June 26 - 30, 2017
1:30 p.m. – 5:00 p.m.
Gives an overview of "multilevel statistical models" and their application in public health and biomedical research. Multilevel models are regression models in which the predictor and outcome variables can occur at multiple levels of aggregation: for example, at the personal, family, neighborhood, community and regional levels. They are used to ask questions about the influence of factors at different levels and about their interactions. Multilevel models also account for clustering of outcomes and measurement error in the predictor variables. Students focus on the main ideas and on examples of multi-level models from public health research. Students learn to formulate their substantive questions in terms of a multilevel model, to fit multilevel models using Stata during laboratory sessions and to interpret the results.
Student Evaluation: Final exam
Upon successfully completing this course, students will be able to: 1) prepare graphical and tabular displays of multilevel data that effectively communicate the patterns of scientific interests; 2) conduct statistical analyses of clustered data by use of multilevel models; 3) interpret parameters of multilevel statistical models; 4) fit multilevel models by use of statistical software packages.
Prerequisite: Previous experience with regression analysis is required.
Auditors Allowed: No
Grading Options: Letter Grade or Pass/Fail