In utilizing hypothesis testing, BWI’s ad campaign that states the average airfares from BWI are lower than averages fares from Dulles. The average (one-way) fares from Dulles over comparable routes is claimed to be $165. If there is a 10% level of significance with over 100 degrees of freedom, there is evidence that the mean sale price is actually not less than those at Dulles. In Figure 1, the t score equals -1.47. Therefore, the null must be rejected because the absolute value of T (1.47) is greater than T of alpha (1.28) for the upper-tail test. The Dulles average fare of $165 is greater than the average fares of BWI of $158.33 (from Figure 2). Average fares at BWI are actually less than the fares at Dulles. Thus, because the null is rejected, there would be a Type I error. If the null is rejected, but BWI accepts the null as true, then the Type I error would prove problematic.
The univariate estimate of fares in question previous aforementioned in the above statement, also known as the point estimate method, can be improved by using another common type of univariate estimation technique. In the point-estimate approach, we randomly choose a fare that would fall under the range most of the fares in the data set fell under. Yet, the interval estimate method is more accurate to use because point-estimate is technically impossible because probability at a specific point is always zero. So, the interval estimates for the average fare with a 90% confidence level would be