A new Gallup poll shows that Asian-Americans are the racial / ethnic group in the United States that is most likely to self-identify as liberal. Noting that a large percentage of African-Americans self-identifies as conservative, Matthew Yglesias takes this poll to mean that polling on ideological self-identificaiton (are you liberal, moderate or conservative?) is useless:
I think that what this pretty clearly shows is that ideological self-identification surveys are not very useful. By any objective measure African-Americans are the most left-wing racial group in America. But African-Americans don't seem to like the term "liberal." And even though very few black people have conservative political opinions, 29 percent of them describe themselves as "conservative." I think this proves beyond a doubt that widely touted results about ideological self-identification are producing garbage numbers. Actual people clearly don't understand these terms the same way political reporters understand them. It's just false that more African-Americans are conservative than liberal-that these surveys produce that result doesn't tell us a surprising new fact about black political views, it tells us that the surveys are badly designed.
 Yglesias is pretty close here. Ideological self-identification is not actually a strong indicator of ideological orientation. For example, according to Pew, 67% of moderate / liberal Republicans believe that when something is run by the government, it is usually wasteful and inefficient, while only 44% of conservative and moderate Democrats think the same thing. Given that the more conservative identifying group actually has a more positive view of government efficiency than the less conservative self-identified group, this shows that partisan self-identification is a stronger indicator of ideological orientation than even ideological self-identification itself.
Further, given that 77% of conservative Republicans believe government is usually wasteful and inefficient (which means 23% don't believe that), and 39% of liberal Democrats hold that view (which means 61% don't believe that), it also shows that there is a lot of division over ideological principles among self-identified liberals and self-identified conservatives.
Additionally, views of government efficiency also depend on which political party is in charge of the federal government. For example, the chart above and to the right shows that conservative Republicans had a much more positive view of government during the Bush administration than they did during either the Clinton or Obama administrations.
Ideological self-identification is also not a very strong indication of policy preference. Political scientist James Stimson has also shown how public support for liberal policies has varied quite dramatically even as ideological self-identification polls remain quite static:
This wide variance in policy support occurred during a 40-year period when ideological self-identification did not change much at all. If ideological self-identification was actually an indicator of policy preference, then it would have changed along with changing views of public policy.
So, if ideological self-identification is neither a strong indicator of political philosophy or support for specific policies, then what good are ideological self-identification polls? The answer is that ideological self-identification is the best indicator of partisan preference available. In the 2000, 2004, 2006 and 2008 elections, Democrats averaged 18% of the self-identified conservative vote, and 86% of the self-identified liberal vote. Republicans scored only 12% of the liberal vote, and 81% of the conservative vote. Other than being African-American, you will not find any demographic indicator of partisan preference stronger than being conservative or liberal.
As such, ideological self-identification is actually a much stronger indicator of which political party an individual supports than it is of either that individual's policy positions or ideological principles. Democrats are able to cancel out the Republican advantage in this area by scoring sweeping margins of victory among Americans who do not self-identify as white, no matter how those Americans self-identify ideologically (identity is simply a subset of ideology, anyway). "Self-identified white conservatives versus self-identified non-whites and white liberals" is actually a pretty good thumbnail description of the two major political coalitions in the United States. About 80% of the former group votes for Republicans, while about 80% of the latter group votes for Democrats. If you can find a better one sentence description of the difference between Republican voters and Democratic voters, I would love to hear it. |
STATE BLOGS
