CALDER Research Method: Hazard Models

One of the advantages of longitudinal data is an ability to follow an individual over time as they change locations.

Example

Consider teacher labor markets as teachers move in and out of teaching and across schools and districts. A hazard rate model can estimate the probability that a teacher exits in a specific year as a function of how long he or she has been teaching.

Research

CALDER researchers have used hazard models to estimate the probability of a teacher exiting teaching, leaving a given district or leaving a given school (Reed, Rueben and Barbour (forthcoming); Boyd, Lankford, Loeb and Wyckoff, 2005; Podgursky, Monroe and Watson, 2000).

We also intend to use such transition models to examine where teachers go once they leave their first job. We have found that teacher transitions vary by level of prior experience. Hanushek, Kain, O’Brien and Rivkin (2005) found that transitions to different types of schools and across districts varied with years of teacher experience.