Interview Scorecard Template (Free), With Rubric Examples
A free interview scorecard template with 1-3-5 rubric anchors, worked examples for three roles, STAR scoring, and how to calibrate your interviewers.
An interview scorecard is a one-page form that lists the three to six competencies a role needs, describes what a weak, solid and strong answer sounds like for each, and gives the interviewer room for a score and the candidate's own words. It turns "I liked her" into a rating someone else can check.
The free template at the top of this page is ready to use, as a spreadsheet or a PDF. It has a first-round screen that suits most roles, plus the three worked examples below: a customer service representative, a registered nurse and a software engineer. Keep the competencies that fit your role and rewrite the rest.
What is an interview scorecard, and how is it different from a rubric?
An interview scorecard is the form every interviewer for a role fills in, the same way, for every candidate. Each competency gets its own score on a fixed scale, each score gets evidence, and the overall recommendation comes last.
An interview rubric is the part of the scorecard that says what each score means. People use the terms interchangeably, but the difference matters: a scorecard without a rubric is a row of numbers that mean whatever each interviewer felt that day. The rubric is what makes one interviewer's 4 mean the same as another's.
Scorecards are the evaluation half of a structured interview. The questions are the other half, and the two are built together: each question exists to collect evidence for one competency on the card.
What goes on an interview scorecard?
| Column | What goes in it | Example |
|---|---|---|
| Competency | One skill or behavior the job needs | De-escalation |
| What good looks like | One sentence describing success in this job | Calms an upset customer and still solves the problem |
| 1 – Weak | What a weak answer sounds like | No real example, or the story ends with a transfer |
| 3 – Solid | What an acceptable answer sounds like | Stayed calm, acknowledged the problem, fixed it within policy |
| 5 – Strong | What an excellent answer sounds like | Specific words, a fair fix, the outcome, and what changed afterward |
| Score (1–5) | The interviewer's rating | 4 |
| Evidence (their words) | What the candidate said that earns the score | "I told her I'd stay on the line until the refund showed up." |
Below the competencies, leave room for the overall recommendation, the interviewer's name and the date.
How do you build an interview scorecard?
1. Pick three to six competencies from the job
Start from what the person will do in their first six months, not from a generic list of virtues. The U.S. Office of Personnel Management's training on structured interviews says they typically assess four to six competencies, unless the job is unusual or very senior. A 15- or 20-minute first-round screen can't cover six well; three or four is enough there.
A competency has to be something you can hear evidence of in an answer. "De-escalating upset customers" works. "Passionate" doesn't. If you're starting from a job description, here's how to turn one into competencies and questions.
2. Write one line on what good looks like
One sentence, in the language of this job. "Notices when a patient's condition changes and escalates it promptly" says far more than "clinical skills". If the hiring manager and the recruiter can't agree on the sentence, they won't agree on the candidates either.
3. Write anchors for 1, 3 and 5
Anchors describe answers, not people. Write what a weak, a solid and a strong answer actually sounds like, in terms an interviewer can match against what they heard. A 2 or a 4 then means "between those two", which keeps the writing manageable without losing the five-point range. OPM's training is in the same spirit: a rating scale typically has three to seven levels, with at least three of them labeled.
Two rules make anchors work. A 3 should be an answer you'd be happy to hire on, so the scale isn't tilted toward perfection. And a 5 should be reachable by real candidates, not reserved for an ideal nobody meets.
4. Add an evidence column
The evidence column holds what the candidate said, as close to their words as you can get. "Good communicator" is a conclusion. "Explained the outage to the client in two sentences and gave a time for the fix" is evidence. OPM's training says the same about interview notes: summarize the actual response, and leave out evaluative remarks and anything about the candidate's demographics.
Evidence is also what lets someone else check your score. A 4 with a quote can be questioned. A 4 with nothing can't.
5. Set the weights before the first interview
Most roles have one or two competencies that matter more than the rest. Give those a weight of 2 and the others a weight of 1, and decide before you meet any candidates, so the weights can't drift toward a favorite.
The overall score is a weighted average: multiply each score by its weight, add the results, and divide by the total of the weights.
| Competency | Weight | Score | Score × weight |
|---|---|---|---|
| De-escalation | 2 | 4 | 8 |
| Resolving the issue | 2 | 3 | 6 |
| Clear communication | 1 | 5 | 5 |
| Accuracy | 1 | 2 | 2 |
| Total | 6 | 21 |
Weighted score: 21 ÷ 6 = 3.5.
6. Keep the overall recommendation separate
The recommendation (hire, don't hire, or hire with a stated reservation) is a judgment, and it should be written after the scores, not instead of them. Usually it follows from them. When it doesn't, write down why. A candidate who averages 3.5 can still be a clear no if they scored 1 on the one competency the job can't do without. Writing the reason down keeps the recommendation from quietly overruling the evidence.
Interview scorecard examples
Three examples follow, all on a 1–5 scale with anchors at 1, 3 and 5. The free template has all three, plus a first-round screen for any role that rates communication, relevant experience, problem solving, motivation for the role and ownership.
Customer service representative
| Competency | 1 – Weak | 3 – Solid | 5 – Strong |
|---|---|---|---|
| De-escalation. Calms an upset customer and still solves the problem. | No real example, or the story ends with a transfer, a raised voice or blaming the customer. | A real example: stayed calm, acknowledged the problem and fixed it within policy. | A specific example with what the customer said and what they said back, a fair fix within policy, the outcome, and what they changed afterward. |
| Resolving the issue. Finds the real problem behind a complaint and resolves it on the first contact where possible. | Follows the script without finding the cause, or escalates by default. | Asks questions to find the issue and resolves common problems without help. | Diagnoses an unusual problem from the account or order history, resolves it or hands it off with full notes, and follows up. |
| Clear communication. Explains policies and next steps in plain words, by phone, chat or email. | Leans on jargon or policy language; the customer would still be unsure what happens next. | Explains clearly; written examples are accurate but generic. | Adjusts tone and detail to the customer, confirms they understood, and puts the next steps in writing. |
| Accuracy. Gets account details, orders and notes right under time pressure. | Can't describe how they avoid errors, or treats mistakes as normal at high volume. | Has a routine for checking details, such as reading back an order or confirming an address. | Gives an example of catching an error others missed and explains the habit that caught it. |
Registered nurse
| Competency | 1 – Weak | 3 – Solid | 5 – Strong |
|---|---|---|---|
| Clinical judgment. Notices when a patient's condition changes and escalates it promptly. | Describes tasks rather than judgment; no example of noticing a change or escalating. | A real example of noticing a change in a patient and escalating to the charge nurse or provider. | A specific example: what they noticed and why it mattered, how they escalated (for example with SBAR), what they did while waiting, and the outcome. |
| Medication safety. Follows the checks that prevent medication errors every time, including on busy shifts. | Can't describe their checks, or describes skipping them under time pressure as normal. | Describes standard checks (patient identity, order, dose, allergies) and uses them consistently. | Stopped an error before it reached the patient, and explains who they told, how it was reported and what changed. |
| Communication and handoff. Gives clear handoffs and explains care to patients and families in plain language. | Describes handoff as reading out a list; no example of a patient or family conversation. | Uses a consistent handoff structure and explains care plainly to patients. | An example where a careful handoff or family conversation prevented a problem, including how they checked understanding. |
| Prioritization under workload. Organizes care across several patients so the most urgent needs come first. | Says they just work faster; no method for deciding what comes first. | Prioritizes by acuity and time-critical tasks, and asks for help when needed. | A specific shift: how they triaged, what they delegated and to whom, how they kept the charge nurse informed, and how it ended. |
SBAR (situation, background, assessment, recommendation) is a handoff and escalation structure many hospitals teach. If your unit uses a different one, put its name in the anchor.
Software engineer
| Competency | 1 – Weak | 3 – Solid | 5 – Strong |
|---|---|---|---|
| Debugging. Finds the root cause of hard problems methodically. | A generic process, such as adding logging until it works; no real bug described. | A real bug: how they reproduced it, narrowed it down and fixed it. | A hard bug traced by forming and testing hypotheses, a root-cause fix, a test or alert so it can't quietly return, and what they shared with the team. |
| Design trade-offs. Makes sensible design choices and can explain what they gave up. | Can't name a trade-off; answers with 'best practice'. | Describes a design choice and at least one alternative they considered. | Names the constraints, compares alternatives by their costs, explains the choice, and says what they would change with hindsight. |
| Collaboration and code review. Gives and takes review well, and settles technical disagreements with evidence. | Describes review as a gate or a fight; no example of changing their mind. | Gives specific, respectful review comments and can describe a disagreement they resolved. | Settled a disagreement with data or a prototype, kept the working relationship, and improved a team practice as a result. |
| Ownership of delivery. Ships working software and stays responsible for it after release. | Describes work as tickets closed; no sense of what happened after release. | Delivered a feature end to end, flagged risks early and fixed problems after launch. | Owned an outcome beyond their tickets, such as monitoring, a fix nobody owned or an incident follow-up, and can show the effect. |
How do you score an interview answer with STAR?
STAR stands for situation, task, action and result. It's the usual shape of a good behavioral answer, and OPM's training uses it to write and probe interview questions. For scoring, each part tells you something different:
| Part | What to listen for | Weak sign |
|---|---|---|
| Situation | A real, specific occasion | "Usually, I would…" |
| Task | What the candidate was responsible for | No clear role of their own |
| Action | What they personally did, step by step | "We" throughout, or a list of duties |
| Result | What happened, ideally with a number or a consequence | No outcome, or "it went well" |
Most of the score comes from the action and the result. The situation and task tell you how hard the problem was, which helps separate a 4 from a 5. Mapped onto the anchors:
- 1: no real example, a hypothetical told as if it were a story, or a story with no action of the candidate's own.
- 3: a real situation and clear actions, with a vague or missing result.
- 5: specific actions they took themselves, a result with numbers or consequences, and what they learned or changed afterward.
When an answer is missing a part, ask for it before you mark it down. "What did you do, specifically?" and "What happened in the end?" are fair follow-ups for every candidate. For questions built to draw out STAR answers, see the competency-based interview questions list, or pick competencies in the free interview question generator.
How do you calibrate interviewers on a scorecard?
Calibration means different interviewers give the same answer the same score; our guide to interview calibration has a one-hour session for it. Four habits do most of the work:
- Score before you talk. Each interviewer scores alone, straight after the interview, before any debrief. OPM's guidance on structured interviews describes the usual panel method the same way: each member rates individually, then the group discusses significant differences.
- Practice on the same answer. Before the first real interview, have every interviewer score the same recorded or written answer. Talk through any scores two or more points apart, and rewrite the anchor that caused the disagreement.
- Read the evidence before the numbers. In the debrief, each interviewer reads what the candidate said, then gives their score. Scores announced first pull everyone else toward them.
- Watch each interviewer's spread. OPM's training lists the familiar rating errors: central tendency (all 3s on a 5-point scale), leniency or strictness (everyone high, or everyone low), the halo effect (one strong area lifting every score) and "similar to me". A look at each interviewer's scores across their first ten candidates shows who drifts.
What are the most common scorecard mistakes?
- Scores with no notes. A number without evidence can't be checked, questioned or compared. No evidence, no score.
- Too many competencies. Ten competencies in a 30-minute interview leaves three minutes for each, so the ratings come from impressions. Keep it to three to six.
- "Culture fit" as a competency. It's too vague to anchor, so it fills up with "people like us", the same "similar to me" error the rating guidance warns about. Name the behavior the team needs instead, such as "gives direct feedback" or "works without close supervision", and write anchors for it.
- Filling the card in after the debrief. By then, the group's view has replaced yours.
- Changing the weights after meeting candidates. Weights set after the fact tend to favor whoever the panel already liked.
- Averaging away a disagreement. If one interviewer gave a 2 and another a 5, the 3.5 hides the interesting part. Talk about it.
How does DIO's scorecard work?
DIO, the AI interviewer Dio makes, gives your team a scorecard built on the same rules. Each competency is rated 1–5 against anchors written for that role, and the scores are weighted into an overall score. Every competency score has to quote the candidate's exact words, and the quote is checked against the transcript. Two different AI models grade each competency independently, and when the evidence isn't there, or the two disagree about whether it is, the score is left blank rather than guessed.
Your team reviews that scorecard alongside the recording, the transcript and the candidate's résumé, and a person makes every hiring decision. To test any vendor's scores before you rely on them, see how AI interview scoring works.