Quote:
Originally Posted by illionaire
I never said that, what I was saying was that this:
"A random sample of 50+ is about where I start stereotyping."
is tilting as ****.
okay, I don't know how your post said that. Apologies on the tilt. Also stereotyping may be the wrong word
And I thought 50 was being prudent. imo you can use the Bernoulli distribution to get significant results with a smaller amount of data
Let's use an example:
Say I see a random sample of Singaporeans, and for lunch 35 out of 50 ate rice. Out of 50 americans, 5 out of 50 ate rice
I will go on to "sterotype" (insert better word if you like) Singaporeans as rice eaters. If I meet a random one, I'll think "There's a good chance he ate rice for lunch today". With an american I will think probably not