Texting for Nutrition
Last registered on May 18, 2017


Trial Information
General Information
Texting for Nutrition
Initial registration date
October 27, 2016
Last updated
May 18, 2017 9:46 AM EDT

This section is unavailable to the public. Use the button below to request access to this information.

Request Information
Primary Investigator
The World Bank
Other Primary Investigator(s)
PI Affiliation
The World Bank
Additional Trial Information
On going
Start date
End date
Secondary IDs
Chronic malnutrition is a persistent and urgent challenge in the rural, mountainous, and mostly indigenous area of Chimborazo, Ecuador. In the Chimborazo province, 49 percent of children under five were stunted in 2013 (ENSANUT). This experiment tests how behavior change can play a role in improving indicators of child nutrition. Caregivers in both rural and urban areas enrolled in the Texting for Nutrition Program led by the World Bank and the Provincial Government of Chimborazo. A randomly selected subset of enrolled caregivers received biweekly text messages that encouraged specific types of caregiver behavior change. Each set of behaviors were designed to improve child nutrition through a different behavioral mechanism: increased attendance to health checkups, increased consumption of micronutrients, improved diet diversity, or improved water sanitation and hygiene practices. We will test whether the text messages affected caregiver behaviors and improved nutrition outcomes for children. We will also be able to test which behavioral channels were more effective because the thematic area of text messages received by caregivers was randomly assigned at the individual level, allowing for a comparison across text message treatment arms.
External Link(s)
Registration Citation
Oviedo, Ana and Megan Rounseville. 2017. "Texting for Nutrition." AEA RCT Registry. May 18. https://www.socialscienceregistry.org/trials/1712/history/17770
Experimental Details
The Texting for Nutrition program was implemented by the World Bank together with the Decentralized Autonomous Government of the Chimborazo Province, and the Ecuadorean Ministry of Social Development Coordination (MCDS). Caregivers were enrolled into the program in urban and rural areas of Chimborazo. The Texting for Nutrition program used text messages to encourage caregiver health behavior change. The content of the text messages varied along five thematically focused treatment arms that each encouraged a unique set of behaviors. In treatment arm one (T1), caregivers were encouraged to bring their children in to their local health center for regular and timely checkups. In treatment arm two (T2), caregivers were encouraged to feed their children micronutrients specifically an iron supplement locally called “Chispas”. In treatment arm three (T3), caregivers were provided information about when they should initiate complementary feeding, and how they could increase their child's diet diversity. In treatment arm four (T4), caregivers were encouraged to ensure that water consumed by their children was potable, to wash their hands regularly, as well as to improve general food preparation and hygiene practices. Lastly in treatment arm five (T5), caregivers were sent a mix of text messages along all four content types from T1, T2, T3 and T4. Additionally, caregivers with children under six months of age were sent text messages about exclusive breast feeding, however since this treatment type was age-specific it is not included in the accompanying evaluation.

The content of the text messages was designed with four types of behavioral mechanisms in mind: providing timely new and relevant information, positively encouraging caregivers, using persuasive social norms language, and sending reminders to keep key messages ‘top of mind’. The nutrition messages were developed in collaboration with the Ecuadorean Ministry of Health, locally developed nutrition messages from the Creciendo con Nuestros Guaguas program, and several international food-security resources including the FAO Complementary Feeding for Children Aged 6-23 Months Recipe Book, among others. Each caregiver received approximately 70 text messages in total over a year and a half of program intervention.
Intervention Start Date
Intervention End Date
Outcomes (end points)
The primary outcomes of interest for the evaluation are nutrition and health indicators for children, including but not limited to anthropometric measures of stunting. The evaluation will also estimate the effects of the program on caregiver behaviors related to the following outcomes: 1. Visits to local health centers with their children for regular check-ups; 2. Consumption of nutritional supplements with a particular focus on the locally available iron supplement Chispas; 3. Boiling or treating water served to children for consumption; 4. Child diet diversity; 5. Exclusive breastfeeding from birth through 6 months of age; 6. Timely integration of solid foods for children; 7. Frequency and timing of hand washing by caregivers (prior to preparing food, after defecation, and prior to interacting with infants); and 8. Maintaining clean cooking and food preparation spaces, and general hygiene.
Outcomes (explanation)
Experimental Design
Experimental Design
Texting for Nutrition was designed and implemented as a randomized control trial. Upon enrollment into the program caregivers were randomly assigned to one of the five treatment arms or to a control group. The evaluation is designed to estimate the effects of receiving the text messages on improving caregiver behaviors and the nutrition and health outcomes for children. The evaluation will also enable a comparison of results across treatment arms (individual level random assignment), to assess whether policy makers can address barriers in one thematic area through behavior change more effectively, as compared to others.
Experimental Design Details
Not available
Randomization Method
The randomization was done using STATA. Random assignment to treatment and control was conducted within matched pairs of parroquias (a geographic unit smaller than a province, larger than a municipality). Pairs were formed using parroquia level characteristics on nutrition, adult education, presence of health centers, and other demographic characteristics. One parroquia from each pair was randomly assigned to treatment and the other to control. Within the treatment group, individuals were randomly assigned to one of five treatment arms.
Randomization Unit
Assignment to treatment and control groups was done at the parroquia level. Within the treatment group random assignment into the five treatment arms was done at the individual caregiver level. In the case of households that had registered more than one caregiver, one caregiver was randomly selected per household, this random selection was stratified by the sex of the caregiver.
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
50 parroquias. (Assignment to treatment and control was clustered at the parroquia level, however, assignment to treatment arms was not clustered.)
Sample size: planned number of observations
3,000 caregivers
Sample size (or number of clusters) by treatment arms
T1: 400 caregivers, T2: 400 caregivers, T3: 400 caregivers, T4: 400 caregivers, T5: 400 caregivers, C: 1,000 caregivers.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
Tufts University: Social, Behavioral, and Educational Research Institutional Review Board
IRB Approval Date
IRB Approval Number