I’m not the best when it comes to stats. But I see some surveys who publish findings based on <30 responses. And I know that these are not valid, and it has something to do with the normal distribution.
In: 0
There is no set minimum to be a valid sample. What matters is confidence intervals. Confidence intervals basically mean ‘how much of the sample needs to align for it to be significant.’
Imagine you only asked 2 people if America is fascist, and they both said yes. Does this mean everyone thinks America is fascist, or is it just random? You have no confidence.
If you ask 30 people, “is America fascist,” and 28 say yes, you can be 95% confident America is fascist.
The larger your sample, the smaller % you need to gain confidence. If you asked 1000 people, and 700 say yes, you can still be pretty confident that America is, in fact, fascist.
Latest Answers