This would entail indicating the variable containing the groups of interests, and knowing if it's binary or not.
If binary, can spread it out using pivot table
If multiclass, then the soln is tricky. It wouldn't be eligible for hypothesis_testing. Actually, the treatment and reference attributes are only used in hypothesis_testing. The object should be created even if not initially given.
Should I pivot all variables to create columns for each object in the columns? That's overkill. But maybe I should make a function for doing that, so then the treatment and reference could be adapted. This is a good idea: helper functions to spread and gather a list of variables in sample_groups.
This would entail indicating the variable containing the groups of interests, and knowing if it's binary or not.
If binary, can spread it out using pivot table
If multiclass, then the soln is tricky. It wouldn't be eligible for hypothesis_testing. Actually, the treatment and reference attributes are only used in hypothesis_testing. The object should be created even if not initially given.
Should I pivot all variables to create columns for each object in the columns? That's overkill. But maybe I should make a function for doing that, so then the treatment and reference could be adapted. This is a good idea: helper functions to spread and gather a list of variables in sample_groups.