Newsweek once wrote [15] that the width of a confidence interval is inversely related to the sample size; for example, if a sample of 500 gives a confidence interval of plus or minus 5, then a sample of 2,500 would give a confidence interval of plus or minus 1. Explain the error in this argument.