8-way Multivariate analysis help required: 1 Output variable, 4 effector variables, 4 influencing variables

Question

Hello,

I have a dataset of 135 Patients with a 4 subtypes of a certain disease measuring the effect of a certain therapy. The effect is measured by the improvement in an numerical ordinal scale (modified rankin scale at manifestation versus end of therapy). The confounding variables are disease subtype (4 types of autoimmune encephalitis, numerical nominal variable), age (numerical continuous), sex (numerical nominal), duration of overall therapy (numerical continuous). The effector variables that I would like to get the Influence of are: total dose (numerical continuous), number of cycles given (numerical ordinal), delay till first time given in months (numerical continuous) and duration of therapy in months (numerical continuous).

How can I analyse this with JMP (I'm in the tryout of this product and not too familiar with it yet) technically (what to click) and derive ideally a outcome table with F and p for an alpha of 0,05 and respectively adapted for multiple testing. Thank you!

Mark_Bailey · Accepted Answer

JMP uses a tall format. Use one column for each variable and one row for each patient.

Select Analyze > Fit Model. Select the data column with the ordinal scale response and click Y. Select the columns with the confounding variables and the four factors and click Add. Actually, I am not sure how you are using the term "confounding." Do you mean that they have an effect along with the factors or do you mean more, that their levels change the effects of the factors?

Click Run and JMP will launch the Fit Least Squares platform. You should find all the results that you are looking for in this platform. Click the red triangle at the top and select Estimates > Multiple Comparisons.