Stated preference studies are often used to provide estimates of non-market values for cost-benefit analysis. Applications of these techniques have been criticised because of the hypothetical nature of the surveys and the possibility of bias. Various types of bias have been cited including strategic, payment vehicle, starting point, interviewer and others. However, one type of bias that has received limited attention, but has the potential to have substantial impacts on value estimates, is that associated with aggregation. In this paper the theoretical basis for aggregation is explored, and alternative approaches for aggregation are trialed. It is demonstrated that the choices made by researchers at this stage of the value estimation process can significantly affect the estimates that are to be used in cost-benefit analyses. Copyright 2000 by Blackwell Publishers Ltd/University of Adelaide and Flinders University of South Australia
