Learning disabilities as functions of familial learning problems and developmental problems.
* In recent years, a consensus concerning the definition of learning disabilities has emerged in the literature (although such agreement does not necessarily prevail in practice, particularly as a definition would affect special education; Furey, 1987; Landers, 1987). This consensus was perhaps best expressed by the National Joint Committee for Learning Disabilities (NJCLD) which defined the term learning disability as follows:
Learning disability is a generic term that refers to a heterogeneous group of disorders manifested by significant difficulties in the acquisition and use of listening, speaking, reading, writing, reasoning, or mathematical abilities. These disorders are intrinsic to the individual and presumed to be due to central nervous system dysfunction. Even though a learning disability may occur concomitantly with other handicapping conditions (e.g., sensory impairment, mental retardation, social and emotional disturbance) or environmental influences (e.g., cultural differences, insufficient/inappropriate instruction, psychogenic factors), it is not the direct result of those conditions or influences. (Hammill, Leigh, McNutt, & Larsen, 1981, p. 336)
Critical elements of this definition are the presumptions that learning disabilities (a) result from central nervous dysfunction and, equally as important, (b) are not secondary manifestations of other primary disorders, most particularly environmental deprivation or emotional disorders.
The definition and its underlying concept are neutral, however, concerning what does cause learning disabilities. Various causes have been suggested--such as genetic defects, intrauterine effects, and laterality. One possible cause that, surprisingly, is omitted from the list of causal candidates is postnatal biological factors. None of these potential contributants to learning disabilities, of course, is exclusive of the others. Other important aspects of learning disabilities, such as their prevalence and incidence, secular changes, and correlates, also remain controversial.
Interest in the possible influence of genetics dates back nearly as far as the first term used to describe learning disabilities and may be implied by the term itself ("congenital word blindness"; Kussmaul, 1877). Empirical evidence supporting the hypothesis that genetic factors had an effect on learning disabilities was first supplied by Thomas (1905). Further empirical support has been provided by family studies (Eustis, 1947a, 1947b; Kagen, cited by Owen, 1977; Norrie, cited by Thompson, 1954; Orton, 1930) and twin studies (Hermann, 1959).
Efforts to identify specific subtypes of learning disabilities may also be traced nearly as far back as the concept itself (Hinshelwood, 1895; Morgan, 1896). Accurate subtyping of learning disabilities is potentially significant to their diagnosis and treatment; but subtyping has a particular bearing on testing the hypothesis that genetic factors contribute to learning disabilities. Particular subtypes of any disorder with distinguishing phenomenological features and specific concomitant markers lend themselves to family studies, medical genetics, and Mendelian analysis.
By contrast, heterogeneous agglomerations of phenomenology are much more difficult to trace through families, let alone through populations. The strategy of combining subtyping with examination of potential genetic determinants has been used to great advantage in the understanding of the etiology of mental retardation, another extremely heterogeneous domain (Childs, Finucci, & Preston, 1977).
Accordingly, strong calls for accurate subtyping of learning disabilities have been issued (Childs et al., 1977), and much recent research designed to test the hypothesis of genetic contributions to learning disabilities has used precisely this strategy. Thus investigators examining causes of learning disabilities have very frequently elected to define them specifically as reading disorders or as some cognate, such as dyslexia or reading disability (Coles, 1978; Institute of Behavioral Genetics Studies, e.g., Foch, DeFries, McClearn, & Singer, 1977; Matheny & Dolan, 1974). Such research testing the degree to which genetics contribute to reading disorders has supported the presence of a genetic determinant. This strategy has not been successful, however, in identifying a single or Mendelian mode of transmission. In addition, the extent to which the findings may be generalized to a wide range of learning disabilities remains unclear.
Unfortunately, the underlying premise on which the strategy is based, that of differentiated, discriminable subtypes of learning disabilities, may not accurately describe the domain. Although it is possible that small specific clusters of learning disabilities, such as visual-motor deficits and verbal deficits, may exist (McCue, Goldstein, Shelly, & Katz, 1986), these clusters taken together may account for only a small proportion of the total domain. Thus several investigators now take the position that learning disabilities are intrinsically undifferentiated disorders (McCue et al., 1986; McClearn, 1977) or, in the words of one group of investigators, "The potential number of subtypes of ... learning disabilities is limited only by the ability of the neuropsychological tests to differentiate efficiently the components and subcomponents involved in any particular neuropsychological or cognitive task" (Hynd, Obrzut, Hayes, & Becker, 1986, p. 469).
Because of heterogeneity inherent in learning disabilities, a strategy that may be used in testing their determination by genetic determinants is one that examines biological relatives of two groups of offspring, a learning-disabled or "index" group and a control group, which are matched on a number of other variables, and compares rates of learning disabilities in the two groups of relatives. This strategy is similar to that of family studies, but uses both a larger index group and a control group; it gives up specificity of the disorder studied for the statistical power obtained by comparing relatively large samples. Critical to the application of such a strategy is obtaining a large sample that is representative of learning disabilities considered generically, particularly because most previous research has used convenience samples drawn from populations presenting for treatment, populations in which variables associated with the disorder are confounded with variables associated with help seeking.
A population in which correlates and potential causes of learning disabilities may be studied is provided by the St. Louis County Special School District (SLCSSD). SLCSSD, a special educational concept and administrative entity, is one of perhaps two of its kind in the United States. This school district, which encompasses the geographical area of St. Louis County, Missouri, is charged with providing diagnostic services and education to all children whose education is not adequately served by the regular public schools within its boundaries. These children are classified as mentally retarded, orthopedically handicapped, behaviorally disordered, emotionally disturbed, and learning disabled. At any one time, SLCSSD has about 6,000 children on its rolls as learning disabled. The diagnosis of learning disabilities and eligibility for services follow the definition of learning disabilities on which the NJCLD definition is based, that is, the implementation of Public Law 94-142 (U.S. Office of Education, 1977), as well as a more recent interpretation of that definition (Mallory, 1984-1986). The diagnosis of learning disabilities and determination of eligibility rests with a multidisciplinary team that makes inferences about children's underlying psychological processes based on a wide variety of data from complete psychological evaluations, classroom observations, academic records, medical records, interviews with parents, and specialized testing.
The purpose of the present investigation was to examine familial learning problems and developmental factors as correlates of and potential contributants to learning disabilities construed as disorders of central information processing in a large, representative sample of children with learning disabilities and in an appropriate control group. Index cases for the study were drawn from SLCSSC. Control cases were selected from a large public school district within the same county. Learning problems in genetically related individuals were assessed by the report of the informant, generally the mother. Developmental factors assessed were gender, length of pregnancy, birth complications, twin status and type, birth order, and laterality as measured by handedness. Concerning its design, therefore, the present investigation was a retrospective longitudinal design with respect to developmental factors, and a consanguinity study with respect to familial learning problems.
All subjects, both index and control, were between 5 and 19 years of age and were enrolled in Grades kindergarten through 12. Each group of subjects was empaneled by respondents from (by coincidence) an initial subject pool of 500, which was in turn randomly drawn from a school enrollment of 6,000 students. Educational characteristics of the two groups, following criteria developed by the Council for Learning Disabilities Research Committee (Smith, Deshler, Hallahan, Lovitt, Robinson, Voress, & Ysseldyke, 1984) are shown in Table 1. Demographic characteristics of the two samples, as well as tests of the differences between them, are provided in Table 2.
Index Group. Subjects numbered 101 pupils enrolled in SLCSSD; thus the participation rate from this group was 20%. All children enrolled in SLCSSD, and thus all in this study, are required to be (re)diagnosed at least every 3 years; all are diagnosed by an interdisciplinary team whose members receive inservice training on diagnosis provided by the district.
Control Group. Subjects numbered 171 pupils enrolled in the Pattonville School District, a large public regular school district located within the boundaries of St. Louis County; thus the participation rate in this group was 34%. Pattonville was used as the most appropriate source of the control group because among the school districts selected by consulting demographers as the best matches for St. Louis County demographically, Pattonville was the school district willing to participate.
As shown in Table 2, the effort to match the two groups demographically was only partially successful; significant differences between the two groups emerged on 9 of the 17 background variables measured. Many of the 9 variables that discriminated between the two groups, however, were related to the same underlying construct of social class. Index subjects were more likely to be enrolled in lower grades (despite equivalent age) and less likely to be only children. Families of index subjects were likely to have more offspring, both biological and stepchildren, and to be Black; parents were more likely to have less education; fathers were more likely to be employed less than full-time; and mothers were likely to have lower occupational status.
Materials consisted of two cover letters and a three-part questionnaire. The cover letter from the investigators requested participation from parents of two groups of pupils, one with identified learning problems and one with none. The letter also informed subjects that the research project pertained to family background as it related to children's learning characteristics and promised confidentiality. The cover letter from the school district introduced the investigators, provided sanction for the study, and promised voluntariness and further confidentiality.
The questionnaire was modeled on one developed by Winter, Cole, and Wright (1983), with modifications made specifically for the current study with advice from researchers at each of the school districts. The first part of the questionnaire inquired about the child's living arrangements, family members, and development, namely age, gender, birth order, number of siblings, and learning difficulties.
The second part of the questionnaire addressed learning problems in biological relatives. Learning problems were arranged in a rationally devised 5-point ordinal scale, which consisted of the following:
1. Learning specific skills or content, such as reading, spelling, arithmetic.
2. Being retained 1 or more years.
3. Being tested by a school counselor or the SLCSSD for academic or learning (not behavioral or emotional) problems.
4. Being diagnosed by a school counselor, the SLCSSD, or other professional as learning disabled.
5. Receiving services (remediation, tutoring, resource room, or developmental bypass) for learning disabilities.
Biological relatives whose learning problems were inquired about were as follows:
1. Monozygtic (MZ) twins.
2. First-degree relatives--dyzygotic (DZ) twins, parents, and full-siblings.
3. Second-degree relatives--grandparents, aunts and uncles, and half-siblings.
4. Third-degree relatives--cousins.
Nonbiological relatives were not inquired about; for example, step-parents and step-siblings were omitted from the list of relatives.
The third part of the questionnaire sought information about parents with whom the child lived, that is, their race, household composition, and variables related to social class: education, employment status, and occupational status. In this case, the focus of the inquiry was the environment in which the child was currently living and not on the environment of his or her biological parents.
Pupils in the initial subject pool were randomly selected from the total enrollment by personnel in each school district (at each district, every 12th name was selected). Address labels corresponding to each child selected for the initial subject pool were drawn by district personnel, and questionnaires were mailed by the district. Stamped envelopes addressed to the investigators were enclosed. Two weeks after the mailing, a postcard was sent to each address to remind the recipient to return the questionnaire as soon as possible. To guarantee mutually exclusive groups, any family who received more than one questionnaire as a result of having children enrolled in both school districts was requested to complete only the first questionnaire received but to return the other also.
Index subjects and their families were compared with control subjects and their families to determine whether there were significant differences between them on any of the variables of interest. Multiple regression analyses were performed in which learning status of the offspring, whether index or control, served as the criterion; and two groups of variables, familial learning problems and developmental factors, served as predictors.
Before multiple regression analyses could appropriately be performed, it was necessary to determine whether any of the nine demographic variables which differed significantly between the two groups was also significantly correlated with any of the predictors; that is, whether any of these variables was confounded. The results of univariate tests of association between demographic variables and predictor variables are shown in Tables 3 and 4. These tables show that eight of the nine demographic variables previously determined to be associated with learning status of the child also proved to be associated with one or more of the predictors. These variables were treated as confounds in multiple regression analyses by entering them as a group first into the regression equation, thus removing variance associated with or controlling statistically for confounded demographic variables initially.
Familial learning problems were defined by measuring three sets of four variables each. The three sets were defined as (a) the most severe problem experienced, following the 5-point ordinal scale of learning problems enumerated previously in the description of the questionnaire; (b) the total number of problems experienced; and (c) a combined index, total severity, of problems experienced. To illustrate, suppose an index case's full-brother was described as having been retained 1 year in school, having been tested for learning problems, and having been diagnosed as learning disabled by a professional. Severity was scored as 4 because diagnosis of learning disability was considered more severe than was either testing (severity score of 3) or retention (severity score of 2). Number of problems was scored as 3, one each for diagnosis, testing, and retention. Total severity of problems experienced was scored as 9 (diagnosis with a score of 4, plus testing with a score of 3, plus retention with a score of 2).
The four variables within each set consisted of (a) MZ twin, if any; (b) first-degree, (c) second-degree, and (d) third-degree relatives. To return to the preceding example, learning problems in the full-brother, however they were measured, would always be accrued to first-degree relatives.
Six multiple regression analyses were conducted. Confounding variables were always entered first hierarchically in each of the six analyses. The first set of three regressions addressed the question of whether familial learning problems made any contribution to the prediction of offspring's learning status, independent of prediction already afforded by demographic variables and developmental problems. This was accomplished by entering developmental problems together as a group hierarchically after demographic variables; the four familial learning problems defined by the four degrees of genetic relatedness were then entered stepwise. A parallel procedure was used to address the question of whether developmental factors made any contribution to the prediction of offspring's learning status, independent of familial learning problems: Familial learning problems taken together as a group were entered after demographic variables hierarchically, and developmental factors were then entered stepwise. In each set of three regressions, familial learning problems were operationalized in the first equation as number of problems; in the second equation, as severity of problems; and in the third equation, as number by severity of problems.
Characteristics of the two study groups in terms of familial learning problems and developmental factors, together with the results of univariate comparisons between the two groups, are listed in Tables 5 and 6. Significant differences between the two groups were found on the majority of familial learning problems and on three out of eight developmental variables.
Because of the large number of univariate comparisons, however, multiple regression portrays more clearly the independent contributions to the variance in learning status of offspring made by familial learning problems and by developmental factors. Table 7 summarizes the results of the six regression analyses, including the change in [R.sup.2], the square of the multiple correlation coefficient and thus the proportion of variance accounted for, added by each variable or group of variables. As shown in Table 7, the set of eight confounded demographic variables accounted for 22% of the variance in the criterion; the set thus proved to be powerful predictors of learning status in offspring, despite the fact that they lay outside the scope of the present investigation.
In the first set of analyses (Analyses I, II, and III), familial learning problems, however they were defined and measured, accounted for 8%-9% of unique variance. In analysis I, familial learning problems were defined and measured as severity; in Analysis III, they were defined and measured as number by severity of problems. Results of these two equations were alike in that problems experienced by third-degree relatives made the most significant contribution in accounting for further variance (about 5%), followed by problems in first-degree relatives (about 3%). In Analysis II, familial learning problems were defined and measured as number of problems; in this equation, ordering of third- and first-degree relatives was reversed, whereas amounts of variance explained by familial learning problems remained quite constant. Comparison of the results of Analyses I, II, and III indicates that number of problems in first-degree relatives, severity in third-degree relatives, and total severity of problems in third-degree relatives proved to be the best predictors, all of approximately equal power in accounting for 5% of unique variance. These three predictors were not pitted against each other in a single subsequent analysis because of the large number of analyses already undertaken.
In the second set of analyses (Analyses IV, V, and VI), familial learning factors were entered into the equation hierarchically after confounded demographic variables, to learn whether developmental factors, like familial learning problems, made a unique contribution to explained variance. In Analysis IV, familial learning problems were defined and measured as severity; in Analysis V, they were defined and measured as number of problems; and in Analysis VI, they were defined and measured as total severity of problems. The amount of variance explained by familial learning problems averaged about 8% across the three equations, despite the fact that it varied slightly because of the differing definitions of learning problems used in the three equations. As when familial learning problems were entered last, number proved to be the best predictor, followed by total severity, followed by severity.
In all three analyses in the second set, only gender explained unique variance (about 5%) when variance attributable to familial learning problems had been removed previously. The significant association between gender and group status indicated that male gender was more frequent in the learning-disabled group. In the three analyses in the first set, developmental factors were entered into the equation immediately after the demographic variables were entered hierarchially. In this position, demographic variables explained nearly 5% of the remaining variance in offspring learning status.
In summary, the results of the six regression analyses indicate that, when learning disabilities are defined as disorders presumed to be due to central nervous system dysfunction, familial learning problems and developmental factors are each significantly and independently associated with offspring's learning disabilities, even when variability in learning disabilities attributable to demographic variables, especially to social class, has been removed statistically. Comparison of results of the two sets of analyses indicates that familial learning problems, regardless of how they are defined, make more powerful predictors of learning status in offspring than do developmental factors, of which only gender was significantly associated. It may also be concluded that both familial learning problems and gender make unique contributions to explained variance. Results also indicate that familial learning problems may be defined in a number of different ways, as findings were robust across three different methods of defining and measuring them used in this study. Finally, the combination of demographic variables, family learning problems, and gender predicts nearly one-third, or a substantial portion, of the variance in learning status in offspring.
These results indicate unequivocally that learning disabilities presumed to be due to central nervous system by dysfunction are familial. Inferences that should be taken from the findings, however, are much more ambiguous. Disorders that are familial may generally be caused by either heredity or environment because, in a consanguinity study, both the hereditarian and environmental positions make the prediction that the disorder will be familial. At the heart of this equivocation is the definition of learning disabilities. On the one had, the definition employed in this study and by the school district in question should have ruled out disorders in neurological processes attributable to emotional or to other environmental causes. On the other hand, proponents of an environmental position, such as Coles (1980), would probably wish to argue that disorders in neurological processes that are relatively irreversible and permanent might nevertheless be caused by environmental determinants, either child-rearing factors or biological environmental factors such as food additives or pollutants, which impinge upon the child early in his or her developmental course, perhaps during a critical period, and leave an indelible mark. Although the present study has probably partially ruled out the effect of environmental and emotional determinants as usually defined, it can go no further in ruling out irreversible effects of very early psychosocial environmental factors or of biological environmental factors. Pertinent to an issue that is raised in the following discussion, the biological environment may very well interact with genetic factors implicated in this investigation.
If the findings of the current investigation are to be called into question, it might be on the basis of the reliability with which the definition used in this study was implemented. It may thus be that the decisions of interdisciplinary teams may deviate considerably in practice from the theoretical guidelines to which they are intended to conform. Relevant to this concern, neither the interjudge reliability nor the validity of these decisions is presently known. Lack of reliability could operate in two ways having potentially opposite effects.
First, according to an experienced examiner in SLCSSD, diagnoses of learning disabilities are frequently made by less well-trained examiners not on the basis of the definition of the U.S. Office of Education (1977), but on the "sever discrepancy" definition, or the presence of a severe discrepancy between the child's school performance and his or her ability as measured by an individually administered intelligence test. Such a practice would presumably work to exclude many children who actually have learning disabilities, but those disabilities are not manifest in such a discrepancy. This effect woulds produce a more homogeneouse group of children with learning disabilities, thus increasing the sensitivity of tests of differences between children with and without learning disabilities.
A second type of unreliability, of diagnostic teams' being erratic in making their diagnoses, would work in the opposite direction to obscure between-group differences: Familial differences would be underestimated (a Type II error).
Lack of validity, however, could create these findings spuriously: If judges were systematically biased toward diagnosing problems caused by child rearing, such as emotional or motivational problems, as learning disabilities, then familial differences between groups with and without learning disabilities would be found, but these differences would not be truly attributable to learning disabilities (a Type I error). Goals for future research, therefore, should probably include (a) using objective judges to make independent diagnoses of records and (b) determining the reliability and validity of the judges' diagnoses.
Percentages of informants for probands with learning disabilities who reported familial learning problems in MZ twins, first-, second-, and third-degree relatives were 4%, 60%, 25%, and 21%, respectively; corresponding percentages for control probands were 0%, 27%, 19%, and 8%, respectively. Rates reported in this study for relatives of probands with learning disabilities are high, but highly comparable with rates identified by other investigators. For example, Rugel and Mitchell (1977) found a family history of reading disorders in 56% of a sample of subjects with reading disabilities; Satyan (1980), problem readers in 49% of immediate family members of a sample of problem readers; and Silver (1971), learning disabilities in 35% of immediate family members of a sample with learning disabilities.
Attention should be called to the apparent greater association with learning disabilities of learning problems in first- and third-degree relatives than with learning problems in second-degree relatives. Although these associations involving third- and first-degree relatives may not be significanly higher than those involving second-degree relatives, they are thought provoking because they involve more biological relatives of the same age as the index cases (siblings, half-siblings, and cousins) than older ages (uncles and aunts). The apparent involvement of younger biological relatives raises the issue of whether a birth-cohort effect may be occurring, that is, whether learning disabilities may in fact be more frequent in persons born more recently. This issue may be related to two possibilities. First, a biological, environmental factor may be interacting with an hereditary factor. Second, because of the current heightened awareness of learning disabilities, both teachers and parents may detect these problems more frequently in children of more recent cohorts.
Research concerning correlates and potential causes of learning disabilities has now reached the stage programmatically at which more sophisticated designs can separate genetic and environmental factors. It is conceivable, for example, that the large size of the St. Louis County Special School District would make possible carrying out an adoption study within this population. Further, the significant association between learning problems in biological relatives and the learning disabled status of offspring indicates that an at-risk design may also be feasible. The at-risk design would have the greatest chance of identifying environmental factors that might interact with heredetary factors to determine learning disabilities. This report ends with a plea that studies of this kind be performed in this or other comparable populations with all deliberate speed.
Childs, B., Finucci, J. M., & Preston, M. S. (1977). A medical genetics approach to the study of reading disability. In A. L. Benton & D. Pearl (Eds.), Dyslexia: An appraisal of current knowledge (pp. 300-309). New York: Oxford University Press.
Coles, G. S. (1978). LEarning Disabilities Test Battery: Empirical and social issues. Harvard Educational Review, 48,313-340.
Coles, G. S. (1980). Evaluation of genetic explanations of reading and learning problems. Journal of Special Education, 14,365-383.
Eustis, R. S. (1947a). The primary etiology of specific language disabilities. Journal of Pediatrics, 31,448-455.
Eustis, R. S. (1947b). Specific reading disability. New England Journal of Medicine, 237,243-249.
Foch, T. T., DeFries, J. C., McCLearn, G. E., & Singer, S.M. (1977). Familial patterns of impairment in reading disability. Journal of Educational Psychology, 69,316-329.
Furey, R. J. (1987, Winter). On the nature of learning disabilities. Missouri Psychologist (Missouri Psychological Association).
Hammill, D. D., Leigh, J. E., McNutt, G., & LArsen, S. C. (1981). A new definition of learning disabilities. Learning Disability Quarterly, 4,336-342.
Hermann, K. (1959). Reading disability: A medical study of word-blindness and related handicaps. Copenhagen: Munksgaard.
Hinshelwood, J. (1895). Word-blindness and visual memory. Lancet, 2,1564-1570.
Hynd, G. W., Obrzut, J. E., Hayes, F., & Becker, M. G. (1986). Neuropsychology of childhood learning disabilities. In A. N. Horton & J. Webster (Eds.), The neuropsychology handbook (pp. 456-485). New York: Springer.
Kussmaul, A. (1877). Disturbance of speech. Cyclopedia of Practical Medicine, 14,581-587.
Landers, S. (1987). LD definition disputed. The Monitor (American Psychological Association), December 1987, p. 35.
Mallory, A. L. Special considerations in the identification and evaluation of students with specific learning disabilities: An in-service guide for educators. Protection in Evaluation Procedure of the Missouri Department of Elementary and Secondary Education State plan. Missouri 1984-1986 State Plan. Jefferson City, Missouri.
Matheny, A. P., & Dolan, A. B. (1974). A twin study of genetic influences in reading achievement. Journal of Learning Disabilities, 7,99-102.
McClearn, G. E. (1977). Review of "Dyslexia--Genetic aspects." In A. L. Benton & D. Pearl (Eds.), Dyslexia: An appraisal of current knowledge (pp. 285-297). New York: Oxford University Press.
McCue, M., Goldstein, G., Shelly, C., & Katz, L. (1986). Cognitive profiles of some subtypes of learning disabled adults. Archives of Clinical Neuropsychology, 1, 3-23.
Morgan, W. P. (1896). A case of congenital word-blindness. British Medical Journal, 2,1378.
Orton, S. T. (1930). Familial occurrence of disorders in the acquisition of language. Eugenics, 3,140-147.
Owen, F. W. (1977). Dyslexia--Genetics aspects. In A. L. Benton & D. Pearl (Eds.), Dyslexia: An appraisal of current knowledge (pp. 265-284). New York: Oxford University Press.
Rugel, R. P., & Mitchell, A. (1977). Characteristics of familial and nonfamilial disabled readers. Journal of Learning Disabilities, 10,308-313.
Satyan, H. S. (1980). Management of children with reading difficulties: A multidisciplinary approach. Journal of Learning Disabilities, 13,435-439.
Silver, L. B. (1977). Familial patterns in children with neurologically-based learning disabilities. Journal of Learning Disabilities, 4,349-358.
Smith, D. D., Deshler, D., Hallahan, D., Lovitt, T., Robinson, S., Voress, J., & Ysseldyke, J. (1984). Minimum standards for the description of learning disabilities research reports. Learning Disability Quarterly, 7,221-225.
Thomas, C. J. (1905). Congenital "word-blindness" and its treatment. Opthalmoscope, 3,380-385.
Thompson, L. J. (1554). REading disability. Springfield, IL: Charles C Thomas.
U.S. Office of Education. (1977, August). federal Register, 42, Part II.
Winter, A., Cole, E., & Wright, E. N. (1983). Characteristics of pupils entering learning disabilities self-contained classes. (Report No. ISBN-0-88881-171-3). Toronto Board of Education (Ontario), Research Dept. (ERIC Document Reproduction Service No. ED 160 019)
J. M. OLIVER is a professor in the Department of Psychology at St. Louis University, Missouri. NANCY HODGE COLE is a Senior Research Associate at Brown University's Alcohol and Drug Addiction Studies of Providence, Rhode Island. HOLLY HOLLINGSWORTH is Director of Academic Computing at St. Louis University, Missouri.
|Printer friendly Cite/link Email Feedback|
|Author:||Oliver, J.M.; Cole, Nancy Hodge; Hollingsworth, Holly|
|Date:||Mar 1, 1991|
|Previous Article:||The effects of initial interval size on the efficacy of DRO schedules of reinforcement.|
|Next Article:||The contribution of skills analysis to curriculum-based measurement in spelling.|