Latest / Hiring & Firing with Fexingo: Recruitment, Performance, and Workforce Decisions / Why Personality Tests Are Failing Your Hiring Process
Transcript
- Lucas: So, roughly four out of five Fortune 500 companies now use some form of personality test during hiring. Myers-Briggs, the Big Five, DISC, Hogan — pick your acronym. And collectively, companies spend well over half a billion dollars a year on these assessments. Luna: And yet, I've seen the data that says these tests predict almost nothing about how someone actually performs on the job. Is that right? Lucas: That is exactly right. There was a major meta-analysis out of the University of Minnesota in 2024, published in the Journal of Applied Psychology, that looked at over 300 studies. The headline finding: personality traits explained, on average, about six percent of the variance in job performance. Luna: Six percent. That means ninety-four percent of what determines whether someone succeeds is not captured. Lucas: And companies are making hiring decisions based on that six percent slice. It gets worse when you consider a specific problem called the Barnum Effect. You know — the phenomenon where people accept vague, generalised personality descriptions as uniquely applicable to them. Luna: Right — 'You have a great need for other people to like and admire you.' That's from a famous study by Forer in 1948. And it works because the statements are true of almost everyone. Lucas: Exactly. Now, fast-forward to 2025. A company called BrightHire — which builds hiring software — decided to run their own controlled experiment. They had two hundred job candidates complete a standard Big Five personality test. Then they gave hiring managers the results, but they randomised which candidate's results went with which resume. The hiring managers still felt the personality profiles were 'spot on' eighty percent of the time. They couldn't tell the difference. Luna: So the managers were essentially reading horoscopes and calling them data. Lucas: That's the polite way to put it. And the problem is incentive: the test vendors charge per assessment, so they have no reason to publish negative results. Meanwhile, HR teams buy the tests because they want something that looks objective and scientific. Luna: But wait — there is some research showing that conscientiousness is a decent predictor of performance, especially for sales roles. So are we throwing out the baby with the bathwater? Lucas: Fair point. Conscientiousness does have the strongest correlation — about point-two-two correlation with performance, which is still weak. The meta-analysis found that adding conscientiousness to a cognitive ability test improved prediction by only about five percent. Meanwhile, the same study found that structured work samples — giving candidates a real task they'd actually do on the job — showed a correlation of point-five-four. That's nearly three times stronger. Luna: So why don't companies just use work samples all the time? Lucas: Because work samples are harder to scale. You can't email a link to five hundred applicants and have them all complete a thirty-minute simulation — or can you? That's exactly what a startup called Micro-Job Sims has been doing. They build short, automated simulations for common roles: customer service, data entry, project coordination. One company — a mid-sized logistics firm called Pinnacle Distribution — used their simulation for warehouse supervisor hires. Luna: And what happened? Lucas: Over twelve months, they hired forty supervisors. Twenty via their old process — resume screen plus personality test plus panel interview. Twenty via the simulation plus a shorter, structured interview. The simulation group had a thirty-four percent lower turnover rate in the first six months, and their managers rated them higher on a ten-point performance scale by an average of one point eight points. Luna: Thirty-four percent lower turnover — that's a huge cost saving. The Center for American Progress estimates replacing a salaried employee costs roughly twenty percent of their annual salary. Lucas: Right. So for a warehouse supervisor making fifty thousand dollars, replacing one costs about ten grand. If you hire twenty, and you save even three of them from quitting, you've already covered the cost of the simulation tool. Luna: But personality tests aren't going away overnight. They're embedded in HR systems. What's the practical advice for a hiring manager listening right now? Lucas: I'd say: don't use personality tests as a screen. Don't use them to reject candidates. If you must use one, use it as a conversation starter — give the candidate their own results and ask, 'Does this match how you see yourself?' Then watch for self-awareness. But the real play is to invest in building even a single, short work sample. If you're hiring for a customer service role, have them respond to three mock emails. It takes twenty minutes and it's far more predictive than a hundred-question personality inventory. Luna: And if you're the candidate on the other side, what do you do when you're asked to take a personality test? Lucas: You take it — you don't opt out, because that's a red flag. But you should know that the test is unlikely to decide your fate. The interview and your work sample matter far more. Also, there's a well-known phenomenon called 'faking' — candidates can manipulate their answers to appear more conscientious or agreeable. And studies show that about thirty percent of candidates do fake to some degree. Luna: So the tests are not only weak predictors, they're also easily gamed. That's a double failure. Lucas: It really is. And it's not just the tests themselves — it's the false confidence they give hiring managers. When a manager sees a 'high agreeableness' score, they subconsciously lower their guard. They stop looking for evidence of conflict or stubbornness. The test becomes a cognitive crutch. Luna: I want to pivot slightly because this conversation — about tools that seem scientific but aren't — it's exactly the kind of thing that makes us think about the resources we rely on as professionals. And honestly, if today's episode helped you think more critically about the hiring tools in your own toolkit, we'd love it if you considered tossing a coffee our way. It's buy me a coffee dot com slash fexingo. Lucas: Yeah, that link is literally just a way for listeners who find value here to keep the show ad-free and independent. No pressure ever. And it genuinely helps us keep digging into topics like this one. Luna: Exactly. So back to the data — Lucas, you mentioned structured interviews earlier. How do they compare to personality tests in terms of predictive power? Lucas: Structured interviews — where every candidate is asked the same questions, scored on a consistent rubric — have a predictive validity of about point-five-one. That's nearly as good as work samples and dramatically better than personality tests. The problem is that most interviews are unstructured: 'Tell me about yourself,' 'Where do you see yourself in five years?' Those have validity of about point-two-zero — no better than a coin flip. Luna: So if you're a hiring team with limited budget, the single highest-impact change you can make is to switch from unstructured to structured interviews. Lucas: Exactly. And do it before you spend a dime on personality testing. I've seen companies spend ten thousand dollars on a test vendor and then still conduct a thirty-minute 'culture fit' chat that carries more weight. The math is backwards. Luna: What about the argument that personality tests help with team composition — like, ensuring you don't hire someone too introverted for a sales floor or too abrasive for a collaborative team? Lucas: That sounds logical, but the evidence doesn't support it. The same Minnesota meta-analysis looked at team-level outcomes and found that personality diversity within teams explained less than one percent of team performance. What mattered more was psychological safety — whether team members felt they could speak up without being punished. That's built through leadership behavior, not hiring filters. Luna: So the test doesn't build the culture. The manager does. Lucas: Exactly. And that's a much harder thing to sell, because it means the person doing the hiring has to get better — not just buy a better test. Luna: We've covered a lot. Let's leave listeners with one concrete takeaway. Lucas: Sure. If you are responsible for hiring, next time you're about to require a personality test, ask yourself: 'What would happen if I replaced this with a thirty-minute work sample?' The answer is probably: you'd make better hires, and you'd save money. And if you're a candidate, don't stress about the test. Focus on demonstrating what you can do. Luna: And maybe bring a coffee for the road. Thanks for listening.