Sample size determination in studies with matched pairs.

Two modifications are made in Schlesselman's procedure for determining the total number of pairs to be studied in a case-control study with matching. One is to take account of the use of the continuity correction in the McNemar test statistic, and the other is to incorporate a measure of the degree of association induced by the matching variables between a case's exposure status and that of his or her matched control. The result of both modifications is to increase the required number of pairs above the number produced by Schlesselman's procedure.