Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
REPORT OF THE PANEL
Introduction: Disproportion in Special Education The overrepresentation of minorities in special education classes is a pressing and volatile issue, not only because of society's continuing con- cern with equality of opportunity and equity of treatment but also because of an increasing number of legal statutes and judicial precedents that have broadened entitlement to needed educational services. Unequal represen- tation in special education is not a new phenomenon. What is at issue is whether it constitutes an inequity, either new or long-standing. The con- troversies that surrounded the earliest programs for children considered unable to profit from regular instruction still dominate the field of special education today: Is there a harmful and enduring stigma associated with placement in special education classes? Is the quality of education in special classes adequate? Can special education students ever return to the regular classroom? Are the methods of assessment and assignment fair and unbiased? Recent legislation attempts to ensure that the benefits of special educa- tion programs are available to all who need them. Both Section 504 of the Rehabilitation Act of 1973 and the Education for All Handicapped Children Act of 1975 (P.L. 94-142) require the formal identification of children with handicapping conditions and the provision of appropriate educational services. At the same time, the equal protection clause of the Fourteenth Amendment and Title VI of the Civil Rights Act of 1964 pro- hibit the classification of persons in such a way that disproportionate harm including the harm of separateness accrues to members of a group identified by race, color, or national origin. The Office for Civil 3
4 REPORT OF THE PANEL Rights (OCR), having enforcement responsibilities under Title VI and Section 504 j routinely examines disproportion in special education and other programs by means of a biannual survey of the nation's school and school district enrollments. An immediate and primary concern of OCR, revealed by the survey data, is a persistent disproportion of minority children and males in classes for educable mentally retarded (EMR) stu- dents. The Panel on Selection and Placement of Students in Programs for the Mentally Retarded was established to aid OCR in identifying factors that account for this disproportion and in developing procedures for rem- edying the imbalance. The panel analyzed the data gathered by OCR through its Elementary and Secondary School Survey to document the nature and extent of dis- proportion in special education classes. The analysis accomplished three purposes: (1) it verified that the relative disproportions cited by OCR do indeed exist, documenting in the process the magnitude and distribution of minotity-white and sex differences in EMR rates; (2) it identified geo- graphic trends in racial and sex imbalances in EMR programs; and (3) it provided an examination of possible correlates of disproportion (e.g., the size and racial composition of a school district, the overall prevalence of EMR classifications in a district, and the desegregation status of a dis- trict) as well as an appraisal of minority-white and sex differences for special education programs other than EMR and for individual racial or ethnic categories. By disaggregating the survey data to the district level, this analysis provided a detailed picture of the disproportion by race or ethnicity, by sex, and by special education classification. The next section of this chapter provides a summary of the results of the panel's statistical investigation; a detailed examination of these analyses is presented in the paper by Finn in this volume. Having confirmed that EMR disproportion is a nationwide phenomenon and that there are clear geographic and demographic conditions under which it occurs to a greater extent, the panel considered a long list of pos- sible "causes." These include characteristics of the legal and administrative systems within which special education programs operate, characteristics of the instruction and of the instructional setting, characteristics of the students themselves as well as possible biases in their assessments, charac- teristics of the students' homes and family environments, and the broader historical and cultural contexts in which they are embedded. It seemed likely that if we could identify the probable causes of dispro- portion, we could then determine effective solutions. However, the panel recognized that disproportion is very probably determined by multiple in- teracting factors that are inextricably confounded in any concrete in- stance. To focus on identifying causes, especially with the hopes of cor
Introduction: Disproportion in Special Education 5 recting or eliminating them to directly reduce disproportion, was deemed insufficient and unfruitful. Furthermore, to continue to focus on factors associated with disproportionate placement rates unduly emphasizes sta- tistical differences that are simply symptomatic of other, more significant issues. Altering placement rates and reducing disproportion in EMR pro- grams may remedy one set of problems e.g., the immediate problem of racial imbalance but it does not attend to the fundamental educational problems that underlie student placement in programs for mentally retarded students. Rather than continuing to explore plausible explana- tions or underlying causes, the panel focused on recurring dimensions of the problem, common to a variety of causes, that cut through the issues in potentially powerful ways. Accordingly, we recast the issue of existing disproportionality by asking why the overrepresentation of minorities in EMR programs is perceived as a problem. The controversy has typically centered on two assumptions. First, it is claimed that assessment procedures may lead to inappropriate placement and services for certain children, especially blacks, who are not really "mentally retarded." This debate has traditionally focused on the use of IQ scores to place children in EMR programs. The second assump- tion, directly related to the first, concerns perceptions of the EMR pro- grams themselves. EMR classes are often perceived as programs offering few valid educational services, channeling students into tracks that im- pede their return to regular programs while isolating them from their reg- ular classroom peers. These negative views of the services offered in EMR classes are in marked contrast to the more positive perceptions of other programs designed to provide special services. For example, the signifi- cant overrepresentation of minorities in Title I programs has not been contested in major court cases, presumably because such children are perceived as obtaining effective remedial services designed to help them achieve the levels attained by their regular classroom peers. Prom this perspective, the key issue is not disproportionality per se but rather the validity of referral and assessment procedures and the quality of instruction received, whether in the regular classroom or in special educa- tion settings. If needed and effective educational services are provided in the least restrictive environment to students validly targeted, then any re- sulting inequality in minority representation in those programs would not constitute an inequity. Emphasizing the validity of referral and assess- ment procedures and the quality of special education programs and out- comes is consistent with legal tenets since all four major laws stipulating entitlements to special education services focus on consequences, either in terms of harm to be avoided or in the types and quality of services to be provided.
6 THE EXTENT OF EMR DISPROPORTION IN AMERICAN PUBLIC SCHOOLS REPORT OF THE PANEL The panel sought to describe the magnitude of disproportion in EMR pro- grams by race or ethnicity and by sex. The survey data collected biannu- ally by OCR were used for this purpose. Although inferences concerning the processes that lead to disproportion and the appropriateness or valid- ity of special education placements cannot be drawn from these data, they do illuminate important differences in placement rates and the context in which these differences arise. Most striking in the description is the ex- treme variability in the magnitude of disproportion; these differences are attributable to ethnic group membership, to geographic region, to specific demographic characteristics of districts, and to handicapping condition. The 1978 OCR survey sampled 6,040 school districts including 54,082 schools, about one third of the districts in the nation. Questionnaires were sent to all district offices and to each school, requesting counts of the total number of students enrolled, the number enrolled in special educa- tion programs, and additional global characteristics of the student popu- lation. All student counts were classified by racial or ethnic identity,2 and some were also classified by sex. Both sex and race classifications were re- quired (but not sex-by-race cross classifications) for students in special education programs for educable mentally retarded, trainable mentally retarded, seriously emotionally disturbed, specific learning-disabled, and Details of the sampling design for 1978 are given in U.S. Department of Health, Education, and Welfare (1978a,b). The survey depends for its accuracy on an adequate count and report from numerous school districts and thus may be subject to some unknown degree of error. (This issue is discussed further in the paper by Finn in this volume.) 2According to the general instructions for the fall 1978 school survey (Form OS/CR 102), the following racial or ethnic categories are identified: American Indian or Alaskan native: A person having origins in any of the original peoples of North America and who maintains cultural identification through tribal affiliation or community recognition. Asian or Pacific Islander: A person having origins in any of the original peoples of the Far East, Southeast Asia, the Pacific Islands, or the Indian subcontinent. This area includes, for example, China, India, Japan, Korea, the Philippine Islands, and Samoa. Hispanic: A person of Mexican, Puerto Rican, Cuban, Central or South American, or other Spanish culture or origin-regardless of race. Black, not of Hispanic origin: A person having origins in any of the black racial groups of Africa. White not of Hispanic origin: A person having origins in any of the original peoples of Europe, North Africa, or the Middle East.
Introduction: Disproportion in Special Education 7 speech-impaired children, as defined by the Office of Special Education and adapted by OCR.3 For purposes of correlating the degree of disproportion with other school-related characteristics, a sensitive "log-odds index" of dispropor- tion was calculated for each special education category.4 The index is positive whenever the odds of minorities being assigned to a special pro- gram is higher than the odds for whites; it is zero if the odds for minorities 3According to the general instructions to the fall 1978 school survey, the following special programs are identified: Educable mentally retarded (or handicapped)-a condition of mental retardation which includes pupils who are educable in the academic, social, and occupational areas even though moderate supervision may be necessary. Trainable mentally retarded (or handicapped)-a condition of mental retardation which includes pupils who are capable of only very limited meaningful achievement in the tradi- tional basic academic skills but who are capable of profiting from programs of training in self-care and simple job or vocational skills. Seriously emotionally disturbed-a condition exhibiting one or more of the following char- acteristics over a long period of time and to a marked degree, which adversely affects educa- tional performance: an inability to learn which cannot be explained by intellectual, sensory, or health factors; an inability to build or maintain satisfactory interpersonal relationships with peers and teachers; inappropriate types of behavior or feelings under normal cir- cumstances; a general pervasive mood of unhappiness or depression; or a tendency to develop physical symptoms or fears associated with personal or school problems. The term includes children who are schizophrenic or autistic. The term does not include children who are socially maladjusted, unless it is determined that they are seriously emotionally dis- turbed. Specific learning disability-a disorder in one or more of the basic psychological processes involved in understanding or in using language, spoken or written, which may manifest itself in an imperfect ability to listen, think, speak, read, write, spell, or to do mathematical calculations. The term includes such conditions as perceptual handicaps, brain injury, minimal brain dysfunction, dyslexia, and developmental aphasia. The term does not include children who have learning problems which are primarily the result of visual, hearing, or motor handicaps; of mental retardation; or of environmental, cultural, or economic disad- vantage. Speech-impaired-a communication disorder, such as stuttering, impaired articulation, a language impairment, or a voice impairment, which adversely affects a child's educational performance. 4The basic element in the log-odds index is the "odds" of being assigned to a particular special education category. For example, the odds of a minority student being assigned to an EMR class is the percentage of minority students classified as EMR divided by the percent- age of minorities who are not in special programs. From Table 1, this is 2.54/92.60, or 0.027. The odds of a white student being classified as EMR is 1.06/94.12, or 0.011. The disproportion index is the ratio of these two odds, scaled by a natural logarithm transforma- tion; that is, In(0.027/0.011) = 0.89. The unscaled odds ratio ranges from zero to infinity; values greater than unity indicate that the EMR odds for minorities is higher than those for
8 REPORT OF THE PANEL and whites is equal; and it is negative if the odds of minorities being assigned to special education classes is lower than the odds for whites. The log-odds index is a linear contrast of the logarithms of the two odds and has a distribution in the population of school districts that closely approx- imates the normal; thus, it is particularly appropriate for analysis by normal-theory methods, e.g., Pearson correlations or analysis of variance. Unfortunately, the index is not simple to interpret since it is unbounded, i.e., it can vary from-oo to +oo, depending on the magnitude of the disproportion. For interpretive purposes, however, the log-odds index can be transformed to a measure of association, Yule's C-statistic, which, like a correlation, is limited to values between-1 and +1.s For EMR pro- grams, the association of race or ethnicity (minority versus nonminority) with placement (EMR versus none) is approximately +.42. Although of some general interest, national aggregate indexes do not provide adequate means to describe the pattern of disproportional enroll- ment in special education classes. Disaggregation is particularly impor- tant since students are placed in special education programs on a district- by-district basis; hence, a wide range of placement rates and racial disproportions may be found among districts operating within the same state guidelines. The OCR survey provides data from which placement rates and the disproportion index may be calculated for each school district. State summary statistics can be obtained by averaging the log- odds measure across districts of similar size, with dispersion measures (e.g., the range or standard deviation) providing an indicator of variability within the larger unit. Such disaggregation prevents results for large districts from obscuring those for smaller districts. Moreover, districts with no students in a particular special education program are eliminated from the respective analyses and thus do not distort summary statistics. The 1978 OCR sample included 4,917 districts with both minority and white students enrolled in EMR programs; these districts provide the data base for statistical analyses of EMR disproportion. Nationwide percentages of students in each of the five special education whites, while values less than unity indicate that the EMR odds for minorities is lower than those for whites. The logarithmic transformation creates an index that is symmetric around zero, ranging from-oo to Too. Furthermore, the log-odds ratio is equivalent to the dif ference between the logarithms of the two odds-i.e., In(0.027/0.011) = In(0.027) - In(0.001)-and the transformation to a logarithmic scale produces linear contrasts. For fur ther information, see Bishop et al., 1975. sThe relationship is given by Q = (a-1)/(a + 1), where a = ex and x is the log-odds index.
Introduction: Disproportion in Special Education 9 program areas, as estimated from the survey data, are given in Table 16; more refined breakdowns are presented in Tables 2 through 5. Despite the fact that a race or ethnicity EMR disproportion appears from Table 1 to be a national phenomenon the average percentage of minority students in EMR classes exceeds the average percentage of whites in every state except four7 massive regional variation in minority representation is evident in the survey data (see the paper by Finn in this volume for a breakdown by state). The average disproportion in southern states (Table 2) is consistently and notably high. Among districts in the South, the median disproportion index is 1.50, which corresponds to an association (Q) of .63. Although substantially lower than in the South, relatively high minority disproportion also pervades the data for the states bordering the South; the median log-odds disproportion value is 0.66, cor- responding to a Q value of .32. Minority disproportion does not appear as a general problem in the Northeast or the Midwest, where the correspond- ing measure of association in each region is .03. Minority disproportion in the West is also relatively low; the association (Q) of race or ethnicity with EMR placement is .17. Dramatic differences in minimum and maximum percentages of minor- ities assigned to EMR classes are also evident in summary regional data (Table 3~. Again, the South exhibits the highest minimum and maximum average EMR placement rates for minority students of any geographic region up to an average of 9.09 percent of minorities enrolled in EMR classes in Alabama. The northeastern and midwestern states show a lower range for minority placement than does the South. At the low extreme, the range of placements for minorities in the West is similar to the relatively homogeneous range for whites throughout the country. In addition, there is a regional tendency for larger disproportions to occur in areas in which the total proportion of children in EMR classes is high. This effect also operates at both the state and district levels. The data indicate that, in general, smaller degrees of disproportion occur in districts, states, and regions that have smaller proportions of students in EMR programs. The average level of racial disproportion in EMR programs is smallest for districts with 1,000-3,000 students. It is somewhat higher for districts with fewer than 1,000 students, higher for districts in the 3,000-10,000 student range, and highest for districts with more than 30,000 students. 6The figures in Table 1 are based on projections to state and national totals obtained by weighting each district in the sample by the inverse of its sampling probability. Details of the procedure are given in the 1976 survey Final File Documentation (U.S. Department of Health, Education, and Welfare, 197Sa). 7The exceptions, New Hampshire, Vermont, West Virginia, and Iowa, have very small per- centages of minority students.
10 ._ C) ._ s~ o C) Ct s: Ct x a, U. C) Ct - o ._ ~_ Ct C) C~ ._ C) V] ._ 3 s: o ._ Ct z x V, ._ C~ . _ ~: ~i s~ o C} e~ 5 = ~4 ~0 C~ 50 oor~C~ __ ..... oO - ..... OO-OO CD - C~ S: C) Ct O - ~_ ~U) Lr ..... . - O O - - U~ CS~ oo ~O O r-1 ~c ~0 . . . . . . _ o O oo _ ~ _ o O 1 - ._ 5 ;- -o s~ oo ~ ~ O g ~ ~ O O O O O .~ e ~;- s~ . o ._ C5 C) ._ ._ Ct - ~ ~2 .= - - ~c~a~0c~c~ o- ~o- ...... -ooc ~ o~ ~o ~oo~ ~oo ~-~ cr :^ - - - < Y-Y ~ a :A.= ~, ~ ~ D ~ D ~ ~ ~ .C~ D ~o ~ - o C~ . ~ _ ~ o - oo C ;: - ~ - - ·_ o4 ~: Cq ~ O O ~ ~ O _ ~ C~ .0 Ct o ~- o o ._ o ~ ~: ·~ Cq C~ O ~0 ·- C .. ~ Z ~ - V' .= C~ ~7 ~n ~: Ct Ct - :^ s~ Ct
Introduction: Disproportion in Special Education TABLE 2 Regional Summary of EMR Disproportion Race or Ethnicity Sex Number Median Median Region of Statesa Log-Odds Q Log-Odds Q Northeast 9 0.06 .03 0.30 .15 Border 6 0.66 .32 0.50 .24 South 11 1.50 .63 0.51 .25 Midwest 11 0.055 .03 0.38 .19 West 12 0.34 .17 0.35 .17 a Hawaii and the District of Columbia, each with only one school district, are not included. 11 The relation of disproportion to the percentage of minority students in a district is not the same for smaller and larger districts. In districts of all sizes, there is an increase from small or nonexistent average disproportion to moderate disproportion as minority enrollment increases from O to 50 percent. In medium and large districts, as the minority enrollment in- creases from SO to 90 percent or more, racial disproportion in EMR pro- grams decreases to close to zero. Among small districts, by contrast, those with 50 percent minority enrollment or greater have still larger dispropor- tions (see Figure 1 in the paper by Finn in this volume). These may involve a significant number of children at a statewide or regional level. Nationwide placement percentages are presented in Table 4 for five spe- cific racial or ethnic groups in each of the five types of special education programs. As is to be expected, since blacks represent approximately two TABLE 3 Minimum and Maximum Average EMR Percentages, by Region Number Minority White Region of StatesU Minimum Maximum Minimum Maximum Northeast 4 1.83 3.35 0.71 1.60 Border 4 2.54 5.20 0.70 2.41 South 11 3.60 9.09 0.84 2.23 Midwest 5 1.57 5.42 1.07 2.46 West 7 0.85 2.51 0.59 1.17 a For 31 states with more than 10 percent minority enrollment each.
12 o C~ ._ ~: ~S ._ C) Ct C) ._ ._ C~ s~ o C) Ct - ~: o ._ Ct C~ - Ct ._ C) V) ._ 3 o ._ Ct z ~2 ~:8 ~ ~ ~ - . ~ ~ ~ ~ C~ _ ~ oO _ o o ~ ~ ._ Ct a' C~ ._ ~: C~ ._ Ct - _ o c, . - ._ ~ .= C~ ~ C~ ~ :L s~ o ~: .= C) 5 .> ~ - _ Ct Ct .o e~ s~ C~ ~ ~ _ ~ r~ 0 . . . . . _ 0 0 r~ O r~ ~ r~ ~ ~ oO . . . . . ~ O O ~ _ Ir)oo ~ a~ oo oO r~ ¢~ ~ u~ r~ .. . . . . ~O O o ~ _ C-11- ~) O ~ u-) _ _ ~ oO .. . . . . -O O O - - c~ ~ ~ a~ ~ oo .. . . . . O- O O ~ - . . £ ~ _ ~ o~i ~ ._~ _ _ _ 5- (~S s~ ~C~ ^, ~ ~C~ - - s.= ~ ~ ) elo- ~ ~ ~- ct c: c) ~ .= ~ s~ ~ ~ - CL - ~ - u) ~ ~ u) 2 - s~ C~ 5 - .0 o :^ - .0 U) ~7 C) ._ - ._ Ct - ~ - ._ ~ _ ~0 ~ ._ ~ ~ ._ s~ ~ c~ CL - ._ C) ~ ·- .~: C) C) C) U) V, ~; o ~ . C~ C~ ._ Ct V, ._ o ._ o o V, ._ o o ._ ._ 5 - C~ ._ U~ E~ +++ + 0- 0 oOC~ . . . . . _ ~ ~ ~r~ . . . . . r~ ~ ~ r~ O r~ a ~c~ . . . . . O r~ _ u ~O C~ _ _ ~ . . . . . 1 1 1 1 1 s~ O Ct ·- ~ et ~ ._ C~ ~ U, ~ C~ o s~ ~ 5) ·O ~ ~ _ _ ._ Z ~ o 0C Ct ._ U) _ ~ ~ r~ oo ~_ . . . .. _ o O O- oc ~ ~ U-)~ O ~ ~ ~ . . . .. _ O O OO - 1 - 1 ~) r~ _ ~ ~ _ V' ~ ~ S o g V, ~ - ~ cr a~ 0 _ , ~ O O ;> 8 8 ~V ° ° _ C~ ~> o :-C) o5- o 8 ~ ~_ ~:
Introduction: Disproportion in Special Education 13 thirds of the minority enrollment in the country's public schools, the pat- tern of black enrollment in each of the special program areas closely par- allels that for the total minority population, as shown in Table 1. Each of the other minority groups identified in the survey is character- ized by some idiosyncratic discrepancies from the total minority group re- sults. For example, students of Asian or Pacific Island origin are typically assigned to special education programs at rates that are considerably below those for whites. In small districts in several western states, how- ever, positive disproportions are found that might reflect a relatively high incidence of recent immigrations. Verification of this hypothesis was not possible from the survey data. Although there is a tendency for American Indian or Alaskan native students to be assigned more frequently than white students to EMR programs, the OCR survey may not provide an adequate data base for evaluating the extent of disproportion, since rela- tively large numbers of American Indians are enrolled in schools or pro- grams outside those sampled by OCR. Despite the fact that the nationwide summary statistics indicate that the proportion of Hispanic pupils enrolled in EMR classes is slightly below that for whites (Table 4), the reverse situation is true in 26 of 31 states reporting 10 percent or more total minority enrollment. To explore this apparent inconsistency, a subsample of school districts was selected in which Hispanic students comprise at least 5 percent of the total enroll- ment with at least 50 Hispanic students enrolled (see Table 5~. Of the 4,917 districts in the survey data, 765 met these criteria. For this subsam- ple, the average EMR disproportion is positive for each of the school dis- trict size intervals presented in Table 5. The striking aspect of the data in Table 5, however, is the broad range of log-odds indexes within each cate- gory of district size from large negative disproportions (many fewer His- panics than whites) to large positive disproportions (many more Hispanics than whites). Unlike disproportion for all minorities combined or for blacks in particular, the small Hispanic-white difference for the nation as a whole is an average of many sizable positive and negative dispropor- tions. Correlates of this phenomenon, including the districts' racial com- position and the availability of bilingual education, are discussed in the paper by Finn in this volume. The March 1980 Current Population Survey published by the U.S. Bureau of the Census estimates that 59.9 percent of the Hispanic population is of Mexican origin, 13.8 percent Puerto Rican, 6.3 percent Cuban, 7.7 percent Central and South American, and 12.3 percent of "other Spanish origin." While there may be noteworthy differences among these groups in school performance or factors affecting performance, research on educational programs fre- quently does not make such distinctions, and the OCR survey instruments obtain only total Hispanic counts. Furthermore, the subgroups are not geographically distinct; the census re- veals that sizable Hispanic populations in most states include two or more of these subgroups.
14 REPORT OF THE PANEL Unlike disproportion by race or ethnicity, the overrepresentation of males in EMR programs is relatively uniform across geographic regions (Table 21. As a consequence of this relative uniformity in the sex dispro- portion, this summary has little distinctive information to impart about the demography of male-female placement in special education classes. Nonetheless, it must be kept in mind that the problems we address con- cerning minorities apply to males as well. POTENTIAL CAUSES OF DISPROPORTION IN EMR PROGRAMS Although the magnitude of the minority-white and male-female dispro- portion in EMR placements rates and the systematic variation in EMR disproportionality as a function of geographic region and demographic characteristics can be clearly documented, the factors that account for this disproportion are less easily analyzed. The multiplicity of potential causes of disproportionate placement rates may be categorized for pur- poses of a brief overview under six main rubrics: 1. Legal and administrative requirements 2. Characteristics of students 3. Quality of the instruction received 4. Possible biases in the assessment process 5. Characteristics of the home and family environment 6. Broader historical and cultural contexts Each of these potential causes is described below. LEGAL AND ADMINISTRATIVE REQUIREMENTS Federal, state, and local legal and administrative requirements establish a network of incentives and constraints within which special education pro- grams operate. Definitions of particular diagnostic categories, policies adopted that establish a particular referral and evaluation system, and policies concerning the funding of special education programs affect which children are referred for special education, how they are evaluated and placed, and the types of services that are available in special educa- tion programs. Some of these factors may contribute to disproportionate placement of minorities in EMR programs. For example, funding schemes that directly tie the number of dollars made available to a special education program to the number of children in that program may en- courage overcounting, and minority children may be more likely to be eli- gible and therefore placed in expanded special education programs. The legal and regulatory structure for the identification, assessment, and
Introduction: Disproportion in Special Education 15 placement of students in special education programs and the fiscal factors that may influence these programs are discussed in more detail in Chapter 2 and in the paper by Magnetti in this volume on potential fiscal incentives. CHARACTERISTIC s OF STUDENTS A variety of causes for disproportionate placement have been proposed that focus directly on the characteristics of students. Students may ex- perience difficulty in school because of undiagnosed or untreated medical and physical problems (see the paper by Shonkoff in this volume); because of difficulties in information processing, comprehension, reasoning, or judgment; because of emotional or motivational disturbances, such as hy- peractivity or anxiety, that disrupt or block effective learning; and be- cause of the absence of adaptive skills and behaviors that are needed in school, etc. Learning deficiencies in the early grades may persist in later years and become barriers to future achievement. QUALITY OF THE INSTRUCTION RECEIVED An almost uniform feature of the selection process for EMR placement is that it begins with an observation of weak academic performance. Poor performance may be accompanied by other behaviors, such as disruptive classroom behavior, but referral for EMR placement seldom occurs in the absence of weak academic performance (see the paper by Bickel in this volume). To the extent that a greater proportion of minority children score below accepted norms on achievement measures used in particular schools, they will be overrepresented in the pool of "potential" special education children. While academic failure is often attributed to characteristics of the learners, current achievement also reflects the opportunities available to learn in school. If such opportunities have been lacking or if the quality of instruction offered varies across subgroups of the school-age population, then school failure and subsequent EMR referral and placement may rep- resent a lack of exposure to quality instruction for disadvantaged or minority children. POSSIBLE BIASES IN THE As SESSMENT PROCESS The measures employed in classification procedures for EMR placement may not yield valid assessments of the cognitive skills of particular minor- ity or disadvantaged groups. Much of the controversy regarding assess- ment has centered on mental ability tests from which IQ scores are de
16 REPORT OF THE PANEL rived. Frequently referred to as "IQ tests," these instruments play a primary role in the determination of eligibility for placement in EMR pro- grams (see the paper by Bickel in this volume). Critics charge that such tests underestimate the skills of minority children- that the items do not tap the same underlying construct for minority groups as for white middle-class children, that particular items are insensitive to minority cultures, that differences exist in the predictive validity of the test for dif- ferent groups. Furthermore, the test-taking situation may artificially depress the scores of minority children compared with those of whites. This position argues that there is a fundamental mismatch between the language and culture reflected in IQ tests and those of various minority groups. Any such mismatch could cause inferior performance on IQ tests by minorities, which in turn has profound implications for later educa- tional experiences, including an increased likelihood of EMR placement. These issues are discussed in detail in Chapter 3 and in the paper by Travers in this volume. In addition, it is possible that features of the placement process may contribute to overrepresentation of minorities in EMR programs. For example, minority students with academic problems may be referred for evaluation more often than other children experiencing similar academic difficulties. An analysis of the placement process and its contribution to minority overrepresentation in EMR programs is presented in the paper by Bickel in this volume. CHARACTERISTICS OF THE HOME AND FAMILY ENVIRONMENT Well-established relationships between parents' socioeconomic status and children's school performance have led to the investigation of variations in home environments and child-rearing styles as possible causes of low achievement among minority and disadvantaged children. Proposed dif- ferences in home environments include the extent to which motivational support is provided for cognitive achievement and the extent to which par- ents and others encourage verbal development and provide appropriate verbal models. Families may also differ dramatically in the degree to which children are encouraged or required to practice the use of complex systems of verbal symbols; the lack of such practice may be related both to the underdevelopment of cognitive skills and to an increased likelihood of EMR placement. BROADER HISTORICAL AND CULTURAL CONTEXTS As noted above, many of the proposed causes of disproportionate EMR placement are attributed to the student directly; so it is not surprising that
Introduction: Disproportion in Special Education 17 to date, studies of mental retardation have generally emphasized charac- teristics of the individual. The problem of disproportion can also be viewed in a broader sociocultural context not just the sociocultural influ- ences on individual students of their familial and street cultures but a per- vasive collective influence of minority status within a dominant majority culture. Discontinuities arise from the child's experiences as mediated by the family and home environment, especially when children from various subgroups are confronted with the curriculum and value structures of the public schools. Discontinuities may also arise from the collective historical confrontation and conflict between minority cultures and the dominant culture. This perspective emphasizes the importance of coping mecha- nisms and survival strategies that have developed in response to the long- term denial of equal opportunity, status, and rewards for minorities. From this analysis, possible societal causes of problems involving to edu- cability of minority children may be identified that in turn contribute to disproportionate EMR placement rates. DISPROPORTION: PROBLEMATIC OR SYMPTOMATIC? The panel agreed that disproportion undoubtedly reflects all of these causes singly and in combination in some school districts some of the time. It became apparent, however, that even if the multiplex causes of EMR disproportion could be identified and disentangled, it is unlikely that remedies could be easily or effectively implemented. Furthermore, an analysis that relies on eliminating the causes of disproportion presup- poses that effective solutions will result in a lack of disproportion in EMR programs. The assumption that effective practices are necessarily ones that reduce disproportion has led individual school districts, and in some cases entire states, to attempt simplistic solutions to the problem of dis- proportion, e.g., by eliminating part or all of the EMR program, by com- bining EMR classes with a program that has fewer nonminorities enrolled so that the overall racial enrollments are more balanced (see Table 4), or by prohibiting the use of IQ tests for EMR placement. Approaches such as these may be misdirected. Each is likely to result in increased disproportion elsewhere in the educational system in place- ment in other special education programs, in over-age grade placements, in disciplinary actions and dropout rates, or perhaps in the number of high school students who cannot read or perform simple numerical tasks proficiently or meet minimum competency standards at graduation. More significantly, such simplistic solutions fail to focus on the needs of the children or on the services that should be provided. Rather than inquiring about the causes of disproportion and how to remedy the problem of disproportion in special education and in EMR
18 REPORT OF THE PANEL classes in particular, a different and more constructive perspective is to ask: Under what circumstances does disproportion constitute a problem? While remedies to disproportion per se are based on an assumption that the disproportions in themselves constitute an inequity, the educational and social conditions under which such an assumption is true should be examined explicitly. Three aspects of the regular and special education programs and placement procedures are most salient in this regard: Dis- proportion is a problem (1) if children are invalidly placed in programs for mentally retarded students; (2) if they are unduly exposed to the likeli- hood of such placement by virtue of having received poor-quality regular instruction; or (3) if the quality and academic relevance of the special in- struction programs block students' educational progress, including de- creasing the likelihood of their return to the regular classroom. Disproportion is a problem if children are invalidly placed in programs for mentally retarded students. If children are systematically assigned to EMR classes when other settings or programs would be more beneficial, then the assessment system for special education is of questionable valid- ity, either for students in general or for particular subgroups that are over- identified. On the other hand, if the assessment system results in dispro- portion for particular subgroups, the assessments may still be successfully defended if their educational utility and relevance can be demonstrated. If not, the procedures should be changed to improve their validity and to lead more directly to appropriate and demonstrably effective educational practices. Prom this perspective, the panel's primary concern is with the validity of the assessment system and its implications for educational practice rather than with the resulting adverse disproportion as such. The validity of assessment practices for placement in EMR programs is inextricably tied to the meaning of the category itself. Educable mental retardation is at least in part a function of the social and educational demands on an individual. The category resists precise definition, allow- ing a wide variety of measurement practices to be employed in the schools. While federal regulations implementing P.L. 94-142 define mental retar- dation as "significantly subaverage general intellectual functioning exist- ing concurrently with deficits in adaptive behavior and manifested during the developmental period which adversely affects a child's educational performance," the translation of these guidelines into assessment prac- tices is neither direct nor uniform. Thus the category EMR is operational- ized in different ways at different times in different areas. For example, adaptive behavior ratings sometimes focusing on achievement-related behaviors and other times not play a variety of roles in special education assessment. In addition, IQ cutoff scores vary from district to district, and different cutoff scores may result in different proportions of students be
Introduction: Disproportion in Special Education 19 ing classified as EMR; a regular student in one district may be classified as mentally retarded in another. At the same time, the resulting category of EMR children is far from homogeneous. To the extent that the use of the label initiates a process of individual diagnosis, planning, and treat- ment, the lack of homogeneity of the category is not troublesome. The use of the EMR label becomes problematic, however, when it is presumed to imply common instructional interventions for children with a wide variety of educational needs or when it leads to inappropriate expectations for the performance of certain children within this diagnostic category. Moreover, the measures used to classify students as mentally retarded may not discriminate among groups of children who require or can profit from different educational settings or programs and hence may not be valid measures for the placements that result. Individually administered IQ tests are a major instrument used in the ultimate classification of re- ferred students. The fact that IQ scores predict a variety of school achieve- ments makes such tests appealing, and their high reliability gives the user confidence in the results. However, the predictive power of the IQ does not necessarily make it a good measure of mental processes; different pro- cesses may underlie the same IQ scores for different groups of children, and different types of remediation may be necessary in cases of poor per- formance. For example, it has frequently been argued that levels of moti- vation and effort of minority students are systematically different from those of white students. Similarly, language factors undoubtedly affect performance more for some groups than others. IQ tests administered en- tirely in English to students for whom English is a second language are an extreme case in point. Because of these and a host of other factors, there is no direct way to infer the source of a child's difficulty from incorrectly answered test items, nor does a test score or a profile of subscores provide the kinds of information needed to design an individualized curriculum for a child in academic difficulty. Furthermore, despite the mandates of federal laws and regulations, impre- cision and looseness in the referral, assessment, and placement systems cannot prevent discretion and personal bias from affecting placement decisions. After all, referral rests largely in the hands of the classroom teacher. If the teacher is distracted by the higher activity level of boys or feels uncomfortable in the presence of minority students, then those groups may be more likely to be referred for possible special placement. Similarly, the choice of assessment instruments and their interpretation remain largely in the domain of the school psychologist. Local discretion at many points in the placement process thus allows a wide range of fac- tors, some of which may be extraneous, to affect placement decisions. Disproportion is a problem if children are unduly exposed to the likeli
20 REPORT OF THE PANEL hood of EMR placement by being in schools or classes with poor-quality regular instruction. Students are referred for special education assessment primarily after they have experienced academic failure. However, children whose regular classroom instruction is poor may experience failure at a higher rate than they would if the quality of instruction were better. Since assessment instruments typically measure the outcomes of learning rather than learning processes, there is a danger that the child who has not learned because of poor instruction will be judged unable to learn from any instruction. The unequal distribution of quality instruction in large urban centers with high minority enrollments, compared with that in higher-income suburban areas, has long been a point of contention and debate. The well- established differences in the outcomes of schooling as a function of socioeconomic and racial or ethnic variation (see, for example, Coleman et al., 1966; Education Commission of the States, 1974) raise significant questions about the quality of instruction in schools serving children from low-income areas. This issue, in turn, has significant implications for the numbers of children who require special education services. Would fewer minority students be classified as mentally retarded if they were exposed to the highest quality instructional practices? Disproportion is a problem if the quality and academic relevance of in- struction in special classes block students educationalprogress including decreasing the likelihood of their return to the regular classroom. There has been much debate over the advantages and disadvantages of separate classes for children diagnosed as EMR. Proponents point to the advan- tages of smaller classes and more individualized instruction for EMR students. Critics argue that expectations for children classified as EMR are low and that behaviors in the classroom are adversely affected by these expectations. In addition, they charge that the EMR curriculum based on the assumption that educable mental retardation is a permanent and unremediable disability is not designed to help students learn the skills necessary to return to a regular instructional setting. Indeed, early concepts of mental retardation were explicit on this issue; Doll (1941) included both "constitutional origin" and "essentially incurable" among the necessary components of the definition of mental deficiency. However, early beliefs that intelligence is predetermined and fixed by genetic endowment have been replaced by the understanding that intelligence is not fixed at birth, that it can be modified through environmental manipulation, and that it partially reflects learned skills and behaviors (Hunt, 1961; Kirk, 1958~. Similarly, current professional definitions and views of mental retardation emphasize observed levels of functioning and behavior rather than perma- nent and unalterable biological conditions inherent in the individual.
Introduction: Disproportion in Special Education 21 Thus, a reasonable goal for many EMR students, children who are consid- ered only mildly mentally retarded, and especially those in the elementary school grades, may be to reenter the regular instructional program follow- ing the provision of effective remedial services. The question as to what constitutes quality instruction for students in special programs is complex, both because there is a variety of outcomes to consider (including the positive and negative effects on the special group, the positive and negative effects on the regular students, and the consequences for the regular-classroom teacher) and because EMR pro- grams frequently serve children with a wide mix of functional needs (in- cluding diverse combinations of cognitive disabilities and adaptive behav- ior problems). Research on the efficacy of EMR classes has generally focused on the effects of particular settings regular classes versus sepa- rate special education classes rather than on the characteristics of effec- tive instruction. Given that children in EMR programs have functional educational needs that are pressing and real, improved educational prac- tices depend on an appropriate match between instruction and each child's individual needs. A significant question also arises as to the mechanism by which special instruction may best be provided. In particular, to what extent must chil- dren be classified and labeled according to a generic class of deficiencies in order to receive special education services? Diagnostic categories such as EMR may be more an administrative convenience than an educational necessity, allowing schools to count the number of children in this and other special programs in accord with federal requirements. If categorical labels are required for administrative purposes, they could be chosen to reflect the educational services provided, thereby emphasizing the respon- sibilities of school systems rather than the failings of the child. A LOOK AHEAD The statistical phenomenon of different percentages of minority and white students in programs for mentally retarded students has a number of po- litical, scientific, and philosophical dimensions. While the sources of dis- proportion are legion, the more basic issues are educational. Dispropor- tion in EMR classes may be indicative of a significant inequity if children are invalidly placed in such programs, if poor instruction in the regular classroom increases the likelihood that certain children more than others will be referred or placed in EMR classes, or if EMR classes do not pro- vide instruction commensurate with the functional needs of the individ- ual. Thus, by focusing on the conditions under which the inequality of placement proportions signals inequity of treatment, two major educa
22 REPORT OF THE PANEL tional issues are highlighted: (1) the validity of referral, assessment, and placement procedures and (2) the quality of instruction received, whether in the regular classroom or in special education settings. These two critical issues are explored in detail in this report. Refocusing attention on the questions of validity and quality- i.e., the valid assessment of students' functional needs and the provision of high- quality, effective instruction has consequences affecting research and practice for students in special education and regular programs alike. If this new focus leads to the formulation of effective instructional pro- grams for individuals in the least restrictive environment, then the statistical issue of disproportion by race or ethnicity or by sex ceases to be a problem.