Information and farming practices in Bangladesh

Last registered on April 28, 2020

Pre-Trial

Trial Information

General Information

Title
Information and farming practices in Bangladesh
RCT ID
AEARCTR-0004807
Initial registration date
April 27, 2020
Last updated
April 28, 2020, 10:56 AM EDT

Locations

Region

Primary Investigator

Affiliation
Yale University

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2019-09-25
End date
2020-05-15
Secondary IDs
Abstract
Potatoes are a staple crop unique in their high nutrient content and susceptibility to the plant pathogen Phytophtera infestans, also known as the fungal late blight. Blight, made infamous by the Irish potato famine of the 1840s, continues to suppress potato yields globally. Losses to blight are ten times greater in low- than high-income countries, limiting the capacity of potato farming to provide food or economic security. Blight is prevalent wherever potatoes are grown, appearing under cool, wet conditions. The spores of the disease spread quickly through wind and water. Blight first infects the leaves of the potato plant, moving down into the tuber, which will rot and deteriorate in the field. Left untreated, blight can destroy an entire crop within a week of infection. The remedy is simple: an application of a prophylactic fungicide prior to infection. However, average losses are high, estimated at between 25 and 57% of potatoes in Bangladesh. In extreme cases, such as the 2006-2007 season, 50 to 80% of all potato crops in Bangladesh were infected with blight, resulting in severe yield losses across the country.

Farmers in Bangladesh have not been able to effectively contain blight because they face a difficult decision in choosing whether or not to
spray their crops. If farmers spray unnecessarily, they pay a cost for no benefit. If farmers do not spray when necessary, then an uncontrolled outbreak of blight can leave them ruined. Blight imposes costs on farmers twice over: first, in the immediate yield losses to
the disease and the financial deprivation that ensues if farmers fail to spray fungicide when necessary. A ruined harvest is not merely lost
income, but also lost investments in seed, fertilizer, and labor. Second, the risk of blight leads farmers to underinvest in technology and inputs for their farm, or leave their fields fallow. Why invest in high yield seeds, or new practices and technologies when they could all be destroyed?

GEOPOTATO is an information and communication technology that provides farmers in Bangladesh with timely alerts about the risk of
blight. GEOPOTATO was built on decades of research by Wageningen University & Research, and has been field-tested in
Bangladesh over the past three years. GEOPOTATO combines satellite and local weather data with an epidemiological and crop growth model to forecast blight pressure at a high geographic and temporal resolution. GEOPTOATO alerts are calibrated using two parallel models: one tracking the growth of the potato crop and the other tracking the lifecycle of blight. The potato growth model is calibrated using data on local potato varietals and sowing dates, and is updated throughout the season using satellite data. The blight model takes local weather station data as an input to predict the likelihood of a blight outbreak. It is at the intersection of these models, when we predict a blight outbreak concurrent with potatoes that are at a susceptible point in their growth cycle, that GEOPOTATO triggers an alert.

When the GEOPOTATO system detects an elevated risk of blight, it sends an SMS and voice message in the local dialect to farmers with at-risk crops, telling them that they should spray a prophylactic fungicide within the next three days. The alert is designed to be clear, timely, and actionable. Farmers, already aware of the importance of controlling late blight, can go to the village dealer to purchase
fungicide and apply it to their fields. Previously, farmers have lacked information about when and when not to spray. Even experienced
farmers who are familiar with the weather conditions conducive to blight cannot directly observe the presence of the fungal spores, or
how small changes in weather patterns affect the risk their crops face based on their precise stage of growth. GEOPOTATO alerts remove the uncertainty and guesswork over the daily risk presented by blight.
External Link(s)

Registration Citation

Citation
Barnett-Howell, Zachary. 2020. "Information and farming practices in Bangladesh." AEA RCT Registry. April 28. https://doi.org/10.1257/rct.4807-1.0
Sponsors & Partners

Partner

Type
ngo

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
Experimental Details

Interventions

Intervention(s)
Sending text message and voice alerts to farmers in the Rangpur region of Bangladesh during the 2019-2020 potato season about the likelihood of late fungal blight (Phytophthora infestans). The alerts are generated by the GEOPOTATO system, a combination of satellite weather data and crop growth model, that estimates the likelihood of a blight outbreak for farmers who planted in a certain region at a certain time. If the likelihood of a blight outbreak exceeds a certain threshold the system triggers an alert to participating farmers, who are notified of the danger and that they should spray fungicide within three days.
Intervention Start Date
2019-09-25
Intervention End Date
2020-05-15

Primary Outcomes

Primary Outcomes (end points)
Primary outcomes include potato yield and revenue; fungicide expenditures and blight damage to the potato crop; the use of inputs, including number of people hired as agricultural labor and fertilizer expenditures.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
A secondary outcome of interest is the degree to which information spillovers occur between those farmers directly receiving GEOPOTATO alerts those not receiving alerts in the same village.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experiment has three arms: (1) control villages, where none of the surveyed farmers receive GEOPOTATO alerts, (2) treatment villages, where all of the surveyed farmers in the study receive GEOPOTATO alerts, and (3) partial treatment villages, where a fraction of the surveyed farmers receive GEOPOTATO alerts. Farmers themselves are randomly assigned whether or not to receive GEOPOTATO alerts within treatment and partial treatment villages.
Experimental Design Details
Randomization Method
Treatment arms (villages) and status (farmers) is assigned using the random number generator in R 3.5.3.
Randomization Unit
Assignment is stratified at the upazila and union level. Villages comprise clusters. The lowest level of observation is a harvest, and farmers are the primary economic actor.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
415 villages
Sample size: planned number of observations
2,500 farmers
Sample size (or number of clusters) by treatment arms
180 control villages; 125 treatment villages, and 110 partial treatment villages
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
5% change in log yields (kg/decimal).
IRB

Institutional Review Boards (IRBs)

IRB Name
Yale University IRB
IRB Approval Date
2019-10-29
IRB Approval Number
2000026712
Analysis Plan

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials