The phenomenon of women's scientific contributions being attributed to male colleagues has a name in the history of science literature and a substantial documented record.
The cases are individually interesting. The mechanism they share is more useful, because it explains why the pattern was systematic rather than a series of coincidences.
The structural features that produced it
Institutional position. For most of the period in question, women in scientific institutions occupied assistant, technician and computer roles rather than principal positions. Authorship conventions in most fields attributed work to the principal investigator, so the same conventions that operated for everyone produced systematically different results depending on who held which post.
Exclusion from the bodies that award credit. Learned societies, editorial boards and prize committees excluded women formally for long periods and informally for longer. Recognition flows through these institutions.
Marriage and employment rules. Many institutions operated formal marriage bars requiring women to resign on marriage. This removed people from the positions where credit accrues at precisely the career stage where it accrues.
Publication conventions. Some fields conventionally listed only senior authors. Some journals and institutions used initials, which obscured attribution in ways that had asymmetric effects when the record was later read.
None of these required individual bad faith, which is why the pattern was so consistent. It was produced by rules applied uniformly to an unequal starting distribution.
The categories of case
The documented instances fall into recognisable types.
Data produced by one person, interpreted and published by another, where the interpretation received the credit and the production did not. This is the most common pattern in the historical record and remains contested in several prominent cases.
Observation and identification work performed by staff whose institutional category placed them outside authorship conventions. Astronomy provides a number of examples, where large classification projects were carried out by staff whose contributions were recorded collectively rather than individually.
Collaborative work attributed to the senior partner, where the collaboration was genuine and the credit was not divided.
Work published under a husband's or supervisor's name because institutional rules prevented publication otherwise.
The correction problem
Historical attribution is genuinely difficult, and the correction effort has its own failure modes worth acknowledging.
Retrospective claims are hard to verify. Laboratory notebooks, correspondence and institutional records are incomplete, and the standard of evidence required to reassign credit is high for good reason.
There is a tendency in popular retellings to overstate. Cases where a contribution was significant but not decisive are sometimes presented as cases where the entire discovery was appropriated, and this is unhelpful — it makes the well-documented cases easier to dismiss by association.
Collaborative science genuinely does involve distributed contribution that is difficult to apportion. Some disputed cases are disputed because the underlying question is hard, not because anyone is concealing anything.
The honest version of this history includes cases that are clear, cases that are contested, and cases where the popular account has outrun the evidence.
The contemporary version
The formal barriers are gone. Attribution problems are not.
Studies of authorship in collaborative research have found differences in how credit is allocated, with women less likely to be listed as authors on work they contributed to in some analyses.
Research on patents has found differences in whether contributions are named as inventors.
Studies of how research contributions are attributed in the memory of colleagues have found that recall of who did what differs systematically.
And studies of co-authored economics papers have found that women receive less credit for collaborative work in tenure decisions than men do, with the effect concentrated in papers co-authored with men.
That last finding is particularly instructive because it concerns interpretation of an unambiguous record. The authorship was correct; the reading of it was not.
What actually addresses it
Contribution statements. Many journals now require explicit statements of who did what, which converts an ambiguous author list into a specific record. Uptake is uneven and the statements are of variable quality, but the mechanism is sound.
Persistent researcher identifiers, which make the record traceable and resistant to name changes.
Documentation habits at the individual level. A contemporaneous record of contribution is the only defence against a disputed recollection years later, and it costs almost nothing to keep.
And an awareness, when reading a historical record, that the record was produced by institutions with rules — and that the rules determined whose name appeared, quite independently of who did the work.