ABSTRACT: Mobile-based informational programs are common across the world, though there is no consensus on how effective they are at affecting behavior. We present causal evidence on the effects of six different mobile-based agricultural information programs implemented in Kenya and Rwanda. All programs shared similar objectives but were implemented by three different organizations and varied in terms of content, design, and target population. With administrative outcome data for over 156,000 people across all experiments, we are sufficiently powered to detect small effects in real input purchase choices. Combining results through a meta-analysis, we find that the odds ratio for following the recommendations is 1.26 (95% CI: 1.19, 1.32). We cannot reject similar magnitudes of effects across all experiments and recommended technologies. We do not find evidence of message fatigue nor of crowd-out of use of other inputs. Providing more granular information, supplementing the texts with an in-person call, and using different behavioral framings did not significantly increase impacts relative to a simple message, but message repetition had a modest positive effect. While the overall effect sizes are small, the low cost of text messages can make these programs cost-effective.
