Tony Fabrizio polls; Marc Ambinder reports; Reihan analyzes.
Home | Atlantic FAQ | Masthead | Site Guide | Subscribe | Subscriber Help
Atlantic Store | Educational Program | Jobs/Internships | Privacy Policy | Terms and Conditions | Feedback | Advertise
Copyright © 2009 by The Atlantic Monthly Group. All rights reserved.





I had a few random observations.
First, a technical point. The segmentation method appears to be (they are not clear about it) either clustering or a decision tree. This is standard pollster “analysis”, up to and including the cutesy names. The within-segment variance vs. inter-segment variance on any question using such methods tends to be so high that the segments are a very misleading way to view the data, so I just ignored them and suggest that you do too.
The most depressing observations are that apparently (1) those over 54 have gone from 28% to 41% of Republicans since 1997, and (2) 93% of Republicans are white.
The most interesting observation (to me, at least) is that the “Other” religious category has gone from 5% to 20% of Republicans since 1997. Data on the 2007 survey breaks out this 20% number, and it is 3% atheist / agnostic and 17% what could be called (had the pollsters insisted on internally-consistent terminology) “Other Other”. What is this growing alternative religious affiliation that now claims 1 in 6 Republicans?
Overall, I’d say this survey is bad news (or, more precisely, confirms previously-known bad news) for Republicans.
Best,
Jim Manzi
Posted by Jim Manzi | July 2, 2007 12:18 AM