US kids’ doodles of scientists reveal changing gender stereotypes

They're referencing 1986-2016 at 72%, versus ~67% from 2010-2017, as a meaningful improvement. That supposedly meaningful improvement is almost entirely wiped out if you adjust just one of the recent outlier data points down to the average. Or alternatively, to show how little actual improvement there has been in a quarter century, drop the data off pre 1992, that 72% vs 67% improvement might even regress (someone want to test that?). That's pretty shaky ground for such a big claim.

Pretty ridiculous article premise given the data in question. There's obviously a lot more work to do on improving the gender stereotype problem the article is claiming has improved, particularly since it has somewhere between barely changed and not changed at all in decades.

For me scientists are certainly far far less visible than they used to be. It's all programmers now.

What would help would be correlating "draw a person", from the same child and seeing if a scientist varies from that.

I would guess they still draw mostly men. Or rather, I would guess they don't even draw them or think about them. I don't hear many calls for equality there.