Nonbiased Assessment: A Need for Operationalism JAMES B. DUFFEY JOHN SALVIA JAMES TUCKER JAMES YSSELDYKE Abstract: The current technical history of nonbiased assessment is highlighted. Definitions of fairness and experts' attempts to alleviate problems associated with bios in assessment are reviewed. Factors relating to nonbiased assessment are emphasized that normally do not receive sufficient attention. Decisions currently made as a result of the assessment process are reviewed and an examination made of how decisions based on inappropriate evaluation can negate the validity of the assessment process. Finally, the utility of many recent efforts to resolve the problems of bias in assessment is questioned and it is suggested that very basic systematic changes are needed before real progress can be mode in reducing bias in assessment. JAMES B. DUFFEY is Director, National Learning Resource Center of Pennsylvania, Montgomery County Intermediate Unit, King of Prussia; JOHN SALVIA is Professor of Special Education, Pennsylvania State University, University Park; JAMES TUCKER is President, Educational Directions, Inc., Austin, Texas; and JAMES YSSELDYKE is Director, Institute for Research on Learning Disabilities, University of Minnesota, Minneapolis. • Comprehensive assessment is required before handicapped children or those with suspected handicapping conditions can receive special education services. However, educators are finding that children identified for special education do not reflect the cultural, economic, and racial composition of society as a whole (Tucker, 1980) and that assessment yields a disproportionate number of minority children. Accusations of racial, ethnic, and cultural bias have followed, with criticism leveled specifically at both the assessment process and the assessors. In the wake of such complaints, the judicial system has become heavily involved, and its findings seem to confirm the existence of discriminatory practices. As a result, the federal district court in northern California, for example, has recently permanently enjoined the use of standardized intelligence tests for the purpose of identifying Black school children for placement in educable mentally retarded classes (Larry P. v. Riles, 1979). Jensen's (1980) recent statements that Black children's low test scores are not the result of cultural bias in tests themselves may revive the controversy once again. Indeed, the concept of biased assessment does suggest images of prejudice, discrimination, and stereotyping. More precisely, biased assessment refers to a constant error in decisions, predictions, and inferences about members of particular groups (Cleary, Humphreys, Kendrick, & Wesmen, 1975). Bias can result in decisions that are too positive as well as too negative; for example, biased test results may Exceptional Children, Volume 47, Number 6. Copyright © 1981 The Council for Exceptional Children. Exceptional Children 427 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 cause some children to be placed in situations where they are likely to succeed. This paper examines the definitions of nonbiased assessment and offers possible solutions to the problem as suggested by the literature. A more pragmatic and perhaps more difficult solution is suggested. DEFINITIONS The literature is replete with information about bias in assessment. Unfortunately, much of this information lacks utility; proposed solutions have been only marginally successful. Attempts to define nonbiased assessment date back long before discriminatory assessment practices gained national attention through legislative and judicial decisions. Unfortunately, there still is little agreement on what constitutes nonbiased assessment. Cole (1973) and Petersen and Novick (1976) have described a number of models of fairness, including the following: The Quota Model-A test is unbiased if the criterion score predicted from the common regression line is consistently neither too high nor too low for members of a subgroup. The Equal Risk Model-A test is unbiased when persons with the same likelihood of success on the criterion task have equal chances of being selected. The Constant Ratio Model-A test is fair when it results in the same selection of proportions of different cultural groups as would be achieved if the person doing the selection had available each subject's exact score on the criterion measure. Thus, a test is fair if it selects the candidates who will be successful in meeting the eventual criteria for success. The Conditional Probability Model-"For both minority and majority groups whose members can achieve a satisfactory criterion score (Y - Yp), there should be the same probability of selection, regardless of group membership." (Cole, 1973. p. 240) Biased decision making may be attributed to incomplete and inconsistent operationalizing of definitions and/or criteria on which decisions are based. The inherent problem is that decisions are made on the basis of situation specific (and perhaps even assessor specific) definitions and rules. For example, mental retardation is not defined by observable behavior, but by intelligence and adaptive behavior, two constructs inferred from unspecified be- havior. Different tests of intelligence sample different behaviors and are imperfectly correlated (Salvia & Ysseldyke, 1978). Consequently, the terms "intelligence" and "adaptive behavior," which define mental retardation, are themselves poorly defined. Even less well defined are the criteria by which decisions about educational placement and special services are made. Without precision and consistency, extraneous variables can inappropriately influence decisions. The decisions facing psychologists, teachers, and administrators are usually complex and often must be made on the basis of ill defined criteria. For example, assigning grades, making referrals, classifying deviance, and recommending services all involve subjective decision making. In these nebulous situations, the preconceived attitudes of- the assessor can influence judgments made about the assessed. A fair amount of research has been devoted to the study of variables that may influence subjective decisions about achievement and the need for special services. Such variables may include: father's occupation, facial attractiveness, residual surgical scars, eligibility for special educational services, and diagnostic labels (Foster & Salvia, 1977; Salvia, Algozzine, & Sheare, 1977; Weierbach, 1977). OPINIONS AND SUGGESTIONS Many of those interested in unbiased assessment have expressed little hope for significant developments in the near future. Adler (1968) has written that tests themselves are cultural artifacts and that to ignore this fact may lead to erroneous assumptions about the meaning of test results. The very taking of tests, according to Adler, seems to be a part of the Western cultural milieu. If a test is used for educational or employment selection, and the educational or employment institution is part of the culture, then the test is going to reflect that culture. Cole (1975) noted that schools represent a culture that must be learned by the children who attend them. To be successful in school, children must deal with the culture represented by the school. Tests, including intelligence tests, measure past learning and predict future accomplishment only to a limited extent. Children who are culturally different reflect culturally different achievement when they take tests. To expect otherwise is folly. March 428 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 1981 Anastasi (1961) noted that "when we use tests to compare different groups, the only question the tests can answer directly is: 'How do these groups differ under existing cultural conditions?' This limitation of tests may appear so obvious as to be trite" (p. 389). However, the fact that this relationship between tests and culture is not so obvious is suggested by the evidence that American educators are deeply involved in a search for culture free measures as they seek to respond to the mandates of Public Law 94-142, the Education for All Handicapped Children Act of 1975. Considerable disparity exists between ways in which bias has been conceptualized and ways in which educators have addressed the issue of bias in assessment. Mercer (1977) has criticized efforts to engage in fair testing as characterized by the following kinds of proposals: Proposed Solutions Development of "culture free" tests. "Culture free" tests appear at first glance to be a plausible solution. Such efforts have met with failure because there is no such thing as culture free learning (Padilla & Garza, 1975). Learning, obviously, occurs in environmental contexts. Development of "culture fair" tests. Efforts have been made to construct tests in which items are "balanced" to represent multiple languages and cultures, but such efforts have been unsuccessful. "Culture fair" tests have not demonstrated good predictive validity (Salvia & Ysseldyke, 1978); they are poor predictors of success in a monocultural school system and have been rejected, for the most part, by the dominant cultural group. Development of "culture specific" tests. As noted earlier, efforts have been made toward developing "culture specific" tests (specific to the Black culture, Hispanic culture, etc.) such as the BITCH Test (Williams, 1972). These tests, like culture fair tests, have demonstrated low predictive validity and have been rejected by the dominant cultural group. In addition, developing such tests has proved difficult because no one test will satisfactorily assess the heterogeneous group of children in anyone cultural or ethnic group. Use of Piagetian tests. The use of Piagetian developmental scales has also been studied (Meyer & Elkind, 1975). Such efforts have not produced fair assessment because specific items are as culturally dependent as items on more traditional scales, and predictive validity is low. Linguistic translation of existing tests. Efforts to administer tests in a child's native language have often consisted of translating existing tests into other languages (Alzate, 1978). Translation changes item difficulty and destroys the applicability of existing norms, which, unfortunately, are too often used, despite item translation. Once again, predictive validity for success in a monocultural school system has been low. Alteration of administration procedures. Assessors often try to achieve fairness by changing administration procedures in testing handicapped youngsters. Verbal tests, for example, are given in sign language to deaf and hearing impaired children. Nonstandardized administration procedures disallow the use of existing norms and, unless social population norms are constructed, norm referenced interpretations are invalid (Salvia & Ysseldyke, 1978). Training children to take tests. Some researchers and practitioners have advocated that children be trained to take tests (Duffy, 1974; Oakland, 1972). Specific procedures have ranged from teaching them to become "test wise" to task familiarization training. The procedure is time consuming, but it may be a worthwhile alternative to traditional assessment procedures. It will not, however, reduce many aspects of bias in decision making. Use of pluralistic norms for existing tests. Mercer and Lewis (1978) developed a System of Multicultural Pluralistic Assessment (SOMPA), which uses existing tests but pluralistic norms. Separate regression equations are used to compute the estimated learning potential of children from Black, Anglo, and Chicano groups. One difficulty with such a procedure lies in accounting for the extremely heterogeneous nature of anyone cultural or ethnic group. Another problem may be predictive validity for success in the majority culture. Criterion referenced assessment. A parallel between job placement and educational placement tests is suggested by Siegel and Leahy (1974). They suggest that tests used for job se- Exceptional Children 429 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 lection should focus on tasks needed on the job. Persons demonstrating the ability to learn a sample of the tasks required on a specific job should qualify for that job. This implies that other factors should be eliminated from tests used in the job selection process. Similar logic could be used in the area of educational selection. Such tests would attempt to measure only the mastery level of children in various content areas and would leave the comparison of children to the standardized testers. Criterion referenced tests offer more utility in the day to day instructional setting and can be helpful in the selection process because they measure mastery of job required tasks. Again, before jumping at this possibility, educators should carefully consider the personnel and dollar resources necessary to implement a criterion referenced mastery system. Learning potential assessment device (LPAD). Feuerstein (1979) and others have proposed assessing an individual's learning potential rather than his or her store of information and existing capacities, factors traditionally tapped by intelligence tests. Specifically, Feuerstein sees a person's potential for modification in a focused learning task as a fairer means of assessment since it may reduce the effects of previous deprivation when compared with traditional testing techniques. There is, however, no reason to believe that potential for modification is any less subject to cultural influences than other constructs we typically attempt to measure. If learning potential assessment devices were to be used as we currently use traditional assessment devices to label, provide placement, and finance programs, they would have the same potential for bias. Testing moratoriums. The NAACP, at its 1974 meeting, proposed a moratorium on standardized testing whenever such tests had not been corrected for cultural bias. A similar position was taken by the CEC Delegate Assembly at its 1978 international convention (1978). Such a position, if generally adopted, would disqualify most tests for general use. These statements are tantamount to calling for an across-theboard cessation of standardized testing. According to Flaugher (1974), many advocates of testing worry about the consequences of not testing, arguing that the use of subjec- tive impressions for selection will make educational and employment institutions less accountable for their selection procedures. Linden, Linden, and Bodine (1974) feel that until tests are improved or replaced, they represent our best objective sources of information for selection purposes. Padilla and Garza (1975) point out that educators will continue to need instruments and techniques for assessing the academic progress of children. But they suggest that they be used for diagnosis, the purpose for which they were originally invented. Padilla and Garza see most abuses of standardized tests arising from the selection process, rather than the diagnostic process. Unfortunately, it is often difficult to distinguish between the two. Special educators must increase th;~ir understanding of the nature of assessment and the importance of technical adequacy in testing. They should consider more carefully the types of decisions made as the result of assessment. Perhaps the entire classification and placement process in special education should be reconsidered and more attention paid to the implications of acculturation. A brief examination of each of these topics follows. ASSESSMENT Assesment in special education is broadly defined as a process of collecting data for the purpose of helping professionals make decisions about individuals. Assessment is not synonymous with testing; testing is simply one part of assessment. Assessment may include direct observation ~f individuals in natural environments, obtaining data from others by means of interviews, and obtaining historical and current information by reviewing records. Clearly, many different kinds of data may be collected in the process of decision making; in its broadest sense, this data collection process is assessment. According to Cromwell, Blashfield, and Strauss (1975), a complete diagnostic-intervention process consists of historical-etiological information, currently assessable characteristics, specific treatments or interventions, and a particular prognosis-all elements required in one form or another by Public Law 94-142. To attain such comprehensiveness, a combination of assessment techniques or models must be used. There is nothing new March 1981 430 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 about this process or the techniques involved. Nor is the process static; it can be adapted to fit local needs. The importance of including parents as an integral part of ongoing assessment deserves special mention. Schools should not presume to make decisions about special placement and programing, other than that provided to every child in a given school system, without involving the parents in the deliberations leading up to the decision, as well as in the decision itself. Parents are important sources for determining cultural factors in assessment and placement decisions. Assessment is also a continuous process, since conditions surrounding a child are always changing. Decisions regarding a pupil's placement or educational programing should always be considered tentative and subject to constant review, as well as to periodic, comprehensive reassessment and possible revision. Technical Adequacy When assessment devices are used to make important decisions about pupils, it is imperative that those devices be technically adequate. Unfortunately, technical adequacy is often ignored by professionals, despite the training generally provided in university programs. Questions of bias may not even be relevant if technical adequacy has not been obtained. Several factors must be considered. The first is the issue of standardization. Scores earned on norm referenced tests reflect the pupil's performance relative to performance of those on whom the test was standardized. If a test is not properly standardized for a particular student, such comparisons will be meaningless. A second consideration is that of reliability. Reliability refers to the consistency with which a device measures a trait or set of behaviors. Reliability indices show how much error is present in any measure. In assessment, a person is interested in obtaining results that adequately reflect student traits, characteristics, or behaviors. Results should be as free from error as possible. Nunnally (1967) provides a yardstick for reliability. He notes that tests used in experimentation must have reliabilities exceeding .50, but tests used to make important decisions about pupils should have re- liabilities exceeding .90. It has been demonstrated that many of the norm referenced devices typically used to make decisions about pupils lack necessary reliability (Salvia & Ysseldyke, 1978; Ysseldyke & Salvia, 1974). When unreliable tests are used to provide data for decision making, decisions may be based more on error than on actual characteristics. The third issue in considering technical adequacy involves validity of the devices and procedures used to collect data about children. Tests must be valid for the particular purposes for which they are used. Assuming that tests measure what they purport to measure, when there is little empirical evidence to substantiate this, can lead to bias in decision making. Types of Decisions Educators need different kinds of data to help make different kinds of decisions. Abuse in assessment can occur when educators select their decision making procedures and devices without regard to the types of decisions necessary; for example, when they use tests for purposes other than those for which they were designed. For example, norm referenced, individually administered intelligence tests were originally designed to help make classification and placement decisions. They are helpful tools in predicting future success. However, the routine practice of engaging in profile analyses of subtest scores earned on norm referenced tests is often observed. Such efforts, as discussed by Kaufman (1979), are made to identify specific ability strengths and weaknesses and to provide data for use in instructional planning. This practice is without empirical support and may actually constitute abuse (Salvia & Ysseldyke, 1978). ASSessment plays a key role in making educational decisions and should provide the necessary information. Salvia and Ysseldyke (1978) identified five different kinds of decisions made in educational settings: 1. Screening. In screening, data are collected to help professionals identify the extent to which a student's behavior differs from "normal" or "average" behavior. Students whose behavior sufficiently deviates from the "normal" are typically identified as candidates for further assessment. Exceptional Children 431 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 2. Placement/classification decisions. Assessment data are routinely collected in educational settings to help professionals decide how to classify students, to declare children eligible for special educational services, and to aid in making placement decisions. 3. Instructional planning decisions. Assessment data are routinely collected to help educational personnel plan instructional interventions for children. Specific efforts are made to identify an individual's educationally relevant strengths and weaknesses, and to plan precisely what to teach and how to teach. Data collected during assessment serves as the basis for planning both long term educational goals and specific instructional objectives. 4. Individual pupil evaluation. Teachers, parents, and students themselves have a right as well as a need to know how a pupil is progressing in his or her educational program. Assessment data enables decision makers to judge the progress being made, both in the achievement of specific instructional objectives and in reference to a local or national sample of chronological or grade level peers. 5. Program evaluation decisions. Data are collected for the purpose of evaluating the effectiveness of educational programs. Typically, the progress of students in two or more programs is compared, or the attainment of program objectives is checked. Only when a placement category carries a cultural or social stigma do the difficulties associated with biased placement emerge. The categories of "mental retardation," "emotional disturbance," "learning disabilities," and, to some extent, "minimal brain dysfunction," give rise to the greatest problems in biased placement, and thus, in biased assessment. These categories create problems because they are based on social norms for ill defined symptoms which appear as much because of normal culture diversity as because of some handicapping condition. It is often impossible to tell the difference between the two. It is sometimes erroneously assumed that placement in a special education class is equal to remedial programing and individualized instruction. While children are placed in self contained special education classes under the assumption that better education will occur, there is little evidence of significant improvement in skill areas, especially when more mildly handicapping conditions are involved. In fact, the social and academic isolation that often accompanies such placement may have a negative effect on the child, as suggested by Sindelar and Deno (1978) in their positive review of resource rooms. It is quite possible, with.present technology and expertise, to develop an educational program that provides for the needs of all children without categorizing any of them beyond their ability to perform various skills. Errors have been made in the past by using assessment data collected for one purpose as a vehicle to meet other assessment needs. As a result, decisions have been made that are both invalid and biased. Acculturation Placement Decisions and Testing Special educators are caught in a dilemma. On one hand, they want to guard against biased labeling of children; on the other hand, they are aware that funds will not be allocated for use with handicapped children unless the equitable expenditure of funds for a specified population is assured. They want to provide specific educational programs to meet the identified needs of children, but when special education is the only instructional situation offering such assistance, they are hesitant to approve placement because of the inherent difficulties that may be involved. Any child's performance on a test is a reflection of past learning in both formal and informal educational environments. As noted by Salvia and Ysseldyke (1978), acculturation is the most important characteristic in evaluating a child's performance on a test. To the extent that a child's acculturation differs from the acculturation of those on whom a test is standardized, norm referenced decisions, based upon test results, may actually be both invalid and biased. When norm referenced tests are used to make decisions about students, assessors must examine the extent to which the assessed student is similar to those on whom the test is standardized. This is especially true when a child exhibits one or more specific handicapping conditions (e.g., deafness and blindness, or cerebral palsy). 432 March 1981 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 CONCLUSIONS The options are clear: Either eliminate condescending labels and categories or dramatically improve the quality of assessment occurring before a child is placed. Assuming that children have been segregated on the basis of biased testing procedures, educators have sought a magic formula to eliminate bias from tests used to place children in special education classes with an implied social stigma. They have not been successful in this undertaking. There is far more evidence that the use of test data has been the biasing factor rather than the tests themselves. Test data are no more valid than the professional judgments made in applying them. The problem of bias in assessment will not be resolved until educators can operationally specify the criteria to be used in decision making. Until this time, attempts to eliminate bias will be as effective as applying a bandaid to a hemorrhage. If the needs of special education children are to be met, the symptoms of handicapping conditions should be stated in operationally defined terms and applied equally to all candidates for special services. Such a move implies the elimination of most, if not all, categories now classified as "mildly handicapping conditions," leaving only severely handicapped categories. This would eliminate many problems identified with biased assessment. Few people object to the more severe labels, probably because any stigma involved in a severely handicapping condition is inherent in the severity of the condition itself and is not a function of the label applied. Education would be dramatically improved by taking the very difficult and possibly utopian step of abandoning as irrelevant such venerable constructs as intelligence, perceptual-motor ability, and adaptive behavior. Professionals should also abandon such imprecise diagnostic classifications as "mental retardation," "emotional disturbance," and "learning disabilities," substituting "learning objectives achieved." Skill development levels should be routinely assessable, not only with less bias but with adequate levels of reliability. Assessment would then be based on the actual service provided by the schools: the teaching of a curriculum. Federal law requires that steps be taken to insure the use of nonbiased assessment pro- cedures for special education students. The path, unfortunately, is blocked by the educational system itself. We cannot achieve meaningful nonbiased assessment by using current norm referenced devices and current classification systems; but it may be possible to make progress in achieving nonbiased assessment by using criterion referenced tests and classification systems based on purely educational criteria. Pursuing these alternatives, however, will create financial burdens and will place innovative responsibilities on the field. The question is: Are taxpayers willing to pay the necessary price? REFERENCES Adler, M. Intelligence testing of the culturally disadvantaged: Some pitfalls. The Journal of Negro Education, 1968,37(4), 364-369 Alzate, G. Analysis of testing problems in Spanish speaking children. Paper presented at the World Congress on the Future of Special Education, Stirling, Scotland, June 1978. CEe Delegate Assembly. Resolutions, Exceptional Children, 1978,45(1), 66. Anastasi, A. Psychological tests: Uses and abuses, Tenchers College Record, 1961, 62, 389-393. Cleary, T., Humphreys, L., Kendrick, S" & Wesman, A. Educational uses of tests with disadvantaged students. American Psychologist, 1975,30, 15-4l. Cole, M. Culture, cognition and I.Q. testing. The Nntionnl Elementary Principnl, 1975, 54(4), 49~ 52. Cole, N. S. Bias in selection, Journal of Educational Measurement, 1973,10,237-255. Cromwell, R., Blashfield, R. K" & Strauss, J. S. Criteria for classification. In N Hobbs (Ed.), Issues in classification (Vol. 1). San Francisco: [ossey Bass, 1975, Duffy, 0, Fnctors affecUng the performance of disadvantaged, rural, Appalachian children on measures of learning aptitude, Unpublished master's thesis, Pennsylvania State University, 1974. Feuerstein, R. The dynamic ossessment of retorded performers: The leorning potenUol assessment device, theory, instrument ond techniques. Baltimore: University Park Press, 1979. Flaugher, R. L. Bias in testing: A review and discussion (TM Report No. 36). Princeton, N,J.: Educational Testing Services, 1974. (ERIC Document Reproduction Service No. ED 099 431) Foster, G., & Salvia, J, Teacher response to the label of learning disabled as a function of demand characteristics, Exceptional Children, 1977, 43, 533~ 534. Jensen, A. Bias in mental testing. New York, Free Press, 1980. 433 Exceptional Children Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016 Kaufman, A. S. Intelligent testing with the WISCR. New York: John Wiley & Sons, 1979. Larry P. v . Riles. No. C-71-2270 RFP, U.S. District Court, Northern District of California (June 1979). Linden, K. W.. Linden, J- D., & Bodine, R. 1. Test bias: Fuss n' facts. Measurement and Evaluation in Guidance, 1974,7(3), 163-168. Mercer, f. SOMPA Workshop. Presented for St. Paul Public Schools, St. Paul, Minnesota, 1977. Mercer, J., & Lewis, J. System of multicultural pluralistic assessment. New York: Psychological Corporation, 1978. Meyer, J- S., & Elkind, D. From figurative operative expectancy in the perceptual judgments of children.Developmental Psychology, 1975,11 (6), 814823. Nunnally, J. Psychometric theory. New York: MeGraw-Hill, 1967. Oakland, T. The effects of test-wiseness materials on standardized test performance of preschool disadvantaged children. Journal of School Psychology, 1972,10(4),355-360. Padilla, A. M., & Garza, B. M. 6 cases of cultural myopia The National Elementary Principal, 1975, 54(4), 5:3-58. Petersen, N. S., & Novick, M. R An evaluation of some models for culture-fair selection. Journal of Educational Measurement, 1976,13(1), 3-29. Salvia, J., Algozzine, B., & Sheare, J- Attractiveness and school achievement Journal of School Psychology, 1977,15,60-67. Salvia, J-, & Ysseldyke, J- E. Assessment in special and remedial education. Boston: Houghton-Mifflin, 1978. Siegel, A. L, & Leahy, W. R. Nonverbal and culturefair performance prediction procedures: III. Cross validation. Wayne, Pa.: Applied Psychological Services, 1974. Sindelar, P. T., & Deno, S. 1. The effectiveness of resource planning. Journal of Special Education, 1978,12(1), 17-29. Tucker, J. A. Ethnic proportions in classes for the learning disabled issuing on non-biased assessment. Journal of Special Education, 1980, 14(1), 93-105. Weierbach, J. Students' obesity as a biasing factor in teachers' academic expectations. Unpublished master's thesis, Pennsylvania State University. 1977, Williams, R 1. The BITCH-lOO: A culture-specific test. Bethesda, Md.: National Institute of Mental Health (DHEW), 1972. (ERIC Document Reproduction Service No. ED 070 799) Ysseldyke, J- E., & Salvia, J. A. Diagnostic-prescriptive teaching: Two models. Exceptional Children, 1974,41, 181-186. 434 March 1981 Downloaded from ecx.sagepub.com at PENNSYLVANIA STATE UNIV on September 17, 2016
© Copyright 2026 Paperzz