March 15, 2010

How are pollsters able to determine whether an American is a liberal, moderate or conservative?

Q. I am always frustrated by polls asking whether one is a liberal, moderate, or conservative. My feeling is that about two-thirds of Americans are liberal on social issues and conservative on economic issues. (In other words they are actually Libertarians.) Can’t you ask this question better? Even laying out “litmus test” questions on gun control, abortion, the effect of more or less taxes and deficits, gay marriage, national defense (foreign adventures), space exploration, size of government, global warming (and what to do about it, assuming it exists), etc. I fear that many people answer “moderate” because they are taking an average, so to speak, while having very strong but inconsistent and diverging opinions — anything but moderate.

As you note, the standard ideology question assumes that most people are arrayed along a single left-right political dimension. For the reasons you point out, that doesn’t work for everyone. Unfortunately there is no simple solution to the problem, because we don’t have the luxury of asking a series of questions on every poll that could be used to classify people more precisely.

However, in a 2006 analysis we attempted to map our survey respondents onto a two-dimensional space with economic issues on one dimension and social issues on the other. Although the selection of questions we had to work with was not ideal, we found that the simple left-right continuum did not work for many people (about one-quarter of the public); 16% of Americans were economically liberal but socially conservative (sometimes called “populists”), while 9% were economically conservative and socially liberal (or “libertarian,” as you describe them). In addition, the analysis showed that about one-in-five people (18%) were “liberal” in both their social and economic views, while 15% were “conservative” on both dimensions. The plurality of respondents (42%) were “ambivalent,” offering a mixture of ideological views or expressing no opinion on several of the items we asked about. Read the report here.

A more extensive effort of this sort is our “political typology” project, conducted periodically (and most recently in late 2004). You can take the test and identify your typology group here and read the full report here.

Scott Keeter, Director of Survey Research, Pew Research Center and Gregory Smith, Senior Researcher, Pew Forum on Religion & Public Life