Rapidly evolving data-centered research themes in many substantive areas still remain vulnerable to missing data. If no sensible action is taken, statistical analyses can be substantially misleading. In addition, techniques specifically tailored to address the unique data structures such as correlated observational units (e.g. longitudinal studies) have been the subject of statistical methods development to obtain objective inferences.
The overall research theme of the Incomplete Data Methods Lab is centered around missing data issues in cases with complex data structures as well as pragmatic missing data solutions that can address a number of practical and computational issues. The underlying methodological research is motivated by substantive issues in public health and health sciences. Building on the existing missing data computation and theory, our research has sought to address these problems. Specifically, our research areas are:
- incomplete data,
- software development,
- measurement error problems and causal inference from a missing-data perspective, and
- novel applications of these methods in public health, health sciences, and the social sciences.
