Hey jickjoker.

The margin of error is used to measure the variance of a mean or proportion (a proportion is a mean).

If you want the margin of error for one candidate, you can treat the problem as a binomial and if this is done then the margin of error of getting one candidate against another does have the margin according to sigma^2 / SQRT(N) where sigma^2 is the variance and N is the sample size.

The above situation applies when you treat a 1 as beloning to the candidate of interest (i.e. they voted for the candidate) and a 0 corresponds to voting for someone else.