Correlation vs Causation

In: Culture Graphic Design (general)

12 Jan 2014

It’s important to know the difference between correlation and causation when using charts. Duh. Below is a good example of why.

image

6 Responses to Correlation vs Causation

Avatar

Joe Blow

January 13th, 2014 at 12:57 am

So are we staying autism causes organic food sales? Or that organic food sales correlate closely to autism? Or that organic food causes autism? ;)

Avatar

jmc

January 13th, 2014 at 11:25 pm

You laugh (and you should) but this not only matches the Jenny McCarthy/vaccines crap, but some guy on Slate’s autism/cable TV “connection

Avatar

LiveFromRussianHill

January 13th, 2014 at 11:33 pm

Umm…Association, anyone?

Avatar

Cornfed

January 13th, 2014 at 11:57 pm

hah! Nice chart! Definitely no causation here. But almost certainly a real relationship.

Avatar

looking closely

January 14th, 2014 at 9:18 am

“Definitely no causation here”

Agreed any connection between these two things seems on its face ridiculous but still, at an R of .9971 the correlation between the two is incredibly robust.

More likely than not the two are either entirely unrelated (just “related” by data set “cherry picking”), or maybe both related to some other independent factor, but how can you state without absolute certainty that there is no causation here?

For example, can you really say for certain that there is no naturally occurring plant or animal toxin present in organic foods thats a contributing factor to autism?

Avatar

lucky

January 15th, 2014 at 3:29 pm

@looking closely, you should develop a theoretical argument for WHY that relationship might be the case and that it should be more compelling (more evidence, more consistent) than arguments for OTHER relationships.

Far more likely explanations include either (a) an increase in ASD diagnoses (NB: not just individuals with an ASD) led to in increase in organic food sales thanks to concerns about environmental exposure to synthetic toxins contributing to autism, (b) increased awareness of the potential impact of environmental exposure to synthetic toxins led to a concurrent increase in the sales of organic produce and ASD diagnoses, or (c) families had more disposable income from the late 90s into the late aughts to spend on more expensive, higher quality foods and identifying and seeking services for previously unidentified ASD.

I find it hard to imagine that there is a significantly higher rate of exposure to naturally occurring toxins than there was in the past. Granted the chart only goes back to 1997 but to support that argument you would also expect to see evidence of a relatively high rate of people exhibiting ASD characteristics before the Green Revolution as well as high rates of people exhibiting ASD behavior in parts of the world that the Green Revolution skipped over and they’ve always eaten food without synthetic fertilizers and pesticides.

Comment Form

What is Chart Porn?

An addictive collection of beautiful charts, graphs, maps, and interactive data visualization toys -- on topics from around the world.

  • Coffey: Not to mean but the theory that Christianity is the one and only true religion is questionable. I s [...]
  • Drumwaster: Good thing we have these snapshots, otherwise I'd have to rely on NASA for such data, and what do th [...]
  • Sean Collins: I'm no economist but how does raising the minimum wage help the unemployed? They have no wages to ra [...]
  • Benjamin: There is actually a difference between items showing up in your timeline (when you like a page and F [...]
  • Westrim: I am a Californian and I approve of this message. [...]