Meta-analysis of the impacts of digital information interventions on agricultural development
Robert Beach, Caleb Milliken, Kirsten Franzen, Daniel Lapidus
Abstract
Digital means of delivering advisory and other information services to small-scale agricultural producers are widely expected to expand the reach of extension and other information provision services while reducing costs. However, there is currently little evidence systematically quantifying these benefits. In this paper, we synthesize quantitative information on the impacts of digital information interventions, which may integrate human support, on adoption of modern farm inputs (fertilizer and improved seeds), yield, and income, covering interventions implemented between 2005 and 2019. Our starting point for the meta-analysis was the Agriculture in the Digital Age Evidence Gap Map dataset, which provided a consistent set of studies of digital interventions, though one limitation is that it excludes more recent studies published after 2021. After applying our criteria for relevance, rigor, and availability of necessary data, we used estimates from 20 studies in our analyses where fifteen were from Sub-Saharan Africa, four from India, and one from Cambodia. Mean impacts of digital information interventions on adoption of fertilizer (+23 %; 95 % CI +6 % to +40 %), yield (+6 %; +2 % to +9 %), and income (+6 %; +2 % to +9 %) are positive while effects on adoption of improved seeds were not statistically significant. Although our analysis indicates that digital farmer services have provided benefits on average, there is considerable variability in estimates across individual studies. However, there are not enough comparable quantitative observations from the literature included within our study population to reliably further disaggregate estimated impacts (e.g., by intervention, level of human assistance, geography, modality of information delivery, farmer type). Expansion of the available evidence base to facilitate quantification of these heterogeneous impacts is needed to better inform program design to maximize effectiveness.