Multilevel and Mixed Models Using Stata
A 3-Day Remote Seminar Taught by
Stephen Vaisey, Ph.D.
To see a sample of the course materials, click here.
This seminar provides an intensive introduction to multilevel models, a class of regression models for data that have a hierarchical (or nested) structure. Common examples of such data structures are students nested within schools or classrooms, patients nested within hospitals, or survey respondents nested within countries.
Using regression techniques that ignore this hierarchical structure (such as ordinary least squares) can lead to incorrect results because such methods assume that all observations are independent. Perhaps more important, using inappropriate techniques (like pooling or aggregating) prevents researchers from asking substantively interesting questions about how processes work at different levels.
Starting September 24, we are offering this seminar as a 3-day synchronous*, remote workshop for the first time. Each day will consist of a 4-hour, live morning lecture held via the free video-conferencing software Zoom. Participants are encouraged to join the lecture live, but will have the opportunity to view the recorded session later in the day if they are unable to attend at the scheduled time.
Each lecture session will conclude with a hands-on exercise reviewing the content covered, to be completed on your own. An additional session will be held Thursday and Friday afternoons as an “office hour”, where participants can review the exercise results with the instructor and ask any questions.
*We understand that scheduling is difficult during this unpredictable time. If you prefer, you may take all or part of the course asynchronously. The video recordings will be made available within 24 hours of each session, meaning that you will get all of the class content and discussions even if you cannot participate synchronously.
MORE DETAILS ABOUT THE COURSE CONTENT
After a brief conceptual introduction (including a discussion of the difference between random and fixed effects), we will begin with simple variance components models that can tell us how much of the variation in a measure can be attributed to different levels of observation. We will then move on to mixed models (random effects models with fixed covariates) that allow us to ask how factors at different levels can affect an outcome.
Next, we will investigate how using random coefficients and cross-level interactions can help us discover hidden structure in our data and help us investigate how individual-level processes work differently in different contexts. We will also briefly consider how these techniques can be applied to cases where we have repeated observations of individuals or other entities over time.
Although the course will focus primarily on the continuous outcome case, we will also cover how these models can easily be extended for use with categorical and limited dependent variables.
The seminar will focus on hands-on understanding and draw from examples across the social and behavioral sciences. After completing the course, students will:
- Know the technical and substantive difference between fixed and random effects
- Understand random intercepts, random coefficients, and crossed random effects models and know when to use each one
- Know how to combine the strengths of random-effects and fixed-effects approaches into a single “between-within” model
- Know how to estimate these models and interpret the result
This remote seminar is held via Zoom, a free video conferencing application. Instructions for joining a session via Zoom are available here. Before the seminar begins, participants will receive an email with the meeting code and password you must use to join.
The vast majority of what you will learn in this course can be applied in any software package. This seminar will use Stata 16 for empirical examples and exercises. (Nearly all commands will work in Stata 14 or 15 as well.) However, no previous experience with Stata is needed. Participants who request it can also get the R notes and syntax as well.
WHO SHOULD Register?
This course is for anyone who want to learn to apply multilevel models to observational data. Participants should have a basic foundation in linear regression.
– what are multilevel models?
– why use multilevel models?
– within and between unit variation
– hierarchical data-generating processes
– complete pooling, partial pooling, and no pooling
– intro to practice data: life satisfaction in Europe
– intraclass correlation and variance components
– estimation in R using lme4
– partial pooling, random effects, and empirical Bayes
– introduction to linear “mixed” models
– introduction to lme4 syntax
– within, between, and total R-squared
– interpreting model coefficients
– random coefficients/slopes
– model selection
– basic diagnostics
– random and fixed effect model assumptions
– linear fixed-effect (no pooling) models
– Hausman and other tests of assumptions
– between-within (BW) or “hybrid” model
– cross-level interactions
– limited dependent variables: binary and count
– workflow and presenting results
“An excellent course. I felt like I went from no knowledge of mixed level models to feeling extremely competent at knowing when mixed level models may be used and how to use these models in just 2 days. Dr. Vaisey is an outstanding educator. This interactive, practical teaching style is extremely effective.”
Oluwaseun Falade-Nwulia, Johns Hopkins University
“Dr. Vaisey’s instruction helped me connect the dots of my previous understanding of relevant statistical methods. I feel more confident with applying what I learned to my future data analysis. Thank you. Dr. Vaisey!”
Yang Yang, Rowan University
“Dr. Vaisey uses techniques to teach MMMs that you cannot find in any single textbook and provides his recommended practices for model fit comparison (e.g. BIC and between within method) and why to use MMMs based on theoretically motivated questions.”
David Rothwell, Oregon State University
“This course was extremely helpful for reacquainting me with the mixed/multilevel model frameworks. It was not super heavy on formulas or syntax and focused mostly on grasping the basic understanding of partitioning variance and how to think about building a model that is most appropriate to a research question.”
Heidi Grunwald, Temple University
“Steve is the professor I wish I’d had in graduate school. He is a black belt at theory and technical details, and has the ability to communicate the materials in a way that helps you to grow an intuition. This is a rare quality in a statistician and teacher, and Steve nails it. He exhibits humor, thoughtful questions and responses, and the ability to anticipate where people get “stuck.” Take a course from Steve and you’ll be glad you did it!”
Andy Bogart, RAND Corporation
“Stephen is an excellent instructor. He has clearly thought deeply about these types of models and is very knowledgeable about the methods. He was available for questions and willing to discuss student’s specific data with them. I am leaving this course with not only knowledge about how to analyze nested data, but also with knowledge about what methods are conventional for certain fields and what methods will be most recognized by others. I appreciate this meta-perspective. I would highly recommend this course for anyone who needs to gain a more thorough understanding of MLM and the differences between fixed and random effects.”
Kelsey Thiem, University of Massachusetts