Created:
| Last Modified:
I have fit a Mixed Model via the Fit Model platform in JMP Pro. When opening up the 'Prediction Profiler', and inspecting the Profiler itself, I can see the as expected predicted response and its as expected confidence interval in [], which reflects the variation due to the Random Effects I have fit.
However, when I run the 'Simulator' with equal experimental variables as above, and after adding noise to my response via the 'Random by Model', the simulated to table response has a much smaller variation which is equivalent to just the RMSE/STDEV of the model fit. It does not seem to incorporate my Random Effects variance. You can see this below in the 'simulated histogram which does not include any bins even remotely close to the [11.9, 30.2] interval.
Is there a way to have the Monte Carlo simulator incorporate the "Random Effect" variance in its predictions like it does in the visual 'Prediction Profiler' tool?