Are people honest with pollsters? According to a new study by the National Bureau of Economic Research, the answer is no, and it cuts both ways. Polls underestimate the number of LGBT people in the U.S., and they underestimate the amount of homophobia as well.
The researchers, economists from Ohio State University and Boston University, used two ways of asking questions related to gay issues. In the first, the respondents answered questions directly, as pollsters usually asked. In the second, the respondents were essentially anonymous, so their responses couldn’t be connected to them personally.
The results showed that people won’t tell pollsters things that are too personal or that they think might put them in a negative light. That includes revealing sexual orientation and owning up to homophobic tendencies.
For example, when directly asked about their orientation, 11%, of the respondents said that they were “not heterosexual.” But given the chance to respond anonymously, that number jumped up to 19%.
Similarly, 15% if the respondents direclty asked said that would be unhappy to have a gay manager at work. But responding anonymously, that number jumped up to 25%.
The researchers are the first to admit that they didn’t exactly come up with a representative cross-section of the country among the 2,500 people they recruited for the surveys. The respondents tended to be younger, less religious, and less likely to be Republican. But the results are a good warning: next time you see a survey results, you may want to take them with a grain of salt. People may just be saying what they want you to hear–or what they don’t.