1

Chapter 6
CORRELATION AND REGRESSION

CORRELATION ANALYSIS

Correlation is a measure which explains the degree of relationship

between variables. It is a statistical technique used to determine the degree to
which two variables are related. In other words, correlation quantifies to which
extent two variables, A and B, go together. If the value of A increases with the
increase in the value of B, the correlation is said to be a positive. If the value of
A decreases when the value of B increases then there exists a negative
correlation. With high values of A being associated to high value of N, a positive
correlation exists between the variables.

Correlation is measured by correlation co-efficient which is denoted by ‘r’

and it always falls between +1 and -1. The direction of changes between
variables is indicated by + or - sign. If r value is any + value then it is known
that there is a positive correlation existing between variables. And if the r is
any negative value, then we can conclude that there is a negative correlation
existing. Hence, the correlation analysis enables us to get an idea about the
degree & direction of the relationship between the two variables under study.

TYPES OF CORRELATION

Type I correlation can be classified as the following:

a) Positive Correlation: The correlation is said to be positive correlation if

the values of two variables changing with same direction.
Example: Public Expenses and sales, Height & weight, Temperature and
water usage, study time and grades, etc.
b) Negative Correlation: The correlation is said to be negative correlation
when the values of variables change with opposite direction.
 2

Example: Price and quantity demanded, Alcohol consumption and

driving ability, etc.

Direction of the Correlation

a) Positive relationship–Variables change in the same direction. It is

indicated by the “+” sign.
 As X is increasing, Y is increasing
 As X is decreasing, Y is decreasing

Example: As height increases, so does weight.

b) Negative relationship–Variables change inopposite directions.It is

indicated by the “-” sign.
 As X is increasing, Y is decreasing
 As X is decreasing, Y is increasing

Example: As TV time increases, grades decrease

Type II Correlation can be classified as the following:

a) Simple correlation: It is also called Pearson's correlation or product

moment correlation
coefficient. Under simple correlation problem there are only two variables
are studied. It measures the nature and strength between two variables
of
the quantitative type. The sign of r denotes the nature of association
while the value of r denotes the strength of association.
b) Multiple Correlation: Under Multiple Correlation three or more than
three variables are studied. It can be further divided into partial and
total.
 3

 Partial correlation: This analysis recognizes more than two variables

but considers only two variables keeping the other constant.
 Total correlation: This is based on all the relevant variables, which is
normally not feasible.

Type III Correlation can be classified as the following:

a) Linear correlation: Correlation is said to be linear when the amount of

change in one variable tends to bear a constant ratio to the amount of
change in the other. The graph of the variables having a linear
relationship will form a straight line.

Ex X = 1, 2, 3, 4, 5, 6, 7, 8,

Y = 5, 7, 9, 11, 13, 15, 17, 19,

Y = 3 + 2x

b) Non Linear correlation:The correlation would be nonlinear if the amount

of change in one variable does not bear a constant ratio to the amount of
change in the other variable.

METHODS OF STUDYING CORRELATION

Correlation can be analysed using any of the following methods:

a) Scatter Diagram Method:

Scatter Diagram is a graph of observed plotted points where each points

represents the values of X & Y as a coordinate. It portrays the relationship
between these two variables graphically. Examples of scatter diagram
depicting various levels of correlation between X and Y:
 4

b) Karl Pearson’s Coefficient of Correlation:

Pearson’s ‘r’ is the most common correlation coefficient. Karl Pearson’s

Coefficient of Correlation denoted by- ‘r’ The coefficient of correlation ‘r’
measure the degree of linear relationship between two variables say x & y.

The coefficient r is defined as below,

r= (x = X - X̅ and y = Y-Y̅)

Or

(Direct Method)

Under direct method, X and Y are the values given in the problem and no
need to calculate mean of the values.
 5

Karl Pearson’s Coefficient of Correlation denoted by “r” ranges from positive

1 to negative 1 i.e -1 ≤r ≥+1. Degree of Correlation is expressed by value of
Coefficient and direction of change is indicated by negative or positive sign.

It summarizes in one value, the degree of correlation & direction of correlation

also.

Interpretation of Correlation Coefficient (r)

 The value of correlation coefficient ‘r’ ranges from -1 to +1

 If r = +1, then the correlation between the two variables is said to be
perfect and positive
 If r = -1, then the correlation between the two variables is said to be
perfect and negative
 If r = 0, then there exists no correlation between the variables.

Assumptions of Pearson’s Correlation Coefficient

There is linear relationship between two variables, i.e. when the two variables
are plotted on a scatter diagram a straight line will be formed by the points.

Cause and effect relation exists between different forces operating on the item
of the two variable series.

CAUSALITY

Causality, also known as causation, is defined as a cause-effect relationship

between two variables. A significant correlation does not necessarily indicate
causality but rather a common linkage in a sequence of events. One type of
significant correlation situation is when both variables are influenced by a
common cause and therefore are correlated with each other. The following are
a few examples of causality.

An individuals with a higher level of income may have both higher levels of
savings and spending.
 6

We might find that there is a positive correlation between level of savings and
level of spending but this does not mean that one variable causes the other.

We should mention the very interesting case where two related variables are
separated by several steps in a cause-effect chain of events.

Example Problem:

The heights (in centimeters) and weight (in kilograms) of 10 athletes in a team
are:

Height (X) 182 185 190 194 197 199 202 205 208 210

Weight (Y) 75 74 80 82 84 88 94 96 96 98

Calculate:
Find out the coefficient of correlation.
HEIGHT WEIGHT
(X) (Y) x y XY

182 75 15.2 231.04 11.7 136.89 177.84

185 74 12.2 148.84 12.7 161.29 154.94
190 80 -7.2 51.84 -6.7 44.89 48.24
194 82 -3.2 10.24 -4.7 22.09 15.04
197 84 -0.2 0.04 -2.7 7.29 0.54
199 88 1.8 3.24 1.3 1.69 2.34
202 94 4.8 23.04 7.3 53.29 35.04
205 96 7.8 60.84 9.3 86.49 72.54
208 96 10.8 116.64 9.3 86.49 100.44
210 98 12.8 163.84 11.3 127.69 144.64
1972 867 231.04 728.1 751.6

= = 197.2
 7

Y̅ = = 86.7 r= =

Rank Correlation
Rank correlation was developed by the British Psychologist Charles
Edward Spearman in the year 1904. It is applicable for the variables which
cannot be quantified or measured in quantity, but which can be ranked such
as beauty, decision making ability, or leadership skill. The rank correlation
finds its application when the data are not normally distributed or when the
pattern of distribution cannot be recognized. Whereas, the Pearson’s
correlation is based on the assumption the data are normally distributed. Rank
correlation is calculated by using the formula,

Spearmen’s Rank correlation = 1- Where, D is R1 – R2

Example:

The ranks of 12 beauty contestants for their beauty and their attitude are tabulated below. Find the
correlation coefficient and interpret the result.

Ranks based on beauty 4 3 7 9 10 1 2 4 12 8 6 11

Ranks based on IQ level 11 9 7 5 2 10 8 12 6 3 1 4

D2
D( )

4 11 -7 49

3 9 -6 36

7 7 0 0

9 5 4 16
 8

10 2 8 64

1 10 -9 81

2 8 -6 36

4 12 -8 64

12 6 6 36

8 3 5 25

6 1 5 25

11 4 7 49

Spearmen’s Rank correlation

= 1-

=1-

= 1-

= 1-

= 1-1.6818

Rank Correlation coefficient = 0.6818

REGRESSION ANALYSIS
 9

Correlation analysis helps us to identify the relationship between

variables and to measure the degree and strength of such relationship. It will
not tell us the quantity of change in one variable for a given change in another
variable. Regression analysis is a statistical process which estimates the
amount of change in the dependent variable with reference to any change in
the independent variable. If the change in the independent variable adequately
explains the change in the dependent variables then the regression equation or
model can be used for prediction of future value. Regression analysis assumes
a linear relationship between variables. Hence, it is a process of forecasting
value of change in variable Y using given value of change in variable X.
According to Morris Hamburg ‘regression analysis refers to the methods
by which estimates are made of the values of a variable from a knowledge of
the values of one or more other variables and to the measurement of the errors
involved in this estimation process’. Ya-Lun Chou explained that it attempts to
establish the ‘nature of relationship’ between variables- that is, to study the
functional relationship between the variables and thereby provide a mechanism
for prediction, or forecasting’.
Some examples of interrelated variables are interest rate and sale of
houses, electricity bill and weather conditions, savings habit and amount of
expenditure, sales and advertisement expenditure etc. The regression
coefficient is the slope of the regression line and tells what the nature of the
relationship between the variables is. How much change in the independent
variables is linked with how much change in the dependent variable. The larger
the regression coefficient the more is the change.

Regression technique
(i) Helps to find equation of X on Y and
(ii) Helps to find equation of Y on X by using the formula,

Regression equation of X on Y
 10

(X-X̅) = r (Y=Y̅)

r =

Regression equation of Y on X

(Y-Y̅) = r (X-X̅)

r =

Example:
The following data gives the experience of Machine operators and their
performance rating given by the no. of good parts turned out per 100 pieces.
Operator Experience (X) Performance Ratings (Y)
1 16 87
2 12 88
3 18 89
4 4 68
5 3 78
6 10 80
7 5 75
8 12 83
Calculate regression lines of performance ratings on experience and estimate
probable performance of the operator who has 7 years of experience.

Solution:
X Y X-X̅ Y-Y̅ X² XY
16 87 6 6 36 36
12 88 2 7 4 14
18 89 8 8 64 64
4 68 -6 -13 36 78
3 78 -7 -3 49 21
10 80 0 -1 0 0
5 75 -5 -6 25 30
12 83 2 2 4 4
 11

_________ _______ _______ ______ ______

_______
80 648 0 0
218 247

X̅ = Y̅ =

= =

= 10 = 81

Regression of equation of Y and X

(Y-Y̅) = r (X-X̅)

r =

= 1.13

Y-81 = 1.13 (X-10)

Y-81 = 1.13X-11.3
1.13X-Y = -69.7

Y = 1.13 X+69.7

If X = 7
Y = 1.13x7+69.7
Y = 77.61

Example Problems:
 12

1. Estimate (a) The sale for advertising expenditure of 100 Lakhs rupees & (b)
The advertising expenditure for sales of 47 cores rupees from the data given
below.
Sales X (in cores rupees) Y Advertisement exp. (Rs in Lakhs)
14 52
16 62
18 65
20 70
24 76
30 80
32 78

Answer:
X X-X̅ X² Y Y-Y̅ Y² XY
14 -8 64 52 -17 289 136
16 -6 36 62 -7 49 42
18 -4 16 65 -4 16 16
20 -2 4 70 1 1 -2
24 2 4 76 7 49 14
30 8 64 80 11 121 88
32 10 100 78 9 81 90
_______ ______ _____ _____ _____
154 288 483 606
384

X̅ = Y̅ =

= =

= 22 = 69

Regression equation if X on Y,

(X-X̅) = r (Y=Y̅)
 13

r =

X-22 = 0.63 (Y-69)

X-22 = 0.63Y – 43.47

X = 0.63Y – 21.47

Regression equation of Y on X

r =

= 1.33

Y-69 = 1.33 (X-22)

Y = 1.33X-29.26+69

Y = 1.33X+39.74

a) X = 0.63Y-21.47
= 0.63X100-21.47
= 41.53 crores

The sale = 41.53 crores

b) Y = 1.33X+39.74
= 1.33X47+39.74
= 102.25 Lakhs

Advertisement = 102.25 Lakhs

2. Information about advertisement Expenditure & Sales.

X Y
Details Advertisement exp. (X) Sales (Y)
(in crores) (in crores)
Mean (X̅) 20 120
 14

5 25
Correlation Coefficient is 0.8
(1) Calculate the 2 regression equations
(2) Find the likely sale when advertisement expenditure is 25 crores
(3) What should be the advertisement budget if the co., wants to attain sales
target of 150 crores.
Answers:
X̅ = 20 Y̅ = 120
X = 5 y = 25
X = 0.8

(1) Regression equations of X and Y

(X-X̅) = r (Y-Y̅)

X-20 = 0.8* (Y-120)

5(X-20) = 0.8 (Y-120)

5X – 100 = 0.8Y - 96
5X = 0.8Y – 96 + 100
5X = 0.8Y + 4
X = 0.16Y + 0.8
(2) Regression equations of Y on X

Y-Y̅ = r (X - X̅)

Y – 120 = 0.8* (X – 20)

Y – 120 = 4(X-20)
Y – 120 = 4X – 80
Y = 4X – 80 + 120
Y = 4X+40
 15

Y = 4X+40
= 4X+40

= 140 crores

X = 0.16Y+0.8
= 0.16X150+0.8

= 24.80 crores

3. The R&D of a Chain Store wants to make a catchment analysis to start its
new branch. As the number of customers visiting the new branch depends
upon the distance of the location of the branch from the city limit, it wants to
make a correlation and regression analysis. It provides you the following data
of number of customers (in thousands) and the distance (in kilometers).

No. Customer (x) 10 8 6 5 3 2

Distance (y) 16 24 28 35 38 40
1. Calculate the correlation co-efficient.
2. If the branch is to be located just 5 kilometers from the city limit, how
many customers, the branch can anticipate?
3. If the branch wants to have 20000 customers, at what distance the
branch is to be started?

Solution:
No of customer 10 8 6 5 3 2
distance 16 24 28 35 38 40

A) Calculate the correlation co-efficient

X Y x y X2 Y2 xy
10 16 4.3 -14.17 18.49 200.79 -60.931
8 24 2.7 -6.17 7.29 38.069 -44.98
6 28 0.3 -2.17 0.09 47089 0.1953
5 35 -0.7 4.83 0.49 23.329 2.3667
3 38 -2.7 7.83 7.29 61.309 57.0807
 16

2 40 -3.7 9.83 13.69 96.629 134.5727

34 181 47.34 424.835 88.3044

x= X-

= = = 5.7

= = 30.17 =

r=

= 0.6226
b) When y is 5km, x (customer) by Regression
Regression of x on y

(X- ) = r (Y- )

r= = = = 0.207

X-5.7 = 0.207 (Y-30.17)

X= 0.3606Y -10.88
X= 0.3606Y -0.57
When Y =5
X= 1.233
c) Regression of Y on X

(Y- ) = (X- )

r = = = 1.86
 17

(Y-30.17) = 1.86 (X-5.7)

Y =1.86X - 10.602+30.17
Y = 1.86 X + 19.568
When X = 2000 Y= 1.86(20) + 19.568
Y= 3739.56 Km

4. The heights (in centimeters) and weight (in kilograms) of 10 athletes in a

team are:

Height (X) 182 185 190 194 197 199 202 205 208 210

Weight (Y) 75 74 80 82 84 88 94 96 96 98

Calculate:
1. Fit the regression line of y on x.
2. Calculate the estimated weight of an athlete whose height is measured
215 cm.
1) Regression line of y on x

(y- (x- )

r. = = = 3.25

y - 86.7 = 3.25(x-197.2)
y = 3.25 x-554.2
2) Calculate y on x = 215 cm
y= 3.25 (215)-554.2
y=144.55cm
5. The marks of six students in science and mathematics are given below:

Mathematics 84 72 90 65 60 55
Science 65 60 78 68 55 52
 18

Determine the regression lines and calculate the expected marks in science for a student who has
secured 95 in mathematics.

Maths (X) Science x Y XY

(Y)
84 65 13 2 169 4 36
72 60 1 -3 1 9 -3
90 78 19 15 361 225 285
65 68 -6 5 36 25 -30
60 55 -11 -8 121 64 88
55 52 -16 -11 256 121 176
426 378 999 448 552

= = 71

= = 63

Calculate repeated mark in Science (Y) What x=95

(Y- (X- ) r = = = 0.58

(Y-63) = 0.58 (X-71)

Y = 0.58 X -41.18+63
Y = 0.58 X + 21.82
Y = 76.92

6. Five diabetics aged 52, 63, 65, 72 and 80 years had blood sugar level of
105, 115, 125, 126 and 130 mg./dL. respectively.
1. Find the equation of the regression line of age on blood sugar level.
2. Based on this data, what is the approximate sugar level of a 85year old
diabetic patient?

Age Sugar Level

 19

x x2 y y2 xy

52 105 -14.4 207.36 -15.2 231.04 218.88

63 115 -3.4 11.56 5.2 27.04 -17.68
65 125 1.4 1.96 4.8 23.04 6.72
72 126 5.6 31.36 5.8 33.64 32.48
80 130 13.6 184.96 9.8 96.04 133.28
332 601 437.2 410.8 373.68

= = 66.4

= = 120.2

A) Regression of age (X) on sugar level (Y)

(X- ) = r (Y- )

r. = = 0.85

(X-66.4) = 0.85 (Y- 120.2)

X= 0.854 – 35.77

b) For 85 year old (X) what is of sugar level (Y)

(Y- ) = r. (X- )

(Y-120.2) = 0.91 (X-66.4)

Y = 0.91X + 59.57
 20

Y = 136.92

7. From the following data of hours worked in a factory (x) and output units (y), determine the
regression line of y on x, the linear correlation coefficient and determine the type of correlation.

Hours (X) 80 79 83 84 78 60 82 85 79 84 80 62
Production (Y) 300 302 315 330 300 250 300 340 315 330 310 240

Hour (X) Units(Y) x x2 y y2 XY

80 300 2 4 -2.6 6.76 -5.2
79 302 1 1 -0.6 0.36 -0.6
83 315 5 25 12.4 153.76 62
84 330 6 36 27.4 750.76 164.4
78 300 0 0 -2.6 6.76 0
60 250 -18 324 -52.6 2766.76 946.8
62 300 4 16 -2.6 6.76 10.4
85 340 7 49 37.4 1398.76 261.8
79 315 1 1 12.4 153.76 12.4
84 330 6 36 27.4 750.76 164.4
80 310 2 4 7.4 54.76 14.8
62 240 -16 256 -62.6 3918.16 1001.6
936 3632 9968.72 2632.8

= = 78

= = 302.6

r. = = 3.50

(a) Regression line of Y on X

 21

(Y- ) = r. (X- )

(Y-302.6) = 3.5 (X-78)

Y = 3.5X + 29.6
(b) Correlation co.eff

r= = = 0.96

The Value shows that there is a high positive correlation

8. Hours for studies (X) Hours for sleeping (Y)

.5 9

1 9

1.5 8.5

2 8

2.5 6

3 5

3.5 5

1. 5
1. Calculate the correlation coefficient.
2. Determine the equation of the regression line of y on x.
3. If a person sleeps twelve hours, how many hours he will allot for his
study?

Solution:
X Y x2 y2
(x- ) x (y- ) y

0.5 9 -1.375 2.0625 1.89 4.25 -2.84

1 9 -0.875 2.0625 0.77 4.25 -1.80
1.5 8.5 -0.375 1.5625 0.14 2.44 -0.59
 22

2 8 0.125 1.0625 0.02 1.13 0.13

2.5 6 0.625 -0.9375 0.39 0.94 -0.59
3 5 1.125 -1.9375 1.27 3.75 -2.18
3.5 5 1.625 -1.9375 2.64 3.75 -3.15
1.0 5 -0.875 -1.9375 0.77 3.75 1.70
15 55.5 0 0 7.89 24.26 -8.73

1.) Correlation co-efficient r =

= = = 1.875

= = = 6.9375

r= =

r= = -1.9815

Correlation co-efficient = -1.9815

2. Regression line of Y on X

(Y- ) = r (X- )

r = = = -1.1065

Y-6.9375 = -1.1065 (X-1.875)

Y-6.9375= -1.1065X+2.0747
Y= -1.1065X+9.0122
 23

3) Regression line of X on Y

(X- ) = r (Y- )

r = = = -0.3599

X-1.875 = -0.3599 (Y-6.9375)

X- 1.875 = -0.3599Y +2.4968
X= -0.3599(12) + 2.4968 + 1.875
x = -4.3188+ 2.4968+1.875
x=8.7
If a person sleeps 12 hours, 8.7 hours he will a cot for his study
9. A researcher wants to study the relation between students leave and his
performance. The following is the details of number of days absent and
marks obtained in an examination.

Marks (X) Number of days of leave (Y)

Below 20 12

20-40 11

40-60 8

60-80 9

80-100 1

Examine what type of correlation exists between X and Y.

Marks X Y M(x) X2 Y2 xy
X=xm- Y=y-

0-20 12 10 -40 3.8 1600 14.44 -152

20-40 11 30 -20 2.8 400 7.84 -56
40-60 8 50 0 -0.2 0 0.04 0
60-80 9 70 20 0.8 400 0.64 16
80-100 1 90 40 -7.2 1600 51.84 -288
 24

41 250 4000 74.8 -480

= = 50

= = 8.2

r= =

r= = -0.877

Very high negative correlation

10. The following table gives the details of salary income and the
money spent in beauty parlor of 40 randomly selected women in a
metropolitan city. Find out the correlation coefficient between salary and
parlor expense.
Salary per month (Rs. in ‘000) Money spent in parlor per month
(Rs. in ‘00)
Below 10 15
10 – 20 20
20 – 30 25
30 – 40 25
40 – 50 35
Above 50 40

X Y M(x) X2 Y2 xy
X=xm- Y=y-

0-10 15 5 -25 -12 625 144 300

10-20 20 15 -15 -7 225 49 105
20-30 25 25 -5 -2 25 4 10
30-40 25 35 5 -2 25 4 -10
40-50 35 45 15 8 225 64 120
50-60 40 55 25 13 625 169 325
 25

160 180 1750 434 840

= = 30

= = 26.6 = 27

r= =

r= = 0.96

Very high positive correlation.

11. The following table summarizes the scores of a sales training provided to six sales men
(x) and sales in the first month (y) in thousand rupees.

X 35 42 32 65 24 76
Y 62 82 58 90 35 94

1 Find the correlation coefficient and interpret the results.

2 Calculate the regression line of y on x and predict the sales of a sales man who obtains 50 in training.

X Y x y X2 Y2 xy

35 62 -10.67 -8.17 113.85 66.75 87.17

42 82 -3.67 11.83 13.47 139.95 -43.42

32 58 -13.67 -12.17 186.87 148.11 166.36

65 90 19.33 19.83 373.65 393.25 383.31

24 35 -21.67 -35.17 469.59 1236.93 762.13

76 94 30.33 23.83 919.91 567.87 722.76

274 2077.34 2552.84 2078.31

 26

= = = 45.67

= = = 70.17

Correlation co-efficient r =

r=

r=

r= = 0.9025

Correlation co-efficient = 0.9025

2. Regression line of Y on X

(Y- ) = r (X- )

r = = = 1.0005

Y-70.17 = 1.0005 (X-45.67)

Y-70.17 = 1.0005X+45.67
Y= 1.0005X+24.48

3) Regression line of X on Y

(X- ) = r (Y- )

X-45.67 = r (Y-70.17)
 27

r = = = 0.8141

X- 45.67 =0.8141 (Y-70.17)

X- 45.67 = 0.8141Y -57.13
X= 0.8141Y-11.46
X=0.8141x50-11.46
X= 29.245

12. Two customers are asked to rank eight varieties of designer sarees.
The ranks given by them are as follows:
Saree Codes: D101 D102 D103 D104 D105 D106 D107 D108
Customer 1: 7 5 8 3 2 4 6 1
Customer 2: 5 6 3 7 3 2 8 4
Calculate Spearman’s rank correlation coefficient.

S.C D2
D( )

D111 7 5 2 4

D102 5 6 -1 1

D103 8 3 5 25

D104 3 7 -4 16

D105 2 3 -1 1

D106 4 2 2 4

D107 6 8 -2 4

D108 1 4 -3 9

R= 1-
 28

=1-

= 1-

= 1-

= 1-0.7619

= 0.2381

Spearman’s rank Correlation coefficient =0.2381

13. Calculate Karl Pearson’s Coefficient of Correlation between

enrolment and playing habits of students of a college from the data given
below.

No. of Students: 240 174 216 302 232 253 132 151

Regular Players: 215 160 180 300 210 220 90 80

X Y y x2 y2 xy
x(x- )

240 215 2.75 33.12 7.5625 1096.93 91.08

174 160 -38.5 21.88 1482.25 478.73 842.38

216 180 3.5 -1.88 12.25 3.53 -6.58

302 300 89.5 118.12 8010.25 13952.33 10571.74

232 210 19.5 28.12 380.25 790.73 548.34

253 220 40.5 38.12 1640.25 1453.13 1543.86

132 90 -80.5 -91.88 6480.25 8441.93 7396.34

151 80 -61.5 -101.88 3782.25 10379.53 6265.62

1700 1455 21795.3125 36596.84 27252.78

 29

= = = 212.5

= = = 181.88

r=

r=

r=

r= =0.9650

r = 0.9650