I ran a repeated measures anova on my paired data groups A,B, and C. It rejected the null hypothesis. How do I compare two groups? Can I use two-way paired student t tests?

  • paired: yes since you have within subject measurements
  • two way: yes since the ANOVA didn't tell you which group it was that rejected the null hypothesis of three equal groups. You might look at group averages and deduce which one it was and in which direction the hypothesis was rejected. However, it would be p-hacking to use your data in this way to define your hypotheses for unilateral tests.
  • Is your sample size large enough ($>30$) so that you don't need to assume normality? In any case, you have already assumed normality for the ANOVA, but an ANOVA is quite robust to most violations of normality.
  • Don't worry about equal variances, since paired data come per definition with equal sample sizes. Unequal variances wouldn't be much of a problem unless there were also unequal sample sizes. Don't do a Welch correction in this case, the paired tests is more powerful.
| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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