I once worked with a boss who used to say, quite openly, that he made up his mind about a candidate the moment the person walked into the room. Before a single question had been asked. He did not say it apologetically. He said it as a statement of professional instinct, and I think he genuinely believed it was one.
I once worked with a boss who used to say, quite openly, that he made up his mind about a candidate the moment the person walked into the room. Before a single question had been asked. He did not say it apologetically. He said it as a statement of professional instinct, and I think he genuinely believed it was one.
I have heard versions of the same claim many times since, from colleagues in business and from colleagues in HR. Some of them are among the shrewdest judges of people I know.
And looking back honestly at my own interviews, I have caught myself doing it too. Not deliberately, and never in those words, but the view was often settled well before the evidence was in.
What has stayed with me is what happened on the occasions when I distrusted that first view and made myself probe further. More often than I am comfortable admitting, the first view did not survive the probing. A candidate I had quietly set aside turned out to have precisely what the role needed. Another I had warmed to early had rather less underneath it than I had assumed.
That is an awkward thing to put in writing when one runs a hiring company. But I have yet to meet a good interviewer who could not say the same, and I have met a few who would not.
The uncomfortable question, then, is not whether we form impressions early. We do, and we cannot really help it. The question is what those impressions do to everything we hear afterwards.
A young lieutenant with a clipboard
In 1955, a twenty-one year old psychology graduate was posted to the psychology branch of the Israeli army. His job was to help decide which recruits should go where. The army had a system already. Interviewers, most of them young conscripts themselves, would sit with a recruit, have a fairly open conversation, form a view, and write down whether the fellow seemed suitable for combat duty.
The interviewers were bright. They were confident. Their assessments were close to worthless as predictions of how the recruits actually performed later.
The young lieutenant was Daniel Kahneman. He was a psychologist, and he would later win the Nobel Prize in Economics, which tells you something in itself. The economists had a tidy account of how people ought to make decisions. Kahneman spent his career showing them how people actually do.
His diagnosis was not that the interviewers were foolish or that they lacked information. It was that they were forming a global impression of the person too early, and then hearing the rest of the interview through that impression. A candidate you have decided you like gives an ambiguous answer, and you hear it charitably. A candidate you have written off gives the same answer, and you hear evasion. Nobody does this deliberately. That is rather the problem.
So he redesigned the interview.
Interviewers were now required to assess six specific characteristics, ordinary things like responsibility and sociability. For each one they had to ask a set of factual questions about the recruit’s life before the army. What work had he done. Had he been regular at it. Did he keep up with friends. They had to score one characteristic fully before moving on to the next. And they were told, in no uncertain terms, not to arrive at any overall view of the man until all six were done.
The interviewers hated it. They complained that they were being turned into robots and that their professional judgement was being taken away from them.
Kahneman made them a concession, and this is the part of the story I find most interesting.
Once you have finished all six ratings, he told them, close your eyes, picture this recruit as a soldier, and give me a number out of five. Whatever your gut says.
Two things came out of it. The structured scores predicted actual performance considerably better than the old free-form interviews had done. In his Nobel autobiography Kahneman records the validity improving from roughly 0.10 to something in the region of 0.30, which sounds modest until you consider that the starting point was nearly useless.
And the closed-eyes number, the pure gut call, turned out to be as good a predictor as the six structured ratings put together. Better still, the two combined outperformed either on its own.
So intuition was not the enemy. Premature intuition was. The instruction that made the difference was not to stop trusting your judgement. It was to wait.
Why this lands rather close to home
I would suggest Indian hiring is, if anything, more intuition-led than most.
Think of a campus drive where a panel sees four hundred students across two days. Think of the walk-in weekend for a retail or BPO expansion. Think of how often "culture fit" is the reason recorded for a rejection, and how rarely anyone is asked to say what exactly was not fitting. Think of how much weight a candidate’s college, or accent, or the confidence with which he shakes hands, quietly carries in that room.
None of these impressions is necessarily wrong. Experienced people do pick up real signals, and my old boss was not a fool. But when the impression arrives before the candidate has sat down, and the evidence keeps arriving until minute forty, we should at least be honest about which one is doing the work.
Ask yourself, after the last interview you conducted: did the second half of the conversation genuinely have a chance of changing your mind?
What the research has established since
Kahneman’s army experiment was not the last word. Structured interviewing became a serious field of study in personnel selection over the following decades.
Campion, Palmer and Campion’s 1997 review pulled together fifteen distinct components of interview structure, covering both what you ask and how you evaluate what you hear. A large review by Levashina, Hartwell, Morgeson and Campion in 2014 went through twenty years of research that followed, looking at question types, rating scales, probing, impression management by candidates, and how candidates themselves react to structured processes.
The findings are not a magic formula. Simply asking every candidate the same questions does not, by itself, produce good hiring. Structure has to be applied to the evaluation as well, not just to the questioning. Poorly designed structure can be worse than none.
But the central principle has held up well for seventy years. Evidence should be gathered and evaluated systematically before it is collapsed into a single verdict.
Which brings us to AI, and to a caution
Technology now allows us to structure hiring at a scale Kahneman could not have imagined with his clipboard. We can define competencies for a specific role. We can put the same questions to every candidate without fatigue setting in by the fortieth interview. We can capture far more of what a candidate actually said, rather than what a tired panel member remembers by evening. We can evaluate against stated criteria and make candidates genuinely comparable.
All of that is real, and it is available today.
Here is the caution, and I say this as someone who runs a company in this space.
A good deal of what is being sold as AI hiring at the moment is selling speed. Screen ten thousand CVs by lunchtime. Interview at scale. Reduce time to hire. Speed is a legitimate benefit and I am not going to pretend otherwise, but speed is not accuracy, and a score reported to two decimal places is not the same thing as evidence.
If we replace an interviewer’s unexplained hunch with an algorithm’s unexplained output, we have not solved Kahneman’s problem. We have automated it. The hunch at least came from someone you could question over a cup of tea.
The opportunity is not to hand over the judgement. It is to structure the evidence on which the judgement rests, and then to hand that evidence to a human being who is accountable for the decision.
Better evidence, then better judgement
At Sainterview this sits at the centre of what we call Structured Hiring Intelligence, and the emphasis belongs on the middle word.
It means being explicit about what is being assessed for a given role. Capturing evidence consistently across every candidate. Interpreting that evidence against criteria that were fixed in advance rather than adjusted afterwards to suit an impression. Keeping the assessments comparable. And then putting all of it in front of the hiring manager rather than putting a verdict in front of him.
The recruiter still decides. The hiring manager still decides. They simply decide later, and on a firmer footing.
Kahneman worked all this out with a pencil, a questionnaire and a group of irritated young conscripts, decades before anyone had heard of a large language model. The tools have changed beyond recognition. The instruction has not.
Structure the evidence. Delay the judgement. Then make the call.
Better hiring may not need less human judgement. It may simply need better evidence to reach it.
________________________________________________________________
What is the earliest you have ever made up your mind about a candidate, and were you right? I would be interested to hear.
The Science Behind Better Hiring Decisions is a series in Hiring Intelligence exploring what psychology, behavioural science, decision science and personnel-selection research can tell us about hiring better.
________________________________________________________________
Sources
Kahneman, D. Autobiography, The Nobel Prize. Kahneman’s own account of designing a structured interviewing system for the Israeli Defence Forces and what it produced. He tells the story again at greater length in Thinking, Fast and Slow (2011).
Campion, M.A., Palmer, D.K. & Campion, J.E. (1997). "A Review of Structure in the Selection Interview." Personnel Psychology, 50(3), 655 to 702.
Levashina, J., Hartwell, C.J., Morgeson, F.P. & Campion, M.A. (2014). "The Structured Employment Interview: Narrative and Quantitative Review of the Research Literature." Personnel Psychology, 67(1), 241 to 293.
