Her source is not the Pew Trust survey (I read all their questions).
Her methodology seems to be pretty much the same as the Pew, judging from above posts. Randomization seems to be nationwide, judging from above, and survey size is quite a bit above the usual politcal polling surveys, which are usually around 1000 randomized subjects, which produce statistically significant results for a population the size of the US.. That should significantly reduce the margin of error.
Her publications are consistent with the usual academic number, actually pretty good.
She is listed on other sites as Edgell Becker for older cites, and Edgell in later cites, but if discussing a name change isn't a red herring for an intelligent discussion, I don't know what is (regarding certain posts above).
And reading her interviews, she is obviously interested in changing roles of religion in people's lives, and changing outlooks shaped by religious views. That's one of the things that sociology of religion is all about. Certainly exploring the views of Americans regarding religious faith or lack of it is consonant with that, and I see no evidence of her personal religious views affecting that.
And doitano (only encountered him/her once before, unsure of name, and everytime I scroll up when in the midst of an answer, I lose everything I've typed, so I'm not gonna do it here, and set calls him d.i., you know who I mean) cites a range of atheism in Sweden because that's what the surveys show
http://www.pitzer.edu/academics/faculty/zuckerman/atheism.html
not because of any suspicious imprecision on his/her part.
I suspect that that post was meant as a semi-humorous post at the anti-atheists, so jumping on him/her, strikes me as inappropriate. Tho we do know that correlation is not the same as causation. Don't we.
Seems to me this topic is suffering from a bad case of "shoot the messenger".