Peer effects of After-school programs on vulnerability, violence and academic outcomes. Evidence from El Salvador.
Last registered on September 20, 2016


Trial Information
General Information
Peer effects of After-school programs on vulnerability, violence and academic outcomes. Evidence from El Salvador.
Initial registration date
September 20, 2016
Last updated
September 20, 2016 1:55 PM EDT
Primary Investigator
Pontifical Catholic University of Chile
Other Primary Investigator(s)
Additional Trial Information
On going
Start date
End date
Secondary IDs
This study aims to measure the impact of after-school clubs (a version of the After-School programs) on measures of violence and vulnerability of students in schools in a developing country, for which there is little evidence of the impact this such interventions. The premise is that clubs improve social skills of children and their ability to handle conflicts, which also allows them to improve their protection factors and their academic performance. This research is relevant to the context in which the clubs are developed: schools located in risky communities in El Salvador, a country with high levels of violence, where it is absolutely necessary to analyze the most efficient way to obtain impacts of these clubs. I also aim to find evidence of peer effects of the intervention, assigning the participants in homogeneous and heterogeneous groups according to their violence level. I want to answer whether tracking participants according to a characteristic could improve the intervention results; or if the outcomes are higher due to peer effects, obtained from heterogeneous groups.
External Link(s)
Registration Citation
Dinarte, Lelys. 2016. "Peer effects of After-school programs on vulnerability, violence and academic outcomes. Evidence from El Salvador.." AEA RCT Registry. September 20.
Experimental Details
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
The key outcomes variables are cognitive and non-cognitive skills. The cognitive skills to be measure are math, reading, science and behavior grades, school assistance and drop-out rates. The non-cognitive skills will be protective and risk factors, such as self-control, self-esteem, risk exposure, conflict management, violent behaviors and emotional regulation.
Primary Outcomes (explanation)
Some of the outcomes will be constructed using the Risk and Protective Factor Scale from Communities that Care.
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
Following the methodology applied in Lafortune, Perticar ́a and Tessada (2014) to measure peer effects on a job training for poor women in Chile, I created an experimental design that take ad- vantage of the composition of the students, but maintaining the normal clubs’ implementation. The first phase was the students’ registration to participate in the clubes. They fill out the usual registration form used by the NGO, which collects personal information from participants and from their households and family. I will complement this self reported information with administrative academic data from the schools, such as enrollment, academic results (GPA), absenteeism, and discipline registry. Using these observables, I estimate the index of vulnerability and violence per child (VV Index) predicting the likelihood to commit a violent act using the data set of explanatory variables from the sample of this study. The total enrolled children is randomly assign between three groups: control (25%) heterogeneous VV Index (25%) and homogeneous VV Index (50%). Then, I use the VV Index to rank the stu- dents: the group above the median VV Index (High VV Index, 25% of the full sample) and the group below the median VV Index (Low VV Index, the remaining 25% of the sample).
Experimental Design Details
Randomization Method
The randomization was made using the STATA software.
Randomization Unit
The randomization unit were individuals, clustered at education level (ciclo)
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
15 cluster, which are education levels, 3 levels per school.
Sample size: planned number of observations
1056 students
Sample size (or number of clusters) by treatment arms
258 students in the control group, 263 in the heterogeneous group, and 535 in the homogeneous group (267 in homogeneous high IVV and 268 in the homogeneous low IVV)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
Ética en Investigación en Ciencias Sociales y Humanidades, PUC
IRB Approval Date
IRB Approval Number
Post Trial Information
Study Withdrawal
Is the intervention completed?
Is data collection complete?
Data Publication
Data Publication
Is public data available?
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers