Does combining traditional and information and communications technology–based extension methods improve agricultural outcomes? Evidence from field experiments in Mali

Abstract Adequate flow of appropriate information to farmers is vital for accelerating the uptake of modern agricultural technologies and improving access to markets, all of which are important for the transformation of African economies. Yet there is limited evidence regarding how information should be disseminated to farmers to achieve the needed impact. Should ICT‐based approaches be used together with traditional methods, or should they be used alone? We used cluster quasi‐random and randomized experiments involving 140 villages in southern Mali to study three chains of causality: does farmer training increase knowledge? Does knowledge lead to change in behavior? Do the different actions generate varying economic outcomes? We show that the role of farmer training in the uptake of extension advice may be overstated in some settings; that procrastination and inattention are more binding constraints to recommendation adherence than information per se; and that such constraints could be mitigated by timely voice SMS reminders. Although timely reminders nudged farmers to take‐up recommendations, leading to meaningful improvements in some intermediate outcomes, these impacts were not large enough to generate improvements in output, prices and profits in the short term. These findings must be interpreted within the context of imperfect compliance with experimental protocols, which pose threats to both the internal and external validity of some of the impact estimates.