Safest Country Maps Hide the Same Footnote
Crowd clicks are not victim surveys. Numbeo versus Gallup versus UNODC crime data in plain terms, plus a 30-second check before you share.
Verdict: missing context.
Primary source: Numbeo Crime Index country rankings, 2026 mid-year →
Safety maps get reposted as fact every month. They mix votes, reported experience and police files, three different things.
Short answer: The claim is not true, and the error is a swap of one word. The maps that go viral are labelled crime data, but they are usually built from a crowd survey of how people feel, not from counts of victims. Three different projects feed these maps, and they answer three different questions. A map is not wrong because it uses one of them. It is wrong when it calls the result crime data.
1. The viral claim
The post usually looks like a colour-coded world map with a handful of countries marked safest and a headline calling it the most dangerous. Nothing in the graphic says whether the underlying numbers are police reports, survey answers or self-selected website votes, and the words crime and crime rate are used for all three.
That single missing word is the whole dispute. A map built on feelings answers a different question from a map built on victimisation records, and the two can rank the same country in opposite directions.
2. Numbeo: the crowd survey
Numbeo’s own crime page states the scale of the input. Its crime survey has been completed by 174,124 people across 8,195 cities worldwide, and the site describes how it works: Numbeo leverages crowdsourcing to collect the data.
The questions it asks are about perception and worry. How serious do you feel the level of crime is. How safe you feel walking during daylight and during the night. How worried you are about being mugged or robbed. How worried you are about being physically attacked by strangers. How worried you are about being insulted or pestered in public. How worried you are about being attacked because of your skin colour, ethnic origin, gender or religion. How much of a problem people using and dealing drugs is. And how much of a problem property crimes and violent crimes are.
Two consequences follow. The answers describe how people feel, not what happened to them, so the index moves when a city’s mood moves rather than when its streets change. And the people who answer are the people who found the site, which skews toward English speakers, urban users and people already interested in the question. The published rankings are titled Crime Index by City 2026 Mid-Year and Crime Index by Country 2026 Mid-Year, and the label “Crime Index” is doing more work than the data supports.
3. Gallup: the survey with a sampling frame
Gallup’s Law and Order Index is the one most safety maps borrow, and it is a different instrument. Gallup says the index is collected through the Gallup World Poll and uses four questions to gauge people’s sense of personal security. The positive answers to those four questions produce the score, and a higher score means a higher share of the population reporting that they feel safe.
The four questions are these. Do you feel safe walking alone at night in the city or area where you live? Do you have confidence in the local police force in the city or area where you live? Within the last 12 months, have you had money or property stolen from you or another household member? Within the past 12 months, have you been assaulted or mugged?
Note what that mix contains. Two questions about feeling safe and trusting police, two about reported victimisation. So the index is not a victimisation rate either. It is a public safety measure built on reported experience and perception.
What makes it usable for comparison is the sampling. Gallup says the results are based on nationally representative, probability-based samples among the adult population aged 15 and older in 144 countries and territories in 2024, gathered by telephone or face-to-face survey with approximately 1,000 or more respondents in each country or territory. The margin of sampling error on the total national adult sample ranges between plus or minus 3.3 and plus or minus 5.4 percentage points at 95% confidence.
The numbers behind the viral maps come from that latest Global Safety Report. The world score is 83 out of 100. The highest scores are Tajikistan at 97, Singapore at 95, Kosovo at 94, and China, Iceland and Vietnam at 93. In Europe the leaders are Kosovo at 94, Iceland at 93, Norway at 92, Austria and Switzerland at 91, and Denmark and Finland at 90. In the Middle East and North Africa, Bahrain, Oman and Saudi Arabia each score 91. The United Kingdom scores 84 and the United States 84, against a world figure of 83, while Canada scores 82.
That pattern is the reason safety maps keep colliding with each other. A perception index that measures walking alone at night and trust in police will rate a well-lit, well-policed city highly, and that is not the same question as how many crimes were recorded.
4. UNODC: the police files
The third source is the one that is actually about recorded offending. The UNODC thematic programme on research, trend analysis and forensics manages global and regional data collections, provides scientific and forensic services, defines research standards, and supports member states to strengthen their data collection, research and forensics capacity.
The input is national crime statistics, which means recording practices, definitions and filing rates differ between countries, and the same act can be classified differently in two places. Homicide is the exception that compares best, because a body count is harder to reclassify than an assault.
5. The 30-second check before you share
- Who answered, and how many? A crowd survey and a probability sample of 1,000 adults are different instruments.
- What exactly was asked? Feeling safe walking at night is not a record of theft.
- Which year, and which vintage? A ranking labelled 2026 mid-year is not the same data as the one you saw last year.
- City or country? A calm capital can mask rough ports in a single national mean.
For a worked example of two credible sources telling different stories about the same country, read our comparison of Sweden crime headlines against survey data. The same method applies to every ranking, and our guide to reading rankings without getting fooled sets out the general rule.
6. The verdict: missing context
The viral claim fails because of the word crime. What these maps measure is closer to safety perception, reported experience, or administrative record keeping, depending on the source, and the three do not agree with each other.
The context that rescues the maps is simple. Each of the three is legitimate for its own question, and each publishes its own method. Gallup publishes the questions, the sampling frame and the margin of error. Numbeo publishes the survey size and the wording. The UNODC publishes its standards. A map that carries the source name and the question asked is not misleading. A map that drops both and labels the result crime data is.
Name the lane, or drop the post.
Sources to check yourself
- Numbeo crime survey: the respondent and city counts and the full list of questions.
- Numbeo Crime Index by Country, 2026 mid-year: the country rankings themselves.
- Numbeo methodology: the crowdsourcing model and what the site says its data can and cannot do.
- Gallup, What Is the Law and Order Index?: the four questions, the sampling method, the margin of error and the country scores.
- Gallup, What Is Global Safety?: how the index is built and how it relates to the Global Safety Report.
- UNODC data and analysis: the office’s data collections, research standards and support for national statistical capacity.
Frequently asked questions
Q: Which map is the real one? A: There is no single real one. Numbeo is a crowd survey of perception, Gallup is a probability survey of reported experience and feeling, and UNODC works from national administrative records. Each answers a different question, and each is valid for its own.
Q: Does the Gallup Law and Order Index measure crime? A: It measures a public safety index from four questions, two about feeling safe and trusting police and two about having money or property stolen and being assaulted or mugged in the past year. It is a measure of reported safety, not a police-recorded crime rate.
Q: How reliable are the country rankings? A: Gallup says its results come from nationally representative, probability-based samples of about 1,000 or more adults per country across 144 countries and territories, with a margin of error of plus or minus 3.3 to plus or minus 5.4 percentage points. A crowd survey has no such frame.
Q: Why do the safest-country maps disagree with each other? A: Because they measure different things. A city can score well on how safe people feel walking at night and badly on recorded assaults, and a country can score well on confidence in police without ranking well on victimisation.
Reader discussion
Conversation for this story opens with memberships. Until then: found an error? Report it — we correct visibly.



