Discusses problems in using multiple-choice items on criterion-referenced tests, offering alternative methods to determine if objectives have been met. These include batteries of true-false items, structural communication grids, concept linkages, and a scoring system which takes partial knowledge into consideration. (JN)