The Effect of Improving Child Care Centers on Early Childhood Development in Bolivia

Intervention:

Public child care centers in Bolivia provide daycare services for children under 4 years of age, whereas preschool-age children (4-5 years) are required to attend formal education facilities. The program provides major physical upgrading and improvements of content and staffing capabilities for existing facilities. It will also strengthen the existing national norms and protocols providing standardized guidelines for the work in child care centers in all intervention centers. Specifically, the intervention will finance and support:

• Facility infrastructure improvements through the upgrading of existing selected centers in line with improved guidelines (refurbishments of buildings, redesigning of space use, improvement of safety conditions, etc.).
• Provision or replacement of equipment at the child care center, including appliances, cookware, furniture, and materials.
• Hiring of new human resources (1 senior caregiver/tutor per center and 1 nutrition and hygiene specialist for every two centers)
• Training of new and existing personnel in line with improved guidelines.
• Elaboration, complementation and reproduction of guidelines, norms and quality standards.

The intervention will be implemented in a total of 79 eligible child care centers. All centers will receive the soft component of the intervention (additional personnel and training), 2 centers will receive, in addition, the equipment component, and 48 centers will receive the complete package of the intervention (infrastructure improvements, equipment, additional personnel and training).

2017-07-03

2017-12-31

The purpose of this research is to assess the impact of the child care center intervention on a variety of service quality outcomes. Our measurement of quality will consider structural factors, such as infrastructure, equipment, availability of trained personnel, protocols, etc, as well as processes indicators that focus on the child-caregiver interactions.

Key Short-term Outcomes
• Availability and adequate physical infrastructure, use of space, furniture, and equipment (Infant Toddler Environment Rating Scale – Revised)
• Quantity and quality of child stimulation and availability of child support in the center’s environment (caregiving routines, activities and interactions) (Caregiver Interaction Scale).
• Caregiver’s knowledge about child development processes and child care practices (Knowledge of Infant Development Inventory).

The study will not measure final child development outcomes in the short run, and will rather focus on quality measures of service provision. The application of ECD standardized tests is considered for subsequent follow up surveys (conditional on the availability of funding) including ECD indicators on cognitive, language, gross and fine motor development, and socio-emotional skills with the ASQ-3 or other similar instruments.

Baseline Survey:

For experimental validity and balance purposes, between February and May of 2014, we collected data from a baseline survey in a sample of 100 child care centers (50 treatments and 50 controls). We surveyed infrastructure, staff and functioning conditions through the application of questionnaires at the center level, including the application of three standardized quality tests: ITERS-R, CIS and COFAS. Within each center, we randomly selected a class/cohort and then applied ECD Standardized Tests to a randomized selection of 10 children. The total sample at baseline includes information on 100 child care centers and 1,000 children under 4 years old. The baseline data analysis provided a comprehensive overview of the situation of pre-program ECD and child care centers quality indicators in the target sample. It also showed no significant differences in observable characteristics between treatment and control groups (Bedregal et.al. 2016).

Follow-up Survey:

The first follow up survey will take place between September and October of 2017 and will collect information from all child care centers where treatment was offered, as well as from an equivalent number of control centers in rural and urban areas of the departments of Chuquisaca and Potosi. Treatment-control pairs will be selected within each municipality; therefore, the evaluation sample will exclude 8 municipalities that have only one child care center. In total, the evaluation sample will contain 158 child centers.

The survey will focus on measuring quality of services provided at the center, including infrastructure conditions, adequate use of space, caregiving routines, defined activities, processes and interactions with children. Specifically, the survey will collect data on:

- Personal care routines, furnishings, language, fine- and gross-motor activities, creative activities, social development, and adult needs (Infant/Toddler Environment Rating Scale ITERS-R).
- Caregiver interactions to evaluate sensitivity, harshness, detachment and permissiveness of caregivers (Care Giver Interaction Scale CIS).
- Knowledge about child care practices and development processes (Knowledge of Infant Development Inventory).
- Infrastructure indicators, including availability and use of space, furniture and equipment.
- Human resources and child care center management indicators.

Contingent on the availability of resources, a subsequent follow-up survey will be implemented in treatment and control communities with the objective of measuring early childhood development outcomes using a sample of resident households with preschool aged children.

Random assignment to treatment and control groups:

The impact evaluation will be based on the design of a Randomized Controlled Trial (RCT). Within each program municipality, eligible child care centers were randomly assigned an order of entrance to the program and listed accordingly. Initially, based on program resources available, 83 child care centers on top of the list were identified for the treatment group and an equivalent number of centers from the bottom of the list were assigned to the control group. In municipalities where there was only one child center, it was directly assigned to treatment; those municipalities, however, will not be part of the experimental sample. Randomization was carried out through public lotteries witnessed by local and UDIT authorities to ensure transparency and legitimacy.


Treatment was assigned at the center level and stratified by municipality.

Planned number of observations: 158 child care centers
- Treatment: 79 centers
- Control: 79 centers

The long run follow-up will include a population based sample of children in the area of influence (community or neighborhood) of treatment and control centers. The sample will include 12 children per area for a total sample of 1900 children.

Public Lottery

Child Care Center

Yes

158 child care centers

158 child care centers and 1900 households with children 3 to 5 years old

Treatment group: 79 centers and 950 children
Control group: 79 centers and 950 children

The minimum detectable effect size, with power =0.8, significance =0.05, intra-cluster correlation =0.012 (ESNUT2012), is 0.14 standard deviations on the aggregate measure of child development.