Constructing exact significance tests with restricted randomization rules

From MaRDI portal





The situation in a clinical trial comparing two treatments A and B is considered. If n patients are assigned randomly to the treatments let \(Y_ j=0\) or 1 according as the j-th patient is assigned to A or B. Let \(x_ 1,...,x_ n\) be the n responses and \(a_ 1,...,a_ n\) the corresponding scores (i.e. \(a_ j\) is a function of the rank of \(x_ j).\) For the test statistic \(T=\sum a_ jY_ j\) the exact distribution conditional on the observed treatment imbalance D is derived where D is the difference between the terminal number of patients assigned to the two treatments. Two examples are considered.











This page was built for publication: Constructing exact significance tests with restricted randomization rules

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3780329)