Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

how is 26 people statistically relevant??


Well, that depends on the effect size etc. right? There's a chocolate bar in front of you and your two buddies. You split it in thirds. As soon as one of your buddies eats his, he drops dead. Your other buddy says "Sample size of 1. Psh. Billions have eaten chocolate.". He eats his. Drops dead. The sample size is now 2. You eat yours or not?

If the answer is yes, well all right.

If the answer is no, weaken the effects until you see what crosses the threshold of "should we study this more?".


It's relevant in that it's enough to cause them to plan to do a bigger study so we can say with more clarity how if it actually helps.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: