Decision-Support Systems at the Frontline: Decision-Support Systems Influence on Frontline Workers’ Assessments and Decision-Making

The public sector is currently experiencing a significant rise in the development and adoption of decision-support systems (DSS). These data-driven systems generate predictive input or recommendations and are intended to inform frontline workers’ decision-making process. Consequently, DSS have the potential to affect a core dimension of the public service delivery by influencing frontline workers’ professional judgements and behavior toward citizens. However, despite the rapid growth in DSS in public organizations, we lack theoretical knowledge and empirical evidence about how information from these systems actually influences frontline workers’ professional decision-making. In this study, we address this gap in the literature by theoretically developing hypotheses about the effects of DSS and empirically testing them. Drawing on psychological research and the street-level bureaucracy literature, we hypothesize that DSS influence frontline workers’ professional assessments, certainty in assessment, and behavior toward clients. We test these preregistered hypotheses in a large-scale survey experiment on a unique sample of social caseworkers (n=772) across 39 Danish unemployment agencies. Our experimental results show that input from DSS directly shape frontline workers’ assessment of clients and their behavior toward them. Importantly, our effects are moderated by the type of clients (complex/simple) that the DSS delivers predictions about. Surprisingly, we also find that information from DSS in most cases decreases caseworkers’ certainty in their own assessments of clients, even when the information from DSS confirms caseworkers’ professional judgement.