Hi Bailey,
First, we must verify that pn and (1-p)n both exceed 10. To do this, we compute p, which some books will call p-hat for p predicted--I'll stick with p to avoid confusion:
p= 255/300
p=0.85
Now, we need to compute np and n(1-p), where
n=sample size=300
np=300*.85=255
n(1-p)=300*.15=45
We can proceed with the z-confidence interval:
CI=p +/- z*sqrt[(p(1-p)/n)]
p=proportion we are testing
z*=z-critical value depending on confidence level
n=sample size
So, for this problem:
p=0.85
n=300
z*=1.645
Note: 1.645 is the z-critical value for all 90% 2-sided z confidence intervals, so you might want to memorize it even though it can be derived from the z-table. Proceeding:
CI=0.85 +/- 1.645*sqrt((0.85*0.15)/300)
CI=0.85 +/- 0.034
CI=(0.816, 0.884)
I notice you have many questions posted here--let me know via Wyzant if you want to meet online and talk through some of these. I hope this helps.