HIV continues to spread unabated in many developing countries. Here we consider the interventions that are currently in place and critically discuss the methods that are being used to evaluate them as reported in the published literature. In recent years there has been a move away from highly individual-oriented interventions towards more participatory approaches that emphasise techniques such as community-led peer education and group discussions. However, this move towards more community orientated intervention techniques has not been matched by the development of evaluation methods with which to capture and explain the community and social changes which are often necessary preconditions for health-enhancing behaviour change. Evaluation research continues to rely on quantitative methodologies that fail to elucidate the complex changes that the newer interventions seek to promote within target communities. In addition, these methods of evaluation tend to rely on the use of highly individualistic and quantitative biomedical indicators such as HIV/STD rates, or knowledge, attitude, perception and behaviour (KAPB) survey questionnaires. We argue that such approaches are inadequate for the task of tracking and measuring important determinants of programme success such as psycho-social changes, features of the community-intervention interface and the degree of trust and identification with which members of target communities regard particular interventions. Rigorously conducted qualitative process evaluations taking account of the above factors could make a key contribution to the development of more successful HIV-prevention interventions.