Table of Contents

1. Why Most Interview Questions Fail
2. The Three Question Types That Predict Performance
3. Building Your Interview Framework
4. Questions to Stop Asking

"Tell me about yourself." "What's your greatest weakness?" "Where do you see yourself in five years?"

These are the most common interview questions in the world. They're also nearly useless for predicting whether someone will actually perform well in the job.

Research is clear on this: unstructured interviews (where each interviewer asks whatever they feel like) predict job performance about as well as a coin flip. Their predictive validity sits around .20 on a scale where 1.0 is perfect prediction. That means 80% of what you're evaluating in a free-form interview has nothing to do with future job performance.

Structured interviews, where you ask the same questions to every candidate and evaluate responses against predefined criteria, nearly double that predictive validity to .55-.70. The questions you ask and how you evaluate the answers determine whether your interview is a reliable assessment tool or expensive theater.

Here's the framework for asking questions that actually predict performance.

Why Most Interview Questions Fail

Traditional interview questions fail for specific, measurable reasons:

They test interview performance, not job performance. "Tell me about yourself" rewards polished presenters and punishes introverts. Unless the job requires giving impromptu presentations to strangers, you're selecting for an irrelevant skill.

They invite rehearsed answers. Everyone who has Googled "common interview questions" knows to say their greatest weakness is "perfectionism" or "working too hard." You're testing preparation, not capability.

They trigger interviewer bias. Vague questions produce vague answers that interviewers interpret through their own biases. Without a scoring rubric, two interviewers will rate the same answer differently based on their own preferences and similarity to the candidate.

They lack construct validity. They don't actually measure what they claim to measure. "Where do you see yourself in five years?" supposedly assesses ambition and commitment, but actually measures whether the candidate tells you what you want to hear.

The Three Question Types That Predict Performance

Research identifies three question categories with demonstrated predictive validity:

1. Behavioral Questions (Past Behavior)

Based on: The principle that past behavior in similar situations is the best predictor of future behavior.

Format: "Tell me about a time when you [specific situation relevant to the job]. What did you do and what was the outcome?"

Why they work: They require candidates to describe real situations they've actually navigated. You're evaluating demonstrated capability, not hypothetical promises. Candidates can't rehearse answers to specific behavioral questions because they require drawing from actual experience.

Examples by competency:

2. Situational Questions (Future Scenarios)

Based on: Goal-setting theory. How people say they would behave in job-relevant situations correlates with how they actually behave.

Format: "Imagine you're in [specific job-relevant scenario]. What would you do?"

Why they work: They test judgment and thought process in situations the candidate will actually face. Unlike behavioral questions, they work for candidates who may not have direct experience in a specific area (career changers, recent graduates) but can demonstrate good reasoning about how to approach it.

Examples by competency:

3. Job Knowledge Questions (Technical Depth)

Based on: Direct assessment of whether the candidate has the technical knowledge required for the role.

Format: "Walk me through how you would [specific technical task central to the role]" or "What factors would you consider when [technical decision]?"

Why they work: They directly test whether the candidate can do the work. Unlike trivia questions or brainteasers, good job knowledge questions ask candidates to demonstrate their thinking process, not recall memorized facts.

Examples:

• For a marketing manager: "Walk me through how you'd design an A/B test for our landing page, including what metrics you'd track and how you'd determine statistical significance."

• For a software engineer: "How would you design a system to handle 10x our current traffic load? Walk me through your thought process."

• For a sales rep: "You're preparing for a call with a prospect who's evaluating us against two competitors. How do you structure the conversation?"

Building Your Interview Framework

Step 1: Define the Competencies That Matter

Before writing a single question, identify 4-6 competencies that actually predict success in the specific role. Not generic "leadership" or "communication." Specific, observable behaviors.

How to identify them:

• Look at your top performers. What behaviors do they consistently demonstrate?

• Talk to the hiring manager. What separates great performers from average ones in this role?

• Review the actual work. What challenges does this person face daily?

Step 2: Write 2-3 Questions Per Competency

Each competency gets at least two questions: one behavioral and one situational. This gives you multiple data points and accommodates candidates with different experience levels.

Step 3: Create a Scoring Rubric

For each question, define what a 1 (poor), 3 (meets expectations), and 5 (exceptional) response looks like. Be specific:

Example rubric for "Tell me about a time you had to work with incomplete information":

Step 4: Train Your Interviewers

Every interviewer should:

• Ask the same questions in the same order

• Score each answer independently before discussing with other interviewers

• Use the rubric, not their gut feeling

• Take notes on specific responses (not just overall impressions)

• Avoid follow-up questions that lead the candidate to a "right" answer

Step 5: Validate Over Time

After 6-12 months, correlate interview scores with performance outcomes. Which questions actually predicted who performed well? Which ones didn't correlate? Update your question bank based on real data, not assumptions.

Questions to Stop Asking

These common questions have weak or zero predictive validity:

Frequently Asked Questions

How many questions should an interview include?

For a 45-60 minute interview, plan 5-7 questions with follow-up probes. This gives enough time for detailed answers while covering the key competencies. Rushing through 15 questions produces surface-level answers. Depth matters more than breadth. Each answer should take 3-5 minutes, plus 1-2 minutes for follow-up clarification.

Do behavioral questions work for entry-level candidates?

They work, but you need to adjust the scope. Entry-level candidates may not have professional examples, but they have academic projects, volunteer work, part-time jobs, and team activities. Phrase questions broadly: "Tell me about a time you had to meet a tight deadline" works for a recent graduate describing a school project as well as for a senior professional describing a product launch.

Should every interviewer ask the same questions?

Each interviewer should cover different competencies with standardized questions for those competencies. Having three interviewers all ask about teamwork wastes time. Assign competencies by interviewer: Person A covers problem-solving and technical skills, Person B covers collaboration and communication, Person C covers ownership and adaptability. This gives you comprehensive coverage without redundancy.

How do you handle candidates who give vague answers to behavioral questions?

Use the STAR probe: "Can you be more specific about the Situation? What was the Task you were responsible for? What Actions did you personally take? What was the Result?" If a candidate consistently can't provide specific examples after probing, that's a data point: they may not have the experience they claim, or the experience may not be as relevant as their resume suggests.

Key Takeaways

• Unstructured interviews predict performance at .20 (barely better than random). Structured interviews predict at .55-.70. The difference is the questions and how you evaluate them.

• Three question types have demonstrated predictive validity: behavioral (past behavior), situational (hypothetical scenarios), and job knowledge (technical depth).

• Define 4-6 competencies per role, write 2-3 questions per competency, and create scoring rubrics that define what good/average/poor answers look like.

• Every interviewer asks the same questions and scores independently before discussing. This prevents anchoring bias and ensures comparable evaluation.

• Validate your questions over time by correlating interview scores with performance outcomes. Update questions that don't predict.

Turn Better Questions Into Better Hires

The right questions are only half the equation. You also need a system that ensures every candidate gets the same questions, every interviewer scores independently, and all evaluations feed into a single decision framework.

HrPanda's applicant tracking system includes built-in scorecard templates with competency-based evaluation rubrics. Assign questions by interviewer, collect scores independently, and compare candidates on actual evidence rather than gut feelings. Start making data-driven hiring decisions today.