Two-Sample Tests for Counts in Two Categories
Where data are arranged simply as counts in two categories, there are still some very useful inferential statistical tests available. They are commonly used in medical research, for example, in assessing the outcome of particular treatments. The important distinction in this case is between paired and independent data. Data are paired if an entry in one category is linked to a particular entry in the other category. A common reason for samples being paired is that the same person is being asked a question twice, before and after some event, samples have been collected in pairs from the same place, or at the same time, or the same sample has been measured twice using different methods.