Master's degree in Statistics and Operations Research

Multi-Level Modeling and Longitudinal Data Analysis with Monte-Carlo Simulations.

Date:

June 22, 23, 25, and 26. MORNING: 9 to 13h (Mo, Tu,Th) and 9 to 12h (Fr).

Instructor

(Din) Ding-Geng Chen, Ph.D. South Africa SARChI Chair in Biostatistics and Extraordinary Professor, Department of Statistics, University of Pretoria, Pretoria, South Africa

Executive Director and Professor in Biostatistics, College of Health Solutions, Arizona State University, Phoenix, USA 

Professor Din Chen received his Ph.D. in Statistics from University of Guelph (Canada) in 1995 and is now the SARCHI research chair (Tier 1) with NRF B1-rating. He is an extraordinary professor in biostatistics at the Department of Statistics, University of Pretoria, and an honorary professor at the University of KwaZulu-Natal, South Africa. He is an elected fellow of the Royal Society of South Africa (FRSSAf), an elected member of Academy of Science of South Africa (MASSAf),  an elected fellow of the American Statistical Association (FASA), and an elected member of the International Statistical Institute (ISI). Professor Chen is also the executive director and professor in Biostatistics at the College of Health Solutions, Arizona State University. He served as the Wallace H. Kuralt Distinguished professor in biostatistics at the University of North Carolina-Chapel Hill, biostatistics professor at the University of Rochester Medical Center, the Karl E. Peace endowed eminent scholar chair in biostatistics from the Jiann-Ping Hsu College of Public Health at the Georgia Southern University. Professor Chen has more than 300 scientific publications and co-authored/co-edited 43 books on clinical trials, survival data, meta-analysis, causal inference and structural equation modeling, Monte-Carlo simulation-based statistical modelling. His research has been funded as PI/Co-PI from NIH R01s and other governmental agencies.

https://search.asu.edu/profile/4022508

Language

English

Description

This short course is designed to introduce the context and intuition with comprehensive treatment on multi-level modeling and longitudinal data analysis with Monte-Carlo simulation-based approach using R.

Course goals

Multi-level modeling (MLM) and longitudinal data analysis are commonly used in public health and social sciences. This short course emphasizes the principles of statistical reasoning, underlying assumptions, and careful interpretation of results for multi-level modeling and longitudinal data analysis. Topics covered include intra-class correlation (ICC), random-intercept and random-slope mixed-effects models for two-level and three-level multi-level modeling, longitudinal data analysis with linear and nonlinear growth-curve models. The general goals of this short courses are 1) to study the principles of statistical reasoning and underlying assumptions with Monte-Carlo simulations, 2) to interpret results for data analysis and multi-leveling modeling and longitudinal data analysis, 3) to analyze real-life data using R and report writing using RMarkdown.

Course contents

  1. Introduction to R/RStudio and RMarkdown (Reference to Chapter 1 in Chen24)
    • Overview of R and RStudio
    • Introduction to RMarkdown for reproducible reporting
  2. Review of Linear Regression (Reference to Chapter 1 in Chen21, Chapters and 2 in Chen24)
    • Review of linear regression and its assumptions
    • Real data analysis with R
    • Monte-Carlo simulation for assumption validation in linear regression
      • Review of commonly used statistical distributions
      • Simulating data from these distributions
      • Validating simulated distributions (uniform, normal, t-distribution, F-distribution, and χ² distribution)
      • Review of parameter estimation, confidence intervals, and hypothesis testing
      • Validating these statistical techniques through Monte-Carlo simulations
  1. Introduction to Multi-level Modeling (Reference to Chapter 2 in Chen21)
    • Multi-level data structure
    • Intra-class Correlation(ICC)
    • Why multi-level modeling with random-intercept and random-slope mixed-effects models

   4. Multi-level Modeling using R package “nlme” (Reference to Chapters 2 and 3 in Chen21)

    • Two-level multi-level modeling using R with real data analysis
    • Three-level multi-level modeling using R with real data analysis
    • Monte-Carlo simulation on why mixed-effects models should be used. 
  1. Longitudinal Data Analysis using R (Reference to Chapter 5 in Chen21)
    • Introduction to longitudinal data analysis
    • Introduction of R package “nlme” for “linear mixed-effects model”
    • Validation of statistical modeling using Monte-Carlo simulations 

   6. Nonlinear Regression and Nonlinear mixed-effects Modeling using R

    • Review of nonlinear regression with logistic growth model
    • Real data analysis using “nls”
    • Nonlinear multi-level mixed-effects modeling with R “nlme”
    • Read data analysis using R

References

    1. (Chen21) Chen, D. G. and Chen, J. K. (2021). Statistical Regression Modeling Using R: Longitudinal and Multi-Level Modeling. Springer.
    2. (Chen24) Chen, J. K. (2024). Financial Data Analysis with R: Monte-Carlo Validation. Chapman and Hall/CRC.

Prerequisites

Students who have basic understanding on: 1. Statistical distributions, point and interval estimation in statistical inference 2. Linear regression 3. R computing (not required, but will be helpful)

Targeted at

Masters, Ph.D students and any statistical analysts who are interested in multi-level modeling and longitudinal data analysis.

Teaching Methodology and Activities

  1. In-person lecturing with guided data analysis step-by-step using R.
  2. Students are required to have R/Rstudio installed and will be guided to use RMarkdown for data analysis during the lectures
  3. R library “nlme” should also be installed and loaded.  
  4. Students will be grouped for R data analysis and exercises and homeworks.

Evaluation

A project will be assigned to work together for each group of students. Each group will prepare a final presentation on Friday

Software requirements

R, which can be freely downloaded and installed at https://posit.co/downloads/.