Structured interviews are usually more accurate predictors of which applicants will make suitable employees than unstructured conversation.
Why it matters when the plan changes
The unstructured interview remains the most used and least predictive selection method in most organisations, kept in place because it feels informative. Structure is the cheapest available improvement to selection quality, and it is resisted because it makes the process feel mechanical to interviewers who believe their judgement is the value they add. Structured interviews are usually more accurate predictors of which applicants will make suitable employees, yet that is the format interviewers trust least.
The tension is that structure and assessment answer different questions and are often treated as substitutes. An interview can probe what someone did and why, in their own account. An instrument measures a disposition without relying on that account. Choosing one and skipping the other leaves a gap that the chosen method cannot see into. Selection is a methodical process, and any method is only as good as the specification it is measured against.
In practice
A hiring panel runs four unstructured conversations and reaches confident, conflicting impressions. The same panel, given a common question set and a rating scale, disagrees on fewer points and can say which evidence produced each rating. Nothing about the candidates changed; the panel's ability to compare them did.
Evidence
Structured interviews ask a predetermined list of questions in a specified order and are usually more accurate predictors than unstructured ones.
Job interview, Wikipedia (2026)Selection is a methodical process, and the methods used are only as good as the specification they are applied against.
Personnel selection, Wikipedia (2026)
What it cannot tell you
The comparison does not tell you which single method to drop. A structured interview still relies on a candidate's own account, verified against criteria rather than independent evidence, and an instrument is silent on how someone will explain or defend a decision under questioning. Neither method compensates for a poorly specified role.
Questions
The same questions in the same order for every candidate, anchored rating scales defined in advance, multiple trained interviewers rating independently, and notes taken against criteria rather than impressions. Removing any one of those recovers most of the unreliability structure was meant to remove.
Because they feel highly informative to the interviewer, even though confidence and accuracy are not the same thing here. The entry on the job interview in Wikipedia (2026) notes that structured formats are usually more accurate, yet the unstructured conversation still feels like insight, and that feeling survives evidence that it predicted nothing.
No. An interview gathers a person's own account of what they did and why, probed in real time. An instrument measures a disposition without depending on that account, which makes it resistant to how well someone presents. They are complements, and each is blind to what the other sees.
The specification. Personnel selection, as described in the Personnel selection entry on Wikipedia (2026), is a methodical process, and both interviews and instruments are only as good as the specification they are measured against. Writing that specification down properly improves both methods more than any change to either one alone.
More than one, each rating independently before any discussion. Independent ratings that are then compared will surface disagreement that a joint conversation would have quietly smoothed over, and that disagreement is usually where the most informative part of the evidence actually sits.