Hasty Generalization

In plain terms

A hasty generalization leaps from a few cases to a sweeping rule. You meet two rude people from a city and conclude the city is rude. A product fails once and you decide the brand is junk. The sample is too small, too skewed, or both, to carry the weight of the conclusion built on top of it.

The leap feels natural. Humans generalize from limited experience constantly; it's a useful instinct. It just isn't a reliable argument.

Why it's fallacious

A generalization is only as good as the sample behind it. To support a claim about a whole group, the evidence needs to be large enough to rule out chance and representative enough to reflect the group rather than a quirky corner of it. A handful of cases, or a self-selected set, fails both tests. The conclusion outruns what the data can justify.

This is the engine behind most stereotypes: a few salient examples, generalized into a rule, then applied to everyone in the category.

Canonical example

"Both times I visited that hospital the wait was over three hours. That hospital is a disaster."

Two visits is a sample of two, possibly on busy days, possibly at busy hours, by one person who may notice waits more than other things. It might be a genuinely troubled hospital. It might be that the visitor hit two bad afternoons. Two data points can't distinguish those, yet the conclusion ("a disaster") is a confident, blanket judgment about the whole institution.

Counter-example (not a fallacy)

"Across 4,000 randomly sampled patient visits over two years, median wait time was 3.2 hours, well above the regional benchmark. The hospital has a wait-time problem."

This isn't a hasty generalization. The sample is large and randomly drawn, so it's plausibly representative, and the conclusion is scaled to what the data shows. Generalizing is fine when the sample can bear it. The fallacy is in the haste: concluding before the evidence is enough.

The line: is the sample big enough and representative enough to support a claim this broad?

How to fix it

If you've been linked here, check the size and shape of your sample against the size of your claim. The fix is usually to shrink the claim to fit the evidence: "the two times I went, the wait was long" is fully supported and honest, where "that hospital is a disaster" is not. If you want the bigger conclusion, get the bigger sample. A modest claim you can defend beats a sweeping one you can't.

If you're on the receiving end, ask about the sample: "How many cases is that based on? Were they typical?" Most hasty generalizations shrink back to reasonable size the moment the sample behind them is examined.