Effects of Experience by Preparation Pathway on Teacher Retention, Tennessee

Cox Proportional Hazard Regression for new teacher retention by preparation pathway
Cox proportion hazard model’s outcome of interest is a binary measure of retention (1=retained). Failure is measured when retention is 0. The observed time interval is experience. Model covariates include preparation pathway individual characteristics (sex, race, age, logged salary), and school characteristics (student race, % special education, % economic disadvantaged, % immigrant or English learner, student sex, school locale, logged enrollment, school level). Standard errors are clustered at the school level. Graph plots the survival function (retention) holding covariates constant at the mean. See TERA working paper Appendix Table 1 for full list of covariates.
Chart: SREB/Vanderbilt UniversitySource: TDOECreated with Datawrapper