How statistics can be misleading - Mark Liddell
1,472,822 Views
27,590 Questions Answered
Let’s Begin…
Statistics are persuasive. So much so that people, organizations, and whole countries base some of their most important decisions on organized data. But any set of statistics might have something lurking inside it that can turn the results completely upside down. Mark Liddell investigates Simpson’s paradox.
Create and share a new lesson based on this one.
About TED-Ed Animations
TED-Ed Animations feature the words and ideas of educators brought to life by professional animators. Are you an educator or animator interested in creating a TED-Ed Animation? Nominate yourself here »
Meet The Creators
- Educator Mark Liddell
- Script Editor Alex Gendler
- Director Mike Foster, Tom Sanders
- Producer Tom Sanders
- Compositor Mike Foster, Tom Sanders
- Animator Mike Foster, Tom Sanders
- Sound Designer Tom Lowe
- Narrator Addison Anderson
by David Lovell
David Lovell
Lesson completed
Can randomisation save us from lurking variables? (I hope so)
I really like this video Mark. I also really like Larry Wasserman's explanation of Simpson's Paradox in "All of Statistics" Chapter 16 in which he uses "potential outcomes" as a lurking variable. If patients were always randomly allocated to hospitals A and B, I think we would be able to rely on the summary survival rates of the hospitals... wouldn't we?
Comments are closed on this discussion.
Mark Liddell
Lesson in progress
Hi David,
If patients were randomly allocated then we would want to rely on the summary survival rates. The real question is, what are the real survival rates? The data contains both a conditional variable and a small group size, which means the survival rates can be presented in two different (and opposite) ways.
butt head
Lesson completed
Yes, theoretically you could, but what if the patient was seeking a surgery that very few hospitals perform and was therefore choosing between a hospital in Pittsburgh and a hospital in Boston? These hospitals serve other people who are not seeking specialized surgery. This would mean, in order to truly randomize results, anytime anyone in Pittsburgh or Boston wanted to go to the hospital, they would randomly get put on a plane (or not.) Ignoring cost, and what that new scenario would do to the hospitals' statistics, some patients would die on the flight. If you want to truly randomize results for all hospitals you would randomize worldwide.
One of the reasons why statistics is a valuable field, is that the real world pretty much always operates in some imperfect way. Statistics can sometimes serve as a way of peering around those obstacles.
Kayode Jonathan
Lagos , Nigeria
Lesson completed
To that extent we know that best decision doesn't lead to the best oucome. In other words the larger the samples the closer to to reality and in so far we can't have 100% samples we shouldn't expect one solution fit it all in randomising.
