If it is a pure statistical evaluation of the yardstick, the extremes of the data should be neglected or their influence on the yardstick should be minimal. How it is treated by the committee is another matter.