You are not lying on your resume. You just can't remember.

Written by someone who interviews for global-company roles, and once sat exactly where you sit.

It is a Tuesday night and the cursor is blinking in the Experience section.

You know 2024 was the hardest year you have had at work. You know there was a stretch around March where you were the only person who understood the thing that kept breaking. You know you fixed it. What you can actually type, right now, is: Supported operational improvements across the team.

You read it back. It is not false. It is just nowhere near what happened, and you cannot get any closer. Most people in that chair conclude something unkind about themselves: that they did not really do very much, that other people have accomplishments and they have tasks. That is almost never what is happening. What is happening is that human memory does not work the way a blank box assumes it does.

The morning after the space shuttle

On 29 January 1986, the day after the Challenger broke apart on live television, a psychology class at Emory University was handed a questionnaire. Where were you when you heard? Who told you? One hundred and six students answered within 24 hours of the event.

Two and a half years later, Ulric Neisser and Nicole Harsch tracked down 44 of those students and asked again. Scored against what each person had written at the time, the average was 2.95 correct out of seven. Only 11 of the 44 even remembered filling in the questionnaire.

The detail worth sitting with is not the score. It is that the students were confident, and their confidence did not predict who was right. A memory that feels like a photograph and a memory that is accurate are two different things, and from the inside they feel identical.

These were people recalling the most memorable news event of their year. Now consider what a resume template asks of you: recall, unprompted, from a standing start, in a text box, what you did across four years of ordinary Tuesdays.

Remembering is rebuilding, not replaying

This is not a modern finding. Frederic Bartlett showed it at Cambridge in 1932: memories do not fade evenly like a photocopy, they reorganise around what the rememberer expects. His conclusion has held for ninety years and is the most useful thing you can know about your own recall: you do not retrieve a memory, you rebuild it, from fragments plus expectation.

Which means what you can reach depends on the cue you are handed. Psychologists call it encoding specificity, and the neatest demonstration is from 1975: divers who learned word lists underwater recalled them best underwater, and lost a large share of the list on land.

Now sit on the other side of the desk

Someone is reading your page. In front of them is Supported operational improvements across the team. Here is their problem: that sentence is produced by a person who did very little, and it is also produced by a person who did a great deal and cannot reach it. From the page, those two people are indistinguishable. There is no follow-up question available, no pause where the real story would have been.

That is not a reviewer being lazy. That is a reviewer with no information. The flat line does not undersell you slightly. It removes the difference between you and someone who was not there.

What police interviewers worked out in the 1980s

The people who took this problem most seriously were not in hiring. They were in criminal investigation, where a witness who cannot recall a detail is a case that does not close. In the mid-1980s, Ronald Fisher and Edward Geiselman built an interviewing method directly out of the memory research and called it the cognitive interview. Four moves: reinstate the context before asking what happened; report everything, including what seems irrelevant; change the order, because chronology smooths over detail; change the perspective.

Across a 1999 meta-analysis of 42 studies and nearly 2,500 interviewees, it produced 41 per cent more correct details than a standard interview, and the accuracy rate held: 85 per cent of reported details correct, against 82 under the standard method. More came out, and the share that was right did not fall.

You can run a rough version tonight, for free

Take one project, not one job, and do this:

1. Reinstate before you recall. Two minutes on where you were. Which desk, which season, who was in the thread, what everyone was worried about that month.
2. Say it out loud, to a person or a voice recorder. Not into a document. Writing edits as it goes, which is exactly the wrong thing at this stage.
3. Tell it out of order. Start at the worst moment, not the beginning.
4. Ask what an interviewer would ask. What would have happened if you had not done it? Who noticed? What did you have to talk somebody into? What did you get wrong first?
5. Write the resume line last, from the transcript. Not from memory. From what you actually said.

The line you get out of that is rarely grander than the one you started with. It is more specific, it has a Tuesday attached to it, and you can be asked about it. That is the whole test: if a line on your page is not something you could answer for out loud, in a room, it should not be on your page. Not because it is dishonest. Because it is not yours yet.

What this does not do

Remembering yourself accurately does not make you a different candidate. It makes you a legible one. What it changes sits entirely inside your own hands: at the point where somebody does read the page, they are reading you, and not the flat residue of you that survived a blank text box on a Tuesday night. That is the whole of the offer, and it is worth the evening.