48984_en-US

liliconcom

Give a warning message when random effects don't have 'Nominal' modeling type

Random effects should (almost) always have a 'Nominal' modeling type. I have seen multiple times where JMP users have random effects in their model with a 'Continuous' modeling type and the results are incorrect and the issue is difficult to spot. If JMP gave a warning message if any random effects were not 'Nominal', it would greatly reduce the problem.

Advanced Statistical ModelingSharing and Communicating ResultsRe: Give a warning message when random effects don't have 'Nominal' modeling type - Status changed to: Acknowledged

Hi @jwiltsie, thank you for your suggestion! We have captured your request and will take it under consideration.

Re: Give a warning message when random effects don't have 'Nominal' modeling type

In correspondence for JMP, the warning message is displayed if your modeling personality is Mixed Model (JMP Pro). However, in base versions of JMP, you tend to fit this model as Standard Least Squares and use REML, where the warning message is not displayed and an incorrect analysis is run. I agree that the warning message should be in both places.