Better Vision for Better Learning?: Evidence from a Randomized Controlled Trial in Vietnam

How to cite

Lee, Jongwook; Glewwe, Paul; Yoon, Sangchul. 2020. "Better Vision for Better Learning?: Evidence from a Randomized Controlled Trial in Vietnam." AEA RCT Registry. AEARCTR-0006533. https://doi.org/10.1257/rct.6533-1.0

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

Jongwook Lee* , Paul Glewwe , Sangchul Yoon

* Primary investigator

Registry ID
AEARCTR-0006533
DOI
https://doi.org/10.1257/rct.6533-1.0
First registered
Last updated
Status
Completed
Trial dates
2016-10-14 – 2017-12-06
Countries
Viet Nam (North Central)
Registry record
http://www.socialscienceregistry.org/trials/6533

Abstract

Vietnam is a country where eye health care infrastructure is inferior to its economic growth and income level. The eye health care for children is especially limited in Vietnam: around one-fifth of the children under 15 years old are suffering from uncorrected refractive errors. Uncorrected refractive errors at school age are not only a vision problem, but also may affect school achievement that ultimately may have a significant impact on labor market outcomes and poverty. This paper presents the effect of vision screening and free eyeglasses on children's academic performance at school in Thanh Hoa province, Vietnam. We use a school-level cluster randomized controlled trial to evaluate providing vision screening and free eyeglasses to primary school students in Grades 4 and 5. We find that approximately 10.7 percent and 9.6 percent of grade 4 and 5 students suffer from vision impairment and uncorrected refractive errors in Thanh Hoa province. Furthermore, while 27 percent of students with refractive errors already had eyeglasses, only 29 percent of them had appropriate eyeglasses. We also find that providing vision screening and free eyeglasses successfully increases children's test scores. The magnitude of the impact ranges from 0.16 to 0.22 standard deviations for reading test scores, 0.16 to 0.23 standard deviations for mathematics test scores, and 0.14 to 0.26 standard deviations for total test scores after one year of providing eyeglasses. This paper also provides heterogenous effect of enhancing low vision on students' test score by gender and vision impairment level.

Intervention

All grade 4 and 5 students in 140 schools in Vietnam receive free school-based vision screening. Vision-impaired students, who need eyeglasses, in the treatment schools receive a pair of free eyeglasses immediately while those in the control schools receive free eyeglasses after one year.

Experimental design

140 primary schools in Thanh How province in Vietnam are randomly selected proportionally based on the number of schools in each district. The 140 schools are paired based on the school size in each district. One school in each pair is randomly assigned to the treatment group, and the other school in each pair is assigned to the control group. Thus, 70 treatment schools and 70 control schools are randomly selected, stratified by district and school size.

A two-step school-based vision screening is implemented for all students in 140 primary schools. Vision-impaired students, who need eyeglasses, in grades 4 and 5 in 70 treatment schools receive a pair of free eyeglasses after the vision screening, and those students in 70 control schools receive a pair of free eyeglasses after one year of the screening. Baseline reading and mathematics tests are conducted with the school-based vision screening, and endline tests are administered one year later.

This study estimates the effect of providing eyeglasses using intention-to-treat (ITT) analysis by comparing the mean outcomes of students in the treatment schools with the mean outcomes of those in the control schools.

Design details

Randomization method
Randomization done in office by a computer from a list of all primary schools in Thanh Hoa province, Vietnam. First, 140 schools were randomly selected. Second, schools were paired based on the school size within each district. Third, one treatment school and one control school were randomly selected from each pair.
Randomization unit
School-level clustered randomization is used.
Clusters
140 schools
Observations
4,200 students
Arms
70 treatment schools and 70 control schools
Primary outcomes
Reading (Vietnamese) and mathematics test scores

Cited by

No citation data available for this registration yet.

Keywords