Wednesday, May 13, 2015

Correlation Found to Imply Causation


For anyone who has taken an introductory statistics class, the phrase: “correlation does not imply causation,” has likely been hammered into your brain. This logic has been considered ironclad for millennia (probably older than statistics, but what the hell; numbers), but a researcher for an online pay-to-play university might be about to change all that. If Dr. Hedupass’s (pronounced Hey-Dew-Pass) recent research is to be believed, correlation may indeed imply causation.

Wait Nancy, do you see that? It's a bunch of bullshit!
“We conducted a correlational study assessing previous correlational studies and their relational to causal outcomes. The data yielded marginally significant p-values (p = .15) suggesting that our model was indeed positively related to causation. That is to say that as the correlational nature of a study increases, so does the relation to causality.” For those who wish to view Dr. Hedupass’s project, he has not posted the data anywhere, but has assured us that the sample size is massive, the data were certainly cherry picked (to assure he got the best ones), and that he has conducted a median split as due diligence for his analysis.

I think I speak for everyone when I say, “Up yours my former stats teacher who failed me in high school for implying that my correlation between dinosaur attacks and genetic modification was not a ‘cause’ for alarm.”

--Ashton Macaulay, Amateur Statistician, and Raptor Alarmist

No comments:

Post a Comment