In: Statistics and Probability
In a study of the effectiveness of physical exercise in weight
reduction, 12 subjects followed a program of physical exercise for
two months. Their weights (in pounds) before and after this program
are shown in the table. Use Wilcoxon's signed-ranks test and a
significance level of 0.05 to test the claim that the exercise
program has no effect on weight.
Before | 162 | 190 | 188 | 152 | 148 | 127 | 195 | 164 | 175 | 156 | 180 | 136 |
After | 157 | 194 | 179 | 149 | 135 | 130 | 183 | 168 | 168 | 148 | 170 | 138 |
What would be the signed rank for the column with values of 175 and
168?
8.5 |
||
10 |
||
8 |
||
9 |
The following table is obtained:
Pair | Sample 1 | Sample 2 | Difference | Abs. Difference | Sign |
1 | 162 | 157 | 5 | 5 | +1 |
2 | 190 | 194 | -4 | 4 | -1 |
3 | 188 | 179 | 9 | 9 | +1 |
4 | 152 | 149 | 3 | 3 | +1 |
5 | 148 | 135 | 13 | 13 | +1 |
6 | 127 | 130 | -3 | 3 | -1 |
7 | 195 | 183 | 12 | 12 | +1 |
8 | 164 | 168 | -4 | 4 | -1 |
9 | 175 | 168 | 7 | 7 | +1 |
10 | 156 | 148 | 8 | 8 | +1 |
11 | 180 | 170 | 10 | 10 | +1 |
12 | 136 | 138 | -2 | 2 | -1 |
Now, the following table is obtained by removing the ties and organizing the absolute differences in ascending order:
Pair | Sample 1 | Sample 2 | Difference | Abs. Difference | Sign |
12 | 136 | 138 | -2 | 2 | -1 |
4 | 152 | 149 | 3 | 3 | +1 |
6 | 127 | 130 | -3 | 3 | -1 |
2 | 190 | 194 | -4 | 4 | -1 |
8 | 164 | 168 | -4 | 4 | -1 |
1 | 162 | 157 | 5 | 5 | +1 |
9 | 175 | 168 | 7 | 7 | +1 |
10 | 156 | 148 | 8 | 8 | +1 |
3 | 188 | 179 | 9 | 9 | +1 |
11 | 180 | 170 | 10 | 10 | +1 |
7 | 195 | 183 | 12 | 12 | +1 |
5 | 148 | 135 | 13 | 13 | +1 |
Now that the absolute differences are in ascending order, we assign ranks to them, taking care of assigning the average rank to values with rank ties (same absolute value difference)
Pair | Sample 1 | Sample 2 | Abs. Difference | Rank | Sign |
12 | 136 | 138 | 2 | 1 | -1 |
4 | 152 | 149 | 3 | 2.5 | +1 |
6 | 127 | 130 | 3 | 2.5 | -1 |
2 | 190 | 194 | 4 | 4.5 | -1 |
8 | 164 | 168 | 4 | 4.5 | -1 |
1 | 162 | 157 | 5 | 6 | +1 |
9 | 175 | 168 | 7 | 7 | +1 |
10 | 156 | 148 | 8 | 8 | +1 |
3 | 188 | 179 | 9 | 9 | +1 |
11 | 180 | 170 | 10 | 10 | +1 |
7 | 195 | 183 | 12 | 11 | +1 |
5 | 148 | 135 | 13 | 12 | +1 |
The sum of positive ranks is:
W+ = 2.5+6+7+8+9+10+11+12 = 65.5
and the sum of negative ranks is:
W- = 1+2.5+4.5+4.5 = 12.5
Hence, the test statistic is
.
(1) Null and Alternative Hypotheses
The following null and alternative hypotheses need to be tested:
H0: Median (Difference) = 0
Ha: Median (Difference) ≠ 0
(2) Rejection Region
The critical value for the significance level α=0.05 provided, and the type of tail specified is , and the null hypothesis is rejected if
(3) Decision about the null hypothesis
Since in this case T=12.5≤13, there is enough evidence to claim that the population median of differences is different than 0, at the α=0.05 significance level.
Please let me know in comments if anything is unclear. Will reply ASAP. Please upvote if satisfied!