Summary

The t-test and the t-distribution that the test is based on, are more useful when sample sizes are very small. Although the t-test is one of the simpler, more well-known tests, it was probably more often needed in the early twentieth century than in this age of larger (sometimes huge) datasets. Nevertheless, we may sometimes find ourselves in need of t. R has dedicated functions for doing t-tests, but it’s easy to fit linear models that produce one-sample, two-sample, and paired versions of the t-test in the summary() table output.