Treatment and sex are collinear in `margex` data

[akgold]

> table(margex$sex, margex$treatment)
        
            0    1
  female    0 1502
  male   1498    0

[anniejw6]

:woman_facepalming:

[nathanielolin]

So should we fix this and rewrite all the tests? ...and am I volunteering by leaving this comment...