Katie Coffman and Lucas Coffman, assistant professors, economics, and Boston University School of Management Professor Keith Marzilli Ericson, are authors of a new study showing that anti-gay sentiment is higher than we think, and current methods for assessing attitudes about sexuality are not as accurate as they should be.

In a study published in the National Bureau of Economic Research, the team of researchers found that current methods may not accurately capture the size of the lesbian, gay, bisexual and transgender (LGBT) population nor attitudes toward them.

“Overall, it’s hard to measure sexual orientation and opinions about sexual orientation because of persistent biases toward more socially acceptable responses,” said Lucas Coffman. “We have found this trend even in computer-generated surveys where responses are anonymous.”

The researchers used a survey technique known as the “veiled elicitation method” to correct for social desirability bias—the tendency of survey respondents to give researchers the answers they think are expected. The veiled method involves asking subjects to respond to a group of questions, and in an indirect manner, which has been shown to reduce the chances that an answer will be biased toward social expectations. The theory is that grouping sensitive and non-sensitive questions together can “veil” how subjects answer the sensitive question, thus reducing the influence of social desirability bias.

Among the 2,516 U.S. volunteers who were randomly assigned to answer questions about their sexuality using one or the other survey method, those taking the more veiled survey were 65% more likely to report a non-heterosexual identity themselves and 59% more likely to report having a same-sex sexual experience than those using the standard survey technique.

The veiled method also revealed more people with anti-gay sentiment than among those taking the other type of survey. The participants were 67% more likely to disapprove of an openly gay manager at work and 71% more likely to admit that it was acceptable to discriminate against people who are lesbians, gay or bisexual.

“Comparing the two methods shows sexuality-related questions receive biased responses even under current best practices, and, for many questions, the bias is substantial,” Katie Coffman said.

Did the enhanced anonymity lead to more truthful responses? The researchers aren’t sure, but previous studies revealed that when responses are blinded, people do tend to express attitudes and opinions that are more raw and closer to their true beliefs.

Researchers cautioned that their results shouldn't be used to extrapolate how many Americans are gay, lesbian or bisexual, since the survey was not nationally representative. (Young adults, for instance, were overrepresented).

But the results raise questions about how surveys have tallied the gay community. And the fact that people shy away from revealing their feelings about gay people, researchers added, “suggests that many other opinions on controversial public issues may not be accurately measured.”