Panel:
Improving Teacher Screening and Hiring in U.S. Public Schools
(Education)
*Names in bold indicate Presenter
The first two papers use data from large, urban districts to study the screening and hiring processes as they are currently implemented in these districts. The first paper leverages data from a new screening system recently adopted by the Los Angeles Unified School District (LAUSD) that allow for analyses both of applicants’ relative employment prospects and of the predictive validity of screening instruments. The preliminary results suggest that measures of prospective teachers’ academic ability and subject area preparation are not predictive of their probability of being hired despite being predictive of future teacher effectiveness, which suggests that meaningful improvements could be made to the screening and hiring process in LAUSD.
The second paper uses three years of applicant and hiring data from a different urban school district that considers detailed background information on applicants, a commercial screening tool, and a video interview in its selection and hiring process. This analysis will investigate the impact of commercial screener scores on principal decision making and hiring outcomes, and will further leverage information on both rejections and acceptances of official interview and job offers to disentangle teacher and principal preferences in the screening and hiring process.
The final two papers consider interventions that have the potential to improve the teacher screening and hiring process. The third paper presents early findings from a randomized control trial designed to assess whether an enhancement to the process through which applicant data are collected from professional references (PRs) can improve teacher selection in Spokane Public Schools in Washington. Specifically, this preliminary analysis focuses on the relationships between PRs’ ratings of applicants and the future outcomes of applicants hired into the district. Moving forward, PR ratings will be provided to hiring officials for a random subset of applicants, which will permit causal estimates of the influence of these ratings on hiring decisions.
The final paper applies machine learning techniques to job application data to develop measures that could predict future performance and turnover. The paper tests these measures on a longitudinal sample of applicants to public school teaching positions and uses them to predict subsequent work outcomes including student evaluations, expert observations of performance, value-added to student test scores, and teacher turnover. The analysis finds that work experience relevance and a history of approaching better jobs are predictive of positive work outcomes, whereas a history of avoiding bad jobs is predictive of negative outcomes.