Skip to main content

Verified by Psychology Today

How to Predict Toxic Behavior

Is it really that hard to assess integrity?

Key points

  • We can and should assess integrity, unless we want to perpetuate toxic and parasitic leadership.
  • There is a well-established psychological science to detect unethical behaviors before it's too late.
Source: fizkes/Shutterstock

Too many leaders lack integrity. This is why antisocial and destructive leaders are all too familiar—Stalin, Hitler, Mao, Madoff, Skilling, Weinstein, and Blatter, to name just a few.

As Dan Hough notes in his brilliant book Analyzing Corruption, $2.6 trillion is lost each year to corruption, and reports by the World Bank estimate that around $1 trillion is paid out in bribes every year. On top of this, unethical leaders cause incommensurable harm and suffering in the form of counterproductive work behaviors, such as harassment, bullying, and intimidation, not to mention disengagement, stress, and alienation at work.

Sadly, there is a general tendency to focus on leaders’ skills, assertiveness, and technical expertise, at the expense of their toxic and parasitic tendencies. And yet, there are measurable and detectable traits that explain the malevolent nature of humans, and they tend to comprise moral dysfunctions or integrity deficits.

Curiously, psychological research has long highlighted integrity as one of the strongest and most consistent predictors of future job performance, including in management and leadership roles. In contrast, there are few signs that either political or corporate leaders are selected on the basis of their integrity, not least because of a lack of awareness on how to reliably measure it, let alone predict it. For instance, ask the average voter or hiring manager whether integrity is a critical leadership trait, and they will no doubt agree. But if you ask them how we can actually measure it, they will likely be left rather clueless.

Assessing Integrity

Is it really that hard to assess integrity? The simplest academic answer to this question would be “not really.” Indeed, over the past 100 years, scientific research has provided a wide range of tools and alternatives for determining a person’s probability to act with integrity, across different situations and relative to other people.

Naturally, there are also known flaws and limitations with any of these approaches, but refraining from using them just because they aren’t perfect is a questionable and risky decision, particularly when it results in ignoring integrity altogether. The aim in science is rarely perfection, but finding better ways of being wrong.

Importantly, we know that there are many better ways of evaluating someone’s integrity than relying on the single subjective opinion or view of one individual, irrespective of how well they know the person in question. Here are a few:

1. Personality tests. Although there are known external influences predicting and explaining why the same individual may act without integrity in some contexts but not others (see point 5), people are generally consistent in their displayed patterns of morality and immorality. Nobody wakes up like Mother Theresa on a Monday but goes to bed like Joseph Stalin on a Friday.

We can model these patterns using personality assessments, and universal personality frameworks such as the Big Five or Five-Factor model contain predictive algorithms for integrity. Most notably, if someone scores low on agreeableness and conscientiousness, their antisocial and unethical tendencies go up.

Unfortunately, it is also true that if we selected people with extremely high scores on these two traits we would end up with highly obedient, conforming leaders who, in the wrong context and with the wrong brief, would be as moral as a concentration camp guard. Note there are also dark side personality traits, such as the dark triad, which predict toxic and destructive leadership behaviors. Unsurprisingly, you can expect narcissistic, psychopathic, and Machiavellian leaders to over-index among corrupt, despotic, and abusive leaders; I have written about them extensively. One of the problems here is that we are often seduced by these characters, who may appear charismatic, charming, and entertaining at first sight. This explains why psychopaths and narcissists often perform well in job interviews, which should not be used to infer integrity.

2. Intelligence tests. Few people like IQ tests, but they have been shown to be a consistent predictor of corrupt and unethical behaviors at work. The smarter you are, the more integrity you tend to show in a work setting. Of course, the correlation isn’t perfect (it never is in the social sciences because humans are more complex than fish or squirrels—at least according to humans). This is why many of the examples I listed at the beginning of this post could be assumed to be highly intelligent.

As Warren Buffet once noted, ideally we want leaders to have intelligence, ambition, and integrity—and if the last one is missing, you should really hope they are low on the first two. Anecdotally this feels right, but the plural of anecdote is not data. And the data strongly suggest that if you select leaders on the basis of intelligence, you could greatly reduce the incidence of antisocial and toxic behaviors at work. Period.

3. Aggregated observer ratings (multisource of 360-feedback). At work, and in any area of life, reputation matters. We may make errors when we try to infer people’s honesty and integrity in a short-term interaction, but once we co-exist with them for a long period of time we should pick up the signals that help us build a consistent model of their reputation. This is why one of the best sources of data to assess a manager’s integrity is their direct reports.

You can think of integrity as the gap between what managers say and do (the smaller the gap, the bigger their integrity). Who is better placed to assess this gap than the people who interact with them every day, and have the largest number of data signals to evaluate their behavior? Ask everyone in a team if they trust their boss, and if they all say yes you can be sure that their boss has high integrity—and vice-versa. Sadly, we rarely look at 360-degree or multisource feedback to decide on leader promotion or selection, but we should. You cannot fool all the people all the time.

4. Past performance (biodata, track record, big data). People are consistent, so past behavior is the best predictor of future behavior. If someone has a track record for being honest, altruistic, and nice, trust them. And if they don’t, please beware of the dangers. Today we have numerous sources of data—references, resumes, past 360s (back to point 3)—to improve our understanding of someone’s integrity. We should ignore these data at our peril.

Everything we do is a product of our choices, and while moral standards are always a mix of subjective, cultural, and objective, we should make the effort to interpret people’s lives in the context of moral choices. Steve Jobs once noted that it is easy to join the dots backward. That is partly because we are too lazy or afraid to join them forwards. To be clear, I am not advocating for a minority report here—just a commonsensical and courageous inspection of people’s track record.

5. Natural language processing. As we have shown, it is possible to translate people’s language use and word preferences into a predictive model of their dark side personality traits. Negative, aggressive, overconfident, and antisocial language signals negative, aggressive, overconfident, and antisocial personality traits. While AI is still not a mainstream methodology to mine these data and infer people’s integrity, we can expect it to become more prevalent in the future. Of course, this would also increase the incentive to “fake good” when it comes to digital communication (be it internet or intranet, social media, or email). And you can always expect smart narcissistic Machiavellian individuals to be good at faking it. But for every false negative we miss, there will still be more true positives we catch.

6. Contextual measures (toxic and parasitic cultures). Oscar Wilde once noted that anyone can be faithful in the countryside—for there are no distractions. In line, even nice people may be nudged to misbehave in a toxic or parasitic environment. Because of this, we need to take into account environmental or contextual measures of integrity, to assess whether we may have a bad barrel, one likely to corrupt or rot even the good apples. This may be one of the most underrated and underutilized uses of climate surveys, but there is time to correct this.

Perhaps the issue is not so much that integrity is hard to measure, but that we overestimate our ability to evaluate it intuitively, or that we don’t care enough about the consequences of picking leaders who lack it. From an ethical standpoint, it is obviously questionable to reject or deselect individuals from leadership roles because of their propensity or potential for behaving in unethical ways, but if the alternative is to assume that everyone is equally risky or risk-free, or that our gut feeling represents an accurate measure of the person’s moral profile, we are surely worse off—and everyone else, too.

LinkedIn and Facebook image: fizkes/Shutterstock