Evaluation of an agricultural technology platform service in India

Last registered on May 24, 2023

Pre-Trial

Trial Information

General Information

Title
Evaluation of an agricultural technology platform service in India
RCT ID
AEARCTR-0011454
Initial registration date
May 20, 2023

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
May 24, 2023, 4:43 PM EDT

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

Locations

Region

Primary Investigator

Affiliation
IFPRI

Other Primary Investigator(s)

PI Affiliation
IFPRI
PI Affiliation
IFPRI

Additional Trial Information

Status
On going
Start date
2023-05-20
End date
2024-09-30
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Small-scale Producers (SSPs) across the developing world frequently lack access to reliable information and extension, crucial for prolonged increases in farm productivity and income. In regions such as South Asia, farmer land holdings are generally small, so that increasing SSP incomes requires sustainable augmentation of farm productivity. In the absence of opportunities to consolidate farm holdings, productivity increases and diversification into high value cash crops are widely regarded as key drivers of small farmer incomes. New forms of
extension using Information and Communication Technologies (ICT) coupled with remote sensing data have emerged as promising interventions that may help to sustainably increase farmer productivity and aid transition to higher value crops. This project aims to evaluate one such innovation in agricultural extension with the potential to improve efficiency and surplus for stakeholders - a proprietary digital platform that provide SSPs with farming advice and facilitate two-way communication to address farmer- and crop-specific concerns. Our
evaluation will study the impact of the introduction of this platform by one client on soybean farmers in Maharashtra, India. We use a stratified clustered randomized controlled trial with two arms, treatment and control, to assess impacts on our outcomes of interest, which include (among others) knowledge and adoption of recommended practices, information seeking, agricultural yields, costs and profits, prices and storage behavior. We also study spillover effects on farmers in the same areas who are not linked with the client, but who might benefit indirectly from this intervention. In addition, we conduct nested experiments relating to measurement error in survey data, recall bias and expectations.
External Link(s)

Registration Citation

Citation
Gautam, Aditi, Sudha Narayanan and Kalyani Raghunathan. 2023. "Evaluation of an agricultural technology platform service in India." AEA RCT Registry. May 24. https://doi.org/10.1257/rct.11454-1.0
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
The main intervention being studied is the rollout of digital platforms. First, a B2B platform to be used by the extension workers to provide timely information and to monitor farmer activities. Second, a B2B2C platform that will be downloaded onto the phones of farmers linked to the client-user, and will deliver information on weather, prices at various client purchase centers and soybean package of practices, while allowing farmers to use two-way communication to receive customized advice. These platforms will be introduced in a randomly selected set of villages where the client-user is active, enabling the use of statistical methods to attribute impact.
Intervention (Hidden)
The main intervention being studied is the rollout of CropIn's digital platforms. First, CropIn Grow, a B2B platform to be used by the extension workers to provide timely information and to monitor farmer activities. Second, CropIn Connect, a B2B2C
platform that will be downloaded onto the phones of farmers linked to the CropIn client, and will deliver information on weather, prices at various client purchase centers and soybean package of practices, while allowing farmers to use two-way communication to receive.
customized advice. These platforms will be introduced in a randomly selected set of villages where the CropIn client is active, enabling the use of statistical methods to attribute impact. We will also conduct three experiments nested within the impact evaluation of CropIn's digital
platforms.
Experiment 1: We will test the effectiveness of digital versus analog modes of data collection on costs of cultivation. Survey data collection is increasingly being administered through Computer-Assisted Personal Interviews (CAPI), however, detailed and often non-linear modules such as cost of cultivation of crops across plots is easier collected using analog methods that enable flexibility in recall and data entry. The costs of cultivation module for the main impact evaluation will be collected using pen and paper. To test the respondent burden
(in terms of time) as well as the accuracy of data collected, we will select at random a subset of farmers from the set of sampled farmers, to whom we will administer a digitized cost of cultivation module. This experiment has the potential to inform choice of method for complex surveys, especially when the potential for respondent burden and cognitive overload is high.

Experiment 2: Survey data on self-reported farmer plot sizes and yields has been shown to suffer from systematic measurement error. Whether these errors arise from misreporting or misperceptions, they have implications for our understanding of farmer decisions around input use and practices. In rural India, where land is largely inherited and careful land size measurements are infrequent, farmers often do not have a chance to update their beliefs about their land size. To assess whether correcting farmer beliefs changes their behaviors, we digitize farmer plots at baseline using GPS mapping software, and inform a randomly selected subsample of farmers of their actual plot sizes. We are then able to compare both farmer reports about plot size as well as their farming behaviors across baseline and endline
surveys.

Experiment 3: A third experiment will focus on the holding and storage behavior of soybean farmers. Ongoing research on farmers' price expectations suggest that farmers may forego remunerative arbitrage opportunities based on their expectations of future prices. Farmers may however have exaggerated expectations that might prompt them to hold stocks for longer than is optimal. We will randomize an information treatment involving sharing price data to assess the impacts.
Intervention Start Date
2023-06-15
Intervention End Date
2024-03-31

Primary Outcomes

Primary Outcomes (end points)
Primary Outcomes (end points)
1. Key outcomes for sustainable and conventional farmers
a. Knowledge and adoption of recommended agricultural practices
b. Information seeking behaviour (enabled by CropIn Connect’s farmer queries feature)
c. Yield, costs of production, profits
d. Storage, holding and sale timing or behaviour
e. Price realization, either because of improved linkages or enhanced output quality
f. Spillover effects – where other farmers not enrolled in the program approach CropIn
Connect users for advice
2. Key outcomes for farmers under the sustainable program
a. Uptake of required practices for certification
b. Progress through the three certification stages
Primary Outcomes (explanation)
Outcome 1(a): knowledge will be constructed as a score based on a knowledge test that will
draw on the content of the agricultural package of practices.
Outcome 1(b): will be constructed based on the frequency and nature of the queries raised
through the farmer phone app
Outcome 1(f): will be constructed using the frequency and nature of the information/advice
farmers not enrolled in the program seek from users of the CropIn phone app

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We propose a stratified cluster randomization design at the village level. The strata are based on combinations of geography and village type based on composition of client registered farmers. The village represents the cluster and will be the level of randomization.
Experimental Design Details
Assessing the impact of CropIn services on participating small-scale producers requires construction of an appropriate counterfactual. Our evaluation will use villages as the basic sampling and treatment unit, aligning with the operations of our partner CropIn client. This
level of treatment will allow us to capture information spillovers within villages. We propose a stratified cluster randomization design at the village level. The strata are based on combinations of geography and village type based on composition of client registered farmers. The village represents the cluster and will be the level of randomization.
The district of Latur, Maharashtra, represents the "core" of the CropIn client's operations - it is where they have had a longer and deeper engagement, and where their processing plant is located. Given this, we propose to stratify by Latur and other neighboring districts,
representing the ‘core’ and ‘periphery’ of the CropIn client’s sphere of influence in the area respectively. Since the client works with both conventional and sustainable farmers, who use significantly different package of practices, we will also stratify by farmer type based on
whether the village has only conventional farmers, or a mix of conventional and sustainable.
Within the four strata so defined – Latur-conventional, Latur-sustainable, Bid/Osmanabad - conventional and Bid/Osmanabad -sustainable - we propose to randomly assign client villages selected for the study into treatment and comparison groups. In the comparison
villages, neither the extension agents nor the registered farmers will receive the digital platforms. In the treatment villages, however, both of these will happen. Extension agents will onboard farmers from treatment villages onto the CropIn Connect platform and will themselves use the CropIn Grow platform to streamline their own extension and monitoring operations.
Randomization Method
Randomization done in office by a computer
Randomization Unit
Village
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
300 villages, stratified into those that fall in the core area of operation of the CropIn client
(closer to processing centers etc) and those that fall in the periphery.
Sample size: planned number of observations
4500 farmers, including sustainable, conventional and spillover farmers.
Sample size (or number of clusters) by treatment arms
15 farmers per cluster
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We use data from the Cost of Cultivation survey conducted by the Government of India; and restrict ourselves to soybean farmers from Maharashtra. With a cluster size of 10 and power of 0.8 we have the following estimates for the Minimum Detectable Impact: Outcome: A2, All actual expenses + Rent paid for leased-in land (INR) MDI = 3418.92 Outcome: C2, All actual expenses + Interest on owned capital assets + rental value of owned land/rent paid for leased-in land + imputed value of family labour (INR) MDI = 4267.49 Outcome: Net Profit (INR, using A2) MDI = 3079.42 Outcome: Yield (quintal/hectare) MDI = 1.19
IRB

Institutional Review Boards (IRBs)

IRB Name
International Food Policy Research Institute (IFPRI) Institutional Review Board
IRB Approval Date
2022-09-28
IRB Approval Number
SAR-22-0947

Post-Trial

Post Trial Information

Study Withdrawal

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

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