Off-campus South Dakota State University users: To download campus access theses, please use the following link to log into our proxy server with your South Dakota State University ID and password.
Non-South Dakota State University users: Please talk to your librarian about requesting this thesis through interlibrary loan.
Dissertation - University Access Only
Doctor of Philosophy (PhD)
Department / School
Mathematics and Statistics
Structural equation models represent the convergence of relatively independent research traditions in psychology, psychiatry, econometrics, and biometrics. It is a multivariate technique which combines the different statistical methodologies, such as regression analysis, simultaneous equations, path analysis, and latent factor analysis under the framework of Structural Equation Modeling (SEM). In the past few years, using SEM, there has been an explosion of research interests in modeling and analyzing multivariate longitudinal data. This dissertation research has four chapters that focus on diffferent topics of structural equation modeling. The first chapter gives a brief introduction to structural equation modeling. The second chapter analyzed factors influencing teaching as a career of choice using structural equation modeling. Data administered to 458 math/science college students were collected at a Midwestern university. The research objective was to investigate whether any motivational and/or social influence factor had any impact in choosing a career in teaching. Results of our analysis show that intrinsic motivation and social influence factors play a great role in choosing a career in teaching. The third chapter, describes comparative study of structural equation models vs. alternative models for multivariate longitudinal data with a research objective of comparing and contrasting the evolution of association and the association of evolution of the dependent responses. The alternative models chosen were linear mixed effect models (LME) and univariate approach to modeling multivariate data. Simulated longitudinal data using the open source R package have been generated for this study. Results of our study show that the univariate model always follows a trajectory that runs below the SEM or LME trajectories. The fourth chapter, analyzes the survey data administered to Americans age 66 and above, on two waves (in 2001 and in 2004) were obtained through the ICPSR (Inter-university Consortium for Political and Society Research) webpage. We investigated the multivariate longitudinal analysis of the effects of depressive symptoms, financial strain, and self-rated health on spiritual connectedness with several demographic covariates such as race, gender, education, region, age and marital status. The results of our analysis show that black Americans are more spiritually connected than their white counterparts and women are more spiritually connected than men. Region and education are observed to have no effect on spiritual connectedness.
Library of Congress Subject Headings
Structural equation modeling
Spiritual life -- Longitudinal studies
Includes bibliographical references (pages 99-107)
Number of Pages
South Dakota State University
In Copyright - Non-Commercial Use Permitted
Padhy, Budhinath, "Analysis of Multivariate Longitudinal Data Using Structural Equation Modeling" (2014). Electronic Theses and Dissertations. 1541.