# Can you use composite z-scores in a repeated measures t-test?

Can you use composite z-scores to perform a t-test. If so, can you use a composite z-scores in a repeated measures t-test?

-
what do you mean by "composite $z$-score"? – Macro Jun 20 '12 at 0:39
Four separate items converted to z-scores and summed. – Matt Jun 20 '12 at 0:40
Then, sure, you can treat this variable the same as you would any other quantitative variable. – Macro Jun 20 '12 at 0:54
Thanks for the quick response. Do I have to do anything special with the z-scores for a repeated measure design? Maybe I should calculate the z-scores based on the total mean and SD for each variable before separating the data into my before and after groups to avoid getting means of zero across the board? – Matt Jun 20 '12 at 1:09
I actually became aware of this website because of your post "Calculating Composite Scores of Ability and Other Tests." Oddly when I calculate the means from my composite measure the two groups always seem to be the same distance from the mean like M1=0.76 and M2=-0.76 or M1=1.79 and M2=-1.79. I've tested the composite measure on two different datasets (one between and the other within subject designs). Is it just coincidence that my results turn out this way? – Matthew Ebben Jun 20 '12 at 12:24