Improving the Quality of Preschool Education in a developing economy

How to cite

Hojman, Andrés; Martínez, Sebastián. 2019. "Improving the Quality of Preschool Education in a developing economy." AEA RCT Registry. AEARCTR-0002340. https://doi.org/10.1257/rct.2340-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Andrés Hojman* , Sebastián Martínez

* Primary investigator

Registry ID
AEARCTR-0002340
DOI
https://doi.org/10.1257/rct.2340-1.0
First registered
Last updated
Status
On going
Trial dates
2016-09-01 – 2021-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/2340

Abstract

This trial evaluates a policy package that was applied to early childhood centers in Panama. The package that all treatment centers received included teacher training sessions, educational materials, a monthly money transfer for the center, and close monitoring and help from the central authority. The package also includes important improvements in infrastructure.

For evaluation purposes, we randomized the order in which 89 child care centers administered by the government will receive the intervention. This will allow us to identify the impacts of the intervention on the centers. We are interested in how the policy affected the quality of classroom interactions, especially given that the literature shows that it is not easy to make improvements in this important outcome. Other potential outcomes of interest are the levels of attendance to the center and measures of structural quality.

Intervention

Our partner, Ministerio de Desarrollo Social, has been implementing different improvements to the centers, starting in 2017. All treatment centers received teacher training sessions. There have been several sessions, covering topics going from general management, quality standards, and teacher-child interactions. They also received a monthly money transfer for the center. These transfers were given in addition to the centers' usual revenues (parents' payments, a per-child food payment, basic educational materials, and minor repairs). The transfers could only be spent on educational materials or other education-related expenditures.

The centers also received close monitoring and help from the central authority for the implementation of 10 quality standards. The package also includes important improvements in infrastructure.

Experimental design

Our original potential sample included all child care centers managed by the government. However, MIDES decided that there were a few (originally 5) that they did not want to be part of the evaluation, because there were previous commitments to improve them. Thus, we ended up with a list of 89 centers to randomize.

Given that we also wanted to run a child-level evaluation in three arms (see the registered trial "The impact of attending a higher quality preschool in a developing economy"), we divided the sample of centers into two groups: the "couples" and the "independent". The couples were pairs of centers that were geographically close to each other. The independent ones were all the rest. We wanted one center from each couple to be part of the treatment group and the other to be part of the control group.

We used a stratified randomization schedule. We included "being in a couple" to define our strata in the randomization. The other variable used to define the strata was urban/rural. For the couples, we randomized which of the centers was to be part of the treatment group. Then, we randomized the priority for all the treated centers, giving their couples the inverse priority (so for example, the control center that was a couple of the first center to be treated received the last priority). For the independent centers, we allocated all of them to be treated after the treated members of the couples, and before the control members of the couples. Within that range, we randomized their priority.

Design details

Randomization method
Public Lottery (in the presence of all of the child care center supervisors)
Randomization unit
Child care centers
Clusters
89 centers
Observations
89 centers
Arms
44 are treated, 45 centers are controls
Primary outcomes
1. Dimensions covered by our quality questionnaire (ITERS): - Interaction (primary outcome) 2. The main score in the Caregiver Interaction Scale (CIS) 3. Attendance to the center
Secondary outcomes
1. Other dimensions covered by our quality questionnaire (ITERS): - Space and Furniture - Personal Care Routines - Talking and Speaking - Activities - Program Structure - Parents and Staff

Cited by

No citation data available for this registration yet.

Keywords