Tendency to Overgeneralize from Small Samples

The persistent cognitive error of drawing broad conclusions from insufficient data—treating a few observations as representative of the whole population. One bad restaurant meal means 'that restaurant is terrible.' Two unsuccessful cold calls mean 'cold calling doesn't work.' Three data points on a chart constitute a 'trend.' The mechanism: the brain's pattern-recognition system doesn't weight by sample size—it finds patterns equally in 3 data points and 3,000. This produces premature certainty: people form confident beliefs from samples too small to support any conclusion. Tversky and Kahneman called this 'the law of small numbers'—the erroneous belief that small samples are as representative as large ones.

When to use it

When conclusions are being drawn from a handful of observations; when 'trends' are based on 2-3 data points; when personal experience (inherently small-sample) is overriding base rate data; when the confidence in a conclusion exceeds what the sample size warrants.

How it can help

Before drawing conclusions: check the sample size against the variability of the phenomenon. High-variability domains (human behavior, market returns, creative success) require LARGE samples before patterns are meaningful. The practice: ask 'how many observations would I need to be confident this pattern is real rather than noise?' For most business decisions: more than you have. The correction: hold conclusions lightly when based on small samples, and actively seek more data before committing to a pattern. The rule of thumb: any pattern based on fewer than 30 observations should be treated as a hypothesis, not a conclusion.