Binary Data in Prerandomized Designs

In 1979 and 1981 Zelen introduced the so-called prerandomized designs as an alternative to the completely randomized trial of therapies: The participants of the trial are first (pre-) randomized to the treatment groups and asked afterwards for their consent. In the case of binary data, tests for selection and treatment effects in the simple and double prerandomized design are derived and then illustrated by a numerical example.