Confidence intervals?

I am having a difficult time answering this question... Could someone help me figure it out?

A direct mail sales company must determine its credit policies quite carefully. Suppose that the firm suspects that advertisements in a certain magazine have led to an excessively high rate of right offs. The firm wants to establish a 90% confidence interval for this magazine's write off proportion that is accurate to within 2% of the true proportion. How large of a sample is necessary?

What have you managed to consider?

1) "quite carefully" and "90% confidence interval" are unlikely to be consistent. 90% is pretty wide!
2) How big is the population we are examining? If it's not big enough, we will need to worry about correcting for a finite population.

3) It is hoped you have seen this guy...

\(\displaystyle n = \frac{t^{2}s^{2}}{d^{2}}\)

No?