Using the Delphi method for determining criteria in agroforestry research planning in developing countries

Abstract

Planning agroforestry research programs requires the participation of people from many disciplines. They may have different interests in allocating financial resources for research purposes and, therefore, the decision-making is complex. Guidelines are needed to improve the decision-making process. Such guidelines may include a list of weighted criteria identified by experts through group processes. This study used a qualitative approach (the Delphi method) to identify criteria to be used when planning agroforestry research programs in the developing countries. The panel of respondents was comprised of agroforestry experts who have experience and expertise in the developing countries. Fourteen (14) criteria were identified in questionnaire 1. These criteria were ranked in questionnaires 2 and 3. A Friedman test of analysis of variance followed by a Mann-Whitney U test of pairwise comparison produced four groups of criteria: group 1, the most highly ranked, comprises local people's needs, sustainability, adoptability and research quality; group 2 includes existing systems and economic criteria; group 3 encompasses biophysical effects, institutional capabilities, partnership and transferability; and group 4, the least ranked, includes diversity of products, tree/crop interface, flexibility and species selection. A Kruskal-Wallis test on each criterion and several respondent characteristics showed that the respondent characteristics of current work, highest degree held, ecosystem experience, geographic area of experience, and number of years experience influenced respondent criteria rankings. The test was not significant for background and respondent origin characteristics.