Discovering Hidden Subgroups

Latent class analysis (LCA) allows researchers to detect the existence of subgroups within populations. Latent transition analysis (LTA) extends LCA to changes in group membership over time.

Read about our research on

Analyzing Rich, Complex Data Sets

Our intensive longitudinal data project provides tools for analyzing data with many observations, like data collected in smartphone studies. Our high-dimensional data project provides tools for variable selection and analysis of datasets with large numbers of variables, like genetics data.

Read about our research on

Designing and Building Health Interventions

The sequential, multiple assignment, randomized trial (SMART) project provides tools for building adaptive interventions so that treatments adapt to meet patients’ needs. The multiphase optimization strategy (MOST) is a framework for engineering interventions for purposes such as maximizing intervention potency within a specified cost limit or streamlining existing interventions for scale-up.

Read about our research on

Inferring Cause Without Randomization

Social, behavioral, and health researchers often face situations where randomization is not possible, practical, or ethical. Our causal inference project develops methods for inferring causality in non-randomized experiments or studies.

Read about our research on