Philip and Jordan emailed me about this study yesterday and it looks like Dr. Three Sticks beat me to it. I am still digesting it and have requested some data from Pew (which I am guessing they have a policy not to release) to do a little analysis of my own on the numbers. Here is the link to Pew.
I do want to offer a few thoughts about research and data. The only way to have a 100% accurate data set is to get every single person in a given population to respond. Because that is impossible research groups like Pew, Barna, etc do polling of a random group of people. That is called a sample. The larger the sample size the less amount of error is typically found in a data set. For instance, if you asked ten people who they were going to vote for president and tried to say that was representative of all of America it just wouldn’t work. The margin for error there is huge because those people are no where close to representing the demographics of voting Americans. 100 would be better and 1000 even better. This Pew study had over 35,000 respondents. That makes for some pretty accurate results. I haven’t run the numbers but you would think a sample of that size would give results that were at least 98-99% accurate. That would mean 1 in 50 calculations have a possibility of being due to random chance. That means if you take 500 measurements and report the results, 10 results with statistical significant differences would probably be in error.
Statistics can be a scary thing. You often see what you want to see and notice the things that confirm what you already thought and miss the parts that go counter to your thinking. I wish this study would have run some cross-tabs calculations on this data to get an even better breakdown. When you run one of those you get one variable viewed in relation with another. So you could ee breakdowns like how many of the conservatives believed in which type of God (personal, impersonal, don’t know) vs. liberals, vs. moderates. You could see things like which type of church attenders think God is like or which political leanings have which types of church attendance. There is so much you could do with it.
So Pew, let me see your data…please?