Individual Matching with Multiple Controls in the Case of All-or-None Responses

The one-to-one individual matching principle of the matched pairs design is generalized to R-to-one individual matching in the case of all-or-none responses and fixed sample size procedures. A test is given; its asymptotic power function is derived; the selection of the matching ratio (R) is considered in relation to the unit costs in the two comparison groups; and finally, procedures for sample size determination are described.