Abstract
Objective: Data help inform research and practice in child protection services. One data category within quantitative datasets is derived from reports made to child protection agencies from others with child abuse concerns. Reports data is used in the development of algorithmic decision- making support tools. Data category coherence is integral to these uses, yet little is known about contributing human judgement processes. This paper outlines a study of police staff decision- making about making reports of concern.
Methods: This paper describes a qualitative study of police ‘reports of concern’, using interviews, focus groups and observations of team decision- making. The methodology, based on critical realism, adopts the decision-making ecology as a theoretical framework, and uses deductive thematic analysis to analyse the results.
Results: Police participants emphasise support responses to families before making reports of concern. Reports are likely when exposure to intimate partner violence is chronic, parents are perceived as culpable and disengaged from services, and reporters are concerned about blame for not reporting. Diverse thresholds for reporting between reporters are linked to differences in role type, accountability concerns, interpretations of reporting policy, experience, and changing policy emphases. Resistance to accepting a report by the statutory agency affected reports. This varied by location and could amplify reporting behaviour over time.
Conclusions: This complexity highlights the lack of internal consistency in ‘reports of concern’ data, and elucidates the feedback effect, challenging their use in predictive tools. Reports cannot be assumed to represent child abuse incidence, nor a consistent level of community-assessed risk of harm.