The Bayesian Analysis of Contingency Tables

摘要: Summary. This paper describes how data from a multinomial distribution, and in particular the form of contingency table, may be studied by using prior distribution parameters expressing results posterior or some aspects thereof, parameters. The analysis used must depend on described here only applies to certain type knowledge but, for reasons given below, it is believed that this frequent occurrence. binomial situation first considered obtained there suggest general result which then established. A few remarks Bayesian enable applied, problems then, with aid another result, tables. method has close connections Analysis Variance these are examined, particularly view simplifying tables involving three more factors. 1. Binomial distributions. Although will appear as special case established situation, instructive begin suggested generalizations. Let N independent trials constant probability 0 success n successes (N - n) failures. likelihood