Structured Interview

Two candidates apply for a service role: one is asked how they handled a difficult customer, while the other is asked mostly about hobbies. Their interviews are hard to compare, whereas a structured method gives both a fair chance to show job skills.

The US Office of Personnel Management publishes a structured-interview guide covering job analysis, question design and rating scales, and its federal process is an evidence-based example, not a mandatory script for every business. Analyse the role first, because a consistent question about an irrelevant trait can still be consistently wrong.

Identify the important tasks and capabilities, then write core questions in advance: behavioural questions ask for past examples, while situational questions ask how a person would handle a realistic case, and both need a link to the work. CIPD selection guidance also discusses structured methods, and the evidence supports careful design, but no interview alone guarantees the right hire.

Use the same core set for every candidate, and allow follow-up prompts only to clarify an answer fairly, since the aim is comparable evidence, not a robotic refusal to ask any question. Define rating anchors that describe what weak, adequate and strong evidence looks like, because a five-point scale without examples invites different interpretations.

Set the weights before the interviews, so that if safety judgment matters more than presentation polish the rubric says so and nobody shifts the weights after meeting a favourite. An illustrative rating is the sum of four equally weighted scores from one to five, so a candidate with ratings 4, 3, 4 and 5 totals 16 of 20.

The number is not a prediction guarantee, and a work sample, reference or test can answer questions the interview cannot, so decide how they fit before selection. Train interviewers to ask, probe, note evidence and avoid prohibited topics, because a form alone does not create skill, and keep the panel focused, since different interviewers can assess different job-related areas but too many people may burden candidates and create noise.

Score independently: panellists should record judgments before hearing a senior manager's view, then discuss differences with examples. Record what was said rather than impressions, since a note such as "seemed confident" is weaker than the steps the candidate described, and review bias from shared background, accent or similar interests, which anchored evidence helps challenge but cannot erase automatically.

Handle one poor answer cautiously, because nerves or a confusing prompt can affect a response, and ask a fair clarifying question before drawing a conclusion. Plan accessibility and fairness: adjustments may be needed for disability or language without changing the essential skill tested, and applicable local rules should be followed.

Tell candidates the format and the role, avoid trick questions and overly abstract hypotheticals, and fix unclear wording for everyone rather than treating it as a hidden test of confidence. Protect notes as personal information, audit results over time (if hires consistently struggle in a skill the interview rated highly, revise the question and anchors, bearing in mind that small samples need caution), and do not promise objectivity: make the process transparent, explain decisions in role terms, and remember that for owners a structured interview is a fairer way to ask and compare job-relevant questions.