A comparison of the precision of three experimental designs employing a concomitant variable