There have been few applications of the contingent valuation method (CVM) to forests in developing countries. When applied, the method is seldom utilized to improve the implementation of development projects. In this paper an application of CVM to the Orissa Social Forestry Project is reported. The results show considerable variation between villages and the analysis of the bid function shows underlying factors to project success that can be used in project planning. The policy implications are to plant new VWL where natural forests are inaccessible, where people have prior experience of VWL, and where fuel consumption relies heavily on market purchases. Such plantations are expected to give greatest benefit to families with few men and many women. An analysis of the reliability of the study concludes that many of the observed biases originate from the enumerators. The study also recommends that one should be careful with the use of a discrete design in a cultural context such as this that is prone to yea-saying, and that strategic bias can arise from an open ended question that follows up the start bid in a culture characterized by bargaining.