Over the summer of 2010 I was involved in a collaborative project to survey a minority Indian community. The community had not been surveyed or enumerated in the national census as a distinct group for over 60 years. The need for up-to-date information in order to plan long-term assistance was recognised by one high profile and influential leader within the community. His lack of research background, however, meant that he had not proceeded with a project to gather this information. As a social scientist working with the group for over ten years I also saw the merits of the project so encouraged him to consider running a survey. A year before we did so we met and agreed to run it together: he had the contacts, was a high profile politician with a national profile, along with all the trappings of power that go with an influential male in an Asian country. For my part I could offer foreign university connections, and expertise. Having worked with the community for the previous ten years I also had sound and trusting research relationships with key community members. There was no reason to feel that we would be anything but an efficient workable team on a unique and valuable project. For a number of reasons the survey did not go as well as we had hoped. In this paper I reflect on one of the key issues to emerge from the experience – that of the effect of uneven power relationships. There is a body of literature referring to the need for social scientists, when working collaboratively with non-academic locals, to be aware of the relationships of power and to ensure that all involved have the opportunity to have some control of the process. The assumption is always that the balance of power lies with the social scientist who must take responsibility to involve other collaborating parties fully. The reality of my experience, however, was anything but that.