Choose one resume bullet and reconstruct the event behind it: the situation, your responsibility, the actions you personally took, and the result you can support. Expand the evidence, not the achievement. If you cannot substantiate a number or a claim of ownership, correct the claim rather than building a longer story around it.
Yale describes STAR as Situation, Task, Action and Result, a way to structure examples from your experience.[50] The original worksheet below adds an evidence check to each part. It is intended for preparing one answer, not collecting an entire library of stories.
Start with a bullet you can actually explain
Pick a bullet relevant to a question you expect to discuss. Read each verb and ask what it commits you to. “Led,” “designed,” “implemented” and “supported” are not interchangeable. A short resume statement may omit context, but the interview answer should not manufacture it.
Before expanding the bullet, identify:
- The specific event, project or period it refers to.
- What you personally owned and what someone else owned.
- The observed outcome and any measurement behind it.
- Details you cannot share or do not remember well enough to claim.
If the bullet combines several projects, choose one real example and say that you are discussing one part of the broader work. Do not stitch different events into a single episode without explaining the distinction.
For example, if a bullet says you supported reporting throughout a placement, an answer can begin, “One example from that reporting work was…” It should not imply that every report involved the same problem and result.
Original evidence-first STAR worksheet
| Part | Fill in from real experience | Evidence check | What not to add |
|---|---|---|---|
| Situation | The context and the relevant problem | What makes you confident this is the right event? | A bigger crisis than actually occurred |
| Task | Your responsibility and the constraint | Who assigned or agreed the work? | Authority you did not have |
| Action | The steps you personally took | Can you describe how you did each step? | Work performed entirely by someone else |
| Result | What happened afterward | Record, observation, feedback or other basis | A percentage with no source |
| Limit | What remains unknown or outside your role | What could a follow-up reveal? | A neat ending that hides the gap |
Use the evidence column to check your memory; it is not an instruction to bring confidential records into an interview. You can establish that a result is supportable without uploading employer documents or identifying a customer.
For an unknown numerical result, leave the number field empty during preparation. Then write the narrower outcome you can honestly describe. “The team used the corrected file for the review” is different from “I improved decision speed by a particular percentage.” Use the former only if it happened; do not treat it as a stock replacement for missing evidence.
A fictional worked example
The following example is invented for explanation. It is not a real candidate's experience, a customer story or a claim about a real employer.
Starting resume bullet: “Supported data quality checks for a monthly operations report.”
The fictional candidate's underlying facts are:
- The report combined exports from different teams.
- Some records used different date formats.
- The candidate was asked to check the combined file before review.
- They flagged the inconsistent fields, confirmed the intended format with the report owner, and reran the affected checks.
- The report owner accepted the corrected file for review.
- No time-saving percentage or downstream business effect was measured.
Notice that none of these facts says the candidate designed a company-wide data platform. The expansion stays inside the original responsibility.
Fill the worksheet before writing the answer
| Part | Fictional preparation note |
|---|---|
| Situation | Monthly reporting file combined team exports with inconsistent date formats |
| Task | Check the combined file before the report owner's review |
| Action | Flag fields, confirm intended format, rerun affected checks |
| Result | Owner accepted the corrected file for review |
| Numerical result | Left empty: no substantiated time or error-reduction figure |
| Limit | Cannot attribute a downstream business outcome to these checks |
Yale's guidance recommends focusing on individual contributions when describing a team situation.[50] Here that means distinguishing the candidate's checks from the report owner's decision to accept the file.
One possible spoken answer
“In a monthly reporting task, I checked a combined file built from different teams' exports. I noticed that the date fields were not using the same format. My responsibility was the pre-review quality check, not ownership of the reporting system. I flagged the affected fields, confirmed the intended format with the report owner, and reran the checks after the correction. The owner accepted the corrected file for review. I did not measure a time saving, so I would describe the result as resolving that inconsistency rather than putting a percentage on it.”
This is a worked example, not a paragraph to copy as your own. In a real answer, the final qualification is needed if it prevents a misleading interpretation; you do not need to recite every possible limitation after a modest factual claim.
Test whether the action survives a follow-up
Take the strongest verb in your answer and ask how you would explain it. If you say “analyzed,” what did you inspect, compare or calculate? If you say “persuaded,” what concern did you address? If you say “implemented,” what did you actually change?
For the fictional reporting example:
“How did you find the issue?” The candidate needs to explain the check they actually used. If the facts do not establish an automated validation script, they should not invent one.
“What changed because of your work?” The supported result is the corrected file being accepted for review, not an unmeasured commercial benefit.
“Who accepted the corrected file for review?” The report owner did. Acceptance for review does not establish approval of every correction or ownership of the reporting system.
For a broader branching drill, see handling follow-up interview questions. The purpose here is narrower: make one resume claim defensible before turning it into a spoken answer.
Repair the bullet if the evidence does not support it
Suppose your actual bullet says “Led a reporting transformation,” but your real contribution was checking one file. Do not use storytelling technique to cover that mismatch. Revise the claim in the relevant application materials through the appropriate process, and use accurate language when discussing it.
If you discover the issue during an interview, you can clarify: “That wording is broader than my role. I supported the data-quality work; the reporting manager led the overall change.” This is an original example phrase, not a guarantee about how a correction will be received.
A correction may feel less impressive than the old wording. That is not a reason to preserve an inaccurate claim. Your preparation should help you explain your contribution clearly, not make it impossible to question.
Make a cue card, not a second resume
Once the worksheet is complete, shorten it to prompts: context, task, action, outcome, boundary. Practice explaining it without reading complete sentences.
If the interviewer asks only about your action, start there. If they ask for a complete example, use the full sequence. STAR is an organizing aid, not a reason to answer every question with all four headings spoken aloud.
Use the paragraph-to-cues exercise for delivery. Use the AI answer verification worksheet if generated wording has added claims that were not in your notes.
Can InterviewRep expand the bullet for me?
InterviewRep's public copy says resume and optional job-description context inform suggestions, which are requested with Space or Generate Answer; it warns that suggestions can be inaccurate.[16] A fluent expansion is not evidence that the events happened. You must supply and verify the facts, and use assistance and capture only where permitted.[16]
The live interview assistance guide covers that workflow. This worksheet also works without a live tool: paper notes and an honest account of one event are enough to do the preparation exercise.
Questions candidates ask
What if the result was not quantified?
State an outcome you can support and identify its basis. Do not add a number because the worksheet contains a result field. An unmeasured result is not permission to invent a measurement.
Can I use an academic or volunteer example?
Yale's guidance includes academic, extracurricular, leadership and internship examples as possible sources of experience.[50] Choose one that answers the question, label its context accurately, and distinguish your contribution from the group's.
Should every bullet become a STAR answer?
This exercise does not require that. Start with one relevant bullet that could support a useful example. Some bullets describe continuing responsibilities rather than one event; you may need to select a concrete episode within that responsibility.
Method and next step
The reporting example, evidence columns and correction language are original editorial material. Yale supports the cited STAR structure and individual-role guidance, not a tested effect from this worksheet. No real candidate outcome, quantified improvement or authenticated product session is claimed.
Choose one bullet now. Complete the evidence column before polishing the wording. If that column exposes a claim you cannot defend, fix the claim first.
Sources
- InterviewRep | Live interview help, in your own words
- The Behavioral Interview – Office of Career Strategy – Yale University
10 minutes for new accounts. No credit card required. Use only where assistance and capture are permitted. Check current plans and terms.