State Departments of Education are turning to the use of criterion referenced, as opposed to norm referenced, models for statewide assessment. The underlying assumption in this turn of events is that results generated by criterion referenced tests within the statewide assessment context permit the drawing of value inferences about the effectiveness of the educational curricula under study. The tenability of this assumption is examined in light of rigorous requirements for test construction and validation. The extent to which the test construction steps can be followed closely to yield a content valid test determines the extent to which the tests can be justifiably used to evaluate the curricula or programs under study. In summary, it is to be concluded that the content validity of measuring instruments must be carefully established in order to ensure meaningful and defensible decision making. The risk involved in using an invalid test must be judged in terms of the costs (psychological, financial, etc.) attendant on making erroneous decisions in a given situation. (MV)