Check a statistic's date and definition

What to do: Read the fine print on every statistic

Every number you see about a country, population, GDP, life expectancy, comes from a specific source, measured in a specific year, with a specific definition. Ignoring that context is the most common mistake in country comparison. This page gives you a repeatable method to check those details before you use the figure.

The three‑step check

Step 1, Find the source and year.
A figure without a source and year is unusable. Look for the attribution in a footnote, a methodology note, or the metadata of a dataset. The most authoritative sources for cross‑national data are the United Nations (population, development indices), the World Bank (economic data, income classifications), the World Health Organization (health metrics), and UNESCO (education statistics). National statistical offices are the primary source for each country’s own census, survey, and administrative data. If the source is missing, do not use the figure.

Step 2, Check the definition.
The same label can mean different things. “Literacy rate” may mean “can sign your name” in one country and “can read a newspaper” in another. “Unemployment rate” often excludes discouraged workers and informal employment. “GDP per capita” can be nominal (at market exchange rates) or PPP (adjusted for local price levels). Every definitional choice changes the number. If the definition is not stated, treat the figure as unverifiable.

Step 3, Note the denominator and unit.
A statistic like “health spending as % of GDP” uses a different denominator than “health spending per capita.” A rate like “infant mortality per 1,000 live births” is not comparable to “infant mortality per 1,000 population.” Units matter: metric tons vs. short tons, international dollars vs. US dollars, square kilometers vs. square miles. Always confirm the denominator and unit before comparing across countries.

What to look for in footnotes

Footnotes are not optional reading. They contain:

• Data lag, Most official economic and social indicators are published 1-3 years after the reference period. A 2024 World Bank figure for GDP per capita is likely based on 2021 or 2022 data. The footnote will state the reference year.
• Revisions, Statistical offices revise figures when methodology changes or new data becomes available. The footnote will note the revision date and explain the change. Always use the latest revision.
• Geographic coverage, Some datasets include disputed territories; others do not. The footnote will specify the geographic scope. If you compare a UN figure that excludes a territory with a national figure that includes it, you are comparing different things.
• Sampling vs. census, Population figures from the UN are often projections based on the last census. The footnote will say “estimate” or “projection.” A census count is an actual enumeration, but censuses are conducted only every 5-10 years and can miss millions of people, especially in rural or nomadic populations.

The source hierarchy

When you have conflicting figures from different sources, use this hierarchy:

1. Primary source, The national statistical office of the country in question. This is the most authoritative for that country’s own data, but it may use definitions that differ from international standards.
2. International agency, United Nations, World Bank, WHO, UNESCO. These agencies standardize definitions and methodologies to allow cross‑national comparison, but they rely on data provided by national offices and may apply adjustments that the country does not recognise.
3. Secondary compilation, CIA World Factbook, OECD, IMF. These are useful but add a layer of interpretation. Always check their methodology notes.

The one question to answer

“What is the exact definition, source, and year of this statistic?”

If you cannot answer that question for every figure you use, you are not comparing countries, you are comparing labels. The same label can hide different realities. For example, two countries with a “literacy rate” of 95% may define literacy so differently that one has a functional literacy rate of 70% while the other has 90%.

How to apply this to common attributes

Country

The definition of “country” varies. The United Nations has 193 member states. Some datasets include Taiwan, Western Sahara, or Kosovo as separate entities; others do not. Always check the geographic coverage footnote. ISO 3166 country codes are stable and change only when countries merge or split. Use the UN membership list as the default definition of sovereignty, and note any exceptions.

Population

Population figures from the UN are estimates updated annually, but they are projections from the last census. The reference year is stated in the dataset. A census is an actual count, but it is conducted only every 5-10 years. The most recent census year varies by country. Always use the most recent census figure for the country in question, and note the census year. Population estimates between censuses are less reliable, especially for countries with high migration or rapid demographic change.

Capital

The capital is the seat of government. It is not always the largest city. For example, Canberra is the capital of Australia; Sydney is the largest city. The capital is defined by law, not by size. Check the official government website or the UN Geospatial Information Section for the authoritative list.

Currency

Currency codes follow ISO 4217. The unit is the local currency (e.g., Japanese yen, Indian rupee). Exchange rates change daily. Do not use nominal exchange rates to compare living standards; use PPP adjustments. The World Bank publishes PPP conversion factors annually. The IANA maintains time zone and country code data, not currency data directly, but the ISO 4217 standard is maintained by the Swiss Association for Standardization.

Language

Official languages are legally recognised. Most spoken languages are different vernaculars or dialects. UNESCO maintains the most comprehensive data on language use and education. National statistical offices publish language data from censuses. Do not assume that an official language is the most widely spoken in daily life.

Common pitfalls

• Mixing nominal and PPP, GDP per capita (nominal) uses market exchange rates and reflects international purchasing power. GDP per capita (PPP) adjusts for local price levels and reflects domestic living standards. They are not comparable.
• Comparing different years, A 2020 HDI figure is not directly comparable to a 2015 HDI figure unless the methodology has not changed. The UNDP revises methodology every few years.
• Ignoring data lag, The most recent World Bank GDP per capita figure for a country may be from 2022, even though the dataset was published in 2024. Always check the reference year.
• Using a single source, The CIA World Factbook is a useful secondary compilation, but it is not authoritative for any single country. Cross‑check with the UN or World Bank.
• Treating estimates as facts, Population projections are not census counts. They are model‑based estimates and can be wrong by millions, especially in countries with high migration or conflict.

When to stop and find another source

If a statistic lacks a source, year, definition, or unit, do not use it. If the source is a blog, a news article without attribution, or a secondary compilation that does not cite its own sources, find a primary or international agency source. If the definition is ambiguous or the denominator is unclear, find a clearer figure. The goal is not to have a number; the goal is to have a number you can trust.

The durable principle

No single number captures a country’s reality. A GDP per capita figure tells you nothing about distribution. A literacy rate tells you nothing about quality of education. A life expectancy figure masks huge disparities by region, income, and gender. The only way to use statistics responsibly is to know exactly what they measure, how they were collected, and when. That is what this page is for.