Randomized Impact Evaluation of a Graduation Program on Livelihoods in Refugee and Host Communities in Uganda
How to cite
Karlan, Dean; Goldberg, Nathanael; Parkerson, Doug; Brune, Lasse; Udry, Christopher. 2019. "Randomized Impact Evaluation of a Graduation Program on Livelihoods in Refugee and Host Communities in Uganda." AEA RCT Registry. AEARCTR-0004080. https://doi.org/10.1257/rct.4080-7.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
The Graduating to Resilience project is a US Agency for International Development (USAID) funded pilot activity, implemented in the Rwamwanja Refugee Settlement and Kamwenge District in the Western region of Uganda by AVSI Foundation, that seeks to improve food and nutrition security and self-reliance among poor households in the refugee settlement and host communities based on the ultra-poor graduation model. In treatment villages, households are randomized into one of three variations of the program: (1) the full graduation program with consumption support, an asset transfer and individual, household-level coaching; (2) the full graduation program with consumption support, an asset transfer and group coaching and (3) the graduation program with consumption support and individual, household-level coaching, but without an asset transfer component.
Intervention
The program provides 6,631 treatment households with three variations of a graduation program: (1) the full graduation program with consumption support, an asset transfer and individual, household-level coaching; (2) the full graduation program with consumption support, an asset transfer and group coaching and (3) the graduation program with consumption support and individual, household-level coaching, but without an asset transfer component.
Experimental design
The first stage of randomization assign villages into treatment (T) and control (C) villages. The second stage of randomization assigns households to one of three types of treatment arms (T1, T2, or T3) or control (C1) in the treatment villages. Households selected for interviews in the control villages are spillover controls (C2).
Participants within each village and experimental arm were randomized into groups of between 23-27 participants. Each treatment village has between four and 15 such groups, depending on the total number of households in each village.
AVSI conducted a public lottery in each village cluster to assign each participant groups to one of the three treatment arms and the control arm.
Design details
Cited by
No citation data available for this registration yet.