Understanding agroforestry tree preferences in Kenya using a Discrete Choice Experiment

Land degradation and increasing climate variability pose significant challenges to smallholder agricultural productivity and rural livelihoods in Kenya. Agroforestry is increasingly promoted as a strategy for addressing these challenges, but the effectiveness and sustained adoption of agroforestry interventions depend on how well proposed tree species and management options align with farmers’ ecological, productive, and livelihood priorities. This dataset was collected to examine smallholder farmers’ preferences for agroforestry attributes and to investigate heterogeneity in these preferences among forest-dependent communities in the Mount Kenya forest ecosystem. The dataset contains survey data from 400 households located in communities adjacent to forests in Meru and Laikipia Counties, Kenya. The household survey collected information on demographic and socioeconomic characteristics, agricultural production and landholdings, household food security, forest-resource use, participation in participatory forest management and community forest organizations, tree planting and agroforestry practices, and perceptions of tree and agroforestry attributes. Farmers also evaluated the importance of agroforestry characteristics using five-point Likert scales. A discrete choice experiment was incorporated into the household survey to elicit preferences for alternative agroforestry systems. The experiment included five attributes: inclusion of fruit trees, effects on crop productivity, value of marketable tree products, adaptability of tree species to local ecological conditions, and seedling price. The experimental design comprised 16 choice sets divided into two blocks, with each household completing eight choice tasks. The dataset therefore contains the household-level characteristics, preference measures, experimental responses, and associated variables needed to reproduce analyses of farmers’ stated agroforestry preferences. The data support analysis of preference heterogeneity across households, including comparisons by gender and participation in participatory forest management. In the associated study, mixed logit models show that productivity and adaptability to local ecological conditions are among the most important determinants of agroforestry preferences, while preferences for fruit-tree inclusion vary substantially across respondents and demographic groups. The dataset provides opportunities for further research on agricultural technology choice, agroforestry adoption, restoration preferences, gender differences in land-use decisions, participatory forest management, and the design of context-specific restoration interventions in smallholder farming systems.