AEA RCT Registry

Hogares Infantiles (HIs) are nurseries run by the Colombian government, Instituto Colombiano de Bienestar Familiar (ICBF), providing partly subsided day care and 60% of daily nutritional requirement to children from low socio-economic backgrounds. Parents pay US$10-25 per month. HIs, of which there are 1,008 across Colombia, are community-based, in the sense that they are typically located in fairly well-equipped community centres and are run by the local parents’ association. They employ between 3 and 6 teachers with some training in early education, each caring for about 30 children aged 2-5 years. Some centres also enrol children below 2 years.
An important component of the new government’s Early Childhood policy is the quality enhancement of HIs. This study analyses the effects of quality upgrades provided by the government and a private NGO in services provided by His on children’s development. The upgrade entails:
• hiring additional and better qualified personnel (one health professional and one professional in socio-emotional development per every 200 children, and one paraprofessional pedagogical assistant per every 50 children. ICBF provides an additional US$20 per child per month to HI to hire the new staff, establishing Hogares Infantiles Mejorado (henceforth, HIM)
In addition, a private foundation, Fundación Éxito (FE), provides extra modules (HIM+FE henceforth), at US$40 per child per month, including:
• pedagogical training of teachers and a reading programme of for teachers
• a modified version of the nutritional component (more proteins and vitamins) along with information on nutrition and food preparation.
We evaluated both interventions to identify their impacts as compared to HIs on:
• children’s nutritional status and development (physical, cognitive and socio-emotional)
We exploited the 18-month expansion phase of the program to design a rigorous evaluation, sampling 120 of the HIs to be upgraded. We randomly assigned them to 3 groups:
• controls, switching to the HIM model at the end of the expansion period
• HI, switching to HIM immediately
• HI, switching to HIM+FE immediately.
Our data will includes:
• a baseline survey of 1,987 children aged 18-36 months and their families.
• a follow-up survey of the same children and their families
• nursery and teacher surveys
We have also investigated the mechanisms yielding the estimated impacts and whether the extra costs involved in HIM and HIM+FE are justified.

External Link(s)

Registration Citation

Citation

Attanasio, Orazio et al. 2019. "Evaluation of Centres for Infant Development: an early years intervention in Colombia." AEA RCT Registry. August 02. https://doi.org/10.1257/rct.1246-3.0

Former Citation

Attanasio, Orazio et al. 2019. "Evaluation of Centres for Infant Development: an early years intervention in Colombia." AEA RCT Registry. August 02. https://www.socialscienceregistry.org/trials/1246/history/51045

Sponsors & Partners

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)

We have evaluated two interventions:
1. The upgrade of HIs to HIM: an important component of the Colombian government’s Early Childhood policy. The intervention involves:
(i) Hiring a team of professionals, which includes one health/nutrition professional per 200 children, one expert in socio-emotional development per every 200 children and one pedagogical assistant per every 50 children to aid with activities in the classroom. The cost of this component is approximately US$20 per child per month.
(ii) Delivery of a one-time pedagogical endowment for a value of US$52 per child for toys, books and other materials.
2. The upgrade of HIs to HIM+FE: a private foundation Fundación Éxito (FE) provided additional modules in some centres. In addition to the above the following was added:
(i) Nutritional improvement programme which increases calorie provision by 15% over what is normally provided.
(ii) Training programme for HI teachers which was planned, coordinated and offered by professionals from the Colombian National University.
(iii) Reading programme for children, teachers and parents.

Intervention Start Date

2013-06-03

Intervention End Date

2014-11-30

Primary Outcomes

Primary Outcomes (end points)

1. Children’s nutritional status
2. Children’s cognitive and language development
3. Children’s socio-emotional development and executive functions
4. Parental practices and quality of care provided in HIs

Primary Outcomes (explanation)

Outcome1:
Children’s nutritional status was assessed by collecting information on height and weight, BMI and arm circumference following World Health Organization (WHO) standards.
Outcomes 2:
Cognitive ability was measured by receptive language using the Peabody Picture Vocabulary Test in Spanish. School readiness was measured using Daberon II Screening for School Readiness. Cognitive achievement and pre-literacy skills were measured using a subset of scales from the Batería III-Woodcock-Muñoz and executive functioning by the Pencil Tapping Test.

Outcome 3:
We have used the Ages and Stages Questionnaire (ASQ) socio emotional scale at both baseline and follow up to collect information on socio-emotional development.

Outcome 4:
We will measure parental and household socio-economic characteristics as well as the quality and extent of stimulation available to the child in the home environment. We will apply Family Care Indicator designed by UNICEF, a version of the Home Observation for Measurement of the Environment (HOME) Inventory (Hamadani et al 2010). We will conduct a household survey involving looking at resources invested in the child such as time, toys, books etc.
Outcome 5:
We will measure the structure and quality of the nursery’s facilities, pedagogical materials, activities etc. With regards to the nursery staff, we will collect information on education, training on child development and knowledge using a reduced version of Knowledge of Infant Development Inventory (KIDI) – (MacPhee 1981, Veddovi et al 2001).

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

The evaluation design was based on a randomized control trial.
We randomly selected 198 His from 670 in our 8 study cities. These 168 HIs were then organised into groups of 3 geographically close HIs. From these groups we selected 40 for the inclusion in the study based on information of attendance by age range. In particular, we needed to guarantee that each HI in our sample had at least 15 children in our target age range of between 18 and 36 months at baseline. Then within each group of 3, we randomly assigned on HI to the control group, one HI to treatment group (HIM) and one HI to treatment group 2 (HIM+FE).
We use a three armed (HI (control), HIM and HIM+FE) design to ensure that we are separately able to estimate three effects:
1. The effect of the HIM upgrades to HIs relative to His without any upgrades.
2. The combined effect of the HIM+FE upgrades relative to His without any upgrades
3. The effect of the HIM+FE upgrades relative to only the HIM upgrade.
We collected baseline data on a random sample of 1,987 children aged 18-36 months at the start of the intervention and attending the 120 HIs in the study. The collection of baseline data allowed us to check whether the three groups were balanced in a variety of observable variables and enabled improvement of the precision of our estimated impacts by allowing us to control for pre-program outcomes and consequently reducing the intra-cluster correlation.
After 18 months, we have collected follow-up data on the same children (36-54 months by then), making sure to track any children who left the centres during this period.
The fact that the allocation to each group is random will guarantee comparability across groups and allow us to estimate impacts by comparing outcomes across groups.

Experimental Design Details

Randomization Method

First nurseries were grouped into threes on the basis of geographical proximity. Then within each triplet one nursery was randomly allocated to each of the three groups using a random number generator.

Randomization Unit

Center (HI)

Was the treatment clustered?

Yes

Experiment Characteristics

Sample size: planned number of clusters

120 centers

Sample size: planned number of observations

1,987 children

Sample size (or number of clusters) by treatment arms

40 centres control (HI), 40 centres treatment (HIM), 40 centres treatment (HIM + FE)

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

We used power calculations to determine the sample size, both in terms of the number of His and the number of children in HI, that we would require to be able to detect moderate positive effects of the two programmes on key outcomes, if they were indeed present, with a probability of 80%. For this purpose we considered effects of the programmes on height for age and cognitive development and we defined moderate positive effects as 20% of one standard deviation for HIM as compared to HI and 20% of one standard deviation for HIM+FE as compared to HIM. Because all of the children in our sample are treated by one modality or the other, i.e. they received some sort of early childhood education, the scope for improvement is lower than if we had an untreated control group. For the power calculations we assumed a moderate intra-cluster correlation co-efficient which was similar to what we observed in the baseline data for cognitive development (0.035 – as measured by ASQ). Given our assumptions, we calculated that we would require 15 children per cluster (HI) to achieve 80% power at 5% significance level. To allow for attrition between baseline and follow-up data collection (of about 10%), we assessed 17 (whenever there were enough) children per cluster. This led to a target sample size for baseline of 2,000 in 120 clusters (since some His only had 15 or 16 children in the target age range).

Supporting Documents and Materials

IRB

INSTITUTIONAL REVIEW BOARDS (IRBs)

IRB Name

Universidad de Los Andes Research Ethics Committee

IRB Approval Date

2013-01-31

IRB Approval Number

Act 204, 2013

IRB Name

UCL Research Ethics Committee

IRB Approval Date