A New Look at the Old South

I hadn't worked with the DW-Nominate scores before my last post:  I just knew in a general way how they were constructed.  According to the online documentation (see the link above),  "the first dimension can be interpreted in most periods as government intervention in the economy or liberal-conservative in the modern era. The 2nd dimension picks up the conflict between North and South on Slavery before the Civil War and from the late 1930s through the mid-1970s, civil rights for African-Americans."   So the calculations in that post were based on the scores from the first dimension.

However, out of curiosity I took a look at scores for individual senators and found something puzzling, as seen in this scatterplot.

Southerners have the highest scores on the second dimension, which makes sense.  What seems less reasonable is that Republicans have substantially lower scores (ie more "liberal") on the second dimension than northern Democrats.  To take a particular example, Barry Goldwater, who is at the intersection of the two lines, has a lower score than all but one northern Democrat, even though he voted against nearly all civil rights bills.  

This means that the interpretation of Dimension 1 as economic issues and 2 as civil rights is problematic:  both dimensions are a mix of both types (at least at this time).  As an alternative measure, I considered the scores from the Americans for Democratic Action, a liberal group.  Their website gives not just the total scores, but votes on each of the items used to compute them.  I distinguished between civil rights issues and other issues.  On the average, northerners ranked 45th on the "other" issues, and southerners ranked 69th.  12 of the 30 most conservative Senators, and only one of the 30 most liberal, were from the south.  That is, the regional differences were almost exactly like those that the DW-Nominate scores show today.  So the conventional wisdom about the political shift being just one of party, not of ideology, seems to be correct.  

It would be interesting to redo the whole analysis using the ADA scores, but that would take more time than I care to devote to a blog post.