|
When you randomize variables you trust to luck. Randomization does not guarantee the absence of systematic differences between conditions on these extraneous variables. (For more on randomization see Sections 9.1.5, 10.4 and 10.6 of the book. It is also discussed in Appendix 2 of the book.) Whenever possible, therefore, measure these extraneous variables and test to see whether they differ between conditions. (Of course, you need to bear in mind here the problem of increasing your type I error rate by increasing the number of separate analyses that you perform: For more on type I error, see Section 11.3 of the book.) Once you are very advanced you may be able to use one of the statistical techniques that enable you to assess the effects of your IVs on your DVs while taking into account the relationships between the IVs, DVs and these other variables (e.g., analysis of covariance – see Section B7.5 of this Web site).
|
