Previous empirical studies have been inconsistent in the identification of cognitive differences between LD and EH children on psychoeducational assessment batteries. Furthermore, studies using multivariate procedures such as discriminant analysis have failed to verify the basic underlying assumption of homogeneity of variance-covariance matrices of the groups. Homogeneity of covariance can be assessed by comparing factor structures of the two groups.The purpose of this study was to investigate the cognitive differences between LD and EH children on a psychoeducational assessment battery Including the WISC-R, WRAT, PIAT, PPVT, and DVMI, typically used to identify the two exceptionalities. Differences In factor structures were investigated in an effort to clarify group differences in cognitive functioning, and to investigate the utility of multivariate analyses such as discriminant analysis with these two groups of children.The subjects were 1165 public school children, aged 6 to 16, referred for assessment and subsequently classified and served in special LD and EH programs. Four and five factor solutions were compared using the congruence coefficient to determine statistical similarities. The first three factors of each solution, Verbal Concepts, Verbal Achievement, and Visual Perceptual Reasoning, were found to be statistically similar in construct and interpretation. These results suggest that If the use of multivariate procedures in studying LD and EH group differences is limited to cognitive performance in the areas of verbal concepts, reading and spelling achievement, and visual perceptual reasoning skills, the results can be accepted as valid. The discriminant analysis would not likely be violating the homogeneity assumption.Differences were found in factor structures Involving the order of the factors extracted in terms of importance In explaining variance. Also, statistically significant differences were found between the factors Involving Math, Visual Motor, and Sequential skills, suggesting that multivariate analyses using these factors would likely violate the homogeneity assumption.Results indicate the presence of heterogeneity in the matrices of LD and EH children, and suggest caution in the interpretation of multivariate statistical analyses with these children. Violation of the assumption of homogeneity can invalidate statistical findings and their application to the study of group differences.
| Identifer | oai:union.ndltd.org:BSU/oai:cardinalscholar.bsu.edu:handle/178083 |
| Date | January 1987 |
| Creators | Mason, Elizabeth M. |
| Contributors | Wenck, L. Stanley |
| Source Sets | Ball State University |
| Detected Language | English |
| Format | vi, 76 leaves ; 28 cm. |
| Source | Virtual Press |
Page generated in 0.0021 seconds