Fit two-level screening doesn't give the same result when replicated. Why?

[#1]

Hello,

I noticed something odd recently when running the "fit two-level screening" analysis in JMP under the DOE:Classical:Two-level Screening menu. With the exact same data table, the analysis does not give the exact same results (see below for three examples). While the A-B-C contrasts themselves are the same each time, the Null contrasts are not. So the half normal quantiles look different. Does anyone know why these come up different each time?

[#2 · Accepted solution]

I think the reason is explained here: Design of Experiments Guide > The Fit Two Level Screening Platform > The Screening Report > Contrast... 

Note: Do not expect the p-values to be exactly the same if the analysis is re-run. The Monte Carlo method should give similar, but not identical, values if the same analysis is repeated.