A Latent Class Analysis of Agricultural Technology Use Behavior in Uganda and Implications for Optimal Targeting

Agricultural productivity is still very low in Africa largely due to low use of improved agricultural technologies. Existing adoption studies are marred by univariate analyses, often focusing on single technologies over a limited scope while assuming uniform effects of the explanatory variables across farm households. In this study, we use a large dataset that covers a wide geographical and agricultural scope to describe use-patterns of improved agro-technology in Uganda. Using latent class analysis, and over 12,500 households collected across the four regions of Uganda, we classify farmers based on the package of improved agro-technologies used. We find that the majority of farmers (61%) do not use any improved agricultural practices ( non-users ) while only 5% of the farmers belong to the class of intensified diversifiers , using most of the commonly available agro-technologies across crop and livestock enterprises. Using multinomial regression analysis, we show that education of the household head, access to extension messages and affiliation to social groups, are the key factors that drive switching from the non-user reference class to the other three preferred classes that use improved agro-technologies to varying levels. Results reveal that different farmer categories with different agro-technology needs, which may have implications for optimal targeting. Acknowledgement : The authors would like to thank the International Institute for Impact Evaluation (3ie) for financial support. This work was undertaken as part of the CGIAR Research Program on Policies, Institutions, and Markets (PIM) led by the International Food Policy Research Institute (IFPRI).