Historically, they'd be very interested in what you'd say to your friends, family, etc. But the allure of internet fame would seem to have a more pursuasive effect.
But more importantly... Participating in online communities, and caring about one's online presence, is statistically normal behavior for working-age people in first world countries with Internet access... Moreso the younger you get. But the point is, it's massively normal behavior.
So if you tried to use your questions to predict potential leakers, you're just playing the "Most serial killers eat bread" game.
In order for your question to be a useful predictor, it would have to have low rates of both false positives, and false negatives. That doesn't seem very likely.
Seems a more difficult question, but a more indicative one, which is why I included it.
Per my experience, you can easily find individuals who spend similar amounts of clock time in online communities, but are very different in how much the community can influence them.
F.ex. HN could offer me dang's karma count, and I wouldn't feel any compulsion to leak classified information here.
They're not included in Standard Form 86, latest revision November 2016: https://www.opm.gov/forms/pdf_fill/sf86.pdf
That's not to say it's excluded from polygraph questions; I can imagine a world where online associations are very relevant to an agency in need of a CI or full scope poly for compartment access.
It's similar because there are lots of false negatives, but few false positives.