Everyone knows this. Everyone does it anyway.
The evidence on structured, demonstrated-work hiring has been clear since before most of us were in the workforce. Both sides of the table default to promises anyway. Practitioners know the research. They hire against it anyway, and the behaviour is symmetrical.
Why the interview feels rigorous from the inside and isn’t.
Confidence is not accuracy. Interviewers exit unstructured rooms with strong reads on candidates they barely measured. The strength of the impression is what makes the read feel like data.
“Culture fit” is a reaction wearing the costume of a reason. When a debrief cannot name a specific behavior, the interview produces a mood and reports it as a verdict.
The house edge: a confident talker beats a quiet expert in an unstructured room, reliably. The format scores fluency in the conversation. Competence at the job is inferred from it. The structural advantage runs to the same kind of person every time.
The black hole and what it costs.
Most candidates never hear back. The application goes into a system that rejects on proxies — school, keyword match, employer prestige — before any human sees them. It is where the most candidates are rejected and the least is demonstrated. We have not found a study ranking it against the other stages, and we are not going to assert one.
Candidates learn to perform rather than demonstrate. That is a rational response to a system that scores performance. Interview coaching, resume optimisation and personal branding are all strategies for winning a promise contest. Those signals measure adaptation. They are read as capability.
A fair-shot mechanism looks like this: a short, realistic task at the top of the funnel, scored blind. It costs the employer minutes per candidate, not hours, and gives every applicant the same chance to be measured on the work itself.
Three findings that explain the gap between what the field knows and what it does.
Interviewers form confident impressions from random answers. In one experiment by Dana, Dawes and Peterson, the people being interviewed secretly answered yes-or-no questions at random. They looked at the first letters of the last two words of each question and said yes if both letters fell in the same half of the alphabet, and no if they did not. The students running the interviews never noticed, and rated those interviews as just as informative as the genuine ones. Students who ran an interview then predicted the person’s grades less accurately than students who simply looked at their past grades — a correlation of .31 against .65.
Dana, Dawes & Peterson (2013), Judgment and Decision Making 8(5).
Practitioners rate their own intuition above the methods that outperform it. In a widely cited review, Highhouse assembled the evidence that hiring professionals keep trusting informal interviews and their own judgement over tests and scoring systems, even when they know the research points the other way. One survey he cites, of 201 human resources executives, found they rated the informal interview more effective than any of the written tests they were asked about.
Highhouse (2008), Industrial and Organizational Psychology 1(3).
When every finalist is credentialed the same, evaluators decide on cultural similarity. Rivera embedded with recruiters at elite law, banking and consulting firms. The interview was where evaluators tested whether the candidate was “someone we’d want to be stuck with.” Ability was assumed from the resume. Fit was the tiebreaker — and fit tracked class background.
Rivera (2012), American Sociological Review 77(6).
“What would you actually bet on this hire?”
Worth asking of any process built on impressions. If the answer is “not much,” the process produces a preference and dresses it as a decision.