Yeah, the article is a bit misleading in its focus on "employment rates". It's using employment rates on the day of graduation, i.e. people who have an offer li

Yeah, the article is a bit misleading in its focus on "employment rates". It's using employment rates on the day of graduation, i.e. people who have an offer lined up before graduating. People who graduate and then interview, starting sometime later, aren't counted as "employed" in these numbers. Ultimately almost all STEM PhDs do end up with a job within a year or so, as the article later acknowledges. Later in the article they quote the unemployment rate for STEM-doctoral holders as ~2%, as you found. I don't think it's therefore a particularly useful analysis. There are a lot of issues with the STEM-PhD job market, but significant risk of unemployment isn't really one of them. Bigger issues are things like whether the opportunity cost is worth it, and the fact that people who go into a STEM PhD really wanting to do research are often disappointed (there are plenty of industry jobs, but many don't really involve scientific research).