Estimation of difference among population proportions
Assume the two proportions be described by P1 and P2, respectively,Then the difference absolute between the two proportions is described by (P1 - P2)
The standard error is described by:-
S(P1 - P2) = √{(pq/n1) + (pq/n2)}  =   where p = (p1n1 + p2n2)/(n1 + n1)  and q = 1 - p
Then described the confidence level, the confidence interval among the two population proportions is described by
(P1 - P2) ± Confidence level S(P1 - P2)
= (P1 - P2) ± Z  √{(pq/n1) + (pq/n2)} 
Whereas P = (p1n1 + p2n2)/(n1 + n1)  always remember to convert P1 and P2 to P.