Competency-Based Interviewing: Framework and Examples

Most guides on competency based interviewing start with a list of clever questions to ask. This one starts at the other end, on purpose. It is written for the people who design the loop, sit on the panel, and have to defend the hire later, not for the candidate rehearsing answers.

Here is the frame I will defend the whole way through, the scorecard-first rule: a competency is only real if you can score it. If you cannot write down what a poor, borderline, solid, and outstanding answer looks like for "handles ambiguity," then "handles ambiguity" is a vibe, not a competency, and your questions about it will collect impressions instead of evidence. So you build the framework backward. The scorecard comes first, the questions get derived from it, and the score is the product the whole process exists to produce.

I lead hiring across several teams and interview on top of everything else, so I have zero patience for process that adds time without adding signal. Competency based interviewing earns its keep only when the scorecard does the heavy lifting. Score the signal, not the vibe, and the rest of the loop falls into place.

What a Competency-Based Interview Actually Is

A competency based interview asks every candidate for a role the same predefined questions, each one tied to a named competency, and scores every answer against an anchored rubric set before sourcing began. That definition has three load-bearing parts, and the scorecard-first rule says the rubric is the one people drop. Skip it and you have a tidy question bank, not a competency based interview, because nothing forces two interviewers to agree on what a strong answer contains.

The rubric is also what separates this from a plain behavioral chat. "Tell me about a time you shipped despite an unclear spec" is only a competency question if "shipping under ambiguity" sits on your scorecard with described levels. Without those levels, the answer becomes a nice story you grade on gut, and gut is exactly the input the scorecard-first rule is built to remove. The anchors are what convert a story into a number you can compare across candidates.

That comparison is the entire reason to do any of this. When the rubric is anchored, the candidate who got the warm interviewer and the candidate who got the skeptical one are still measured against the same described bar. The same bar for every candidate is the point, and the scorecard is the thing that holds it in place when the conversation drifts, which it always does. Hold that idea, because the validity research explains why the scorecard, not the question list, is where the predictive power lives.

Why the Scorecard Is Where the Validity Lives

The strongest evidence in hiring points at scoring discipline, not at clever phrasing. In 2022 Philip Sackett, Charlene Zhang, Christopher Berry, and Filip Lievens revisited decades of personnel-selection data and found structured interviews carried the highest mean operational validity at .42, ahead of cognitive ability tests (SIOP, 2022). That .42 is the headline, and the scorecard-first rule is the reason it lands where it does: the structure being measured is the rubric, not the wording.

The .42 figure comes with a caveat worth keeping honest about. Sackett and colleagues report an 80% credibility interval of .18 to .66, so the real reading is .42 plus or minus .24 (SIOP, 2022). That spread of .18 to .66 is not noise to ignore, it is the gap between a structured interview run with a real anchored scorecard and one run with a scorecard in name only. The teams sitting near .66 are the ones who built the rubric first.

How far below that ceiling you land tracks how loose the scoring is, and the contrast with unstructured interviews makes the point. Sackett et al. put unstructured interviews near .20, less than half the structured figure (SIOP, 2022). The earlier Schmidt and Hunter 1998 meta-analysis drew the same shape, .51 for structured versus .38 for unstructured (Schmidt & Oh, 2016). The number that changes between those two columns is the scoring method, which is why the rubric is the focal point.

The focal-point language is not mine, it is the field's. As Sackett and colleagues put it, "one might propose structured interviews as the focal predictor against which others are evaluated" (SIOP, 2022). Treat the scorecard as that focal predictor and the score becomes the thing you defend, which is also the cheapest way to make the loop faster, the topic the next section builds out.

How to Build the Framework Backward From the Scorecard

Building a competency based interview is a design task, not a conversation skill, and the scorecard-first rule fixes the order of operations. Here is the build, scorecard first, then questions, then everything else.

Name 3 to 5 competencies from the first 90 days, not a generic list. Pull them from what the person will actually do, then write a one-line behavioral definition for each. For a support lead that might be "de-escalates angry customers," "writes reusable macros," and "triages by impact." These become your scorecard columns.
Anchor each competency on a 1 to 5 scale before writing a single question. For every competency, describe what a 1 (no evidence), a 3 (solid, specific example), and a 5 (mastery with measurable impact and self-reflection) answer contains. This is the step the scorecard-first rule exists to protect, and the one teams skip when rushed.
Derive one or two questions per competency from those anchors. Write each question to pull the evidence your anchors ask for. Past-behavior form works best: "Describe a specific time you..." rather than "How would you...". The question is the delivery mechanism, the anchor is the target. For prompts already mapped to competencies with scoring notes, 50 Structured Interview Questions by Competency is a bank you can derive from.
Split competencies across the panel so coverage does not overlap. Give each interviewer a competency to own and probe deeply. Two people grading the same competency wastes a slot and leaves another competency unmeasured.
Score independently, before the debrief. Every interviewer commits a number on the rubric before hearing anyone else's read. This single rule moves the score more than any question you could add.
Debrief on evidence against the anchor, not impressions. Resolve disagreements by pointing at what the candidate said versus the described level, not by who feels more strongly. Hire on what they said, not how you felt.

That order is deliberately backward from how most loops get built, and step 5 is the one that quietly decides whether the scorecard means anything, which is the failure worth naming next.

The Mistake That Quietly Wrecks the Scorecard

The most common failure in competency based interviewing is not a weak question, it is scoring after the debrief instead of before it. When interviewers walk in without committed numbers, the loudest or most senior voice anchors the group, and the scorecard turns into a form people backfill to match the room. The scorecard-first rule has a procedural twin here: lock numbers first, talk second, and the fix costs nothing.

This is where I have a first-party read worth sharing honestly as a qualitative observation, not a clean stat. Across the structured loops the Asked team has watched scored from live transcripts, the pattern is consistent: panels that commit a rubric number before the debrief reach a clear hire-or-pass call noticeably faster and reverse themselves far less often than panels that talk first and score after. Same candidates, same questions. The only thing that moved was when the number got written down, which is exactly what the independent-scoring research predicts.

There is a quieter version of the same trap, treating the rubric as a form to fill rather than a definition to agree on. If two interviewers read "solid" differently, the anchors are not real yet. Calibrating on one or two recorded answers before the role opens fixes that faster than any amount of generic interviewer training, and it sets up the comparison most teams are actually trying to make between formats.

Competency-Based vs Behavioral vs Unstructured

The three interview styles people lump together score very differently, and the gap is all in the scorecard, not the questions. Here is how they compare on the dimensions that change a hiring outcome.

Dimension	Unstructured	Behavioral (loose)	Competency-Based
Questions	Improvised per candidate	Same "tell me about a time" prompts	Each mapped to a named competency
Scoring	Overall impression at end	Notes, graded on gut	Anchored 1 to 5 rubric, committed first
Predictive validity	Lower (~.20)	Between the two	Higher (~.42)
Defensibility	None	Thin notes	Score tied to a described anchor
Best for	Almost nothing	Quick screens	Most roles, most of the time

The pattern in that table is the scorecard-first rule restated: the column that improves as you move right is scoring, while the questions barely change. The middle column is where most teams actually sit, running behavioral interview questions without mapping each one to a defined competency or scoring against anchors. Competency based interviewing is a structured interview with the competencies made explicit, so it sits inside the broader practice covered in Structured Interviews: The Complete Guide for Hiring Teams. That guide is where the format choice gets decided, and the anchors below are where you make any chosen format actually score.

A Worked Example: One Competency, Anchored

Take "handles ambiguity" for a product manager, the competency most teams claim to test and least often score. The scorecard-first rule says start with the anchors, so here is a 1 to 5 scale before any question gets written.

1, no evidence: Describes a fully specified project, or cannot name a time the spec was unclear.
3, solid: Names a specific ambiguous situation, the call they made, and why, with a real outcome.
5, mastery: Adds measurable impact, names the tradeoff they accepted, and reflects on what they would change, mapping cleanly to STAR (Situation, Task, Action, Result) (Predictive Index).

Only now does the question get written, derived straight from those anchors: "Describe a specific time you had to ship with an incomplete spec. What did you decide, and how did it turn out?" The question exists to surface the evidence the anchors grade, which is the whole inversion. Run the same anchored question against two candidates and the scores are comparable, because both were measured against the described levels, not against each other on the day.

Build-Day Checklist

Before a competency based interview goes live, run this list. Each item protects the scorecard.

Each competency has a one-line behavioral definition pulled from the first 90 days.
Every competency has anchors written for at least the 1, 3, and 5 levels.
Each question maps to exactly one competency and uses past-behavior phrasing.
Competencies are split across the panel with no overlap.
Independent scoring is required before the debrief, in writing.
The team calibrated on one real recorded answer before the role opened.

FAQ

How many competencies should a competency-based interview test?

Three to five for most roles. Fewer than three and you are not measuring enough of the job, more than five and each one gets a shallow read because no single interviewer has time to probe deeply. Split them across the panel so each interviewer owns one or two and scores them properly, which keeps every competency on the scorecard from going unmeasured.

What scoring scale works best for competency-based interviews?

A 1 to 5 anchored scale, where 1 means no evidence of the competency and 5 means clear mastery with specific, measurable examples and self-reflection. The number itself matters less than the written anchors behind each level. An unanchored 1 to 5 scale is just a feeling with a number taped to it, which the scorecard-first rule exists to prevent.

Are competency-based and behavioral interviews the same thing?

Close, but not identical. Behavioral interviews ask about past situations, which is good, but many run without mapping each question to a defined competency or scoring against anchors. A competency based interview adds both, which is what moves it from the loose-behavioral column near gut grading toward the .42 validity that structured scoring earns (SIOP, 2022).

Does this actually save time, or just add process?

It saves time once the scorecard exists. Google reports structured interviewing saves an average of 40 minutes per interview when teams reuse pre-made questions, guides, and rubrics, on a framework of vetted questions, recorded answers, standardized rubrics, and calibration (Google re:Work). The build is front-loaded, the savings repeat on every candidate after.

Do This Next

Pick one role you are hiring for this week and write down the three competencies that decide success in its first 90 days. Build a 1 to 5 anchored scorecard for each one before you touch a single question, then derive the questions from the anchors. Score your next two candidates against that scorecard independently, before anyone debriefs, and watch how much faster the call gets. Start today: try Asked free and let it join the call, transcribe live, and draft the self-scoring scorecard straight from the transcript so the score is written down before the room can talk you out of it.