First Advisor

Holighi, Khalil Shafie

Second Advisor

Lalonde, Trent L.

Document Type

Dissertation

Date Created

5-1-2017

Department

College of Education and Behavioral Sciences, Applied Statistics and Research Methods, ASRM Student Work

Embargo Date

5-23-2019

Abstract

In analysis of longitudinal data, a number of methods have been proposed. Most of the traditional longitudinal methods assume that the independent variables are not dependent on time and are the same across study. However, one of the main advantages of longitudinal studies is the ability to observe outcomes and covariates at the same time, and a researcher can define whether changes in a covariate lead to changes in the outcome of interest. In addition, the methods focused on a predetermined observation time that does not carry information about the response variable. Moreover, it is possible in real research to have time-varying covariates, unbalanced observation time, and the observation times may be informative. The usual longitudinal statistical analysis might be biased if their assumptions are not valid. The purpose of this study was to develop a joint model of a longitudinal outcome and informative time with time-dependent covariates. In this study, a joint model and analysis of longitudinal data with possibly informative observation times and time-dependent covariates via joint probability distributions has been proposed. The maximum likelihood parameter estimates of the proposed model were obtained from Monte Carlo simulated data by employing a nonlinear optimization in R. Furthermore, the model selection criteria and likelihood ratio test statistic were computed to select the best fitting model and for comparing nested models. Additionally, the R codes were developed for the proposed model and an application is presented on the bladder cancer data used for explanation purposes. In the application, the results show that the time-dependent covariate appear to be important predictor in the longitudinal data.

Keywords

Informative Time, Longitudinal Data, Time-Dependent Covariate

Extent

130 pages

Local Identifiers

Alomair_unco_0161D_10566

Rights Statement

Copyright belongs to the author.

Download

Share

COinS