Measurement guide
How is democracy measured?
Researchers translate a contested political idea into observable questions, coded evidence, statistical estimates, and composite indices. The result can reveal patterns—but only if definitions, sources, uncertainty, and value choices stay visible.
STEP 1 / DEFINE
Start by saying what democracy means.
No dataset measures “democracy” without first deciding what democracy includes. A narrow measure may focus on competitive elections. A broader measure may add basic freedoms, checks on government, participation, public reasoning, equality, welfare, or local power.
Good measurement makes that boundary explicit. Democracys.org uses five published measures and clearly states its own choice to give each one equal weight.
STEP 2 / OBSERVE
Combine facts with structured judgment.
Observable facts
Examples include whether elections occurred, who could vote, turnout, constitutional rules, or whether an office is elected.
Expert-coded conditions
Questions such as media freedom, election integrity, or judicial independence often require contextual assessment by multiple country experts.
Standards-based coding
Researchers may evaluate reports, laws, decisions, and documented events against consistent criteria.
Each approach has strengths and error sources. Formal rules may differ from practice; raw outcomes may lack context; and expert judgment requires careful sampling, calibration, and statistical modeling.
STEP 3 / ESTIMATE
Turn disagreement into an estimate with uncertainty.
V-Dem uses a Bayesian measurement model to combine ratings from thousands of country experts while accounting for differences in reliability and scale use. Its modeled estimates are accompanied by uncertainty information. That does not eliminate judgment; it makes the aggregation systematic and the uncertainty visible.
A precise-looking decimal is therefore not a direct physical measurement. It is an estimate produced from defined questions, evidence, coders, assumptions, and a model.
STEP 4 / COMBINE
Every combined score makes choices.
Researchers must decide which measures belong together, how much each contributes, and whether strength in one area can compensate for weakness in another. A familiar average allows full compensation; a weak-link average makes very low results matter more.
Democracys.org blends those two approaches. Equal weights are not handed down by the source dataset; they are a visible beta choice open to criticism.
See how the score worksSTEP 5 / SHOW UNCERTAINTY
A rank is not a finish-line photograph.
Adjacent countries may have overlapping uncertainty. Historical estimates can also be revised when evidence, coders, coverage, or models change. V-Dem cautions against comparing absolute scores across different dataset versions, so Democracys.org keeps every chart within v16 and will recalculate all years when adopting a later release.
The shaded area on each chart is an uncertainty guide built from the source’s lower and upper estimates. It is not a guarantee that the final score falls inside that area.
READING CHECKLIST
Six questions to ask of any democracy index
- What definition?Which democratic principles are inside and outside the measure?
- What evidence?Who recorded it, from what sources, and through what questions?
- What aggregation?Which weights and rules turn inputs into a headline score?
- What uncertainty?Are small gaps meaningful or within a plausible range?
- What version?Were all countries and years calculated from one consistent release?
- What claim?Does the chart describe a pattern, or overreach into an unsupported causal story?
PRIMARY REFERENCES