Hasty Generalization
From SkepticWiki
Contents |
[edit] Definition
Hasty Generalization (sometimes known as "false generalization," "leaping to conclusions," or "the fallacy of insufficient sample size") is an informal logical fallacy where a participant observes a fact in a small group and applies it to a much larger group, where the observational group was too small or was otherwise not representative. In its general form, it often runs something like this
- By observation of group G, I know that P is true of that group
- G is a subset of H
- Therefore, P is true of H as well
A simple variation on this is
- By observation of group G, I know that P is true of that group
- Therefore, P is true
This is a fallacy because G may not be large enough (or varied enough) to reveal the counterexamples in the full population.
[edit] Examples
Example 1:
- Antagonist: I was reading that women Ph.D.'s often don't make very much money. But all the women who taught at my college make a fortune?
- Protagonist: So?
- Antagonist: Obviously, then, women Ph.D.s make a fortune and the article was wrong.
Example 2:
- Antagonist: When my uncle had that nasty disease, he took cider vinegar mixed with egg white for it, and got better. So why waste money on expensive penicillin tablets?
Example 3:
- Antagonist: Astrology doesn't work, and I can prove it!
- Protagonist: How?
- Antagonist: Last Tuesday, I looked up my horoscope in the paper, and it didn't come true. Obviously astrology doesn't work.
In the first example, Antagonist's college might be a particularly wealthy and well-paying college (and not all women Ph.D.'s teach at college in the first place). In the second example, people sometimes get better from diseases by themselves, but drugs often work faster and more reliably. In the third example, the local astrologer may be incompetent, while the astrologer in the next town over might be brilliant. In any case, Antagonist's observations don't justify the general conclusion drawn.
[edit] Exceptions to the Rule
No rule can be drawn to describe when a sample is absolutely beyond a doubt sufficiently large and representative unless you have sampled and tested the entire population. The term "Hasty Generalization" is usually reserved for situations where the sample is demonstrably inadequate in some way, either by bias, unrepresentativeness, or simply being too small.
However, it is unfair to raise issues of "Hasty Generalization" when the number of samples available for study is itself too small, but no other data can reasonably be made available. For example, biologist (and particularly paleographers) often have to speculate about processes based on a single specimen of an animal. It is possible, for example, that the "average" height of a T. rex is twice what we currently believe it to be, but that all fossils found were from particularly short individuals. However, there is no evidence to support this speculation and studying all available evidence suggests that we have the correct height. To suggest otherwise, and to take the suggestion seriously, is to commit the argument from ignorance.
