SPARC has six years of experience carrying out research and collecting meaningful data in some of the world’s most fragile places.
Fragile and conflict-affected settings (FCAS) are by definition unpredictable, dynamic, risky and usually dangerous. Much SPARC-funded research took place in such contexts, in Mali, Nigeria, Somalia, Sudan and Yemen. So what have we learned from these experiences that can inform how future FCAS research studies are designed and implemented, and how conventional research approaches must be adapted?
While it is important to have clarity about the overall objectives of the research and the proposed approach at the outset, a predefined, rigid and linear project design can quickly become unworkable, even irrelevant, in dynamic and volatile FCAS. Instead, iterative planning and adaptive management are essential to remain relevant and responsive.
In Darfur, Sudan, this meant letting the context determine the focus, frequency and type of outputs. At times we shifted from planned market briefs to urgent updates on the escalating crisis in El Fasher, where the humanitarian response was severely lacking. These pivots were driven by local researchers, drawing on their experience and the insights of key informants in and around the city.
SPARC’s programme managers have reflected on the importance of agile contracts that allow for such flexibility, noting that milestone-based contracts worked better than fixed timelines. As described above, adaptability in what constituted a milestone so that it was relevant to the context was as important as a flexible timeline.
At the proposal stage researchers are usually asked to describe their data collection methods in some detail. This can even include examples of questionnaires and checklists for semi-structured interviews. In FCAS this is a misplaced and possibly counter-productive investment. There is a strong likelihood that conventional data collection methods are inappropriate. For example, in contexts where there are high levels of suspicion and distrust, a local researcher following a questionnaire on a tablet or a piece of paper will attract attention and may be at risk.
Instead, researchers need the space to develop and adapt their data collection methods iteratively, working closely with local researchers who are best placed to judge what is feasible, safe and insightful. This often means adjusting approaches as conflict dynamics shift: for example, changing the geographic focus, and collecting information remotely if local researchers relocate for security reasons or are themselves forcibly displaced. In one SPARC project, new locations were added to better understand the impact of conflict.
The safety and security of local researchers in FCAS is paramount, and approaches must be highly context-specific. As one researcher noted, standardised security procedures imposed from outside can inadvertently increase risks. In one case, seeking official permission or information from actors involved in the conflict exposed researchers’ planned movements and made them more vulnerable. Such experiences highlight the need for flexible, locally informed security practices rather than blueprinted protocols.
The aid system tends to favour quantitative data, treating numbers as more ‘scientific’ and credible. This bias appears to have deepened despite the difficulties of collecting reliable quantitative data in protracted crises which are mostly in FCAS. In such contexts, quantitative surveys often break down: sampling frames cannot be followed, distrust leads to superficial or inaccurate answers, and poor communications make data transmission difficult or impossible. Despite their appearance of precision, these numbers may obscure more than they reveal.
Many SPARC researchers have therefore relied more on qualitative data. This offers flexible, sensitive and semi-structured conversations that allow unexpected issues to surface; it is better suited to understanding causality in fragile contexts; and it captures local knowledge and lived experience that quantitative surveys cannot reflect when normal patterns of life are disrupted by conflict. It is also easier to transmit - often orally or via WhatsApp - where internet access is limited.
In all of this local researchers have a crucial role to play, from identifying the research questions, to designing and adapting the data collection methods, to collecting information and making sense of it. In some of the most extreme FCAS contexts, SPARC projects have found it necessary to recruit experienced researchers who can self-manage and carry out their own preliminary analysis, particularly where it is not possible for the whole research team to gather in person, either at the design or final analysis stages of the project.
Valuing this local expertise requires an ethical and practical commitment, whether through named authorship or, in contexts like Sudan, ensuring anonymity to protect researchers’ safety.
In a conflict environment characterised by deep suspicion and fear, the importance of fostering relationships of trust when carrying out research cannot be overemphasised. This includes trust between in-country and external researchers, among local researchers themselves, and between researchers and community members. It is easiest to build trust on long-standing relationships - as in Darfur, where some team members had worked together for decades - but where this is not the case, deliberate effort is needed. Clear and sensitive communication, both within teams and across locations, helps build mutual understanding, prevents miscommunication and rumours, and is vital for researchers’ safety in volatile settings.
This piece was based on inputs from a number of SPARC researchers at the SPARC Annual Learning Forum in September 2025, and was coordinated by Margie Buchanan-Smith.
This blog forms part of a wider series of SPARC's learnings about how to conduct research more effectively in fragile contexts. Please see our accompanying blogs on the importance of effective procurement, building strong relationships and trust, and in-programme learning processes in fragile and conflict-affected settings.