Biased sample

A biased sample falsely claims to be typical of the whole group. Someone saying "Everyone liked that movie!" might not mention that the "everyone" was them and three of their friends, or a group of the star's fans.

Online and call-in polls are particularly at risk of this error, because the respondents are self-selected. At best, this means you get the people who care most about an issue; at worst, people listening to a particular radio host, or on a political mailing list, flood the poll.

Biased samples aren't always an attempt to mislead: in 1936, in the early days of opinion polling, the American Literary Digest magazine called two million random telephone numbers, questioned the people who answered, and predicted the election result. They got it wrong because, at the time, telephones were far from universal, and telephone owners weren't a good sample of the electorate as a whole. In contrast, a poll of only 50,000 citizens selected by George Gallup's organisation successfully predicted the result, leading to the popularity of the Gallup poll[?].

See also : Logical fallacy