It's possible. (I suppose new members are nominated and elected by existing members, and people may tend to favour candidates who resemble themselves and be influenced by politics, religion, skin colour, etc., etc., etc.)

Yes, that's the sort of thing I'm thinking of. People (in general) are usually more comfortable associating with people who share their opinions.

The last paragraph of the Larson-Witham letter to Nature looks to me like (weak) evidence against a strong atheistic bias in the NAS, in that if there were such a bias I would expect its public utterances and those of its leaders to be a bit less conciliatory. As I say, weak evidence only.

Very weak evidence; it's easy to be conciliatory if one can also be smugly superior in pointing out how wrong the other party is (which is one possible, not necessarily correct interpretation of the last sentence of that paragraph).

(There could also be a bias in responses; maybe atheists are more cooperative in surveys or something. I would expect any such bias to be small and it's not obvious to me which way it's more likely to go.)

That is a point which I had not considered. I'm not sure which way it would go either (unless they did the survey by phoning people at their homes on a Sunday morning, when many Christians would be at church, but that would just be stupid)

Yup, I did. I've linked the right one now. Sorry about that.

Ah, thanks.

...that Gallup evolution poll at the start seems quite telling. It suggests that the difference between scientists and the general public is entirely in the (much larger) rejection of young-earth creationism. This fits with my expectations (which is probably why I draw attention to it).