The Design of School Accountability Systems

Accountability became fashionable in the 1960s. In the 1970s, there was a heated discussion about the pros and cons of ac­countability. We discussed the need for accountability. We also discussed problems and limitations. During the 1960s and 1970s state account­ ability systems began to emerge. For ex­ample, Michigan, Florida, and New York instituted statewide accountability sys­tems. These varied in philosophy, empha­sis, and effectiveness. What is new today is a widespread recognition that increased state funding of public education will inevitably trigger new calls for state ac­countability. Faced with demands for improvements in the schools, legislatures are increasingly attempting to affect pol­icy and outcomes by instituting statewide controls on the schools.

Current trends are to rely on testing, more specifically on standardized true­–false testing. There are many reasons for this tendency. Standardized tests are available and useful. They have increas­ingly been used to point to the successes and failures of U.S. education. In the last decades they have become the best known and the most documented tool to evaluate the state of U.S. schools. Because the tests are not directly linked to the curriculum, they can be used across the board to evaluate many different stu­dents or many different schools. They provide an egalitarian tool for centralized control without obliging the state to inter­vene in the daily affairs of the schools. Or, at least, they seem to.

Centralized control, however, inevita­bly results in some form of intervention. The current call for statewide accounta­bility systems linked with economic in­centives means that standardized testing is to take new importance in the schools. For example, in California, most if not all the proposed new school-based account­ ability systems under discussion in the legislature use the schoolwide scores of the California Assessment Program (CAP) to provide school comparisons that would be used in a statewide incentive scheme whereby schools achieving high scores would be rewarded financially. It is diffi­cult to imagine that tests that are useful to diagnose students' educational needs and potentialities could be used to eval­uate schools without distorting what hap­pens in the schools. The fact that the tests are not linked to the curriculum makes them uniquely useful instruments to as­sess students or even schools. But the moment financial incentives are linked with the tests they are bound to yield goal displacement: the tests will become goals; the curriculum, the teachers, and even the textbooks will start to look like the tests. The main argument of this paper is that central control is inevitable. But central control cannot operate without cen­tral responsibility. If the state needs to assess achievement, it will have to ad­ dress the question of state definitions of a desirable curriculum and will have to in­volve educators in these definitions.

This article reviews many of the prob­lems associated with accountability. It goes beyond to suggest a set of principles for statewide accountability systems. It argues that teachers are im­portant. We begin by describing teacher performances we would all like to en­ courage. We want to establish a consen­sual nonpolemical view of the teacher's role, a starting point for discussing how to design accountability systems. We then proceed to discuss accountability and its uses: (a) to inform (i.e., provide feedback), (b) to reorient action, and (c) to justify action. This leads us to a more detailed discussion of how accountability actually works. We examine the importance of establishing a linkage with teacher re­ wards or sanctions and the greater impor­tance of rewards over sanctions in motivating teachers. We come to the in­evitable conclusion that the teaching profession, as presently structured, does not provide sufficient incentives. Ac­countability with little incentive leads to little change. We recognize also that ex­cessive use of accountability, i.e., exces­sive use of testing of one kind or another, tends to lower the status of the profession. We believe that accountability systems should be parsimonious. They should en­hance the quality of life among teachers and not require excessive paperwork.

We distinguish between top-down and bottom-up accountability and suggest a role for both. We discuss the kinds of measures we might want to collect and the uses we might put them to. We argue that standardized achievement testing is unsuitable in schoolwide accountability for four main reasons: (a) when such test­ing is tied to economic incentives it inev­itably leads to goal displacement, even when this is not the intent; (b) such test­ing, when used in schoolwide scores, masks real issues such as student turn­ over; (c) standardized testing does not es­tablish a minimum standard and does not tell us how to reward schools trying to deal with difficult students; and (d) stand­ardized testing is not designed to provide sufficient incentives. Accountability sys­tems should provide positive encourage­ment but standardized tests, by definition, discourage half the population.

We end the article by presenting a set of principles and the outline of a statewide accountability system that would rely on both top-down and bottom-up accounta­bility.

This article was originally published in Educational Evaluation and Policy Analysis by the American Educational Research Association and Journal Storage (JSTOR).