9.59 Selective reporting and p-hacking In the webcomic on the link http://xkcd.com/882/, a girl claims that jelly beans cause acne. Scientists investigate and find no link between the two (p > 0.05). They are asked to check if jelly beans of a particular color cause acne. They test 20 different colors each at a significance level of 5%

and find a link between green jelly beans and acne.

This leads to a newspaper headline, “Green Jellybeans Cause Acne” where the 5% chance of the link is mentioned as 95% confidence. When the scientists repeat the same experiment, they are unable to find any link between acne and color of jelly beans.

They conclude that the earlier result might be coincidental.

Using this example, explain why you need to have some skepticism when research suggests that some therapy or drug has an impact in treating a disease.