Hacker News Reader

The lower rates weren't statistically significant.

I'm not agreeing or disagreeing with the study, just pointing out the table. In fact, it kind of me of the Texas Sharpshooter Fallacy - https://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy#Exa....

AKA the multiple comparison problem, AKA the look elsewhere problem, AKA the green jelly bean problem [1].

Edit: they looked at 17 cancers and one effect was significant. Is this exactly the jelly bean scenario?

Edit: reading the paper now. There are overall results that are probably more interesting.