This article discusses the methodological challenges of a service delivery survey in Khayelitsha, a township in Cape Town, South Africa. The survey aimed to gain a better understanding of the relationship between citizen participation and the environmental challenges facing residents in this urban area. Khayelitsha is a township prone to violence. Encountering violence during fieldwork can alter the way researchers execute research, yet this often remains unacknowledged in ‘objective social science’, especially in probability sampling. The article examines the effects of the risk felt by researchers on the research method employed in quantitative surveys. It shows that deliberating on this aspect of the research process is both necessary and useful, especially in terms of recognising the need to factor fear or uncertainty into the ways in which research processes unfold.