N-way Manova for 4 dependent variables and 3 independent variables?

My data is in a table with 7 columns. X1: within-subject factor, 16 levels [frequency at which measurement is made] X2: within-subject factor, 3 levels [amplitude at which measurement is made] X3: between-subject factor, 7 levels [time at which measurement is made] Y1-Y4: response variables, all continuous, (real numbers) The experiments were done in triplicate, i.e. there are three data points (rows) for each unique combination of levels of all three factors.

How should I set up fitrm() and manova()? Any help appreciated

Answer 1

[Rey9]

Hi,

You may design within-subjects table to show which factors an how you want to compare within-subjects. then you need to note in "fitrm", the factors for between subjects.

Overall, you have two tables one is your data , it should be in a "table" structure, and one for strategy "withindesign" or "withinmodel". Finally, you will be able to do "MANOVA" via Matlab finction.

Simple example the repeated measurment design can be provided following;

rm = fitrm(T,'y1-y4 ~x3','WithinDesign',within)

T is table for data and factors, if I understood your plan, "y1-y4" is your observation factors and, "x3" between subjects factors. then you need plan for withinsubjet in "within" table with x1 and x2, (size depends on how you want to test your hypotheses like 4x2, if you have two types of within subject factors analysis).

Later you can run this function:

manova(rm)

I hope it helps,

See matlab help exampe in https://www.mathworks.com/help/stats/fitrm.html

load repeatedmeas
rm = fitrm(between,'y1-y8 ~ Group*Gender+Age+IQ','WithinDesign',within)