How we work out the numbers
AboutAverage is meant to be fun, but the numbers are serious. Here's exactly what we do with them.
Where the data comes from
Every statistic comes from an official source, such as a national statistics office or Eurostat, under a licence that allows reuse. We keep a copy of every file we download and record when we got it. Before a figure appears here it is checked automatically (for example, that a country's areas add up to its published total) and then approved by a person. The sources page lists every dataset used on the site, with its licence.
What kind of number is this?
Every answer tells you which of these it is:
- Published by the source
- Calculated by AboutAverage by adding up published figures (exact)
- Calculated by AboutAverage from published figures that were rounded or perturbed (approximate)
- An estimate from a survey or statistical model
We don't use one vague label such as "estimated". A survey estimate (with sampling error) is different from a figure we added up ourselves from rounded census counts, and we say which is which.
How "you're higher than X%" is calculated
We only place you in a group when the source publishes how the whole group is spread out: either percentiles (for example, the pay below which 10%, 20% … of employees fall) or counts in bands (for example, how many homes use 2,500 to 5,000 kWh). We never estimate your position from an average, because an average alone says nothing about how spread out people are.
Between two published percentiles we draw a straight line. Inside a band we assume people are spread evenly. Past the last published point, or in an open-ended band such as "85 and over", we only say "at least". Each result states which of these applied.
Your inputs are used in your browser and never sent to us.
Things we refuse to do
- Average averages, or add up percentages or medians.
- Compare countries when they measure something differently. Working from home during different lockdown phases, or incomes in different currencies and definitions, are examples; we explain why instead of showing a misleading answer.
- Rank or quiz on sensitive topics such as religion, ethnicity or health. Those appear only in area profiles for places with at least 10,000 residents, and very small counts are withheld.
- Fill a gap with a guess. If a total is incomplete, we leave it out.
- Treat our own visitors' answers as if they represented a country. If we ever show visitor comparisons, they will be labelled as a self-selected sample.
Quizzes and games
Spelling, maths and general knowledge questions are written and checked by hand. Guess the Average answers come from the same approved statistics as the rest of the site. None of our tests is a validated measure of intelligence, ability or health.
Found a mistake?
Please tell us. We correct errors and note significant corrections on the page.