One of the clearest trends I observed is the inverse correlation between religiosity and verbal IQ (the GSS includes a verbal IQ test). In other words, people who are more religious tend to be less intelligent. [Emphasis his.]
He then goes on to compare people's belief in God, the nature of the Bible, and feelings on some church-state questions with their scores on the verbal IQ test. Although only 13.4% of all people with high verbal IQ scores (which he defines as having a WORDSUM(9-10)) were atheists or agnostics, twice as many people who are atheists, agnostics, or don't believe in a personal God fall into the high verbal IQ category than those who believe in God "some of the time," believe despite doubts, or believe without doubts.
Of the people who believed the "Bible is an ancient book of fables, legends, history, and moral precepts recorded by men," there were almost twice as many with high Verbal IQs as there were in the group which believed the "Bible is the inspired word of God but not everything in it should be taken literally, word for word" and almost six times as many as in the group that believed the "Bible is the actual word of God and is to be taken literally, word for word."
In the comments, I wondered if this could simply reflect higher education levels (more education means less religiosity, on average) or racial differences (Blacks are more religious and have significantly lower average IQ scores, probably due to socioeconomic circumstances), but Half Sigma points out that the results hold even when correcting for these factors. (Ashkenazi Jews have significantly higher IQs, on average, but I suspect that we are correspondingly more likely to be atheists than most other ethnic groups. There seem to be a lot of Jewish atheists out there. Also, we make up a tiny percentage of the American populace.)
Theists will no doubt argue that people at the high end of the verbal IQ curve are simply too smart for their own good, or else are correspondingly lacking in other areas. One emailer told me he believes that atheists have something akin to a learning disability -- a spiritual disability. Others will take comfort in the fact that even among the high verbal IQ crowd, a majority believe in some sort of God. There's also the possibility that being a non-believer raises your IQ rather than the other way around, but that seems unlikely.
Still, I think the fact that atheism, agnosticism, and not-believing-in-a-personal-god-ism (Einsteinism?) are correlated with both higher verbal IQs AND more education is suggestive that maybe we non-believers are onto something. :-)