The standard error of a proportion measures how much a sample percentage would vary from sample to sample. It is largest when p̂ = 0.5 and shrinks with the square root of the sample size.
SE(p̂) = √( p̂(1 − p̂) ÷ n )
If you sampled more than about 5% of a finite population, tick in a population size to apply the finite population correction.
Example
In a survey of 200 people, 42% agree. SE = √(0.42 × 0.58 ÷ 200) = 0.0349, so the 95% margin of error is 1.96 × 0.0349 = ±6.8 percentage points.