If 95% and 98% confidence intervals were developed to estimate the true cost of an mp3 player with a known population standard deviation, what differences would they have?