1

I have a pre-treatment (T1) and post-treatment (T2) DV measure, and 4 IVs. I am interested in how (and which) IVs predict change in DV from pre- to post-treatment, or predict post-treatment outcome (I'm happy to answer either).

I'm currently doing a stepwise multiple linear regression of the 4 IVs and pre-treatment measurement, with post-treatment measurement as the DV.

However, I've received feedback from my supervisor that this is confusing as I've mentioned 'change' in my hypothesis, yet I am using an outcome rating at T2, in which case my regression is looking for predictors of outcomes at T2. He recommended there should be a "separate section for predicting CHANGE scores (T2-T1 scores)", and that I need to do "change score analyses".

My questions are:

  1. Should I just remove the word "change" entirely and keep to only this current regression i.e. "predictors of outcome measure"?
  2. Or should I do BOTH this regression and the suggested additional section of predicting change scores? Does it make sense to do both, and would that add huge value to my thesis?
  3. What is a change score analysis? How do I go about doing it? I did a repeated measures t-tests showing that change from pre- to post-test is significant, but apparently that's not what he was referring to.

Thank you SO very much for your help! I'm absolutely clueless and this has been giving me a massive headache. Any reply would be much appreciated!

qwertypo
  • 21
  • Do both. There's a substantial and complex literature comparing both methods, and results can vary surprisingly between the two, so you'll want to be aware of any difference. Change score = gain score = T2 - T1. Also see https://stats.stackexchange.com/questions/17724/how-should-one-control-for-group-and-individual-differences-in-pre-treatment-sco – rolando2 Mar 28 '18 at 19:16

0 Answers0