Dr.

Gururaj S Barki

CORRELATION

Correlation is a statistical measure that indicates the extent to which two or more variables
fluctuate together.
In other words “Correlation is a statistical technique that can show whether and how strongly
pairs of variables are related”
Example: i) Income & Expenditure.
ii) Rainfall & Production .
iii) Sales & Price.

Types of Correlation

1. Positive Correlation

If the two variables correlated are moving in the same direction is called “Positive
Correlation”.

2. Negative Correlation

If the two variables correlated moving in the opposite direction is called Negative
Correlation”.

3. Zero Correlation:

If there is no correlation between the two variables then the correlation is called “Zero
Correlation”.

4. Linear Correlation:

If the ratio of change in two variables is same, then it is called “Linear Correlation”.

5. Non-Linear Correlation

If the ratio of change in two variables is not same, then it is called “Non-Linear
Correlation”.

Methods of Studying the Correlation

1. Scatter Bar Diagram:

It is a non-mathematical method of studying correlation between two variables.

If is a simple method of ascertaining of scatter of various points.

a) Perfect Positive Correlation:
y

o x
 Dr. Gururaj S Barki

b) Perfect Negative Correlation. c) High Degree Positive Correlation.

y y

o x o

d) High Degree Negative Correlation. e) Low Degree Positive

y
y

o x x

f) Low Degree Negative Correlation. g) Negative Correlation.

y
y

o
o
2. Graphic Method.
x
By plotting the points on graphs two separates lines or curves are got for two variables
by studying the direction and closeness of the said two curves it is derived whether correlation
exits.

If the both the curves are moving in the same direction correlation said to be Positive.
 Dr. Gururaj S Barki

Diagram-1

If the curves are moving the opposite direction, correlation is said to be Negative.
Diagram-2

3. Karl Pearson Co-efficient of Correlation:

It is widely used mathematical way of finding correlation between variables found by
karl Pearson.

Assumption or Merits

 There is a linear relationship between the variables.

 Variables understudy are effected by large number of independent courses.
 There is a cause and effect relationship between forces and effecting the series.

Demerits

1. The assumption of linear relationship always does not exits among variables.
2. The value of correlation is unduly effected items.
3. It takes more time to compute.

Properties of Correlation Co-efficient

 The value of co-efficient of correlation lies between  1.

 The correlation co-efficient is independent of change of scale and origin of variables
x&y.
 The co-efficient of correlation is geometric mean of two regression co-efficient i.e.
r= bxy + byx
 Dr. Gururaj S Barki

Formula............
I. Karl Pearson
Individual observation Discrete & Continues observation

ndxdy - (dx) (dy) Nfdxdy - (fdx) (fdy)

r=
r= Nfdx2 - (fdx)2 Nfdy2 - (fdy)2

ndx2 - (dx)2 ndy2 - (dy)2

Where...
n = Number of observation. N = Number of frequency.
x = First variable y = Second variable.
f = Frequency. d = Deviations.

4. Rank Method of Correlation:

It is found by “Charles Spearman”. The requirement of normal distribution for study
of correlation by Pearson Co-efficient method is overcome by this method given observation
will be assigned ranks. Using ranks correlation co-efficient of ranks is found.

Formula......
6D2
rs = 1 -
(n3 - 1)

Where...

d = Difference of ranks. n = Number of rank.

In case of equal ranks assigned to some entries formula is adjusted as follows...

6 [D2 + 1/12 (m3-m) + 1/12 (m3-m) + 1/12 (m3-m).......]

rs = 1 -
(n3 - 1)

If the ranks are given 3 or more judges

Formula is............
6D2
rs = 1 -
n3 - n
m = Number of ranks repeated.

5. Concurrent Deviation Method

It is a simple method of studying correlation considering only change in variables

rather than degree.
 Dr. Gururaj S Barki

rc =   (2c - n)

n
Where...

C = Number of positive signs.

n = Number of observation - 1

Degree of correlation

0 - 0.30 Poor degree correlation.

0.30 - 0.50 Low degree correlation.
0.50 - 0.70 Moderate degree correlation.
0.70 - 0.99 High degree correlation.
0.99 - 1 Perfect degree correlation.
 Dr. Gururaj S Barki

INDEX NUMBER

Index number is a ratio which gives the average change in the level of a phenomenon
between two different periods of time or two different places.

Uses of Index Numbers:

 To frame suitable economic and wage policies.
 To forecast business conditions.
 To disclose general tendency.
 To measure the production, prices, exports, imports.
 To determined the real value of money.

Steps to construct of Index Number

1. Purpose and Scope:

A precise statement of the object and the adequate coverage of geographical area
must be mentioned.

2. Selection of Base Year:

If is the year of comparison. It should be very close to current year and should be a
normal period.

3. Selection of number of items:

A few representative items based on the sample survey, should taken.
4. Price list:
Accurate prices must be obtained. Decision must be made whether to get retail prices
or whole-sale prices.
5. Selection of an average:
A suitable average should be selected in the light of nature of the data available object,
scope.
6. Selection of suitable weights:
Relative importance attached to different items according to their use are called
weights.
7. Selection of suitable formula:
A suitable formula must be selected depending on the object, scope, nature of index
number.

Types of Index Numbers

[Link] Index Number: ( it measure the change in price)

It measures the general changes in the prices like wholesale price index number, retail
price index number or consumer price index number.

[Link] Index Number: (it measure the change in quality)

It measures the changes in the valume of goods produced (manufactured), consumed

or distributed like the indices of agricultural production, industrial production, imports and
exports etc.,

[Link] Index Number: (it measure the change in value)

It measures the changes in the total value ( price multiplied by quantity) of production
such as indices of retail sales or profits or inventories.

Formula.............
 Dr. Gururaj S Barki

Name Price Index Numbers Quantity Index Numbers

P01 =  p 1q 0 Q01 =  q 1p 0
Laspeyre’s X100 X100
 p 0q 0  q0p 0
 p 1q 1  q 1p 1
Paasche’s P01 = X100 Q01 = X100
 p 0q 1  q 0p 1
 p 1q 0 +  p 1q 1  q 1p 0 +  q1p1
Marshall- P01 = X100 Q01 = X100
Edgeworth  p 0q 0 +  p 0q 1 q0p 0 +  q 0p 1
 p 1q 0  p 1q 1  q 1p 0  q1p 1
+
Dorbish & P01 =  p 0q 0  p0q1X100 Q01 =  q 0p0 +  q0p1 X100
Bowley’s 2 2
Fishers  p 1q 0  p 1q 1  q 1 p 0 x  q1p 1
Ideal
P01 = x X100 Q01 = X100
Index  p 0q 0  p 0q 1  q 0p0  q0 p 1
Number

Where..
P0 = Base Year Price.
P1 = Current year price.
Q0 = Base year quantity.
Q1 = Current Year Quantity.

3. Value Index Number:

 p1q1
V= X100
 p0q0
Test of Index Number:

1. Unit-Test:
It requires that the index number should be independent of statistical units. All the index
number satisfy this test.

2. Time-Reversal Test:
It requires that product of the index numbers calculated by reversing the time periods
should be unity.

P01 X P10 = 1

3. Factor-Reversal Test:
It requires that product of the price index number and Quantity index number should
be equal to value.( or total change in the value)

 p 1q 1
P01 X Q01 =

4. Circular Test:  p 0q 0
It is the extension of time reversal test and requires that
P01 X P12 X P20 = 1
The formula can be extended to any number of years.
 Dr. Gururaj S Barki

Consumer Price Index number ( Cost of Living Index Number )

It is an index number used to measure the average change in the cost of living of a
particular class of people between two points of time.

Uses of C.L.I or C.P.I.

1. They are used to find purchasing capacity of rupee.

2. They are used to frame suitable wage and economic policies.
3. They are used in market analyses of goods and services.
4. They used by the employers to grant dearness allowances to their employees from
time to time.

Steps of C.L.I or C.P.I.

1. Weight:

Based on samples survey suitable weights must be assigned to the commodities

selected.

2. Price List:
Accurate retail price of the items selected should be calculated.

3. Method of Constructions:
These are two methods both methods give same results.
Aggregate Expenditure Method Family Budget Method
 p1 q 0  WP
C.P.L = X100 C.P.L =
W
 p 0q 0
Where....
1) 2)W = p0 q0
P= P1 X100
P0

Conducting the family Budget survey and suitable weights.

[Link]. [Link]. [Link] and lighting. [Link] Rent.
 Dr. Gururaj S Barki

TIME-SERIES

Meaning:

 The past data observed and recorded over different points of time is known as time
series. Or
 A time series is a set of statistical observations arranged in chronological order. Or
 A time series consist of statistical data which are collected, recorded, observed over
successive increments.

Uses of time series

1. It helps planning and estimating / forecasting the business activities
2. It helps in understanding past behavior
3. It helps to compare the data over a period of time
4. It helps to identify the economic activities present and future.
Components or Types Time-series:

[Link] Trend. [Link] variation.

[Link] Variation. [Link] Variation.

1. Secular Trend:( Long-Term Variation)

It is the general tendency of the time series to increase or decrease or to remain
constant over a long period of time.

Secular Variation:
The variation that occurs in time series simply because of lapse of long time is known
as “Secular Variation”.
Examples:
a) Increase in the price of gold in the past many years.
b) Decrease in the death rates in the past many years.

2. Seasonal Variation: (Short-term Variation)

The regular and periodic changes that occurs in time series due to the repeated
occurrences of seasons are known as “Seasonal Variation”.
Examples:
a) Increase in sales of woolen cloth during winter season
b) Increase in sales in umbrella in rain season.
c) Increase in sales of cold or cold drinks during summer reasons.
d) Change in customs and tradition based on the seasons.

3. Cyclical Variation

The changes due to the repeated occurrences of the periods (period of dispersion,
period of improvement, period of prosperity, period of decline in the business activities) over
a long period of time are known as “Cyclical Variation”.
Example:

Prosperity Decline

Normal

Depression Recovery / Improvements

 Dr. Gururaj S Barki

4. Random Variation or Irregular or Erratic Variation

The changes that occur in time series due to the random occurrences like earth
quakes, wars, etc.,
Examples:
a) Sudden increase in the death rate.
b) Unexpected increase in the price of vegetables
c) Flood, war, earthquake, Tsunami etc,.

Methods measuring the Secular Trend

I. Graphic Method

A straight line or a smooth curve giving the regular movement of the time series is
drawn along the given data by eliminating fluctuation.

II. Methods of Least Squares

A straight line in the form of Yc = a + bx is fitted to the time series to find the trend,
taking Y as the observed variable and X as the time periods.
The two constants a and are determined of lest squares i.e. by solving the following
normal equations.

Formula..........
Trend line is...... Y = a + bx

Normal Equations are supported trend line

y = na + bx.......................I equation.
xy = ax + bx2....................II equation.

When x = 0 it is easy to apply the following formula

y xy
a= b=
n x2
Merits:
1. It is a mathematical method.
2. It is the line best fit.

Demerits:
1. Omission or addition of any value will change the position of the line.
2. It is difficult method as compared to the other methods.
3. It is always not practical.

III. Methods of Moving Averages

It is a non-linear trend. The moving averages are simple arithmetic means calculated
successively for the consecutive values of observations of time series by taking a specific
period say, 2years, 3years, 4years, 5years etc.,
These values are plotted on the graph and joined by smooth curve giving general
tendency, of the movement of the variable.

Merits
 Dr. Gururaj S Barki

1. It is simplest to calculate and easy to understand.

2. Cyclic fluctuations are easily removed.
3. It is bases on all the observation.
4. It is suitable when the data contains short term cyclic variations.
5. The general trend can be shown clearly.

Demerits:

1. It is applied only to eliminate periodic variations.

2. Trend values for all the periods cannot be calculated.
3. Choice of the period of the moving averages is difficult is difficult.

REGRESSION

Meaning:
Regression is the theory of estimation of unknown value of a variable with the help of
known value of the variables.
Regression is a statistical methods of estimating the value of one variable for the given
of the other variable if the two variables given are correlated.

Properties of Regression co-efficient

1. The co-efficient of correlation is geometric mean of two regression co-efficient i.e.

r= bxy + byx
2. The regression co-efficients cannot be of opposite signs.

 If r is positive, both the regression co-efficients will be positive.

 If r is negative, both the regression co-efficients will be negative.
 If r is zero, both the regression co-efficients will be zero.

3. The regression co-efficients bxy is the change occurring in x unit change in y. The
regression co-efficient byx is the change occurring in y for unit change in x
 Dr. Gururaj S Barki

4. The regression co-efficients are independent of the origin of measurment of the

variables. But, they are independent on the scale.

Properties of Regression Lines

1. The regression lines intersect at x and Y .

2. If the correlation is positive the slope of the variable also positive. ( or vice versa )
3. If the perfect correlation, the regression lines coincide.

Differences Between Correlation & Regression

Correlation Regression
[Link] is the statistical method [Link] is the statistical method
with the help of more than one with the help of only one variable.
variable.
[Link] does not clearly indicate which [Link] clearly indicates the which variable
variable is the cause and which is cause and which variable is effect.
variable is the effect.
[Link] gives the degree and direction of [Link] gives the nature of the relationship
the correlation between the two that exits between the two variables.
variables.
Regression Equation:
1. Regression Line of X on Y
The regression line of X on Y is represents by the linear equation as,
X = a + by

For solve the above equation the following normal equation are necessary.
x = na + by..............................I equation.
xy = ay + by2...........................II equation.

2. Regression Line of Y on X

The regression line of Y on X is represents by the linear equation as,

Y = a + bx

For solve the above equation the following normal equation are necessary.
y = na + bx.........................................I equation.
xy = ax + bx2....................................II equation.

Regression line in terms of correlation Co-efficient

Following are the regression co-efficient formula’s….

1. X on Y Line

The regression equation obtain with the assumptions that Y is depending on

x is called Regression Y on X.
x
X-X= r (Y-Y)
y
2. Y on X Line
The regression equation obtain with the assumptions that x is depending on
y is called Regression X on Y.
 Dr. Gururaj S Barki

y
Y-Y=r (X-X)
x

PROBABILITY

If a coin is tossed the toss may result in Head or Tail. Here the chance of Head or tail are
equal.

Meaning

Probability is numerical which indicates the chance of occurrence.

Events

Event is a subject of Sample space. OR Any out comes of the experiments is known
as Event.

Sample Space

The set of all possible outcomes of a random experiment is called Sample Space.

Exhaustive Events

The total number of possible outcomes of an experiments is known as ‘Exhaustive

Events’
 Dr. Gururaj S Barki

Favorable Events

The events which entail the occurrence of the desire events in trial are called
‘Favorable Events’

Mutually Exclusive Events

Two or more events mutually exclusive if any of them can occur at time.

Complementary Events

Complement of A is the event of non-occurrence of A.

The complementary of A is denoted by A’
A’ = 1 – (none) or P (A) + P (A’) = 1

Mathematical or Classical or Priori Probability:

Number of Favorable Cases m
P(A) = OR P ( A ) =
Total number of Outcomes n

Addition Theorem :
1) For mutually exclusive events
P(AB)=P(A)+P(B)
2) When the events are not mutually exclusive events:
P(AB)=P(A)+P(B)–P(AB)

Multiplication Theorem:

P(AB)=P(A).(B) or P(AB)=P(A)x(B)