You are on page 1of 31

PRIORITY RANKING

____a. money and gadgets


____b. education
____c. family
____d. friends
____e. popularity in social media
____f. girlfriend/boyfriend
SPEARMAN RANK
CORRELATION COEFFICIENT
SPEARMAN RANK CORRELATION COEFFICIENT

o also called as Spearman’s rho, in honor of the


psychologist Charles Edward Spearman who
developed the formula.

o a statistical tool that measures the strength of


association between two categorical variables on the
ordinal scale.
FORMULA: Spearman’s rho

σ 2
6 𝑑
𝜌 =1− 2
𝑛(𝑛 − 1)
where:
n = number of persons or objects being ranked
𝒙 = rank of person/object with respect to first variable
𝒚 = rank of person/object with respect to second variable
d = difference of ranks (d = x - y)
Spearman’s rho Interpretation
Note: The value correlation coefficient (𝜌) ranges from −1 to
+1.
𝜌 (rho) Verbal Interpretation
±0.01 - ±0.20 slight correlation
±0.21 - ±0.40 low correlation
±0.41 - ±0.60 moderate correlation
±0.61 - ±0.80 high correlation
±0.81 - ±0.99 very high correlation
Suppose that the top 5 high school students of a
graduating batch have been ranked based on
standardized test scores. The school head wants to
know if the batch performance based on grades and
extra-curricular agree with the standardized test
ranking?
Student Batch Performance(x) Test Rank (y)

A 99.87 1

B 96.75 5

C 97.50 4

D 99.33 3

E 98.00 2
EXAMPLE 1

In a study of the relationship between age and the


electroencephalogram (EEG), data were collected on
20 subjects between ages 20 and 60 years. The table
below shows the age and particular EEG (ordinal
pattern frequency) output value for each of the 20
subjects. The investigator wishes to know if it can be
concluded that this particular EEG output is inversely
correlated with age.
EXAMPLE 1
EEG EEG
Subject Age (x) Output Subject Age (x) Output
Value (y) Value (y)
A 20 98 K 40 68
B 21 75 L 42 66
C 22 95 M 44 71
D 22 100 N 46 62
E 27 99 O 48 69
F 30 65 P 51 54
G 31 64 Q 55 63
H 33 70 R 55 52
I 35 85 S 55 67
J 38 74 T 60 55
EXAMPLE 1: SOLUTION

STEP 1: Hypothesis

𝑯𝒐 : There is no significant relationship between the age and


the EEG output value. (𝜌 = 0)

𝑯𝒂 : There is a significant relationship between the age and


the EEG output value. (𝜌 < 0)
EXAMPLE 1: SOLUTION

STEP 2: Level of Significance


𝛼 = 0.05
STEP 3: Critical Value (one-tailed)
𝑑𝑓 = 20 − 2
= 18
𝒕𝒂 = −𝟎. 𝟑𝟗𝟗
EXAMPLE 1: SOLUTION

STEP 4: Compute the correlation coefficient (𝜌)

6σ𝑑 2
𝜌 =1−
𝑛(𝑛2 − 1)
EXAMPLE 1
EEG EEG
Subject Age (x) Output Subject Age (x) Output
Value (y) Value (y)
A 20 98 K 40 68
B 21 75 L 42 66
C 22 95 M 44 71
D 22 100 N 46 62
E 27 99 O 48 69
F 30 65 P 51 54
G 31 64 Q 55 63
H 33 70 R 55 52
I 35 85 S 55 67
J 38 74 T 60 55
EXAMPLE 1
Subject x-rank y-rank Subject x-rank y-rank
A 20 3 K 10 11
B 19 6 L 9 13
C 17.5 4 M 8 8
D 17.5 1 N 7 17
E 16 2 O 6 10
F 15 14 P 5 19
G 14 15 Q 3 16
H 13 9 R 3 20
I 12 5 S 3 12
J 11 7 T 1 18
EXAMPLE 1
Subject x-rank y-rank d 𝐝𝟐
A 20 3 17 289
B 19 6 13 169
C 17.5 4 13.5 182.25
D 17.5 1 16.5 272.25
E 16 2 14 196
F 15 14 1 1
G 14 15 -1 1
H 13 9 4 16
I 12 5 7 49
J 11 7 4 16 continuation in the next slide
EXAMPLE 1
Subject x-rank y-rank d 𝐝𝟐
K 10 11 -1 1
L 9 13 -4 16
M 8 8 0 0
N 7 17 -10 100
O 6 10 -4 16 σ 𝑑 2 = 2 348.50
P 5 19 -14 196
Q 3 16 -13 169
R 3 20 -17 289
S 3 12 -9 81
T 1 18 -17 289
EXAMPLE 1: SOLUTION
STEP 4: Compute the correlation coefficient (𝜌)
σ 2
6 𝑑
𝜌 =1−
𝑛(𝑛2 − 1)
6 2 348.50
𝜌 =1−
20 202 − 1
𝜌 ≈ - 0. 766 High negative correlation
EXAMPLE 1: SOLUTION

STEP 5: Make a Decision


Since the value correlation coefficient (𝜌 ≈ -0.766)
is less than the critical value (𝒕𝒂 = −𝟎. 𝟑𝟗𝟗),
therefore we reject the null hypothesis.
EXAMPLE 1: SOLUTION

STEP 6: Interpretation
At 5% level of significance, there is enough
evidence to claim that there is a significant
relationship between the age and the EEG output
value.
EXAMPLE 2

Nozawa et al. (A-10) calculated the Japanese


Orthopaedic Association score for measuring back
pain (JOA). The results for the 20 subjects along with
the duration of follow-up are shown in the following
table. The higher the number, the lesser the degree
of pain.
EXAMPLE 2
Duration of Duration of
Subject Follow-Up JOA Score Subject Follow-Up JOA Score
(Months) (Months)
A 103 21 K 38 13
B 68 14 L 36 24
C 62 26 M 34 21
D 60 24 N 30 22
E 60 13 O 19 25
F 54 24 P 19 23
G 49 22 Q 19 20
H 44 23 R 19 21
I 42 18 S 17 25
J 42 24 T 16 21
EXAMPLE 2: SOLUTION

STEP 1: Hypothesis

𝑯𝒐 : There is no significant relationship between the duration


of follow-up and JOA score. (𝜌 = 0)

𝑯𝒂 : There is a significant relationship between the duration


of follow-up and JOA score. (𝜌 ≠ 0)
EXAMPLE 2: SOLUTION

STEP 2: Level of Significance


𝛼 = 0.05
STEP 3: Critical Value (two-tailed)
𝑑𝑓 = 20 − 2
= 18
𝒕𝒂 = ± 𝟎. 𝟑𝟗𝟗
EXAMPLE 2: SOLUTION

STEP 4: Compute the correlation coefficient (𝜌)

6σ𝑑 2
𝜌 =1−
𝑛(𝑛2 − 1)
EXAMPLE 2
Duration of Duration of
Subject Follow-Up JOA Score Subject Follow-Up JOA Score
(Months) (Months)
A 103 21 K 38 13
B 68 14 L 36 24
C 62 26 M 34 21
D 60 24 N 30 22
E 60 13 O 19 25
F 54 24 P 19 23
G 49 22 Q 19 20
H 44 23 R 19 21
I 42 18 S 17 25
J 42 24 T 16 21
EXAMPLE 2
Subject x-rank y-rank Subject x-rank y-rank
A 1 13.5 K 11 19.5
B 2 18 L 12 5.5
C 3 1 M 13 13.5
D 4.5 5.5 N 14 10.5
E 4.5 19.5 O 16.5 2.5
F 6 5.5 P 16.5 8.5
G 7 10.5 Q 16.5 16
H 8 8.5 R 16.5 13.5
I 9.5 17 S 19 2.5
J 9.5 5.5 T 20 13.5
EXAMPLE 2
Subject x-rank y-rank d 𝐝𝟐
A 1 13.5 -12.5 156.25
B 2 18 -16 256
C 3 1 2 4
D 4.5 5.5 -1 1
E 4.5 19.5 -15 225
F 6 5.5 0.5 0.25
G 7 10.5 -3.5 12.25
H 8 8.5 -0.5 0.25
I 9.5 17 -7.5 56.25
J 9.5 5.5 4 16 continuation in the next slide
EXAMPLE 2
Subject x-rank y-rank d 𝐝𝟐
K 11 19.5 -8.5 72.25
L 12 5.5 6.5 42.25
M 13 13.5 -0.5 0.25
N 14 10.5 3.5 12.25
O 16.5 2.5 14 196 σ 𝑑 2 = 1 438
P 16.5 8.5 8 64
Q 16.5 16 0.5 0.25
R 16.5 13.5 3 9
S 19 2.5 16.5 272.25
T 20 13.5 6.5 42.25
EXAMPLE 2: SOLUTION
STEP 4: Compute the correlation coefficient (𝜌)
σ 2
6 𝑑
𝜌 =1−
𝑛(𝑛2 − 1)
6 1 438
𝜌 =1−
20 202 − 1
𝜌 ≈ - 0. 081 Slight negative correlation
EXAMPLE 2: SOLUTION

STEP 5: Make a Decision


Since the value correlation coefficient (𝜌 ≈ -0.081)
is greater than the critical value (𝒕𝒂 = −𝟎. 𝟑𝟗𝟗),
therefore we fail reject the null hypothesis.
EXAMPLE 2: SOLUTION

STEP 6: Interpretation
At 5% level of significance, there is enough
evidence to claim that there is no significant
relationship between the duration of follow-up
and JOA score.

You might also like