|
When there are more than two levels on your IV or you have a nonexperimental variable with more than two categories, strictly you need to run and report further analyses to locate any statistically significant differences. Statistically significant outcomes to your Kruskal-Wallis, or your Friedman test, or a statistically significant main effect or interaction in your ANOVA, mean that you have evidence that the values differ sufficiently from each other for you to be prepared to reject the null hypothesis. However, this could happen for a variety of reasons. For example, with a three level IV it could be because only one of the conditions differs from the other two (e.g., condition A differs from B and C), or because all three differ from each other.
If you have been taught how to conduct such analyses, you need to report them here. Discussion of how to do this can be found in Sections B3.3, B3.4 and B7.1 of this Web site.
|
