Statistical Methods Course Description
An introduction to using multilevel analysis for grouped and longitudinal data, including analysis of categorical data.
This course is an introduction to multilevel techniques for both researchers and students. The course considers multilevel regression models in detail. It starts with an introduction to the basic two- and three-level regression model, estimation methods, and interpretation of results. Next, it discusses longitudinal models, and models for non-normal data such as multilevel logistic regression models. Although this is not a computer course, references are made to multilevel software packages, and there are software demonstrations on example data sets. The general structure of the course is a lecture in the morning and a computer lab in the afternoon for the first two days. The third day treats a variety of statistical issues, such as explained variance, standardizing coefficients, assumptions and robust estimation. Participants from a variety of fields—including psychology, education, human development, public health, prevention science, sociology, marketing, business, biology, medicine, political science, and communication—will benefit from the course.
Participants will receive an electronic copy of all course materials, including lecture slides, practice datasets, software scripts, relevant supporting documentation, and recommended readings. Participants will also have access to a video recording of the course.
- Multilevel regression for continuous outcomes, with two or three levels.
- Multilevel regression for longitudinal data.
- Random slope coefficients and variance structures.
- Multilevel regression for non-normal outcomes, such as dichotomous, ordered categorical and count data.
- Choices for different estimation methods, robust estimation.
- Strategies for stepwise analysis of multilevel data.
Note: This course focusses on observed variables. It does not cover multilevel analysis including latent variables, i.e. multilevel structural equation modeling. The emphasis is on multilevel regression models, it does not cover the more general multilevel path model. (##Reference to Summer Stats Camp Mplus course here?##)
Instructor: Joop Hox
Dr. Joop Hox (email@example.com) is emeritus professor of social science methodology and statistics at Utrecht University, the Netherlands. He has taught at Summer Stats Camp and at the first Stats Camp Euro in Utrecht. He has written a well-received handbook on multilevel modeling and more than 100 papers on a variety of methodological and statistical topics, including multilevel regression and structural equation modeling, survey methodology. He is a founding member of the European Association for Methodology (EAM) and serves on the EAM executive Board as past president. Joop Hox is a frequent consultant for colleagues from other fields in social and behavioral science, and known for his enthusiastic and accessible teaching style in introducing advanced analysis techniques for substantive audiences.
Course Software and Computer Support
Participants need to bring a laptop computer with Wi-Fi capabilities. The course will use the software HLM in the computer labs: a student version of HLM can be downloaded from www.ssicentral.com. Most techniques covered in this course can be applied in other statistical packages. Participants who own a statistical package that supports multilevel regression (e.g., SAS, SPSS, Stata) may use that, but will receive limited support in the computer lab.
All statistical software used at Stats Camp will be available, free to participants, on our SMORS (statistical modeling on remote servers) system for the duration of camp.
- A basic understanding of statistical inference, and some experience with analysis of variance and multiple regression analysis
- Intermediate proficiency with at least one statistical software package (e.g., SPSS, Stata, SAS, R, etc.).
Not required but advantageous:
- At least limited experience (e.g., graduate-level course) with multivariate data analysis.
- Intermediate proficiency in multiple linear regression, including the use of categorical independent variables.
- Some experience with binary logistic regression.
No level of proficiency beyond basic awareness is assumed for skills related to:
- Multilevel modeling in general.
- Advanced mathematical or statistical topics such as matrix algebra or likelihood theory.