Public opinion polls likely understate opposition to same-sex marriage, a new study has found.
A "social desirability bias" may come into play with many people answering polling questions in support for same-sex marriage, skewing the results of pre-election studies in favor of the issue, says political scientist Richard Powell, who authored the study set to be released in the November issue of the journal
American Politics Research.
His research found that such surveys underestimate public opposition to same-sex marriage by 5 to 7 percentage points as respondents offer views to pollsters that seem socially acceptable at the time, rather than their honest feelings about issues that are sensitive, according to the
Pew Research Center, in reporting on the study.
A Gallup poll, the first since the U.S. Supreme Court struck down the Defense of Marriage Act and released in July, found that 52 percent of Americans in 50 states backed a law supporting gay marriage, while 43 percent said they would vote against such a measure.
If Powell's research is correct, it means a slight majority is against same-sex unions.
Powell, a University of Maine professor, looked at 33 statewide races with same-sex marriage ballot initiatives that took place from 1998 to 2012. In each race, pre-election polling was conducted, and he looked at only polls that took place within 90 days of elections.
Controlling for several factors that could be related to social desirability bias, he then compared the pre-vote polling on same-sex laws with the votes in the election, finding the underestimation, which has not been taken into account in the politically divisive debate over gay marriage.
As a concept, such "social desirability bias" is difficult to detect for researchers.
Pew cited the "Bradley Effect" in the 1982 California governor's race, in which Los Angeles Mayor Tom Bradley polled to win the election but was beaten narrowly by challenger George Deukmejian.
Bradley, who died in 1998, was black, while Deukmejian is white, leading some to ponder whether prospective voters gave a pollster an answer they deemed socially acceptable before the election, but later voted their true opinion at the ballot box.
Harvard researcher Daniel Hopkins later nailed down the existence of the social bias, Pew noted. He studied Senate races featuring black and white candidates from 1989 to 2006.
Black candidates did better in opinion polling than they did in the final vote counts through 1996.
That phenomenon remains unknown.
Some pondered in the 2008 and 2012 presidential election about whether then-candidate
Barack Obama would suffer the Bradley Effect, but in his case it did not materialize.
Related Stories:
© 2025 Newsmax. All rights reserved.