referred (but at risk) on reading achievement, mobility (number of times a family moved prior to referral), and the number of times the child was late to school. The narrative interview data revealed that teachers agreed that the two groups of students performed similarly at the beginning of the school year and that the problem children were on their “watch” list for possible referral. Teachers described one group of referred students in terms of their having “given up and not making an attempt to learn,” while those they did not refer still seemed motivated to learn. Finally, students whose behavioral problems prompted referral were characterized as having a history of misbehavior, but a single episode (“the last straw”), a critical event, precipitated the referral. It should be noted that one-eighth of the teachers made two-thirds of the referrals, suggesting differences among teachers in their tolerance for low achievement or misbehavior.
What has been clear for many years is that the public schools do not screen with tests to “catch” children with disabilities. We have long recognized that many children presenting similar psychometric profiles (IQ scores below a cutoff score, IQ-achievement discrepancies) are never referred by classroom teachers or parents and are never at risk for special education identification. Few datasets illustrate this more clearly than that of Mercer (1970) in the Riverside Desegregation Study. In addition to her description of the stages that children went through at that time on the way to being labeled mentally retarded, she also surveyed classrooms from which referred students came. Individual tests of intelligence (WISC) were administered to all Spanish-surname (n = 509) and black (n = 289) children in three segregated elementary schools and a random sample of white (n = 500) students in predominantly white schools. None of the students assessed was identified as having a disability or referred for evaluation. At the time of this study, the upper cutoff score used in California for MR was 80. The percentage of each racial/ethnic group scoring below IQ 80 in these samples was as follows: white, 1.2 percent; black, 12.2 percent, and Spanish-surname, 15.3 percent. Were the schools to have screened with IQ tests, very few additional white students would be “caught,” but had IQ alone defined MR, substantially more black and Hispanic students would have been identified as mentally retarded, but were not. While Judge Peckham opined in his famous trial opinion (Larry P., 1979) that IQ was “primary and determinative” in the identification process for MR, clearly professional judgment of teachers in deciding which children to refer reduced the number of minority group children formally considered for eligibility, even at a time when minority overrepresentation was much higher than it is today.