Question 1161977
The difference between two proportions test is
z=(p1-p2)/sqrt(p1(1-pi)/n1+p2(1-p2)/n2))

The half-interval of the CI is the z*SE, where z(0.95) =1.645 for a 90% CI
That is added to and subtracted from the difference of the means, which is -0.03 if we make it p(A)-p(B)

Ho:p(A)-p(B)=0
Ha: p(A)-p(B) does not equal 0 
alpha=0.10 p{reject Ho|Ho true}
SE=sqrt (.05*.95/80+.08*.92/50)
=0.04545
the z value is -0.66, p-value=0.49 so Ho is not rejected

The half-interval is z*SE=1.645*0.0454=0.075
the interval is (-0.105, +0,.045). Because 0 is in the confidence interval, we are 90% confident that the true parameter of the difference is in the interval, and because 0 is in the interval, that is another way of saying the two proportions could be part of the same population and therefore not different,