To debate religion, we must first find out what people believe

 Julian Baggini is doing an interesting series of articles in the Guardian's CiF (comment is free) belief section, and this one raises a very important point, namely that a lot of debate invokes the kind of believer/non-believer out there, with very little empirical evidence to back it up. This raises the risk of stereotypes and anecdotes being treated as fact, and with movements being epitomised by their more vocal , radical (and hence less representative) members.

His full article is :  http://www.guardian.co.uk/commentisfree/belief/2011/nov/14/debate-religion-what-people-believe?INTCMP=SRCH