Skip Navigation

Improving Science Assessments

In recent years, the emphasis on educational reform and accountability has compelled many states to rely on high-stakes tests to assess students’ knowledge. When these assessments are administered to every student, they must be highly efficient with regard to administration and scoring. This in turn restricts what can be measured and consequently what students are encouraged to learn.

Another limitation is that teachers receive results from statewide assessments several weeks after the tests are administered. Therefore, results cannot formatively guide learning in a way that is responsive to students’ present needs.

CALA researchers are investigating alternatives to reduce these problems. Our approach involves three components:

  • Use authentic performance assessments for real-world skills that cannot be measured with conventional tests. Because the per-student cost is high, this type of statewide assessment would have to be administered to carefully selected samples of students, not every student.
  • Use teachers’ assessments to establish how each individual student performs on these real-world skills. We are examining ways to help ensure that performance assessments administered to statewide samples and those used by teachers are comparable.
  • Use teachers’ performance assessments formatively, administering them not just at the conclusion of instruction, but during instruction to guide learning.

The focus of our research is science taught at the middle school level. The intent, however, is to establish procedures that are useful at other grade levels and in other subject areas.

This project is supported by a three-year grant from the Institute of Education Sciences, the research arm of the U.S. Department of Education. View our news story at