Presenter:
|
Mike Daniels |
Abstract |
|
This course provides a survey of modern model-based approaches to handling dropout in longitudinal studies, and illustrates the use of newly-developed methods for sensitivity analysis and incorporation of prior information. The emphasis is on Bayesian approaches but the models and methods discussed can be implemented in non-Bayesian settings as well. The course will be roughly divided into two parts:
- Part 1 will focus on formal classification of dropout and missing data mechanisms, describe classes of models that can be used to adjust for biases caused by dropout, and the logistics of model fitting;
- Part 2 will deal with specification and fitting of models to handle non-ignorable (informative) dropout, with emphasis on the role of sensitivity analysis and informative prior distributions for encoding key assumptions.
Integrated into the course will be three case studies that illustrate many of the concepts introduced during the course. We will build on each case study to illustrate progressively more complex analyses (e.g. progressing from analysis under MAR, to analysis under MNAR, to use of informative priors and sensitivity analyses). Code for model fitting using WinBUGS software (and the R-to-WinBUGS interface) can be found on the instructor's webpage.
|
| |
Who Would Benefit? |
| Attendees should have working knowledge of generalized linear models and statistical inference at the master's level and some familiarity with the basics of Bayesian inference. |
| |
|
Biography |
|
Mike Daniels is the Chair and Professor in the Department of Statistics at the University of Florida. Mike has published extensively in the statistical literature on methods for (incomplete) longitudinal data with articles appearing in Biometrika, Biostatistics, Statistics in Medicine, and Biometrics. He recently completed a book, coauthored with co-presenter Joe Hogan, titled Missing Data in Longitudinal Studies: Strategies for Bayesian Modeling and Sensitivity Analysis published by Chapman & Hall/CRC. Since 2001 he is the Principal Investigator on an R01 statistical methods grant on analysis of longitudinal data and has been a co-Principal Investigator on several other methodology grants for incomplete longitudinal data.
He has taught a graduate-level course on (incomplete) longitudinal data at the University of Florida several times and has given several short courses on missing data and dropout at national conferences and government agencies. |
| |
|
2010 Dates |
|
| April 5-7, 12-13 |
| Or weeks of
|
|
