Physical Sciences and Mathematics Commons^™

Determining Sample Size To Bound The Probability Of Classifying A Sample Into The Wrong One Of Two Multinomially Distributed Populations, C. Philip Cox

Conference on Applied Statistics in Agriculture

The problem considered is that of choosing between the two specifications . . . of known multinomial probabilities on the basis of sample values x _j, the observed counts in the j = 1,... ,k, classes, with . . . The particular question examined is 'how large should N be to achieve reliable differentiation?'. It is shown how to find N such that the probability of misc1assification does not exceed a prescribable value. The method is exemplified in a genetic context.