Test-Based Educational Accountability. Research Evidence and Implications

In recent years, many nations, including Germany, have begun to use students\u27 scores on achievement tests to monitor the performance of schools and educational systems. Such systems have been in place for a considerable time in a few nations, notably the United States, and numerous studies of their effects have been conducted. While these studies are limited, they are sufficient to reveal serious problems that should be confronted as new systems are put in place in other nations. Research in the U.S. has shown two related types of problems in test-based accountability (TBA) systems. Studies have revealed a mix of positive and undesirable effects on teaching and other aspects of educational practice. Research has also shown that increases in scores can become seriously inflated. That is, scores can increase by a larger amount - in some cases, a far larger amount - than actual improvements in student learning warrant. This paper summarizes studies of score inflation, describes several mechanisms that produce it, and notes implications for evaluation, testing, and the design of accountability systems. (DIPF/Orig.