I was reading a Digg article about atheism, and I was sure the percentages of the population were incorrect. So Googleing up an answer gave me the best answer yet.
There is a great deal of room for debate as to the accuracy of any method of estimation (of atheism), as the opportunity for misreporting (intentionally or not) a category of people without an organizational structure ( read church) is high.
That about said it all, no way to count them!
( (editorial) )