I've seen some numbers from the US and some from Japan, but there's an incredibly high level of unreliability in surveys on kink because people both lie and do not respond at higher rates with them, and it's hard to remove the selection biases such studies produce.