Yes, but the secrecy around such matters means I have no way to know if these potential abuses aren't actual. We only ever hear about such data collection when someone points out yet another instance of the alphabet agencies stretching or outright breaking the limits of their authority and subverting attempts at civilian oversight. The only recourse is to trust that the people with access to the data are using it properly, but my faith in the government has never been lower than it is now.
While I'm inclined to agree with the results they do strike me as overly broad and the sort of thing that needs further investigation.
Well of course there are always outliers, but anecdotal evidence doesn't trump consistent data trends.
As far as the reasons for this go, I suspect that the highly intelligent are more likely to ignore nonreligious cultural biases as well because the unintelligent are more likely to just do what everyone else seems to be doing. I further suspect that the highly intelligent with religious beliefs have their faith rooted in entirely different ways than their peers, and likewise that unintelligent atheists are just in it as a form of rebellion, parroting a different group unquestioningly rather than really thinking things out personally.