Originally Posted by
travelkid
1000 people asked seems to be a magic number in many professional surveys.
I think that such a result would be pretty representational.
you're indeed right.
the magic number (the sample size) is much closer to low 900s than 1000 but from statistics class i had long ago it's allows a 95% significance level on a 50/50 split in choice of A or B *had the sample size been random*.