4
$\begingroup$

I have n subjects. I perform an experiment on all the subjects and measure the outcome in two different ways. I want to evaluate the difference between these two types of measurements. I would use a matched-pairs t-test but I will be adding in a blocking variable.

It seems like ANOVA wouldn't take into account that they are paired. 

outcome = measureType + blockingVar + measureType*blockingVar

Should I use MANOVA where the two types of measurements are the two dependent variables?

Improve this question
$\endgroup$
8
  • $\begingroup$ What exactly is a "blocking variable" in this case? Several groups of subjects? $\endgroup$
    – amoeba
    Commented Aug 10, 2016 at 19:48
  • $\begingroup$ @amoeba Yes, the blocking variable is a factor with levels that separate the subjects into several groups. $\endgroup$
    – Michael Webb
    Commented Aug 10, 2016 at 20:03
  • $\begingroup$ Then you have a repeated measures ANOVA (RM-ANOVA) with one between-subject factor (blockingVar) and one within-subject factor (measureType). It's not a MANOVA. $\endgroup$
    – amoeba
    Commented Aug 10, 2016 at 20:04
  • $\begingroup$ That makes sense. So how does RM-ANOVA take into account that measure1 and measure2 should be paired within each subject? For a matched-pairs t-test, I would have to create a new variable that is measure1 - measure2. $\endgroup$
    – Michael Webb
    Commented Aug 10, 2016 at 20:15
  • 1
    $\begingroup$ Asking R coding questions is kinda off-topic, but if it's heavily statistical in nature then sometimes it is tolerated. You can try posting a new question, asking how to code this RM-ANOVA in R, but please provide very clear description of the test you want to run, some of your current attempts (even if unsuccessful) and be sure to check if this perhaps was already asked before. $\endgroup$
    – amoeba
    Commented Aug 12, 2016 at 13:12

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.