Fitting models when response varaible distribution is something other than Normal
Aug 1, 2013 3:38 PM(1013 views)
I use menu version of JMP Pro (10.0.1). I see how to change a variable's distribution in the column properties. Then when I test the distribution under Analyze function it automatically fits the one I have selected. However, when I go to Fit Model and run say a Stepwise model with an AIC model competition, the distribution I have selected for the column does not seem to apply. The column file in the data spreadsheet still indicates the Distribution I have selected, but it seems to have no influence on the conduction of the model (i.e. I get the same output as if no distribution was selected and column is thus treated as normally distributed by default). So the bottom line is: does anyone know (with menu driven prompts) how to conduct a Stepwise model when the response variable is something other than normally distributed?