A hybrid paired and unpaired analysis for the comparison of proportions.

A method is presented for making inferences about the difference between two binomial proportions when there is partial overlap of the two samples. The method will simplify to the usual cases of two independent samples or paired samples (McNemar's method) depending on whether there is no or complete overlap. The method is used to compare sensitivities, specificities and Youden indices from a comparative study of trauma triage (screening) instruments.