pbinom((200*.95)-1, size=200, prob=.9, lower.tail=FALSE)
## [1] 0.00807125

or

sum( dbinom((200*.95):200, size=200, prob=.9) )
## [1] 0.00807125
n=670
p=.85
CI=.95
z.star=round(qnorm(CI+(1-CI)/2),2)
SE=sqrt( (p*(1-p))/n )
p+c(-1,1)*z.star*SE
## [1] 0.822962 0.877038

We are 95% confident that 82.3% to 87.7% of all Americans have good intuition about experimental design.

newError=0.01
n=((z.star**2)*p*(1-p))/(newError**2)
print(n)
## [1] 4898.04

we need at least 4899