Date of Award
2018
Document Type
Open Access Dissertation
Department
Psychology
Sub-Department
College of Arts and Sciences
First Advisor
Amanda Fairchild
Abstract
Integrating discrete time survival and mediation analytic approaches, discretetime survival mediation models (DTSM) help researchers elucidate the impact of predictors on the timing of event occurrence. Though application of this model has been gainful in various applied developmental and intervention research contexts, empirical work has yet to consider how DTSM models operate with a mediator that has a varying effect over time. The importance of examining this situation has important impacts for application of the model, given more complex statistical models are required, and subsequent interpretation of model parameters differ from the basic DTSM model. The overarching purpose of this dissertation was to understand how the addition of a mediator with a time variant effect impacts parameter estimation and fit of the DTSM model estimated in a mixture modeling framework. This investigation was done within the context of an applied example (Study One) to simultaneously inform applied considerations in timing to onset of youth alcohol use, as well as to evaluate statistical performance of the model in a related single-cell Monte Carlo study (Study Two) and an expanded simulation study (Study Three). Results are presented with discussion of future directions for this research and considerations for application of this modeling approach
Rights
© 2018, Heather Lasky McDaniel
Recommended Citation
McDaniel, H. L.(2018). Advancing Understanding Of Dynamic Mechanisms In Onset To Event Models: Discrete Time Survival Mediation With A Time Variant Mediator. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/4673