Subscribe Bookmark RSS Feed

Help me understand using Manova with JMP on a mixed (1 within, 1 between) experimental design


Community Trekker


Jun 30, 2013

I’m teaching a graduate psychology Anova course using JMP, and I want to try to teach my students something I don’t yet understand or know how to do myself. If someone can help here, we’d all be grateful. (My background is regression, and I never learned how to analyze mixed designs. Note: I have learned how to analyze the data using the Mixed Models option in JMP, i.e., via multilevel modeling, and I understand that approach is superior. But I still want to understand the Manova approach better.)

The data file is attached. Here’s the explanation (real data, although that doesn’t really matter). There is a between subjects factor, Group, with 3 values (A, B, and C). Each subject is measured on three DVs: Within1, Within2, and Within3. (These are different experimental conditions.)

I have been teaching them to use planned orthogonal contrasts, and I’d like to use that approach here as well. Let’s say, then, that I’m interested the following contrasts:

Contrast 1 (Between Groups): Contrasts A versus the average of B and C

Contrast 2 (Between Groups): Contrasts B versus C

Contrast 3 (Within Groups): Contrasts Within1 versus the average of Within2 and Within3

Contrast 4 (Within Groups): Contrasts Within2 versus Within3

Contrast 5 (Interaction): Contrast 1 * Contrast 3

Contrast 6 (Interaction): Contrast 1 * Contrast 4

Contrast 7 (Interaction): Contrast 2 * Contrast 3

Contrast 8 (Interaction): Contrast 2 * Contrast 4

I can, and have, constructed the two Between Groups contrasts in the JMP data file. But I can’t do the same with the others, because the data are in wide form.

How do I conduct this analysis using “Fit Model: Manova Personality?”