Shift Analysis for Large Number of Tests

[#1]

I’m trying to check for run-to-run variation across a large number of tests — e.g. >1,000 tests

The data is in long format, with all test values in a single column (TestValues). Each unit has measurements across multiple runs (e.g., WW1, WW2, WW3), identified by a WW column, and we have a UnitID to track which unit it is. I want to:

• Check if there’s a statistically significant shift across runs for each test,

• And get the results in a summary table format (p-values, confidence intervals, etc.), or a formate where I can track big number of tests. 

I’m using base JMP (not Pro), so I don’t have access to the Mixed Model option. What’s the best way to do this efficiently across all tests?

[#2]

What might work is Fitting a model with Y = TestValues and WW as Model Effect and fit By Test Id. This will Fit Model for each test and analyze each test separately.  Then you can right click on any table and Make Combined Data table which will give you the results you clicked on (e.g. effect summary) for all tests in a table. Let me know if this works. 

Spyros