The Mystery Of Internet Survey IQs

Scott investigates why internet surveys (LessWrong, ClearerThinking) report implausibly high IQs (138, 130). Working with Spencer Greenberg's data, he isolates three problems: the most-used SAT->IQ conversion site overestimates; only the smartest people report/remember their SATs (selection); and self-reported IQ-test scores are inflated while official IQ tests are 'gobbledygook over 135.' Concludes the real figures are ~128 (LessWrong) and ~111 (ClearerThinking).

Quality 70 · Strong. Strong: a careful, numerate psychometrics detective story that genuinely resolves a real puzzle with sound methodology (the three problems, the demographic norming); kept mid-Strong as a self-contained empirical resolution rather than a broad framework.

Claude’s paradigm shift 42 · Moderate. Moderate-low: the findings (IQ tests break down over 135, SAT selection) are a useful synthesis, not a novel framework.

Real-world impact 2 · Minor. Minor/within-discourse: informs how rationalist/ACX surveys interpret self-reported IQ; a niche methodological contribution.