Learning Digitally: Evaluating the Impact of Farmer Training via Mediated Videos
Kathryn Vasilaky  1, 2, *@  , Kentaro Toyama  3@  , Tushi Baul  4@  , Dean Karlan  5@  
1 : Columbia University
2 : California Polytechnic State University [San Luis Obispo]  (CAL POLY)  -  Site web
San Luis Obispo, Californie 93407 -  États-Unis
3 : University of Michigan [Ann Arbor]  -  Site web
500 Church Street Ann Arbor, MI 48109-1090 -  États-Unis
4 : University of Notre Dame  -  Site web
Indiana -  États-Unis
5 : Yale University [New Haven]  -  Site web
157 Church Street, New Haven, CT 06510-2100 -  États-Unis
* : Auteur correspondant

Individualized training in agriculture is costly, but necessary if newer and more sustainable methods of cultivation are to be adopted. Technology enables us to disseminate information quickly and at a low cost. Of course, whether that information will be read, used, or even misunderstood is not easily tractable, nor well documetned. We examine the effectiveness of mediated video-based training on individuals' adoption rates of a new agricultural technology (alternate wetting and drying) in rural Bihar, India. In a 3-arm clustered randomized control trial (RCT) framework, villages are randomly assigned to the treatment arm where villagers view videos developed by Digital Green (DG) with the aid of their local Self Help Groups (SHGs) mediated by a selected village representative (VRP). In addition, all participants receive standard agricultural training provided by India's National Livelihood Rural Mission (NLRM). The videos are tailored to local norms, and scalable in their dissemination. Our results indicate that adoption increases above and beyond standard training as a result of tailored and mediated videos, which feature individuals similar to themselves operating in their local setting. The probability of adoption increases by 0.05 for those who viewed DG videos, where the average adoption rate in the control group is 0.10, a 50\% increase between treatment and control groups. In two sub-arms, we address farmers' uncertainty regarding the costs of adopting the new technology as well as their uncertainty regarding their personal ability to implement the technology (self-efficacy). We find that such subtler messages have a more variable and weaker effect. Messages addressing labor costs had more consistent impacts on adoption, while messages addressing self-efficacy had little to no effect.


Personnes connectées : 3