A Program for Computing Fisher's Exact Probability Test and the Coefficient of Association λ for n × m Contingency Tables

Fisher's exact probability test is a nonparametric statistical technique for determining the significance of a difference for independent groups with discrete data. Whenever computing the exact test for tables larger than 2 × 2, two problems arise (Hays, 1973): (1) the computations become very tedious and (2) there is a need to measure the strength of statistical association in each sample. The purpose of this paper was to provide a mathematical algorithm for Fisher's test (1) which is adaptable to n × m contingency tables and (2) which provides the user with lambda (λ), an index of predictive association designed for crosstabulations of nominal-level variables.