This paper delineates several problems which arise when criterion-referenced test results are used to evaluate the effects of a specific educational treatment. Focus is on these topics: (1) the aggregation of individual students' data on objectives, (2) the sensitivity of the instrument to instruction, (3) the interpretation of criterion-referenced group data, and (4) comparison of crition-referenced test results and standardized test results. The results show that the methods used in reporting individual objective data affect the outcomes. In addition, comparisons of gains show that the criterion-referenced tests are not necessarily more sensitive to student growth. (Author/RC)