The Pattern That Keeps Repeating

Every few years, a new study confirms what women in STEM already know from their own lives: we leave. The pipeline loses us at multiple stages — after undergraduate degrees, after PhDs, after postdocs, after a few years in industry. The explanations offered are usually structural: hostile culture, pay gaps, mentorship deficits, the second shift at home. All of that is real. I've lived most of it. But I've been sitting with a theory that doesn't get as much attention, and I want to put it on paper.

The Belonging Problem Is Different Than We Describe It

When people talk about women leaving STEM, the conversation often focuses on explicit discrimination or hostile incidents. Those happen. They're important. But what I've observed — in my own trajectory and in conversations with dozens of women across fields — is something more ambient than that. It's the feeling of explaining yourself. Not your work — your existence in the room. There's a particular tax that comes with being one of few women in a space: the labor of managing other people's perceptions, of deciding whether to correct assumptions, of reading every room to figure out how much of yourself to bring. This tax is not usually distributed equally. And it is not nothing. Over a career, it's exhausting in a way that's hard to account for on any official ledger. I don't think most of the men I've worked with were aware they weren't paying it. I genuinely believe that. But awareness doesn't affect the balance.

What the Research Finds and What It Misses

Investigators at the American Association of University Women published a comprehensive analysis tracking women's departure rates from STEM careers over a twelve-year period. The primary predictors of attrition were not test scores, family status, or even salary gaps — they were measures of workplace belonging and the perception of whether the field valued contributions like theirs. Women who felt their work was seen as peripheral, or who didn't see themselves reflected in leadership, left. Women who had a mentor who had stayed — who showed that staying was possible and worthwhile — left at significantly lower rates. This is a different problem than "there aren't enough women in the pipeline." It's a problem about what happens to the women already in it.

The Competence Trap

Here's my theory, and I want to be honest that this is a theory, not a study: women in STEM are often subject to what I'd call a competence trap. We know that we're evaluated more harshly on technical errors and less harshly on interpersonal qualities that actually matter for good science — collaboration, communication, intellectual honesty about uncertainty. So many of us adapt by becoming relentlessly, visibly competent. Ironclad preparation. Nothing can be wrong. That posture is exhausting to maintain. And it subtly warps what work feels like. Science, done well, is supposed to involve a lot of being wrong — proposing things that don't pan out, following a thread until it goes cold, changing your mind in public. The women I've watched leave were often the ones who, consciously or not, couldn't afford the psychological cost of that kind of public fallibility. Because they knew the failure would be read differently.

A Tangent Worth Following

There's an interesting comparison in law. Women left large law firms at predictable rates for decades before the profession did serious structural analysis of why. The interventions that worked weren't diversity statements or unconscious bias trainings — they were concrete changes to billing credit systems, assignment processes, and whose work got visibility at the partnership level. The representation problem was downstream of an economic structure problem. STEM is different in important ways, but the lesson transfers: looking at the pipeline without looking at the structure of the field it feeds is looking in the wrong place.

Why I'm Still Here, and What That Took

I want to be careful not to make this essay about resilience as a virtue. Staying in a field that makes you pay a toll is not always the right choice, and framing departure as failure misreads the situation entirely. But I've stayed, and what's made that possible is mostly other people. A mentor who said out loud that the thing I was describing was real. Colleagues who redistributed some of the invisible labor. A team where my judgment was treated as judgment, not a position to be checked. Researchers at MIT's Laboratory for Innovation Science and Policy have found that women's persistence in STEM correlates strongly with whether they experienced at least one period of clear, unambiguous intellectual respect — someone treating their ideas as simply worth engaging with, without subtext. That shouldn't be the bar. But it's the bar. And knowing it's the bar is useful information.