Machine-Readable Lives or ‘Troubled Families’? Classification, categorisation and stereotyping in data collection and sharing in children’s social care in England

Critical data studies examines how the collection and use of data interact with systems of power: they shape who can know what about the world, and to what uses this knowledge can be put. This thesis uses feminist and queer approaches to consider the human rights impact of the collection and sharing of data in children’s services in England. I draw on a wide range of critical literature to situate and explore a case study: the ‘Troubled/Supporting Families Programme’ which plays a key role in children’s social care, and in the wider project of public sector datafication in England. In this thesis, I use the concepts of classification - segmentation of the world - and categorisation - naming the segments - to surface how the collection and sharing of data is not neutral but the outcome of human decision-making. I situate the collection and sharing of data within the history of information-gathering and decision-making in children’s services and with the political choices which have shaped service delivery and datafication. Classification and categorisation are used to define the ‘family’ as a unit of analysis, which enables the identification of the ‘problem family,’ and further its definition as implicitly outside of the norm. Through examining the ways in which data systems classify, categorise and stereotype individuals who are known to social services based on their gender, I show how the expectation that individual and family lives are legible to computers is used to normalise certain forms of families, and stereotype those who do not comply as ‘troubled.’