As someone who knows a thing or two about statistics, this looks as a painfully wrong interpretation of the data. Essentially this is p-hacking, although it mig

As someone who knows a thing or two about statistics, this looks as a painfully wrong interpretation of the data. Essentially this is p-hacking, although it might be inadvertant by the authors. This XKCD explains it well: https://xkcd.com/882/ https://xkcd.com/882/ So, if you have a lot of data, you don't get to try many things and then report the ones with high significance. I really doubt this result will stand in a dedicated followup study.