Improving Causal Claims in Observational Research: An Investigation of Propensity Score Methods in Applied Educational Research
Date of Award
Doctor of Philosophy (PhD)
This study used existing institutional data from a large, urban, public, very high research university to compare sixteen matching schemes, built from three separate datasets, to estimate the propensity score, achieve balance between groups and test the sensitivity of the average treatment effect (ATE). For each PS model, four different conditioning strategies were applied. The first four matching schemes used commonly collected data available within a student information system (referred to as SIS dataset). The next four matching schemes combined the SIS dataset with data from an entering student survey (referred to as ESS dataset). The next four matching schemes, again, combined the SIS dataset with data gathered from a noncognitive survey (referred to as NCS dataset). The final four matching schemes included data from the SIS, ESS and the NCS datasets. Each model builds upon the next, offering additional covariates for the model building process.
Wren, Julie Diane, "Improving Causal Claims in Observational Research: An Investigation of Propensity Score Methods in Applied Educational Research" (2016). Dissertations. 2603.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.
Copyright © 2016 Julie Diane Wren