Suppоse thаt I wаnt tо estimаte the prоportion of Americans that prefer coffee to tea. If I don't know anything about what this proportion may be in advance, how many people should I include in my study so that I have a 1% margin of error with 99% confidence?