Lizardman's Constant is 4% – Estimating the Troll Contingent on Surveys

> I really wish polls like these would include a control question, something utterly implausible... Whatever percent of people answer yes to the hippo question get subtracted out from the other questions.

This is standard in appropriate cases. The official survey of young people's drug taking in England says:

"The survey asks about awareness and usage of a fictional drug called Semeron. Responses to these questions provide a guide as to how much pupils are over-estimating so it is reassuring to see that in the 2018 survey only 24 pupils (0.2% of all pupils)admitted to taking Semeron which is in line with previous surveys and suggests that only a tiny minority of pupils maybe falsely admitting to drugs use. However, 1,717 pupils (13%) said they had heard of Semeron which suggests awareness of drugs could be over-estimated. Hence the majority of the tables in this report focus on drug usage rather than drug awareness."

https://files.digital.nhs.uk/B8/8B8160/sdd-eng-2018-qual.pdf...