A story bank that covers the job

Past-behaviour questions come from Janz (1982), credited by Levashina and colleagues as the origin of the approach, which rests on the premise that past behaviour predicts future behaviour. Situational questions come from Latham and colleagues (1980). Levashina's review puts the two side by side using the studies that compared them: uncorrected validity of .31 for past-behaviour questions against .25 for situational ones in Taylor and Small (2002), and .31 against .26 in Huffcutt and colleagues (2004). Their summary is that situational questions have slightly lower validity than past-behaviour ones, especially for complex jobs. Interrater reliability is close for both: .79 for situational and .76 for past-behaviour questions.

The authors do not recommend picking one. They recommend using both, and many structured loops do. For you the message is plain: "tell me about a time" is asking for an event, with you in it, that an interviewer can write down. It is not asking for your philosophy of stakeholder management. If the story you reach for has no date, no decision and no result, it is a view, not evidence.

One more finding matters if you are early in your career. Levashina reports, from Gibb and Taylor (2003), that applicants with little job experience can draw on non-work examples. A volunteer project, a course group, a club you ran: these are real events where you made decisions, and they count.

A situational question tests your reasoning, and a good answer does reason: what you would check first, who you would talk to, what you would refuse to do and why. The strongest answers then anchor that reasoning in something you did. "I would ask what decision the chart is meant to support before drawing anything. I did exactly that last spring, when..." turns a hypothetical into evidence without dodging the question.

Two things to avoid. Do not answer a "what would you do" question with an unrelated story you happen to like; answer the hypothetical first, briefly, then anchor. And do not invent a past event to anchor in. If you have never faced the situation, say so and reason it through; an honest "I have not been in that exact spot, here is how I would approach it" scores better than a story that collapses at the first follow-up.

Situation and task set the scene in a sentence or two: where you were, what was at stake, and what was yours to do. The task matters more than people think, because it tells the interviewer which decisions were yours. Action is the biggest part and it says "I": what you checked, what you decided, what you said to whom, and why. This is the part a structured interviewer's rubric is written against. Result is something you can state: a number, a decision that changed, a practice the team adopted, or an honest "it did not work, and here is what I learned".

Ren's story card above is constructed down to the last number, and yours will differ. Notice its proportions. The context takes two short sentences. Four actions, each one a verb Ren did. The result leads with something checkable and then says what changed afterwards.

The consulting firm DDI's own page says it introduced the STAR method in 1974 with its Targeted Selection interviewing system, and defines the S as "Situation/Task". That is a company describing its own history. No independent history of who coined the acronym was found, so the honest form is "DDI says it introduced STAR in 1974", not "STAR was invented in 1974".

The peer-reviewed story is separate. Behavioural, past-behaviour interviewing is credited in the research literature to Janz (1982), as cited by Levashina and colleagues; the 1982 paper itself was not read. So there are two different claims: a research origin for asking about past behaviour, and a company's claim to the acronym used to answer. The variants you meet on career sites, SAR, PAR and CAR, have no known origin at all. They are folk rearrangements of the same idea, and choosing between them changes nothing that matters.

The "we" story. "We noticed the test was peeked at, we re-ran it, we held the launch." Every sentence is true and none of it is evidence about you. The interviewer cannot score a team. Say "we" for the context and "I" for the action, and if your part was small, tell the small part precisely.

All situation, no action. Two minutes of background on the product, the reorganisation and the difficult stakeholder, and thirty seconds of what you did. It feels like necessary context to the teller. To the listener it is waiting. The figure's proportions are the fix: context in two sentences.

No result. The story stops at "so I sent the note". The interviewer is left guessing whether it worked. If you do not know the outcome, say what you know and what you would check; if it went badly, say that and what you changed. A stated result, including a bad one, beats a missing one every time.

Write the competencies down the side and the stories across the top, then mark each cell: strong if the story contains a specific action of yours and a result you can state for that competency, thin if it touches the competency but the evidence is weak, empty if it does not. Ren's fits are constructed; S3, "Choosing the one metric for a launch", was written for metric choice and turns out to be strong for explaining results and thin for turning a vague ask into a question. That is three rows from one story, at no cost.

A stretch is different. Stretching is telling a story for a competency it does not evidence, because you like the story or have nothing else. A structured interviewer scores each question against its own rubric line, so a great story about experiments, told in answer to "tell me about a time you learned from a mistake", earns nothing on the mistake line. Tagging finds the honest extra rows. Stretching invents rows that are not there, and the empty cell is still empty.

Behavioural rounds have a clock, and a four-minute story eats the next question. A common coaching rule of thumb says about two minutes per answer. It is a rule of thumb: no study behind a particular answer length was found, and an interviewer who wants more will ask for it. Trim toward the limit by cutting situation first, keep the actions, and keep the result. The build-along trims Ren's two longest stories this way.

Impression management is the research name for shaping how you come across, and Levashina and colleagues' 2014 review is frank about it. Across 13 studies and 2,202 people, self-promotion correlated r = .26 with interview ratings. Citing Ellis and colleagues (2002), they write that almost all applicants use some form of impression management in structured interviews. So presenting your work well is normal, expected and rewarded. In their summary of Barrick and colleagues (2009), the link between verbal impression management and ratings was .27 in low-structure interviews and .17 in high-structure ones: structure dampens it without removing it.

Deceptive impression management is something else: inventing or inflating what you did. Levashina and Campion (2007), as summarised in the review, found that interview structure decreased faking, and that probing follow-up questions increased it. The second finding is a warning, not a licence. Probing gives a faker more room to elaborate; it also gives an interviewer more places to notice that the details do not hold together.

The working line: choosing which true story to tell, putting your actions first, and stating your result plainly is presenting well. Claiming the team's work as yours, rounding a result up, or describing a decision you did not make is deceiving. A practical test is the follow-up. If "what exactly did you do on the second day?" would make you uncomfortable, the story needs editing toward the truth, not more rehearsal.