Good Data, Bad Conclusions

Tortured Data Provides Poor Intelligence

January 18, 2014

This started off as a routine day with the company statisticians fooling around (as usual), mass-emailing to our colleagues funny graphs found randomly on the internet:

Did you know that the real cause of the rising prevalence of autism was organic food consumption?:

 Lean Six Sigma Statistical Analysis: Spurious Correlation

Or that you could reverse global warming if you just changed your profession?:

Lean Six Sigma Statistical Analysis: Spurious Correlation Pirates

These examples got us thinking about why flawed statistical reasoning (of which these silly graphs are extreme examples) persists. “Correlation does not equal Causality” is an oft-repeated truism. Unfortunately, it doesn't stop some people from continuing to make the same mistakes in interpreting data (albeit of more subtle kinds) over and over again.

It may be de rigueur to blame bad conclusions on statistics and that much-maligned group, statisticians. They are an easy target. But the real culprit is lack of sound critical thinking that is grounded in scientific (and statistical) principles. Our mission is to help people sharpen their thinking about cause and effect: you can’t make things better if you don’t know what really makes things better.

Smita Skrivanek, Principal Statistician

1. Autism prevalence vs. organic food sales (image1): http://www.abc.net.au/science/articles/2013/04/29/3740590.htm
2. Global warming vs. Number of pirates (image2): http://www.forbes.com/sites/erikaandersen/2012/03/23/true-fact-the-lack-of-pirates-is-causing-global-warming/

+ View/Add a Comment (0 comments)

Hide Comments

You must be logged in to comment on this blog.

Please login above or

Create an Account

Need to Report Abuse?