Methodological issues in group-matching designs: alpha levels for control variable comparisons and measurement characteristics of control and target variables.