1

UNIT 7 RANK CORRELATION Rank Correlation

Structure
7.1 Introduction
Objectives
7.2 Concept of Rank Correlation
7.3 Derivation of Rank Correlation Coefficient Formula
7.4 Tied or Repeated Ranks
7.5 Concurrent Deviation
7.6 Summary
7.7 Solutions / Answers

7.1 INTRODUCTION
In second unit of this block, we have discussed the correlation with its
properties and also the calculation of correlation coefficient. In correlation
coefficient or product moment correlation coefficient, it is assumed that both
characteristics are measurable. Sometimes characteristics are not measurable
but ranks may be given to individuals according to their qualities. In such
situations rank correlation is used to know the association between two
characteristics. In this unit, we will discuss the rank correlation and calculation
of rank correlation coefficient with its merits and demerits. We will also study
the method of concurrent deviation.

In Section 7.2, you will know the concept of rank correlation while Section 7.3
gives the derivation of Spearman’s rank correlation coefficient formula. Merits
and demerits of the rank correlation coefficient are discussed in Sub-section
7.3.1. There might be a situation when two items get same rank. This situation is
called tied or repeated rank which is described in Section 7.4. You will learn the
method of concurrent deviation in Section 7.5.

Objectives
After reading this unit, you would be able to
• explain the concept of rank correlation;
• derive the Spearman’s rank correlation coefficient formula;
• describe the merits and demerits of rank correlation coefficient;
• calculate the rank correlation coefficient in case of tied or repeated ranks;
and
• describe the method of concurrent deviation.

7.2 CONCEPT OF RANK CORRELATION

For the calculation of product moment correlation coefficient characters must
be measurable. In many practical situations, characters are not measurable.
They are qualitative characteristics and individuals or items can be ranked in
 2

Correlation for Bivariate Data order of their merits. This type of situation occurs when we deal with the
qualitative study such as honesty, beauty, voice, etc. For example, contestants
of a singing competition may be ranked by judge according to their
performance. In another example, students may be ranked in different subjects
according to their performance in tests.
Arrangement of individuals or items in order of merit or proficiency in the
possession of a certain characteristic is called ranking and the number
indicating the position of individuals or items is known as rank.
If ranks of individuals or items are available for two characteristics then
correlation between ranks of these two characteristics is known as rank
correlation.
With the help of rank correlation, we find the association between two
qualitative characteristics. As we know that the Karl Pearson’s correlation
coefficient gives the intensity of linear relationship between two variables and
Spearman’s rank correlation coefficient gives the concentration of association
between two qualitative characteristics. In fact Spearman’s rank correlation
coefficient measures the strength of association between two ranked variables.
Derivation of the Spearman’s rank correlation coefficient formula is discussed
in the following section.

7.3 DERIVATION OF RANK CORRELATION

COEFFICIENT FORMULA
Suppose we have a group of n individuals and let x1, x 2 ,..., xn and

y1 , y2 ,..., yn be the ranks of n individuals in characteristics A and B

respectively.
It is assumed that no two or more individuals have the same rank in either
characteristics A or B. Suppose both characteristics X and Y are taking rank
values 1, 2, 3,…, n. Then sum of ranks of characteristics A is
n

x
i=1
i = x1 + x 2 + ... + x n

= 1 + 2 + ... + n (since X is taking values 1, 2,…,n)

n
n(n + 1)
x
i=1
i
=
2 …(1)

From, the formula of sum of n natural numbers.

Mean of variable X is
1 n
x =  x = 1 n(n + 1)
i
n i=1 n 2
(n + 1)
x=
2
Since both variables are taking same values 1, 2, …, n then
(n + 1)
x=y=
 3

2
 4

If variance of X is denoted by  then

2
x Rank Correlation
n
1
n
 =
2
x (xi − x)2
i=1

 (x
1
 = + x 2 − 2x x)
2 2
x i i
n i=1
n
1 n n
 =
x + 
2
( x − 2x  xi )
2 2
x i
n i=1 i=1 i=1

n
1
 =
x
2
x
(
i
2
+ nx 2 − 2nx 2 )
n i=1

n
1
 =
x
2
x
(
i
2
− nx 2 )
n i=1

x
1 1
 = − x2 = (x 2 + x 2 + ... + x 2 ) − x 2
2 2
… (2)
x i

n i=1 n 1 2 n

Substituting the value of x in equation (2), we have

2
1 n + 1
x =
2
(1 + 2 + ... + n ) − 
2 2 2
 (Since X is taking values 1,2,…,n)
n  2 
2
1 n(n + 1)(2n + 1)  n + 1
 =  − 
2

 2 
x
n 6

(From the formula of sum of squares of n natural numbers)

(n + 1)(2n + 1) (n + 1)
2

x
2
= −
6 4

 2n + 1 (n + 1) 
 x = (n + 1) −
2

6 4
 
2(2n + 1) − 3(n + 1) 
 x = (n + 1) 
2

 12 
4n + 2 − 3n − 3
 x = (n + 1)
2

12
 n − 1 
 x = (n + 1)
2

 12 
2 n −1
2

σx = (from the formula (a - b)(a + b) = a 2 − b 2 )

12
Since both variables X and Y are taking same values, they will have same
variance, thus
2 2 n 2 −1
σ y = σx =
12
5
 6

Let di be the difference of the ranks of ith individual in two characteristics,

Correlation for Bivariate Data
then

di = xi − yi
di = xi − yi − x + y Since x = y
di = (xi − x) − (yi − y)

Squaring and summing d2 iover i = 1 to n, we have

n n

d =  (x i − x) − (yi − y)

2 2
i

i=1 i=1

(x − x) + (y 
n n

 d i=  − y)2 − 2(x i − x)(yi − y)

2 2
 i
i=1 i=1
i

n n n n

  d =  (x i − x) +  (y i − y) − 2 (x i − x)(yi − y)
2
i
2 2
… (3)

i=1 i=1 i=1 i=1

Dividing equation (3) by n, we have

1 n 2 1 n 1 n 1 n
 d =  (x −
i
x) 2
+  (y − y) 2
− 2  (x − x)(y − y)
i i i i

n i=1 n i=1 n i=1 n i=1

1 n d2 = 2 + 2 − 2Cov(x, y)

n
 i … (4)

x y
i=1
Cov(x, y)
We know that, r = , which implies that Cov(x, y) = r  .
xy
xy
Substituting Cov(x, y) = rxy in equation (4), we have

1 n d2 = 2 + 2 − 2r 
n
 i
x y x y
i=1
Since,  =  , then
2 2
x y

1 n d2 = 2 + 2 − 2r 
n
 i x x x x
i=1
n
1
d = 22 − 2r2
2

i x x
n i=1
n
1
d = 22 (1 − r)
2
i x

n i=1
n
1 2
 7

2n 2
d i
= (1− r)
x i=1

d
2
i

r = 1 − i=1 2
2nx
 8

6d i
2 Rank Correlation
n 2− 1
r = 1 − i=1 2 (Since x = )
n(n −1) 2 12

We denote rank correlation coefficient by rs , and hence

n

6d i
2


rs = 1 − i=1 2 … (5)
n(n −1)

This formula was given by Spearman and hence it is known as Spearman’s

rank correlation coefficient formula.
Let us discuss some problems on rank correlation coefficient.
Example 1: Suppose we have ranks of 8 students of B.Sc. in Statistics and
Mathematics. On the basis of rank we would like to know that to what extent
the knowledge of the student in Statistics and Mathematics is related.
Rank in Statistics 1 2 3 4 5 6 7 8
Rank in Mathematics 2 4 1 5 3 8 7 6

Solution: Spearman’s rank correlation coefficient formula is

6  d i2
n

rs = 1 − i =1

n ( n 2 − 1)
Let us denote the rank of students in Statistics by R x and rank in Mathematics
by R y . For the calculation of rank correlation coefficient we have to find
n

d
i=1
2
i which is obtained through the following table:

Difference of
Rank in Rank in
Ranks
Statistics Mathematics
(R ) (d = R − R ) d2i
(R x ) y
i x y

1 2 −1 1
2 4 −2 4
3 1 2 4
4 5 −1 1
5 3 2 4
6 8 −2 4
7 7 0 0
8 6 2 4
d 2
i
= 22

Here, n = number of paired

i observations = 8
n(n 2 22 63
6 d 2 − 1) 6 
8
rs = 1 − = 1− = 1−
 9

132 =
372 =
504 504 0.74
 10

Correlation for Bivariate Data Thus there is a positive association between ranks of Statistics and
Mathematics.
Example 2: Suppose we have ranks of 5 students in three subjects Computer,
Physics and Statistics and we want to test which two subjects have the same
trend.
Rank in Computer 2 4 5 1 3
Rank in Physics 5 1 2 3 4
Rank in Statistics 2 3 5 4 1

Solution: In this problem, we want to see which two subjects have same
trend i.e. which two subjects have the positive rank correlation coefficient.
Here we have to calculate three rank correlation coefficients
r12s = Rank correlation coefficient between the ranks of Computer and Physics
r23s = Rank correlation coefficient between the ranks of Physics and Statistics

r13s = Rank correlation coefficient between the ranks of Computer and

Statistics
Let R1 , R 2 and R3 be the ranks of students in Computer, Physics and
Statistics respectively.
Rank in Rank in Rank in d12 = d1 22 d23= d2 32 d13= d1 32
Computer Physics Statistics
R1−R2 R2−R3 R1−R3
(R1) (R2) (R3)
2 5 2 −3 9 3 9 0 0
4 1 3 3 9 −2 4 1 1
5 2 5 3 9 −3 9 0 0
1 3 4 −2 4 −1 1 −3 9
3 4 1 −1 1 −3 9 2 4
Total 32 32 14
Thus,

d 2
12 =32, d 2
=32 and d 2
=14.
23 13
Now

6 d12
2
6  32 = 1 − 8 = − 3
r12s = 1 − =1− = −0.6
n(n − 1)
2
5  24 5 5

6 d 223 6  32 = 1 − 8 = − 3
r23s = 1 − =1− = −0.6
n(n − 1)
2
5  24 5 5

6 d 13
2
6  14 = 1 − 7 = 3 = 0.3
r13s = 1 − = 1 −
n(n 2 −1) 5  24 10 10

r12s is negative which indicates that Computer and Physics have opposite
 11

trend. Similarly, negative rank correlation coefficient r23s shows the opposite
 12

trend in Physics and Statistics. r13s = 0.3 indicates that Computer and Statistics Rank Correlation
have same trend.
Sometimes we do not have rank but actual values of variables are available. If
we are interested in rank correlation coefficient, we find ranks from the given
values. Considering this case we are taking a problem and try to solve it.
Example 3: Calculate rank correlation coefficient from the following data:
x 78 89 97 69 59 79 68
y 125 137 156 112 107 136 124

Solution: We have some calculation in the following table:

x y Rank of x Rank of y d = Rx-Ry d2
(Rx) (Ry)
78 125 4 4 0 0
89 137 2 2 0 0
97 156 1 1 0 0
9 112 5 6 -1 1
59 107 7 7 0 0
79 136 3 3 0 0
68 124 6 5 1 1
n

d 2
i
=2
i=1

Spearman’s Rank correlation formula is

n

6d i
2

rs = 1 − i=1 2
n(n −1)
62 12
rs = 1 − = 1−
7(49 −1) 7  48

1 27
=1− = = 0.96
28 28

Now, let us solve a little exercise.

E1) Calculate Spearman’s rank correlation coefficient from the following
data:
x 20 38 30 40 50 55
y 17 45 30 35 40 25

7.3.1 Merits and Demerits of Rank Correlation Coefficient

Merits of Rank Correlation Coefficient
1. Spearman’s rank correlation coefficient can be interpreted in the same way
as the Karl Pearson’s correlation coefficient;
2. It is easy to understand and easy to calculate;
 13

Correlation for Bivariate Data 3. If we want to see the association between qualitative characteristics, rank
correlation coefficient is the only formula;
4. Rank correlation coefficient is the non-parametric version of the Karl
Pearson’s product moment correlation coefficient; and
5. It does not require the assumption of the normality of the population from
which the sample observations are taken.
Demerits of Rank Correlation Coefficient
1. Product moment correlation coefficient can be calculated for bivariate
frequency distribution but rank correlation coefficient cannot be
calculated; and
2. If n >30, this formula is time consuming.

7.4 TIED OR REPEATED RANKS

In Section 7.3, it was assumed that two or more individuals or units do not
have same rank. But there might be a situation when two or more individuals
have same rank in one or both characteristics, then this situation is said to be
tied.
If two or more individuals have same value, in this case common ranks are
assigned to the repeated items. This common rank is the average of ranks they
would have received if there were no repetition. For example we have a series
50, 70, 80, 80, 85, 90 then 1st rank is assigned to 90 because it is the biggest
value then 2nd to 85, now there is a repetition of 80 twice. Since both values
are same so the same rank will be assigned which would be average of the
ranks that we would have assigned if there were no repetition. Thus, both 80
will receive the average of 3 and 4 i.e. (Average of 3 & 4 i.e. (3 + 4) / 2= 3.5)
3.5 then 5th rank is given to 70 and 6th rank to 50. Thus, the series and ranks of
items are

Series 50 70 80 80 85 90

Ranks 6 5 3.5 3.5 2 1

In the above example 80 was repeated twice. It may also happen that two or
more values are repeated twice or more than that.
For example, in the following series there is a repetition of 80 and 110. You
observe the values, assign ranks and check with following.

Series 50 70 80 90 80 120 110 110 110 100

Ranks 10 9 7.5 6 7.5 1 3 3 3 5

m(m2 −1)
When there is a repetition of ranks, a correction factor is added to
12

d 2
in the Spearman’s rank correlation coefficient formula, where m is the
number of times a rank is repeated. It is very important to know that this
correction factor is added for every repetition of rank in both characters.
 14

In the first example correction factor is added once which is 2(4-1)/12 = 0.5, Rank Correlation
while in the second example correction factors are 2(4-1)/12 = 0.5 and

3 (9-1)/12 = 2 which are aided to  d 2 .

Thus, in case of tied or repeated rank Spearman’s rank correlation coefficient

formula is
 
d 2 + m(m12− 1) + ...
2


6

r = 1−  
n(n − 1)
s 2

Example 4: Calculate rank correlation coefficient from the following data:

Expenditure on
10 15 14 25 14 14 20 22
advertisement
Profit 6 25 12 18 25 40 10 7

Solution: Let us denote the expenditure on advertisement by x and profit by y

x Rank of x y Rank of y d = Rx-Ry d2
(Rx) ( Ry )
10 8 6 8 0 0
15 4 25 2.5 1.5 2.25
14 6 12 5 1 1
25 1 18 4 −3 9
14 6 25 2.5 3.5 12.25
14 6 40 1 5 25
20 3 10 6 −3 9
22 2 7 7 −5 25
d 2
=83.50

 
d 2 + m(m12− 1) + ...
2
6

r = 1−  
n(n − 1)
s 2

Here rank 6 is repeated three times in rank of x and rank 2.5 is repeated twice
in rank of y, so the correction factor is
3(32 −1) 2(22 −1)
+
12 12
Hence rank correlation coefficient is
 3(32 − 1) 2(22 − 1) 
683.50 + + 12
12
r = 1−  
8(64 − 1)
s
 15
 3  8 2 3
6 83.50 + +
Correlation for Bivariate Data  
12 
rs = 1 − 
12
8 X 63

6(83.50 + 2.50)
rs = 1 −
504
516
r =1−
s
504
rs = 1−1.024 = − 0.024
There is a negative association between expenditure on advertisement and
profit.
Now, let us solve the following exercises.
E2) Calculate rank correlation coefficient from the following data:
10 20 30 30 40 45 50
y 15 20 25 30 40 40 40

E3) Calculate rank correlation coefficient from the following data:

x 70 70 80 80 80 90 100
y 90 90 90 80 70 60 50

7.5 CONCURRENT DEVIATION

Sometimes we are not interested in the actual amount of correlation coefficient
but we want to know the direction of change i.e. whether correlation is
positive or negative, coefficient of concurrent deviation serves our purpose. In
this method correlation is calculated between the direction of deviations and
not their magnitudes. Coefficient of concurrent deviation is denoted by rc and
given by
(2c − k)
rc =   … (6)

where, c is the number of concurrent deviation or the number of + sign in the

product of two deviations, k = n − 1 i.e. total number of paired observation
minus one. This is also called coefficient of correlation by concurrent
deviation method. Steps for the calculation of concurrent deviation (see the
Example 5 simultaneously) are:
1. The first value of series x is taken as a base and it is compared with next
value i.e. second value of series x. If second value is greater than first
value, ‘+’ sign is assigned in the Column titled Dx . If second value is less
than the first value then ‘-’ sign is assigned in the column Dx .
2. If first and second values are equal then ‘=’ sign is assigned.
3. Now second value is taken as base and it is compared with the third value
of the series. If third value is less than second ‘-’ is assigned against the
 16

third value. If the third value is greater than the second value ‘+’ is Rank Correlation
assigned. If second and third values or equal than ‘=’ sign is assigned.
4. This procedure is repeated upto the last value of the series.
5. Similarly, we obtain column D y for series y.

6. We multiply the column D x and Dy and obtain column D x Dy .

Multiplication of same sign results ‘+’sign and that of different sign is ‘-’
sign.
7. Finally number of ‘+’ sign are counted in the column D x Dy , it is called c
and we get coefficient concurrent deviation by the formula (6).
8. In the formula, inside and outside the square root, sign ‘+’ and ‘-’ depends
on the value of ( 2c − k ). If this value is positive than ‘+’ sign is taken at
both places if ( 2c − k ) is negative ‘-’sign is considered at both the places.
Let us discuss the following problem.
Example 5: We have data of income and expenditure of 11 workers of an
organization in the following table:
Income 65 40 35 75 63 79 35 20 80 60 50
Expenditure 60 55 50 66 30 71 40 35 80 75 80

Find whether correlation is positive or negative by coefficient of concurrent

deviation.
Solution: Coefficient of concurrent deviation is given by

(2c − k)
rc =  

Let us denote the income by x and expenditure by y and we calculate c by the

following table:
x Change of direction y Change of direction DxDy
sign for x (Dx) sign for y (Dy)
65 60
40 − 55 − +
35 − 50 − +
75 + 66 + +
63 − 30 − +
79 + 71 + +
35 − 40 − +
20 − 35 − +
80 + 81 + +
60 − 75 − +
50 − 80 + −
c=9

Here, c = 9 and k = n −1 = 10 then we have

 17

2 9 −10 8
=+ +
Correlation for Bivariate Data
r = 
10 10
(Both signs are + because 2c − k is positive)
= 0.8 = 0.89
Thus correlation is positive.
Now, let us solve the following exercises.

E 4) Find the coefficient of correlation between supply and price by

concurrent deviation method for the following data:
Year 1970 1971 1972 1973 1974 1975
Supply 114 127 128 121 120 124
Price 108 104 105 106 100 99

E5) Calculate coefficient of concurrent deviation for the following data:

x 368 384 385 360 347 384
y 122 121 124 125 122 126

7.6 SUMMARY
In this unit, we have discussed:
1. The rank correlation which is used to see the association between two
qualitative characteristics;
2. Derivation of the Spearman’s rank correlation coefficient formula;
3. Calculation of rank correlation coefficient in different situations- (i) when
values of variables are given, (ii) when ranks of individuals in different
characteristics are given and (iii) when repeated ranks are given;
4. Properties of rank correlation coefficient; and
5. Concurrent deviation which provides the direction of correlation.

7.7 SOLUTIONS /ANSWERS

E1) We have some calculations in the following table:
x y Rank of x Rank of y d= d2
(Rx) (Ry) Rx-Ry
20 17 6 6 0 0
38 45 4 1 3 9
30 30 5 4 1 1
40 35 3 3 0 0
50 40 2 2 0 0
55 25 1 5 −4 16
n

d 2
i
=26
i=1
 18

6  26
rs = 1 − Rank Correlation
6(36 −1)
26 9
= 1− = = 0.26
35 35

E2) We have some calculations in the following table:

x Rank of x y Rank of y d = Rx-Ry d2
(Rx) (Ry)
10 7 15 7 0 0
20 6 20 6 0 0
30 4.5 25 5 −0.5 0.25
30 4.5 30 4 0.5 0.25
40 3 40 2 1 1
45 2 40 2 0 0
50 1 40 2 −1 1

d 2
= 2.5

 
d 2 + m(m12− 1) + ...
2
6

r = 1−  
n(n − 1)
s 2

Here, rank 4.5 is repeated twice in rank of x and rank 2 is repeated

thrice in rank of y so the correction factor is
2(22 −1) 3(32 −1)
+
12 12
and therefore, rank correlation coefficient is
 2(22 − 1) + 3(32 − 1) 
62.5 + 12
12
r = 1−  
7(49 − 1)
s

2  3 3  8 
62.5 + +
 12 12 
rs = 1 −
7  48

6(2.5 + 2.5)
rs = 1 −
336
30 306
rs = 1 − =
336 336

rs = 0.91
E3) We have some calculations in the following table:
19
 20

Correlation for Bivariate Data x Rank of x y Rank of y d = Rx-Ry d2

(Rx) (Ry)
70 6.5 90 2 4.5 20.25
70 6.5 90 2 4.5 20.25
80 4 90 2 2 4
80 4 80 4 0 0
80 4 70 5 −1 1
90 2 60 6 −4 16
100 1 50 7 −6 36

d 2
= 97.5

Rank correlation coefficient is

 
d 2 m(m12− 1) + ...
2

 +
6

r = 1−  
n(n − 1)
s 2

Here, rank 4 and 6.5 is repeated thrice and twice respectively in rank of
x and rank 2 is repeated thrice in rank of y, so the correction factor is
3(32 −1) 2(22 − 1) 3(32 −1)
+ +
12 12 12
and therefore, rank correlation2 coefficient2 is
 3(3 − 1) 2(2 − 1) 3(32 − 1) 
+ 
697.5 + + 12
12 12
r = 1−  
7(49 − 1)
s

697.5 + 4.5
rs = 1 −
7  48
6(102)
rs = 1 −
336
102
r=1− = −0.82
s
56
E4) Coefficient of concurrent deviation is given
(2c − k)
rc =  
Let us denote the supply by x and price by y and we calculate c by the
following table:
x Change of Direction y Change of Direction DxDy
sign for x (Dx) sign for y (Dy)
114 108
127 + 104 − −
128 + 105 + +
121 − 106 + −
120 − 100 − +
124 + 99 − −
c=2
 21

Now c = 2 and k = n -1 = 5 Rank Correlation

2 2 − 5
r=  =−
(Both signs are ‘-’ because 2c − k is negative)
= − 0.45

Thus, correlation is negative.

E5) Coefficient of concurrent deviation is given
(2c − k)
rc =  
Let us denote the supply by x and price by y and we calculate c by the
following table:

Change of Direction Change of Direction

x y DxDy
sign for x (Dx) sign for y (Dy)
368 122
384 + 121 − −
385 + 124 + +
360 − 125 + −
347 − 122 − +
384 + 126 + +
c=3
Now c = 3 and k = n −1= 5

2 3 − 5
r =  =

(Both signs are ‘+’ because 2c−k is positive)

r = 0.45
Thus, correlation is positive.