Thinking, Fast and Slow · chapter 10 · id thinking-fast-and-slow-c10-01-even-trained-researchers-have-
“Even trained researchers have poor intuitions about sampling variability and expect small samples to be representative of the population ('belief in the law of small numbers').”
holdsconfidence: high✓ ZHP-verified + red-teamed
Receipts
Tversky & Kahneman (1971), Psychological Bulletin 76(2)DOI registry: valid
Even attendees at mathematical psychology meetings underestimated sample sizes needed for adequate statistical power and overestimated the replicability of results from small samples.
Button et al. (2013), Nature Reviews Neuroscience 14DOI registry: valid
The median statistical power of studies in neuroscience is approximately 21%. This confirms that researchers systematically underestimate required sample sizes, exactly as Tversky and Kahneman predicted.
This claim is a stable, citable object. If you can falsify a verdict, tell us — corrections are loud here.