Mann-Whitney U Test Both spring-cut strips (2 & 4) had a lower % cover of Bluebells than the
autumn-cut strips (1 and 3). This difference was most evident in strip 4. However, the
difference in % Bluebell cover between strip 2 and strips 1 and 3 was not particularly
marked. It is really necessary to test the data statistically to determine whether the
observed differences are significant or not, before drawing any firm conclusions.
In order to determine whether the differences are
statistically significant, a Mann-Whitney U Test has been performed on the data.
Calculations and Results
Comparison between results obtained for sections 1 (autumn cut) and 2 (spring cut)
Data arranged in order and ranked
The data on % Bluebell cover from both section 1 (orange) and 2 (green) is arranged in ascending order of % Bluebell cover, from
lowest to highest and then ranked.
Sample |
5 |
5 |
. |
. |
10 |
. |
. |
. |
. |
. |
. |
20 |
20 |
. |
. |
. |
. |
. |
. |
. |
. |
Rank |
1.5 |
1.5 |
. |
. |
6 |
. |
. |
. |
. |
. |
. |
15 |
15 |
. |
. |
. |
. |
. |
. |
. |
. |
Sample |
. |
. |
8 |
8 |
. |
10 |
10 |
15 |
15 |
15 |
15 |
. |
. |
20 |
20 |
20 |
20 |
20 |
25 |
25 |
25 |
Rank |
. |
. |
3.5 |
3.5 |
. |
6 |
6 |
9.5 |
9.5 |
9.5 |
9.5 |
. |
. |
15 |
15 |
15 |
15 |
15 |
20.5 |
20.5 |
20.5 |
Sample |
. |
30 |
. |
45 |
50 |
70 |
. |
. |
90 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
Sum |
Rank |
. |
23 |
. |
25 |
26 |
27 |
. |
. |
30 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
170 |
Sample |
25 |
. |
40 |
. |
. |
. |
80 |
85 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
Sum |
Rank |
20.5 |
. |
24 |
. |
. |
. |
28 |
29 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
295 |
R1 = 170
n1
= 10
R2
= 295 n2 =
20
To calculate values of U, where n1 and n2 are the
number of samples from each section.
U1 = n1 x n2 + ½
n2(n2 + 1 ) - R2
U1 = 10 x 20 + ½ 20(20 + 1) -
295
U1 = 200 + 210 - 295
U1 = 115
|
U2
= n1
x n2
+ ½ n1(n1 + 1 ) - R1
U2 = 10 x 20 + ½
10(10 + 1) - 170
U2 = 200 + 55 - 170
U2 = 85 |
The smallest figure of U obtained in the calculations is then
compared with the critical value of U for
n1 = 10 and n2 = 20. This critical value is looked up in a table of critical values for the
Mann Whitney U Test. The critical value in this case is 55.
If the smallest figure of U obtained is less than, or equal to this
critical value, then the results are significantly different. The smallest value (U2)
was 85, which is larger than this critical value. The observed difference between
the data for % Bluebell cover between Section 1 and 2 is therefore not significant at the
p = 0.05 level.
Comparison between results obtained for sections
2 (spring cut) and 3
(autumn cut)
Sample |
8 |
8 |
. |
. |
. |
. |
10 |
10 |
. |
15 |
15 |
15 |
15 |
. |
20 |
20 |
20 |
20 |
20 |
. |
.. |
Rank |
2 |
2 |
. |
. |
. |
. |
6.5 |
6.5 |
. |
11.5 |
11.5 |
11.5 |
11.5 |
. |
16.5 |
16.5 |
16.5 |
16.5 |
16.5 |
. |
.. |
Sample |
. |
. |
8 |
9 |
10 |
10 |
. |
. |
13 |
. |
. |
. |
. |
20 |
. |
. |
. |
. |
. |
24 |
25 |
Rank |
. |
. |
2 |
4 |
6.5 |
6.5 |
. |
. |
9 |
. |
. |
. |
. |
16.5 |
. |
. |
. |
. |
. |
20 |
23 |
Sample |
25 |
25 |
25 |
25 |
. |
. |
40 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
80 |
85 |
. |
Sum |
Rank |
23 |
23 |
23 |
23 |
. |
. |
27.5 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
38.5 |
40 |
. |
343.5 |
Sample |
. |
. |
. |
. |
27 |
40 |
. |
45 |
45 |
45 |
49 |
55 |
60 |
60 |
67 |
70 |
80 |
. |
. |
. |
Sum |
Rank |
. |
. |
. |
. |
26 |
27.5 |
. |
30 |
30 |
30 |
32 |
33 |
34.5 |
34.5 |
36 |
37 |
38.5 |
. |
. |
. |
476.5 |
R1
= 343.5 n1
= 20
R2
= 476.5 n2 =
20
U1 = n1 x n2 + ½
n2(n2 + 1 ) - R2
U1 = 20 x 20 + ½ 20(20 + 1) -
476.5
U1 = 400 + 210 - 476.5
U1 = 133.5 |
U2
= n1
x n2
+ ½ n1(n1 + 1 ) - R1
U2 = 20 x 20 + ½
20(20 + 1) - 343.5
U2 = 400 + 210 - 343.5
U2 = 266.5 |
The critical value of U for n1 = 20 and n2 = 20 is 127
Neither of the U values above are equal to or below this figure. The
differences between section 2 and section 3 are therefore not significant at the p = 0.05
level.
Comparison between results obtained for sections
3 (autumn cut) and 4 (spring cut)
Sample |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
8 |
9 |
10 |
10 |
13 |
. |
. |
Rank |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
14.5 |
16 |
17.5 |
17.5 |
19 |
. |
. |
Sample |
0 |
0 |
0 |
0 |
0 |
1 |
1 |
1 |
1 |
2 |
4 |
6 |
6 |
8 |
. |
. |
. |
. |
. |
15 |
. |
Rank |
3 |
3 |
3 |
3 |
3 |
7.5 |
7.5 |
7.5 |
7.5 |
10 |
11 |
12.5 |
12.5 |
14.5 |
. |
. |
. |
. |
. |
20 |
. |
Sample |
20 |
24 |
25 |
27 |
. |
40 |
45 |
45 |
45 |
49 |
55 |
60 |
60 |
. |
67 |
70 |
. |
80 |
. |
. |
Sum |
Rank |
21 |
22 |
23 |
24 |
. |
26 |
28 |
28 |
28 |
30 |
31 |
33 |
33 |
. |
35 |
36 |
. |
38 |
. |
. |
520.5 |
Sample |
. |
. |
. |
. |
30 |
. |
. |
. |
|
. |
. |
. |
. |
60 |
. |
. |
75 |
. |
. |
. |
Sum |
Rank |
. |
. |
. |
. |
25 |
. |
. |
.. |
|
. |
. |
. |
. |
33 |
. |
. |
37 |
. |
. |
. |
220.5 |
R1 =
520.5 n1 =
20
R2
= 220.5 n2
= 18
U1 = n1 x n2 + ½
n2(n2 + 1 ) - R2
U1 = 20 x 18 + ½ 18(18 + 1) -
220.5
U1 = 360 + 171 - 220.5
U1 = 310.5 |
U2
= n1
x n2
+ ½ n1(n1 + 1 ) - R1
U2 = 20 x 18 + ½
20(20 + 1) - 520.5
U2 = 360 + 210 - 520.5
U2 = 49.5 |
The critical value of U for n1 = 20 and n2 = 18 is 112
The lowest U value (U2) of 49.5 is well below this
figure. The differences between section 3 and section 4 are therefore significant at the p
= 0.05 level.
Comparison between results obtained for sections
1 (autumn cut) and 4 (spring cut)
Sample |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
5 |
5 |
. |
. |
. |
10 |
. |
20 |
20 |
30 |
Rank |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
12.5 |
12.5 |
. |
. |
. |
17 |
. |
19.5 |
19.5 |
21.5 |
Sample |
0 |
0 |
0 |
0 |
0 |
1 |
1 |
1 |
1 |
2 |
4 |
. |
. |
6 |
6. |
8. |
.. |
.15 |
.. |
. |
.. |
Rank |
3 |
3 |
3 |
3 |
3 |
7.5 |
7.5 |
7.5 |
7.5 |
10 |
11 |
. |
. |
14.5 |
14.5 |
16 |
. |
18 |
. |
. |
.. |
Sample |
. |
45 |
50 |
. |
70 |
. |
90 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
Sum |
Rank |
. |
23 |
24 |
. |
26 |
. |
28 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
203.5 |
Sample |
30 |
. |
. |
60 |
. |
75 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
.Sum |
Rank |
21.5 |
. |
. |
25 |
. |
27 |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
. |
202.5. |
R1 =
203.5 n1 =
10
R2
= 202.5 n2
= 18
U1 = n1 x n2 + ½
n2(n2 + 1 ) - R2
U1 = 10 x 18 + ½ 18(18 + 1) -
202.5
U1 = 180 + 171 - 202.5
U1 = 148.5 |
U2
= n1
x n2
+ ½ n1(n1 + 1 ) - R1
U2 = 10 x 18 + ½
10(10 + 1) - 203.5
U2 = 180 + 55 - 203.5
U2 = 31.5 |
The critical value of U for n1 = 10 and n2 = 18 is 48
The U2 value of 31.5 is below this figure. The
differences between section 1 and section 4 are therefore significant at the p = 0.05
level.
Summary of Results
% Bluebell Cover |
Section 1 (autumn) |
Section
2 (spring) |
Section
3 (autumn) |
Section 1 (autumn) |
|
Section
2 (spring) |
Section
3 (autumn) |
Section
4 (spring) |
Section
4 (spring) |
p = 0.05 level |
Not significant |
Not significant |
Significant |
Significant |
Click here to return to
the exercise answers |