******************** THIS BLOG HAS MOVED TO WWW.LEGALINSURRECTION.COM ********************

This blog is moving to www.legalinsurrection.com. If you have not been automatically redirected please click on the link.

NEW COMMENTS will NOT be put through and will NOT be transferred to the new website.

Friday, March 11, 2011

Polling 101, Last Class: You Can’t Learn Much From Asking People Their Ideology

This is the fifth in a series of GUEST POSTS by Matthew Knee, a Ph.D. candidate at Yale University specializing in campaigns and elections, ethnic voting patterns, public opinion, and quantitative and experimental approaches to political science.  While this is the last post in the Polling 101 series, Matt will continue to do guest posts on polling issues in the future.

You Can’t Learn Much From Asking People Their Ideology

I get frustrated every time someone on the right declares that we are winning the ideological war or an extreme candidate can win, just because some poll showed that many more people in the country/state/congressional district call themselves conservative than liberal.

First of all, many people do not know enough about themselves or the political spectrum to have a consistent ideological system. As I explained elsewhere in the primer, individual views on specific issues are usually assembled on the spot from values and considerations activated by questions. People’s answers often vary across question wording or even multiple askings of the same question. If views on individual issues are so unstable, how can one expect individuals to be ideologically consistent? Those not on the higher end of the political knowledge spectrum are thought by many not to have consistent ideological systems, even if they may identify consistently with an ideological group.

Truly, purely, ideological Americans are firmly in the minority. Relatively few of those who report that they are conservative are True Conservatives in the sense that purists mean it. More methodologically rigorous surveys ask many questions and construct an ideological scale by combining that information. There is a good deal of ideology to be found beneath the surface, but it is found through data-mining the responses of those who are far from ideologically pure.

Those who believe that conservatives consistently outnumber liberals by a wide margin should ask themselves why the parties are so evenly matched in the average election.

It is not because there are all these conservatives who do not turn out because a candidate is not conservative enough. In 2008, Gallup had the national ideological breakdown at 37% conservative to 22% liberal. The exit polls showed a basically identical 34% - 22% in a race where the right was about as demoralized as possible.

Moderates often rival conservatives in number, and more conservatives still vote Democrat than liberals vote Republican. Twice as many did so in the 2008 presidential race, and a year later Gallup found that 4% of Republicans were liberal but 21% of Democrats were conservative (numbers that have remained very similar since at least 2000).

We must be realistic about our prospects. We live in a closely-divided country with a non-ideological, often low-knowledge center whose swings largely dictate the balance of power. Conservatives can rejoice that the label “conservative” is in a lot better shape that the label “liberal,” but we should not confuse that information with the prevalence of actual conservatives and liberals in the country. We certainly should not confuse that with the prevalence of “pure” conservatives and liberals, who are rarer still.

Follow me on Twitter, Facebook, and YouTube
Visit the Legal Insurrection Shop on CafePress!
Bookmark and Share


  1. Thanks for a great series! Valuable and informative.

    I'm curious about the validity of "standard" polls, the ones that get asked over and over. For example the Consumer Sentiment Index http://www.businessweek.com/news/2011-03-11/u-s-michigan-consumer-sentiment-index-fell-to-68-2.html.

    Do fluctuations in this number mean anything at all? It looks like an odd kind of opinion poll that masquerades as economic data.

  2. Matthew, thank you for this series of posts. I've saved them as a Word document for future reference. I'm always trying to increase our daughter's interest in politics, and as she loves quantitative approaches to social phenomena and number crunching, these posts should help me in my never ending crusade. Hope to see you back soon posting again.

  3. This is why the hardcore conservatives and liberals that give post-mortum analysis and say "if our guy was just more pure we would have won," they are hopelessly delusional and self defeating.

    A lot of the time, the American people just go with whoever sounds more confident and has more cohesive sounding ideas.