Quality Counts: Analysis of the Report

K-12 Testing

Quality Counts: A Report Card on the Condition of Public Education in the 50 States, a special issue of Education Week published in January 1997, presents probably the most coherent and comprehensive summary of data available about U.S. K-12 education. Written to help evaluate progress in school reform at the state level, Quality Counts focuses on a few areas: achievement, the existence of standards and assessments, teaching quality, adequate and equitable funding, and school climate.


The report supports performance assessments and prefers criterion-referenced to norm-referenced testing, and it recognizes that the quality of assessment needs improvement. However, the report also supports not only state exams tied to standards, but also high-stakes exams. It gives states a failing grade if they do not have a state assessment system. It also gives higher scores to states which use rewards or penalties to hold schools accountable for student performance and for graduating or promoting students based on whether they "master the standards." While accountability or mastering standards do not require use of tests, in practice this is what Education Week supports: Only states with high school exit tests score high on this indicator.


With this approach, Education Week traps itself in a quandary. To assess complex forms of thinking (e.g., divergent thinking, synthesis, evaluation or use of knowledge in a field), schools require more complex forms of assessment (e.g., performance exams with extended responses or projects). The more complex the form of measurement, the more difficult it is to ensure high reliability.


However, to pass legal muster, an exit exam must have a high reliability. The result is that imposing high stakes tends to force a narrowing in what can be measured and in the measurement methods (e.g., only multiple-choice). Once the tests are high stakes, teachers are even more likely to teach to them and to tailor instruction to their methodology. Students thus tend to be taught not the whole domain, but what is tested, and to be taught in ways that emphasize low-level drill. Scores on the tests may go up, but real learning likely will not. As a result, tests designed to establish accountability in light of standards can undermine the learning called for in the standards.


The problems of inadequate and inequitable financial and instructional resources and very limited professional development, which Quality Counts highlights, will all too likely cause the tests to simply confirm the class and race divisions in U.S. schools, but with more serious and more harmful consequences for students. Such results could become the final twist in failed school reform, in which the results on tests that helped undermine the reform are used to once again justify the very inequities that the tests reveal.