Strengthening Community Resilience in Conflict-affected Societies: A Randomized Controlled Trial of a CDD Intervention with a Conflict Resolution Dimension in Eastern DRC

The selected communities received a budget of up to $100,000 to finance an infrastructure project. Furthermore, the members of each CDD community received training to select and manage the project in an inclusive and participatory way. A random half of CDD communities received a third component, namely the conflict mediation component, which consisted of conflict prevention and management activities, identified and led by NGOs specialized in the matter.

2015-09-26

2020-06-30

The study focuses on two primary outcomes of interest. The first is related to infrastructure while the second one relates to social cohesion. We divide each of them in more precise outcome families to capture a more detailed picture of the intervention effect on the main outcomes of interest.
Regarding infrastructure access and quality, we measure
a) Existence/creation of socioeconomic infrastructure as well as its quantity and quality;
b) Quantity & type of infrastructure obtained on request of village or NGO (vs. government);
c) Household access to this infrastructure;
d) Use of this infrastructure by household (and frequency);
e) Satisfaction with the infrastructure;
f) Health indicators;
g) Education indicators.
This study thus does not only focus on whether infrastructure has been built, but also in how far individuals (particularly the most vulnerable) have access to this infrastructure, make effectively use of it, and whether it has led to improvements in outcomes such as health and education. At the village level we will also study the effect of the intervention on the provision of infrastructure by different actors (village member, government, NGO).
The second main outcome of interest is social cohesion. We decided to divide it in the following families:
a) Trust in another village member;
b) Community organization;
c) Ethnic/social cleaves:
d) Social cohesion;
e) Information transmission;
f) Conflict (within community members and between villages)
g) Inclusion of outsiders (e.g., IDPs; refugees; ex-fighters);
h) Participation in community meeting/collective action;
i) Civic engagement/political participation

To capture the effect of the intervention on the economic life of the participants, we will perform additional analyses on variables characterizing socioeconomic well-being. Those include measures of economic welfare, income generating activities, and subjective well-being.

The CDD program was randomly assigned to about two thirds of 400 communities with eligible project proposals. A random half of CDD communities received the conflict mitigation component.

The CDD program was implemented on a phased-in yearly basis, with a new enrollment period at the beginning of each year. Project implementation started each time upon the community’s proposal passing the quality threshold set by DRC’s Social Fund and obtaining approval from the provincial authorities (Comité Consultatif Provincial). As a result, in some months only one new project started, while in other months ten projects started. In order to keep randomization logistically feasible, all provincial headquarters of Fonds Social received a “randomization list”. New communities were added to this list as they came in (row 1, row 2, etc.). Every subsequent three rows on each of the province lists were assigned randomly to CONTROL, CDD or CDD+. In total, 400 villages were assigned across these three categories. For 35 among these 400 villages, project implementation was impossible due to security, inaccessibility and other operational challenges.
The distribution across 'primary' CONTROL, CDD or CDD+ was as follows: 127, 138, 135. (The imbalance of communities across the three experimental groups relates to the phase-in design by province and specificities of the randomization: from 2016 to 2019, 4 randomization occurred (over the 6 provinces). Moreover, the randomization was stratified by project type, to make sure we would have an even number of project type by territories.)
In addition to the primary list of randomized CDD and CDD + communities, more CDD and CDD+ projects from the list were randomly selected as replacement projects, to be used in case a primary community dropped. Following a drop-out, a replacement project would be picked in the same geographic area, and in order of the random rank allocated to it.

All provincial headquarters of Fonds Social received a “randomization list”. New communities were added to this list as they came in (row 1, row 2, etc.). Every subsequent three rows on the list were assigned randomly to CONTROL, CDD or CDD+.

Communities

Yes

400 communities

4000 households

The distribution across 'primary' CONTROL, CDD or CDD+ was as follows: 127, 138, 135 communities, respectively. (The imbalance of communities across the three experimental groups relates to the phase-in design and specificities of the randomization: from 2016 to 2019, 4 randomization occurred (over the 6 provinces). Moreover, the randomization was stratified by project type, to make sure we would have an even number of project type by territories.)

Pre-analysis plan CDD in RDC