Curriculum-Based Measurement and Developmental Reading Models: Opportunities for Cross-Validation
The merit of basic or theoretical versus applied or practical research is a long-standing topic of debate among researchers and practitioners. Harris (1988) has argued not for a greater emphasis on either basic or applied research, but, rather, for more emphasis on the interface between the two--between the "why" and the "how" questions. Similarly, Bernice Wong (1988), editor of a 15-article series on basic versus applied research in learning disabilities, concluded that both kinds of research are necessary for advancement of knowledge and urged increased cooperation between both camps of researchers. In this vein, we propose that models of reading development constructed by basic research provide support for the applied methods of curriculum-based measurement (CBM), while the effectiveness of CBM in the field provides further verification for the validity of the reading models.
Curriculum-based assessment (CBA) is the generic term for procedures that sample a student's skills in the student's actual curriculum. A variety of models of CBA tend to emphasize measurement that is brief, is frequent, is based on the student's classroom curriculum, and can be used to monitor instructional progress and effectiveness (Frisby, 1987; Tucker, 1985).
Most models of CBA rely on informal, non-standardized procedures; however, CBM is an empirically derived, standardized form of CBA developed by Stanley Deno and his colleagues at the University of Minnesota. With CBM, academic skills are assessed through repeated 1- to 3-minute (min) rate samples using stimulus materials taken from a student's curriculum. CBM has been applied to several academic areas, but most research has dealt with reading, which is the focus of this article. Perhaps because of the variety of educational decisions for which CBM has been used, CBM has become one of the more controversial forms of CBA (Allen & Marston, 1988; Deno, 1988; Lombard, 1988a, 1988b).
The two reading models addressed here, Chall's (1983) stages of reading development and LaBerge and Samuels' (1974) model of automaticity in information processing, both view learning to read as a developmental process consisting of component skills that build on each other. These models emphasize decoding skills at the beginning levels of reading, in contrast to holistic models that emphasize whole-word meanings.
We first outline the rationale and development of CBM and its empirical support. Then we summarize two reading models before we discuss how CBM and the reading models may validate each other.
Development of Measures
One of the goals of the University of Minnesota Institute for Research on Learning Disabilities (IRLD) from 1977 to 1983 was to develop a measurement system that would allow teachers to routinely monitor student achievement. The basic requirements of such a system were that it be (a) reliable and valid, (b) simple and efficient, (c) easily understood, and (d) inexpensive--allowing for repeated measurement (Deno, 1985). Measures were developed in three areas: reading, spelling, and written language (Deno, Marston, & Mirkin, 1982; Deno, Mirkin, & Chiang, 1982; Deno, Mirkin, Lowry, & Kuehnle, 1980); but as previously noted, reading is the focus of this article.
In approaching the question of what in reading should be measured and how it should be measured, the Minnesota IRLD research group started with the assumption that the goal of reading is comprehension. They soon decided that although directly assessing comprehension through answering comprehension questions might be a task high in content validity, this approach was not simple, efficient, or economical (Deno, 1985).
Through a review of the literature on reading, other possibilities for assessment methods were determined and then tested in field trials. …