Liberal or Conservative?: The Measure is the Meaning By Nat Ehrlich

The terms "liberal" and "conservative" are used quite ubiquitously, in politics, news, survey research, and everyday conversation, to describe people's ideological leanings, as though everyone knows what they mean; but these classifications can be interpreted in any number of ways. In the spring of 2002, the University of Michigan's Survey Research Center conducted a survey that lent itself to an exploration of the ways people in different populations use these labels to describe themselves.

The original purpose of the study was to update a mail survey that had been done some seventeen years earlier for the U.S. Sportsmen's Alliance (USSA), the central focus of which was to learn how hunting was viewed by five groups: elected officials, wildlife professionals, outdoor writers, media (print and TV), and leaders of sportsmen's groups. Having fulfilled that assignment for the sponsor, we decided to use the data collected in the survey to learn more about what political orientation meant to these groups.

The survey was conducted via the worldwide web. We targeted every state and federal elected official-representatives and senators-and every governor in the United States as of September 2002-about 8,000 all told-for a total of 908 respondents, and used samples provided by the client for the other four groups. We conducted a telephone follow-up of a randomly selected group of nonrespondents to see if the responses we got over the phone were different from the those obtained via the internet; if they had been, we would have weighted the results to reflect the opinions offered by the nonrespondents where they differed from the respondents, but there were no substantial differences, so the data remained unweighted.

Survey respondents were asked to click on a button along a seven-point scale, from liberal on the left to conservative on the right, to describe their positions on social issues and again on economic issues. We found that the five occupational groups differed among themselves more than within themselves (that is, each group was more homogeneous than the total sample), and also that the relationships between social and economic orientation were different for each group.