The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Knowing What Students Know: The Science and Design of Eduacational Assessment
conveyed by state and national curriculum standards, would be enhanced by facets-type research on learning. Current standards based on what we want our students to know and be able to do could be improved by incorporating findings from research on what students know and are able to do along the way to competence. By using a matrix sampling design, facet clusters could be covered extensively, providing summary information for decision makers about specific areas of difficulty for learners—information that would be useful for curriculum revision.
Use of Large-Scale Assessment to Signal Worthy Goals
Large-scale assessments can serve the purposes of learning by signaling worthwhile goals for educators and students to pursue. The challenge is to use the assessment program to signal goals at a level that is clear enough to provide some direction, but not so prescriptive that it results in a narrowing of instruction. Educators and researchers have debated the potential benefits of “teaching to a test.” Proponents of performance-based assessment have suggested that assessment can have a positive impact on learning if authentic tasks are used that replicate important performances in the discipline. The idea is that high-quality tasks can clarify and set standards of academic excellence, in which case teaching to the test becomes a good thing (Wiggins, 1989). Others (Miller and Seraphine, 1993) have argued that teaching to a test will always result in narrowing of the curriculum, given that any test can only sample the much broader domain of learning goals.
These views can perhaps be reconciled if the assessment is based on a well-developed model of learning that is shared with educators and learners. To make appropriate instructional decisions, teachers should teach to the model of learning—as conveyed, for example, by progress maps and rubrics for judging the quality of student work—rather than focusing on the particular items on a test. Test users must understand that any particular set of assessment tasks represents only a sample of the domain and that tasks will change from year to year. Given this understanding, assessment items and sample student responses can provide valuable exemplars to help teachers and students understand the underlying learning goals. Whereas teaching directly to the items on a test is not desirable, teaching to the set of beliefs about learning that underlie an assessment—which should be the same set of beliefs that underlies the curriculum—can provide positive direction for instruction.
High-quality summative assessment tasks are ones for which students can prepare only through active learning, as opposed to rote drill and practice or memorization of solutions. The United Kingdom’s Secondary School Certification Exam in physics (described in more detail later in this chapter) produces a wide variety of evidence that can be used to evaluate students’ summative achieve