Digital Information Provision and Farmer Behavior: Lessons from Six Experiments in East Africa

Last registered on December 01, 2020

View Trial History

Pre-Trial

Trial Information

General Information

Title

Digital Information Provision and Farmer Behavior: Lessons from Six Experiments in East Africa

RCT ID

AEARCTR-0006837

Initial registration date

November 30, 2020

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

December 01, 2020, 1:28 PM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country

Kenya

Region

Country

Rwanda

Region

Primary Investigator

Name

Raissa Fabregas

Affiliation

University of Texas - Austin

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Matthew Lowes

PI Affiliation

PI Name

Michael Kremer

PI Affiliation

PI Name

Robert On

PI Affiliation

PI Name

Giulia Zane

PI Affiliation

Additional Trial Information

Status

Completed

Start date

2014-09-01

End date

2019-06-30

Keywords

Agriculture

Additional Keywords

text-messages, agricultural extension

JEL code(s)

Secondary IDs

Abstract

We present causal evidence on the effects of six different text message-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, which allows us to learn about the extent of impact heterogeneity.

External Link(s)

Registration Citation

Citation

Fabregas, Raissa et al. 2020. "Digital Information Provision and Farmer Behavior: Lessons from Six Experiments in East Africa." AEA RCT Registry. December 01. https://doi.org/10.1257/rct.6837

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

This project examines the impact of mobile-based agricultural extension messages on farmer behavior change. The evidence we present is drawn from six different programs: a program implemented by the Kenya Agriculture and Livestock Research Organization (KALRO), a Kenyan public organization; two programs implemented by One Acre Fund (OAF), a social enterprise; and two programs implemented by a partnership between Precision Agriculture for Development (PAD) and Innovations for Poverty Action (IPA), both non-profit organizations. All programs were delivered via text-messages and encouraged farmers in areas with acidic soils to experiment with agricultural lime, an input used to increase soil pH. Four programs also sent messages about relatively unknown chemical fertilizers. Despite sharing similar objectives, programs varied in terms of farmer recruitment, message content and design, agricultural seasons, and complimentary services.

Intervention Start Date

2014-12-01

Intervention End Date

2018-06-30

Primary Outcomes

Primary Outcomes (end points)

purchase of recommended agricultural inputs (as measured by administrative and survey data)

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

use of non-recommended inputs
knowledge about agricultural lime

Secondary Outcomes (explanation)

Experimental Design

In each one of the trials, farmers were randomly assigned to a control group (did not receive messages) or treatment groups. We summarize main treatment arms below:

(1) KALRO project: 834 farmers divided into a treatment or a control group
(2) IPA/PAD1-K: 1,897 farmers randomized into two treatment arms (general SMS or specific SMS) or a control group
(3) IPA/PAD2-K: 5,890 farmers randomized into three different treatment arms (SMS only, SMS and phone call, SMS and phone call offer) or a control group
(5) OAF1-K: 4,884 farmers randomized into two treatment arms (broad SMS or detailed SMS) or a control group
(4) OAF2-K: 32,572 farmers randomized into two arms (Lime messages or Lime and fertilizer messages) or a control group. Message framing, number of repetitions, and frequency was cross-randomized.
(4) OAF3-R: 110,400 farmers randomized into two arms (Same messages or Different messages) or a control group. Message framing, number of repetitions, and frequency was cross-randomized. Randomization was conducted at the group level.

Experimental Design Details

Randomization Method

Randomization done in an office by a computer.

Randomization Unit

For all experiments except OAF3-R randomization was done at the individual level. For OAF3-R the unit of randomization was the farmer group.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

Sample size: planned number of observations

(1) KALRO: 834 (2) IPA/PAD1-K:1,897 (3) IPA/PAD2-K:5,890 (4) OAF1-K:4,884 (5) OAF2-K:32,572 (6) OAF3-R:110,400

Sample size (or number of clusters) by treatment arms

(1) KALRO: 417 control, 415 treatment
(2) IPA/PAD1-K: 632 control, 632 specific SMS, 633 general SMS
(3) IPA/PAD2-K: 1470 control, 1475 sms, 1473 sms and call, 1472 sms and call offer
(4) OAF1-K: 1559 control, 1684 broad sms, 1641 detailed sms
(5) OAF2-K: 8142 control, 19558 lime only, 4872 lime and fertilizer
(6) OAF3-K: 19743 control, 18988 same SMS, 18821 different sms

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

IRB