Barnard Statistics Concepts Practice Test 2026 - Free Statistics Practice Questions and Study Guide

Fisher's exact test is designed for situations with sparse data where sample sizes are small and expected counts in the cells are below about 5. In this regime, the usual chi-square approximation isn’t reliable, so Fisher’s test provides an exact p-value by calculating the probability of the observed table (and more extreme ones) under the null hypothesis, using the hypergeometric distribution with fixed margins. This exactness is why it’s preferred when data are scarce. When the sample is large and all expected counts exceed 5, the chi-square test works well and is more straightforward. Normal distribution of the data isn’t a requirement here, since Fisher’s test is about categorical counts, not continuous data. And having more than two categories isn’t the defining factor; Fisher’s test is especially common for small-sample 2x2 tables, though it can be extended to larger tables.