The ethics of conducting research on family violence in low resource settings: experiences of embedding social workers within a longitudinal study in South Africa

Ethical guidelines for violence research emphasise the importance of safeguarding participants through referral to appropriate services. However, little empirical evidence exists on how these ethical obligations can be operationalised in resource-constrained settings where statutory social services remain limited or overburdened. This study explored the implementation of an embedded social work model within a longitudinal study on intergenerational violence in South Africa, examining the implications for participant safeguarding, fieldworker wellbeing and research practice. This qualitative study was conducted in Mpumalanga Province between July and September 2023 as part of the Interrupt_Violence project. Data were collected through two focus group discussions with fieldwork staff, one with researchers, and in-depth interviews with eleven purposively selected participants who had received support from the embedded social workers, case reviews, as well as one social worker. Data were analysed thematically. Findings highlighted the critical role that an embedded social work model provided including active case management for 1,399 cases over 18 months, addressing intersecting psychosocial, child protection, legal and socioeconomic needs. Participants reported that sustained engagement with the social workers reduced feelings of shame and isolation, facilitated access to services and provided valued emotional support. Fieldworkers described reduced emotional burden, strengthened professional boundaries and greater confidence that participants would receive appropriate care. However, the effectiveness of the model was constrained byheavy caseloads, bureaucratic barriers and limited capacity within statutory and social services. Embedding social workers within violence research represents a practical mechanism for operationalising ethical safeguarding in resource-constrained settings. While this approach strengthened participant support, protected fieldworker wellbeing and enhanced the ethical conduct of research, it also highlighted the substantial resources required to provide meaningful safeguarding where statutory services are limited. Future violence research should consider embedded safeguarding models as a core component of ethical study design rather than relying solely on passive referral pathways.