Why Stress Makes You Say ‘No’ To The Right Things.

What a laboratory found out about pressure and refusal, and why the caution that built your company is now capping it.

The proposal that goes back for more work.

Every growing business has a version of this meeting: the senior hire who joined eight months ago brings a proposal, they are still on their second slide, and the founder has already found the flaw. They say so, reasonably, with two or three sharp questions attached, and it goes back for more work. Nobody thinks anything has gone wrong. That founder has a long record of being right about this sort of thing, which is most of how the company got to eighty people.

Then some version of it happens most weeks for a year, and the company slows down without anyone being able to name the day it started. We file it under judgement, or standards, or commercial discipline.

At the end of 2025 a team publishing in Communications Psychology went looking for something else, and asked a question that probably should have been settled forty or fifty years ago. What does acute stress actually do to the quality of a hard decision?

The team settled on logic puzzles of a type where the difficulty can be calculated rather than eyeballed, and where each puzzle either has a valid solution or it does not. The participant's entire job was to say which. Forty-two people did it twice, once calm and once acutely stressed by a standard laboratory procedure described at the footer of this piece, spitting into tubes at intervals so their cortisol (stress) levels could be tracked, with an eye tracker recording where they looked.

Then the researchers lined accuracy up against the cortisol and found essentially nothing. One full standard deviation of extra stress hormone cost about 1.7 percentage points of accuracy. The journal's own summary describes the effect of acute stress on its own as surprisingly limited, which is a polite way of saying the stressed participants were solving hard problems at more or less the rate they solved them relaxed.

You can imagine the mood in the lab; stress had zero effect on decision making. That was until they looked again.

Then they found the clock.

What mattered was not how much time someone had, it was whether they felt rushed. Among the participants whose cortisol had genuinely risen, that feeling of being under the gun changed the shape of their mistakes, and it changed it in one clear direction, they started saying no.

On the puzzles where 'no' happened to be the correct answer this made them look unusually sharp, and accuracy went up by seven percentage points. On the puzzles where a solution genuinely existed, accuracy fell by eighteen points against the same benchmark, and in absolute terms dropped twenty-five points, down to roughly what you would get from a coin.

Which raises an obvious question. Why say 'no'? Most people assume stress leads to recklessness, so why does it lead to refusal?

Amy Arnsten spent a career demonstrating that the prefrontal cortex, the part of you that holds three possibilities open at once and weighs them against each other, is peculiarly fragile under stress chemistry. When the catecholamines flood in its grip loosens and older, faster circuitry takes up the slack. That circuitry has no interest in weighing anything. It runs whatever has worked before, at speed, which is what it was built for and why you are still alive today to read this. Joseph LeDoux describes the same machinery as a defensive system engineered to be quick rather than correct.

So a stressed brain carries on making decisions by handing the job to the part that decides by reflex precedent rather than accuracy. Or in SSC language: it decides from a defensive, safety position rather than an open, opportunity-led one.

The ‘no’ that built the company.

Now take all of that into a company that is growing, where it meets a scoring system so lopsided it might as well have been designed to keep old habits and old decision making alive.

Refuse something that would have failed and the proof arrives promptly, with an audience. The competitor who launched the same idea and lost their shirt. The hire you passed on who blew up somewhere else eighteen months later. You were right, everybody watched you be right, and it goes into the story you tell about your own judgement, probably for years.

Refuse something that would have worked and nothing arrives at all. The product does not exist, so it cannot sell. The person who asked for more responsibility does the job they already have perfectly adequately and leaves two years later, and the exit interview records the reason as personal. In the Netherlands 86 per cent of employees think this is a good moment to find another job, the highest figure anywhere in the world.

In the first three years the founder's no was almost always correct. There was no money, no slack and no second chance, and a policy of refusing anything unproven was not caution, it was survival, and it worked, which is why the company still exists. The pattern was forged in conditions where it was the right answer, several thousand times over.

Then the conditions change. There is a balance sheet, a team, and a market that will punish you for moving slowly rather than for moving wrong. The policy does not update, because nothing has arrived to tell it to. The founder who was right at twelve people is running the same reflex at eighty, in a company whose entire value now depends on other people being allowed to decide things.

And the physiology is working against them the whole way. Gallup's State of the Global Workplace 2026, drawing on interviews with more than 141,000 employed people through 2025, found 46 per cent of leaders reporting a lot of stress the previous day, ahead of the people they lead and ahead of them on anger, sadness and loneliness as well. Those same leaders top the report on engagement and on thriving, so this is a portrait of load rather than of unhappiness; they are committed, they are doing well, and they are running hot.

Put it together and you have somebody working a diary purpose-built to manufacture the sensation of being rushed from half past eight in the morning, holding the decision rights for most of what the company is trying to do next. They are not visiting the state that laboratory created, they live there.

Why knowing about it changes nothing.

The usual response to a finding like this is to tell the leader about it: send them the article, book the offsite, put them in a founder peer group, and hope somebody who understands the pattern will catch themselves next time.

The trouble is timing. A reflex laid down over a decade and rehearsed several thousand times is quicker than the noticing of it, so by the time the founder clocks what they are doing the no has already left the building and three people are rewriting the deck. Insight turns up reliably and slightly too late, which is why so many founders can describe their own pattern beautifully and go on producing it every Tuesday.

What actually moves is whatever is holding the reflex in place, and it is nearly always protective in origin. Somewhere back there refusal became the safe move, and being right about it often enough has been topping up the lesson ever since. Release that and the same person walks into the same meeting with their standards intact and their commercial instincts intact, and finds a choice where the reflex used to be.

If any of this is uncomfortably familiar there is an easy place to start. Go back over the last quarter, find the proposals that went back for more work, then count how many of them ever came back.

Addendum: what the study can and cannot hold.

Forty-two students, a screen and a stressor lasting minutes, which is not a Tuesday in a scale-up. The time-pressure result was post-hoc, meaning nobody set out to test it and they found it while poking about afterwards, so it wants replicating before anyone builds a business case on those numbers. What it gives us is a mechanism, one that lines up with thirty years of work on what stress does to the front of the brain, and with what anybody who has spent time around founders has already watched happen. Pressure rarely makes a capable person stupid, it narrows what they can reach for, and what they reach for is the move that has never visibly cost them anything.

Addendum: how you stress somebody on purpose.

The participants were put through the Trier Social Stress Test, introduced by Clemens Kirschbaum and colleagues at the University of Trier in 1993 and still the most widely used way of inducing acute psychosocial stress in a laboratory. It goes like this.

You are given ten minutes to prepare a short talk explaining why you are the right person for a job you badly want. When the ten minutes are up, someone takes your notes away and walks you into a room where three people sit behind a table, and you deliver the talk standing, from memory, while they watch. They have been trained not to react. No nodding, no smiling, no encouragement of any kind, and if you dry up before your time is done one of them informs you that you still have several minutes and waits. You are told the whole thing is being filmed and will be analysed afterwards, which is true, and does not help. Then they ask you to count backwards from 1,022 in steps of thirteen, out loud, and every time you slip they stop you and send you back to the start.

Somebody had to design this in 1993 and then talk an ethics committee into it. Across six independent studies the original paper reported salivary cortisol rising to between two and four times baseline, along with significant increases in heart rate and in ACTH, growth hormone and prolactin. Individual variation is large, which is why studies using it commonly analyse the cortisol responders separately from everybody else.

Sources.

• Acute stress impairs decision-making at varying levels of decision complexity (2025). Communications Psychology 3:179. DOI 10.1038/s44271-025-00355-x. Within-participants laboratory experiment, n = 42. The time-pressure interaction described here was a post-hoc exploratory analysis.

• Kirschbaum, C., Pirke, K. M., & Hellhammer, D. H. (1993). The 'Trier Social Stress Test': a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology 28(1-2), 76 to 81. DOI 10.1159/000119004.

• Gallup (2026). State of the Global Workplace: 2026 Report. Data collected January to December 2025, 263,810 respondents of whom 141,444 were employed, via the Gallup World Poll. Netherlands job-market figure from the same report's country-level data.

• Arnsten, A. F. T. (2009). Stress signalling pathways that impair prefrontal cortex structure and function. Nature Reviews Neuroscience 10(6), 410 to 422.

• LeDoux, J. E. Work on defensive survival circuits, summarised across Anxious (2015) and subsequent papers.