In: Statistics and Probability
) A sample of 15 graduating students, 15 females and 5 males, was randomly selected. Their GPAs were recorded, and the ranks of these GPAs for the pooled sample are presented below. Test the following hypothesis: H0: There is no difference in GPAs between the genders? Use the Mann-Whitney test.
Males | Females |
11 | 1 |
12 | 2 |
3 | 13 |
14 | 4 |
6 | 5 |
15 | |
7 | |
8 | |
9 | |
10 |
Calculated test statistic:
Critical value of the test statistic:
Decision rule:
Decision:
GIVEN:
Sample size of males
Sample size of females
HYPOTHESIS:
The hypothesis is given by,
(That is, there is no significant difference in GPA's between males and females.)
(That is, there is significant difference in GPA's between males and females.)
LEVEL OF SIGNIFICANCE:
Let us assume significance level to be .
TEST STATISTIC:
where
CALCULATION:
Let us first arrange GPA's of males and females in ascending order separately.
Males | Females |
3 | 1 |
6 | 2 |
11 | 4 |
12 | 5 |
14 | 7 |
8 | |
9 | |
10 | |
13 | |
15 |
Now we should give combined ranks to both groups.
Males | Females | ||
3 | 3 | 1 | 1 |
6 | 6 | 2 | 2 |
11 | 11 | 4 | 4 |
12 | 12 | 5 | 5 |
14 | 14 | 7 | 7 |
8 | 8 | ||
9 | 9 | ||
10 | 10 | ||
13 | 13 | ||
15 | 15 | ||
Now
CRITICAL VALUE:
From the Mann Whitney U test critical value table, the two tailed (since ) critical value with and at significance level is .
DECISION RULE:
CONCLUSION:
Since the calculated test statistic value (19) is greater than the critical value (8), we fail to reject null hypothesis and conclude that there is no significant difference in GPA's between males and females.