ABSTRACT
T-tests answer a very simple question-are two scores the same or different? In spite of the apparent simplicity of this test, it is widely useful in a variety of settings. All of the statistical tests investigated in this chapter involve looking at two mean scores on the same testing instrument. Owing to random chance, scores from two groups on the same test are unlikely to be exactly the same, even if the two scores come from the same set of people. A t-test can tell us if the differences between the two scores are big enough that we can assume they come from two different groups, or that scores at Time 1 differ from scores at Time 2. If the differences are not big enough we will assume that the scores come from the same group or that there are no differences between Time 1 and Time 2 scores. This chapter will look at two main types of t-tests which are very frequently used in second language research studies:
1 the independent-samples t-test 2 the paired-samples (or matched-samples) t-test
This section will help you understand how to choose which t-test to use. The independent-samples t-test is used when you perform an experiment and you have obtained mean scores from two independent groups. You have measured both scores yourself, and you only have two of them (if you had three groups, you could not use a t-test). For example, you might assign students randomly to one of two treatments, and then at the end of the experiment you would test them and have two mean scores. You want to know whether one group performed better or worse than the other group.
