Scott Alexander investigates why self-reported IQ scores in online surveys are implausibly high, identifying issues with conversion tools, reporting bias, and unreliable tests.
Longer summary
Scott Alexander investigates why self-reported IQ scores in online surveys seem implausibly high. He analyzes data from two surveys (Less Wrong 2014 and Clearer Thinking 2023) and identifies three main issues: 1) A popular SAT to IQ conversion tool was inaccurate, inflating scores. 2) Only the smartest people tend to report their SAT scores, skewing the sample. 3) Self-reported IQ test scores are often inflated or based on unreliable tests. After accounting for these factors, he estimates the average IQ of the Clearer Thinking sample at 111 and the Less Wrong sample at 128, which he considers more plausible.
Shorter summary