Religion statistics by country

Religion statistics are not a single set of numbers you can compare neatly across borders. Every figure comes from a question that was asked differently, answered voluntarily or not at all, and then interpreted by a respondent who may hold multiple identities. To use this data responsibly, you must first understand the measurement itself.

Why census wording makes direct comparison unreliable

The United Nations recommends a standard question: “What is your religion, if any?” with an optional follow-up on denomination or sect (UN 2021). But most countries do not follow this. Some ask about “religious affiliation” (a group membership), others about “religion” (a belief or practice), and a few ask about “ethnicity and religion” in the same question. The wording changes what people report.

For example, a census that lists “Christian” as a category will get a different count than one that lists “Catholic,” “Protestant,” and “Orthodox” separately. A question that includes “No religion” as an explicit option captures more non-affiliated responses than one that requires the respondent to write it in. A question placed after detailed ethnicity questions may lead respondents to treat religion as a proxy for cultural background.

What to do: Before comparing two countries’ religious statistics, find the exact wording of the census or survey question used in each. If the wording is not available from the national statistical office, treat the comparison as uncertain. The United Nations Statistics Division maintains a repository of census questionnaires for many countries; that is your best starting point.

Nonresponse: the silent skew

Not everyone answers a religion question. Some countries make it voluntary; others require an answer but allow “no religion” or “prefer not to say.” The nonresponse rate varies widely. In England and Wales, the 2021 census included a voluntary religion question and 5.5% of respondents did not answer it (ONS 2023). In countries where the question is mandatory, nonresponse can be below 1%. But in surveys where religion is sensitive, nonresponse can reach 15% or more.

Nonresponse is not random. People who are less religious, belong to minority faiths, or live in regions where religious identity is politically charged are more likely to skip the question. This means the official figures systematically overrepresent the majority religion and underrepresent the non-affiliated and religious minorities.

What to do: Look for the nonresponse rate in the metadata of any religion dataset. If it is not reported, treat the figures as unreliable for minority groups. The UN Demographic Yearbook tables often include a “not stated” column; use it to calculate the effective response rate.

Identity complexity: one person, many labels

A person may report “Christian” on a census, attend a mosque for family events, and describe themselves as “spiritual but not religious” in conversation. They may be culturally Jewish but atheist, or Buddhist by family tradition and secular by personal belief. The 2014 Pew Religious Landscape Study found that 37% of U.S. adults had switched religious affiliation since childhood (Pew Research Center 2015). That is not a static identity.

Census and survey questions force a single answer. They cannot capture dual affiliation, private belief without public practice, or cultural identity without belief. Some countries now allow multiple selections (e.g., “Christian and Buddhist”), but most still require one choice. This means the data systematically flattens lived reality.

What to do: Use religion statistics to understand declared affiliation at a point in time, not to infer belief, practice, or cultural belonging. When writing a country report, add a sentence like: “These figures reflect census affiliation, not necessarily belief or practice; surveys that ask about behavior typically show lower rates of regular attendance.”

A practical checklist for using religion statistics by country

1. Identify the source and year. Is it a census (full count) or a survey (sample)? What year was it conducted? Data from 2011 is not comparable with 2021 without accounting for demographic change.
2. Find the exact question wording. Was it “What is your religion?” or “What is your religious affiliation?” Were response options provided or was it open-ended? Was “No religion” listed as an option?
3. Check the nonresponse rate. If more than 5% of respondents did not answer, note that the distribution may be skewed.
4. Look for multiple-response options. Did the form allow selecting more than one religion? If not, know that mixed identities are invisible.
5. Compare only identical metrics. Do not compare “percentage Christian” from a census with “percentage who attend church weekly” from a survey. They measure different things.
6. Contextualize with a second source. Cross-reference with a survey that asks about practice (e.g., Pew Research Center, World Values Survey) to see how declared affiliation compares with reported behavior.

Population, capital, currency, language, visa

These are the basic identifiers for any country comparison. Each one carries its own measurement caveats.

• Population: The United Nations Department of Economic and Social Affairs (UN DESA) publishes annual estimates, but these are projections based on the most recent census. Censuses are conducted every 5 to 10 years, so the estimate may be 2-5 years old by the time you see it. Always check the reference year. For example, UN DESA’s 2024 World Population Prospects includes estimates for 2023, but the underlying census data may be from 2011 for some countries.

• Capital: The capital is the seat of government, not necessarily the largest city. Canberra (population ~470,000) is the capital of Australia; Sydney (population ~5.3 million) is the largest city. The CIA World Factbook and the UN Geospatial Information Section are reliable sources for capital city identification.

• Currency: Every country has an official currency, but not all currencies are freely convertible. The IMF publishes a list of member country currencies and exchange rate arrangements. IANA maintains the currency code (e.g., USD, EUR, JPY) under ISO 4217.

• Language: The official language is legally recognized, often by constitution or statute. The most spoken language may be different. For example, India’s official language is Hindi (in the Devanagari script) plus English for official purposes, but the most spoken language by number of native speakers is Hindi, followed by Bengali. UNESCO’s World Atlas of Languages provides data on both official and spoken languages.

• Visa: Visa policies change frequently and vary by nationality. There is no single global source for current visa requirements. The IATA Travel Centre maintains a database used by airlines, but it is updated daily. The best practice is to check your own country’s foreign ministry website for the destination you plan to visit. Do not rely on third-party visa advice sites; they can be outdated.