Using machine learning to translate applicant work history into predictors of performance and turnover

Abstract:

Work history information reflected in resumes and job application forms is commonly used to screen job applicants; however, there is little consensus as to how to systematically translate information about one’s work-related past into predictors of future work outcomes. In this article, we apply machine learning techniques to job application form data (including previous job descriptions and stated reasons for changing jobs) to develop interpretable measures of work experience relevance, tenure history, and history of involuntary turnover, history of avoiding bad jobs, and history of approaching better jobs. We empirically examine our model on a longitudinal sample of 16,071 applicants for public school teaching positions, and predict subsequent work outcomes including student evaluations, expert observations of performance, value-added to student test scores, voluntary turnover, and involuntary turnover. We found that work experience relevance and a history of approaching better jobs were linked to positive work outcomes, whereas a history of avoiding bad jobs was associated with negative outcomes. We also quantify the extent to which our model can improve the quality of selection process above the conventional methods of assessing work history, while lowering the risk of adverse impact. (PsycInfo Database Record (c) 2020 APA, all rights reserved)

In this paper, we drew on the strategic HR literature and applied machine learning techniques to job application data (including previous job descriptions and stated reasons for changing jobs) to develop innovative, valid, and fair predictors of work outcomes. These predictors included work experience relevance, tenure history, a history of involuntary turnover, a history of avoiding bad jobs, and a history of approaching better jobs. Using these predictors,we then applied econometrics methods to develop a predictive model of employee performance and risk of voluntary and involuntary turnover that circumvents the typical limitations of selection programs. We then compared the performance of our employee selection model with the selection model implemented in the organization in terms of performance, turnover, and adverse impact on minorities and showed that our model significantly outperformed the traditional selection model implemented in the organization in all three areas