> We use SED data to measure the number of PhD graduates in our focal STEM fields who were non-US citizens at the time of graduation.

[...]

> Emigration rates across PhD cohorts correlate strongly with the foreign national share of graduates ( = 0.86 to 0.95) [...] In all cohorts, the 5-year (15-year) emigration rate is approximately 25% (50%) of the foreign national share

I am not sure if they did this on purpose or not but they missed putting that critical part in the title or right in the abstract. The majority here are not US citizens but foreign nationals. And, most importantly, I couldn't find where they mentioned (or maybe they don't) that these students are studying in US on non-immigrant visas. They're not supposed to or expected to stay after they are done studying. Some stay if they find a company to sponsor them for an internship (Optional Practical Training) but unless they change their visa type they're still expected to leave for their home country.

Without that part highlighted it makes it sound like these US citizens who were born and grew up in US, went to universities here, and then graduated and went to work in China or Europe or something. There is a number of those but, it's not the majority. Maybe they can study just that cohort separately, I think that would be a more interesting thing to look at.