ABSTRACT
I analyse how the inclusion of working condition data influences the estimated marginal effects of student demographics on teacher mobility. Using detailed administrative data on public schools, combined with unique data describing teachers’ perceptions of working conditions, I estimate a model characterizing the multinomial outcomes of teachers choosing to work at different schools. By comparing the estimated marginal effects of student characteristics with and without incorporating detailed working conditions, I find that excluding these additional data results in overestimating the effect of student characteristics on the probability a teacher moves to a different school, but has no influence on estimated effects for the probability of leaving teaching altogether.
Acknowledgement
I thank the North Carolina Education Research Data Center for access to the data. Comments are welcome at [email protected].
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 I limit the sample to full-time teachers who are matched to a single school within a given academic year.
2 Estimated coefficients from the models excluding and including the working condition survey responses are in and , located in the appendix.
3 The likelihood ratio test rejects the null hypothesis that the coefficients are jointly equal to zero at the 1% level ().