Social learning, 'online' versus in-person: Experimental Evidence from Indian Farmers

How to cite

Sharma, Vanisha. 2021. "Social learning, 'online' versus in-person: Experimental Evidence from Indian Farmers." AEA RCT Registry. AEARCTR-0007851. https://doi.org/10.1257/rct.7851-6.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Vanisha Sharma*

* Primary investigator

Registry ID
AEARCTR-0007851
DOI
https://doi.org/10.1257/rct.7851-6.0
First registered
Last updated
Status
Completed
Trial dates
2021-08-01 – 2022-07-30
Countries
India (Andhra Pradesh)
JEL codes
Q1, O1, O2, O3
Registry record
http://www.socialscienceregistry.org/trials/7851

Abstract

With increasing internet accessibility in India, channels of information exchange have widened more than ever before. A by-product of this development has been an increase in the rural uptake of smartphones for various purposes, including agriculture. In this paper, I study the effects of exogenously created farmer-Whatsapp groups on agriculture outcomes, including yields, costs and response to weather and pest shocks. I also measure the effect of online interactions on farmer beliefs of willingness to exchange information with other farmers in-person versus over the phone. More specifically, by collecting data from a unique field experiment spanning 108 Indian villages, I test whether long-distant (online) connections lead to better diffusion of information and adoption of new agricultural technology by connecting farmers across space on several online groups. Due to the endogeneity of within-village networks, I randomize treatment at the village level, into two treatment interventions. In the first arm, farmers from non-neighboring villages are connected on twelve moderated Whatsapp groups (one group per 3 villages) to facilitate online information sharing of farming practices. In the second arm, along with the Whatsapp groups I also simulate in-person connections by inviting the Whatsapp group participants meet physically to discuss several farming related topics over lunch, and attend a poster-presentation. I then measure the differential effects of online farmer-interactions in these two types of groups. In addition, I track individual Whatsapp chat participation rates to test whether in-person interactions substitute or complement the online interactions. By facilitating online discussions on topics that are important to local farmers, I expect to see better agricultural outcomes for treated farmers. Preliminary analysis indicates that the mechanism through which this may happen is increasing willingness to exchange agriculture information with unknown farmers for treated groups.

Intervention

The main goal is to connect farmers on social media groups to test traditional social learning and technology adoption theories in a new dimension of virtual interactions. We will test the influence of connecting farmers from different villages onto the same platform, a moderated Whatsapp group, to enable engagement in conversations about new technology. This influence will be compared to another treatment group, where farmers, before being added to the moderated Whatsapp group, will be offered to meet in-person with their Whatsapp peers- thereby simulating an in-person connection.

Experimental design

We will create 2 treatment arms with across-village Whatsapp groups to enable farmer interaction about smartphone agriculture technologies. Sampled farmers from non-neighboring villages will be connected on a Whatsapp group in both arms. In one treatment group, other than the Whatsapp group, farmers will also be encouraged to meet their Whatsapp peers in-person on a monthly basis to simulate an in-person connection.

Design details

Randomization method
Randomization will be conducted through the statistical software Stata.
Randomization unit
This study uses a two-stage randomization method. First, 109 villages were randomly selected in the Chittoor district of India. Then within each village, a random sample of at least 10 farmers were surveyed.
Clusters
A total of 109 village clusters.
Observations
We plan to survey at least 1000 farmers, 10 per each village.
Arms
Approximately 30 clusters (villages) and at least 300 farmers per treatment arm (100 farmers per cluster), 1000 farmers in total.
Primary outcomes
Smartphone agriculture technology adoption and usage: Are farmers more likely to adopt and engage in agriculture mobile phone applications if they are connected to other peers using these applications? Learning from virtual vs in-person peers: Is learning different for farmers interacting on Whatsapp with virtual peers from other villages, or do virtual interactions with known farmers from the same village complement learning?
Secondary outcomes
Beliefs about the phone application- Do farmers update their baseline beliefs about smartphone agriculture after engaging in virtual interactions with their peers? Crop loss due to pests: Does using a pest-diagnostic smartphone application reduce crop loss due to pests?

Papers from this trial

  1. Sharma, V. (2023). Social (Media) Learning: Experimental Evidence from Indian Farmers https://cornell.app.box.com/s/mdf52g78ah9o1useuct0q6lktal1jirt

Cited by

No citation data available for this registration yet.

Keywords