Social theorists, above all Duke University’s Timur Kuran, have drawn attention to the phenomenon of “preference falsification.” The basic idea is that when people speak in public, they aren’t always truthful about their preferences. What they say is different from what they really think.

In unfree societies, people may be too frightened to disclose their actual views in opinion surveys. But preference falsification can also afflict democracies, if social pressures lead people to misdescribe their real views and behavior.

Recent research uncovers strong evidence of preference falsification in the U.S. When people are assured of anonymity, it turns out, a lot more of them will acknowledge that they have had same-sex experiences and that they don’t entirely identify as heterosexual. But it also turns out that when people are assured of anonymity, they will show significantly higher rates of anti-gay sentiment.

These results suggest that recent surveys have been understating, at least to some degree, two different things: the current level of same-sex activity and the current level of opposition to gay rights.

The research, conducted by Ohio State University economist Katherine B. Coffman and her colleagues, involved 2,516 participants, all from the U.S. About half of the participants were randomly assigned to take a standard survey, employing the “best practices” in widespread use today.

Survey Findings

In this survey, people were asked to respond to several innocuous questions, not involving sensitive issues, and then to answer questions about sexual orientation, designed to elicit both their views and their reports about their own behavior. This approach gives apparently credible assurances of anonymity to those surveyed, but it remains possible, in practice, for the experimenters to link particular answers to particular questions.

The other participants were assigned to what Coffman and her colleagues call a “veiled report” treatment. The details are a bit technical, but the basic point is to design the survey so that the experimenters can’t learn, and can’t even make inferences about, any individual’s answers to particular questions. They can calculate answers only at the aggregate level.

The two approaches produced significantly different results. In the best practices survey, 17 percent of participants said they had had a sexual experience with someone of the same sex (12 percent of men, 24 percent of women). For the veiled report, the corresponding number was 27 percent (17 percent of men and 43 percent of women) -- an increase of 58 percent.

In the best practices survey, 11 percent of the population said that they didn’t consider themselves to be heterosexual. In the veiled report, that number jumped to almost 19 percent -- an increase of 65 percent.

Did participants believe that discrimination on the basis of sexual orientation should be illegal? In the standard survey, only about 14 percent said no. That number increased to 25 percent in the veiled report.

In best practices, only 16 percent of participants said they would be uncomfortable having a manager at work who was lesbian, gay, bisexual or transgender (LGBT for short). The number jumped to 27 percent in the veiled report.

The effect of assuring anonymity varied significantly across demographic groups. The veiled survey had no effect on the answers of young people to questions about their sexual orientation, apparently because social norms don’t much discourage young people from revealing the truth.

Veiled Approach

But among Christians and older people, the effect of the veiled approach was especially large, increasing their reports of non-heterosexuality and of same-sex experiences by more than 100 percent.

In best practices, only a minority of Republicans (35 percent) said they would be unhappy with an LGBT manager. Under the veiled report, most Republicans (67 percent) said they would be unhappy.

It is important to emphasize that Coffman and her colleagues didn’t have a representative sample, so the total percentages can’t be taken as reflective of what the general American population thinks and does. Among other things, the participants in their study were younger, more liberal and better educated than the general U.S. population.

But the researchers’ real interest was the effect of assuring anonymity, and on that question the absence of a representative sample doesn’t undermine their conclusions. On the contrary, the impact of assured anonymity on the answers would almost certainly be even bigger with the American population as a whole, because the demographic groups that show the largest effects from the veiled report were underrepresented in their survey.

In recent years, the U.S. has experienced rapid shifts in popular attitudes toward same-sex relationships. Americans increasingly disapprove of discrimination against gay men and lesbians. That disapproval is likely to grow over time.

But social norms continue to matter. We have good reason to believe that there is more same-sex activity, and also more homophobia, than current surveys suggest.

(Cass R. Sunstein, the Robert Walmsley University professor at Harvard Law School, is a Bloomberg View columnist. He is a former administrator of the White House Office of Information and Regulatory Affairs, the co-author of “Nudge” and author of “Simpler: The Future of Government.”)

To contact the writer of this article: Cass R. Sunstein at csunstei@law.harvard.edu.

To contact the editor responsible for this article: Katy Roberts at kroberts29@bloomberg.net.