Must Instructionally Useful Performance Assessment Be Based in the Curriculum?

Article excerpt

During the past decade, traditional, commercial, norm-referenced tests have come under sharp scrutiny and have been criticized for both their lack of authenticity and their lack of utility in helping teachers improve the quality and effectiveness of their instruction (e.g., Shepard, 1991; Wiggins, 1992). A great deal of interest and activity has been generated around developing alternative assessment methods that are relevant to the primary purposes of learning and that can be used to enhance teachers' instructional planning. The result is that portfolio assessment (e.g., Hebert, 1992; O'Neil, 1992; Paulson, Paulson, & Meyer, 1991), authentic assessment (e.g., Archibald & Newman, 1988; Wiggins, 1989), and performance assessment (e.g., Bembridge, 1992; Nuttall, 1992; Szetela & Nicol, 1992) have become the focus of attention as schools move away from traditional approaches to assessing outcomes.

Curriculum-based assessment is one performance assessment approach that has become increasingly popular over the past decade among special educators. Two major reasons for this popularity can be identified. First, the notion that assessment can be conducted in the curriculum of the local school is appealing to educators and school boards, because it helps preserve the sense of local control. With curriculum-based assessment, decision makers can feel that student learning is being appraised in the curriculum designed or selected for their school or district. A second reason for the popularity of curriculum-based assessment is that it allows teachers to determine directly the extent to which a student is learning what is being taught. In aligning the curriculum and assessment in this way, curriculum-based assessment may help teachers index student progress, evaluate the effectiveness of their instruction, and design better programs (Fuchs & Fuchs, 1986).

Researchers have described a variety of curriculum-based assessment methods (e.g., Blankenship, 1985; Deno, 1985; Gickling, 1981; Rosenfield, 1987; Shapiro, 1987). The specific procedures of these different forms of curriculum-based assessment vary dramatically. Nevertheless, as defined by Tucker (1987), these methods share three features: Student proficiency must be sampled in material from the school's curriculum; assessments must recur over time; and information must be used to formulate instructional decisions.

Given current debates about assessment, we think it is important to reflect on the most salient, distinguishing assumption of curriculum-based assessment: To be instructionally useful, measurement must be based on the materials in which instruction occurs (Tucker, 1987). To examine this issue, we first describe the literature that explores the importance of assessing a student's proficiency in his or her instructional curriculum. (We use the word curriculum to denote the materials in which instruction occurs, not the goals, objectives, or outcomes inherent in school curricula.) Then, we identify the potential advantages and disadvantages of curricular sampling of academic performance. Finally, we propose a set of features for ensuring the instructional utility of measurement.

EVIDENCE ON THE EFFECTS OF

CURRICULUM SAMPLING

WITHIN TESTING

A diverse literature addresses the effects of curriculum sampling on student test performance. This literature can be organized in terms of (a) analysis of the match between commercial norm-referenced tests and curricula and (b) comparisons of measurement validity and the effectiveness of instructional decisions when student performance is measured directly across different curricula.

Analysis of Match Between Commercial Tests

and Curricula

Evidence of Mismatch. Most available analyses of the overlap between tests and curricula indicate mismatch. Two early relevant studies were conducted at the University of Illinois Center for the Study of Reading. …