The methodology for the Faith Matters survey is included in the paperback edition of American Grace that will be out early in 2012, but here are the basic facts.
The questions we have asked on our survey are are listed elsewhere on this blog site, and borrowed from standard questions used in other national polls, about the Tea Party and American politics in general. Our 2006 survey was a large (N=3000, as we reported in the op-ed), randomly chosen, nationally representative sample—all ages, all classes, all races, all parties, all religions, and so forth, and each in the right proportions. The survey firm we employed drew the sample from the national population using random digit dialing, whereby every phone number in the country has an equal probability of being called. In 2011 we interviewed everyone from the original sample who is still alive and traceable, and virtually everyone we could find this year agreed to the re-interview. (Because young people move often and change phone numbers even more often, we could not trace as many of them five years later, but the final weighted sample is nationally representative and the 2011 results correspond with all other credible national surveys that we’ve found.) The 2011 interviews were all done this summer. The margin of error varies from item to item, but the sample is several times larger than most national polls reported in the media, so the margin of error is correspondingly smaller. All the generalizations in our “Crashing the Tea Party” op-ed are statistically significant, that is, well beyond the bounds of sampling error. This is a large, high-quality, scientific study.
Download the PDF of this graphic here 
Share
