In: Statistics and Probability
Question 4: The HR department of a large company needs to assign a newly hired director to either their marketing development or their marketing operations division. To help in their decision, they organize two discussion groups with randomly selected employees from each division. During the talks, the director lays out his marketing vision and employees ask questions relevant to their daily work. At the end, each employee has to rate the director on a scale from 1 to 10 (1=very bad; 10=very good). The HR department wants to know if the distribution of ratings of the marketing development employees is different among the employees of the two divisions. Ratings data for the two groups of employees are included in the Minitab file HR_Ratings.mtw and in the Excel file Assign2.xlsx under the HR_Ratings tab.
a) Examine the distributions of the ratings (show histograms) by the two groups of employees and explain why a non-parametric test is justified to perform the analysis.
b) Perform an appropriate non-parametric test using a 5% significance level to determine if the distribution of ratings of the marketing development employees is different than that of the marketing operations employees. Specify any assumptions and/or conditions you need to make to apply the test and state your hypothesis clearly. Show your manual calculations.
c) Use Minitab to perform the test in b) above and compare your results here is the data for HR_Ratings.mtw
Datafrom HR_Ratings.wtw below
Marketing Development Employees | Marketing Operations Employees |
8 | 9 |
7 | 8 |
6 | 7 |
2 | 8 |
5 | 10 |
8 | 9 |
7 | 6 |
3 |
a.
From the histogram it is observed that the data are not obtained from normal distribution. Hence here we can't assume normal distribution. So for answering the query that whether the distribution of ratings of the marketing development employees is same or different among the employees of the two divisions, we perform non parametric test.
b.
c. Minitab output:
Mann-Whitney Test and CI: Marketing Develo, Marketing Operat
N Median
Marketing Development Employees 8 6.500
Marketing Operations Employees 7 8.000
Point estimate for ETA1-ETA2 is -2.000
95.7 Percent CI for ETA1-ETA2 is (-5.000,0.001)
W = 45.5
Test of ETA1 = ETA2 vs ETA1 not = ETA2 is significant at
0.0372
The test is significant at 0.0346 (adjusted for ties)
Conclusion: Since p-value=0.0372<0.05 hence we reject null hypothesis and conclude that the distribution of ratings of the marketing development employees is different than that of the marketing operations employees.