Designing conflict-sensitive M&E frameworks in fragile contexts
Standard results frameworks assume a stable delivery environment: fixed target populations, safe enumeration access, and indicators whose meaning does not shift with the political ground. None of these hold in conflict-affected Somalia. This note describes a practical approach to conflict-sensitive M&E design.
How standard frameworks fail
The failures are predictable. Targets set against a baseline population that has since displaced. Enumeration plans that assume access to districts that change hands. Indicators, such as attendance at government services, whose interpretation inverts when the authority providing the service changes. And perhaps most damaging, data collection that itself alters conflict dynamics: enumerator selection signalling clan alignment, or beneficiary lists that become targeting lists.
A conflict-sensitive framework does not abandon rigour. It relocates rigour to the assumptions layer, making explicit what standard frameworks leave implicit.
Design principles that survive contact
First, indicator definitions carry a context annotation: what this indicator means under each plausible authority and access scenario, agreed at design stage rather than argued at reporting stage. Second, every outcome indicator has a designated fallback measurement method for inaccessible areas, third-party monitoring, remote sensing, or phone-based panels, with the precision downgrade documented honestly rather than hidden. Third, the results framework is versioned: a formal revision protocol with the donor, triggered by defined context thresholds, replaces the fiction that targets set at design remain meaningful throughout.
Fourth, the M&E system itself is conflict-analysed. Enumerator recruitment, data storage, and list management are assessed for the harms they could cause, with the same seriousness applied to programme activities. In fragile contexts, the monitoring system is an intervention.
The political economy of honest reporting
The hardest constraint is not technical. Implementers underreport access failures because access is read as performance; donors accept smoothed data because portfolio reporting rewards it. Breaking this equilibrium requires contractual cover: reporting requirements that explicitly request access maps, confidence ratings on each indicator, and revision histories. Where donors have adopted this, in our direct experience, data quality conversations shifted from defensive to diagnostic within two reporting cycles.
- 01Annotate every indicator with its meaning under each access scenario.
- 02Pre-agree fallback measurement methods and document the precision cost.
- 03Version the results framework with defined revision triggers.
- 04Conflict-analyse the M&E system itself; monitoring is an intervention.