Test-Based Accountability: The Promise and the Perils

Reform movements in American education are based on theories of social change. The standards and accountability movement is based on the theory that a sequence of three activities will improve education: first, defining what students should learn (setting standards); second, testing to see what students have learned (measuring achievement); third, making the results count (holding educators and students accountable). Most analysts date the standards and accountability movement to the early 1990s, when states began establishing standards in academic subjects. States then instituted testing programs and implemented incentives for schools and students based on pupil test scores. The systems are mature enough to have produced some preliminary results.

What is known so far about the effects of accountability systems on student achievement? Do they work? Are there any unintended consequences? In general, evaluations of accountability systems have been quite positive. In raising student achievement, states that have implemented such systems are outperforming states that have not done so. Although the potential for serious unintended consequences cannot be ruled out, the harms documented to date appear temporary and malleable.

Promising results, however, do not guarantee the longevity of an education reform.¹ Various threats to accountability exist, in particular, the political perils that state systems face when policies are implemented. What do these threats portend for the future of test-based accountability in the United States? That question is especially relevant today as the No Child Left Behind Act, the landmark legislation that federalized what had been primarily a state and local policy domain, comes under increasingly heavy fire from critics. "An Education Rebellion Stirring," headlined the Christian Science Monitor in February 2004. In March of the same year, an article in the New York Times reported that legislation or resolutions criticizing the law had passed at least one house in twelve state legislatures (both chambers in seven states) and described the coalition of anti-No Child Left Behind Act forces in Oklahoma as "liberal Democrats and states' rights Republicans, angry over what they see as a cumbersome federal intrusion on local schools."² A reasonable idea of the future of the No Child Left Behind Act can be gleaned from analyzing the politics of accountability in the states.

What Does the Evidence Show?

Because many states did not have systems in place before 2000, studies of accountability are relatively recent. The preliminary results are positive, but researchers are not unanimous. Audrey L. Amrein and David C. Berliner of Arizona State University are the authors of two 2002 studies sharply critical of high-stakes testing.³ The researchers examine twenty-eight states with a mixture of student and school accountability policies. States are categorized as having high stakes for students if test scores are linked to promotion in grade level, graduation from high school, monetary awards or scholarships, or the freedom to transfer out of poorly performing schools. States are considered as having high stakes for schools if the state offers monetary awards to educators in high-performing schools or can take over low-scoring schools, replacing principals or teachers, based on test scores. One of these studies investigates the effects of high-stakes testing on various indicators of academic achievement. The other looks at the effects of a high school exit exam on dropout rates and enrollments in General Educational Development (GED) programs.

In the first study, Amrein and Berliner find that states showed no clear academic gains after adopting high-stakes testing. The findings from the second study are decidedly negative. In the sixteen states with exit exams, dropout rates increased, graduation rates declined, and GED rates went up relative to national changes on the same measures. Amrein and Berliner conclude that there is no solid evidence that high-stakes tests produce achievement gains and considerable evidence that high-stakes tests produce negative consequences.

The Amrein and Berliner studies drew withering criticism, primarily on methodological grounds. Comparing data from high-stakes states with national averages...