Keynote Speakers

Bengt Muthén:

Keynote: Advances in Mixture Modeling

After a brief overview of the many uses of finite mixture modeling, applications of new mixture modeling developments are discussed.  One major development goes beyond the conventional mixture of normal distributions to allow mixtures with flexible non-normal distributions.  This has interesting applications to cluster analysis, factor analysis, SEM, and growth modeling.  The talk focuses on applications of Growth Mixture Modeling for continuous outcomes that are skewed.  Examples are drawn from national longitudinal surveys of BMI as well as twin studies.  Extensions of this modeling to the joint study of survival and non-ignorable dropout are also discussed.

 

Sophia Rabe-Hesketh

Keynote: Simple methods for handling non-randomly missing data– Sophia Rabe-Hesketh and Anders Skrondal

In multiple linear or logistic regression, multiple imputation has become increasingly popular for handling missing covariate values. The much simpler approach of listwise deletion or complete-caseanalysis is often dismissed as making overly strong assumptions. However, I will point out that complete-case analysis is consistent and performs better than multiple imputation for many types of non-random missingness mechanisms. In longitudinal data analysis, dropout or intermittently missing responses are typically dealt with by specifying a joint model for the responses, such as a growth-curve/hierarchical/multilevel model, and estimating the parameters by maximum likelihood. This approach is consistent if missingness of a response depends on observed responses for the same individual but not if if it depends on the response itself or on the random effects in the model. One way of handling such non-random missingness is to model missingness jointly with the response variable of interest, but these joint models are complex, require specialized software, and make unverifiable assumptions. I will suggest simple fixed-effects approaches that are consistent if missingness depends on the random effects and, in the case of binary responses, if missingness depends on the response itself or previous (observed or unobserved) responses.

Jamie Robins

The principal focus of Dr. Robins’ research has been the development of analytic methods appropriate for drawing causal inferences from complex observational and randomized studies with time-varying exposures or treatments.

Edward Vytlacil

Edward Vytlacil received his PhD in Economics from the University of Chicago in 2000. He is currently a Professor of Economics at New York University, having previously been a faculty member at Stanford University, Columbia University, and Yale University. He is a Co-Editor of the Journal of Applied Econometrics, and an Associate Editor for Econometrica and the Journal of Econometrics.

Keynote: Accounting for Individual Heterogeneity in Treatment Effect Analysis