Please fill out this short user survey of only 3 questions in order to help us improve the site. We appreciate your feedback!
Testing Means to Scale Early Childhood Development Interventions in Rural Kenya
Last registered on January 15, 2021


Trial Information
General Information
Testing Means to Scale Early Childhood Development Interventions in Rural Kenya
Initial registration date
April 18, 2018
Last updated
January 15, 2021 12:48 PM EST

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

Request Information
Primary Investigator
RAND Corporation
Other Primary Investigator(s)
PI Affiliation
RAND Corporation
Additional Trial Information
In development
Start date
End date
Secondary IDs
Recent neurobiological and psychological research has established that vital development occurs in language, cognitive, motor and socio-emotional development during the first few years of life, and early life outcomes are key determinants of adult outcomes such as educational achievement, labor market outcomes, and health. Yet more than 200 million children under age five in low and middle income countries (LMICs) will fail to reach their developmental potential as adults, predominantly due to poverty, poor health and nutrition, and inadequate cognitive and psychosocial stimulation. Early childhood development (ECD) interventions that integrate nutrition and child stimulation activities have been proposed as a powerful policy tool for the remediation of early disadvantages in poor settings, and numerous field studies have shown they can be effective in improving children's developmental and health outcomes, at least in the short-term. Key questions remain on what models of delivery are the most effective and cost-effective that can be potentially scalable in LMICs, as well as how to sustain parental behavioral changes over time, which can lead to long-term improvements in child development and the possibility of positive spillovers to benefit younger siblings. Having a better understanding of the underlying behavioral pathways leading from intervention, to parental behavior changes, to child impacts, is also key to inform policy about the optimal design of interventions to maximize their scalability and sustainability. This study will conduct a multi-arm clustered randomized controlled trial across 60 villages and 1200 households in rural Western Kenya that tests different potentially cost-effective delivery models for an ECD intervention with a curriculum that integrates child psychosocial stimulation and nutrition education. Selected households will undergo baseline and follow-up surveys to measure short-term impacts in parental behaviors and children's developmental outcomes, and the study will collect data on potential mediators of parental behavioral change to uncover the pathways leading to impacts. Two follow-up surveys, one immediately after the end of the planned intervention and a second two years later, will enable testing of the short term and midterm sustainability of impacts, as well as the presence of any spillovers onto younger siblings. In collaboration with a local NGO, the Safe Water and AIDS Project (SWAP), community health volunteers (CHVs) will be trained to implement the intervention by introducing the ECD curriculum in their villages.

The goal of this study is to provide policymakers with rigorous evidence of how best to expand ECD interventions in low-resource rural settings.
External Link(s)
Registration Citation
Garcia, Italo and Jill Luoto. 2021. "Testing Means to Scale Early Childhood Development Interventions in Rural Kenya ." AEA RCT Registry. January 15. https://doi.org/10.1257/rct.2913-1.1.
Former Citation
Garcia, Italo, Jill Luoto and Jill Luoto. 2021. "Testing Means to Scale Early Childhood Development Interventions in Rural Kenya ." AEA RCT Registry. January 15. http://www.socialscienceregistry.org/trials/2913/history/83963.
Experimental Details
Arm 1 ("group" sessions): Group meetings only
Intervention for Arm 1:
Behavioral: Group sessions
Mother-child dyads in Arm 1 households will receive biweekly ECD sessions for a total of 16 sessions over 7 months. CHVs will record attendance at each session. To maximize participation, prior to each session the CHVs will send an SMS reminder of the session's topic, time and location to all participants.

After the end of the 16 biweekly sessions, we will re-randomize across the 40 intervention villages, stratified by Arms 1 and 2, and half of each of Arm 1 and Arm 2 villages will receive bimonthly booster visits for the period between endline and follow-up surveys.

Arm 2 ("group+home" sessions): Group meetings with a limited number of individual home visits and booster sessions
Intervention for Arm 2:
Behavioral: Group+Home sessions
Households in Arm 2 will receive a total of 16 sessions with identical content similar to Arm 1, but 4 of those sessions will replace group sessions held at the level of villages for personalized home visits, in which the CHV will visit each participant household to deliver these sessions. These home visits will cover identical material and topics as the group sessions in Arm 1 villages, but will be delivered on a personalized basis in the home of the mother and child. Personal barriers to the practices will be discussed and an active resolution strategy developed in concert with the CHV.

After the end of the 16 biweekly sessions, we will re-randomize across the 40 intervention villages, stratified by Arms 1 and 2, and half of each of Arm 1 and Arm 2 villages will receive bimonthly booster visits for the period between endline and follow-up surveys.

Arm 3: No Intervention
This arm will serve as the control group to identify the effects of a parenting intervention and the most effective mode of delivery, as well as the sustained impacts from the intervention

Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Changes in child development, and changes in parental behavior
Primary Outcomes (explanation)
The Bayley Scales of Infant Development 3rd edition (Bayley's III), is validated in African settings and provides measures for all dimensions of child development up to 42 months of age. In the baseline and follow-up surveys, the study will also use the Malawi Developmental Assessment Tool (MDAT), which can be applied to children up to 6 years old and is culturally relevant in Sub Saharan African (SSA) settings, with good reliability.

At baseline, the study will use the Family Care Indicators, a self-reported scale of parenting practices which measures the quality time spent with children in learning and playing activities for young children at home. Examples of questions are how often parents take children out to the park, or other recreational activities, whether there is always an adult looking after children, the frequency of learning and play activities with children, and the amount and variety of play and learning materials. In the first follow-up survey the study will complement the Family Care Indicators with observational measures of practices using the Home Observation for Measurement of the Environment (HOME) inventory, a 45 item structured survey combining mother self report and observational measures.
Secondary Outcomes
Secondary Outcomes (end points)
1. Changes in child growth
2. Changes in preventive health practices
3. Perceived social support
4. Problem solving/social support
5. Maternal Knowledge
6. Maternal Beliefs
Secondary Outcomes (explanation)
- Anthropometrics will be measured using techniques for the WHO Multicenter Growth Reference Study (MGRS).
- Food security will be measured using the Household Food Insecurity Access Scale (HFIAS). This scale measures food insecurity. The questionnaire consists of nine questions that represent a generally increasing level of severity of food insecurity (access). Each HFIAS question consists of the stem (timeframe for recall), the body of the question (refers to a specific behavior or attitude), and four response options (0 to 3) that represent a range of frequencies (0 = never, 1 = rarely, 2 = sometimes, 3 = often). The HFIAS score is a continuous measure of the degree of food insecurity (access) in the household (in the past 30 days). The maximum score for a household is 27 (the household response to all 9 questions was "often", coded with response code of 3); the minimum score is 0. The higher the score, the more food insecurity (access) the household experienced. The lower the score, the less food insecurity (access) a household experienced.
-The study will measure perceived social support using the Multidimensional Scale of Perceived Social Support (MSPSS), which has the advantage of measuring distinctly social support from family, friends and a significant other, which is key to analyze the differential impact of our experimental arms.
-The study will adopt measures from investigators' work in Uganda to measure ways of coping with interpersonal conflicts and daily stressors. Parental stress will be assessed using the Parenting Stress Index (PSI). The study will measure maternal psychological well-being using the widely used Center for Epidemiologic Studies Depression Scale (CESD) with proven psychometric properties
-The study will elicit maternal knowledge about child development through asking mothers about the ages at which they think the child would be able achieve certain developmental milestones, which are then compared with the expected ages reported in the literature.
-The study will adapt and measure the scale to elicit beliefs developed by Cunha et al. (2013)with the target of eliciting parental beliefs regarding the benefits of providing children better cognitive and non-cognitive stimulation. The instrument asks parents about developmental milestones in language and socio-emotional development under different home scenarios, which are constructed using data from the Family Care Indicators.
Experimental Design
Experimental Design
The evaluation design is a factorial assignment and consists of a clustered Randomized Control Trial (cRCT) in which 60 CHVs and their associated villages will be randomly assigned to one of three equally-sized treatment arms. Arms 1 and 2 comprise a 2x2 factorial design, while Arm 3 will serve as a control group. Our experimental design features 3 treatments: group meetings only with some booster sessions (Arm 1), group meetings with a limited number of individual home visits and booster sessions (Arm 2), and the involvement of fathers in the integrated intervention (one half of villages in Arms 1 and 2). All households, including the control group (Arm 3), will receive basic information about child feeding during a baseline survey.
Experimental Design Details
Not available
Randomization Method
Randomization done in office by computer
Randomization Unit
Village Randomization.

After all baseline surveys are complete, we will randomly assign the 60 CHVs and their associated villages to one of three study arms. Each arm will have 20 CHVs and 400 households. After assignment to study arm, villages assigned to Arms 1 or 2 will undergo a secondary randomization
procedure to determine those villages that additionally will invite fathers to the sessions focused on fathers’ involvement (in a 1:1 ratio among the 40 villages assigned to Arms 1 and 2). All randomizations will be stratified by Jamii center to ensure balance across treatment arms on any village-level characteristics that have the potential to have some relationship with intervention effects.
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
60 villages
Sample size: planned number of observations
3802 observations (1200 households)
Sample size (or number of clusters) by treatment arms
For the full evaluation, a total of 60 villages-- each arm will have 20 CHVs and 400 households
o 20 in Arm 1 (groups only), 20 in Arm 2 (groups plus home visits), and 20 in Arm 3 (Control group)
o Among the 40 villages in Arms 1 and 2, 20 will have fathers invited; 20 without fathers invited and only mothers and children are invited
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
In a side-by-side comparison between study arms, the sample size of 400 mother-child dyads in each arm provides 80% power to detect an increase in children’s cognitive and receptive language development of 0.30 SD at the 5% level of statistical significance. We conservatively assume 75% compliance among those invited to the sessions, 15% attrition, and an ICC of 0.07 within 60 CHV catchment areas.
IRB Name
RAND Human Subjects Protection Committee
IRB Approval Date
IRB Approval Number