Cookies

We use cookies to ensure that we give you the best experience on our website. You can change your cookie settings at any time. Otherwise, we'll assume you're OK to continue.

Durham University

Wolfson Research Institute for Health and Wellbeing

Previous Events

WASP 4: Multi-level model (4 day course)

29th June 2014, 10:00 to 2nd July 2014, 16:30, D026, Ebsworth Building, Queen's Campus Stockton

Registration for this training will open shortly

This module will discuss analysis of longitudinal and clustered data. Most real life phenomena do not occur as a single point event, but as repeated occurrences over time or groups of occurrences with defined and meaningful clusters. Analysis of correlated data using generalised linear models will be discussed. Unlike in generalised linear models, longitudinal data analysis involves investigation of average profiles as well as changes in individual profiles relative to the average profile. This module will introduce linear mixed effect models for continuous data with an emphasis on the interpretation of results for both the fixed and random components. The mixed effect model will be extended to incorporate both random effects and correlated errors. The last part of the module will focus on generalised linear mixed effect model (GLMM) for binary and count data with an emphasis on the conditional interpretation of results.

Specific Topics

Gaussian Longitudinal Data

  • Regression model with independent errors
  • Regression model with correlated errors
  • Exploring correlation structures
  • Linear mixed effect model (LMM)
    • LMM with random intercept model and uncorrelated errors
    • LMM with random intercept model and correlated errors
    • LMM with random intercept, random slope and uncorrelated errors
    • LMM with random intercept, random slope and correlated errors
  • Model building and variable reduction

Gaussian Clustered Data

  • Regression model with exchangeable correlation
  • LMM with random intercepts and uncorrelated errors
  • LMM with random intercepts and correlated error

Binary Longitudinal Data

  • Generalised estimating equations
  • Introduction to generalised linear mixed model (GLMM)
  • Binomial-Normal random intercept model
  • Binomial-Normal random intercept and random slope

Binary Clustered Data

  • Generalised estimating equations
  • Generalised linear mixed model (GLMM)
  • Binomial-Normal random intercept model
  • Binomial-Normal random intercept and random slope model

Counts Longitudinal Data

  • Generalised estimating equation
  • Generalised linear mixed model (GLMM)
  • Poison-Normal random intercept model
  • Poison-Normal random intercept and random slope model

Prerequisite

This module requires participants to have good conceptual understanding of regression model and generalised linear models. Participants would be required to demonstrate they have taken our courses on Applied Linear Regression and Generalised Linear model or any other relevant quantitative course.

Time Day 1 Day 2 Day 3 Day 4
10:00 –12:00 Overview of GLM Multilevel Gaussian data (II) Multilevel binary data (I) Multilevel binary data (III)
12:00 – 13:00 Lunch Lunch Lunch Lunch
13:00 – 15:00 Multilevel Gaussian Data (I) Multilevel Gaussian data (III) Multilevel binary data (II) Multilevel count data
15:00 – 15:30 Coffee break Coffee break Coffee break Group presentation
15:30 – 16:30 Group work Group work Group work