By Dylan Wiliam
Cherryholmes, Berlak, and others suggest that assessments are, inevitably, value-laden. Berlak says, “But as Cherryholmes argues, it is not only that subjective judgments are involved; it is also that constructs themselves are products of power and their use is an exercise of power.” I myself have written that there is no such thing as “assessment degree zero” and that “assessments reify the constructs they purport to assess.”
However, there is another way to think about assessments, and that is by making a clear separation between the constructs that are to be assessed and the assessments themselves.
Within this alternative view, assessments are value-free, or, more precisely, are no more value-laden than the constructs they assess. This can be illustrated by considering the assessment of history.
For some, being good at history means knowing lots of facts and dates. Proponents of this view consider multiple-choice tests perfectly adequate for assessing history because using a multiple-choice test, one can assess a large number of facts and dates, and score the responses reliably.
For others, history is about evaluating conflicting sources of evidence, constructing historical argument, understanding chronology, and cause and effect. Adherents of this view consider multiple-choice tests inadequate because they underrepresent the construct of history. These people see constructed-response tests, preferably including writing extended essays, as more valid.
On the other hand, proponents of the view that history is about facts and dates would regard essay-based tests as deeply flawed, since they assess students’ writing and linguistic skills as well as historical knowledge (what is sometimes called “construct-irrelevant variance”).
The point is that these different views come into opposition when we try to assess history, but they are not really arguments about assessment. They are arguments about what history is, or should be, about. In other words, they are arguments about constructs.
Of course, assessments can be bad. Assessing students in a language other than that in which they have been taught a subject will lead to results that are impossible to interpret; was the student’s response weak due to a lack of knowledge? Was it because the student was unable to understand the question or express a response in the language required in the test?
But if they are well constructed, assessments will be no more value-laden than the constructs they assess. The correct “objects of history” are constructs, and the “exercise of power” is the decision to assess certain constructs rather than others. By being clear about this—by being clear that arguments about assessment are usually arguments about the underlying constructs—we are likely to have far more fruitful and effective discussions.
Have a Question for Me?
I’m ready to answer it! Send me any questions you have about embedding formative assessment, college and career readiness, Common Core State Standards, or any other hot topic in education. In next month’s issue, I’ll offer my thoughts on the subjects that matter most to you.
Berlak, H. (1992). Toward the development of a new science of educational testing and assessment. In H. Berlak, F. M. Newmann, E. Adams, D. A. Archbald, T. Burgess, J. Raven & T. A. Romberg (Eds.), Towards a new science of educational testing and assessment (pp. 181-206). Albany, NY: State University of New York Press.