ABSTRACT
This chapter explores how the specific need to produce measurable and standardised evaluative indicators and outcomes in order to demonstrate accountability and document efficacy inhibits the deployment of inductive, context-based modifications to intervention design. One of the best-known and longest-running targeted interventions in India is an HIV prevention project with sex workers in Kolkata, West Bengal. The Sonagachi project has been widely viewed as a model for effective community-led structural intervention. The chapter focuses on two key issues: the trans ferability of different evidentiary forms; and the respective roles and value of experientially and statistically based varieties of evidence. It explores the potential value of what can loosely be seen as a focus on plausibility, rather than reliability, in analysing why interventions do or do not work' and discerning their potential trans ferability to new populations and settings. Community engagement in HIV prevention offers a useful exemplar through which to illustrate the tensions between measurement and meaning in evaluation work.
