Methodology

How we compute civilizational stress.

Every score is the output of a published formula running over sourced indicators. Below is the formula, the inputs, and the choices we made — including the ones that could reasonably go another way.

The shape of the index

For every country we track, on every cron run, the system produces:

  • A normalized score for each of 31 indicators on a 0–100 stress scale (lower means less stress).
  • A score for each of 5 meta-indexes, averaged from its constituent indicators.
  • One composite score per country, a weighted average of the 5 meta-indexes.

Everything else on the site — Pulse, Research, Glossary, Country detail pages — reads from those three layers.

The composite formula

The composite is a weighted average of the five meta-indexes. The weights below were chosen to reflect how immediately each domain shows up in daily life — economic stress hits first, environmental stress accumulates last.

composite = 0.25·economic + 0.20·social + 0.20·mental + 0.20·technological + 0.15·environmental
Meta-indexWeightWhy this weight
Economic25%Felt first — income, jobs, housing are the most immediate pressure surface.
Social20%Sustained communal trust and cohesion underpin every other meta.
Mental20%Population-level mental load — anxiety, suicide, loneliness — measured separately from economic stress.
Technological20%Automation exposure and digital displacement are real but slower to translate into lived stress.
Environmental15%Accumulates over decades; weighted lowest only because year-on-year movement is slow.
Total100%

From raw measurement to stress score

Each indicator has its own native unit — Gini coefficient, GDP per capita, temperature anomaly °C, suicide rate per 100k. To make them comparable, we map each to a 0–100 stress scale using fixed bounds that are documented in the source code per indicator.

  • Stress-positive indicators (higher = worse, e.g. unemployment, homicide rate, temperature anomaly): mapped linearly between a low-stress bound and a high-stress bound.
  • Stress-negative indicators (higher = better, e.g. life expectancy, life satisfaction): inverted so the resulting score still aligns with the stress direction.
  • Bounds are reviewed periodically as part of the normalization sanity sweep audits.

Stress bands

Score bands convert continuous numbers into editorial categories. They are calibrated so that what we have actually been measuring across 25 countries spreads roughly evenly across the lower four bands, with the critical band reserved for genuinely uncommon situations.

  • Low
    025
    Functional society — stressors present but bounded
  • Moderate
    2645
    Strain visible in headline indicators
  • Elevated
    4665
    Recurring strain across multiple meta-indexes
  • High
    6680
    Acute stress; institutional buffer eroding
  • Critical
    81100
    Sustained crisis-level signals

Freshness and confidence

Not all data points are equal. Two visible signals tell you how much to trust any given number on the site.

Freshness

How recently the underlying observation was made.

  • Freshwithin 2 years. Counts at full weight.
  • Aging2 to 3 years. Counts at full weight, visibly tagged.
  • Stale3 to 5 years. Counts with downweight + warning.
  • Very staleover 5 years. Excluded or held until refresh.

Confidence

Where the data point came from determines how it is used.

  • VerifiedFiled with a regulator or official statistical agency (SEC EDGAR, WARN Act, World Bank, Eurostat, IMF, OECD).
  • ReportedWire service or established news outlet, single source.
  • RumoredTracked but not surfaced to scores until corroborated.

When multiple sources publish the same indicator

For indicators with more than one independent source (e.g. inflation from IMF WEO and World Bank, temperature from NASA GISS and Berkeley Earth), the orchestrator does three things on every cron run:

  1. Compare the latest reading from each adapter. If they agree within a per-indicator threshold, the primary source wins.
  2. Flag any divergence that exceeds the threshold. The divergence is recorded and visible on the transparency dashboard — we never silently pick a side without leaving a record.
  3. Fall backto the next available source if the primary fails, so the country's composite continues to update even during an outage.

What this is — and is not

The Human Index is a directional, peer-comparable stress scoreboard. It is intended for editorial framing, prioritization, and longitudinal comparison.

It is not a clinical instrument, not a policy optimization target, not a substitute for the original sources. The absolute number means less than its movement over time and relative to peers.