Term of Award
Spring 2016
Degree Name
Doctor of Public Health (Dr.P.H.)
Document Type and Release Option
Dissertation (restricted to Georgia Southern)
Copyright Statement / License for Reuse
This work is licensed under a Creative Commons Attribution 4.0 License.
Department
Department of Biostatistics (COPH)
Committee Chair
Hani Samawi
Committee Member 1
Daniel Linder
Committee Member 2
Haresh Rochani
Abstract
Mediation is a hypothesized causal chain among three variables. The mediation analysis for continuous response variables is well developed in the literature, and it can be shown that the indirect effect is equal to the total effect minus the direct effect (ab=c-c'). However, mediation analysis for categorical responses is still not fully developed. The purpose of this dissertation was to develop a simpler method of analyzing the mediated effect among three variables when the dependent and mediator variables were both dichotomous. This approach made mediation analysis more convenient and efficient. This dissertation proposed and developed a new approach by using the latent variable technique to adjust for ab=c-c'. Our intensive simulation study for different probabilities of success and different parameter settings showed that on average the proportion was b1/b=0.412. Another aim of this study was to use the proposed method and the Winship and Mare method to test the potential mediators on the relationship between depression and obesity among children and adolescents. We found that low level of physical activities and binge TV watching were the mediators on the relationship between childhood depression and obesity and that the proposed method was more efficient than the Winship and Mare method.
Recommended Citation
Cai, Jingxian, "A Simpler Approach for Mediation Analysis for Dichotomous Mediators in Logistic Regression: An Application to Children’s Health Conditions Associated with Obesity" (2016). Electronic Theses and Dissertations. 1385.
https://digitalcommons.georgiasouthern.edu/etd/1385
