- Kathryn Patterson
My Reaction to John Oliver's Take on "Studies"
I enjoy watching "Last Week Tonight with John Oliver" every week. I actually subscribe to HBONow, but I don't stay up late enough to watch on Sunday nights. Instead, my husband and I watch it together every Monday. This last episode, John Oliver covers the problems present with all of the "scientific studies" referenced by media, specifically how the media misreports science. I'm including the video of this segment below if you want to watch it, though I caution you that it's not necessarily safe for work. After watching this, I felt a bit guilty because I don't include links to all the studies I read for various posts, mostly because the studies themselves are quite boring to read. There's nothing like a bunch of scientific jargon to put a person to sleep. Note: If I ever cite a study and you want to read it, let me know. I will try to include links from now on. But I think that John Oliver could have explained a few more bits:
The 'p' in p-hacking stands for 'probability'.
Scientists who do p-hacking typically work on meta studies. That's a study where the scientist doesn't do new research, instead he or she looks through other people's work to find commonalities and correlations.
Correlation does not mean causation. Basically, just because a group declares that eating cabbage is statistically correlated to having an innie belly button does not mean that one causes the other one. Eating cabbage doesn't changing a person's belly button; nor does the type of belly button influence what a person eats.
Scientific study that finds no commonality or correlation can be as just important as ones that do. Or to paraphrase Thomas Edison, you haven't failed. You just found a way that doesn't work.
We need to set up rewards for people who find no correlations as well as for people verifying the results on exploratory studies. Maybe we need a Nobel prize for fact checking.
Basically, I agree with this segment. I wish that morning shows and other media did a bit more homework before presenting scientific studies, because there is enough craziness out there without mass media adding more.