Welcome to SREE's online HLM course. We created this course by organizing videos from our past HLM summer intensives and augmenting them with additional insights and information to aid those interested in learning HLM methods.

About This Course »
Hierarchical linear models have evolved as an important tool in education research. Many studies in education involve data that have a complex structure. They may contain multiple levels, such as students clustered within schools, or they may be longitudinal. Because of these structures, simpler linear regression models may not be appropriate. Instead, we ought to analyze these types of data in a way that accounts for their structure. Doing so with hierarchical linear models not only provides correct inferences, but it can also shed light on relationships within the data that might otherwise go unnoticed.

This course provides an introduction to the theory and application of hierarchical linear models. It discusses key considerations for fitting and interpreting 2-level, 3-level, and longitudinal models. These include a general approach to the mathematical notation and terminology of hierarchical models, a discussion of the logic and theory of the models, and practice applying them to real data. Completing this course will train you not only to think and talk in terms of hierarchical models, but also to fit and interpret them.
Approaching the Course »
The course is designed to allow you to learn at your own pace. Key concepts are divided into learning modules, and later modules tend to build on previous ones. Modules contain video lectures and written lessons, and take 30-60 minutes to complete. Lectures are edited from SREE summer intensives run by Steve Raudenbush (University of Chicago) and Tony Bryk (Carnegie Foundation), and run between 30 and 60 minutes in length. Lessons are designed to work in concert with the lectures, so that they augment and clarify the material presented in the videos. Moreover, the self-assessments and other prompts within the lesson provide extra practice to master important concepts.

A key component of this course is actually fitting hierarchical linear models, and multiple modules provide guided practice to do so. The lectures and lessons both use the HLM7 software to do this. A free student version of HLM7 is available for download so that you can follow along with examples or practice fitting other models.

Finally, there are two potential resources that might help you as you progress through the modules. First, each module has a comment section, where you can ask questions that will be answered by SREE personnel, as well as other users. In fact, if you see that someone has a question that you can help them with, we encourage you to do so. It not only gives you a chance to work through the material, but it helps build an active learning community within the course. Second, the textbook Hierarchical Linear Models: Applications and Data Analysis Methods, 2nd Edition (Sage, 2002) by Raudenbush and Bryk gives a deeper dive into theory and more examples, and it describes other potential applications.

CLICK A MODULE TO START



Module 1: Introduction to HLM

Brief overview of the history, development, and utility of HLMs
35 – 40 minutes (approximate time)

Module 2: 2-Level HLM Notation

Notational conventions for defining 2-level HLMs
45 – 50 minutes (approximate time)

Module 3: HLM Notation with High School and Beyond

Defining a 2-level HLMs for the High School and Beyond (HSB) dataset
30 – 35 minutes (approximate time)

Module 4: HLM Software for 2-Level HLMs I

Getting started with HLM7 software for estimating 2-level HLMs
35 – 40 minutes (approximate time)

Extension Module 4: R for 2-Level HLMs I

Getting started with R for estimating 2-level HLMs
35 – 40 minutes (approximate time)

Module 5: HLM Software for 2-Level HLMs II

Understanding and interpreting HLM7 output for 2-level HLMs
45 – 50 minutes (approximate time)

Module 6: Data Analysis Exercise I

Practice analyzing data from the Tennesee class size experiment with 2-level HLMs
35 – 45 minutes (approximate time)

Module 7: Centering, Fixed vs. Random, & More

Discussion of important modeling and coding choices
30 – 35 minutes (approximate time)

Module 8: 3-Level HLM Notation

Notational conventions for defining 3-level HLMs
45 – 50 minutes (approximate time)

Module 9: HLM Notation with Tennessee STAR Data

Defining 3-level HLMs for the Tennesee slass size experminent (STAR project) dataset
30 – 35 minutes (approximate time)

Module 10: Data Analysis Exercise II

Practice analyzing data from the Tennesee class size experiment with 3-level HLMs
35 – 40 minutes (approximate time)