Paul Krugman writes,
I read a lot of polls, and they suggest that the center of public opinion on the issues is, if anything, left of the center of the Democratic Party.
Is this true? Here is a graph based on National Election Study data from 2004. Each dot represents where a survey respondent places him or herself on economic and social issues: positive numbers are conservative and negative numbers are liberal, and “B” and “K” represent the voters’ average placements of Bush and Kerry on these scales:
Most voters tend to place themselves to the right of the Democrats on economic and on social issues, and most voters tend to place themselves to the left of the Republicans in both dimensions. (See here for more about our research on this topic; the Annals of Applied Statistics article is here.) Just to be clear: I’m talking about survey questions asking people on their opinions on various issues and policies, not about self-identified liberalism or conservatism.
Other data?
This makes me wonder what the basis was of Krugman’s comment above. It possibly arises from choices of which issues to include in measuring public opinion, or maybe how you define the “center of the Democratic Party,” or maybe changes between 2004 and 2007? I dunno, though, because Joe Bafumi and Michael Herron found something similar to what we found, that the average Democratic congressmember was to the left of the average voter, and the average Republican was to the right. I’d be interested to see Krugman’s data in order to resolve the discrepancy.
The difference seems like it might be nothing more than "public opinion" vs "voter opinion". Aren't non-voters more concentrated on the left?
The graphic is not nice! There must be some graphics where jittering is effective, but this isn't one of them. An area-based graphical approach would be better. If you mkae the data available, I'll show you what I mean.
Glen,
Including people who are eligible to vote but choose not to would move things a bit to the left, but I don't think so far as to move the center to the left of the Democrats.
Antony,
OK, I'll send you the data!
This data compares voters' perceptions of their own positions to the same voters' perceptions of the political parties (and/or office holders). What's needed to challenge Krugman's view is data comparing voters' perceptions of their own positions with officeholders' responses to the same questions.
So I think the Bafumi/Herron study is far more relevant. It suggests that — contra Krugman — the median voter is very close to the median member of Congress, but — supporting Krugman — the whole curve of the electorate is to the left of the whole curve of the legisators.
As suggested in your comments in the previous post, this is mainly the result of the votes-seats equation, the regional distribution of partisan views, and (to a lesser extent) gerrymandering.
I think Krugman is more right than he is wrong.
I think Krugman is more right than he is wrong.
On second thought, my comment on the Bafumi/Herron data doesn't lead to this conclusion at all. Krugman is comparing all voters to Democratic legislators; I was comparing all voters to all legislators.
I'll try to be more careful in the future.