Longitudinal Data Analysis and Visualization

In this course, you will learn how to analyze data gathered over multiple occasions and how to visualize the results of your analysis. Topics include:

• Brief orientation to linear mixed effects (a.k.a. multilevel) models
• Model estimation, random effects, and strategies for modeling residual error
• Alternative ways to model the passage of time
• Modeling non-linear change  
• Incorporating time-invariant predictors of change
• Incorporating time-varying predictors of change
• Brief overview of model adaptations for longitudinal categorical outcomes
• Strategies for visualizing change over time   

The course meets in person in Calgary July 15-16, 2025, from 9.00am to 5.00pm each day.

We will make use of R statistical software, primarily using the lme4 package and the ggplot2 package for data visualization. For those who prefer not to use R for data analysis, supplementary code will be available for SAS software. Students are encouraged to bring a laptop to class with a recent installation of R software, the RStudio integrative development environment (download for free at https://posit.co/downloads/), and the lme4 and ggplot2 packages already installed. Prior experience using R software is helpful but not required.

This course is meant for researchers and data users in any field who use repeated measures data to better understand individual growth and change. Learners should have a background in introductory statistics and regression analysis at the graduate level. Prior knowledge of mixed effects or multilevel modeling is not required, but students with this background may follow along more easily. Some prior experience with R software is also helpful, but not required. Ideally, participants have some background working with data and running statistical analyses in at least one type of software suited for linear modeling (e.g., R, SPSS, SAS, or STATA).

Upon completion of this course, you will:

• Understand the requirements for a model assessing change over time
• Learn how to interpret and use random effects in models of change
• Be able to choose and implement linear and non-linear forms of change over time
• Understand how to separate within- and between-person effects in time-varying covariates
• Know how to estimate trajectories of change for people with different characteristics
• Know how to test whether people depart from their own trajectories when they encounter different time-specific events
• Be able to create high-quality visualizations of longitudinal change using output from any software program
• Learn what elements of longitudinal data analysis are similar and different when the outcome variables are categorical.

A certificate of completion from the Canadian Centre for Research Analysis and Methods is provided at the end of the course.

Longitudinal data analysis is an extensive topic with many advanced applications. We will not be discussing longitudinal analysis in the context of structural equation modeling, for which I recommend that students first complete an introductory SEM course. We also won’t have time to discuss approaches for non-parametric change (e.g., generalized additive models (GAM), time-varying effect models (TVEM)) or dynamic systems and time-series models appropriate for intensive longitudinal data, though I will provide links to suggested resources for further reading. Our coverage of non-linear outcomes will also be brief. The foundational skills you acquire in this course will prepare you well to explore more advanced topics on your own.

"I absolutely loved this course and thought Andrea was great. I loved all the hands on work in R and being able to either type along with her or open the already complete file and just run it. Her slides were so clear and I am so glad I can keep going back to them. I also really liked how she handled questions and incorporated what participants asked into the presentation."