Structured Interviews: A Better Way to Predict Performance

By Don Georgevich

Phase Three of the Scientific Hiring Method: Discovery

A structured interview does not make the conversation robotic. It makes the evidence comparable.

Most interviewers believe they can tell when a conversation is going well. That is exactly where the risk begins.

A candidate who speaks easily, shares familiar interests, and tells a convincing story can feel stronger than someone whose evidence is better but whose delivery is less polished. A structured interview prevents the conversation itself from quietly changing the standard.

What makes an interview structured

OPM defines a structured interview as an assessment method that measures job-related competencies through a standardized questioning and scoring process. Candidates are asked predetermined questions and evaluated using the same standards. See OPM’s Structured Interviews guidance.

In practice, a structured interview has six parts:

1. The competencies and outcomes come from the role analysis and scorecard.
2. The central questions are written before interviews begin.
3. Every candidate receives the same core questions.
4. Interviewers use relevant follow-up probes to complete the evidence.
5. Answers are scored against defined anchors rather than general impressions.
6. Interviewers record independent ratings before the group discussion.

Begin with the outcome you need to examine

Do not begin by collecting popular interview questions. Begin with the Job Scorecard.

Suppose an Operations Manager must create a reliable weekly operating rhythm within the first ninety days. The underlying capabilities may include process diagnosis, prioritization, influence, follow-through, and practical judgment.

The interview should gather evidence about those capabilities in a situation similar enough to the role to matter.

Turn the outcome into a question

Weak question: “Are you good at improving processes?”

The candidate knows the preferred answer, and “yes” provides no evidence.

Better behavioral question: “Tell me about a time you inherited an operating process that people did not trust or follow. What was happening, what did you change, and what result did you produce?”

This asks for past behavior connected to the role. The answer can be examined for context, ownership, decisions, resistance, and results.

Possible situational question: “You join a small company where each manager runs the weekly operating meeting differently, important problems move between teams without clear ownership, and the owner keeps stepping in. How would you diagnose the problem during your first month?”

This tests how the candidate would approach a realistic situation, although proposed behavior should not be treated as proof that the person has done the work before.

Use probes to complete the evidence

Standardization does not mean asking each question once and refusing to clarify. It means using follow-up questions for the same purpose.

Useful probes include:

“What was the condition when you started?”

“What part did you personally own?”

“What alternatives did you consider?”

“Who disagreed with your approach?”

“What did you measure?”

“What was the result?”

“What would you do differently now?”

Candidate-specific probes may differ because answers differ. The competency being tested and the evidence required should remain the same.

Score the evidence, not the performance of being interviewed

Before the first candidate arrives, define what weak, acceptable, and strong evidence would look like.

The anchors should fit the specific competency. A strong score for process diagnosis should not use the same language as a strong score for customer communication or financial judgment.

A worked answer and score

Question: “Tell me about a time you inherited an operating process that people did not trust or follow.”

Candidate answer: The candidate describes a scheduling process used by three service teams. Each supervisor kept a separate tracker, jobs were assigned before equipment was confirmed, and customer delays were rising. The candidate mapped the handoffs, reviewed four weeks of failures, and found that the teams used different definitions of “ready.”

The candidate created one shared readiness definition, tested it with one team, and met resistance from a sales manager who feared slower booking. The candidate compared booking time and delay rates during the pilot, showed that booking time remained stable, and gained agreement to expand the change. Late starts fell from fifteen per month to five during the next quarter.

Independent rating: 4 — Strong.

Reason: The example is relevant and specific. The candidate explains personal ownership, diagnosis, resistance, measurement, and a sustained result. The evidence is stronger than acceptable but does not yet show repeated success across several comparable settings.

Open question: The candidate had analyst support. A later Candidate Test Drive should test whether the person can simplify the approach with fewer resources.

Do not let the group create the score

Each interviewer should record evidence and a provisional rating before hearing the opinions of others.

When the most senior or enthusiastic interviewer speaks first, other ratings often move toward that judgment. Independent notes preserve disagreement long enough for the team to understand it.

After the independent ratings are recorded, interviewers can compare evidence. The goal is not to average away every difference. It is to identify why the ratings differ and whether one person heard evidence others missed.

Assign interview responsibilities

A panel does not need every interviewer to ask every question.

Assign each person a competency, outcome, or question set. One interviewer may examine process diagnosis while another examines influence and a third examines practical judgment. Everyone should understand the full role standard, but individual ownership reduces repetition.

OPM notes that structured interview panels commonly include two to four members. The right number for a small company depends on the role and available expertise, but adding interviewers without clear responsibility usually adds noise rather than evidence. See OPM’s panel-size guidance.

Ask the same central questions in the same order

Using the same central questions improves comparability and helps prevent one candidate from receiving an easier interview than another.

The conversation may still feel natural. Interviewers can acknowledge answers, clarify meaning, and use relevant probes. What should not change is the central evidence the company requires.

Do not give a favored candidate helpful hints, extra chances, or a different version of the standard.

Separate note-taking from interpretation

Write what the candidate said before writing what you think it means.

Observation: “Candidate said the team resisted the new process and that the vice president required adoption after the pilot.”

Interpretation: “May have relied on executive authority rather than personal influence.”

The distinction matters because interpretations can be tested. A later probe may show that the candidate built support first, or it may confirm that the change depended entirely on authority from above.

Use more than one type of evidence

A structured interview is stronger than an unstructured conversation, but it should not carry the entire hiring decision.

OPM’s assessment guidance treats structured interviews as one method that may be combined with other job-related assessments. Work samples can reveal how a candidate handles realistic tasks, while references can test whether claimed patterns were visible to people who observed the work. See OPM’s Assessment and Selection guidance.

Use the interview to identify what should be demonstrated in the Candidate Test Drive and what should be verified during Reference Checking.

Keep questions job-related

Interview questions should focus on the work and be applied consistently. Avoid questions about age, family status, pregnancy, disability, religion, national origin, medical history, or other protected information unrelated to job performance.

The EEOC states that selection procedures can violate federal law when they disproportionately exclude protected groups and are not job-related and consistent with business necessity. See the EEOC’s Employment Tests and Selection Procedures guidance.

This page is educational and is not legal advice.

The interview is finished when the evidence is complete

A good structured interview should leave the team with specific examples, independent ratings, open questions, and a clear reason for the next step.

It should not leave the team arguing about who seemed more confident, who felt like a better cultural fit, or who told the most enjoyable story.

Use Phone Screening and the Phone Screening Questions before this stage. Then move the most important unresolved capability into the Candidate Test Drive and verify major claims through Reference Checking.

Return to Discover What a Candidate Can Really Do for the complete Phase Three path.

See the Scientific Hiring Method.

Related Phase Three resources: Structured Interview Questions That Produce Comparable Evidence, Structured vs. Unstructured Interviews, Interview Scorecard: How to Turn Answers Into Evidence, and Interview Rubric: How to Grade Every Candidate the Same Way.