Behavioral interview questions ask a candidate to describe something specific they did in a real past situation, not what they would do in a hypothetical one. The reasoning is plain: someone who has already handled an angry customer or a colleague who missed a deadline has shown you how they handle it, and habits repeat. This guide is for the employer side of the table: how to write behavioral interview questions that resist rehearsed answers, how to probe, how to score with anchors instead of impressions, and how the same questions work in a recorded one-way round.
Key takeaways
- Behavioral interview questions ask for one real past event, because what a candidate did is better evidence than what they say they would do.
- STAR is not a script for the candidate, it is a checklist for you: if the Action or the Result is missing, you have a story and not an answer.
- Write the probes and the scoring anchors before the first interview, or you will score on how confident the story sounded.
- Behavioral prompts record well, so a one-way round can put the same questions to every applicant before anyone books live time.
What makes a question behavioral
A behavioral question names a competency and asks for one instance of it. The opener is usually 'Tell me about a time when' or 'Walk me through the last time you'. The second is stronger: 'a time' invites the best rehearsed story, 'the last time' asks for whatever actually happened most recently.
The test is whether the answer requires a real memory. If a competent stranger could answer well without ever having done the job, it is a discussion prompt, not an assessment. 'How do you handle pressure?' fails that test. 'Walk me through the last week where you had more work than time. What did you drop?' does not.
- It asks for one event, not a policy or a general approach.
- It ties to a competency the job needs, taken from the job description rather than a list of nice traits.
- It is answerable by someone who has done the work and awkward for someone who has not.
- It has a written note of what a strong answer contains, made before anyone is asked.
How to write behavioral interview questions
Start from the job, not from a list you found. Write down the four or five things this person will be doing in their first three months, then turn each one into a past-tense question. A support role needs a question about a day the queue got away from them. A first sales hire needs one about the last deal they lost.
- 1
Pick four to six competencies, no more
One question per competency. Six probed questions beat twelve skimmed ones, and interviewer attention runs out long before the list does.
- 2
Write each as a single past event
Name the situation type and ask for the most recent occurrence. Do not stack two questions in one sentence: the candidate answers the easier half.
- 3
Write the probes at the same time
Fix the two follow-ups you will use for every candidate, so the help you give is identical for the person you liked and the person you did not.
- 4
Write the anchors before the first interview
If you cannot describe a strong answer in two sentences, the question is too vague to score.
- 5
Fix the order and keep it
Same questions, same order, every candidate. Changing the order changes what people warm up on, and that alone moves scores.
Fifteen behavioral interview questions by competency
Adapt the wording to your role. The competency column is the part that matters: pick the rows that describe the job you are filling and drop the rest rather than asking all fifteen.
| Competency | Question |
|---|---|
| Ownership | Walk me through the last mistake you made that reached a customer. What did you do about it? |
| Ownership | Tell me about something you fixed that was not your job to fix. |
| Prioritization | Describe the last week you had more work than time. What did you drop, and who did you tell? |
| Prioritization | Tell me about a deadline you knew you would miss. When did you raise it? |
| Conflict | Tell me about the last real disagreement you had with a colleague about how to do something. |
| Conflict | Describe a time you had to give someone feedback they did not want. |
| Customer focus | Walk me through the angriest customer you have dealt with. What did you say first? |
| Customer focus | Tell me about a time you had to say no to a customer. |
| Problem solving | Describe a problem you solved where the first thing you tried did not work. |
| Problem solving | Tell me about a decision you made with information you knew was incomplete. |
| Collaboration | Tell me about a project where someone else was not delivering their part. |
| Learning | What is the most recent thing you had to learn quickly for work, and how did you learn it? |
| Resilience | Tell me about a target or goal you missed. What happened next? |
| Attention to detail | Describe the last error you caught in someone else's work. |
| Adaptability | Tell me about a time your priorities changed halfway through the week. |
STAR and the probes that make it work
STAR stands for Situation, Task, Action, Result. It is taught to candidates as an answering structure, which is exactly why you should treat it as a completeness check and not a quality signal. A polished STAR answer can still be empty. Notice which of the four parts is missing, because the missing part is usually the informative one. Worked STAR examples and a cheat sheet are in the STAR method guide.
| Part | What is usually missing | The probe to use |
|---|---|---|
| Situation | Vague timing and scale | When was this, and how big was it? |
| Task | Their own responsibility is unstated | What was yours to decide, and what was someone else's? |
| Action | 'We' instead of 'I' | What did you personally do first? What did you say? |
| Result | No outcome, or only a good one | How did you know it worked? What would you do differently? |
Use at most two probes per question, agreed in advance. Unscripted follow-ups are how a structured interview quietly becomes an unstructured one: you probe the candidate you like until they produce a good answer, and the candidate you do not like never gets a second chance. That difference never appears in your notes.
Scoring anchors instead of impressions
An anchor is a written description of what each score contains, in the specific things a candidate would have to say. Anchors stop a score from meaning 'I enjoyed that answer'. Score each answer immediately, before the next question, and write one line of evidence next to the number.
| Score | What the answer contains |
|---|---|
| 1, weak | A general policy rather than an event, or a story with no stated action of their own. Nothing you could check. |
| 2, adequate | A real event with a clear personal action, but no outcome, or an outcome with no reflection on what they would change. |
| 3, strong | A dated, sized event, their own decision stated in the first person, what they said or did, the outcome including what went wrong, and what they changed afterwards. |
Two rules keep the anchors honest. Do not raise a score for delivery, because fluency mostly reflects how many interviews the person has already sat this month. And do not compare candidates while interviewing: score against the anchor, then rank, or every candidate is measured against whoever came before them.
Interview scorecard templateA blank scorecard with criteria, anchors and evidence columns. No email required.Behavioral interview questions in a recorded one-way round
Behavioral questions record well because most of their value survives without a live interviewer, and the probes can be built into the question: 'Walk me through the last mistake you made that reached a customer. Say what you did first, and what you changed afterwards.' That one sentence gets you the Action and the Result with nobody on the call.
Consistency also comes for free: same wording, same visible time limit, no accidental coaching. In springhire the candidate clicks one link, with no account and no app, records on a phone or a laptop, and the answers are transcribed and scored against your rubric, with a person deciding. Four to six questions and under fifteen minutes is where completion holds up. See the one-way interview page.
Transcripts are the underrated part. Reading is faster than watching, and reading is where you notice that an answer never named a single action. Answers, transcripts and scores land on one report per candidate, and the sample report shows what that looks like.
Mistakes that make behavioral answers useless
- Asking a hypothetical by accident. 'How would you handle an angry customer' is a situational question, and it belongs in a situational interview, not in a behavioral round.
- Accepting the first story. The first answer is the prepared one. The probe is where the evidence is.
- Asking about a situation the candidate could not have been in. A graduate cannot tell you about managing a team, and asking anyway measures imagination.
- Scoring at the end of the day. Notes written after five interviews record who you liked, not what was said.
- Treating a fluent answer as a true one. Ask for the date, the size and the name of the thing.
