If the confidence interval and hypothesis test both used the same data, how did they come up with different standard errors