What actually biases a first-round interview (and what doesn't)

Ask a hiring manager what swayed a first-round call and you will rarely hear “their accent.” You will hear “strong communicator,” “great energy,” “we just clicked.” Those phrases feel like judgment. Often they are cover for something older and less flattering: a reaction to how a person sounds, looks, and carries themselves that lands before the content of their answer ever does.

Most first-round bias is not malice. It comes from signals that feel informative and are not — and the harder you try to ignore them by willpower, the more they leak in through the side door. The fix is not to be a better person in the room. It is to change what the room measures.

The signals that sway us

Accent and name. In a controlled study of interview judgments, an applicant’s accent and name shifted how favorably they were rated: an applicant with an ethnic-sounding name speaking with an accent was viewed less positively than the same profile without those cues, and those judgments fed straight into the hiring decision (Purkiss et al., 2006). None of that is competence. It is the sound of a voice doing work that only the answer should be doing.

Confident delivery. A candidate who speaks in clean, certain sentences reads as capable. But fluency is a skill of its own — trained by privilege, practice, and a first language that happens to match the interviewer’s. The nervous candidate who gives the better answer loses to the smooth one who gives a worse one, and everyone in the room calls it a gut feeling.

Shared background and rapport. The interview that “flows” is usually the interview with someone like you — same schools, same hometown, same references. Rapport feels like signal because the conversation is easy. Ease is not evidence. It is similarity, and similarity is exactly the thing a fair process is supposed to see past.

The résumé halo. A familiar logo at the top of the page tilts everything that follows. Once you have decided someone from a brand-name company must be good, you hear their answers as confirmation. The halo is not evidence of the competency in front of you; it is a prior you forgot you were holding.

Why unstructured interviews amplify all of it

Here is the mechanism. When every conversation is different — different questions, different order, different follow-ups — there is no shared standard, so each interviewer fills the gap with a private one. And private standards are exactly where accent, rapport, and polish do their quiet work. The looser the structure, the more room the noise has to move.

The research is unusually consistent on this. In Schmidt and Hunter’s landmark synthesis of 85 years of selection research, structured interviews predicted job performance far better than unstructured ones — a validity of .51 versus .38 (Schmidt & Hunter, 1998). More recently, after Sackett and colleagues re-examined decades of that research and corrected a systematic overcorrection in how the numbers had been adjusted, structured interviews came out ranked at the very top of the methods we use to predict performance (Sackett et al., 2022). Same job, same candidates — the difference is whether the interview holds its shape.

We wrote a practical guide to running that structure across hundreds of candidates without turning hiring into a bottleneck: how to run structured interviews at scale.

What actually predicts performance

If accent and polish are noise, what is the signal? The content of the answer, scored against behavior you can actually observe. A concrete decision and the constraint that drove it. A tradeoff the candidate named on their own, including what they gave up. The failure case — what went wrong, and what they did about it. Ownership: what they did, distinct from what the team did around them.

These are things two evaluators can read in the same transcript and land in the same place on — which is exactly why they make scores comparable, and exactly why accent does not. Anchor every rating to one of these moments and the vibe has nowhere to hide. We laid out the full discipline — what to cite, and what to deliberately leave out — in evidence-based scorecards.

Designing bias out instead of training it away

The instinct, when you notice bias, is to train it out of people. Sit everyone down, name the biases, ask them to try harder. It is a popular fix because it asks nothing of the process — no new questions, no rubric, no change to how the decision actually gets made.

The trouble is that the signals we have been describing are implicit. They act before deliberate judgment gets a vote, which is precisely why asking people to deliberate harder does so little. You cannot will yourself deaf to an accent. What you can do is change the instrument: ask every candidate the same questions, score them against the same versioned rubric, and require that every rating cite a specific moment in the transcript. Fairness becomes a property of the process, not a favor the interviewer is asked to perform.

And one line that should be non-negotiable: no biometrics. How a candidate looks, their tone of voice, their facial expressions, their “energy” — none of it belongs anywhere near a score. Judge the content of what someone says, never how they look or sound. The moment a system starts reading faces or voices, it has quietly rebuilt the exact bias you were trying to design out.

Where Hure fits

This is the whole idea behind how Hure runs. Every candidate for a role gets the same structured, 8–10 minute voice-to-voice interview in their own language, covering the same versioned rubric the same way. The result is a scorecard where each rating cites the moment in the transcript that earned it — evidence you can audit, not an impression you have to trust.

And Hure’s interviewer cannot do the thing this whole post is about. It never sees a face and never judges an accent; it scores the content of what a candidate said, and nothing else. That is not a feature bolted on the side — it is the point. A human still makes the final call. The job of the structure is only to make sure that when they decide, they are looking at what the candidate actually said, not at how it happened to sound.