Mike D. answered 12/21/20
Effective, patient, empathic, math and science tutor
Margin of error = 1.96 sqrt (pq/n)
1.96 (Z score for a 95% confidence interval)
p=q=0.5 as stated by Jon (given no other information)
n is the sample size
So 0.02 = 1.96 sqrt (0.5 x 0.5 / n), giving n=2401 as Jon calculated. (This is correct theory).
Obviously intuitively this isn't sensible, as if we polled all 1600 students we would have perfect information, so a better than 95% confidence interval !