They could be p-hacking, though getting 6 sigmas through p-hacking without outright fraudulent data is actually some pretty impressive level of p-hacking. I'd s

They could be p-hacking, though getting 6 sigmas through p-hacking without outright fraudulent data is actually some pretty impressive level of p-hacking. I'd say it's more likely they either made an error or there really is a correlation.