Debate currently rages over the development potential of Non-Timber Forest Products (NTFP) in tropical forests. Proponents of particular "solutions" can refer to evidence (case studies, data) which tend to support their interpretation of events and relationships. Recommendations thus frequently depend on how data are classified and interpreted. Inaccurate or incomplete classification leads to defective subsequent theories, models, and recommendations. We present a method for classifying very divergent case-study data, and some initial results as a basis for general understanding of key factors that influence a given result. Crucial issues determining outcomes of NTFP development include the nature of government involvement, distribution of property rights, the ability of local people to claim and enforce such rights, market transparency, and pressure on the resource. This paper concludes with suggestions for further testing and development of the methodology.