If P is the probability of yes, then (1-P) is the probability of no. If you known P with error e, then you know (1-P) with exactly the same error. If yes is coded a 1 and no as 0, then the mean is the probability of yes, so the error bar on the mean is the error bar on the probability of yes.