How Is Testing Supposed to Improve Schooling?

摘要

Validation research for educational achievement tests is often limited to an examination of intended test score interpretations. This article calls for an expansion of validation research in three dimensions. First, validation must attend to actual test use and its consequences, not just score meaning. Second, validation must attend to unintended as well as intended testing consequences. Third, validation must attend to indirect as well as direct testing effects. Indirect effects include the effects of score-based incentives in prompting actions intended to raise test scores (directing student effort or focusing the system) as well as messaging effects associated with a testing program per se but not dependent on specific scores (shaping perceptions). This expanded program of test validation can best be accomplished by measurement professionals working in collaboration with scholars from other social science disciplines.