Interview Rubric Guide
A practical step-by-step guide to building BARS-anchored interview rubrics for any role.
What is a BARS Interview Rubric?
A Behaviorally Anchored Rating Scale (BARS) interview rubric is a scoring framework that attaches specific, observable behavior examples to each point on a numerical rating scale. Unlike a generic "rate this candidate 1-5," BARS rubrics define what a 5 looks like, what a 3 looks like, and what a 1 looks like — in concrete, behavioral terms.
The effect on interview consistency is significant. Two interviewers who've been calibrated against the same BARS rubric will converge on similar scores for the same candidate response — not because they're agreeing in the room, but because the rubric did the calibration work in advance. Meta-analyses on structured interview validity consistently show this improvement over unstructured methods.
Step 1: Identify the Scoring Dimensions
Start by listing 4–6 dimensions that genuinely predict success in the role. For a Software Engineer: Technical Depth, Problem Framing, Communication, Collaboration, and Role Fit are common starting dimensions. Keep the list short enough to score rigorously — 8 dimensions per interviewer is unmanageable.
For each dimension, write a single question that best elicits behavioral evidence of that dimension. Use the STAR framework as a starting point: Situation, Task, Action, Result. The question should require the candidate to describe something they actually did, not what they would do.
Step 2: Write the Behavioral Anchors
For each dimension, write anchors for score levels 1, 3, and 5. Level 2 and 4 are implied midpoints. The anchors should describe observable candidate behaviors — what they say, what they demonstrate, what evidence they provide — not your inference about their capabilities.
Example for Technical Depth (Software Engineer role):
- Score 5: Candidate articulates the constraints they were working under, documents specific trade-offs with measurable justification, and cites a measured outcome. Can point to a specific decision and explain why it was made.
- Score 3: Candidate identifies that trade-offs existed but doesn't quantify them or provide evidence for the final choice. Correct direction, incomplete depth.
- Score 1: Candidate describes what they built or what happened, without providing reasoning about design decisions or demonstrating awareness of trade-off space.
Step 3: Calibration Before You Interview
Before any interviews begin, have all interviewers independently score two or three example responses — either from past interview transcripts or from synthetic examples you write. Compare scores. If two interviewers are more than 1 point apart on the same example response, discuss. Usually the divergence comes from different interpretations of what "evidence" means in the rubric. Fix the anchor language until you converge.
Step 4: Independent Scoring During the Loop
Each interviewer submits scores independently — without discussing with other interviewers — before the debrief session. This is the most important rule in structured interviewing. Social anchoring and seniority bias contaminate scores the moment anyone hears another person's assessment. Lock the scores first.
Step 5: Debrief with the Calibration View
The debrief should open with all scores visible simultaneously. Start with dimensions where there's disagreement, not consensus. Where you see an outlier — one interviewer significantly higher or lower than the others — ask them to read their evidence note aloud. Often, the outlier heard something the others didn't. Sometimes the outlier was anchored to a different interpretation of the rubric. Either way, the discussion is productive because it's grounded in specific evidence, not general impressions.
The final hiring decision is logged in the debrief record with a brief rationale. This is your EEOC documentation.