|
As I say in the book, knowing what to report can be bewildering. Even when you have chosen the correct analysis and run it properly it may not be easy to detect the bits you need to extract from the output and report in the RESULTS. This is especially so when you are a novice, but it can be mystifying even when you are more experienced.
The main thing, especially early on in your career, is to demonstrate to your marker that you realize what needs to be reported. The basic rule when reporting inferential statistics is that you need to provide sufficient information to enable someone else to check whether the outcome of a test is or is not statistically significant. In most cases, the minimum you need to provide is:
 the obtained value of the statistic,
the information needed to look this up in tables of critical values,
(preferably) the exact probability associated with this obtained value,
if you are testing for statistical significance, you need also state whether or not this outcome is statistically significant.
You can find out more about tables of critical values in Appendix 3 of the book. You can also find there what to do if you cannot provide the exact probability, such as when your statistical print out says p = .000.
The examples here and in the book give you illustrations of how to provide such information for the statistics that you are most likely to use as students of psychology. The illustrations here also include some of the things to watch out for when reporting these statistics. Before using this material you should be familiar with the issues involved in choosing tests (see Section A of this Web site).
If a specific test you have used is not mentioned here, it should nevertheless be possible to find a reasonable equivalent on which to model your reporting of that statistic. With experiments, inferential statistics are usually tests of difference. However, you may also have used measures of association and need to report these. The material here therefore contains advice on how to report these as well. In preparing these examples I have assumed that you will be testing for statistical significance. (See Chapter 12 of the book for a discussion of alternatives to significance testing.)
For more on the distinction between nonparametric and parametric tests, see Section A2 of this Web site.
Here and there I make reference to the two statistics textbooks that are paired with the latest edition of Designing and Reporting Experiments in Psychology : Greene and D'Oliveira's Learning to Use Statistical Tests in Psychology 3/e is the more basic and introductory and is suitable for students at the beginning of their careers as undergraduate psychologists. Pallant's SPSS Survival Manual 3/e is somewhat more advanced and appropriate for more experienced undergraduate students. Both are best sellers in their own right. |
