Great comic by xkcd.
Originally posted in 2009.
Lisa Wade, PhD is an Associate Professor at Tulane University. She is the author of American Hookup, a book about college sexual culture; a textbook about gender; and a forthcoming introductory text: Terrible Magnificent Sociology. You can follow her on Twitter and Instagram.
Comments 6
NL — March 7, 2009
I about died laughing when I read this the other day.
A recent strip that looks at "The Princess Bride" was also interesting -- sort of turning common romantic tropes, especially with such a well-loved movie, on its head.
http://xkcd.com/549/
mordicai — March 7, 2009
Oh yeah-- very very funny.
leontine — March 7, 2009
Don't forget the famous xkcd rollover text: "Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing, 'look over there.'"
Sociological Images » WHAT WE’VE BEEN UP TO BEHIND YOUR BACK (APRIL 2009) — May 1, 2009
[...] Total Drek revised an xkcd cartoon on the difference between causation and correlation. So we added it to our original post. [...]
Alice — April 27, 2010
Haha! I feel kind of lame that I both get it and appreciate it.
Bill R — September 28, 2014
HA! I have found correlations in social science to be low and generally unreliable, especially over time. Our risk management models for insurers, healthcare businesses and lenders constantly need to be refreshed or they go stale.
Statisticians are typically not gurus here but if you ignore their validation tools you swim at your own risk. And when they want to help you understand the limitations of parametric statistics on skewed data, let them.
In the end I'm left will precious few guidelines, like:
1. LOOK AT THE DATA. VISUALIZE!
2. Use very large and varied datasets. Forget conducting studies using a couple hundred observations that are obvious biased subsets of the real population you're interested in. (Studying student behavior with self-reported data in a college or two gets you what you deserve.)
3. Think casually, but realize that its elusive for most of us.
4. Don't invent theoretical constructs and then assume they're real without repeated assessment and criticism. All models are wrong and some are useful; develop a deep understanding of the truth in that statement.
5. Triangulate.