3 credits
Fall 2026 Lecture
This course gives students a basic grounding in the class of statistical techniques known as multilevel modeling (MLM), also known as hierarchical linear modeling (HLM), mixed models, or random coefficient models. Primary discussions will be on applications of these models to the study of marriages, relationships, families, aging, and child and adult development, but also will touch on biomedical, educational, and economic examples. Nesting in both the contextual and longitudinal data situations are examined. Three-level MLMs are also covered. Data preparation, hypothesis testing and estimation approaches for MLMs are introduced throughout the course. While mathematical basics of statistical methods are covered, emphasis is placed on model development, the conceptual understanding of models, and interpretation of model results. Students are assumed to have taken at least two graduate statistics courses and have a solid understanding of regression analysis. Prerequisites: STAT 50100 and STAT 50200 or HDFS 61300 and HDFS 61700, or other similar sequence with instructor approval.
Learning Outcomes1Develop a basic understanding of multilevel models including proper application, interpretation, and evaluation of the models.
2Learn the benefits and limitations of MLMs including when it is advantageous to use this modeling approach.
3Develop an understanding of the data structure and visualization techniques related to MLM.
4Develop an understanding of the underlying statistics including model notation, model structure, and hypothesis testing.
5Run MLM data analysis using R statistical software.
6Improve analytic and critical thinking skills.
7Improve written and verbal communication of analytic results.
8Use MLM methods to test research questions using real data and statistical software.
Restrictions
made by Purdue students.