binomial test

Quick Reference

In inferential statistics, a simple test of whether a sample of dichotomous scores, labelled success/failure (or male/female, remember/don't remember, and so on) comes from a binomial probability distribution. The binomial distribution represents the probability, given a sequence of Bernoulli trials with a fixed probability of success on each trial, of achieving a number of successes as extreme as the number observed in a specified number of trials. For example, a string of guesses (hits/misses) with a pack of 25 ESP or Zener cards, which contains five cards with each of five different symbols, entails a 1/5 probability of a hit on each trial. Application of the binomial test shows that, for a statistically significant result (conventionally, a probability of less than p < .05), leading to a rejection of the null hypothesis that the scores come from a binomial distribution, more than eight (or fewer than two) hits in 25 guesses would be required. [From Latin bis twice, bini two by two + nomen a name]

Subjects: Psychology.

Reference entries