Maths · GCSE · Statistics

Correlation is not causation

Correlation is not causationFree3:53Heddy Score 3 of 7, below the standard
Next clip

What you need to know

  • Overall percentages can hide a different pattern within subgroups, as Simpson’s paradox shows.
  • A lurking variable can affect both the grouping and the result, so check how data have been divided before drawing a conclusion.

In this clip

Runs 3:53, from 0:06 to 3:59 of the full video.

Heddy Score3/7, Below the standard
See the seven stars
  • Trusted source · passed. The creator's own upload, live and playable here.
  • On the map · not judged yet. A teacher watched it: it mainly teaches this map topic.
  • Spec-true · not judged yet. Accurate, and right for the current UK specification.
  • Focused · passed. One idea, short enough to watch in one go.
  • Accessible · passed. Good captions, clear sound and pictures.
  • Safe for children · not judged yet. Right for 11 to 16 year olds, with no sales pitch.
  • Proven · not judged yet. Earned from our students: they learn from it, find it helpful and watch it to the end.
★ passed · ☆ pending · ✕ failed
Our students

Not enough students yet

Our own students' answers. These are what move ★7 Proven.
On YouTube

1.76M views · 35.3K likes · TED-Ed 23M subscribers

As reported by YouTube on 2026-11-21. Not part of the Heddy Score.
How the Heddy Score works

Why we picked it

The hospital and other examples show how subgrouping can reverse an apparent statistical trend.

TED-Ed on YouTubechecked 2026-10-20

Plays from youtube-nocookie.com. No YouTube cookies until you press play.