Drivers of diversified legume adoption among smallholder farmers in Singida, Tanzania: A Heckman two-stage model approach

Abstract Background : Diversified legume production enhances smallholder farmers’ food security, income, and soil fertility in semi-arid regions of Tanzania. Understanding the factors that drive both adoption and the extent of adoption is critical for promoting sustainable agricultural practices. Methods : This study examines the socio-economic, institutional, and motivational drivers of diversified legume adoption in the Singida Region using data from 204 smallholder farmers. A Heckman two-stage model was applied to correct for potential sample selection bias. Results : Key drivers of adoption and intensity include farm size allocated to legumes (β = 0.203, p < 0.01), educational attainment (β = 0.154, p < 0.05), occupational status (β = 0.388, p < 0.01), access to extension services (β = 0.318, p < 0.05), perceived economic benefits (β = 0.157, p < 0.01), motivation from Farmer Research Networks (β = 0.095, p < 0.05), market expectations (β = 1.031, p < 0.001) and learning from peers (β = 0.171, p < 0.01). Older farmers exhibited lower adoption intensity, and risk-averse farmers were less likely to diversify. The model was robust, with Chi-square = 258.06 ( p < 0.001) and a significant inverse Mills ratio. Conclusion : The findings demonstrate that land endowment, human capital, institutional support, market expectations, and behavioural factors jointly determine adoption and diversification levels. Policies targeting information access, credit provision, and risk mitigation can enhance both participation and the extent of diversified legume production.