Evaluating New Approaches to Assessing Learning

We present a framework for evaluating cognitive claims to the interpretation of assessment scores and provide evidence of its applicability to science achievement. We adapted the idea of the “assessment triangle” (Pellegrino, Chudowsky, & Glaser, 2001), in the form of an assessment square with four tightly linked corners: construct (definition), assessment (task/response/score analysis), observation (cognitive and statistical data), and interpretation (link between observation and construct). In an iterative process of assessment review, the model focuses on four analyses that feed back on one another: conceptual, logical, cognitive, and statistical and/or qualitative. The heart of the model is a knowledge framework consisting of declarative (knowing that), procedural (knowing how), schematic (knowing why), and strategic (knowing when knowledge applies) knowledge that underlie achievement. Concrete examples of the model’s application are provided.