Project Title and Abstract

Using Multi-site Multiple Mediator Instrumental Variables to Explore Compliance and Mediator Effects in the NSLM

Multisite trials are designed to estimate the average as well as heterogeneous treatment effects of interventions across sites. However, in many cases, compliance with treatment assignment is imperfect. In this proposed study, I will use instrumental variable (IV) approaches to (1) analyze “compliance” to treatment and (2) explore the various mediating pathways through which the intervention affects student achievement. Specifically, I will develop a conceptual framework using the multisite multiple-mediator instrumental variables models to identify and analyze the various hypothesized mediators of the growth mindset intervention, thereby illuminating how the intervention affects student achievement. Broadly, the goals of this project are to contribute to new understanding regarding the consequences of differential compliance to the treatment in social psychological interventions, and to analyze the effects of differential mediating processes across multiple sites using IV approaches.