The general theme of this lab is the development of innovative solutions to novel methodological and applied problems in statistics. In so doing, we seek to push the frontier of statistics in practice. Current domains of focus and example projects include:
- Social network analysis: The development and application of social networks to gain insights into the diffusion of medical technology and practices; The development of models for relational data in longitudinal and hierarchical contexts.
- Multivariate-multilevel models: The development of methods to estimate the covariance matrix of a multivariate outcome at a group level (e.g., health plan or hospital) when observations are obtained from sub-units (e.g., patients).
- Causal inference: The development of methods to distinguish social influence (peer effects) from social selection (homophily); The development of methods for causal inference in randomized trials with departures from the study protocol and in observational studies subjected to unmeasured confounding when the outcome is a possibly censored time-to-event.
- Proximity to food establishments and neighborhood effects: The application of a novel model to estimate the effect of proximity to various type of food establishments on health outcomes accounting for neighborhood effects and other forms of heterogeneity and clustering of individuals.
The Statistical Problem Solving Laboratory is directed by James O’Malley.