Prior research has established the statistical significance of several predictors of disproportionate representation of children from certain minority groups in special education programs for students with mental retardation and emotional disturbance (e.g., Finn, 1982; Oswald, Coutinho, Best, & Singh, 1999). The variables investigated thus far can be classified into two broad categories: demographic and economic characteristics of children and/or school districts. Educators have little control over these variables. A significant weakness in the research is the exclusion of variables that are more directly related to special education eligibility decision making and to factors that are alterable through the efforts of educators. In this study, achievement variables were added to the demographic and economic variables used previously in an effort to better understand patterns of disproportionate representation of minority students in special education. Moreover, additional racial/ ethnic groups (i.e., American Indian, Asian/Pacific Islander, and Latino) were added in this analysis as well as the disability category of learning disabilities (LD), the most prevalent of all special education disabilities.
PATTERNS OF DISPROPORTIONATE REPRESENTATION
The disproportionate representation of minority students in special education has been a constant and consistent concern for nearly 4 decades. National patterns of disproportionality have been documented and demonstrated to be robust and steady over time. Using the Elementary and Secondary Schools Civil Rights Compliance Report of the U.S. Department of Education's Office of Civil Rights (OCR), the patterns have been documented every other year since 1968 (see Chinn & Hughes, 1987; Donovan & Cross, 2002). With each updated OCR survey, African American students are found to be overrepresented in the categories of mental retardation (MR) and emotional disturbance (ED); American Indian students are overrepresented in the learning disability (LD) category; Asian/Pacific Islander students are underrepresented in almost every category; and African American, Latino, and American Indian students are underrepresented in the gifted and talented (GT) category. In addition, in 1998, the Office of Special Education Programs began collecting these data and the patterns found in them corroborate those from the OCR survey (Donovan & Cross).
The consistency of the findings, despite variations in sampling procedure and more than 25 years of attention to the issue, demonstrates its importance and the urgency with which solutions are needed. Calls to shift the focus of research in this area away from documenting patterns toward taking action and developing solutions have been increasing (see Donovan & Cross, 2002; Heller, Holtzman, & Messick, 1982; Markowitz, Garcia, & Eichelberger, 1997; Serna, Forness, & Nielsen, 1998). Most of the solutions proposed address the provision of appropriate and preventive interventions for students who are experiencing difficulty in school and improving teacher training in working with students from culturally and linguistically diverse backgrounds.
Although solutions and interventions are certainly desirable, developing them based on the current body of research on disproportionate representation may be problematic. Most examinations of disproportionate representation focus on general patterns. What is often missing from the discussions, however, is an analysis of the variables that produce or are related to the overrepresentation patterns (MacMillan & Reschly, 1998). In order to develop effective and efficient strategies to address disproportionate representation, a deeper understanding of the variables that may affect it is needed. Some researchers (Artiles, Aguirre-Munoz, & Abedi, 1998; Coutinho & Oswald, 1998; Finn, 1982; Hosp & Reschly, 2002) have done this using various levels of data (e. …