Brackets, pots and kettles
Bland and Altman 'correct' Huber-Wagner's version of the logit
transformation to ln[1/(p-1)]
Sorry but that should surely be (1-p) in there!
if p = 0.2 then p-1 = -0.8 and the result comes out as ln(-1.25)
which is not a real number.
Competing interests: No competing interests