Your sharpest woman may be rated the hardest to work with. The men with the same scores were not.

It is the end of a quarter and you are reading peer feedback on your team. One name keeps turning up with the same words: abrasive, intense, hard to work with. You know she is also the person who finds the error in the model before it ships. You are about to decide what to do about the first part without asking whether the second part is causing it.

A paper in the Journal of Applied Psychology gives you a reason to ask. It was first published online in May 2026 — PubMed lists May 18 — and PsyPost covered it on October 10, which is why it is in front of you now. The title is "Understanding the intersection of gender and cognitive ability on interpersonal outcomes: A multistudy investigation." The authors are Nitya Chawla and Seo-Hyun Kwon (University of Minnesota, Carlson School of Management), Trevor M. Spoelma (University of Iowa), Allison S. Gabriel (Purdue University), Aleksander P. J. Ellis (University of Arizona) and Wen Wu (Beijing Jiaotong University).

How far I got: I read the published abstract through PubMed and the PsyPost write-up. I did not read the full paper. Everything below about the method comes from those two, and the details I attribute to coverage are PsyPost's.

What they did. The abstract describes two multisource studies covering 826 people nested in 140 work groups, using objective cognitive ability scores plus a mix of round-robin and self-report data. According to PsyPost, the first study followed 432 undergraduate business students in small teams on a semester-long consulting project, and the second followed 394 full-time employees in small teams at a Chinese financial company, with cognitive scores drawn from the company's hiring test. In both, teammates rated each other on hostility; PsyPost reports that victimization and help received were self-reported. These are field-style observational data, not an experiment, so the paper can show a pattern, not that anyone caused anyone else's treatment.

What they found, per PsyPost. For women in the first study, higher cognitive ability was linked to higher hostility ratings from teammates, more reported victimization, and less help received. In the second study, women with higher scores were again viewed as more hostile, were also rated as less communal, and reported more victimization and less task-related and personal help. For men the two studies differ, and the difference matters. In the first, higher scores did not predict higher perceived hostility, and PsyPost says victimization and help did not change either. In the second, higher scores were linked to lower hostility ratings, less victimization and more help. So the penalty for women replicated across both samples; the bonus for men appeared in only one.

The authors' explanation, as stated in the abstract, draws on social role theory and the idea of an implied communality deficit: higher cognitive ability reads as a gender norm violation for women but not for men, producing perceptions of hostility and low communality, which are in turn associated with victimization and less help. That is their theory of the mechanism. The data are associations.

Our results paint an important, and difficult, picture for higher cognitive ability women, highlighting several theoretical issues that warrant further investigation and practical challenges to be addressed by organizations.— Chawla et al., published abstract, as listed on PubMed

Why this lands differently for the person at the top. You are the person whose read on people becomes the company's read on people. When peers describe a colleague as hostile, you hear a data point, and it feeds promotions, pay, who gets the hard account and who gets quietly managed out. Nobody above you checks that input. If a perception bias is sitting in the ratings, your signature is what converts it into a career outcome. And in a company like yours the sample is small: a handful of strong performers, one of whom gets the label, and the label sticks because there is no volume of data to wash it out.

What the study does not tell you. It does not say women with high scores are never hostile, or that every hostility rating is bias. It rests on peers' perceptions and on self-reported victimization and help. Both samples were specific settings, business students and one company, and PsyPost notes the moderate sample sizes. Those cautions on generalizing are mine; I did not read the authors stating them in the abstract.

The part the research doesn't say, but that I'll say plainly: if your feedback process is a stack of peer ratings, you cannot tell a hard person from a person the room has decided is hard. The study does not show how to tell them apart. It shows why you have to try before you act.

One thing to do this week. Take the last round of peer or manager feedback and pull every comment that uses a temperament word: abrasive, aggressive, intense, not a team player. For each, write down one specific behavior with a date. If you cannot, mark it as a perception, not an observation. Then check one more thing: whether the people labeled this way are also getting less help on their work. Count it. Do not guess.

How you read people is itself wiring, and it is measurable. The Leadership Profile maps how you process people and pressure.