A discussion of the assumptions when using the t procedure to construct a confidence interval for the population mean mu. The assumptions are discussed, and the effect of different violations of the normality assumption is investigated through

simulation.