What is the clearest way to differentiate between a binomial probability problem and a two outcome sample distribution problem that should be solved on the basis of a normal distribution?