A validation method that estimates how well a chosen estimator will perform by accounting for the sampling design used in evaluation.