2
$\begingroup$

I have a two-group pre-test/post-test design with one dependent variable. I want to find out whether or not an intervention has been effective. Would it be more appropriate to do a t-test on the gain scores or a two-way mixed ANOVA?

Thanks!


It's also worth noting that the means of both groups at pre-test were identical.

$\endgroup$
  • $\begingroup$ Means were identical? That sounds fishy to me. $\endgroup$ – Jeremy Miles Apr 12 '16 at 18:06
2
$\begingroup$

It doesn't matter. These are the same test. (Try it, you'll get the same p-value, and $F$ will be equal to $t^2$.

However, most people would suggest doing a regression analysis, and controlling for baseline scores. If you don't, you're acting as if pre-and post scores are correlated 1.0 (within groups).

See, for example: http://www.ncbi.nlm.nih.gov/pubmed/16895814

$\endgroup$

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.