Information and collective action in Angolan schools: inside the black box of community-driven development

How to cite

Leeffers, Stefan; Maro, Vincenzo Di; Serra, Danila; Vicente, Pedro. 2019. "Information and collective action in Angolan schools: inside the black box of community-driven development." AEA RCT Registry. AEARCTR-0003930. https://doi.org/10.1257/rct.3930-2.4000000000000004

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

Stefan Leeffers* , Vincenzo Di Maro , Danila Serra , Pedro Vicente

* Primary investigator

Registry ID
AEARCTR-0003930
DOI
https://doi.org/10.1257/rct.3930-2.4000000000000004
First registered
Last updated
Status
Completed
Trial dates
2014-10-01 – 2019-03-29
Countries
Angola (Kwanza Sul)
Registry record
http://www.socialscienceregistry.org/trials/3930

Abstract

Community-driven development interventions blend multiple components difficult to identify separately. The literature reports mixed impacts of these interventions. We designed and conducted a field experiment in 126 schools of the Angolan province of Kwanza Sul, aiming to distinguish the impact of the information and the collective action components of a scorecard system. Specifically, we explore a design with three treatments: (i) decentralized information to parents on the performance of schools, (ii) organization of meetings with parents without pre-set information on performance, and (iii) centralized information on performance through meetings with parents. Our measurement of outcomes includes: parent, school, teacher, and school principal surveys, as well as lab games and test scores. In the games, apart from standard lab games (trust, public-good), we develop a novel dictator game where teachers individually decide about sharing their endowments with multiple parents.

Experimental design

We designed and conducted a field experiment in 126 schools of the Angolan province of Kwanza Sul, aiming to distinguish the impact of the information and the collective action components of a scorecard system. Specifically, we explore a design with three treatments: (i) decentralized information to parents on the performance of schools, (ii) organization of meetings with parents without pre-set information on performance, and (iii) centralized information on performance through meetings with parents.

Design details

Randomization method
The 126 schools were randomly allocated by a computer across the four treatment groups using blocks. The schools were organized in blocks employing village and a set of school characteristics, which included the number of classrooms, number students, number of teachers, and whether there are elections for a parent committee.
Randomization unit
School
Clusters
126 Schools
Observations
30,000 students, 2,000 parents, 2,000 teachers
Arms
32 schools control, 32 schools information treatment, 31 schools meetings treatment, 31 schools information and meetings treatment
Primary outcomes
Our measurement of outcomes includes: parent, school, teacher, and school principal surveys, standardized student tests, as well as lab games and test scores. In the games, apart from standard lab games (trust, public-good), we develop a novel dictator game where teachers individually decide about sharing their endowments with multiple parents. Important outcome variables from the parent survey are self-reported involvement at the school in terms of time and money invested and the relationship with the teacher, satisfaction with the school, teachers, management and behavior of other parents, as well as collective action indicators. Teacher-related outcome variables from both the teacher and director surveys are self-reported performance outcomes (i.e., attendance), satisfaction with school, other teachers, management and behavior of parents, collective action indicators, as well as questions regarding the parental involvement of the parents of the 40 selected students answered by the relevant teacher. Other outcome variables related specifically to the director survey and include information about tasks and responsibilities, as well as the quality and quantity of relations with other institutions that support the school. For the school level surveys, important outcome variables are those related to investments in the school facilities and resources (by parents), administrative data on student, teacher and school board performance (i.e., number of students that passed grade 6, attendance of teachers and school directors etc.), as well as parent involvement (i.e., existence of a parent committee).

Cited by

No citation data available for this registration yet.

Keywords