TABLE OF CONTENTS The first sets of test we will look at are known as tests of difference. In these tests you want to know if there is any significant difference between people's score on the independent variable (if you can't remember what the independent variable is, have a quick look back at Chapter 3). The first two tests to be discussed in this section are the related t-test and the unrelated t-test. These are parametric tests. The second two tests that are discussed, the Mann-Whitney U and the Wilcoxon, are non-parametric tests.
This test is commonly used when you want to compare the differences between two groups of numbers that have come from a repeated measures design or matched-pairs design. For example, imagine you have carried out a study comparing people's performance when using two versions of a mobile cinema listing service: speech input and keypad input. In order to assess user performance you timed how long it took each participant to complete a specific task (e.g. finding the cinema opening times). In the example we are using here, our data has come from a repeated measures design as our participants used both versions of the mobile phone service. If you are unsure about these terms go back to Chapter 3 and re-fresh your memory. It may actually be worth your while to book mark that chapter while you go through this chapter, just in case.
Let's go through how to carry out...
