AbstractThe asymptotic relative efficiency of theRto one matching design is compared to a standard one to one matching design for both continuous and dichotomous data. An additional consideration which separates this work from previous work is that in this work the pairwise correlations of the matched design are not necessarily equal. For both types of data, it is seen that the efficiency of theRto one design relative to the matched pairs design is influenced by the correlation among theRmatching “controls” relative to the correlation between theRmatching “controls” and the