Child protection threshold talk and ambivalent case formulations in ‘borderline’ care proceedings cases

Abstract

This paper focuses on cases about children who were the subject of child protection plans and designated by children's social care services as 'borderline' for compulsory intervention by way of care proceedings. It moves beyond abstract language, into the everyday vocabularies of practice, with the aim of better understanding decision-making in such cases. The majority of these cases had been categorised as neglect (34/47 children). While social workers and managers clearly invoked a threshold or line for compulsory action, their discussions demonstrated a range of factors about why it was not always easy to identify when this line had been crossed.

Towards a critical decision-making ecology approach for child protection research

“…Some explain how judgements are formed through a combination of individual and social sense-making based in intuitive, social and analytical processes, linking these to organisational elements and external factors such as resource availability (Nyathi already links this to DME concepts) (Cook, 2017;Helm, 2011;Nyathi, 2018). Others show that thresholds are more like a 'fuzzy juncture' than a firm line based on individual factors, instead they are influenced by site specific social meaning-making processes, for example managers' perceptions (rather than front-line workers') of what types of cases should meet a threshold for action (Doherty, 2016;Keddell and Hyslop, 2020;Platt and Turney, 2014). Still others show how different contexts (such as states) implement decisionmaking tools (in this case the SDM) with greater or lesser levels of strictness, creating differences in the relative influence of actuarial tools and practitioner discretion (Bosk and Feely, 2020).…”

Algorithmic Justice in Child Protection: Statistical Fairness, Social Justice and the Implications for Practice

“…As is described above, attention to precisely what feedback the model is referring to in its selected outcome is important and challenges the ability of the model to learn accurately. It not only does not 'see' some abuse, the lack of consensus of the definition of abuse creates further distorted data, as decisions to investigate or substantiate abuse can be highly variable (Doherty 2016;Munro 2002). In the classic Dawes et al (1989) study of predictive abilities, the three outcomes they used to compare human and actuarial predictions were: major mental illness, the detection of brain damage, and the prediction of survival times relating to illness.…”

The Definitions Are Legion: Academic Views and Practice Perspectives on Violence Against Children

“…While some severe acts of commission and omission against children are unambiguously violent, the distinction between still acceptable educational behavior and mildly violent acts that trespass the threshold into child maltreatment is often difficult to establish and not clear-cut -for both academia and practice (Doherty, 2017;Platt & Turney, 2014). Epidemiological definitions therefore differ on the thresholds of severity, if a single event (of belittling a child, or mildly slapping its fingers) should be counted, or if certain types of behavior should be included or excluded (e.g., verbal sexual aggression).…”