3 ms·
Be careful with statistics: In this experiment, 3000 eggs were boiled, showing a difference of 10% versus 12% cracked eggs with/without pricking before cooking.
by mkesper 11y ago
Be careful with statistics: In this experiment, 3000 eggs were boiled, showing a difference of 10% versus 12% cracked eggs with/without pricking before cooking.
Statistically, this does not have _any significance_ at all, though:
https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=de&ie=UTF-8&u=http%3A%2F%2Fwww1.wdr.de%2Ffernsehen%2Fwissen%2Fquarks%2Fsendungen%2Fei132.html&edit-text=&act=url https://translate.google.com/translate?sl=auto&tl=en&js=y&pr...
(Sorry, that translation is hilarious).
- tokenadult 11y agoFrom the article: "Here's the truth: there is no 100% fool-proof method, and anybody who tells you different is selling something. And I do believe I've tried them all, many, many times over. The number of eggs I've boiled over the last several years in carefully controlled circumstances numbers well into the thousands, but despite that, the best boiled egg I cooked this year is no better than the best boiled egg I cooked twelve years ago in that Beacon Hill kitchen."
- GFK_of_xmaspast 11y agoOn the other hand, if you actually bother to do the math, using the normal approximation you get that a 95% confidence interval around p=0.1 with n=3000 has an upper bound of 0.111, which is well-separated from p'=0.12. (in fairness, the confidence intervals do overlap each other)