18.25. Telephone communications. A management consultant was engaged by a firm to improve the cost-effectiveness of its communications. A~ part of the study. the consultant selected 10 home-office executives at random from each of the (!) sales, (2) production. and (3) research and development divisions, and studied the communications of these executives during the past 10 weeks in great detail. Among other data, the consultant obtained the following infOlmation on weekly dollar costs of long-distance telephone calls to branch offices by the executives: j 2 3 4 5 6 7 8 9 10 666 920 495 602 1,499 960 796 343 894 813 2 488 362 156 546 216 542 345 291 516 126 3 391 450 609 910 705 472 645 496 763 1,309 The consultant decided to employ a nonparametric approach to test whether or not the mean telephone expenses for the three divisions are equal. a. What feature of the data may have suggested the use of a nonparametric test? b. Conduct the nonparametric rank F test, controlling the risk of Type I elTor at a == .05. State the alternatives, decision rule. and conclusion. What is the P-value of the test? c. Conduct multiple pairwise tests based on the ranked data to group the three divisions according to mean telephone expenditures; use family level of significance a == .05. Describe your findings.