ABSTRACT

Contrasts are comparisons between means in the case of parametric tests. To discuss them I am going to return to the word-recall experiment, described in Chapter 16, which involves a group using pegwords, a group using the method of loci and a control group. The mean recall of participants in the three conditions were found to be significantly different. The types of contrast you can perform are almost infinite in number. The main distinction is between pairwise contrasts, where you compare two means at a time, and other forms of contrast, such as a comparison of the control condition with the mean of the two other conditions. There are numerous tests of contrasts. Rather than describe all of them I will give a set of tests that cover most situations and an explanation of when each is appropriate.