451124_en-US

liliconcom

"Spurious Correlation" Demo

Hi all,

 

happy new year!

 

A champion at one of my customers asked, if we have any Examples of "Spurious Correlation". The idea is to open the eyes of new users when analyzing data.

Some initial ideas and inspiration were Anscombes Quartett, or looking at only subsets of a dataset etc.

 

I actually think it is a great idea, but so far I don't have any good examples (except the two mentioned above).

 

Would be happy about inspiration or exchange on this topic!

 

Best regards from Heidelberg,

Florian

Re: "Spurious Correlation" Demo

I used https://www.tylervigen.com/spurious-correlations in the six-sigma classes I taught. Some of the examples are pretty funny. My students and I both had a lot of fun with it. 

Re: "Spurious Correlation" Demo

I also used Tyler's "Spurious Correlations" web site in JMP classes. No one had trouble with the concept of "correlation does not necessarily imply causation" afterward.

 

(It would be great if journalists spent time on this web site...)

Re: "Spurious Correlation" Demo

Thanks!

Tyler Vigen is great - STIPS actually has an example from him. Good reminder that we already have this resource in place, @mia_stephens  

Re: "Spurious Correlation" Demo

Hi Florian,

I created the attached example for a customer workshop. It was based on an example from a Discovery keynote by Chris Nachtsheim. It is more about the idea of a lurking variable / indirect causation.

It might be interesting for you to use.

Cheers,

Phil