Structured vs Unstructured Interviews, According to the Research

Structured vs unstructured interviews: which one actually predicts who will do well on the job? For decades the honest answer was "structured, but not by as much as you'd think." A 2022 correction to the underlying research changed that. Once a long-standing statistical error was fixed, structured interviews came out on top of the entire list of hiring methods, ahead of cognitive ability tests, the format psychology had treated as the gold standard since the 1980s.

That is not a small revision. It reorders the methods employers have relied on to decide who gets hired, and it explains why the interview format matters far more than most hiring processes treat it.

What actually makes an interview structured

A structured interview asks every candidate the same questions, in the same order, and scores each answer against a fixed rubric written before anyone walks in the room. An unstructured interview lets the conversation wander: different questions per candidate, no shared scoring guide, a decision that often comes down to a feeling about the person by the end.

The rubric is the part that matters. Scoring an answer against described levels of a good, adequate, or poor response is the same logic behind a behaviourally anchored rating scale: a described level of observable behaviour instead of a vague number in an interviewer's head. It replaces "I liked them" with a graded response to a defined question, which is also why it holds up better than an unscored gut read of a candidate.

Structured vs unstructured interviews: the 2022 correction

For years, meta-analyses ranked general cognitive ability tests as the single best predictor of job performance, with structured interviews running a respectable second. That ranking rested on a statistical correction called restriction of range, applied the same way across old and new studies. In 2022, Paul Sackett, Charlene Zhang, Christopher Berry, and Filip Lievens published "Revisiting Meta-Analytic Estimates of Validity in Personnel Selection" in the Journal of Applied Psychology (107, 2040 to 2068), arguing that correction had been systematically overapplied for decades, inflating some methods' scores more than others.

Once the correction was redone, the ranking changed. Per SIOP's summary of the paper, the corrected validity estimates put structured interviews first at r=.42, followed by job knowledge tests at .40, empirically keyed biodata at .38, work sample tests at .33, and cognitive ability tests at .31.

A method that gets treated as a soft, subjective step in hiring turned out, once the arithmetic was fixed, to outrank the test everyone assumed was the objective one.

Why structure closes the gap

Structure works because it removes the parts of a normal conversation that predict nothing about job performance: how much small talk two people share, whether the interviewer's first impression forms in the opening minute, whether a candidate happens to remind someone of a good hire from years back. None of that is scored in a structured format, because the rubric only rewards what the candidate actually said in response to a fixed question.

That is the same argument the research on self-assessment accuracy makes about self-ratings: an unscored impression is a weak instrument, however sincere the person forming it. A situational judgement test applies the identical fix to a written scenario instead of a live conversation, scoring a candidate's actual choice against what practice experts agree is the better response.

What this means if you are on either side of the table

For a hiring team, the practical takeaway is specific: write the questions and the scoring rubric before the first candidate walks in, and use both for everyone in the role. That single change is most of what separates a structured interview from an unstructured one.

For a candidate, it means the fairest interviews are the ones that feel the most repetitive from your side. The same questions asked of everyone are not a lack of interest in you specifically. They are the format that the corrected research now ranks above every other common hiring method, including the tests that used to outrank it. The check is built on the same principle: six non-technical skill domains, scored against a rubric, not against how well a conversation happened to go.