Interview questions that surface real evidence: 7 to ask, and how to score them
The best interview questions do one thing well: they get candidates to show how they actually work. So the follow-up and the scoring matter more than the wording.
Search for interview questions, and you'll find endless lists of clever prompts. Before you spend hours trudging through them, know that most aren’t tied to the job. And a memorable answer is not the same as useful evidence.
A better interview starts with the role:
- Decide what good performance looks like
- Ask questions that bring out relevant examples
- Judge every candidate against the same standard
The seven questions below are practical starting points rather than a universal ranking. Adapt them to the role, use natural follow-ups, and agree on the scoring before the first interview.
For role-specific examples, browse the Metaview interview questions library.
Why most question lists fall short
Most lists focus on questions that sound interesting on paper (and when asked). But even an interesting conversation may tell you very little about whether someone can do the job.
Structured interviews take the opposite approach. The team defines the skills the role needs, asks comparable job-related questions, and scores the evidence against a shared rubric. Research and assessment guidance consistently find that this produces more reliable decisions than an informal chat.² ³
That doesn't mean interviews need to feel scripted. Feel free to ask follow-ups. Clarify vague answers or let the conversation flow naturally. Just make sure each candidate is being assessed on the same things.
Seven questions that surface useful evidence
Treat the seven questions below as a menu. Each one targets a different competency, and no single interview needs all of them.
Pick the questions that match the role. Three or four well-chosen questions with good follow-ups will usually tell you more than rushing through the full list.
1. Walk me through a project you owned from start to finish. What were you personally responsible for?
Why ask it: Use this to assess ownership, delivery, and decision-making.
Follow up with:
- What was the goal?
- Which decisions were yours?
- What was the hardest trade-off?
- What happened in the end?
Listen for: A clear account of what the candidate did, why they did it, and what changed as a result. Strong answers separate “I” from “we” without pretending the work happened alone.
2. Tell me about a decision you made without having all the information you wanted.
Why ask it: This shows how someone deals with uncertainty and risk.
Follow up with:
- What was missing?
- What did you do to reduce the uncertainty?
- Which trade-off did you accept?
- What did you learn?
Listen for: A sensible process, clear assumptions, and an honest view of the risk. Do not give extra credit for a lucky outcome.
3. Describe a time your work fell short. What did you change afterward?
Why ask it: Look for ownership and learning. A dramatic confession is not the point.
Follow up with:
- What was your part in the outcome?
- When did you realize there was a problem?
- What changed the next time?
Listen for: Specific ownership and a concrete change. Weak answers spend most of the time blaming the circumstances or other people.
4. How would you approach this realistic problem from the role?
Why ask it: A good situational question shows how the candidate applies their knowledge to your context.
Follow up with:
- What would you want to know first?
- Which risks would you check?
- Who would you involve?
- What would you do first?
Listen for: Useful clarifying questions, clear priorities and reasoning that fits the constraints. Avoid rewarding someone simply for already knowing your internal systems.
5. Show me how you would perform one of the role’s core tasks.
Why ask it: A short work sample gives you direct evidence of how someone approaches the job.
Follow up with:
- Talk me through your approach.
- What assumptions are you making?
- How would you handle an edge case?
- What would you check before finishing?
Listen for: A workable process, sound judgment, and attention to the details that matter. Give every candidate the same task and score it against the same criteria.
6. Explain a complex idea you know well to someone who is new to it.
Why ask it: Use this when teaching, explaining or influencing is part of the role.
Follow up with:
- How would you explain it to a customer, an executive, or a new teammate?
- What would you leave out?
- How would you check they understood?
Listen for: A clear explanation pitched to the listener’s level. Complexity is not the same thing as expertise.
7. Tell me about a time new evidence made you change your approach.
Why ask it: This reveals adaptability without asking the candidate to describe themselves as “open-minded.”
Follow up with:
- What did you believe at first?
- What changed your mind?
- How quickly did you respond?
- What happened next?
Listen for: A real change in reasoning and a clear explanation of how the new evidence affected the decision or work.
A simple scoring guide
The question is only useful if interviewers agree on what the scores mean. Write the anchors before the interviews start, then tailor them to the competency you are testing.
| Score | What it means | What the answer sounds like |
|---|---|---|
| 1 - Weak | Little useful evidence | Vague, hypothetical or unrelated. The candidate cannot explain their own actions. |
| 2 - Partial | Some evidence, but important gaps | A relevant example with limited detail, unclear reasoning or no clear result. |
| 3 - Strong | Meets the bar | Specific actions, sound reasoning and a result that connects to the competency. |
| 4 - Exceptional | Clear evidence above the bar | Strong evidence, thoughtful trade-offs and insight that goes beyond the basic requirement. |
Ask interviewers to score independently before the debrief. If the scores are far apart, go back to the evidence and the rubric instead of averaging the disagreement away.
Why documentation matters
Good questions usually break down in one of two places:
The scoring is inconsistent
Different interviewers apply different standards to the same answer. Fix this with clearer anchors and occasional calibration.
The evidence is missing
Notes arrive late, the scorecard stays empty, or the debrief depends on memory.
In Metaview’s research across 5.2 million candidate interviews, 31.2% had a scorecard attached. A separate measure, covering 296,555 advances recorded in connected ATS data, found 41.9% had no submitted scorecard on file when the advance was recorded.¹
Feedback may well have existed. What the data shows is that the reasoning was not visible in a submitted scorecard at the point measured. A clear interview scorecard gives the evidence somewhere to go and makes the decision easier to review later.
How Metaview helps
Metaview records interviews run by customer hiring teams and organizes the notes around the questions and competencies in the team’s template. It can also draft the scorecard and link the evidence back to the conversation.
The scoring stays with the panel; what changes is how much of the answer they still have in front of them. The interviewer reviews the draft, corrects anything that is wrong, and applies the rubric.
It's unimaginable to go back to a time before Metaview where we'd have to take notes during interviews and rely on guesswork for improving our processes.”
How to put the questions to work
Start with one role and build from there. Agree on what good looks like, give each interviewer a clear focus, and score the evidence while it’s still fresh. Once the process works well for one role, you can apply it to others.
- Define four to six competencies: Use the intake call to agree on what each one means in practice.
- Choose one or two questions for each competency: Keep the core questions consistent and plan a few useful follow-ups.
- Give each interviewer a clear focus: One or two competencies per person is usually enough.
- Score while the evidence is fresh: Ask interviewers to submit their scores independently during the interview or immediately after.
- Review what is working: Use talent calibration to tighten unclear scoring, and Metaview Reports to check whether questions are producing useful evidence across the team.
- Go deeper when a role needs it: Pair the question set with competency-based interviewing.
We recommend you don't remove a question just because most candidates score well. First check whether the bar is too loose, the applicant pool is unusually strong, or the competency is simply a baseline requirement.
Keep the evidence and the scoring in one place.
Live notes, structured scorecards, and ATS sync across your hiring process.
Frequently asked questions
What interview questions predict job performance?
No specific wording has been shown to predict performance on its own. What the published research on structured interviewing supports is the method: job-related questions, asked consistently, with the evidence scored against anchors agreed in advance. A short work sample adds direct evidence of how someone approaches the work.
Do structured interviews work better than unstructured interviews?
Generally, yes. Research and assessment guidance find that structured interviews tend to produce more reliable and valid assessments. Structure means comparable job-related questions and a clear scoring method. It does not mean a robotic script.
Should every candidate be asked exactly the same questions?
Use the same core questions for candidates applying to the same role and stage. Interviewers can still clarify answers and ask natural follow-ups, as long as everyone gets a fair chance to show the same competencies.
How many questions should an interviewer ask?
For a 30- to 45-minute interview, three or four core questions usually leave enough time for follow-ups. Across the full process, cover the role's most important competencies without making candidates repeat the same evidence several times.
How should interview answers be scored?
Define what weak, partial, strong and exceptional evidence looks like before the interview. Score the answer against those anchors, not against the previous candidate or the interviewer's general impression.
Does Metaview score or reject candidates?
No. Metaview captures and organizes evidence from the conversation and can draft the scorecard. Recruiters and hiring managers review the draft, apply the rubric and make every decision.
Sources and methodology
¹ Metaview platform research, 2026: a corpus of 5,519,828 captured conversations, including 5,207,185 candidate interviews. The 31.2% figure is the share of those interviews with at least one scorecard attached. The 41.9% figure is a separate measure covering 296,555 advances recorded in connected ATS data, so the two have different denominators. This research covers what happens inside the hiring process. It does not track how people do once they are hired, so nothing here is an outcome study.
² U.S. Office of Personnel Management, “Structured Interviews”: official guidance.
³ Sackett, Zhang, Berry and Lievens, “Revisiting meta-analytic estimates of validity in personnel selection” (2022): research record | SIOP summary.
