STATISTICS COLLOQUIUM
Fernanda Lang Schumacher
Campinas State University, Brazil
Robust mixed-effects models for longitudinal data
Abstract
In clinical trials, studies often present longitudinal or clustered data and are frequently affected by missing data. These studies are commonly analyzed using linear mixed models (LMMs), and for mathematical convenience, it is usually assumed that both random effect and error term follow normal distributions. However, these restrictive assumptions may result in a lack of robustness against departures from the normal distribution and invalid statistical inferences. In this talk, a flexible extension of LMMs considering the scale mixture of skew-normal class of distributions will be presented, accommodating skewness and heavy-tails and accounting for a possible within-subject serial dependence. The model estimation and evaluation using the R package skewlmm will be illustrated, and two applications to longitudinal data sets, regarding schizophrenia and mouse diet clinical trials, will be discussed. Additionally, some recent and future extensions will be introduced.
Bio: Fernanda Schumacher recently completed a Ph.D. in Statistics at the University of Campinas, Brazil, where she also obtained a master’s degree in Statistics in 2016. Her research interests include robust models, models for censored data, longitudinal data, EM algorithm, and scale mixture of skew-normal distributions.
For more information, contact: Tracy Burke at tracy.burke@uconn.edu