y'know this entire intellectual exercise could be resolved quite easily. It does raise a valid question regarding the correlation of belief vs action. Simple way to solve --

Part One: take a clipboard or three, one or two friends, stand in Harvard Yard (or some busy streetcorner by a University) and conduct your own survey. Make sure that you ask the requisite demographic questions for cross-tabulation (age, sex, religious preference, political party, annual income). Three to four hours should be sufficient for a good sample. Then take the same clipboard and hi y'selves to a different part of town frequented by a different group of people (Southie?) and repeat the process. For improved accuracy repeat the process one more time.

Part Two: Calculate the results and cross reference results by age / religious preferences / income. Some fundamental statistics should yield publishable results.

Part Three: Call your local news media and announce your results

Part Four: Wait for the deluge of interviews, emails, phone calls, etc.

dod