This interview scorecard template pairs a question with observable response criteria and space for independent notes. A rating needs a specific part of the answer beside it. The filled example shows how a panel can work through a disagreement, and the blank version lets you prepare a question for your own role.
Start with one question and a usable scale
For this fictional archive coordinator practice interview, the question is:
A delivery manifest lists 12 boxes. The inventory entry lists 11. What would you do before marking the delivery complete?
The panel defines the following illustrative scale before reviewing an answer. It measures the handling of this discrepancy, not the candidate's overall suitability for the job.
| Rating | What the answer needs to describe |
|---|---|
| 1 | Marks the delivery complete or changes a count without checking the source records. |
| 2 | Compares the manifest with the inventory entry and records the unresolved discrepancy. |
| 3 | Does everything in 2, identifies a specific next check, and says who will receive the unresolved discrepancy and next step. |
| Not assessed | The question was not asked, or there is no response evidence to rate. |
“Good communicator” would give the panel too little to work with. The description for 3 asks for something an interviewer can actually hear: a next check and a person to tell. Not assessed stays separate from the numbered scale. It does not mean the candidate gave a poor answer.
The U.S. Office of Personnel Management's structured interview guidance describes using predetermined questions in a consistent order and evaluating responses against the same scale and standards. The archive question and its three rating descriptions here are original teaching material.
Record the answer before discussing a score
The practice response is:
“I would compare the delivery manifest with the inventory entry and note which box is missing before closing the delivery.”
Jo and Kai record their ratings independently:
| Interviewer | Original rating | Evidence note |
|---|---|---|
| Jo | 2 | Compares the manifest and inventory entry; records the missing box before closing. |
| Kai | 3 | Checks the records and notes the missing box. |
Kai's evidence supports the checking and recording criteria. It does not identify the next check or the person who would receive the discrepancy. Neither appears in the response.
In the example, open Compare evidence to put both notes next to the rating descriptions. Then show the proposed calibration note:
Proposed revision for Kai: 2. The response describes checking the two records and noting the missing box. It does not yet describe the additional actions required for 3.
Keep Kai's original entry visible beside that proposal. A panel member can then explain whether the original note missed part of the answer or whether the rating needs to change. The page presents this worked comparison; it does not evaluate arbitrary candidate responses.
If the panel's procedure permits a follow-up, use it consistently and record the new response separately. An interviewer should be able to tell which evidence came from the original answer and which came later.
Prepare a blank scorecard for your question
Choose Blank scorecard and enter the question, the rating descriptions, and the response evidence you want to discuss. Keep the original rating and any proposed revision in separate fields. Build review copy should identify missing information before producing the review view.
Write rating descriptions around the work the role requires. For this example, “names a next check” is more useful than “shows initiative.” For a different role, the behavior might be explaining a trade-off or identifying the source needed to answer a customer's question. Define that behavior before the interview.
Use the scorecard to keep each question's evidence readable. A total across unrelated criteria can conceal a missing answer, so review the individual notes before making a decision through your team's hiring process.
Make your own HTML scorecard
Copy this prompt into an AI assistant and replace the brackets. Supply the question and criteria yourself; the assistant's job is to arrange them into a page you can review.
Create one complete, self-contained HTML interview scorecard.
Role and work task: [FILL IN]
Question, exactly as it will be asked: [PASTE]
Rating labels and observable descriptions: [PASTE]
Meaning of Not assessed: [FILL IN]
Fictional practice response, if wanted: [PASTE OR LEAVE BLANK]
Independent interviewer notes and original ratings: [PASTE OR LEAVE BLANK]
Proposed calibration note, if supplied: [PASTE OR LEAVE BLANK]
Keep my question, rating descriptions, and response wording intact.
Label any practice example fictional. Do not invent candidate experience,
response evidence, validated criteria, or a hiring recommendation.
Keep Not assessed separate from numeric ratings.
Put the question and scale beside the response and independent notes.
Show original ratings separately from proposed revisions. Add a
Compare evidence view that displays the supplied notes against the
supplied criteria. Do not use keyword matching to grade free text.
Do not rank candidates or turn a missing answer into a zero.
Include a blank scorecard with labeled fields for the question, scale,
response evidence, interviewer, original rating, and proposed revision.
Build review copy must identify missing question, scale, or evidence.
After an edit, remove or clearly mark any old generated review as stale.
Keep typed notes when switching views. Ask before resetting typed work.
Do not claim that notes are saved or submitted.
Make the page readable at 360px with no horizontal page scrolling.
Use body text of at least 16px, keyboard-accessible controls, and
inline CSS and JavaScript. Use no network, storage, or external assets.
Return the complete HTML file.
Read the generated scale against the question before using it. If a criterion asks for something the question never invites the candidate to discuss, revise the question or the criterion while preparing the interview.
Agree on the wording before the next interview
The archive panel may decide that “specific next check” is still too vague. They can clarify the preparation copy to say “identify a source record or person you would check next,” then make sure the question gives candidates a fair opportunity to describe that action.
For an HTML preparation copy, NOS lets you edit that static wording on the page without asking the AI to rebuild the layout. Share the agreed question and rating descriptions with the panel. If you correct the saved page later, colleagues can reopen the existing active link to read the correction.
Keep actual candidate notes and hiring decisions in your team's authorized hiring system. The shared preparation page gives interviewers a common question and clear criteria to use when they get there.