VikeMan
It ain't all burritos and strippers, my friend.
- Joined
- Aug 24, 2010
- Messages
- 5,982
- Reaction score
- 6,247
I've been arguing since almost day one that they should be using a p<0.20 or 0.15 rather than p<0.05 as a threshold of *maybe* there's a difference, rather than declaring "tasters were unable to reliably distinguish...". Their goal of 95% confidence is too great. This is joe schmo science, not a laboratory. I look at the p data to know whether the experiment might be onto something or not.
What really gets me is that an overwhelming majority of "tasters were unable to reliably distinguish" experiments has more than 1/3 of the tasters identifying the different beer. If there was any validity at all to this misleading label of "tasters were unable to reliably distinguish," there should be just as many where less than 1/3 got it right, yet they are pretty hard to find.