Skip to content. | Skip to navigation

Personal tools

Navigation

You are here: Home / Research / Selected Research / Using propensity scores for causal inference with covariate measurement error

Using propensity scores for causal inference with covariate measurement error

Faculty Associate Frauke Kreuter's project, an R01 funded by the National Institute of Mental Health, seeks to develop and assess new statistical methods

Dr. Kreuter's project will develop and assess new statistical methods for estimating causal effects in the presence of measurement error and differential measurement error in the covariates across treatment groups. By enabling better estimation of causal effects in settings where randomized trials are infeasible or unethical this will help address specific NIMH strategic objectives. It is also connected to recent initiatives regarding data and resource sharing, by developing methods that will help researchers use propensity score methods when combining data sources. The work is motivated by three studies in mental health evaluating the effectiveness of (1) a perinatal depression prevention program, (2) depression screening and treatment in primary care, and (3) early intervention for children with autism. In each study the team is comparing the group receiving intervention to a comparison group from an external data source. They are using simulation studies to assess the performance of the methods and, in two of the three motivating examples they will have a randomized trial of the intervention to which they can compare their results to determine how well the method is working.

See Dr. Kreuter's bio.

Navigation