Playing it safe? Using sandbox experiments to explore predictive risk modelling

This article examines how individual and collective sensemaking in child and family welfare services can be fostered through a sandbox experiment exploring predictive risk modelling (PRM). Using various qualitative research methods as artefacts, we designed a sandbox that simulated social workers’ sensemaking processes of past, present and future decision making. This approach provided valuable insights for us and the participants, particularly into the complexities and entanglements of PRM in social work and the situatedness of decision making in child and family welfare. The sandbox experiment was conducted in Denmark, where projects aiming to study the use of PRM in child and family welfare services have been paused or significantly modified due to ethical and legal concerns. We argue that research involving child welfare, safeguarding and the full implementation of predictive risk models entails high-stakes decisions for social workers and organisations, which come with significant implications for the collaborating parties in practice due to ethical, legal and cost-related uncertainties. We propose sandbox experiments to tackle these challenges, using new technology to transform high-stakes decision making into low-stakes, simulated decision making and sensemaking. This approach makes PRM more accessible for study while ensuring adherence to ethical and practical considerations.