Group-by-Time Interactions

Abstract

Whether we deal with different plots of land that receive different treatments in agricultural experiments or we have groups of human participants who are given different treatments in an educational or psychological experiment, when we measure these plots of land or groups of participants at two or more occasions on the same outcome variable of interest we speak of a split-plot design and we can use split-plot analytic methods to analyse differences between plots of land or groups of participants and how these differences increase or decrease from occasion to occasion. In some cases, there are two measurements (i.e., two occasions) one of which takes place immediately after treatment (i.e., post-test, sometimes also referred to as immediate post-test) and the second sometime after (i.e., follow-up). In other cases, one of the measurements takes place before the treatment (i.e., pre-test) and the other takes place after treatment (i.e., post-test). The latter type is also known as pre-test post-test control-group design. In yet other cases, there are three or more measurements; these measurements may either all take place after treatment, or one or some (though not all) of them take place before the treatment. In this chapter, each of these types of situations is discussed with an example experiment.