Sampling Distributions with Central Limit Theorem (Proportions)

I know how to use the TI-83 for the Central Limit Theorem with means in a word problem such as this:

The mean number of words per minute (WPM) rad by sixth graders is 90 with a variance of 441. If 149 sixth graders are randomly selected, what is hte probability that the sample mean would differ from the population by greater than 4.47 WPM.

normalcdf(-999999,85.53,90,21/sqrt(149) X 2

However, I'm having trouble figuring out word problems when no standard deviation is given, and proportions are used instead of means such as this one :

A carpet expert believes that 2% of Persian carpets are defective. If the expert is right, what is the probability that the proportion of defective in a sample of 671 Persian carpets would be greater than 3%.

I'm assuming you can still use normalcdf but I'm not grasping how to get the values. Thanks in advance!