Exact Inference for Categorical Data
摘要
In practice statistical inference for categorical data is based mainly on large sample approximations of test statistics. This asymptotic theory is valid only if the sample sizes are reasonably large and well balanced across populations.