Assessment as a Lever in Education Reform
Diez, Mary E., National Forum
Assessment has become a key element of educational reform. After all, the critical question that follows a description of what we want our children to know and to be able to do as a result of their experiences in school is, How will we know when they have attained it? For nearly seventy-five years, Americans have been content to let standardized, norm-referenced tests serve as a proxy for looking more directly at the ends of education. The current movement of educational reform has created a tension between the desire to address more complex and performance-oriented standards and the practice of using easy-to-administer and score testing methods. Underlying that tension is a conflict in views about what we understand by learning and teaching, and conflicts in the varying purposes and uses of assessment data. Achieving the goals of reform will require changing the relationship between assessment and learning by making assessment integral to learning.
CONFLICTING VIEWS OF TEACHING, LEARNING, AND ASSESSMENT
Educated in the humanities as an English teacher and accustomed to engaging my high-school and college students in grappling with the meaning of literature for their lives, I remember being startled when, as a graduate student in communication, I was told that learning is a conditioned response. The experience was to be one of many that revealed to me the importance of epistemology in the study of a discipline. Gradually I began to see the conflict between competing views of the world - the behaviorist/positivist on one side and the constructivist/interactionist on the other, with many variations in between.
Is knowledge out there to be discovered or passed on as though it were an object? Or is it an interaction between the knower and the known? What difference does it make? A great deal, as it turns out.
The behaviorist/positivist position is probably best represented in American education by Edward L. Thorndike, whose focus on the quantitative made measuring, counting, and predicting central to educational research. The positivist view of knowledge privileges facts and leads to a model of teaching as the process of disseminating information. In this view, learning has occurred when students give back the information in the form of "right answers," most often in a multiple-choice format as the most efficient way to gather them. The unit of analysis is the test item, rather than the learner.
Thorndike's emphasis on quantitative testing and prediction led to the development of norm-referenced standardized tests, which spread the norming population out over the 100 point scale. By definition, 50 percent of those in the norming group will be below average and 50 percent will be above. To successfully create this "spread," test developers eliminate questions that the majority get right and privilege questions that discriminate among respondents. Interestingly, the single clearest correlation with scores on this type of test is a student's socio-economic status.
The use of norm-referenced standardized tests has had another effect: students who score low on the scale are given the clear message that they have less ability. Their parents and their teachers accept that judgment and expect less of them. A revealing study of American and Japanese mothers recently uncovered a telling difference in the two cultures: Japanese mothers faced with a child who does not perform well say that he or she did not try hard enough; American mothers in the same situation say that he or she must not have the ability.
Sylvia Farnham Diggory, in her 1990 book Schooling, notes that the constructivist/interactionist position has been less dominant than the behaviorist/positivist position in U.S. education, but it has emerged regularly throughout this century. John Dewey probably best represents the view that information only becomes knowledge when it is used or applied. …