Many people rely on local accountants to handle their tax work. A local television station, which prides itself on doing investigative reporting, decided to do a study to determine the extent to which tax errors are made by the accounting firms.

The station selected eight people to have their taxes figured at each of three accounting offices in its market area. The following data show the tax bills (in dollars) as figured by each of the three accounting offices:

Return Office 1 Office 2 Office 3 1 4,376.20 5,100.10 4,988.03 2 5,678.45 6,234.23 5,489.23 3 2,341.78 2,242.60 2,121.90 4 9,875.33 10,300.30 9,845.60 5 7,650.20 8,002.90 7,590.88 6 1,324.80 1,450.90 1,356.89 7 2,345.90 2,356.90 2,345.90 8 15,468.75 16,080.70 15,376.70

a. Discuss why this test was conducted as a randomized block design. Why did the station think it was important to have all three offices do the returns for each of the eight people?

b. Test to determine whether blocking was necessary in this situation. Use a significance level of 0.01.

State the null and alternative hypotheses.

c. Based on the sample data, can the station report statistical evidence that there is a difference in the mean taxes due on tax returns? Test using a significance level of 0.01. State the appropriate null and alternative hypotheses.

d. Referring to part

c, if you did conclude that a difference exists, use the appropriate test to determine which office has the highest mean tax due.