The feedback problem on your team isn't honesty. It's who you decide can handle the truth.

You've got the number in front of you. It's a real number, tied to something that actually happened this quarter, and it isn't flattering. Before you send it, you pause. Not because you're unsure it's accurate. Because you're running a quick calculation about how the person on the other end is going to take it, and somewhere in that calculation you decide to round it up, soften the language, or leave it out of the conversation entirely. You tell yourself you're being considerate. You move on to the next review.

A new study in Management Science — “The Signals We Give: Performance Feedback, Gender, and Competition,” by Alexander Coutts at York University, Boon Han Koh at the University of Exeter Business School, and Zahra Murad at the University of Portsmouth — put that moment under controlled conditions and watched what people actually do with it.

The design is a lab experiment, not a workplace study, and it's worth being precise about that. The researchers recruited 2,620 people online and randomly assigned them to play manager or worker. Workers took a short cognitive-ability quiz where only top scorers earned a bonus. Managers were shown a worker's profile — including gender — and given that worker's score with varying precision: exact, vague, or none at all. The manager then chose whether to pass the information along as received, blur it, or withhold it. The worker used whatever they got to decide between a safe fixed payout and a riskier, higher-paying competitive one. There's no ongoing relationship here, no reputation at stake, no real job on the line — strangers evaluating strangers on a quiz, once. That's a real limit on how far this travels into your building, and I'd rather say so than let the number do more work than it's earned.

Inside that limit, though, the pattern is clean. About one in four managers withheld feedback even when it was positive. Negative feedback was withheld more often when it was vague and wouldn't change the worker's decision anyway — consistent, the researchers say, with managers protecting a worker's self-esteem rather than protecting themselves. And the withholding wasn't evenly distributed: when negative feedback was imprecise, managers withheld it from women at nearly twice the rate they withheld it from men. Men, meanwhile, tended to discount negative feedback and chase the competitive payout regardless — sometimes to their own financial loss — while women adjusted their choices to match what they'd actually been told.

Feedback is one of the most powerful tools that organisations and educators have to improve worker performance, increase student learning, and spark individual growth, yet our results show that a substantial amount of feedback is still withheld, including positive feedback.Dr Boon Han Koh, as quoted in the University of Exeter news release, July 9, 2026

Zahra Murad put the gender pattern in terms worth sitting with: the effect showed up specifically when feedback is vague and non-instrumental, which she said points to the importance of designing feedback processes that reduce ambiguity and limit room for selective withholding. The researchers also note, based on prior work, that women are not less eager than men to receive hard feedback — so a manager's instinct to spare someone the number isn't reading the room correctly. It's substituting for it.

Here's why this lands differently for you than for anyone else in the building. In a company like yours, there's no HR department standing between your judgment and what your people hear about their own work. You are the feedback process. Whatever instinct you follow when you decide someone can't take a hard number doesn't stay contained to that one conversation — it becomes the model every manager under you copies, because it's the only model they've watched up close. If you round up for the people you've quietly decided are fragile, you've just taught your management layer that rounding up is what care looks like.

The part the research doesn't say, but that I'll say plainly: softening a hard number rarely protects the person receiving it. It protects the person delivering it from an uncomfortable ninety seconds. The study didn't test that motive directly — the managers' own stated reasons, in a follow-up survey, were transparency and a sense of duty. But the behavior split by gender anyway, which is exactly what you'd expect if the real driver were the manager's own discomfort dressed up as consideration. Keep the study's finding and my read of it separate: they found the pattern and a plausible paternalistic bias behind part of it; the claim that it's usually self-protection is mine, not theirs.

The thing to do this week: pull the last three pieces of feedback you gave, to anyone, and check them for precision. Not tone — precision. Did you give an actual number, an actual example, an actual next step, or did you give something soft enough that the person could read it either way? If one of those three was vague and you can tell yourself a story about why that person needed the softer version, that's the one to go back and redo. Send the real number.

If you want to know whether that instinct to soften runs through how you're wired more broadly — not just in feedback, but in how you handle the harder calls under pressure generally — the Leadership Profile is built for exactly that question.