Please see the attached. Best wishes, David Booth

Hello Jeffrey,

It's not clear to me what you have (or what you seek). At first blush, it sounds to me as if extracted color scores is the outcome/DV (with three variable values for each instance) and reference color is the IV.

Are the paired data sets independent? (In other words, are you confident the the results for RC instance (211,24,175) observed as EC instance (191,12,212) are independent of those for the next pairing: (72,88,15) and (119,48,128)?) If not, then you'll need to consider a model which takes dependencies into account.

If what you wish to know is, are differences (e.g., 211 - 191; 24 - 12; 175 - 212) distributed with a mean centroid value of (0, 0, 0), then: (a) form the R differences, G differences, and B differences as a set of three variables, (b) evaluate whether the centroid = (0, 0, 0) (this is sometimes called the test of the intercept).

If this misses the mark, then perhaps you could elaborate your query. Doing so might help to get a more focused recommendation. In any event, Professor Booth's suggestions can point you to the mechanics of manova/multivariate regression in R.

Good luck with your work.

Dear Lee,

try to read:

Dear Lee,

try to read