An academic researcher has a data set with 30 binary columns and wants to do every possible pairwise chi-square test among them (30 x 29 / 2 = 435 tests) using FDR control. Looking at Response Screening, there doesn't appear a way to set up this specific set of tests. Including all 30 columns as both X and Y causes symmetrical tests to be run (e.g., A-by-B and B-by-A; 30 x 29 = 870 tests), making the FDR results overly conservative.
The output does helpfully include a Sym column in the data table to identify the symmetrical tests, but there doesn't appear to be a way to tell response screening to run only tests where e.g., Sym = 1. It also won't accept the same column in the X or Y roles multiple times, so there seems no way to set it up using the Paired X and Y option. (This can be scripted using the Paired X and Y option, but we intend to use this for an entry in the Academic Case Study Library and want to avoid JSL.)
Thanks for your thoughts, even if all you can do is confirm that Response Screening can't handle this particular problem.
