| David Kaplan
University of Delaware
Quantitative Approaches to Educational Policy Analysis Utilizing Multilevel Structural Equation Modeling
FINAL REPORT:
Study 1: A Multilevel Structural Model of Science Achievement from an Indicator System Perspective: Implications for Educational Policy Analysis (co-Author P.R. Elliott).
This paper considers the estimation, testing, and policy utilization of a structural model of science achievement from an indicator system perspective. Specifically, this paper extends previous research on science achievement by first recognizing and modeling the multilevel organizational structure of schooling and, second, by choosing variables for the model that have been considered essential indicators for gauging the health of science education. Muitilevel structural equation modeling was applied to a subsample of the First Follow-Up of the National Longitudinal Study of 1988 (NELS:88). A within school model of science achievement was linked to a between school model of the academic press of the school. Separate estimation of these models revealed adequate fit to the data after minor modifications. The multilevel model also showed adequate fit to the data. The model was then used to explore various policy options for increasing science achievement. Combined policies related to class grouping, teacher background, and time devoted to science instruction seem to have the greatest effect on increasing the predicted value of a number of indicators, including science achievement, compared to individual policies. In all cases, however, the effects were very small. The paper closes with a discussion of the limitations of the model, the potential for future model development, and the implications of the study for quantitative modeling within the domain of education policy.
Study 2: Multilevel Structural Equation Modeling for Organizational Studies: The Case of Education (Co-author P.R. Elliott).
This paper provides a didactic presentation and application of new developments in structural equation madeling that allow for the modeling of multilevel data. Such data often arise naturally from organizational structures wherein within group units (employees, students, etc.) are observed in larger between group units (firms, schools, etc.) The paper begins by overviewing the basic ideas of structural equation modeling and multilevel linear regression. The synthesis of both methods is then presented in the simple case of multilevel path analysis model wherein within group level parameters are modeled as a function of between group variables following their own path model. An application fallows that is motivated by a real problem in the field of education having to do with the monitoring of the health of science education in the United States. It is shown that it is possible to statistically capture the salient complexities of the educational organization through the application of multilevel structural equation modeling. The paper concludes with a discussion of the utility of multilevel structurai equation modeling for organizational studies.
Back to Funded Research Grants Page |