Conference Agenda

The digitalization of welfare administration has transformed how citizens interact with the state. Across many countries, welfare systems increasingly rely on automated decision-making and digital-first interfaces to process benefits and enforce compliance requirements (Alston, 2019). The United Kingdom’s Universal Credit (UC) program represents one of the most ambitious examples of this shift, consolidating six means-tested “legacy benefits” into a single system administered primarily through an online infrastructure (Department for Work and Pensions, 2010). While proponents argue that digitalization improves efficiency and consistency while reducing administrative costs, a growing body of research highlights the potential for such systems to generate new forms of administrative burden, financial hardship, and psychological stress among claimants (Alston, 2019; Moynihan, Herd, & Harvey, 2015; Herd & Moynihan, 2018).

Existing research on Universal Credit relies on surveys, interviews, or administrative datasets, which, while valuable, offer limited visibility into how claimants collectively interpret and navigate the system over time. At the same time, large claimant-run social media groups have emerged as informal infrastructures where individuals seek advice, interpret bureaucratic rules, and share experiences of navigating UC. These online communities provide a unique opportunity to observe claimant experiences at scale and to examine how digital peer networks serve as mediators to the welfare state (Prescott et al., 2020).

Novel Methods, Novel Data

This paper’s novelty lies in analysis of claimant-generated content from Universal Credit-related Facebook groups to better understand the types of issues claimants encounter, how these issues evolve over time, and how peer communities function as informal support systems. The study examines claimant discussions using a mixed-methods approach that combines qualitative coding, computational classification, and sentiment analysis.

The analysis uses a dataset of posts collected from large UC-focused Facebook support groups. The corpus contains over 150,000 posts produced by more than 36,000 users. These groups have memberships of hundreds of thousands of claimants and function as spaces where individuals ask questions about eligibility, reporting requirements, payment calculations, sanctions, and other aspects of UC administration. The dataset includes post text, timestamps, and engagement metrics such as comments and reactions, enabling analysis of both thematic content and community interaction dynamics.

Our analysis proceeds in three stages. First, a stratified sample of posts was manually coded to develop a thematic classification framework grounded in administrative burden theory (Moynihan, Herd, & Harvey, 2015). Categories capture distinct dimensions of claimant experience, including learning costs (e.g., confusion about eligibility or calculations), compliance costs (e.g., uncertainty about reporting requirements or sanctions), and psychological costs (e.g., reports of anxiety, frustration, or uncertainty). Second, supervised machine learning models are used to classify posts across the full dataset, allowing examination of issue prevalence and temporal dynamics at scale. Third, sentiment analysis and engagement metrics are used to assess the emotional tone of discussions and the extent to which different types of issues mobilize peer support within the communities.

Findings

Findings suggest that claimant issues of administrative complexity dominate the posts, particularly uncertainty regarding eligibility rules, payment calculations, and communication with welfare administrators. Emotional analysis indicates persistent expressions of anxiety and confusion in posts related to administrative processes. At the same time, engagement patterns suggest that posts describing acute hardship or bureaucratic challenges attract the highest levels of peer response. These results highlight how informal digital communities function as intermediaries, translating complex, abstract, and often poorly explained welfare rules into everyday claimant experiences.

Policy Relevance

By analyzing claimant-run social media groups as sites of collective sense-making, this research contributes to several strands of scholarship, including studies of digital welfare states (Alston, 2019), administrative burden theory (Herd & Moynihan, 2018), and online peer support networks (Prescott et al., 2020). The policy implications of the findings demonstrate how analyzing posts in platform-mediated communities can provide policymakers with valuable insights into the lived experiences of welfare reform. These informal support infrastructures may help identify areas where policy-makers could improve policy design, communication, and digital service delivery to reduce user burden and enhance claimant support.