Marginal Effects in Ordinal Regression

[#1]

Hi there,

I am running an ordinal regression with 7 independent variables for my bachelor thesis and i cannot figure out how i can see the marginal effects for each variable.

The only thing I can find is the profiler, where I can see the probabilities for the output categories for a specific combination of all independent variables, and the same concept applies when I use output tables or the prediction formulas. JMP will give me the probabilities and predicted category for all possible combinations of variables. I would like to see the marginal effects for each variable, so I can interpret my results better (something like: Age has a positive effect on the dependent variable, all else equal one added year in age ads [...] percentage points).

Is this possible to create in JMP and if so, how do I get there?

Thanks!

[#2 · Accepted solution]

The odds ratio is the usual way to obtain and interpret the marginal effect of each predictor. Odds ratios are available only from the Nominal Regression platform with binary responses.