Jun 14, 2019 12:31 PM
| Last Modified: Jun 14, 2019 12:44 PM(528 views)
I ran an analysis in JMP, took the create SAS Code from the model dialog and input into SAS and ran. However, the analysis spits out different results that what JMP spit out, again same code tyring with PROC GLM and PROC MIXED. I tried different solution DDFM to see if that was the issue with no luck. Wondering if anyone has any recommendations as some of my interactions are no longer significant in SAS as they were in JMP. Thanks, SAS code and images below.
PROC MIXED DATA=Active_Inactive_Only_5_locations ALPHA=0.05;
CLASS Group WgtCat Sex Trt BehavCategory3;
MODEL Log10_0_001_PropBudget = Group(WgtCat Sex Trt ) WgtCat Sex WgtCat*Sex Trt WgtCat*Trt Sex*Trt WgtCat*Sex*Trt BehavCategory3 LocationCat2 Trt*LocationCat2 BehavCategory3*LocationCat2 LocationCat2*LocationCat2 LocationCat2*LocationCat2*Trt LocationCat2*LocationCat2*LocationCat2;
See Help > Books > Fitting Linear Models. The statistical details will explain the difference in the parameterization of the models in SAS and JMP. They represent different effects and so, the tests are different. That is, The SAS effect and the JMP effect are not the same, so neither are the tests.