Balancing Accuracy and Acceptance of Algorithmic Hiring Decisions: Put the Human Touch into the Equation
Citation. Matić, J. J., Neumann, M., Kerscher, F., Linn, F. H., & de Vries, R. E. (2025). Balancing accuracy and acceptance of algorithmic hiring decisions: Put the human touch into the equation. Journal of Business and Psychology, 40(6), 1425–1443. https://doi.org/10.1007/s10869-025-10026-7
Summary. This article tests ways to combine human judgment and an algorithm when predicting job performance in a hiring task. Across two studies, a structured hybrid approach called mechanical synthesis preserved some decision-maker input while producing more accurate predictions than purely intuitive judgment. The findings suggest that design choices about how human input enters a selection system matter for both decision quality and whether people will accept the system.
Key claims
- In Study 1 (N = 261), clinical synthesis and mechanical synthesis produced more accurate job-performance predictions than holistic prediction, and mechanical synthesis outperformed clinical synthesis.
- In Study 2 (N = 610), mechanical and clinical synthesis again produced much more accurate predictions than holistic prediction and were viewed more positively than a prescribed algorithm with no decision-maker input.
- Mechanical synthesis adds a decision maker’s overall judgment to the algorithm as a consistently weighted input, rather than allowing ad hoc changes to the algorithm’s final prediction.
- The authors frame the central design problem as an autonomy–validity dilemma: strict algorithms can improve prediction but reduce perceived control, while unrestricted judgment can preserve control at a cost to consistency and accuracy.
Evidence & limitations
- Two preregistered experimental studies asked adult participants to predict the performance of real airline ticket-agent applicants using an archival dataset; the studies used simulated selection decisions rather than live organizational hiring outcomes.
- The study compares defined information-combination procedures and reports prediction accuracy and participant perceptions, but it does not establish whether the preferred designs improve actual organizational adoption, legal compliance, applicant experience, or fairness across demographic groups.
- The article is a peer-reviewed Journal of Business and Psychology publication; the ingested Drive PDF matches the published journal version identified by its DOI, title page, volume, pages, and publication metadata.