“It is remarkable how much long-term advantage people like us have gotten by trying to be consistently not stupid, instead of trying to be very intelligent.”
In 2015, Europe’s largest daily newspaper, Bild of Germany, ran a front-page story with the headline “Slim by chocolate!” that suggested the latest new diet fad. The story was then picked up by media outlets across the world from the Times of India to an Australian morning talk show. A team of German researchers at the Institute of Diet and Health had come up with a weight-loss formula that was both easy to implement and effective. Chocolate was the weight-loss accelerator that the world had been looking for.
The problem was that the research wasn’t robust. Journalist John Bohannon was proving a point—pretty successfully, it turned out. He was proving that there are a lot of bad data, and people (and indeed journalists) don’t often separate the wheat from the chaff. John and his team recruited 15 people and measured 18 different variables (weight, cholesterol, sodium, blood protein levels, sleep quality, well-being, etc.) over three weeks. This gave them a slightly better than 60 percent chance of finding a significant data point, even if the data were random, and in fact they found two to help prove their point.
Fooled by Randomness
The research was done to show that findings that look robust can carry limited actual meaning. ...