simply need firmer handling.
What she is actually looking at is two people who respond to entirely different things. The anxious one responded to the warning immediately and completely, and the response did not survive the next load of pressure. The calm one did not respond to the warning at all and instead responded immediately to a change in the rule that made the act impossible to perform alone. One lever reached one of them and a different lever reached the other, and the only one she thought to pull reached nobody for long.
That is the phenomenon this article is about, and you can state it without any technical vocabulary at all. Two people look identical from the outside, and only one responds to anything you do.
1.3 Why Reaching for the Word Gets in the Way
The word most readers will have reached for by now is psychopath, and I would ask you to put it down for three reasons before going further. The first is that it is the wrong grain: psychopathy, as it is studied in ordinary working populations, is a set of traits on which every one of us sits somewhere, most of us low, and a trait score describes a tendency across people rather than a verdict on one of them.
The second reason is that the word carries a single picture, and the picture is almost always the calm one in 1.1. The research this article draws on has argued for decades that the traits come in two broad forms. The anxious, reactive, easily overwhelmed form is not a milder version of the calm one but a different pattern with a different engine. A word that calls up only one of them will send you looking in the wrong place.
The third should matter most to a practitioner. To call a colleague a psychopath is to make a claim about what they are, which they can deny with complete confidence and which you cannot substantiate, and it moves the conversation from what happened to whether you were entitled to say it. That is a fight you lose. Everything that follows is designed so you never have to make that claim at all, because the useful information is in how somebody responds rather than in what they are.
Chapter 2: One Label, Two Patterns
2.1 An Old Distinction That Keeps Coming Back
The idea that psychopathy takes two forms dates back to clinical writing in the 1940s, which argued that one form is a fixed feature of the person and the other a symptom of distress or adverse circumstances [NEEDS SOURCE: Karpman (1941), volume and pages unconfirmed against a primary index]. The two labels come in their modern form from a short questionnaire built in the 1990s for ordinary adults outside any institution (Levenson, Kiehl & Fitzpatrick, 1995), and they have stuck.
Primary psychopathy is the cold form: low anxiety, low fear and a willingness to use people. Secondary psychopathy is the reactive form: impulsive, quick to anger and, crucially for this article, often anxious rather than calm. A later account prefers three ingredients to two types, boldness, meanness and disinhibition, which is why the calm one in 1.1 and the anxious one can commit the same breach from two different engines (Patrick, Fowles & Krueger, 2009).
2.2 Two Groups of Prisoners With the Same Records
The strongest evidence that the split is real comes from studies that sorted people by personality and then compared the groups. Among 367 violent offenders high in psychopathy, two groups emerged with comparable antisocial records, and the secondary group was more anxious, with more features of borderline personality, poorer functioning with other people and more mental-health symptoms, while carrying fewer of the core traits (Skeem, Johansson, Andershed, Kerr & Louden, 2007).
A smaller study of 96 male prisoners found the same pattern: one group emotionally stable and socially dominant; the other high in negative emotion, low in self-restraint, and low in warmth, which the researchers labelled aggressive (Hicks, Markon, Patrick, Krueger & Newman, 2004). Two samples, two methods, one division.
That is worth holding on to, because it means the anxious relationship manager in 1.1 is not a failed version of the calm one. He is a recognisable pattern in his own right, and researchers have been finding him in prisons for twenty years.
2.3 What This Does Not Show
Both subtype studies in 2.2 were run in prisons, and much of the early work separating the two forms was forensic. The step from a prison to an office is an extrapolation that the next chapter only partly supports. The questionnaires are imperfect too, and one has been criticised for measuring, under the heading of secondary psychopathy, a good deal of ordinary negative emotion, a dispute Chapter 7 takes seriously.
What matters for a manager is not where the boundary sits. It is the difference in what each pattern responds to.
Evidence status: evidenced as to the two patterns in forensic samples; contested as to their boundaries. The split is replicated across independent prison samples using different methods. Whether the secondary form is a distinct thing or largely general negative emotion is disputed, and Chapter 7 sets out the dispute.
Chapter 3: The Cold One Is Not Doing Most of the Damage
3.1 One Hundred and Sixty-Six Samples
If the popular picture were right, the cold form would be the one wrecking organisations and the reactive form a comparatively minor irritation, which is very nearly the opposite of what the largest synthesis of the workplace evidence found, and that finding is the one I would most like readers of this article to carry away. Lenke Roth and Ute-Christine Klehe pooled 166 independent samples, covering 49,350 psychopathy scores from studies published between 2008 and 2024, on the relationship between psychopathy and three kinds of workplace behaviour: how well people do their jobs, how much discretionary help they give beyond their job description, which researchers call organisational citizenship behaviour, and how much they do that actively damages the organisation or the people in it, which researchers call counterproductive work behaviour (Roth & Klehe, 2025).
Psychopathy was associated with worse outcomes on all three, and more substantially than an earlier and more tentative pooling had suggested (O’Boyle, Forsyth, Banks & McDaniel, 2012). What matters here is which form was doing it. The effects were mostly more pronounced for secondary psychopathy, and the paper’s own summary describes the detrimental effects as holding for secondary psychopathy and, to a lesser extent, for primary psychopathy. That is the reverse of the popular picture.
3.2 What That Does and Does Not Say
I want to be exact about this, because it is easy to overstate and the overstatement is tempting. The finding does not say the cold form is harmless at work; it says that when the two forms are measured separately, the reactive one carries more of the measurable harm to performance, helpfulness, and conduct, while the cold form still carries some. I have not been able to read the paper’s own figures for each form, so I am reporting the direction of the difference and not its size. I would treat anyone quoting a precise number with caution until you have seen where it came from.
The evidence is also built from associations rather than experiments, so it tells you what travels together in large groups and nothing about any particular colleague. Roth and Klehe say as much themselves. What it does tell you is that the picture most organisations carry around, in which the dangerous colleague is the composed one, is pointing at the smaller half of the problem.
Evidence status: evidenced as to direction. A peer-reviewed meta-analysis of 166 samples in the leading applied journal. The size of the difference between the two forms is not reported here because the paper’s own figures could not be read during drafting; see publication status footnote 1.
3.3 The Same Trait, Two Careers
One workplace study shows the split operating inside a real organisation, and it adds a twist the meta-analysis cannot. Gerhard Blickle and Nora Schütte measured two factors in 523 German employees and then asked each person’s co-workers, rather than the person, to rate how they actually behaved (Blickle & Schütte, 2017). The first factor, which they called self-centred impulsivity and which sits close to the reactive form, predicted misconduct directed at the organisation. The second, fearless dominance, which sits close to the bold ingredient of the cold form, did something more interesting: among people with less education and weaker social skill it went with misconduct, and among the highly educated and socially skilled it went with better job performance.
That is a finding worth sitting with, because it describes the calm relationship manager in 1.1 better than any label does. The same fearless, dominant disposition is a liability in one setting and an asset in another, depending on whether the person has the skill to channel it, which may be one reason the cold form survives and advances in organisations rather than being weeded out by them, although that is my inference and not the study’s finding.
3.4 Why the Popular Picture Points the Wrong Way
If the reactive form does more of the measurable damage, why is it the calm one everybody pictures? I think there are three reasons, and the third matters for what follows.
The calm form is more memorable. A composed person who harms others without apparent discomfort is frightening in a way that a volatile person is not, and fear is what makes a story stick. The reactive form is also easier to explain away, because an anxious, overwhelmed colleague who lashes out looks like somebody having a hard time, and usually is, so the harm gets filed under stress rather than under conduct.
The third reason is that the reactive form responds, and responsiveness reads as manageability. The anxious relationship manager was contrite; he improved immediately, and he looked like somebody who had learned a lesson, which is why the relapse in week six surprised everybody except, in my experience, the people who had worked alongside him longest. An organisation that measures whether a warning landed by how it was received in the room will conclude that the anxious one is fine and the calm one is the problem, and on this evidence it will be wrong about both.
PART II. WHY THE SAME LEVER LANDS DIFFERENTLY
Chapter 4: What a Threat Actually Does
4.1 A Brake and an Accelerator
The most useful way I know of thinking about responsiveness is a model with two systems in everybody: a brake, which notices possible punishment or loss and slows you down, and an accelerator, which notices possible reward and pushes you on. The literature calls them the behavioural inhibition system and the behavioural activation system, and almost every organisational response to bad behaviour presses one or the other.
A warning, a sanction, a performance improvement plan and a disciplinary hearing are all attempts to press the brake, while a bonus, a promotion, public recognition and a stretch assignment are all attempts to press the accelerator, and in my experience most organisations have far more of the first kind of tool written into their procedures than of the second. Most reach for the brake first. Which of those will work on a given person depends, at least in part, on how sensitive each system is in that person, and this is where the two forms separate.
4.2 Where the Two Forms Part Company
In 2005 Newman and colleagues set out a simple prediction: the cold form should have a weak brake and an ordinary accelerator, while the reactive form should have an ordinary brake and an overactive accelerator (Newman, MacCoon, Vaughn & Sadeh, 2005). For the cold form, the results matched predictions. People scoring high on it reported a brake that engaged less readily when punishment was in prospect, a finding about what people say of themselves rather than what they do.
A study of 326 young adults outside any institution sharpened the picture and complicated it in the same move (Ross et al., 2007). Both forms were related to the accelerator, so both reported being drawn to reward. Only sensitivity to the threat of punishment separated them.
I would put the practical consequence in one sentence. If you want to reach both forms at once, this evidence suggests, though it does not establish, that you reach for the accelerator, because that is the system they share, and that a response built entirely on the brake is built on the one system where the cold form is least likely to be listening.
4.3 Fear That Depends on Where You Are Looking
The obvious conclusion from 4.2 would be that people high in psychopathy simply do not feel fear, and that a threat therefore has nothing to act on. The evidence does not support that, and the correction is more useful than the original idea.
Newman, Curtin, Bertsch and Baskin-Sommers measured a reflex, the startle blink that grows larger when a person is afraid, in 125 prisoners facing a signalled threat during a laboratory task (Newman, Curtin, Bertsch & Baskin-Sommers, 2010). When the prisoners high in psychopathy were attending to the threat itself, their fear response was normal. When their attention had been directed to something else in the task, it was reduced. The study did not separate the two forms.
That changes the question entirely. The problem, for people high in the trait, is not that a consequence cannot register; it is that it registers fully only when it is the thing in front of them, and the moment their attention is on the goal, the target, the deal or the win, the consequence drops out of view, which is the same shape an older study of 60 prisoners found in behaviour rather than in reflexes: when reward and punishment were both in play, the psychopathic group was relatively poor at learning to hold back from the rewarded response, which is precisely the situation of somebody chasing a number while a control stands in the way (Newman & Kosson, 1986).
Evidence status: evidenced as to the laboratory finding, in prisoners; inference as to the workplace. The attention finding is an experimental result in 125 offenders and was not split by form. The step from a laboratory threat to a credit control is mine.
4.4 What Goes Missing Is Restraint
A second correction comes from a pooled analysis that asked whether psychopathy is really about low anxiety at all. Its answer was that fear and anxiety were only negligibly related to psychopathy as a whole, and that what showed a moderate relationship was low restraint, the ordinary capacity to hold back (Derefinko, 2015). I would not lean on the exact wording here, because I have been able to confirm the paper’s conclusion but not its figures. The direction, though, fits everything else in this chapter.
It also explains why the anxious relationship manager relapsed. He was not short of fear; he had plenty of it, and it was visible in the room when the warning was delivered. What he lacked, under load, was the brake that turns fear into holding back.
4.5 How Big Is Any of This
A reader entitled to be sceptical will want to know how large these effects are, and the honest answer is that they are real and modest. The idea that people high in psychopathy find it hard to change course once they are set on a goal has been tested across 94 samples and 7,340 people, and pooled together the difficulty is there but it is small to medium in size by the conventions researchers use to describe such things (Smith & Lilienfeld, 2015), with two further findings attached: how anxious the person was changed the size of the effect, which is the two forms showing up again in a different guise, and there were signs that the more striking results were the ones more likely to have been published.
The authors conclude that this account is unlikely to explain psychopathy on its own. I agree with them, and I would draw a narrower lesson from their numbers than popular writing usually does: the difficulty is real enough to change how you design a response, but not large enough to justify treating anybody as beyond the reach of one.
Chapter 5: Why Getting Tough Does Not Reach the Cold One
5.1 Happier Under a Bad Boss
If the cold form has a quieter brake, the standard managerial instinct, which is to raise the stakes when behaviour does not change, should work poorly on it. One study went further and asked whether harsh treatment might actually suit people high in the cold form, and found that, on some measures, it did.
Charlice Hurst, Lauren Simon, Yongsuhk Jung and Dante Pirouz ran a scenario study and a time-lagged field study of employees working under supervisors who treat them with sustained hostility, which researchers call abusive supervision (Hurst, Simon, Jung & Pirouz, 2019). Under that kind of boss, employees high in primary psychopathy reported more positive mood and more engagement with their work, and in the field study they were less angry than their peers under the same supervisors; under supervisors who did not behave that way, that advantage shrank or reversed.
I find that one of the most practically important findings in this literature, and I would state its implication as bluntly as I can. The treatment most people find corrosive is, for people high in the cold form, a setting in which they do comparatively well. The authors note that abusive supervisors may therefore empower employees whose characteristics could damage the organisation.
Evidence status: evidenced, from two studies. One scenario study and one time-lagged field study, published in a peer-reviewed business ethics journal. Only the primary form was measured, so the finding says nothing about how the reactive form fares under the same treatment.
5.2 Clear Rules, in Particular
If raising the temperature does not reach the cold form, what does? The most directly relevant study I have found asked exactly that question about leaders, and its answer is not the one most organisations act on. L. Maxim Laurijssen, Barbara Wisse, Stacey Sanders and Ed Sleebos started from the common view that leaders high in primary psychopathy cannot be controlled and tested it across three studies: an experiment, a survey of leaders paired with their own subordinates, and a survey of teams (Laurijssen, Wisse, Sanders & Sleebos, 2024), and they compared three features of the organisational context for how far each weakened the link between a leader’s primary psychopathic traits and their self-serving and abusive behaviour: clear rules and policies, the ability to sanction misconduct, and the visibility of what the leader was doing.
Of the three, clear rules in particular weakened the link. I have not read the paper’s figures for the other two, so I would not claim they did nothing, only that they did not stand out in the same way. That is almost exactly what the lending scene in 1.1 describes, where a warning backed by the possibility of a sanction achieved nothing and a rule built into the system, which made the act impossible to perform alone, stopped it within the month.
Evidence status: evidenced, from three studies. One experiment and two field surveys, in a peer-reviewed journal. Only the primary form was measured. The finding is about leaders’ self-serving and abusive behaviour, and applying it to other conduct is inference.
5.3 Why a Rule Works Where a Threat Does Not
Putting 4.3 and 5.2 side by side suggests an explanation, which I offer as reasoning rather than as a finding. A threat of sanction is a consequence that sits somewhere in the future and has to be held in mind while the person is doing something else, which is exactly the condition under which the fear response in 4.3 was reduced. A clear rule, especially one built into a process, sits at the point of decision itself, where attention already is.
That is the difference between the warning and the second signatory in 1.1. The warning asked the calm relationship manager to remember a consequence while chasing a deal. The second signatory put an obstacle in the deal's path just as he was reaching for it, and he didn't have to remember anything. Nobody did.
There is a second part to the explanation. A rule does not ask anybody to feel anything. It works on the calm one because he does not need to be afraid for it to operate, and it works on the anxious one too, because it removes a decision that he would otherwise have to make under pressure, with a brake that fails him under exactly that pressure.
Chapter 6: The Anxious One Under Load
6.1 A Different Engine
The reactive form has been the poor relation throughout this literature and it deserves a chapter of its own, partly because it carries more of the harm and partly because what works on it is close to the opposite of what an organisation instinctively reaches for. The forensic studies in Chapter 2 describe it consistently: high negative emotion, low self-restraint, more anxiety rather than less, and more difficulty with other people (Hicks, Markon, Patrick, Krueger & Newman, 2004; Skeem, Johansson, Andershed, Kerr & Louden, 2007). A study of 129 young adults outside any institution, whose title states its finding, reports that secondary psychopathy, and not primary, went with risky decision-making on a gambling task (Dean et al., 2013).
The picture I draw from this, and it is my reading rather than any one study’s finding, is of somebody whose brake works well enough at rest and fails under provocation or strain, and whose accelerator is quick to engage. That describes the anxious relationship manager’s five good weeks and his one bad afternoon almost exactly. The timing is the tell.
6.2 What a Warning Does to Somebody Already Anxious
Here I move from evidence to reasoning, and I want to mark the transition clearly. If the reactive form is driven by negative emotion that overwhelms restraint, then a response that adds negative emotion, which a formal warning always does, may produce exactly the pattern in 1.1: an immediate and sincere improvement while the fear is fresh, followed by a relapse the first time the fear is joined by pressure from somewhere else. On that reading the warning did not fail because it failed to register. It registered so completely that it became one more source of the strain that produces the behaviour in the first place.
I have watched that sequence more times than I can count, in more than one sector, and I cannot point you to a study that has tested it inside an organisation. What the evidence does support is that the reactive form carries the most measured workplace harm (Roth & Klehe, 2025), that it is characterised by high negative emotion and low restraint, and that people high in it report being drawn to reward like everybody else (Ross et al., 2007).
Evidence status: practitioner judgement, consistent with the mechanism evidence. That the reactive form is marked by high negative emotion and low restraint is evidenced in forensic samples. That a warning can worsen the pattern by adding strain is my reading from practice. No study has tested it at work.
6.3 What This Means for the Levers
Put the two forms side by side and the difference in what each responds to becomes a practical map rather than an abstraction. The cold form is reached least by threat and most by structure, and it is reached by reward like everybody else; the reactive form, in my experience, is reached by threat instantly and briefly, is destabilised by adding strain, and is also reached by reward, so that the one lever the two genuinely share is the one most disciplinary processes use least.
The other lever that works on both is structure that removes the decision from the moment of temptation or pressure, and the one that works least reliably on either is the escalating warning. That is not a comfortable conclusion for most organisations I have worked in, because the escalating warning is the tool most of them have built their processes around. Practitioner judgement.
Chapter 7: Where the Evidence Disagrees With Itself
7.1 A Profile That Does Not Fit
I have presented the two forms more cleanly than the evidence allows, and the newest work shows it. A 2025 study of 446 adults drawn to represent the American population found four profiles rather than two, and the one nearest the reactive form showed a weak brake rather than an ordinary one, with a reduced response to reward (Castagna & Kinrade, 2025), which does not fit Chapter 4 and suggests that the two-form account may not hold cleanly in the general population your colleagues come from.
7.2 The Brain Does Not Agree With the Questionnaires
A 2026 study measured the brake and the accelerator twice in the same people, by questionnaire and by brain-based measures, and only the questionnaires related to psychopathy as expected (Dickison, Neo, Young, McNaughton & Sellbom, 2026). The questionnaire findings did not show up in the brain-based measures, and every finding in 4.2 rests on what people say about themselves.
7.3 Is the Reactive Form a Thing at All
The longest-running complication goes to the heart of the distinction. One line of work finds that the questionnaire’s secondary scale relates more strongly to ordinary negative emotion and low conscientiousness, meaning unhappiness and disorganisation, than its primary scale does (Miller, Gaughan & Pryor, 2008), while a pooled analysis questioning whether boldness belongs in psychopathy at all (Miller & Lynam, 2012) was answered directly by researchers who hold that it is precisely what marks the primary form (Lilienfeld et al., 2012). That argument remains unsettled, and I am not going to pretend to settle it here.
7.4 Does Any of This Change What You Should Do
The test I apply to a dispute like this is whether its outcome would change what a reader should think or do, and here the answer is mostly no, for one reason. Every practical recommendation in Chapter 10 is chosen because it reaches both forms, which means it does not depend on the boundary between them being real.
What the dispute does change is how confidently anybody should talk about types. If the reactive form is partly ordinary distress, then some of the people carrying the most measured harm in 3.1 are struggling rather than dangerous. That should make everybody slower to attach the word to them. It is also the strongest argument I know for the approach this article takes: read responses rather than classify people.
Evidence status: contested. The two-form distinction is disputed at its boundaries, in the general population and at the level of biology. The dispute does not change the recommendations in Chapter 10, which are chosen to work whichever side is right.
Chapter 8: Can Either of Them Change
8.1 The Received Wisdom
The belief that people high in psychopathy cannot change, and that treatment makes them worse, is one of the most widely repeated claims in this territory, and it rests on a real finding. A maximum-security therapeutic community for psychopathic and other mentally disordered offenders was evaluated against a comparison group, and treatment was associated with lower reoffending among the non-psychopathic patients and with higher violent reoffending among the psychopathic ones (Rice, Harris & Cormier, 1992). In my experience, that study is still quoted as proof that nothing works. It proves much less.
The comparison group was matched rather than randomly assigned; the patients were forensic inpatients rather than anybody you are likely to employ. A single programme failing tells you about that programme, not about every possible one.
8.2 What Happened When Somebody Measured Change
A more useful question is whether people who change more do better afterwards, and researchers have asked it directly. Mark Olver, Kathy Lewis and Stephen Wong followed 152 high-risk adult offenders with psychopathic traits through a treatment programme aimed at reducing violent reoffending and then for about five years afterwards (Olver, Lewis & Wong, 2013). Those who changed more during treatment went on to reoffend violently less, and that held after accounting for how psychopathic they were to begin with.
The same study found where the limit sits. Change was smaller among people highest in the cold, emotionally detached core of the trait. So the honest summary is that change is possible, that it matters when it happens, and, on my reading of which core that is, that it is hardest for the form most people are worried about.
8.3 Matching the Response to the Pattern
One study went a step further and tried matching the intervention to the pattern. Baskin-Sommers, Curtin and Newman gave offenders cognitive training aimed at the specific difficulty each subtype showed in the laboratory: for the psychopathic group, training in attending to the wider context while pursuing a goal, and for a group characterised by impulsive, emotionally reactive behaviour, training in regulating emotional reactions (Baskin-Sommers, Curtin & Newman, 2015). Each group improved on the tasks it was trained on and on tasks it was not.
I would not overstate what that shows, because it was a prison sample, the outcomes were laboratory tasks rather than behaviour, and the reactive group is not identical to the secondary form. It is nevertheless the closest thing in this literature to a direct test of the idea this article is built on, which is that the two patterns respond to different things, and it came out the right way.
Evidence status: evidenced in forensic samples; no evidence base at work. Change and its benefits are demonstrated in offender programmes. No study has tested any intervention for either form among employees, and nothing here should be read as saying an organisation should attempt treatment, which is a matter for clinicians.
PART III. BEFORE YOU USE ANY OF THIS
Chapter 9: The People You Will Mislabel
Everything up to here describes two patterns that produce similar behaviour for different reasons. This chapter is about the people who produce the same behaviour for reasons that have nothing to do with either pattern, and in my experience they outnumber the people the research describes by a wide margin. If reading this article makes you more willing to decide that a colleague is psychopathic, of either form, it has failed.
9.1 The One Who Is Simply Overwhelmed
The reactive form in Chapter 6 is anxious, impulsive and prone to lashing out under pressure, and so is a very large number of ordinary people having a very bad year, whether the cause is burnout, a failing relationship, a sick parent, an unmanageable workload or an undiagnosed health problem. None of those is a personality trait.
The question I would ask is about history rather than intensity: a pattern that appeared recently, alongside a change in the person’s circumstances, is far more likely to be strain. If the research instruments struggle to tell distress apart from the secondary form, which is the dispute in 7.3, a manager watching from across the floor should not expect to.
9.2 The Calm, Blunt, Competent One
The cold form is composed, unafraid and willing to put the objective ahead of feelings, and so are many decent executives who are simply unflappable and direct. Among educated, socially skilled employees that disposition went with better performance (Blickle & Schütte, 2017), so the test is whether the calm comes with harm, and a colleague unmoved by your warning may simply have judged, correctly, that it was not warranted.
9.3 The One Responding Reasonably to an Unreasonable Setting
Some rule-breaking is a rational response to a rule that does not work, and some hostility is a proportionate response to being treated badly. The question to ask is whether the behaviour would continue if the setting were fixed. You can often find out by fixing the setting.
9.4 The One Who Needs a Clinician
This boundary is categorical, and the whole series holds it firmly. Psychopathy as studied in the workplace literature is a set of traits on which everybody sits somewhere. It is not a diagnosis, and neither of the two forms described here is a clinical category; antisocial personality disorder, borderline personality disorder, substance problems and a range of other conditions can produce behaviour that overlaps with both forms, and telling them apart is the work of a qualified clinician who has assessed the person, never of a colleague who has observed them.
An observation is not a diagnosis. A questionnaire score is not a diagnosis either, and nothing in this article can be turned into one.
9.5 And the Numbers Do Not Describe a Person Anyway
Every figure here describes large groups, so a finding pooled across 166 samples tells you where harm concentrates and nothing about the relationship manager on your team. That is not a technicality. That is why this article offers a way to read responses rather than a checklist of signs.
9.6 How Common Is It, Really
I have looked for a defensible figure for the proportion of employees who would score high on either form and have not found one I would put my name to; estimates circulate, some of them precise to a decimal place, and I would treat any percentage you are offered as unsupported until you have seen the study behind it and the threshold it used. The earlier article in this series on the bad apple effect reached the same conclusion about the figures most often quoted for psychopathy in organisations, and nothing in this sweep has changed it.
What can be said is narrower, and it is my judgement rather than a measured figure. High scores on either form are uncommon; difficult behaviour at work is not; and the distance between those two statements is where most mislabelling happens.
PART IV. WHAT TO DO
Chapter 10: Watch the Response, Not the Person
Evidence status: mixed and stated per item. Two items rest on peer-reviewed organisational evidence for the primary form only. The rest are reasoned from laboratory and forensic findings or from practice, and no intervention for either form has been tested among employees. Each item carries its own status.
10.1 You Do Not Need to Know Which One It Is
This is the principle the whole chapter rests on, and it is the reason the article has spent nine chapters on a distinction it now asks you not to use. You cannot see which form a colleague has, if either, and you are not entitled to guess. What you can see is how they respond to what you have already done.
A person who improves immediately after a warning and relapses under pressure is showing you something; a person who is unmoved by a warning and changes at once when a process changes is showing you something else, and neither of them needs to be named, labelled or discussed with anybody for that difference to be useful to you. A person who improves and stays improved has shown you that the warning worked, which is the outcome most of the time, and which this article is not about.
So the first move is simply to keep a record of responses rather than of impressions: what you did, when, and what changed afterwards, with dates. That record is evidence about the response, and it is usable without anybody having to be classified. Practitioner judgement.
10.2 Write the Rule Down Before You Raise the Stakes
When behaviour does not change after a first conversation, the instinct is to escalate: a formal warning, then a final one, then a hearing. The evidence in Chapter 5 suggests a different first move, which is to check whether the rule being broken is clear, written, specific and built into the process at the point where it is broken.
In the leader studies, of the three features tested, it was clear that rules in particular weakened the link between primary psychopathy and self-serving or abusive behaviour (Laurijssen, Wisse, Sanders & Sleebos, 2024). A rule that says overrides above the limit require a second named signatory, recorded in the system does work that a rule saying act with integrity cannot, and a rule built into the booking screen does work that a rule in a policy manual cannot. Evidenced for leaders high in the primary form; inference as to other conduct and to the reactive form.
10.3 Put the Consequence in Front of the Decision
If people high in psychopathy register a consequence fully only when attention is on it (Newman, Curtin, Bertsch & Baskin-Sommers, 2010), then a consequence described once in a meeting and then left in a file will be absent at exactly the moment it matters, and the practical translation is to move the consequence to the point of decision: a prompt that appears when the override is attempted, a reviewer who sees the deal before it closes rather than after, a hold that requires the person to stop and do something before proceeding.
None of that is punitive, and it has the advantage of reaching the reactive form as well, since a pause at the point of decision is exactly what a failing brake needs. Inference from laboratory evidence in offenders, not split by form. Untested in organisations.
10.4 Take Load Off Before You Add Pressure
For the pattern in which a warning produces contrition followed by relapse under strain, adding a second, sterner warning adds more of what produced the relapse. The better first question is where the pressure is coming from, and whether any of it can be removed, redistributed or made more predictable: a target that arrives without warning, a deadline stacked on another deadline, a manager who escalates at the moment of greatest load.
I would stress what this is not, because it is easily misread. It is not excusing the behaviour, which should be dealt with on its merits through whatever process the organisation has, and it is not therapy, which is not a manager’s job; it is recognising that the conditions that produce a relapse are partly within an organisation’s control, and that the reactive form carries more of the measured harm than the cold one (Roth & Klehe, 2025). Practitioner judgement, consistent with the mechanism evidence. Untested as an intervention.
10.5 Reach for the Accelerator Where You Would Reach for the Brake
Reward is the shared lever. Both forms report being drawn to reward (Ross et al., 2007), and the cold form reports the least sensitivity to threat. That suggests making the behaviour you want the rewarded route rather than making the behaviour you do not want the punished one: tie the recognition to the clean file rather than tying the sanction to the dirty one, and make the second signatory quick to obtain rather than slow to bypass.
I am aware that rewarding people for doing what they should have done anyway sits badly with most managers, and I have some sympathy with that. The point is not fairness but effectiveness, and on this evidence, a system built entirely on the threat of punishment rests on the lever least likely to reach the person you are most concerned about. Inference from laboratory evidence. Untested in organisations.
10.6 Three Refusals
Do not get tougher as the default. Employees high in the cold form reported more positive mood and engagement under abusive supervisors, and in a field study less anger than their peers under them (Hurst, Simon, Jung & Pirouz, 2019). Harshness is the setting in which that form does comparatively well, which makes it a poor tool for reaching it. Evidenced for the primary form, from two studies.
Do not screen for either form. I would not use either instrument for selection. They were tested in anonymous, low-stakes conditions that selection does not resemble, nobody can say with confidence what the secondary scale measures, and the same disposition predicts misconduct in one person and strong performance in another. Practitioner judgement, reasoned from Chapters 3 and 7.
Do not tell anybody what they are. The conversation that works is the one built from the record in 10.1: what was agreed, what happened, on which dates, and what it cost the work. Not you do not seem to care about consequences, which is an argument about your judgement that you will probably lose, but on the fourth and the eleventh the limit was exceeded without a second signatory, and the file now needs to be reviewed by credit risk before the client’s facility can be renewed. The second version is true whatever is going on inside the person, which is the only kind of claim about somebody else worth making. Practitioner judgement.
Chapter 11: What This Changes
11.1 The Short Version
If you read nothing else in this article, read this. The word psychopath covers two patterns, and they respond to different things. One is cold, calm and relatively unresponsive to the threat of punishment; the other is anxious, impulsive and prone to lose control under strain. Both are drawn to reward.
The cold one is not doing most of the measurable damage. Across 166 samples and 49,350 psychopathy scores, the harm psychopathic traits do to performance, helpfulness and conduct at work was mostly more pronounced for the anxious, impulsive form. Fear is not absent in psychopathy; it depends on attention. In 125 prisoners high in the trait, the fear response was normal when attention was on the threat and reduced when it was elsewhere, which is why a consequence left in a file does less than one placed at the point of decision.
Clear rules help, and harshness suits the wrong person. Across three studies of leaders, clear rules in particular weakened the link between the cold form and self-serving or abusive behaviour, and in two studies employees high in the cold form reported more positive mood and engagement under an abusive boss. Change is possible and hardest where people most expect it. Among 152 high-risk offenders followed for about five years, those who changed more reoffended violently less, and change was smaller among those highest in the cold core of the trait.
11.2 So What Do You Do
Three things, and none of them requires you to decide what anybody is. Read the response rather than the person. Improvement followed by relapse under load and indifference followed by instant change when a process changes are two different signals, and a dated record of what you did and what followed is the only reliable way to see which you are getting.
Change the structure before you change the temperature. Make the rule clear, specific and present at the point where it is broken, and put the consequence in front of the decision rather than in a file. Take load off before you add pressure, and reach for reward before threat. Both are cheap to try, both reach both patterns, and the escalating warning, which reaches neither for long, is the thing to reach for last rather than first.
11.3 Why Any of This Matters
Two reasons, and the second is the one I would leave you with. The first is practical: most organisations have built their response to difficult behaviour around a single escalating ladder of warnings and sanctions, and on this evidence that ladder is well designed for the colleague who improves after the first rung and badly designed for both of the patterns this article describes. People spend months climbing it and arrive at a hearing with a file that proves only that the ladder did not work.
The second is less comfortable. The popular picture of the dangerous colleague, the composed one who feels nothing, has sent attention towards the smaller half of the problem and away from the larger one, which is somebody anxious and struggling whose behaviour, in my experience, gets worse under exactly the pressure organisations apply to fix it. That is not an argument for sympathy instead of accountability. It is an argument that accountability designed around the wrong picture does not work.
11.4 What I Still Do Not Know
Four things remain open, and I would rather state them than resolve them by omission. Whether the laboratory findings on attention and on the brake operate in an office at all, since every one of them was established in prisons or in student samples and none was tested on employees. Whether any of the moves in Chapter 10 improves anything, since not one of them has been tested against either pattern inside an organisation.
How large the difference between the two forms actually is in the workplace meta-analysis, since I have been able to confirm its direction and not its size. And whether the secondary form is a distinct pattern or largely ordinary distress measured by an imperfect scale, which is the dispute in Chapter 7 and which bears directly on how many of the people carrying the most measured harm are dangerous and how many are simply struggling.
A closing note on how I have handled the material. Psychopathy is the most sensational word in this literature and the one where the temptation to write about monsters is strongest. I have tried throughout to keep the forensic findings and the workplace inferences visibly apart, to say where the evidence disagrees with itself, and to give you a way of reading responses rather than a way of deciding what your colleague is. What I would most like to have made harder is a confident verdict about the person you had in mind while you were reading.
