Repeated-measures ANOVA and correlated variables

#1
Hello everyone,

To analyze the results of an experience for my master's thesis (at the crossroads between business and social sciences), I believe I need a repeated-measures GLM design. Basically, I want to test that "x [y, z] when A is consistent with B is significantly different from x [y, z] when A is inconsistent with B", so I measured subjects' reactions twice (to both situations).

The problem is, x, y and z are, in each case, very strongly correlated. I used a MANOVA to analyze the same variables in another experiment. Is there any way I could/should take this into account, or is it ok to just do the repeated-measures ANOVA three times? The issue doesn't seem to be addressed in the books I normally use. I am in SPSS.

Thanks a lot in advance!
 

bugman

Super Moderator
#2
Have you looked into RM MANOVA for this? This will deal with both your correlated responses and the coorelation within subjects for repeated measurements.

P