Marc Ambinder, commenting on the recent Pew poll:
bq. When we think of Republican-leaning independents, our binary conception of politics draws the mind to conclude that these folks tend to be more centrist than the average Republican. But that’s not correct. In fact, the evidence from this poll is that they are more conservative; they reject the Republican identity not because it’s too conservative, but because it did not reflect their values enough. The GOP is shrinking as a party, but the number of people who’ll vote for Republican candidates is fairly constant. The tranche of Americans who occupy the space between Republican self-identifiers and the extreme reactionary right are the most politically engaged, and the most angry.
Here are some data from the 2008 American National Election Study that speak to Ambinder’s specific point about the ideology of Republican-leaning independents. Here, the ideology measure is rough-and-ready. Respondents place themselves on a seven-point scale from liberal to conservative. So the measure reflects self-identification, not positions on actual issues. Let’s ignore for the moment those who do not place themselves on the scale. The scale is coded 0-1 where 1 equals the strongest conservative, .5 equals “moderate,” and 0 the strongest liberal. Here are the means, for each category of party identification.
Strong Dem: .36
Weak Dem .40
Ind leaning Dem .40
Ind .52
Ind leaning Rep .74
Weak Rep .68
Strong Rep .79
Of course, these data can’t be read as indicating anyone’s literal location in ideological space. But spacing of parties relative to each other is the point. The data suggest that, first, Republican-leaning independents do tend to identify as conservative, and slightly more so than weak Republicans. “Slightly” is the key word, however: there are few differences among Republicans of any stripe. And, second, Democratic-leaning independents are close to the midpoint of the scale than are Republican-leaning independents. Of course, the midpoint of the scale is an artifact of measurement, not necessarily the “center” per se.
Caveats abound when measuring ideology, so take this quickie analysis as preliminary. But its provisional conclusion lines up with Ambinder’s.