A couple weeks ago, I was remarking how I like Coke Zero much more than Diet Coke, and Uschi said, "Oh, isn't it just the same thing, but just with a different label on it?" To which I responded "No way, it's completely different!" So we did a little blind taste test right there on the spot and I of course got them mixed up. :)
Ah, but that was just n=1. So with that in mind, we decided to take a more, umm, scientific approach at our lab holiday party today. Uschi set up two cups for all 13 participants, one with Coke Zero and one with Diet Coke. Everyone then had to say which they thought was which. Results? 11 of 13 got it right! (And I was one of them, thankfully...)
Verdict: Coke Zero and Diet Coke are distinguishable, with a p of 0.0017.
Ah, but much better to have three cups and you have to identify which ones are which. Much harder to get lucky. Of course, who has what in their cups will have to vary. Just two cups (50:50) is a little "light" isn't it.
ReplyDeleteOk, you've got that impressive p-value so there you go, although assuming it's a binomial distribution I'm getting a different (less impressive but still significant) one. Just wondering if the results were truly independent...
Hmm, I could have screwed up the p-value. I just did binocdf(11,13,0.5,'upper') in MATLAB. I suppose I should have multiplied by 2 for deviations in either direction.
ReplyDelete