مشکلات کلیدی در تشخیص تاثیرات توانایی گروه بندی برروی دستاورد دانش آموز

The paper presents empirical evidence that earlier research may have overstated the impact of ability grouping and tracking on inequality in student achievement. We list six key difficulties facing research on the effects of grouping on student achievement. Each of these difficulties offers opportunities for further research and for collection of more appropriate data sets. Strong conclusions as to the differential effect of ability grouping on high-achieving and low-achieving students are probably not yet warranted

Ability grouping is a widespread practice in American schools. For well over a decade, researchers have investigated how the grouping of students into classrooms by achievement levels (ability grouping) has affected the average level and the dispersion of achievement. The purpose of this paper is to enumerate some of the major difficulties in distinguishing the impact of ability grouping on student achievement. Section 2discusses the comments made by Rees, Brewer and Argys (1999)(henceforth RB and A) and presents new evidence that omitted ability bias has likely led to an overstatement of the differential effects of grouping in the previous literature. Section 3lists six key difficulties that confront all researchers in this area. We conclude that based on the existing evidence it is difficult to make a clear policy prescription as to whether “detracking” America's schools will lead to gains or losses for all, some, or even any students.

Based on the evidence presented in Section 2, we believe that the estimated effects of ability grouping on inequality in test scores in some of the earlier literature is substantially overstated. At the same time, the six problems listed in Section 3raise serious questions about whether our earlier research, or any of the earlier research, has completely captured the true effects of ability grouping and tracking. Definitional problems (ability grouping versus curriculum tracking) are compounded by a series of measurement issues. No data set adequately deals with all of these problems. An optimal data set would include data on all students at a school, allowing one to measure both mean achievement and the dispersion of achievement within each classroom. It would also include information on how teachers grouped students within the classroom, and detailed information on curriculum differences between and within classrooms. In the absence of such a data set, perhaps our wisest conclusion is that we still do not fully understand what “detracking” schools would do.