Reflections on Creating New International Aid Network Datasets
A Q&A with RIPIL team members building three original international aid datasets through the multi-year NSF grant project Networks of Influence and Support between War and Peace
November 7, 2023 | Interview conducted by Madison Shomaker and written with support from Christina Harris and the project team
The Research on International Policy Implementation Lab (RIPIL) employs dozens of students and collaborators from across the globe to research pressing issues in international peace, security, development, and humanitarian response. For this piece, I met with RIPIL research managers and PhD candidates Alex Bruens, Ariana Schmidt, and Hatem Zayed* to discuss their experiences in data collection, coding, and lab management for one of RIPIL’s most ambitious projects: Networks of Influence and Support between War and Peace.
Made possible by a multi-year grant from the National Science Foundation, “Networks of Influence” utilizes new and open-source network data in 20 countries where there were active United Nations peacekeeping or special political missions (SPMs) between 2005 and 2021 to create three unique datasets. The datasets aim to illustrate the vast networks of influence among international and domestic aid actors. Over the past two years, RIPIL has employed more than 20 BA, MA, and PhD student researchers to assist Professors Susanna Campbell (PI) of American University and Jessica Maves Braithwaite (Co-PI) of the University of Arizona on the project’s execution.
The three datasets underway focus on areas often overlooked in the international aid sector: (1) formalized coordination efforts of peacebuilding, humanitarian, and development actors, (2) relationships among peacebuilding, humanitarian, and development actors beyond donor-recipient ties, and (3) the presence and role of domestic non-governmental organizations (NGOs) and civil society organizations in these networks as well as less traditional peacebuilding actors like private firms. These datasets will help researchers better understand how dynamic characteristics of aid networks, like density or centrality, affect peace and conflict indicators.
This project is based on the assertion that the aid community is more than the sum of its parts; it is a product of interactions and contracts among aid actors with tangible implications for those in conflict-affected contexts. The project is now in the final stages of data collection, coding, and preparation of scholarly and policy papers based on the findings.
Below, Alex, Ariana, and Hatem share what makes this project, the data collection process, and these new datasets truly groundbreaking.
Can you tell me about these new datasets and what they add to the current peacebuilding literature?
Alex: There are three main datasets. One of the datasets is project-level, focused on contractual relationships between aid donors and their partner organizations. This speaks very much to the literature on aid, generally, because it’s what has been most focused on: how do donors give aid to countries and how do partner organizations implement aid projects?
We’re also collecting data on coordination among aid actors which is entirely new. It speaks to how these actors work together outside of their contractual relationships. This has not been examined in the aid, peacekeeping, or peacebuilding literature. This is a major contribution, and we’re collecting these data in a few different ways. These data allow us to answer questions about how aid actors—whether bilateral aid donors, intergovernmental organizations (IGOs), host governments, international non-governmental organizations (INGOs), or civil society actors— work together outside of the formal contractual relationships that they sign to deliver and distribute aid. When do representatives of these organizations attend coordination meetings? What is the sectoral focus of these meetings? When do they talk with one another and coordinate aid projects? When do they coordinate in the same sectors? In different sectors?
Our third dataset captures information about all of the related humanitarian, development, and peacebuilding actors active in our sample of countries. We have compiled a massive list of organizations active in all of the 20 countries in our sample – about 20,000 rows capturing over 10,000 unique organizations cross-nationally and over time. There just isn’t a lot of data out there on this range of organizations. For example, national NGOs, host government ministries, and civil society organizations are often completely missing in existing datasets on international organizations or aid actors. When datasets do explore this range of actors, they look at each type of actor in isolation of one another in literature on aid. The datasets we are developing put them all in conversation, more closely mirroring the real aid world.
What sets the Networks of Influence project and RIPIL’s approach to it apart?
Ariana: Even after working in humanitarian aid [in Iraq, Greece, Moldova, and Doha], I am still amazed by the diversity of organizations we capture in these datasets. I am excited for when we can present the final products to a larger audience as even those who know the aid world well will likely be surprised by the many players involved. We go beyond tracing financial links to capturing how diverse organizations coordinate in complex environments.
Alex: We’ve presented this research at many conferences over the last year and have been engaging with other scholars on it throughout the process. I even just got a text from a grad student saying, “Can you tell me more about your project? I want to cite it in my prospectus because it’s at the forefront of the literature.”
Hatem: I think the project is doing something really special. From the beginning, we’ve been trying to essentially see what others are doing with data on peacebuilding, development, and humanitarian actors in conflict settings. Time and again, we’ve discovered that we’re contributing something that’s very unique. There’s very little research out there that brings together data on this diverse range of aid actors and cleans it in a way that can be easily used for research.
The theoretical argument that we are making is also something that we believe responds to a very clear gap in the literature. Very rarely do people talk about the role of different aid actors in conflict settings, how they work together, and how this affects peace and conflict outcomes. This is especially true when it comes to less visible actors, such as national NGOs and civil society actors. I think some international actors, in particular, are typically more visible just because the literature focuses on them and, perhaps, because the data on them is easier to gather. We’ve been looking at actors that may not be center stage in the literature even though they might be playing a central role in humanitarian, development, and peacebuilding work.
I understand RIPIL has brought on many students to support the project’s data collection and coding. What has the management and data collection process been like?
Hatem: We’ve had quite a big team over the last two years. We have hired around 25 research assistants (RAs) since the start of the project. We needed so many RAs because the scope of our project required coding thousands of aid projects for hundreds, usually thousands, of organizations in each of our 20 countries over 16 years. RAs’ contributions to the project have been substantial, too. They have helped us to refine our coding rules and we even changed the scope of the project because of RA feedback in the initial stages of coding.
Before we started hiring RAs, Alex and I were doing all of the coding. We know firsthand how much doing the coding rules affects how we understand the data. We used our initial pilot coding to refine the codebook and then further refined it based on feedback from the RAs. This initial coding process enabled us to develop a robust codebook and coding scheme that benefited from the insights and contributions of the entire team.
Alex: We have three different teams of RAs, one each for our main datasets. We have weekly meetings with each of our teams. This helps us to build a community and address the specific concerns of each coder. When meeting with the RAs, we always highlight the theoretical motivation behind each of our decisions. When RAs raise issues in our weekly meetings, Ariana, Hatem, and I figure out the options to respond to them, discuss these options with the PIs, and then we make theory-driven choices from there. This has always been the process. This is how we've built out the three datasets and our extensive codebook. We focus on the overall theoretical motivation and specific concepts that we are trying to capture with the datasets and ensure that our coding rules enable us to create valid measures of these concepts.
How will Networks of Influence support policy and practice? What impact do you expect these datasets to have?
Alex: I think there are clear applications here to the United Nations (UN). Their focus over the last few years now has been on data harmonization and analysis.
Ariana: Building on what Alex said, the UN or other interested parties could use the contractual and coordination data to evaluate progress in key initiatives like localization, which aims to increase the proportion of aid going directly to domestic actors in aid-recipient countries. The network maps and characteristics go beyond tracing financial commitments. They can show where national NGOs and government actors feature prominently or peripherally within aid networks over time. By combining three distinct datasets, practitioner and academic researchers alike can ask dynamic and, possibly, uncomfortable questions about how aid communities function.
Hatem: I agree. A new trend in both aid policy and some of the literature on aid effectiveness is aid localization. Each aid actor has its own data on what it does, but I don’t think there’s one dataset that brings together what’s going on across these different types of donors. By comparing the relative distribution of aid across national NGOs, INGOs, IGOs, and bilateral donors, we could assess the degree to which localization policies are being implemented. In cases where they are implemented, we could make inferences about the effects of aid localization. Again, I think there’s an assumption that aid is not that localized – that it’s controlled by international actors. But in our datasets there are so many domestic actors who clearly play central roles in aid distribution. This may challenge the assumption that aid is not localized. These domestic actors are there in the data. They’re getting aid money. They have influence on aid projects, in one way or another. Our data enable us to understand what their effect might be. We actually have data on how national NGOs, civil society, and government ministry officials contribute to international humanitarian, development, peacebuilding, and peacekeeping outcomes.
*Since this interview was conducted, Hatem Zayed received his PhD from American University’s School of International Service.
For more on RIPIL’s work, follow us on X @ripil_au.