Detecting effect modification due to the variables controlled in multiple matching

A new approach is presented to the study of heterogeneity of the odds ratio in a case-control design when the possible interaction effects are associated with the variables used for multiple matching. This analysis does not require stratification on the matching covariates.