A framework for assessing online community resilience under disruption

Resilience of online communities is frequently inferred from activity or network metrics, although their functional meaning changes across disturbances. We develop a phase-aware, multidimensional assessment framework linking a bounded community, a security-relevant shock, a protected function, and phase-specific evidence. The framework separates structural connectivity, participation, communication, adaptive recovery, and governance; distinguishes relative indicators from resilience outcomes; and explicitly represents uncertainty and missing evidence. Two descriptive Wikipedia applications examine contrasting shocks: governance restrictions following the English Wikipedia Palestine–Israel articles 5 arbitration case (30 deterministically selected pages, three 55-day windows) and the June 2025 Iranian Internet shutdown during the Iran–Israel war (five daily Persian Wikipedia series). In the first case, contribution volume and co-editing connectivity display divergent trajectories; bootstrap intervals reflect the sensitivity of findings to the page cohort. In the second, pageviews return closer to baseline than editing and participant-entry indicators after access restoration. An exploratory supplementary Telegram analysis assesses transferability of the phase-based descriptive workflow without equating posting or emoji frequency with resilience. These applications demonstrate descriptive feasibility, not causal effects, construct validation, or generalizability beyond the observed systems.