ConflictMeter Methodology
How ConflictMeter works
ConflictMeter is an independent operationalisation of Professor David Betz’s published civil-conflict theories. It measures how closely current conditions in a country align with the structural, mobilisation and accelerant conditions that Betz argues make large-scale civil unrest and internal political conflict more likely. The production model is frozen as Betz v13.15 and is designed to be inspected rather than treated as a black box.
A higher ConflictMeter score means stronger alignment with those conditions. It is not a percentage probability, a prediction that conflict will occur, or a forecast of when it might occur.
ConflictMeter is not produced or endorsed by Professor Betz. See Betz, “Civil War Comes to the West” (2023) and Part II: Strategic Realities (2025).
The model in one minute
The headline score is built in three stages. First, three structural preconditions are combined into Pressure. Second, Pressure is combined with Activation using symmetric convergence. Finally, economic and narrative conditions can add a bounded accelerant effect.
1. Structural Pressure
Geometric convergence of Downgrading / Status Reversal, Factionalism and Loss of Faith in Normal Politics.
2. Activation
Current general mass mobilisation plus positive-only change over roughly ten years. It measures mobilisation intensity/capacity, not political violence itself.
3. Accelerants
Economic stagnation and narrative fragmentation can raise the Base score modestly. Each has a frozen 10% headroom coefficient.
Frozen production formula
v13.15 makes no methodological change to the v13.14 production candidate; it is the final QA/reproducibility freeze.
Missing inputs remain NULL. ConflictMeter never substitutes zero or regional averages simply to keep a country in a ranking.
The three structural pillars
P1 — Downgrading / Status Reversal
Global: demographic-change velocity as a globally scalable proxy.
Enhanced: recent comparable subjective nativism/status evidence is added where available.
Actual political-power transitions remain context rather than a substitute for perceived status reversal.
P2 — Factionalism
Global: V-Dem political polarisation current level plus long-run change.
Enhanced: direct 2025 immigrant/native-born and inter-ethnic tension evidence where available.
P3 — Loss of Faith in Normal Politics
Global: Democracy Perception Index evidence for perceived democratic/institutional performance.
Enhanced: Ipsos Society is Broken Index 2025 where available.
What the accelerants do — and do not do
Economic pressure
World Bank WDI growth shortfalls are zero-preserving: no adverse shortfall means zero pressure. GDP-per-capita and household-consumption-per-capita shortfalls are combined arithmetically.
Inflation remains context-only rather than a scored accelerant.
Narrative fragmentation
V-Dem online-media fractionalisation provides the scored narrative-fragmentation channel using current adverse level and positive long-run deterioration.
Party hate speech remains context-only because it overlaps more strongly with Factionalism.
Containment
The universal headline is the Gross index. Democratic containment is already represented through legitimacy; there is no generic democratic safety brake.
Autocratic Net25 is a secondary sensitivity view, not the primary score.
Global and Enhanced are separate products
The Global model uses globally scalable inputs and is the basis for worldwide comparison. Enhanced uses additional recent direct-survey evidence where comparable coverage exists. They are never mixed into one ranking universe.
Global
96 countries currently have Pressure + Activation coverage.
81 have FULL current Global coverage including both accelerants.
Enhanced
30 countries have Enhanced Core + Activation coverage.
30 have FULL Enhanced coverage.
Coverage is information
A country without FULL coverage can still have legitimate partial current evidence or historical component history. ConflictMeter labels that state explicitly rather than manufacturing completeness.
Historical methodology
ConflictMeter separates what could have been known at the time from what today’s source databases now say about the same historical period.
Vintage
Uses the source release that was available by the historical cutoff. This is the public movement baseline from 2025 onward.
Reconstructed
Uses later revised source databases for the same historical period. It is kept separate so later source revisions are not mistaken for real historical change.
No synthetic backfill
Pre-2025 history remains component-only where full direct comparability is not authorised. ConflictMeter does not infer a retrospective headline score simply because most components exist.
2025 Vintage is the first authorised public headline baseline for forward score/rank movement. No pre-2025 headline history is published.
Source diagnostics are not risk components
Current Revision Sensitivity
Measures how much later source revisions changed the latest comparable historical observation. Public country pages show the classification; paid products can show the driver and magnitude.
Historical Revision Exposure
Summarises revision exposure across 2021–2025: maximum change, affected families and affected years.
Why expose this?
Many public datasets revise history. ConflictMeter distinguishes source revision from substantive historical movement rather than silently rewriting the record.
These are source diagnostics. They are not ConflictMeter components, confidence scores or measurement-uncertainty scores.
Versioning and change control
The active production methodology is Betz v13.15, frozen on 28 September 2026. Any future change to pillar definitions, source placement, Activation, interaction form, accelerant coefficients, containment treatment or missingness requires a new version and an explicit audit.