Ensure consistent column name prefixes

[MattsonCam]

This pr attempts to fix a problem mentioned in #19, where the column name prefixes will change in the resulting dataframe when calling the pairwise compare manager. Thanks for agreeing to see if the same problem occurs again with your notebook @axiomcura. Please let me know if this doesn't fix the issue.

[axiomcura]

LGTM. However, I'm wondering if we should add tests to verify that the column ordering remains consistent—especially since similar set() operations are used elsewhere in the code. This could be addressed in a separate issue.

[MattsonCam]

LGTM. However, I'm wondering if we should add tests to verify that the column ordering remains consistent—especially since similar set() operations are used elsewhere in the code. This could be addressed in a separate issue.

I've thought about this as well. There are two reasons I don't think this should be included as a test:

1. I see this issue as falling under the category of input validation and correction, which is already implemented in the pairwise compare manager and are not included as tests.
2. The column prefixes (i.e. antehoc, posthoc) largely don't matter in the output, so mixing these prefixes is more of just an inconvenience and shouldn't impact the final result. These prefixes are useful because the code needs a way to distinguish columns.

[MattsonCam]

Thanks for your help with this issue @axiomcura, merging now!