I need to compare three groups of five subjects each (between-subjects design) where a variable is measured between time 0 and time 1. The interest of the study is on relative or raw difference.

I wonder whether I should use repeated measures ANOVA or one-way ANOVA. My guess is that I am removing any within subject variation because my research interest is in the difference between time 0 and time 1 and not in absolute value. So a one-way approach would be right. I wonder whether my reasoning is flawed anywere?

The same reasoning could be applied on paired vs. one sample t-test when the interest is on differences.


1 Answer 1


You have a choice between using a two-way mixed ANOVA on the absolute values or a one-way ANOVA on the differences. I would go with the former because of what would happen if there was no difference between groups.

If the two groups did not differ then there could be two things going on:

  1. There is no difference between time 0 and time 1
  2. There is a difference between time 0 and time 1, but it is the same magnitude of difference for all groups.

With the one-way ANOVA on the differences you could not test between these two hypotheses, all you would have is the single null effect (i.e. the groups don't differ). With the two-way mixed ANOVA you would be able to test for effects of group, but also for whether there was a difference overall between time 0 and time 1. This would allow you to test between the two hypotheses: If there was a main effect of time, then this would be evidence for hypothesis 2.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.