CORRELATION
And
REGRESSION ANALYSIS
Measures of Association
• Statistical technique used to check the association
between two variables.
• Correlation measures the degree of association between
two variables.
Measures of Association
Correlation is only a number (r), which measures the direction and
magnitude of the association/relationship.
– Relationship between Sales of a Company and Expenditure on
Advertisement or Investment on Research & Development(R&D)
– Relationship of Sales of a Company and Earning per Share or
Price-Earning Ratio of its stock
– Relationship of Price and Demand of a Product
– Relationship between Rate of Inflation and Gold Price
Numerical Value of Correlation Coefficient
and Its Interpretation
• The value of correlation coefficient r lies between –1 and
+1.
Karl Pearson’s correlation coefficient
Karl Pearson’s coefficient of correlation is a quantitative measure of
the degree of relationship between two variables. Suppose these
variables are x and y, then Karl Pearson’s coefficient of correlation is
defined as
Karl Pearson’s correlation coefficient
The sales revenue and advertisement expenses of a company for the
past 10 months is given below. Find the coefficient of correlation
between sales and advertisement.
Q. A major supermarket chain wants to
determine the association between promotion
and sales. Data were obtained from 5 states
on the promotional expenses and sales
expenses. You are assigned the task of
telling the manager whether there correlation
exists between advertising and Sales or not?
STATE PROMOTION SALES
1 45 90
2 50 75
3 65 70
4 60 75
5 35 80
• Question: The following data gives the Sales and Net
Profit for Top Telecom companies during the quarter July-
September 2014. Find out the Karl Pearson Correlation
coefficient and interpret your results.
Company Average Sales (Rs Average Profits
Crores) (Rs Crores)
Airtel 65 47
Idea 22 22
Aircel 24 35
Tata Docomo 10 4
Reliance 5 6
Vodafone 16 9
BSNL 24 20
Spearman’s Rank Correlation
• When the data is ordinal in nature, we use
Spearman’s Rank Correlation Coefficient.
• Spearman defined the rank correlation as
2
6 di
rs 1
n(n 1)
2
where, di is the difference in the ranks of ith individual or
unit, and n is the number of individuals or units.
Spearman’s Rank Correlation
A social science researcher wants to find out the degree of association
between sugar prices and wheat prices. The researcher has collected data
relating to the price of sugar and wheat in 14 randomly selected months from
the last 20 years. How can he compute the Spearman’s rank correlation from
the data provided in the table below:
Spearman’s Rank Correlation
Sugar and wheat prices for 14 randomly selected months from the last 20 years
Question: As per a study, the following are the ranks
of priorities for ten factors taken as ‘Job Commitment
Drivers’ among the executives in Asia Pacific (AP) and
India. Calculate the Correlation between priorities of
‘Job Commitment Drivers’ among executives from
India and Asia Pacific.
Favorable Rank
INDIA ASIA PACIFIC
Job Satisfaction 1 1
Work Environment 2 2
Team Work 3 4
Communication 4 3
Performance Management 5 5
Innovation 6 6
Leadership 7 9
Training and Development 8 7
Supervision 9 8
Compensation 10 10
Regression Analysis
• Correlation analysis deals with exploring the correlation that
might exist between two or more variables.
• The regression analysis, on the other hand, studies the
relationship among two or more variables.
Regression Analysis
• Regression analysis is the process of developing a
statistical model, which is used to predict the value of a
dependent variable by at least one independent variable.
• The variable whose value is influenced or is to be predicted
is called as the dependent variable.
– Also called as Regressed or explained variable.
• The variable which influences the value or is used for
prediction is called as independent variable.
– Also called as Regressor or Predictor or explanatory variable.
Types of Regression Analysis
• There are two types of regression analysis
viz. Simple and Multiple.
– Simple Regression Analysis deals with only
two variables
– Multiple Regression Analysis deals with more
than two variables
Simple Linear Regression
• The equation for the population from which sample is
drawn is:
y = α + βx
• The equation for the SAMPLE is made as:
y = a + bx
• The values of the parameters α and β are estimated
by a and b, respectively. The exercise of estimating
the relationship amounts to determining the values of
‘a’ and ‘b’.
For Example
The regression equation of Marks on I.Q.
of the type
Marks = a + b ( I.Q.)
by representing Marks by the variable y
and I.Q. by the variable x .
Two Regression Equations
• Regression equation indicates the change in
y with respect to change in x.
y a bx
• Similarly, the equation
x c dy
is called the regression equation of x on y.
A Deterministic and Probabilistic Model
ε is the error of the regression line in fitting the points of the regression equation. If
a point is on the regression line, the corresponding value of ε is equal to zero. If the
point is not on the regression line, the value of ε measures the error.
It can be noticed that in the deterministic model, all the points are assumed to be
on the regression line and hence, in all the cases random error ε is equal to zero.
Probabilistic model includes an error term which allows the value of y to vary for
any given value of x.
Error in simple regression
Summary of the estimation process for simple
linear regression.
Value of a and b for equation
y=a+bx
A cable wire company has spent heavily on advertisements. The sales and
advertisement expenses (in thousand rupees) for the 12 randomly selected
months is given below. Develop a regression model to predict the impact of
advertisement on sales.
Q. A major supermarket chain wants to
determine the association between promotion
and sales. Data were obtained from 5 states
on the promotional expenses and sales
expenses. You are assigned the task of
telling the manager whether there is any
impact of advertising on
STATE
Sales or not?
PROMOTION SALES (Units
(Crore) Sold)
1 45 90
2 50 75
3 65 70
4 60 75
5 35 80
Q. The marketers of a University wanted to
explore whether student performance is
dependent on their IQ or not? They collected
data from respondent (Table). Please find the
relationship between these two variables.
Also, Predict the Performance of a student
with IQ 90. STATE Performance IQ
(Marks)
1 87 70
2 76 75
3 77 80
4 45 55
5 76 70
Thank You
