Mark Regnerus had a thoughtful article this Wednesday on public attitudes toward same-sex marriage. Public opinion is far more nuanced than many observers realize. As Regnerus points out, slight variations in the wording of survey questions, the population being surveyed, or the preceding questions can cause significant changes in the results. He is also correct that when asked about controversial issues, many respondents will offer the response they feel is the most socially acceptable, regardless of their actual opinion. Regnerus shows that mainstream media have emphasized polls that show substantial approval for same-sex marriage, while offering scant coverage of surveys showing only mixed support.
The media have shown a similar bias when covering surveys dealing with abortion. Surveys asking people to describe themselves as either “pro-life” or “pro-choice” typically receive lots of media coverage, in part because, until a few years ago, the pro-choice position usually received plurality support. Surveys that offered a range of possible positions received less coverage, in part because they showed that public attitudes toward abortion were more nuanced. And survey research firms only rarely ask about broadly supported incremental pro-life laws. For instance, since the mid-1990s, Gallup has asked the “pro-life or pro-choice” question over 35 times but has asked about asked about waiting periods before abortions only three times and about parental-involvement laws four times.
Commentators and activists often read way too much into the results of individual surveys. When analyzing public-opinion data, a good rule of thumb is to consider the same question asked by the same survey research firm over an extended period of time. For instance, shortly after the 2012 election, Rasmussen released a poll showing that only 37 percent of Americans described themselves as “pro-life.” The media seemed pleased, and many pro-lifers were dismayed. Right-to-lifers had become used to seeing multiple Gallup polls showing pro-life sentiment approaching 50 percent. Had the inopportune comments by Richard Mourdock and Todd Akin coupled with the Democratic party’s “War on Women” rhetoric reduced pro-life sentiment by over ten percentage points?
In actuality there was no significant decline in pro-life sentiment. Polls conducted by Rasmussen — which, unlike Gallup, contacts only likely voters — consistently show lower pro-life sentiment. The postelection Rasmussen poll showed a decline in pro-life sentiment by about three percentage points, near the margin of error. Of course, this received scant coverage from the mainstream media.
— Michael J. New is an assistant professor of political science at the University of Michigan–Dearborn. Follow him on Twitter @Michael_J_New.