The Inside Story Of How An Ivy League Food Scientist Turned Shoddy Data Into Viral Studies (Brian Wansink, Cornell)

https://www.buzzfeed.com/stephaniem...cking?utm_term=.cb9BaqMQN&bftwnews#.brXZqwz1x

 

Egads

 

Really brings home the perils of changing the question (or just numbers in the question) after scouring the data for answers to a question you did not originally ask, and maybe not properly controlled for. Presumably there will sometimes be genuine cases where someone stumbles upon an apparent correlation of interest in the data post hoc, but would then need to do another independently designed study with freshly acquired data to prove if a fluke or not.

 

This is such an elegant story - a beautifully perfect train-wreck, so to speak.

The Cornell prof in question got caught out by blogging about how to fish for relationships in data sets. He'd had a 30-yr academic career, and apparently had no idea that his normal research methodology (p-hacking) is unscientific; a recipe for false positives. (Why else would he tell the world that this is how to do research?)

 

Yep. A mind set so steeped in bad methodology, that he had no recognition of that fact. Sort of sounds familiar.

 

I feel for the "dogged" unpaid grad student from Turkey "who never said 'no'"

 

My sympathy is somewhat limited by having read so many crappy CFS dissertations.

There are so many PhD students working underneath quacks churning out dodgy work and when they play along with it they can do real harm to others. The money made from PhD students seems to be another corrupting aspect of academia (although I don't really know enough about it to be sure).

 

Good point. I suspect I wouldn't have made that comment if the research hadn't been about subsidised pizza eating.