Structured vs. Unstructured Interviews: What the Research Says

The academic case for structured interviews has been strong and consistent for four decades. Meta-analyses of predictive validity routinely find that structured interviews are significantly better predictors of job performance than unstructured ones — often two to three times as accurate. Despite this, surveys consistently find that the majority of companies still rely primarily on unstructured interviews. The research gap and the practice gap have never fully closed.

What "structured" actually means

A structured interview has three defining features: the same questions are asked of every candidate in the same order, responses are evaluated against a defined scoring guide rather than general impression, and interviewers are trained on how to ask and evaluate questions consistently. Notably, "structured" does not mean robotic or impersonal — it means standardised enough to enable fair comparison.

An unstructured interview, by contrast, follows no fixed format. Questions emerge organically from conversation, evaluation is holistic and impressionistic, and the information gathered from two candidates for the same role may be almost entirely different. The conversation may be engaging and warm — and still be a poor predictor of performance.

Why unstructured interviews persist despite the evidence

Most hiring managers prefer unstructured interviews because they feel more natural and produce a richer sense of the "fit" between a candidate and a role. This preference is itself a bias — the feeling of connection in an unstructured interview reflects social comfort, not job-relevant capability. The people who interview well in unstructured conversations are often people who are confident, articulate, and socially skilled — which correlates imperfectly with most of the things that actually predict performance in specific roles.

There is also an institutional inertia problem. Changing to structured interviews requires upfront investment: defining questions, building scoring guides, training interviewers. For organisations without a centralised recruiting function, that investment feels disproportionate to the perceived problem.

The bias problem in unstructured interviews

The validity gap between structured and unstructured interviews is driven partly by random variation and partly by systematic bias. Unstructured interviews amplify affinity bias (favoring candidates similar to the interviewer), halo and horn effects (first impressions dominating subsequent evaluation), and attribution errors (crediting candidates for team outcomes or blaming them for organisational failures).

For organisations with diversity hiring goals, this is a material problem. Structured interviews do not eliminate bias, but they significantly reduce the channels through which it operates by standardising what information is collected and how it is weighted. See our guide on reducing unconscious bias in recruiting.

How to transition without losing interview quality

The most common objection to structured interviews is that they make conversations feel like tests and reduce the ability to assess cultural fit. Neither needs to be true. Well- designed behavioural questions ("Tell me about a time when...") elicit rich narrative responses that reveal how candidates actually think and behave. A scoring guide can be designed to evaluate both technical and interpersonal dimensions.

The practical transition: start with a pilot on one role or function, define four to six core competencies the role requires, write two questions per competency, build a simple scoring guide, and train interviewers on the basics of behavioural interviewing. AI coaching tools can help interviewers apply the structure consistently during live interviews. Learn how Koachly supports structured recruiting interviews.