Insensitivity to sample size

Estimation biases biases

Insensitivity to sample size is when you treat a small number of cases as if they were a reliable guide to the whole. Small samples vary a lot by chance, but we under-expect that variation and draw strong conclusions anyway. It is largely the same idea as the fallacy of hasty generalization: one is the mental tendency, the other is the argument that generalizes from too few cases. Related: Hasty generalization.

Examples

  • Two of your friends got sick after eating at the new café, so you conclude the café makes people sick—without considering that two people is too small a sample to tell.

  • Your new exercise routine seems to work after one week, so you decide it is clearly effective, ignoring that short-term results can be random.

  • A politician wins two by-elections and pundits say they are on course to win the next general election, despite the tiny sample of races.

← Back to Biases