This is a representative subsample of my n=970,000 Big Kink Survey. This survey is a very large, comprehensive survey about sexual fetishes that went viral online. The original demographics were heavily skewed for young, female, liberal, and non-cis. The sample I'm providing here is a small, more representative subsample. I did, however, cut the age off at 32, as I had significantly fewer older responders. I also only included people from the US, Canada, and Europe. Balancing was done based on more total demographic information than is included in the provided set; I removed a bunch of the demographic columns after balancing for anonymity reasons. This dataset has been been anonymized using a few rounds of similar-row demographic swapping and noise across the entire dataset. In general, correlations are about 25% less strong than they are in the original subsample, pre-noise. Base rates of most things are less affected. Here's a folder with a bit more info, including the back end of the original survey, including the full questions and the way they were labeled internally. Original cleaning was minimal, and dropped: people who reported answering the survey dishonestly, people who finished the survey way too fast, people who were extreme outliers (e.g. marking every answer as yes), people who gave inconsistent responses (e.g. reporting being both very into and not at all into a fetish at different points in the survey), and all the people who said their age was 69. I also did some structural cleaning, because the original file got kinda messed up from the survey software. Mostly this was columns got duplicated, and half the answers went into the first half, and half into the other. I think the questions about dirty talk and cunnilingus may be reversed, make sure you double-check those.