“I don’t believe it!” was the exasperated catchphrase of Victor Meldrew, played by Richard Wilson in the television show One Foot in the Grave. While I didn’t find the show particularly funny, Father Ted brilliantly mocked Wilson’s distaste towards the catchphrase in The Mainland, thus cementing the catchphrase’s place in pop-culture history.
I found myself shouting “I don’t believe it!” earlier today when I clicked on a Fairfax “news” article about a “study” that claims men overeat in the presence of competitive environments. I could see where this was going straight away, I was not surprised to discover that the study was from none other than Brian Wansink‘s group at Cornell University.
Why was I not surprised? Because Wansink has form: It all started in 2016, he published a horrific post extolling the praises of a graduate student who was able to churn out 5 papers(!) from some data about restaurant eating habits. Furthermore, he derided a post-doc who refused to have any part in the whole affair for wasting too much time on “Facebook, Twitter, [and] Game of Thrones”.
The post aroused suspicion within the scientific community. Wansink’s work was subjected to further scrutiny, where it was discovered that Wansink was practicing P-hacking. P-hacking is where one tests a data set looking for the slightest correlations between any sets of variables. By convention, a statistical significance (P) below 0.05 is deemed to confirm a hypothesis. The trouble with P-hacking is that no hypothesis exists to motivate the design of an experiment, or to collect data.
Even worse, researchers found that Wansink’s statistical analysis was atrociously poor and that some of his work was self-plagiarised. It comes as no surprise that the retractions are starting to rack up.
The paper that the article was referring to* was one produced by the P-hacking effort outlined in Wansink’s brag post and has had to be corrected. Was any of this reported in the Fairfax article? Of course not, Wansink’s claims were simply parroted by the journalist. There wasn’t even a link to the paper that the article is supposed to be about! That the study was the result of academic misconduct is an important piece of information that readers should know about. More importantly, the poor quality of Wansink’s research means we can’t be sure that the competitive overeating phenomenon he claims exists is even true!
So we have an badly written article about a study so poor that it doesn’t warrant being written about. Why does this happen? The corporate media’s insatiable demand for ad revenue would be my guess. P-hacked studies that claim ridiculous sounding correlations between variables are typical clickbait that draws the attention of the non-scientific masses. Real science articles are usually less folksy and only appeal to a limited audience, hence don’t feature in mainstream news.
This approach is bad for science. The media is the only interface between scientists and much of the population. Superficial reports about dodgy studies presented as fact create a picture of scientists as bumbling, grant-hungry fools; whose findings are either bloody obvious, or obviously rubbish. The dumbarse comments below the offending Fairfax article prove my point.
While scientists such as Wansink are incompetent and their findings are ridiculous, most scientists are highly knowledgeable and produce robust work. But the public doesn’t get to see these people because their work isn’t “sexy” enough. Other fields of science such as climate change and vaccine research are then attacked by charlatans who are only too happy to exploit the public’s distrust in science to further their own agenda. The media has created the unbelievable situation where the public are lied to and then won’t believe the scientists who try to set the record straight.
It is absurd that the media these days is about profit first and disseminating information second. It is absurd that Fairfax think it is acceptable to publish claims by a fraudulent academic without letting it’s readers know that the claims are supported by shoddy research. It is absurd that our society has to suffer because confidence in science is undermined by media only interested in its bottom line. Down with this sort of thing.