Or worse, people on Mechanical Turk. Bad sample populations are a huge problem. There's only one really good study on sexual behavior in in the US, called "Sex

Or worse, people on Mechanical Turk. Bad sample populations are a huge problem. There's only one really good study on sexual behavior in in the US, called "Sex in America, a definitive survey".[1] It was expensive, but they did it right. They used a random process to select a large number of regions across the US, then random individuals within those regions. There was a mailed survey, followed up by a phone survey, followed up by visits by interviewers, followed up if necessary by paying people to do the survey. The result was 90%+ participation. That's how you overcome selection bias. The main result is that, for the overall population, sex is rather mundane. [1] https://archive.org/details/sexinamericadefi00mich/mode/2up https://archive.org/details/sexinamericadefi00mich/mode/2up