That study do reference a bunch of other study, some which did also polls rather than just census data. However, note the mass use of the words "could be", "potentially", "suggests" and so on. I choose to interpret those words as "further study is needed" each time they are used, limiting the article to only what has the word "finding" on it.
I'm in my lab now, but if you insist I can transcribe the reference section of my favorite psychology of women textbook when I get home.
I don't think it's unfair to call out people with obvious agendas (in this case, single-gender education) when they make broad claims about entirely uncontroversial topics within a field itself. Nobody on HN would brook that over an evolution debate; it's sexist, frankly, to allow that to happen here.
