PARTIAL CORRELATION

when we measure correlation between a dependent variable and one particular independent
variable by keeping all other independent variables constant, it is called partial correlation.
Partial Correlation Coefficient of First Order:
Partial correlation coefficient between 𝑋1 and 𝑋2 by keeping 𝑋3 as constant = 𝑟12.3
Partial correlation coefficient between 𝑋1 and 𝑋3 by keeping 𝑋2 as constant = 𝑟13.2
Partial correlation coefficient between 𝑋2 and 𝑋3 by keeping 𝑋1 as constant = 𝑟23.1

Formula
𝑟12 − 𝑟13 𝑟23
𝑟12.3 =
2 ) (1 2)
√(1 − 𝑟13 − 𝑟23
𝑟13 − 𝑟12 𝑟23
𝑟13.2 =
2 ) (1 2)
√(1 − 𝑟12 − 𝑟23
𝑟23 − 𝑟12 𝑟13
𝑟23.1 =
2 ) (1 2)
√(1 − 𝑟12 − 𝑟13
Partial correlation Coefficient of Second Order: It is used in case of four variables, wherein
correlation coefficient is calculated between two variables by keeping two other variables as
constant. For example, if correlation coefficient is calculated between 𝑋1 and 𝑋2 by keeping
𝑋3 and 𝑋4 as constant, it will be denoted as 𝑟12.34 and its formula will be as follows:
𝑟12.4 − 𝑟13.4 𝑟23.4
𝑟12.34 =
2 ) (1 2 )
√(1 − 𝑟13.4 − 𝑟23.4
Similarly, 𝑟13.24 , 𝑟14.23 , 𝑟23.14 , 𝑟24.14 , 𝑟34.12 can also be calculated.
Interpretation of Partial Correlation
Partial correlation coefficients vary between -1 and +1 and they are interpreted just as simple
correlation coefficient.
Example 4.73: Find out the values of 𝑟12.3 , 𝑟13.2 and 𝑟23.1, if
(i) 𝑟12 = 0.8, 𝑟13 = −0.4, 𝑟23 = −0.56
(ii) 𝑟12 = 0.28, 𝑟23 = 0.51, 𝑟13 = 0.51
Solution:
𝑟12 − 𝑟13 𝑟23 0.8 − (− 0.4) (− 0.56)
(i) 𝑟12.3 = =
2 ) (1− 𝑟 2 )
√(1−𝑟13 √[1− (− 0.4)2 ][1− (− 0.56)2 ]
23

0.8−0.224 0.576 0.576

= = = 0.759 = + 0.7589
√(1−.16) (1−.3136) √0.5766
 𝒓𝟏𝟑 − 𝒓𝟏𝟐 𝒓𝟐𝟑 − 𝟎.𝟒 − (𝟎.𝟖 × − 𝟎.𝟓𝟔)
𝑟13.2 = =
√(𝟏 − 𝒓𝟐𝟏𝟐 ) (𝟏 − 𝒓𝟐𝟐𝟑 ) √[𝟏 − (.𝟖)𝟐 ] [𝟏 − (− 𝟎.𝟓𝟔)𝟐 ]

− 0.4 + 0.448 0.048 0.048

= = = 0.497 = +𝟎. 𝟎𝟗𝟔𝟔
√(1 − .64) (1 − .3136) √0247

𝒓𝟐𝟑 − 𝒓𝟏𝟐 𝒓𝟏𝟑 − 𝟎.𝟓𝟔 − (𝟎.𝟖 × − 𝟎.𝟒)

𝑟23.1 = =
√(𝟏 − 𝒓𝟐𝟏𝟐 ) (𝟏 − 𝒓𝟐𝟏𝟑 ) √[𝟏 − (.𝟖)𝟐 ] [𝟏 − (− 𝟎.𝟒)𝟐 ]

− 𝟎.𝟓𝟔 + 𝟎.𝟑𝟐 −𝟎.𝟐𝟒 −𝟎.𝟐𝟒

= = = 𝟎.𝟓𝟒𝟗𝟗 = −𝟎. 𝟒𝟑𝟔𝟒
√(𝟏 − .𝟔𝟒) (𝟏 − .𝟏𝟔) √𝟎.𝟑𝟎𝟐𝟒

𝑟12 − 𝑟13 𝑟23

(ii) 𝑟12.3 =
2 ) (1− 𝑟 2 )
√(1−𝑟13 23

.28 − .49 × .51 0.28 − 0.25

= [1 − (.49)2 [1 − (.51)2]] =
√(1 − 0.24) (1 − 0.26)

.03 .03
= = .75 = + 0.04
√.76 × .74

𝑟13 − 𝑟12 𝑟23

𝑟13.2 =
2 ) (1− 𝑟 2 )
√(1−𝑟12 23

.51 − 28 × .49 .51 − 0.1372

= =
√[1 − (.28)2 ][1 − (.49)2 ] √(1 − 0.0784) (1 − 0.24)

0.3728 0.3728
= = 0.8369 = 0.445
√0.9216 × .76

𝑟23 − 𝑟12 𝑟13

𝑟23.1 =
2 ) (1− 𝑟 2 )
√(1−𝑟12 13

0.49 − .28 × .51 .49 − 0.1428

= =
√[1 − (.28)2 ] [1 − (.51)2 ] √(1 − 0.0784) (1 − 0.26)

0.3472 0.3472
= = 0.8258 = 0.42
√0.9216 × .74

MULTIPLE CORRELATION
Multiple correlation is a statistical measure of the combined relationship of two or more
independent variables with one dependent variable.
(a) Multiple Correlation Coefficient with 𝑋1 as dependent variable and 𝑋2 and 𝑋3 as
independent variables = 𝑅1.23
(b) Multiple Correlation Coefficient with 𝑋2 as dependent variable and 𝑋1 and 𝑋3 as
independent variables = 𝑅2.13
(c) Multiple Correlation Coefficient with 𝑋3 as dependent variable and 𝑋1 and 𝑋2 as
independent variables = 𝑅3.12
 2 2
𝑟12 + 𝑟13 − 2𝑟12 𝑟13 𝑟23
𝑅1.23 = √ 2
1 − 𝑟23

2 + 𝑟 2 − 2𝑟 𝑟 𝑟
𝑟12 23 12 13 23
𝑅2.13 = √ 2
1−𝑟13

2 2
𝑟31 + 𝑟32 − 2𝑟12 𝑟13𝑟23
𝑅3.12 = √ 2
1 − 𝑟12
• Interpretation: Multiple correlation coefficients can never be negative. It lies between
0 and +1, 𝑖. 𝑒.
0 < 𝑅1.23 < 1

Example 4.70: Calculate 𝑅1.23 from the following information:

𝑟12 = 0.4, 𝑟13 = 0.3, 𝑟23 = 0.9

Solution:
2 +𝑟 2 −2𝑟 𝑟 𝑟
𝑟12 13 12 13 23
𝑅1.23 = √ 2
1−𝑟23

(.4)2 +(.3)2 − 2 (.4) (.3) (.9)

=√ 1−(.9)2
.16 + .09 − .216 0.034
=√ =√ = 0.423
1−.81 .19

Example 4.72: The following zero-order correlation coefficients are given:

𝑟12 = 0.98, 𝑟13 = 0.44, 𝑟23 = 0.54
Calculate multiple correlation coefficient treating first variable as dependent and
second and third variables as independent.
Solution:
We have to calculate multiple correlation coefficient treating first variable as dependent
and other variable as independent, it means the value of 𝑅1.23
2 + 𝑟 2 − 2𝑟 𝑟 𝑟
𝑟12 13 12 13 23
𝑅1.23 = √ 2
1− 𝑟23
(.98)2 + (.44)2 − 2 (.98 × .44 × .54)
=√ 1 − (.54)2

0.9604 + 0.1936 − 2 × 2328

=√ 1 − 0.2916

0.6884
= √0.7084 = √. 9718 = 0.9858