Response to Wang and Walberg

Abstract

Response to Wang and Walberg Wang and Walberg characterize their ALEM research as descriptive, not experimental (p. 134, this issue). They write, for example, that teachers . . . are not randomly selected by an investigator for study participation (p. 134); instead, parental and teacher choice decides (p. 132). However, Wang and Walberg claim our demand for high quality experimental research is unfair and impractical. They complain, Such research presumably would obtain representative samples and randomly assign students to experimental and control They end their protestation with confidence: School-based studies using criterion of a true experimental design, of course, are nearly impossible; such studies are extremely rare (p. 131). As a valued scholar-farmer at University of Minnesota would say, Hog-wash! Schweinhart and Weikart (1985) described from seven of earliest and longest ongoing evaluations of preschool programs for children from poverty. Of these valuations, 4 employed experimental designs, despite that most were begun 20 to 26 years ago (see Schweinhart & Weikart, 1985, p. 546). In an evaluation of Title I programs, Slavin (1987) reviewed pertinent literature in search of rigorously evaluated and effective models (p. 111). In his Tables 1 through 5 (see pp. 112-119), he described 20 programs that have been researched with matched and/or randomly assigned control groups. In fact, Slavin has conducted several of his ownlarge-scale, school-based investigations of a mathematics program. One study (Slavin, Madden, & Leavey, 1984) compared this program to control methods in a 24-week experiment involving 1,371 students in 59 classrooms. More than 8% of pupils were designated handicapped. Examples of scientifically rigorous, school-based investigations are not difficult to fiond in research literature, which is not to say that such studies are easy to accomplish. Indeed, are not. But neither is it easy for a general educator to implement effective instruction for all students in a mixed ability classroom or for a special educator to engineer successful transfer of pupils from resource to regular classrooms. Despite difficulty of their jobs, we expect these educators to aim for, if not achieve, excellence. Likewise, conducting high-quality, school-based research is challenging, but do-able. Teachers, parents, administrators, and policy makers as well as educational research community should demand it, especially in connection with proposals such as GEI, which call for fundamental sweeping change. ALEM RESEARCH: IF NOT EXPERIMENTAL, THEN HOW MUCH? Although Wang and Walberg concede ALEM research is largely nonexperimental, contend that their descriptive research and replications of program implementations . . . are fundamental to building a research base (p. 132). This prompts question, How many ALEM replications, or implementations, have Wang and colleagues conducted? Because chunk data obtained from one implementation into smaller segments, and then publish these as multiple studies, there is an appearance of many ALEM implementations. This impression is strengthened by their tendency to report some of same data in different places. Bryan and Bryan (1988) have argued, the danger is that this [practice of duplicative reporting] creates an impression of massive data and replicated results (p. 3). In fact, we found no more than four and possibly as few as two documented implementations, depending on interpretation of which published study belongs to what implementation effort. Wang and Walberg argue that this seriously underestimates size of ALEM data base. They charge, the Fuchs have overlooked a vast number of ALEM studies (p. 133), conducted by them and others, and claim one reason for this purported oversight was that they did not come to us directly as most knowledgeable source of information (p. …