A Choosing a statistical test: the six questions in Section 11.4 of the book

Choosing the appropriate statistical test is obviously key to the whole process of finding out exactly what your data have to tell you. Of course this task can seem quite bewildering, especially when you are new to the job, but even later on you can find yourself uncertain as to exactly which of the various available tests you should choose. Indeed, sometimes there are several statistical tests you could choose, each of which is potentially appropriate for your data. At the end of Chapter 11 of Designing and Reporting Experiments you will find references to helpful coverage about choosing statistical tests in the statistics textbooks paired with it: Greene and D’Oliveira’s Learning to Use Statistical Tests in Psychology 3/e and Pallant’s SPSS Survival Manual 3/e . The material in section A of this website is also designed to help you with this task.

Choosing an appropriate statistical test can be a difficult task, but it need not be as daunting a process as some of you think. For those of you who find this task a challenge, the important thing is not to panic: if you keep a clear head and take things step by step, most of the time you should have few problems in arriving at an appropriate test for your data.

In order to decide which test to use, you need to be able to answer a number of questions about the type of study you conducted, the nature of the data you obtained, and the precise questions you wish to ask of these data. Below are the key questions you should consider. Consider these in turn:

1. Do you want to (a) compare conditions or (b) correlate variables?
2. Which type of data do you have? Are they (a) nominal (b) ordinal or (c) at least interval?
3. If at least interval, can you assume that the data come from a population of scores that is normally distributed?
4. If you wish to compare conditions, did you manipulate more than one IV in your study?
5. If you wish to compare conditions, how many do you wish to compare at any one time?
6. Do you have the same or different participants in these conditions?

I attempt in this part of the website to explain the main concepts in the questions above. If you do not understand them after reading this section, you will find fuller discussions in all decent textbooks of statistics, including the textbooks paired with the book, Greene and D’Oliveira’s Learning to Use Statistical Tests in Psychology 3/e and Pallant’s SPSS Survival Manual 3/e.