Estimating NonResponse Bias in Mail Surveys: A Replication Study