The 2-Sample Mann-Whitney (or two-sample Wilcoxon rank sum test) is a nonparametric test of the equality of two population medians. It also calculates the corresponding point estimate and confidence interval. Its parametric counterpart is the Paired t test.
This procedure assumes that the samples are randomly and independently drawn from two populations that have the same shape.
|
Test |
Null Hypothesis, H0 |
Alternate Hypothesis, H1 |
|
One-tailed |
m1 = m2 |
m1 < m2 or m1 > m2 |
|
Two-tailed |
m1 = m2 |
m1 ¹ m2 |
where m1 and m2 are the population medians
1. Choose ProcessMA > Statistics > Nonparametric > 2-Sample Mann-Whitney.
2. In Sample 1, select the column containing the first sample.
3. In Sample 2, select the column containing the second sample.
4. Click OK.
Optional
5. In Alternate, select the type of test.
Note To select a column of data into a textbox, double-click on any of the column names shown in the list on the left of the dialog box while in the textbox.
Sample 1: Numeric.
Sample 2: Numeric.
The number of printers sold by two teams of salespeople was recorded for the previous month. You want to find out if there is a difference in the population median of sales between the teams.
1. Open worksheet ProcessMA > Tools > Data Files > Stat.xls.
2. Choose ProcessMA > Statistics > Nonparametric > 2-Sample Mann-Whitney.
3. In Sample 1, select W – Sales A.
4. In Sample 2, select X – Sales B.
5. In Alternate, select Not equals.
6. Click OK.
Interpretation
The median for both sales team A and B are 51 and 52 respectively. For a desired a = 0.05, since p = 0.7995 > a, we fail to reject H0. Therefore, we conclude that there is no significant evidence that the sales of the two teams are the different.