There's a weird press release circulating about how Natural scientists are less likely to believe in God than are social scientists (though MSNBC tried to stay positive, with the subheading Study debunks notion that science is incompatible with religion). The researcher will be presenting her perliminary findings at the American Association of the Sociology of Religion conference in Philadelphia in the 14th of August between 2:30 and 4:45 PM (in case anyone wants to swing by).

She got a 75% response rate from over 2100 surveys sent to "21 of the top U.S. research universities." How you select your sample is very critical here obviously, but in any case, the researcher found that, 38 percent of natural scientists surveyed said they did not believe in God vs. 31 percent of the social scientists. The range seems to be as follows, 41 percent of the biologists and 27 percent of the political scientists said they don't believe in God.

A 1969 survey of American academics with a sample of 60,000 seems to have suggested the opposite result, that social scientists are more secular than natural scientists (see all the results I'm talking about here). This has led sociologist of religion Rodney Stark to claim that the perception by the public that natural science is godless is false, and has been concocted by social scientists who were strongly influenced by secular ideologies like Marxism. Stark has had some issues with defending this thesis lately because of two surveys performed by Edward J. Larson and Larry Witham within the past 10 years which found two broad results: 40% of natural scientist Ph.D.s affirmed a belief in a personal God (a lower frequency than can be inferred from the survey above), while less than 10% of National Academy of Science affirmed a belief in a personal God. Stark has pointed out that there is selection biasing of these samples in comparison to the 1969 survey, obviously in the case of the latter study, but also in the scope of those who responded to the first one. The same criticism applies to the new research from what I can gather, but it must be kept in mind that a small minority of researchers tends to drive the majority of the advances in any given field. Ultimately, these studies keep coming out because of the whole Religion vs. Science angle that is always circulating in the popular press, but as long as 99% of scientists aren't atheists they'll have something positive to write about (for obvious reasons the 1996 study that showed 40% belief among scientists got a lot more publicity than the 1998 finding which showed only 7.5% belief among elite scientists).

Addendum: When I posted the numbers for the 1969 research godless capitalist offered in the comments to the effect that the higher secularity of the social scientists was tapping into anti-organized religion moods as opposed to genuine secular humanist sentiments. If you look at the questions none of them ask straight out about opinions relating to God, so you had to infer from the frequency of those who self-declared as "religious conservatives" or avowed "no religion." If this new data holds up to be robust, and even taking into account the more select sample, I think there might be something to what godless capitalist suggested, especially considering that the date of the survey was the late 1960s when many of the faculty might have been rather radical and oppositional in relation to institutions.