4.1 Introduction

Chapter 3 looked at how to fit a General Linear Model and understand the parameter estimates. Estimating parameters and obtaining the precision of those estimates is the most fundamental thing you do in inferential statistics, which is why it was the focus. However, people often want their statistical model to provide a second, related thing: a statistical test. A statistical test is a procedure for deciding, given the evidence of the data, whether to accept or reject some hypothesis about the population parameter of interest. They arise so often in inferential statistics because science is often hypothesis-driven; we did the study to see if the predictions from some hypothesis were supported or not, and so want the test in order to know whether they are.

There are two main types of tests: tests of difference (do the data suggest that some parameter in the model is different from some value of interest, usually but not always 0); and tests of equivalence (do the data suggest that some parameter in the model is the same as some value of interest, again, usually but not always 0). This chapter covers both kinds of test. It uses the same behavioural inhibition dataset as chapter 3, and the same models.