- Four causality conditions for E(Y|X,Z) - The experimental design technique of conditional randomization - Covariate selection based on the causality conditions - The example of nonorthogonal analysis of variance: Conditional and average total treatment effects