Lesson 34 of 44
Overview
This episode unpacks why polished interviews so often miss the mark and how unstructured hiring can reward charm over actual job behavior. It also explores how structured interviews, clear scoring, and careful use of AI can improve fairness and prediction without scaling bias.
Welcome to The Science of Leading. I’m Claire Monroe with Edwin Carrington. Let me start with something painfully familiar. Candidate walks in, great handshake, smooth story, confident eye contact, says everything you want to hear… and six months later, you’re quietly wondering how you got it so wrong. And the uncomfortable truth is, most people in that room thought they were being careful. They weren’t. They were being impressed. And those are two very different mental processes. “Impressed” is doing a lot of work there. Because interviews do seem to reward polish. Fluency. Likeability. Sometimes just… being similar to the interviewer. Not necessarily whether this person can actually deliver results when things get messy. Exactly. Most hiring still operates on a kind of quiet assumption that you can infer future behavior from a short conversation. Thirty, maybe forty-five minutes, and suddenly we believe we’ve assessed judgment, resilience, integrity. In reality, we’ve mostly observed how someone performs in an interview setting. That’s such a subtle shift. So the candidate might be excellent at the very specific job of interviewing… not the actual job they’re being hired to do. Yes. Interviews often turn into auditions for confidence. And confidence is persuasive to the point of being misleading. Without evidence, it’s just presentation. The deeper issue is that organizations still rely on intuition as if human behavior reveals itself quickly and cleanly. It doesn’t. And once you believe you can “read people,” you stop demanding proof. You just… trust your impression. Which, to be fair, a lot of leaders say with absolute certainty. They do. I’ve said it myself in the past. And I was wrong often enough to stop trusting that instinct. Strong hiring systems don’t ask, “Did I like them?” They ask, “What did they actually demonstrate, and did we evaluate that consistently?” So this is where structure comes in. Because without it, interviews are basically… improvisation. Sometimes insightful, sometimes all over the place, and almost impossible to compare. That’s a good way to frame it. In an unstructured interview, one candidate gets deep questions about conflict, another talks about career goals, and a third ends up telling a story about a marathon. Then we line them up and pretend the comparison is fair. It isn’t. And when people talk about “FBI-style” interviewing, what they really mean is discipline. Same questions, clear signals, observable behavior… not just gut feeling. Exactly. Strip away the branding and it’s very straightforward. Ask comparable questions. Look for concrete examples. Understand what happened, what the person did, and what changed as a result. Then evaluate that against defined criteria. That’s closer to investigation than impression. That structure matters. Because if someone says, “I’m a strong leader,” that’s… nice, but meaningless. If they say, “I took over a team of 12, turnover was high, I changed how we ran one-on-ones, clarified roles, and retention improved,” now you actually have something to work with. Exactly. Past behavior isn’t perfect, but it’s far more reliable than charisma. And it improves fairness. Because when decisions are driven by instinct, instinct tends to favor what feels familiar. That’s where bias quietly enters the system. And then we layer technology on top of that and call it progress. That’s where “algorithmically mediated inequality” comes in, right? Bias, but now it’s automated and harder to see. Yes. If you digitize a flawed process, you don’t remove bias, you scale it. And if the system is opaque, managers don’t even understand why decisions are being made. That’s not rigor. That’s outsourcing responsibility. So let’s make this practical. If someone actually wants to improve hiring, what do strong systems do differently? They define success before the interview begins. They standardize questions. They focus on job-relevant behaviors. And they score answers based on evidence, not delivery style. The research on this is consistent. Structured interviews outperform intuition. And when we say structured, we’re not just talking about rigid scripts. There are specific types of questions that work better, right? Yes. Three categories tend to be effective. Past behavioral questions, where candidates describe what they’ve actually done. Situational questions, where they explain how they would handle a realistic scenario. And background questions, which explore relevant experience in a disciplined way. Across multiple studies, these approaches predict performance more reliably than informal conversations. So asking someone to explain a concept perfectly… isn’t necessarily the best signal? Correct. Knowledge and performance are related, but they’re not the same. Someone can sound highly competent and still struggle with execution, prioritization, or decision-making under pressure. Evidence of behavior tends to be more predictive than polished explanations. That explains why some interviews feel like trivia contests. You’re testing recall, not real-world capability. Exactly. And those formats can be appealing because they make the interviewer feel sharp. But hiring isn’t about validating the interviewer. It’s about making a better prediction. Which brings us back to AI. Because there’s a lot of excitement around speed and scale. But if the underlying criteria are weak, you’re just accelerating bad decisions. Precisely. Poor inputs at scale produce poor outcomes at scale. If algorithms are built on vague proxies or biased historical data, you get faster, more consistent mistakes. Efficiency without clarity is not progress. So the real value of AI isn’t replacing judgment. It’s forcing clarity in how decisions are made. That’s the useful application. Define criteria clearly. Standardize evaluation. Make scoring transparent. Keep humans accountable. When tools support that process, they reduce noise instead of hiding it. That’s where platforms like OAD.ai can be useful, if they’re used to structure evidence and improve comparability. So for anyone hiring next week, the move isn’t to “trust your gut more.” It’s to define what good looks like, ask consistent questions, and evaluate actual evidence. Not delivery, not confidence… evidence. Exactly. And one useful mindset shift is this: you’re not trying to feel certain after a short conversation. You’re trying to reduce the odds of being wrong. And that’s a much higher bar. If you want to put that into practice, you can test tools like OAD’s behavioral assessments for free at o-a-d-dot-a-i. It’s a straightforward way to bring more structure and clarity into hiring. Thanks for listening. See you next time on The Science of Leading.