Methods Section

This research used a quasi-experimental design for third through fifth-grade students from 12 intact classrooms. There were 207 students in the study, which eliminated the possibility of the random assignments of participants. Students were examined in three groups: 68 students were in Group 1 from four noninclusion classrooms; 34 students were in Group 2 from two clustered inclusion classrooms; and 105 students were in Group 3 from six random inclusion classrooms.

Six variables were defined: 1) parental concerns; 2) teacher and parent-reported instances of problem behaviors on the part of students; 3) the academic performance of the students; 4) and student self-reported self-esteem. The researchers used three standardized instruments to measure the variables of interest, these being the CBCL, SEI, and SAT (mathematics, reading, spelling, and language subscales. The researchers also used their own parent concern questionnaire. The researchers explain each of the tests and what each measures as well as how each is to be scored. The results are set forth in tables for easier comparison along with indicating the means of analysis, the analysis of variance (ANOVA) to show where the differences observed are significant and where they are not. When the results showed the need for some adjustment, the researchers used gain scores and pretest scores to achieve greater reliability.

The parent concerns questionnaire was a 22-item questionnaire with responses marked on a seven-point scale, ranging from "not true at all about me at this time" to "very true about me at this time." Previous psychometric data were not available for this study. The researchers therefore used factor-analytic procedures to gain information about the construct validity of the data generated, and so long as meaningful factors were found, the scores would be used to address the...

SAT data was collected from archival records kept in the schools involved. SEI data came directly from students who completed the instruments during normal classroom sessions. Teachers completed their portion and returned the protocols to the researchers, while parents were mailed their part with an explanation for the purpose of the study and a request for participation as well as an assurance of confidentiality.
Discriminant analysis was used to analyze the data. The response rate from parents was low, so that data was subject to separate analysis.

Results and Conclusion

The results included complete TRF, SEI, and SAT data for 178 of the 207 students. The researchers found a moderate degree of correlation between the type of program and the dependent variable cohorts, specifically in terms of reading gain scores, parent concerns, and subscales of the SEI. The grade level of the students also showed a difference so that third-grade-inclusion students showed higher gains in reading scores while fourth-grade-inclusion students showed smaller gains in mathematics. Consistent academic gains were not found to be the product of inclusion, suggesting a more careful approach to shifting to inclusion of improved academic performance is the goal. Students in inclusion classrooms exhibited more bad behavior than students not in such classrooms. Some evidence suggest that teachers in inclusion classrooms are more distracted, allowing for more bad behavior. Children in inclusion classrooms report uniformly lower self-esteem, shown for the group as a whole, and reduced self-esteem may inhibit academic achievement.

The results raise more questions about the usefulness of inclusion and question the rationale that has been offered for why inclusion will help these students. In fact, it seems that inclusion can create more problems than…