That graph (which is from here: [1]) lists [2] as its data source; none of the tables under "326 Completion and…" match the data they're graphing; it never brea

That graph (which is from here: [1]) lists [2] as its data source; none of the tables under "326 Completion and…" match the data they're graphing; it never breaks it out by degree+sex, AFAICT; there are a lot of tables, and I did not exhaustively search them, nor did the article include enough information on their methodology. The phrasing of "Over the weekend, I…" implies some exogenous data source, but it's not shared. One can't even begin to replicate the conclusion reached. > Any class with sub 10% women is a big outlier. You claim, but what is being asked for here is evidence to support such; Occam's razor implies that not only should it not be, that it would be an outlier in the other direction. Hence the desire for something that lays out its methodology well enough that we can tell that it's not in that lovely third category of "statistics". [1]: https://www.aei.org/carpe-diem/chart-of-the-day-the-declining-female-share-of-computer-science-degrees-from-28-to-18/ https://www.aei.org/carpe-diem/chart-of-the-day-the-declinin... [2]: https://nces.ed.gov/programs/digest/current_tables.asp https://nces.ed.gov/programs/digest/current_tables.asp