CivStat
Methodology & Sources
CivStat normalizes 7,000 years of human civilization data into 6 composite indexes. All indexes are standardized to a 0–100 scale, and population-adjusted relative metrics are used to enable meaningful historical comparisons.
Executive Summary
CivStat is a quantitative framework for measuring the trajectory of human civilization across 7,000 years. It tracks 76 individual indicators and scores 33 of them into 6 core dimensions: Conflict, Welfare, Knowledge, Freedom, Sustainability and Cooperation. These six dimensions combine into a single Civilization Score — a 0–100 index of the overall state of human civilization at any point in history.
Why CivStat? Existing composite indexes like the Human Development Index (HDI, 3 dimensions, since 1990), Social Progress Index (SPI, 3 pillars, since 2014), and Global Peace Index (GPI, since 2008) provide valuable but limited perspectives. They cover only recent decades, use narrow dimensional frameworks, and largely ignore environmental sustainability and international cooperation. CivStat extends measurement across millennia, broadens the dimensional scope to 6 domains with 76 indicators, integrates conflict as a core (negative) dimension following Steven Pinker's per-capita violence methodology, and normalizes all indicators by population for meaningful cross-era comparison.
Key design principles: population-normalized metrics (so ancient tribal warfare at ~15% violent deaths correctly scores higher on Conflict than WWII at ~2.9%); every value labelled Observed, Published or Estimate; fixed normalization bounds per indicator; and a 3-layer aggregation framework inspired by the OECD Handbook on Constructing Composite Indicators (2008).
7,000+
Years of data
76
Indicators
6
Dimensions
3
Aggregation layers
Explore the Methodology
6 Dimensions
Deep dive into each dimension — Conflict, Welfare, Knowledge, Freedom, Sustainability, Cooperation. Definitions, indicators, historical narratives, and country comparisons.
Read more →76 Indicators
Complete catalog of all indicators with definitions, data sources, historical ranges, and comparative examples (Korea, Norway, Sierra Leone).
Browse catalog →FAQ
15 frequently asked questions covering data quality, interpretation, country scores, academic citation (BibTeX), data licensing, and the roadmap.
View FAQ →Civilization Score Calculation
How the Civilization Score is computed — 3-layer aggregation, dimension weights, normalization, and a live worked example computed from the published data.
Jump to section ↓Our Approach
CivStat builds on the following intellectual traditions to quantitatively measure the long arc of civilization.
Cliodynamics
Peter Turchin
The mathematical study of history. Quantitatively models and predicts the rise and fall of societies.
Big History
David Christian
Interdisciplinary study of the long arc of 13.8 billion years — from the Big Bang to the present.
Composite Indicators
OECD Handbook 2008
Standard methodology for synthesizing multidimensional data into a single index (min-max normalization and weighted aggregation).
Planetary Boundaries
Rockström et al. 2009
Defines the Earth system boundaries within which human civilization can safely operate.
Core Methodological Principles
- ①Population-normalized metrics — All indicators are calculated as a proportion of the world population at the time, not as absolute figures. In particular, the Conflict Index follows Steven Pinker's methodology, which is why ancient tribal warfare (~15% violent deaths) scores higher than WWII (~2.9%).
- ②Interpolation — Years between anchor points are interpolated using an ease-in-out curve. Data gaps use trend-based estimates informed by historical context.
- ③Data tiers — Modern (1900–): based on empirical data / Early Modern (1500–1900): historical documents + archaeology / Ancient (~1500): expert estimates + archaeological evidence.
- ④Transparency — All calculation logic is open source. Raw data and algorithms are available in the GitHub repository.
Historical Data Estimation
CivStat spans 7,000 years — far beyond the reach of modern statistics. This section explains how we estimate, interpolate, and classify data confidence for pre-modern periods.
Before 1900: the estimate curve
Before 1900 the six dimension scores are not computed from indicators. They come from a CivStat estimate curve: dimension values set at anchor years from the historical literature, joined by a smooth cosine ease between anchors. The curve is then rescaled per dimension so it meets the computed 1900 value. Every pre-1900 year is labelled Estimate on the site.
Where each value comes from
Observed
Fetched from the source
Values pulled from the publisher's own data feed for the recent years it covers.
Typical era: Latest published years
Published
Published series
Values from the indicator's published series, from the first year that series covers.
Typical era: From each series' start year
Estimate
CivStat estimate
Values before a series' first published year, and every dimension score before 1900. Labels only: they do not change the weights.
Typical era: Before each series starts
Uncertainty Bands by Period
Judgement-based ranges to read the long curve with; they are not statistical confidence intervals.
0 – 500 CE
Uncertainty: Very High
±40–60%
500 – 1500 CE
Uncertainty: High
±20–40%
1500 – 1800 CE
Uncertainty: Moderate
±10–20%
1800 CE – present
Uncertainty: Low
±1–10%
Historical Data Sources
Primary databases and scholarly sources used to reconstruct pre-modern indicator values.
Maddison Project Database
↗GDP per capita, 1 CE – present
Angus Maddison's pioneering long-run economic estimates, continued by Jutta Bolt and Jan Luiten van Zanden. Provides comparable GDP per capita in 2011 international dollars.
Clio-Infra
↗Multiple indicators, 1500 – present
Historical indicators on wages, heights, education, and inequality compiled by economic historians. Core source for pre-industrial welfare estimates.
V-Dem Dataset
↗Democracy indices, 1789 – present
Varieties of Democracy project providing multidimensional democracy measurement with expert-coded historical data for 200+ countries.
Our World in Data
↗Multiple indicators, various periods
Comprehensive open-access data platform curated by Max Roser. Aggregates and harmonizes data from WHO, World Bank, UN, and academic datasets.
SESHAT: Global History Databank
↗Social complexity, 10,000 BCE – present
Systematic database recording hundreds of variables about past societies, from population size to information systems, enabling quantitative analysis of social evolution.
Ice Core Records
↗CO₂ & temperature, 800,000 BCE – present
Antarctic and Greenland ice core data (EPICA, Vostok, Law Dome) providing atmospheric CO₂ concentrations and temperature proxies. Key source for pre-industrial Sustainability Index.
Angus Maddison & Jan Luiten van Zanden
↗Economic history scholarship
Foundational scholarly work establishing methods for estimating pre-modern economic output, living standards, and population figures across civilizations.
Indicator Addition Methodology
How new indicators are incorporated into CivStat's 3-layer hierarchy.
Hierarchical Structure: L1 → L2 → L3
Layer 1 (Raw Indicators): Individual measurable data points (e.g., GDP per capita, literacy rate, battle deaths per 100k). New indicators are added at this level after source validation and coverage assessment.
Layer 2 (Composite Indexes): Thematically grouped L1 indicators are normalized to 0–100 (100 = best) and averaged within their dimension; each dimension's headline indicator counts double.
Layer 3 (Civilization Dimensions): Scored indicators roll up into the 6 civilization dimensions (Welfare, Knowledge, Freedom, Cooperation, Sustainability, Conflict).
Weighted Arithmetic Mean
At each aggregation level, indicators are combined using a weighted arithmetic mean:
where wᵢ is the weight assigned to indicator i and xᵢ is the normalized value. Weights reflect theoretical importance, data quality, and coverage completeness. Weight assignments are documented and open to community review.
Min-Max Normalization
All indicators are normalized to a 0–100 scale using min-max scaling before aggregation:
x_min and x_max are fixed bounds set per indicator (for example life expectancy 20–90 years), not the observed minimum and maximum of the series; results are clamped to 0–100. For indicators where lower is better (e.g., war deaths, CO₂), the result is reversed so that higher scores always indicate better outcomes.
3-Layer Aggregation Framework
CivStat adopts the methodology of the OECD Handbook on Constructing Composite Indicators (2008) to synthesize raw data in three stages. Each layer represents a progressively higher level of abstraction.
Layer 1
76 Raw Indicators
“What happened?”
Layer 2
33 Scored Indicators
“What counts?”
Layer 3
6 Civilization Dimensions
“How is civilization doing?”
76 Raw Indicators
Measurable raw data — 76 individual indicators including GDP per capita, war deaths, literacy rate, CO₂ emissions, and treaty counts.
Examples
Method: Data collection from OWID, UCDP, UNESCO, WHO, World Bank, V-Dem, etc.
33 Scored Indicators
Each indicator belongs to one dimension. Indicators with a clear good direction, a published source and post-1990 data are scored; the rest are shown as context.
Examples
Method: Min-max to 0–100 (log for multiplicative series), 100 = best
6 Civilization Dimensions
Converges into the final 6 civilization dimensions. Each dimension represents a core aspect of human civilization, and their weighted average forms the Civilization Score.
Examples
Method: Weighted mean of scored indicators, headline ×2; before 1900 a labelled estimate
Indicator Methodology
Civilization Score
The Civilization Score (0–100): the average of all six dimensions under the chosen lens.
DetailsCollapse
Civilization Score
The Civilization Score (0–100): the average of all six dimensions under the chosen lens.
Calculation Methodology
Balanced lens: equal-weighted mean of the six dimensions, with Conflict entering as Peace (100 − Conflict). Other lenses reweight the same six numbers.
Conflict Index
Organized and interpersonal violence (0 = peaceful, 100 = maximum violence). Per-capita rates, not absolute counts, following Steven Pinker's approach.
DetailsCollapse
Conflict Index
Organized and interpersonal violence (0 = peaceful, 100 = maximum violence). Per-capita rates, not absolute counts, following Steven Pinker's approach.
Calculation Methodology
Each scored indicator is put on a 0–100 peace scale against fixed bounds, averaged (the representative indicator counts twice), and the result is reported as violence (100 − peace).
Scored indicators (5)
Welfare Index
Health, material living standards and basic needs, from life expectancy to child mortality and poverty.
DetailsCollapse
Welfare Index
Health, material living standards and basic needs, from life expectancy to child mortality and poverty.
Calculation Methodology
Each scored indicator is put on a 0–100 scale against fixed bounds (log scale for GDP per capita) and averaged; life expectancy, the representative indicator, counts twice.
Scored indicators (8)
Knowledge Index
Education and the capacity to create and share knowledge.
DetailsCollapse
Knowledge Index
Education and the capacity to create and share knowledge.
Calculation Methodology
Each scored indicator is put on a 0–100 scale against fixed bounds and averaged; years of schooling, the representative indicator, counts twice. Publications, patents and AI adoption are shown as context, not scored.
Scored indicators (5)
- Years of Schooling · representative ×2
- Education
- Tertiary Education
- Internet
- R&D Spending
Freedom Index
Political rights, civil liberties, press freedom and equal participation (0–100).
DetailsCollapse
Freedom Index
Political rights, civil liberties, press freedom and equal participation (0–100).
Calculation Methodology
Each scored indicator is put on a 0–100 scale against fixed bounds and averaged; V-Dem Liberal Democracy, the representative indicator, counts twice. The EIU Democracy Index overlaps V-Dem and is shown as context only.
Scored indicators (5)
Sustainability Index
The state of the Earth system that civilization depends on: climate, oceans, forests, biodiversity and air.
DetailsCollapse
Sustainability Index
The state of the Earth system that civilization depends on: climate, oceans, forests, biodiversity and air.
Calculation Methodology
Each scored indicator is put on a 0–100 scale against fixed bounds (higher = safer) and averaged; atmospheric CO₂ concentration, the representative indicator, counts twice.
Scored indicators (6)
Cooperation Index
How connected and institutionally linked the world is: trade, international organizations and diplomacy.
DetailsCollapse
Cooperation Index
How connected and institutionally linked the world is: trade, international organizations and diplomacy.
Calculation Methodology
Each scored indicator is put on a 0–100 scale against fixed bounds and averaged; trade volume, the representative indicator, counts twice. Treaty counts and aid flows are shown as context, not scored.
Scored indicators (4)
- Trade Volume · representative ×2
- Institutions
- Diplomatic Relations
- Globalization Index
Key References
The Better Angels of Our Nature: Why Violence Has Declined
Pinker, Steven · 2011
Core reference for population-normalized violence methodology. Theoretical foundation for the Conflict Index design.
Our World in Data
Roser, Max et al. · 2023
Primary data source for Welfare, Knowledge, and Freedom indexes.
Varieties of Democracy Dataset v14
V-Dem Institute · 2024
Historical time-series data source for democracy and freedom indexes.
Armed Conflict Dataset
UCDP/PRIO · 2024
Primary source for modern conflict fatality data.
Atlas of World Population History
McEvedy & Jones · 1978
Source for ancient-to-early-modern population estimates. Used to compute the Conflict Index denominator.
Maddison Project Database
Maddison Project · 2023
Long-run economic data and living standard estimates.
Academic References
Key academic papers and books providing the scholarly foundation for each CivStat dimension.
⚔️ Conflict & Peace
Pinker, S. (2011). The Better Angels of Our Nature. Viking.
Conflict dimension: population-normalized violence rates
“Violence has declined over millennia due to Leviathan, commerce, feminization, and reason.”Source ↗
Gleditsch, N.P. et al. (2002). Armed Conflict 1946–2001. Journal of Peace Research.
Conflict Index data source: UCDP/PRIO dataset
“Systematic measurement of armed conflict enables rigorous study of war and peace.”Source ↗
💰 Economic Development
Bolt, J. & van Zanden, J.L. (2020). Maddison Project Database 2020. Journal of Economic History.
Welfare dimension: GDP per capita from 1 CE to present
“Long-run GDP estimates provide comparable economic growth data across centuries.”Source ↗
Sen, A. (1999). Development as Freedom. Oxford University Press.
Freedom and Welfare dimensions: human capabilities beyond GDP
“Development is the expansion of human freedoms, not merely economic growth.”Source ↗
📐 Composite Index Methodology
UNDP (2023). Human Development Report. United Nations Development Programme.
Precedent for combining normalised dimensions into one 0–100 index
“The HDI combines health, education and income into a single summary measure of human development.”Source ↗
Nardo, M. et al. (2008). Handbook on Constructing Composite Indicators. OECD Publishing.
CivStat's 3-Layer normalization and weighting methodology
“Min-max normalization transforms indicators to 0–100 scale for cross-indicator comparability.”Source ↗
🌍 Environmental Sustainability
Rockström, J. et al. (2009). A safe operating space for humanity. Nature, 461, 472–475.
Sustainability dimension: planetary boundaries framework
“We have identified nine planetary boundaries that define a safe operating space for humanity.”Source ↗
Steffen, W. et al. (2015). Planetary boundaries: Guiding human development. Science, 347.
Updated planetary boundaries — Sustainability thresholds
“Four of nine planetary boundaries have been crossed due to human activities (2015 assessment; seven of nine by the 2025 Planetary Health Check).”Source ↗
📜 Historical Dynamics
Turchin, P. (2003). Historical Dynamics. Princeton University Press.
Secular cycles — long swings in Cooperation and Conflict
“Societal instability follows predictable mathematical cycles driven by population and elite dynamics.”Source ↗
Tainter, J. (1988). The Collapse of Complex Societies. Cambridge University Press.
Complexity ceiling — why CivStat tracks diminishing returns
“Societies collapse when investments in complexity yield diminishing marginal returns.”Source ↗
🏛️ Democracy & Institutions
Acemoglu, D. & Robinson, J. (2012). Why Nations Fail. Crown Publishers.
Freedom and Cooperation dimensions: inclusive vs extractive institutions
“Inclusive economic and political institutions are the fundamental cause of prosperity differences.”Source ↗
Marshall, M.G. et al. (2019). Polity5 Project. Center for Systemic Peace.
Freedom dimension: democracy scoring (Polity)
“Authority characteristics of states can be measured on a spectrum from autocracy to democracy.”Source ↗
Related Projects & Data Sources
Related projects that CivStat draws data and inspiration from.
Our World in Data
↗Max Roser's comprehensive long-run data platform — primary source for many CivStat indicators
Cross-domain long-run trendsMaddison Project Database
↗Historical GDP per capita reconstruction from 1 CE to present
Economic historySeshat: Global History Databank
↗Systematic database of historical societies measuring social complexity
Civilizational complexityV-Dem Institute
↗Multidimensional democracy measurement across 200+ countries
Democracy & governanceUCDP (Uppsala Conflict Data Program)
↗World's main provider of data on organized violence since 1946
Armed conflictCLIO-Infra
↗Historical welfare and inequality indicators from 1500 to present
Historical welfareCivilization Score Calculation
The Civilization Score is the headline metric — a single 0–100 number that captures the overall state of human civilization. Here's how it's computed.
3-Layer Aggregation Process
Each scored indicator is normalized with min-max scaling against fixed bounds set for that indicator. For multiplicative quantities (GDP per capita, war death rate), a log transformation is applied first. Indicators where lower is better (e.g., conflict deaths, CO₂) are reversed so higher = better. The other raw indicators are shown as context and do not enter the score.
Each scored indicator is normalised to 0–100 and averaged into its dimension (the headline indicator counts twice). Example: Welfare = weighted average of life expectancy, child mortality, extreme poverty and the other scored welfare indicators.
The six dimensions are averaged into the Civilization Score using the formula below — an arithmetic mean, so a gain in one dimension can offset a loss in another. The Taleb lens adds a weakest-link term for readers who reject that trade-off.
How Civilization Score is computed (method v3.0)
Each dimension is scored 0–100 where higher is better (Conflict is inverted to Peace = 100 − Conflict). The default Balanced Civilization Score is the plain average of the six. Value lenses re-weight the same six numbers; the Taleb lens additionally takes half of its score from the weakest dimension. Civilization Fragility = mean(Conflict, 100 − Sustainability, 100 − Cooperation).
Worked example: World, 2026
| Dimension | Raw | Score used |
|---|---|---|
| Peace | 40.6 | 59.4 |
| Welfare | 77.4 | 77.4 |
| Knowledge | 65.7 | 65.7 |
| Freedom | 45.6 | 45.6 |
| Sustainability | 39.5 | 39.5 |
| Cooperation | 77.2 | 77.2 |
Average = (59.4 + 77.4 + 65.7 + 45.6 + 39.5 + 77.2) ÷ 6 = 60.8 → Civilization Score 60.8. Fragility = 41.3. No adjustments, no rounding fudge.
Same data, five value lenses (2026)
- Balanced60.8
- Pinker65.1
- Sen61.6
- Rockström55.8
- Taleb48.6
- PeaceWelfareKnowledgeFreedomSustainabilityCooperation
| Lens | Peace | Welfare | Knowledge | Freedom | Sustainability | Cooperation | Civilization Score |
|---|---|---|---|---|---|---|---|
| Balanced | 17% | 17% | 17% | 17% | 17% | 17% | 60.8 |
| Pinker | 22% | 28% | 22% | 11% | 6% | 11% | 65.1 |
| Sen | 12% | 24% | 18% | 29% | 6% | 12% | 61.6 |
| Rockström | 13% | 13% | 6% | 6% | 44% | 19% | 55.8 |
| Taleb | 28% | 11% | 6% | 11% | 28% | 17% | 48.6 |
Taleb additionally blends 50% weakest-dimension score.
Normalization Methods
Used for linearly distributed indicators (life expectancy, literacy rate, freedom scores).
Used for multiplicative quantities among the scored indicators (GDP per capita, war death rate). The log transform prevents extreme values from dominating the scale.
Indicators with no value for a year are left out of that year's average. Provenance labels (Observed / Published / Estimate) are disclosed, but they do not change the weights.
Between two data points of a series, values are interpolated linearly. After a series' last data point its value is carried forward for up to 10 years, and the share of the score resting on carried values is shown. The cosine ease applies only to the pre-1900 estimate curve described above.
Dashboard Ideas — Coming Soon
We're developing 15 new visualization dashboards. Each provides a unique lens into civilizational data.
Country Comparison Tool
In DevelopmentSide-by-side radar charts comparing any two countries across all 6 dimensions with time-series overlay.
Gapminder-Style Bubble Chart
PlannedAnimated bubble chart showing countries moving through multi-dimensional space over time, inspired by Hans Rosling.
Dimension Deep-Dive
In DevelopmentDrill down into any single dimension to see all sub-indexes and raw indicators with interactive filtering.
Historical Event Impact
PlannedVisualize how major events (wars, pandemics, inventions) impacted each dimension with before/after comparison.
Turchin Cycles Overlay
ResearchOverlay Peter Turchin's secular cycles (demographic-structural theory) on CivStat time series.
Planetary Boundaries Radar
PrototypeReal-time radar chart showing Earth's 9 planetary boundaries and how close we are to each threshold.
Income vs. Freedom Scatter
PlannedInteractive scatter plot exploring the relationship between economic development and political freedom.
Knowledge Diffusion Timeline
PlannedHow fast do technologies spread? From printing press (centuries) to smartphone (decade) to ChatGPT (months).
Conflict Heatmap
ResearchGeographic heatmap of conflict intensity over time, showing how violence has shifted across regions.
Sustainability Trajectory
ResearchForward projection: what happens to Civilization Score under different climate scenarios (1.5°C, 2°C, 3°C, 4°C)?
Small Multiples Grid
PlannedTufte-inspired small multiples showing all 76 indicators as miniature sparklines for pattern recognition.
Confidence Map
PlannedVisualize data confidence across time and dimensions — where do we know the most? Where are we estimating?
Development Archetype Clusters
ResearchK-means clustering of countries by their dimensional profiles: Nordic, East Asian, Petro-state, etc.
Rate of Change Dashboard
In DevelopmentWhich dimensions are improving fastest? Where is progress stalling? First and second derivatives of Civilization Score.
What-If Simulator
PlannedAdjust dimension weights and see how the Civilization Score changes. Explore alternative weighting philosophies.
Data Limitations & Disclaimer
- • Ancient data (~1500 CE) relies on archaeological estimates and carries high uncertainty.
- • Interpolated year values are trend-based estimates, not observed measurements.
- • Collapsing “civilization level” into a single index inevitably simplifies complex history.
- • CivStat is a visualization tool for educational and exploratory purposes, not academic research.