Question: Vitamin C Study. In one vitamin C study, subjects were randomly assigned to one of two treatment groups. The treated group received daily pills with vitamin C; the control group received daily placebo pills with no active ingredient. It was found later that the taste of the pills allowed many subjects to break the blind and guess their treatment group. Let X be an indicator of those that correctly guessed their group.
(a) How could this indicator be used as a covariate? Explain the model and what it can answer.
(b) Let V be an indicator of whether subjects believed they were in the treated group. Explain why V might be a better covariate than X.