PRCHN seminar: “Approaches to Multi-level Modeling”

The Prevention Research Center for Healthy Neighborhoods (PRCHN) will welcome national speaker Don Hedeker for the next monthly seminar series event.

Hedeker, professor of biostatistics in the Department of Public Health Sciences of the University of Chicago, will present a talk titled “Approaches to Multi-level Modeling” Wednesday, June 14, from noon to 1:15 p.m. in the BioEnterprise Building’s ground-floor conference room.

Hedeker’s expertise is in the development and use of statistical methods for clustered and longitudinal data, with particular emphasis on mixed-effects models. He will discuss study design and modeling approaches for multilevel data, including examples and applications of strategies.

Multilevel models—also known as hierarchical linear models and mixed models—were developed to analyze data from clustered designs, where individuals are members of a larger group or context. A common example of a clustered design is from school-based studies, in which students are assessed within grades and schools. The culture of the classroom, as well as the school, may have an impact on student responses. Another example would be familial studies, where one might be interested in the responses of both parents and children to interventions delivered to the families. Other examples include studies of individuals within social clubs, neighborhoods, wards, households or worksites.

Unlike ordinary regression analysis of clustered data, multilevel models do not assume that each observation is independent, but rather assumes data within clusters are dependent to some degree. The degree of this dependency is estimated along with estimates of the usual model parameters, thus adjusting these effects for the dependency resulting from the clustering of the data.

This seminar will describe and illustrate the use of multilevel models, focusing on an analysis of a dataset where students are clustered within classrooms and schools. The results from the multilevel model will be compared to both individual-level analysis, which ignores the clustering of the data, and classroom-level analysis that aggregates the individual data.

Hedeker is the co-author (with Robert Gibbons) of Longitudinal Data Analysis (Wiley, 2006). He was named a fellow of the American Statistical Association in 2000 and became an elected member of the International Statistical Institute in 2015. Also in 2015, Hedeker received a Long-Term Excellence Award from the Health Policy Section of the American Statistical Association. In 2016, he was elected to the Society of Multivariate Experimental Psychology. He has served as an Associate Editor for Statistics in Medicine and Journal of Statistical Software since 2006.

All PRCHN seminars are free and open to the public.