Orange-Fleshed Sweet Potato Sustainability Project in the Lake Region in Tanzania

Last registered on June 30, 2025
View Trial History

Pre-Trial

Trial Information

General Information

Title
Orange-Fleshed Sweet Potato Sustainability Project in the Lake Region in Tanzania
RCT ID
AEARCTR-0016110
Initial registration date
June 28, 2025

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
June 30, 2025, 6:12 AM EDT

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

Locations

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

Request Information

Primary Investigator

Name
Juan Bonilla
Affiliation
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Rosa Castro-Zarzur
PI Affiliation
AIR
PI Name
Andi Coombes
PI Affiliation
AIR
PI Name
Frederick Grant
PI Affiliation
International Potato Center (CIP)
PI Name
Ariel BenYishay
PI Affiliation
College of William and Mary
PI Name
Katherine Nolan
PI Affiliation
College of William and Mary
PI Name
Kwame Ogero
PI Affiliation
International Potato Center (CIP)

Additional Trial Information

Status
In development
Start date
2024-12-01
End date
2026-11-30
Keywords
Agriculture, Gender, Health
Additional Keywords
Technology adoption, Agriculture, Social networks, Gender discrimination
JEL code(s)
012 J16
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The adoption of orange-fleshed sweet potato (OFSP) among smallholder farmers in sub-Saharan Africa (SSA) is a highly effective strategy for addressing vitamin A deficiency, a condition that can impair immune function, increase the risk of severe infections, and cause preventable blindness. However, adoption rates remain low in many parts of SSA, and farmers who initially adopt OFSP often dis-adopt it within a few production cycles following the end of agricultural interventions that promote it. Existing evidence indicates that adoption is influenced by a combination of farmers’ socio-economic characteristics, consumption preferences, and agronomic traits; however, the reasons behind dis-adoption remain largely unexplained and underexplored in the literature.

In this study, we assess the relevance of key potential mechanisms to enhance the sustainability of OFSP adoption over time among smallholder farmers with children under 5 in the Lake Region in Tanzania. The intervention focuses on strengthening the entire value chain—from production to consumption by conducting the following activities: (1). Strengthen seed systems and input markets to ensure that farmers have consistent access to quality planting materials. This involves supporting vine multipliers, establishing decentralized vine nurseries, and facilitating private sector participation in seed distribution; (2). Provide agricultural and nutrition trainings to promote OFSP adoption among female and male farmers and conducting awareness campaigns highlighting the nutritional benefits of OFSP to stimulate consumer demand; (3). Provide commercialization trainings to farmers, local bakers, and processors to support the creation of a robust demand for OFSP roots and processed products (e.g., bread, chips, puree) to cater to diverse consumer preferences.

To assess the program impacts, we designed a 2-year, mixed-methods experimental study in 120 villages in the Lake Region. The evaluation randomizes access to inputs and trainings at the village level with 75 villages receiving trainings on OFSP agricultural production, nutrition, and commercialization as well as connecting OFSP producers with local OFSP buyers and processors; 25 villages receiving only the agricultural and nutrition trainings; and 20 villages in the pure control group. A key aspect of the design is that it explores the role that commercialization has on sustainability.

In addition, in the 75 villages receiving trainings on OFSP agricultural production, nutrition, and commercialization, we will leverage aspects of the selection of the community-based extension agents (CBEAs) -who are trained farmers- to investigate how the social diffusion of the information provided by the project occurs and to what extent the number and gender of the trainers disseminating information influences program impacts. Furthermore, we also collect qualitative data to assess the quality and extent to which the program is implemented as designed, identify challenges and facilitators to program implementation, and understand how beneficiaries in the program areas experience the program.
External Link(s)

Registration Citation

Citation
BenYishay, Ariel et al. 2025. "Orange-Fleshed Sweet Potato Sustainability Project in the Lake Region in Tanzania." AEA RCT Registry. June 30. https://doi.org/10.1257/rct.16110-1.0
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 adoption of orange-fleshed sweet potato (OFSP) among smallholder farmers in sub-Saharan Africa (SSA) is widely recognized as an effective strategy for addressing vitamin A deficiency—a condition that impairs immune function, increases the risk of severe infections, and can cause preventable blindness. Despite its proven benefits, OFSP adoption rates remain low in many parts of SSA. Moreover, even among initial adopters, dis-adoption is common within a few production cycles after the conclusion of supporting agricultural interventions. While existing evidence identifies factors such as farmers’ socio-economic characteristics, consumption preferences, and agronomic traits as key determinants of adoption, the drivers of dis-adoption remain largely underexplored in the literature.

This study investigates mechanisms that could improve the sustainability of OFSP adoption among smallholder farmers in Tanzania’s Lake Region among households with children under 5 years old. The intervention targets persistent barriers to the uptake and sustained use of new crop varieties by resource-constrained farmers, including credit constraints, risk aversion, limited access to agronomic and market information, and entrenched gender norms that shape the roles crops play within households (Carter et al., 2013; Duflo et al., 2006; Foster & Rosenzweig, 1995).

The intervention seeks to support the OFSP value chain—from production to consumption—through three core components:

* Strengthening seed systems and input markets to ensure reliable access to quality planting materials by supporting vine multipliers, establishing decentralized vine nurseries, and fostering private sector involvement in seed distribution.

* Delivering agricultural and nutrition trainings to both women and men to promote OFSP cultivation and use, coupled with awareness campaigns highlighting its nutritional benefits to stimulate consumer demand.

* Providing commercialization training to farmers, local bakers, and processors to expand markets for OFSP roots and value-added products such as bread, chips, and puree. The program will then connect OFSP farmers with local bakers and OFSP processors.
Intervention Start Date
2024-12-01
Intervention End Date
2026-06-30

Primary Outcomes

Primary Outcomes (end points)
The key outcome variables of interest in the experiment are:

1. OFSP Adoption
Whether farmers decide to plant orange-fleshed sweet potato (OFSP), particularly commercially viable varieties.

2. Knowledge Gains from the nutrition and agricultural trainings
Increases in farmers’ knowledge of OFSP agricultural production and nutrition.

3. OFSP Production and Productivity
Changes in the levels of OFSP production and crop management practices.
Productivity gains.

4. OFSP Consumption and consumption of Vitamin A rich foods
Household-level consumption of OFSP, especially among mothers and children under 5.
Changes in dietary diversity using the Diet Quality Questionnaire and vitamin A intake

5. Child morbidity
Probability of being sick in the last 7 days before the survey and probability serious illness that requires hospitalization
Probability of death of young children in the household

6. Commercialization Outcomes
Likelihood of farmers engaging in OFSP commercialization activities
Sales in local currency and quantity sold
Sales revenue

7. Norms and Behaviors
Shifts in community and individual beliefs, norms, and behaviors that affect sustained OFSP adoption.

Primary Outcomes (explanation)
Most outcomes referenced above are not constructed. The only outcomes that will be constructed from the data collected is the diet diversity outcomes. For this, we will apply the Diet Quality Questionnaire, developed by the Global Diet Quality Project, a standardized tool to estimate dietary patterns. With data from the DQQ module, we can construct a number of different diet measures, including the minimum dietary diversity for women (MDD-W), which is an indicator of micronutrient sufficiency that reflects whether respondents ate at least five of 10 specific food groups during a day. With the DQQ, we will also construct the Dietary Diversity Score and the All-5 recommended good groups consumed.

For children aged 6 to 23 months, we also collect data on the number of servings consumed of focus foods over the past 7 days before the survey. We complement these data with the DQQ for infant and young child feeding (IYCF). The main IYCF-DQQ indicators assessed are: (1). Breastfeeding indicators (i.e., ever breastfed (EvBF); early initiation of breastfeeding (EIBF); exclusively breastfed for the first two days after birth (EBF2D)); and (2). MDD for 6-23 months.

Secondary Outcomes

Secondary Outcomes (end points)
Social diffusion: Adoption of OFSP by an external sample of farmers that are not part of project activities but who live in the catchment areas where the program is implemented. The evaluation will collect data from up to 40 members (beyond the program participants) in all villages to assess OFSP production and consumption.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experimental design involves a randomized controlled trial conducted across 120 villages in Tanzania, encompassing a total of 2,000 farmers. Of these, 20 villages (with 333 farmers) serve as a pure control group, while the remaining 100 villages (1,667 farmers) are assigned to treatment conditions.

The treatment group is further divided into two main arms: T1 and T2. T1 includes 25 villages who will receive nutrition and cooking trainings along with access to OFSP vines. T2 comprises 75 villages who will receive the same interventions as T1, plus additional agricultural and commercialization trainings. Within T2, the design explores the impact of varying the number and gender of trained community-based extension agents (CBEAs) per village—either 1, 3, or 5 trained farmers—with some groups having gender-specific assignments (e.g., only female CBEAs) and others allowing any gender. This layered structure enables the evaluation of both the direct effects of commercialization support and the influence of social diffusion mechanisms, particularly the role of the number of CBEAs and their gender in information dissemination and adoption behaviors.
Experimental Design Details
Not available
Randomization Method
CIP team identified 120 villages in the lake region. Evaluation team conducted the randomization in office by a computer using Stata.
Randomization Unit
the unit of randomization is at the village level
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
120 villages for the main sample
Sample size: planned number of observations
2000 farmers distributed evenly in the 120 villages Team will also collect information from 30 farmers in the catchment areas of the 120 villages to measure social difussion
Sample size (or number of clusters) by treatment arms
120 villages (2,000 farmers total). Of these,
*C: 20 villages (333 farmers) in the pure control group
* T: 100 villages (1,667 farmers) in treatment groups

The 100 villages in the treatment group are further divided into:
* T1 (25 villages; 417 farmers): Nutrition and cooking trainings + access to OFSP vines.
* T2 (75 villages; 1,250 farmers): Same as T1 + agricultural and commercialization trainings.

and the 75 villages from T2 are divided into:
* T2-1: 25 villages with 1 trained farmer to serve as the CBEA for the commercialization component
* T2-2: 25 villages with 3 trained farmers to serve as CBEAs for the commercialization component
* T2-3: 25 villages with 5 trained farmers to serve as CBEAs for the commercialization component

In each one of the T2-x groups (with x=1,2,3), half of the villages will have female CBEAs (gender reservation) and in the other half of the villages CBEAs of any gender can be selected.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
To assess overall impact of the program (T vs C) on the probability of OFSP adoption, we calculate that, for a sample of 2000 farmers from 120 villages, the minimum detectable effect size (MDES) is 0.22 standard deviations (SD), which is equivalent to a 10-percentage-point increase in OFSP adoption probability. We assume an intra-cluster correlation (ICC) in OFSP adoption of 0.09 based on a previous project in a nearby district in Tanzania. We also assume a 5% level of confidence and 80% power. For the comparison between T1 and T2, which includes a total of 100 villages, to assess the additional impact of the commercialization component relative to the other program components, we calculate an MDES of 0.21 SD using the same assumptions. To assess the difference between one of the T2-x (x=1,2,3) treatment groups relative to any of the other two groups, the estimated MDES is 0.26 SD. Lastly, to assess differences based on the gender of the CBEAs, we will aggregate all the villages by the randomized gender of the CBEA. Thus, we will have a total of 75 villages of which 38 will have a gender reservation condition and 37 will not. Using the same assumptions, we calculate an MDES of 0.21 SD.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
American Institutes for Research in the Behavioral Sciences
IRB Approval Date
2025-06-19
IRB Approval Number
8F164
Analysis Plan