Those who believe in the general accuracy of pre-election polls will want to read this article Gerald Wasserman of Purdue University’s Department of Psychological Sciences, written in 1997. (Hat tip: Polipundit)
Wasseman notes that in 1996, "Almost all of the major pre-election presidential polls missed the final result by being off in the same direction, differing only in the degree of their prediction failures." He argues that chance alone would produce such a result only once in 4,900 elections.
