digestVidSign in
summary

TED - Why you should love statistics | Alan Smith

Thumbnail of the video “TED - Why you should love statistics | Alan Smith”
Community upload. AI output has not been independently verified.
Author: TED
Categories: Science Education
Tags: science, education
Views: 1 · Last viewed: 2026-10-13 11:24 UTC

Short Summary

Main recommendation: Treat statistics as a tool for understanding ourselves as groups, and use numbers to test your own assumptions about the world around you, since people are consistently poor intuitive statisticians. Speaker Alan Smith argues that the "two kinds of people" split (those who can do numbers and those who can't) is a false dichotomy, and that anyone can be inspired by numbers. He backs this up with surveys showing large gaps between public perception and official figures, and with a local quiz that let people check their own guesses.

Key Points

  • The UK's 2003 numeracy survey found 47 of every 100 working-age adults lacked level one numeracy skills (fractions, percentages, decimals). The 2011 survey found 49, showing policy changes had not moved the figure.
  • OECD figures on young people's numeracy show the USA near 40% with low numeracy, several countries above 20%, and the Netherlands and Korea in single figures.
  • Public estimates were far off official figures: UK adults guessed about 24 of every 100 people in England and Wales were Muslim (official figure about 5); Saudis guessed just over 25 of 100 adults were overweight or obese (official about 34); Japanese guessed 56 of 100 lived in rural areas (official about 7).
  • Speaker cites Daniel Kahneman and Amos Tversky's research: people are poor intuitive statisticians, and media reporting by exception can distort perception.
  • The "How well do you know your area?" quiz, built from census data, drew about a quarter of a million players within 48 hours and crashed the statistics website, with people sending it to politicians.
  • Even experts misjudged their areas: the National Statistician John Pullinger scored 44 and Jeremy Paxman scored 36 (after a glass of wine), per the speaker.

Details Worth Noting

  • Concrete instructions: The speaker gives no step-by-step method for the reader to follow. The actionable suggestion is to test your own local assumptions: enter your postcode into the quiz, guess the figures for your area (for example, the share of under-16s, median age, and home ownership with a mortgage or loan), then compare against census data. Also treat striking statistics as a prompt to check your perception against official figures before drawing conclusions.
  • Quiz design: each question asks for a number from 0 to 100 per 100 people, answered with a slider over repeating icons (Isotypes, from Otto Neurath's 1920s–30s work), which lets people engage without fractions or percentages.
  • Example from Exeter: the speaker guessed 5 of 100 people under 16 (accuracy not confirmed in the transcript), and guessed an average age of 35, which he found too high because of the university's influence.
  • The speaker notes that easily visible traits like age are easier to judge than hidden ones like home ownership, so people fall back on their biases for the latter.
  • Caveats: the census data behind the quiz was already a few years old at launch, and the speaker concedes the underlying data was not new, only presented in a gamified way.

Bottom Line

Statistics is less about being good at maths and more about checking what you think you know against the data, which is why the speaker calls it "the science of us."