Business Statistics Made Easy in SAS

Numbers of Categories to Compare: Two vs. More than Two

There are two different sets of techniques depending on whether you are comparing only two categories to one another (e.g. males to females) or more than two (e.g. New York to Chicago to Miami).

• Comparisons of two categories: For comparisons of two categories only, we usually compare using “t-tests.” I discuss this approach in the next major section.
• Comparisons of more than two categories: We call these techniques “Analysis of Variance.” You should not use t-tests for this scenario. Some people wonder why you cannot just compare all possible pairs of groups via t-tests (e.g. New York to Miami, Miami to Chicago, etc.) but you cannot, for reasons which I mention later.