Effects of preschool integration for children with disabilities.
Research examining the educational and developmental effects of integration has generally found little or no difference between integrated students and segregated students on language, social, and general developmental measures (e.g., Cooke, Ruskus, Apolloni, & Peck, 1981; Ispa & Matz, 1978; Jenkins, Odom, & Speltz, 1989; Jenkins, Speltz, & Odom, 1985). These studies, however, were designed to evaluate group differences between integrated and segregated students, using statistical methods such as analysis of variance or covariance. None of these studies examined the relationship of children's pretest functioning and the effect of classroom integration.
This absence of research on individual differences is notable, because several lines of research suggest that students may respond differentially to integration depending on their specific developmental characteristics. Galloway and Chandler (1978), for example, found that developmental progress in integrated classes was related to degree of disability; lower functioning students made the smallest gains, and students with mild disabilities demonstrated larger gains. Similarly, Guralnick (1980) found differential social performance related to degree of disability in a study of the social interactions of normally developing students, those with mild disabilities, and those with severe disabilities. The normally developing students and those with mild disabilities interacted with each other more extensively than either did with students with severe disabilities. The impact of varied amounts of social and communicative interaction available to students as a function of their disability level is not yet known. It has been demonstrated, however, that normally developing children as young as 4 years of age adjust the complexity of their language when speaking to children with less developed language skills (Shatz & Gelman, 1973). Furthermore, exposure to language adjusted to the developmental level, particularly comprehension level, of the child, appears to play a facilitating role in the early language development of normally developing (Furrow, Nelson, & Benedict, 1979) and language-delayed (Fey, 1986) children.
Taken together, these lines of research imply that differential amounts of exposure to language and social interaction from more highly skilled children might differentially influence development in children with delays in these areas. Children who are functioning at a higher level may receive more language and social stimulation, and benefit from the interactions more readily due to their higher skills levels. More delayed children, who generally require more trials for learning to occur, may get less stimulation from the higher functioning students. In summary, integration may have differential effects on student progress; these effects are related to student aptitude at the onset of intervention. Thus the examination of mean group differences between students in integrated and segregated settings may not be sensitive to important differences in the effects of integration.
The purpose of the present study was to investigate the effects of integration and segregation in a special education preschool program for children with mild to moderate disabilities to determine the degree to which initial level of development influenced the gains achieved in the two conditions. The design of the study permitted random assignment of children with disabilities to integrated and segregated classes that were otherwise equivalent. The design also included extensive pretest and posttest assessment of children's performance in the areas of language, cognitive, and academic development. Based on previous research, we predicted that integration alone would not have a pronounced effect on the mean level of posttest performance of students. We hypothesized instead that higher functioning students would gain more in the integrated setting than in the segregated setting, while lower functioning students would gain more in the segregated setting. These hypotheses were evaluated using multiple regression analyses, including aptitude-by-treatment interaction terms.
A total of 124 children, ages 3-6 years, participated in this study. Of these children, 100 (71 boys and 29 girls) had mild and moderate disabilities, according to Washington State criteria for special education eligibility as "developmentally delayed." This eligibility classification includes children with deficits of at least 1.5 standard deviations below the mean in two or more of the areas of cognitive, language, social, gross motor, and fine motor development, or 2 standard deviations below the mean in any one area. Regarding specific diagnostic categories, approximately 80% of the subjects were delayed in language, 50% were cognitively delayed, 65% exhibited fine motor deficits, 60% demonstrated delays in gross motor skills, and 60% exhibited social-emotional delays. The percentages sum to more than 100% because the majority of children had significant deficits in more than one area of development. Twenty-four normally developing children (15 boys and 9 girls) served as models in integrated classrooms.
This study was conducted over a 4-year period, using preschool classes that met for half-days (2 hours [hr] per day). Two classes each year were integrated (4 normally developing students, 8 students with disabilities), and two were segregated (12 students with disabilities). Students with disabilities were randomly assigned to integrated and segregated classrooms; normally developing children were randomly assigned to specific integrated classrooms. Pretest performance for the three groups of subjects is presented in Table 1.
Teachers received no specific instructions regarding the typically developing, or model students in the classroom. This was done to isolate the effect of integration, rather than assess the combined effect of integration paired with a specific integration curriculum. Teachers engaged the models in a variety of ways across the classrooms. The typically developing students were sometimes called on to provide responses for new language, social, and academic tasks being introduced to the class, to provide correct exemplars within a natural context. More frequently, however, the model students were simply placed in
[TABULAR DATA OMITTED]
proximity to students with disabilities, without a specific structure overlaid on their interaction. In no case were the typically developing students coached as confederates, which would have emphasized that they were "models" and other students were "targets."
This study was conducted at a Pacific Northwest university laboratory school for young children with disabilities. Thirty-nine students with disabilities were included in the first year, 23 in the second year, 17 in the third year, and 21 in the fourth year of intervention. Although some children were enrolled in the school program for more than 1 year, only data from the child's first year in program were included in this study. Approximately 40% of students within each condition remained in the program longer than 1 year.
Each class was staffed by a head teacher and an assistant teacher. In addition, classes were served by a speech-language pathologist, occupational or physical therapist, graduate students, and work-study students. To control for teacher effect, each head teacher and assistant teacher served either an integrated or segregated classroom in the morning, and the opposite model in the afternoon. Classes met 5 days a week for the standard 180 school days.
Subjects were participants in a larger research study comparing the effects of two preschool curricula: an academically-based model and a cognitively-based model (Dale & Cole, 1988; Jenkins, Cole, Dale, & Mills, 1989). Within the segregated group, 30 students had been assigned to the academic curriculum, and 28 to the cognitive curriculum. The integrated group consisted of 20 students assigned to the academic curriculum, and 22 to the segregated curriculum. Jenkins, Cole, Dale, and Mills (1989) found only modest main effects of program for the larger sample. For purposes of the present study, program differences may have added some variance to posttest measures within the integrated and segregated groups. The equivalent division within both groups to program, however, eliminates the possibility of confounding of program with level of classroom diversity.
The following measures were administered pretest and posttest for the study:
McCarthy Scales of Children's Abilities. The McCarthy Scales of Children's Abilities (McCarthy, 1972) is an individually administered intelligence test for children ranging in age from 2 1/2 to 8 1/2 years. Subtests include Verbal, Perceptual, Quantitative, Memory, Motor, and the General Cognitive Index (GCI), composed of a combined set of subscores. Substantial reliability data are reported for the test. The GCI has an average split-half reliability of .93. Norms were gathered on approximately 1,000 children.
Peabody Picture Vocabulary Test-Revised (PPVT-R). The PPVT-R (Dunn & Dunn, 1981) is a nonverbal, multiple-choice test that evaluates comprehension of semantic information. The range of the test is 2 1/2 years through adulthood. Split-half reliability for the age range of 2 to 6 years ranges from .67 to .88 years. The test was normed on 4,200 children between the ages of 2.5 and 18 years.
Test of Early Language Development (TELD). The TELD (Hresko, Reid, & Hammill, 1981) is designed to assess both the semantic and syntactic aspects of children's language through expressive and receptive abilities. The age range is from 3 to 8 years. Split-half reliability is reported as averaging .90. The normative sample included 1,184 children.
Test of Early Reading Ability (TERA). The TERA (Reid, Hresko, & Hammill, 1981) is designed to assess reading ability for children from 3 to 8 years of age. It assesses the areas of construction of meaning (comprehension of written language), knowledge of the alphabet and its function, and conventions of written language (i.e., left-right progression). Split-half reliability ranges from .87 to .96. Standardization procedures included 1,184 children.
The test battery was administered in October through December as a pretest, and again in May through August as a posttest. In all cases, at least 6 months passed between pretest and posttest. Assessments were administered by speech-language pathologists and research staff. The latter included graduate students in speech-language pathology, special education, and psychology. Testing staff were not informed of the hypotheses involved in the study, although it was impossible to keep the testers blind to the classroom make-up (segregated or integrated) because the testing staff was housed in the school where the intervention took place.
Analysis of Pretest Performance
To determine the degree to which random assignment was successful in forming comparable groups, one-way analyses of variance (ANOVAs) were performed on the McCarthy GCI, PPVT-R, TELD, and TERA pretests. No significant differences were found between students in segregated and integrated classes at pretest. These results are reported in Table 1. There was a minor trend for the integrated subjects to perform slightly higher on measures, although this difference did not approach statistical significance for any measure.
Analysis of Posttest Performance
Differences in effectiveness of the two classroom-diversity configurations (integration and segregation) were assessed by the interaction terms in repeated-measures ANOVAs with Diversity as a between-subjects factor, and Time of Testing as a within-subjects factor. These results are summarized in Table 2. Raw scores, as well as quotient scores, are presented for the TERA and TELD because a number of students had raw scores on these pretests and posttests that were below the tabled quotient scores for the measures. Rather than extrapolate a quotient score, both types were analyzed and data reported. The difference in the Ns for the raw and quotient scores for these measures results from this scoring difference. As anticipated, no significant differences were found between the integrated and segregated subjects on any of the measures.
Aptitude-by-Treatment Interaction Analyses
Because previous research has suggested the possibility of differential effectiveness of classroom integration related to student aptitude (Galloway & Chandler, 1978; Guralnick, 1980), two potential aptitude measures were evaluated on their ability to predict program performance. A general
[TABULAR DATA OMITTED]
cognitive measure (McCarthy GCI) and a language knowledge measure (PPVT-R) were selected as measures that conceptually could influence a student's ability to benefit from classroom integration.
Following Pedhazur (1982), multiple regression analyses were used to evaluate Aptitude-by-Treatment interactions (ATI). Multiple regression techniques use the full range of pretest values, rather than arbitrarily blocking them as is the case for ANOVAs. For each ATI analysis, the appropriate pretest variable was entered first, followed by condition (integrated vs. segregated), followed by the interaction term. A significant interaction term reflects a significant difference in the regression slopes of the posttest variable on the pretest variable for the two groups.
The results of these analyses are presented in Table 3. Three significant ATIs were found. These consisted of the pretest McCarthy GCI as a predictor of posttest PPVT-R standard score and TELD language quotient, and the pretest PPVT-R raw score as a predictor of the McCarthy Verbal subtest. For all three significant ATIs, the direction of the interaction revealed that higher performing students gained more from the integrated classes, whereas lower performing students gained more from the segregated classes. Figures 1, 2, and 3 illustrate the significant interactions with the regression lines for the two groups.
Furthermore, the same direction of interaction--steeper slope for the integrated group than for the segregated groups--was observed for 13 of the 15 nonsignificant analyses reported in Table 3. Thus a total of 16 out of 18 analyses produced similar results. The failure of analyses to attain significance when the pretest and posttest measures were the same (though they were in the appropriate direction) reflects the high correlation between pretest and posttest scores on the same variable (r = .77 and .66, respectively, for the McCarthy GCI and PPVT-R). In this case, there is less variance to be accounted for by other measures. In contrast, other correlations between pretest and posttest variables in Table 3 ranged from .23 to .72, median 4 = .53.
The purpose of this study was to examine the effects of classroom integration and segregation for young children with disabilities. The major hypothesis was that children's initial levels of development for language and general cognitive development
Summary of Aptitude-by-Treatment Multiple Regression Analyses for Students in Integrated and Segregated Classes
Pretest Aptitude Measure McCarthy Posttest Measure GCI PPVT-R McCarthy GCI N.S. N.S. Verbal N.S. F<.05 Perceptual N.S. N.S. Quantitative N.S. N.S. Memory N.S. N.S. Motor N.S. N.S. PPVT-R Standard Score F<.01 N.S. TELD Quotient Score F<.05 N.S. TERA Quotient Score N.S. N.S.
Note: McCarthy = McCarthy Scales of Children's Abilities (McCarthy, 1972); GCI = General Cognitive Index, a McCarthy subtest; PPVT-R = Peabody Picture Vocabulary Test--Revised (Dunn & Dunn, 1981); TELD = Test of Early Language Development (Hvesko, Reid, & Hammill, 1981); TERA = Test of Early Reading Ability (Reid, Hresko, & Hamill, 1981).
would influence their performance within the two conditions. After randomly assigning children to experimental conditions, we provided center-based preschool instructions for 1 school year. A battery of tests administered pre-intervention and postintervention indicated no significant overall differences between the two treatment conditions (integration and segregation). Multiple regression analyses, with the pretest McCarthy GCI and PPVT-R as predictor variables, revealed a pattern of ATIs in which lower functioning children made greater gains in segregated settings, and relatively higher functioning children made greater gains in integrated settings.
These interactions may have been mediated by the models, the teachers, or both. It is likely that the models provided differential amounts of language and social interaction for the higher functioning and lower functioning students. Cavallaro and Porter (1980), Blackmon and Dembo (1984), Rogers-Warren, Ruggles, Peterson, and Cooper (1981), Guralnick (1980), and Ispa (1981) have noted that normally functioning models tended to select either other models or students with mild disabilities with whom to interact. It is possible that greater access to models by the students with milder disabilities facilitated their cognitive and language development. For the students with a greater degree of disability, the segregated classrooms may have actually provided more language stimulation because of the presence of a larger number of students who were interested in interacting with the more involved children.
A second potential causal factor for the ATIs may be the general level of instruction provided by the teaching staff. Perhaps instruction in classes containing model students is more complex and fast paced to keep the attention of the normally developing children. This higher level of instruction could facilitate learning in students with mild disabilities who may be able to benefit from increased complexity and higher expectations. For the more delayed students, however, instruction that will stimulate the model students may be too complex and fast paced to be effective.
These findings have several implications for research and program development in early childhood special education. It is important to realize that there are measurable differences in student performance that result from classroom integration. Earlier studies have examined the overall effect of integration for groups of students, and found only minimal effects of integration. This study suggests that the effects of integration may be complex, and that individual student profiles can influence performance within integrated and segregated settings.
These results should not be interpreted prematurely as an indication that all lower functioning students should be placed in segregated settings, and all higher functioning students in integrated settings. Although the statistically significant ATIs suggest that lower functioning children may be more successful in segregated settings, the amount of variance accounted for by the interaction between aptitude measures and test performance ranged from 4% to 5%. This trend is consistent enough to be statistically significant, but it represents only a small difference in actual test performance.
For comparison, in those ATI analyses for which the interaction term was significant, the pretest scores generally accounted for 16% to 27% of the variance in the posttest scores; and classroom configuration (integrated vs. segregated) accounted for less than one percent.
As mentioned earlier, the legal and ethical justifications for integration may be viewed as conditional on the determination that integration does not harm children. Whereas the findings of this study indicate some detrimental effect of integration (relative to segregation) for more delayed students, the amount of difference in language and cognitive development may be counterbalanced by the more subjective benefits of integration. These may include increased parental satisfaction and comfort with integrated classes, increased tolerance of diversity by students, and promotion of a general social attitude of inclusion, rather than exclusion, of individuals with disabilities.
These findings also imply that early childhood special educators in integrated settings should be careful to monitor the instructional and social environment, as well as student performance, to be sure that lower functioning students are receiving appropriate stimulation. In addition, successful integration may require structuring of the amount and nature of interaction between the students with marked disabilities and those with more highly developed skills.
Several areas of future research are indicated by this study. Observational studies examining potential causal factors for the differential effects of integration, particularly the interactions among students with disabilities, model students, and teachers, may contribute to our understanding of why these differences occur. This information could guide the development of programs to promote integration successfully among students of varying ability.
Other related areas of research might also include examination of individual differences in the social domain. Our study found individual differences in language development, which suggests that social development may also be influenced. Finally, knowledge of the effects of different levels of classroom diversity (the ratio of models to children with disabilities) would be useful.
Allen, K. (1980). Mainstreaming in early childhood education. Albany, NY: Delmar Press.
Allen K., Benning [Mills}, P., & Drummond, W. (1972). Interaction of normal and handicapped children in a behavior modification preschool: A case study. In G. Semb (Ed.), Behavior analysis and evaluation (pp. 127-44). Lawrence, KS: University of Kansas Press.
Blackmon, A., & Dembo, M. (1984). Prosocial behaviors in a mainstreamed preschool. Child Study Journal, 14, 205-215.
Bricker, D. (1978). A rationale for the integration of handicapped and nonhandicapped preschool children. In M. Guralnick (Ed.), Early intervention and the integration of handicapped and nonhandicapped children (pp. 3-26). Baltimore: University Park Press.
Bricker, D., & Sandall, S. (1979). Mainstreaming in preschool programs: How and why to do it. Education Unlimited, 1, 29.
Cavallaro, S., & Porter, R. (1980). Peer preferences of at-risk and normally developing children in a preschool mainstream classroom. American Journal of
Mental Deficiency, 84, 357-366.
Cooke, T., Ruskus, J., Apolloni, T., & Peck, C. (1981). Handicapped preschool children in the mainstream: Background, outcomes, and clinical suggestions. Topics in Early Childhood Special Education, 1(1), 73-83.
Dale, P., & Cole, K. (1988). Comparison of academic and cognitive programs for young handicapped children. Exceptional Children, 54, 439-447.
Dunn, L., & Dunn, L. (1981). Peabody Picture Vocabulary Test-Revised. Circle Pines, MN: American Guidance Service.
Fey, M. (1986). Language intervention with young children. San Diego: College-Hill Press.
Furrow, D., Nelson, K., & Benedict, H. (1979). Mothers' speech to children and syntactic development: Some simple relationships. Journal of Child Language, 6, 423-442.
Galloway, C., & Chandler, P. (1978). The marriage of special and generic early education services. In M. Guralnick (Ed.), Early intervention and the integration of handicapped and nonhandicapped children. (pp. 261-287). Baltimore: University Park Press.
Guralnick, M. (1978). Early intervention and the integration of handicapped and nonhandicapped children. Baltimore: University Park Press.
Guralnick, M. (1980). Social interaction among preschool handicapped children. Exceptional Children, 46, 248-253.
Hresko, W., Reid, D., & Hammill, D. (1981). Test of Early Language Development. Austin, TX: Pro-Ed.
Ispa, J. (1981). Social interaction among teachers, handicapped children and nonhandicapped children in a mainstreamed preschool. Journal of Applied Developmental Psychology, 1, 231-250.
Ispa, J., & Matz, R. (1978). Integrating handicapped and preschool children within a cognitively oriented program. In M. Guralnick (Ed.), Early intervention and the integration of handicapped and nonhandicapped children (pp. 167-190). Baltimore: University Park Press.
Jenkins, J. R., Cole, K., Dale, P., & Mills, P. (1989). A longitudinal comparison of two preschool instruction models (Grant no. G008400646). Washington, DC: Department of Education; Special Education Programs.
Jenkins, J. R., Odom, S., & Speltz, M. (1989). Effects of social integration on preschool children with handicaps. Exceptional Children, 55, 420-428.
Jenkins, J. R., Speltz, M., & Odom, S. (1985). Integrating normal and handicapped preschoolers: Effects on child development and social interaction. Exceptional Children, 52, 7-17.
MacMillan, D. (1982). Mental retardation in school and society. Boston: Little, Brown.
McCarthy, D. (1972). McCarthy Scales of Children's Abilities. New York: The Psychological Corporation.
Odom, S., & McEvoy, M. (1988). Integration of young handicapped children and normally developing children. In S. Odom & M. Karnes (Eds.), Early intervention for infants and children with handicaps: An empirical base (pp. 241-267). Baltimore: Paul H. Brookes.
Pedhazur, E.J. (1982). Multiple regression in behavioral research (2nd ed.). New York: Holt, Rinehart and Winston.
Reid, D., Hresko, W., & Hammill, D. (1981). Test of Early Reading Ability. Austin, TX: Pro-Ed.
Rogers-Warren, A., Ruggles, T., Peterson, N., & Cooper, A. (1981). Playing and learning together: Patterns of social interaction in handicapped and nonhandicapped children. Journal of the Division for Early Childhood, 3, 56-63.
Shaltz, M., & Gelman, R. (1973). The development of communication skills: Modification in the speech of young children as a function of the listener. Monographs of the Society for Research in Chld Development No. 152, 38(5).
KEVIN N. COLE (CEC Chapter #389) is a Research Scientist, PAULETTE E. MILLS is a Project Coordinator, PHILIP S. DALE is an Associate Professor, and JOSEPH R. JENKINS (CEC Chapter #318) is a Professor at the Child Development and Mental Retardation Center at the University of Washington, Seattle.
|Printer friendly Cite/link Email Feedback|
|Author:||Cole, Kevin N.; Mills, Paulette E.; Dale, Philip S.; Jenkins, Joseph R.|
|Date:||Sep 1, 1991|
|Previous Article:||Cooperative teaching project: a model for students at risk.|
|Next Article:||Vocational technical programs: follow-up of students with learning disabilities.|