1

UNIT — I& II
STATISTICS

Definitions: - “The classified facts relating the condition of the people in a state
specially those facts which can be stated in members or in tables of members or
in any tabular or classified arrangements.”
-Webster
“Statistics may be regarded as (i) the study of population (ii) The study of
variation (iii) The study of method of reduction of data”
-R.A. Fisher

2
 Nature /Features /Characteristics of statistics
It is an aggregate of facts

It is estimated according to reasonable standard of accuracy

It is numerically expressed

Analysis of multiplicity of causes.

It is collected for pre-determined purpose

It is collected in a systematic manner

Statistical Methods
•By this method we mean methods
specially adapted to the elucidation
of quantitative data affected by a
multiplicity of causes. Few Methods
are:-
• Collection of Data (2) Classification
(3) Tabulation (4) Presentation (5)
Analysis (6) Interpretation (7)
Forecasting.
Theoretical
• : Mathematical theory which Applied
is the basis of the science of
• It deals with the application
statistics is called theoretical
of rules and principles
statistics.
developed for specific
Division problem in different
disciplines. Eg: - Time series,
of Sampling, Statistical Quality
control, design of
Statistics experiments.

Functions of Statistics:-
 It presents facts in a definite form.
 It simplifies mass of figures
 It facilitates comparison
 It helps in prediction
 It helps in formulating suitable & policies.

Scope of Statistics:-
1. Statistics and state or govt.
2. Statistics and business or management.

3
  Marketing
 Production
 Finance
 Banking
 Control
 Research and Development
3. Statistics and Economics
 Measures National Income
 Money Market analysis
 Analysis of competition, monopoly, oligopoly,
 Analysis of Population etc.
4. Statistics and science
5. Statistics and Research

Limitations:-
(i) It is not deal with items but deals with aggregates.
(ii) Only on expert can use it
(iii) It is not the only method to analyze the problem.
(iv) It can be misused etc.
STATISTICAL INVESTIGATION
Meaning: In general it means as a statistical survey.
In brief, it is Scientific and systematic collection of data and their analysis with the help of various
statistical method and their interpretation.

STAGES OF STATISTICAL INVESTIGATION:-

4
 Interrelation
of Data or
Presentation
and analysis Report
of Data Preparation
Editing of Data

Collection of
Data

Planning
of
Investigati
on

Experiment
or survey
investigation

Original or Confidential
repetitive or open
investigation investigation

Types of
Statistical
Investigation

General
purpose and Complete or
specific sample
purpose investigation
investigation

Official,
semi-official,
Non official
investigation

PROCESS OF DATA COLLECTION

Data: - A bundle of Information or bunch of information.
Data Collection: Collecting Information for some relevant purpose & placed in relation to each other.

5
Collection of Data: - It means the methods that are to be employed for obtaining the required
information from the units under investigations.

Methods of Data Collection:- (Primary Data)

- Direct Personal Interviews
- By observation
- By Survey
- By questionnaires

Preparation of Questionnaires:-
This method of data collection is quit popular, particularly in case of big enquires, it is adopted by
individuals, research workers. Private and public organization and even by government also.
A questionnaires consists of number of question printed or type in a definite order on a form or set of
forms. The respondents have to answer the question on their own.

Importance:-
i. Low cost and universal
ii. Free from biases.
iii. Respondents have adequate time to respond
iv. Fairly approachable

6
 Basis of Primary
Difference Secondary Data
Data

Secondary data are not original, i.e..,

Primary data are original i.e.,
Originality they are already in existence and
collected first time.
are used by the investigator.

Secondary data are in the from of

Organisation Primary data are like raw finished product. They have passed
material. through statistical methods.

Primary data are according to the Secondary data are collected for
Purpose object of investigation and are some other purpose and are
used without correction. corrected before use.

Secondary data are easily available

The collection of primary data
Expenditure from secondary sources
require large sum, energy and
(published or unpublished).
time.

Precautions are not necessary in Precautions are necessary in the use

Precautions
the use of primary data. of secondary data.

Demerits:-
(i) Low rate of return
(ii) Fill on educated respondents
(iii) Slowest method of Response

Steps in construction of a questionnaire:It is considered as the heart of a survey operation.

Hence it should be very carefully constructed

Technical
Emphasize Personal
Ask Logical terms and
Prepare it Prepare on question questions
and not vague
in a general sequence of formulation should be
misleading expressions
form question and left to
questions. should be
wordings the end.
avoided.

7
Example :

Classification & Tabulation of Data

After collecting and editing of data an important step towards processing that classification. It is
grouping of related facts into different classes.
Types of classification:-
i. Geographical:- On the basis of location difference between the various items. E.g. Sugar Cave,
wheat, rice, for various states.
ii. Chronological:- On the basis of time
e.g.-
Year Sales
1997 1,84,408
1998 1,84,400
1999 1,05,000
iii. Qualitative classification: - Data classified on the basis of some attribute or quality such as,
color of hair, literacy, religion etc.

iv. Quantitative Classification: - When data is quantify on some units like height, weight, income,
sales etc.
Tabulation of Data
A table is a systematic arrangement of statistical data in columns and Rows.
Part of Table:-
1. Table number
2. Title of the Table
3. Caption
4. Stub
5. Body of the table
6. Head note
7. Foot Note
Types of Table:-
(i) Simple and Complex Table:-
(a) Simple or one-way table:-

8
 Age No. of Employees
25 10
30 7
35 12
40 9
45 6

(b) Two way Table

Age Males Females Total
25 25 15 40
30 20 25 45
35 24 20 44
40 18 10 28
45 10 8 18
Total 97 78 175

2) General Purpose and Specific Purpose Table:-General purpose table, also known as the reference
table or repository tables, which provides information for general use or reference.
Special purpose are also known as summary or analytical tables which provides information for one
particular discussion or specific purpose.

METHODS OF SAMPLING
Meaning: - The process of obtaining a sample and its subsequent analysis and interpretation is known
as sampling and the process of obtaining the sample if the first stage of sampling.

The various methods of sampling can broadly be divided into:

i. Random sampling method
ii. Non Random sampling method

Random Sampling Method

9
I Simple Random Sampling: - In this method each and every item of the population is given an equal
chance of being included in the sample.
(a) Lottery Method (b) Table of Random Numbers

Merits:
Equal opportunity to each item.
Better way of judgment
Easy analysis and accuracy
Limitations:
Different in investigation
Expensive and time consuming
For filed survey it is not good

II Stratified Sampling:- In this it is important to divided the population into homogeneous group
called strata. Then a sample may be taken from each group by simple random method.

Merit:- More representative sample is used.

10
Grater accuracy
Geographically Concentrated
Limitations: Utmost care must be exercised due to homogeneous group deviation. In the absence of
skilled supervisor sample selection will be difficult.

III Systematic Sampling:- This method is popularly used in those cases where a complete list of the
population from which sampling is to be drawn is available. The method is to be select k th item from
the list where k refers to the sampling interval.

Merits: - It can be more convenient.

Limitation: - Can be Baised.

IV Multi- Stage Sampling: - This method refers to a sampling procedure which is carried out in several
stages.

Merit: - It gives flexibility in Sampling

Limitation: - It is difficult and less accurate

11
Non Random Sampling Method:-
I. Judgment Sampling: - The choice of sample items depends exclusively on the judgment of the
investigator or the investigator exercises his judgement in the choice of sample items. This is an
simple method of sampling.

II. Quota Sampling: - Quotas are set up according to given criteria, but, within the quotas the
selection of sample items depends on personal judgment.

III. Convenience Sampling: - It is also known as chunk. A chunk is a fraction of one population taken
for investigation because of its convenient availability. That is why a chunk is selected neither by
probability nor by judgment but by convenience.

12
Size of Sample:- It depends upon the following things:-
Cost aspects.
The degree of accuracy desired.
Time, etc.
Normally it is 5% or 10% of the total population.

Limitation of overall sampling Method:-

Some time result may be inaccurate and misleading due to wrong sampling.
Its always needs superiors and experts to analyze the sample.
It may not give information about the overall defects. In production or any study.
It Becomes Biased due to following reason:-
(a) Faulty process of selection
(b) Faulty work during the collection of information
(c) Faulty methods of analysis etc.

13
 UNIT-III
MEASURES OF CENTRAL TENDENCY
The point around which the observations concentrate in general in the central part of the data is called
central value of the data and the tendency of the observations to concentrate around a central point is
known as Central Tendency.

Objects of Statistical Average:

 To get a single value that describes the characteristics of the entire group
 To facilitate comparison

Functions of Statistical Average:

 Gives information about the whole group
 Becomes the basis of future planning and actions
 Provides a basis for analysis
 Traces mathematical relationships
 Helps in decision making

Requisites of an Ideal Average:

 Simple and rigid definition
 Easy to understand
 Simple and easy to compute
 Based on all observations
 Least affected by extreme values
 Least affected by fluctuations of sampling
 Capable of further algebric treatment

• arithmatic mean
MEAN • geometric mean
• harmonic mean

MEDIAN

MODE

ARITHMETIC MEAN ( )
Arithmetic Mean of a group of observations is the quotient obtained by dividing the sum of all
observations by their number. It is the most commonly used average or measure of the central

14
tendency applicable only in case of quantitative data. Arithmetic mean is also simply called “mean”.
Arithmetic mean is denoted by .

Merits DeMerits Uses

• It is rigidly defined. • The arithmetic mean is • When the frequency
• It is easy to calculate and highly affected by extreme distribution is symmetrical.
simple to follow. values. • When we need a stable average.
• It is based on all the • It cannot average the ratios • When other measures such as
observations. and percentages properly. standard deviation, coefficient
• It is readily put to • It cannot be computed of correlation are to be
algebraic treatment. accurately if any item is computed later.
• It is least affected by missing.
fluctuations of • The mean sometimes does
sampling. not coincide with any of the
• It is not necessary to observed value.
arrange the data in • It cannot be determined
ascending or descending by inspection.
order. • It cannot be calculated in
case of open ended classes.

MEDIAN (M)
The median is that value of the variable which divides the group into two equal parts, one part
comprising of all values greater and other of all values less than the median. For calculation of median
the data has to be arranged in either ascending or descending order. Median is denoted by M.

Merits DeMerits arithmetic mean.

• It is easily understood and easy to calculate. •For calculation, it is
• It is rigidly defined. necessary to arrange
•It can sometimes be located by simple data in ascending or
inspection and can also be computed descending order.
graphically. •Since it is a positional
•It is positional average therefore not affected average, its value is not
at all by extreme observations. determined by each
•It is only average to be used while dealing and every observation.
with qualitative data like intelligence, •It is not suitable for
honesty etc. further algebric
•It is especially useful in case of open end treatment.
classes since only the position and not •It is not accurate for large
the value of items must be known. data.
• It is not affected by extreme values. •The value of median is
more affected by
sampling fluctuations
than the value of the

15
Uses
•When there are open- ended classes
provided it does not fall in those classes.
•When exceptionally large or small values
occur at the ends of the frequency
distribution.
•When the observation cannot be
measured numerically but can be
ranked in order.
•To determine the typical value in the
problems concerning distribution of
wealth etc.

16
 MODE (Z)
Mode is the value which occurs the greatest number of times in the data. The word mode has been
derived from the French word ‘La Mode’ which implies fashion. The Mode of a distribution is the value
at the point around which the items tend to be most heavily concentrated. It may be regarded as the
most typical of a series of values. Mode is denoted by Z.

Merits DeMerits Uses

• It is easy to understand • It is not well defined. • When a quick
and simple to calculate. • It is not based on all approximate
• It is not affected by extreme the values. measure of
large or small values. • It is suitable for large values central tendency
• It can be located only by and it will not be well is desired.
inspection in ungrouped data defined if the data consists of • When the
and discrete frequency small number of values. measure of
distribution. • It is not capable of further central tendency
• It can be useful for qualitative mathematical treatment. should be the
data. most typical
• Sometimes, the data has
• It can be computed in open- value.
one or more than one mode
end frequency table. and sometimes the data has
• It can be located graphically no mode at all.

GEOMETRIC MEAN (G.M)

The geometric mean also called geometric average is the nth root of the product of n non-negative
quantities. Geometric Mean is denoted by G.M.

Properties of Geometric Mean:

 The geometric mean is less than arithmetic mean, G.M<A.M
 The product of the items remains unchanged if each item is replaced by the geometric mean.
 The geometric mean of the ratio of corresponding observations in two series is equal to the
ratios their geometric means.
 The geometric mean of the products of corresponding items in two series.

Merits of Geometric Mean:

 It is rigidly defined and its value is a precise figure.
 It is based on all observations.
 It is capable of further algebraic treatment.
 It is not much affected by fluctuation of sampling.
 It is not affected by extreme values.

Demerits of Geometric Mean:

 It cannot be calculated if any of the observation is zero or negative.
 Its calculation is rather difficult.
 It is not easy to understand.
 It may not coincide with any of the observations.

Uses of Geometric Mean:

17
  Geometric Mean is appropriate when:
 Large observations are to be given less weight.
 We find the relative changes such as the average rate of population growth, the average
rate of intrest etc.
 Where some of the observations are too small and/or too large.
 Also used for construction of Index Numbers.

HARMONIC MEAN (H.M)

Harmonic mean is another measure of central tendency. Harmonic mean is also useful for quantitative
data. Harmonic mean is quotient of “number of the given values” and “sum of the reciprocals of the
given values”. It is denoted by H.M.

Merits of Harmonic Mean:

 It is based on all observations.
 It not much affected by the fluctuation of sampling.
 It is capable of algebraic treatment.
 It is an appropriate average for averaging ratios and rates.
 It does not give much weight to the large items and gives greater importance to small items.

Demerits of Harmonic Mean:

 Its calculation is difficult.
 It gives high weight-age to the small items.
 It cannot be calculated if any one of the items is zero.
 It is usually a value which does not exist in the given data.

Uses of Harmonic Mean:

 Harmonic mean is better in computation of average speed, average price etc. under certain
conditions.

18
 UNIT IV
DISPERSION
The Dispersion (Known as Scatter, spread or variations) measures the extent to which the items vary
from some central value. The measures of dispersion is also called the average of second order (Central
tendency is called average of first order).
The two distributions of statistical data may be symmetrical and have common means, median or
mode, yet they may differ widely in the scatter or their values about the measures of central tendency.

Significance/ objectives of Dispersion-

 To judge the reliability of average
 To compare the two an more series
 To facilitate control
 To facilitate the use of other statistical measures.

Properties of good Measure of Dispersion

 Simple to understand
 Easy to calculate
 Rigidly defined
 Based on all items
 Sampling stability
 Not unduly affected by extreme items.
 Good for further algebraic treatment

Range
(coefficient of
Range)
Based on
selected Items
Inter-quartile,
coefficient of
Range (IQR)
Dispersion

Mean Deviation
Based on
all items
Standard
Deviation

1. Range: - Range (R) is defined as the difference between the value of largest item and value of
smallest item included in the distributions. Only two extreme of values are taken into
considerations. It also does not consider the frequency at all series.
2. Quartile Deviation: - Quartile Deviation is half of the difference between upper quartile (Q3) and
lower quartile (Q1). It is very much affected by sampling distribution.

19
3. Mean Deviation: - Mean Deviation or Average Deviation (Alpha) is arithmetic average of
deviation of all the values taken from a statistical average (Mean, Median, and Mode) of the series.
In taking deviation of values, algebraic sign + and – are also treated as positive deviations. This is
also known as first absolute moment.
4. Standard Deviation:- The standard deviation is the positive root of the arithmetic mean of the
squared deviation of various values from their arithmetic mean. The S.D. is denoted as  Sigma.

Distinction between mean deviation and standard deviation

Base Mean Standard

Deviation Deviation
Actual signs +, - are not ignored
Actual +, - Signs are ignored and
Algebric Sign whereas they are squared
all deviation are taken as positive
logically to be ignored.

Use of Measure Mean deviation can be computed Standard deviation is computed

from mean, median, mode through mean only

It is not capable of further It is capable of further algebraic

Further algebraic Treatment
algebraic treatment. treatment

M.D is simple to understand and S.D is somewhat complex than

Simplicity
easy to calculate mean deviation.

It is based on simple average of It is based on square root of the

Based
sum of absolute deviation average of the squared deviation

Variance
The square of the standard deviation is called variance. In other words the arithmetic mean of the
squares of the deviation from arithmetic mean of various values is called variance and is denoted as 2.
Variance is also known as second movement from mean. In other way, the positive root of the variance
is called S.D.
Coefficient of Variations- T o compare the dispersion between two and more series we define

coefficient of S.D. The expression is x 100 = known as coefficient of variations.
X
Interpretation of Coefficient of Variance-
Value of variance Interpretation
Smaller the value of Lesser the variability or greater the uniformity/ stable/ homogenous of
2 population
Larger the value of 2 Greater the variability or lesser the uniformity/ consistency of the population

20
 DISPERSION
RANGE = R
Individual Series Discrete Series Continuous Series
Range = L-S RLS RLS
Where L=Largest,
S=Smallest Observation
Coefficient of Range LS LS
LS LS LS
LS

QUARTILE DEVIATION - Q.D.

Individual Series Discrete Series Continuous Series
Q.D.  Q3  Q1 Q.D.  Q3  Q1 Q.D.  Q3  Q1
Q3  Q1
Coefficient of Q.D.  Q Q Q Q
 Q 3  Q1  Q 3  Q1
Q3  Q1 3 1 3 1

MEAN DEVIATION - M.D.  (“Through actual Mean, Mode, Median)

Individual Series Discrete Series Continuous Series
M Median 
dM  fdM  fdM
N N N
  
Coefficient of      
M M M
Mean  X 
 dx
 fdx  fdx
N N N
  
Coefficient of X 
X X X
(Mode) Z 
dz  fdz  fdz
N N N
  
Coefficient of Z
Z Z Z

Standard Deviation =  can be calculated through mean only

Individual Series Discrete Series Continuous Series
Direct
(Through actual 
dx 2 fd 2   fd 2

mean) N f  f
2 2 2
Indirect (Through
assumed mean)
 dx 2
  dx 
 
 fdx 2
  fdx   fdx 2
  fdx 
N  N    
   

f   f  f   f 

21
 SKEWNESS

Skewness is a measure of symmetry, or more precisely, the lack of symmetry. A distribution, or data set,
is symmetric if it looks the same to the left and right of the center point.

Skewness is positive if the tail on the right side of the distribution is longer or fatter than the tail on the
left side. The mean and median of positively skewed data will be greater than the mode. Skewness is
negative if the tail of the left side of the distribution is longer or fatter than the tail on the right side. The
mean and median of negatively skewed data will be less than the mode. If the data graph symmetrically,
the distribution has zero skewness, regardless of how long or fat the tails are.

Karl Pearson developed two methods to find skewness in a sample:

1. Pearson’s Coefficient of Skewness #1 uses the mode. The formula is:

Where = the mean, Mo = the mode and s = the standard deviation for the sample.

2. Pearson’s Coefficient of Skewness #2 uses the median. The formula is:

Where = the mean, Mo = the mode and s = the standard deviation for the sample.
It is generally used when you don’t know the mode.

22
 TIME SERIES ANALYSIS

“A Time Series” is a series of statistical data recorded in accordance with their time of occurrence. Here
it is noted that it is a set of observation taken at specified times usually (but not always) at equal
intervals. Thus a set of data depending on the time (which may be year, quarter, month, day etc.) is
called a “Time Series”.
Today the use of time series analysis is not merely confined to economists and businessmen, but it
extensively used by scientists, sociologist, biologists, geologists, research workers etc.

Some example of time series are

i. The population of a country in different years.
ii. The annual production of coal in India over the last ten years.
iii. Deposits received by bank in a year.
iv. The daily closing price of a share in the Bombay Stock Exchange.
v. The monthly sales of departmental store for the last six months.
vi. Hourly temperature recorded by the store for the last six months.

According to Patterson “A timeseries consists of statistical data which are collected. Recorded or
observed over successive increments.

Utility or importance of Time Series

The very important use of time series analysis is its use in forecasting future information and behavior.

i. It enables us to predict or forecast the behavior of the phenomenon in future, which is very
essential for business planning. On the basis of past information, the trend can be estimated and
projections can also be made for the uncertain future. It assists in reducing, the risk and uncertainties
of business and industry.

ii. It helps in the evaluation of current achievement by review and evaluation of progress made
through a plan can be done on the basis of time series.

iii. It helps in the analysis of past behavior of the phenomenon under consideration. What changes
had taken place in the past, what factor were responsible for these changes, under that conditions these
changes took place, etc. are certain issues which could be studied and analyzed by time series.

iv. It helps in making comparative studies in the values of different phenomenon at different times
or place. It provides a scientific basis for making comparison by studying and isolating the effects of
various components of a time series.

v. The segregation and study of the various components of time series is of paramount importance
to a businessman in the planning of future operations and the formulation of executive and policy
decisions.

vi. On the basis of the past performance of the various sectors of economy, we can determine
future requirements and a suitable policy can be formulated to get desired and predetermined
objectives.

Causes of variation in time series

23
 If the values of a phenomenon are observed at different periods of time, the values so obtained will
show appreciable variations.
The following factors are generally affecting any time series are :
i. Changing of tastes, habits and fashions of the people.
ii. Changing of customs, conventions of the people.
iii. Rituals and festivals.
iv. Political movements, government policies.
v. War, Famines, Drought, Flood, Earthquakes and Epidemic etc.
vi. Unusual weather or seasons.

Components of Time Series

A time series may be defined as a collection of readings belonging to different time periods of some
economic variable or composite of variable.
Eg. The retail price of a particular commodity are influenced by a number of factors namely the crop
yield which further depends on weather conditions, irrigation facilities, fertilizers used, transportation
facilities, consumer demand etc.

The various forces affecting the values of a phenomenon in a time series may be broadly classified into
the following four categories, commonly known as the components of a time series.
i. Secular Trend (i.e. long-term smooth, regular movement)
ii. Seasonal Variation (periodic movement, the period being not greater than one year)
iii. Cyclical Variation (periodic movement with period greater than one year)
iv. Irregular or Random Variation.
1. Secular Trend: - It is the matter of common sense that there might be violent variations in a time
series during a short span of time, however in a long run, it has a tendency either to rise or fall. This
tendency or trend of variation may be either upward or down set on over a long time period. This is
known as ‘Secular trend’ or ‘Simple trend. It is but natural that population growth, Technological
progress medical facilities production, prices etc. are not judge over a day, month or year they shores.
The movement are upward, downward or constant over a fairly long period.

Broadly the trends are divided under two heads:

1. Linear Trends: - If we plot the values of time series on graph it shows the straight line i.e. growth
rate is constant. Although in practice linear trend is commonly used but it is rearely found in
economics and business data.
2. Non-Linear Trends: In business or economics generally growth is slow in the begging and them it
is rapid for some time period after which it becomes stable for some time period and finally retards
gradually. It is not linear it forms a curve known as non linear trends.

Seasonal Variation: As we read season the first things comes in our mind is spring, summer, autumn
and winter. Generally seasonal variations occur due to changes in weather condition, customer,
tradition fashion etc.
Seasonal variations represent a periodic movement where the period is not longer than one year. The
factors, which mainly cause this type of variation in time series, are the climatic changes of the different
seasons. For example
i. Sale of woolensgo up in winter.
ii. Sale of raincoat and umbrella go up in rainy season.
iii. Prices of food grains decrease with the arrival of new crop.
iv. Sale of cooler, refrigerator etc. rise during the summer season.
Another variation occurs due to man-made convention and customs, which people follow at different
times like DurgaPooja, Dashehra, Deepawali, Ide. X-Max etc. The seasonal variations may take place per
day per week or per month. For example:
i. Sale of departmental stores go up in festivals.

24
 ii. Sale of cloths and Jewelry pick up in marriages.
iii. Sale of Paint, furniture and electronics goes up during festivals like, Deepawali, Ide, X-max etc.
iv. Sale of vehicles increase considerably during DurgaPooja and Dasherhra.

Cyclical Variations: Most of the business activities are often characterized by recurrence of periods of
prosperity and slump constituting a business cycle. Cyclical variations are another type of periodic
movement, with a period more than one year. Such movements are fairly regular and oscillatory in
nature. One complete period is called a ‘cycle’ cyclical variations are not as regular as seasonal
variation, but the sequence of changes, marked by prosperity, decline, depression and recovery,
remains more of less regular.

PHASES OF A BUSINESS CYCLE

Irregular or Random Variation: Irregular or random variation are such variation which are
completely unpredictable in character. These are caused by factors which are either wholly
unaccountable or caused by such unforeseen events like Earthquakes, flood, drought famines, epidemic
etc, and some man-made situations like strikes lock-outs wart etc.

Mathematical Models for Analysis of Time Series

Though there are many models by which a time series can be analyzed, two models commonly used for
decomposition of a time series into various components are
1. Additive Model: - According to the additive model, the decomposition of time series is done on the
assumption that the effect of various components are additives in nature, i.e. U = T+S+C+R
Where, U, is the time series value and T, S, C, and R stand for trend seasonal, cyclical and random
variation.
In this model ‘S, C and R are absolute quantities and can have positive or negative values. The model
assumes that the four components of the time series are independent of each other and non-has any
affect whatsoever on the remaining three components.

2. Multiplication Model: According to the multiplication model, the decomposition of a time

series on the assumption that the effects of the four components of a time series (T, S, C and R) are not

25
 necessarily independent of each other. In fact, the model presumes that their effects are interdependent
U=T×S×C×R
Measurement of Trend or Secular Trend
The different methods of determining the trend component of a time series are:

1. Moving Average Method: Moving average method is very commonly used for the isolation of trend
and in smoothing out fluctuations in time series. In this method, a series of arithmetic means of
successive observation, known as moving averages, as calculated from the given data, and these

a+b+c3 b+c+d3 c+d+e3 d+e+f3

moving average are used as trend values. Yearly moving average is given by :

Illustration1 Calculate 3 yearly moving averages:

Years : 1979 1980 1981 1982 1983 1984 1985 1986
Earning(Lakhs) : 80 90 70 60 110 50 40 30

Working Rule
i. Add the values of the first3 years (namely 1979, 1981 i.e., 80+90+70=240) and place the total
against the middle year1980.
ii. Leave the first year’s value and add up the values of the next 3 years (i.e., 1980, 1981, 1982, viz.,
90+70+70+60 = 220) and place the total against the middle year i.e., year 1981.

Illustration2 Calculate 5 yearly moving averages and seven year moving average for the following
data:
Year : 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
Sales (‘000 Rs.) : 123 140 110 98 104 133 95 105 150 135

Calculation of Moving Averages when the Period is Even:

If the period of the moving average is even, centre point of the group will lie between two years. It is,
therefore, necessary to adjust or shift (technically known as centre) these average so that they coincide
with the years. For example
4-yearly moving average is calculated as:
Step 1 : Add the values of first four year, and place the total between the 2 and 3 year.
nd rd

Step 2 : Leave the first year value and then add the for values of the next four years and place the total
in between the 3 and 4 year Continue this process until the last year is taken into account.
rd th

Step 3 : Divide 4 yearly moving totals 4. It will give 4 yearly moving average.
Step 4 :Add first two moving averages and divide it by 2 to get the moving average centered. Place it
against 3 year. Leave the first moving average and then add next two moving average and divide by 2
rd

to get the next moving average centered. Place it against the 4 Year. Continue this process till the last
th

moving average is included.

Alternative Procedure: In this procedure step 1 and 2 are same as above.
Step 3: Addfirst two 4 yearly moving total place it against 3 year. Leave the first moving total and then
rd

add nexttwo moving total to get the next moving total centred. Place it against the 4 year. Continue this
th

process till the last moving total is included.

Step 4 : Diving these centered moving totals by 8. It will give 8 yearly moving average. This procedure
will more clear by following illustration.

Illustration Construction a four-yearly centered moving average from the following data :
Year : 1970 1975 1980 1985 1990 1995 2000

26
 Imported Cotton (in ‘000) : 129 131 106 91 95 84 93

Method of Least Squares

It is an appropriate mathematical technique to determined an equation which best fits on a given
observation relating to two variables. In this procedure for fitting a live to a set of observation the sum
of the squared deviations between the calculated and observed values in minimised. Therefore the
technique is named as “Least-Squares method.” And the line so obtained is known as ‘Best fit line”.
We know that sum of the deviations from the arithmetic mean is zero. Therefore the sum of the

(y-c)=0, i.e., the sum of the deviations of the actual values of y and computed values of y is zero.
deviations from the line of the best fit is zero.

(y-y)2=is least, i.e., the sum of the squares of deviations from the actual and the computed
i.
ii.
value of y is least.
That is why it is called the method of least squares and the line obtained by this method is called the
‘line of best fit’
This method may be used either to fit a straight line trend or parabolic trend straight line trend is
represented by the equation y= a + bx where y represents the estimated values of the trend x
represents the deviations in the time period. A and b are constants.
‘a’ represents intercept of the line of the y no is and ‘b’ represent the slope of the line i.e. it gives the
changes in the value of y for per unit change in the value of x if b>0 it show and growth rate and if b<0 it
shows decline rate.

Merits:
1. This is the only method of measuring trend which provides the future values authentically very
convincing and reliable.
2. This method is used for forecasting the series for example.
3. If other factors are not so effective no share market, this method can provide very reliable
information about the movement of the share of a company.
4. This method has no scope for personal bias of the Investigator.
5. It is only method which gives the rate of growth per annum.

Demerits:-
1. The method required mathematical ability. Some items it involves tedious and complicated
calculations.
2. The method has no flexibility i.e. if even a single term is added to series it makes necessary to do
all the calculations again.
3. Estimations and predictions by this method are based only on long term variations and the
impact of cyclical, seasonal and irregular variations are completely ignored.

Computation of Trend Values by the Least Squares Method

We know straight lines trend is given by y= a+bx in order to determine the values of the constants and
b the following two normal equations are to be solved.

27
 ∑Y=na+b∑X

∑XY = a∑X+b∑X 2

y sum of actual values of y variable.

Where n represents number of years (months or any other period) for which data are given:

y represents sum of deviations from the origin.

y x represents sum of deviations from the origin.
xy represents sum of the deviations from the origin and actual values.
2

taken as origin and deviations are taken from the middle time period it provides x=0 the above
Remarks :-The variable x can be measured from any point of time as origin. But if middle time period is

y=na+xy=na+0=naThus a = Yn
normal equation would be reduced to the

xy=ax+bx2+0+bx2= xy=b=x2 Thus b = xy=x2

UNIT-V
CORRELATION

Introduction
1. Correlation is a statistical tool & it enables us to measure and analyse the degree or extent to
which two or more variable fluctuate/vary/change w.e.t. to each other.
2. For example – Demand is affected by price and price in turn is also affected by demand.
Therefore we can say that demand and price are affected by each other & hence are correlated.
the other example of correlated variable are –
3. While studying correlation between 2 variables use should make clear that there must be cause
and effect relationship between these variables. for e.g. – when price of a certain commodity is
changed ( or ) its demand also changed ( or ) so there is case & effect relationship between
demand and price thus correlation exists between them. Take another eg. where height of
students; as well as height of tree increases, then one cannot call it a case of correlation
because neither height of students is affected by height of three nor height of tree is affected by
height of students, so there is no cause & effect relationship between these 2 so no correlation
exists between these 2 variables.
4. In correlation both the variables may be mutually influencing each other so neither can be
designated as cause and the other effect for e.g. –
Price  Demand 
Demand  Price 
So, both price & demand are affected by each other therefore use cannot tell in real sense which
one is cause and which one is cause and which one is effect.

DEFINITIONS OF CORRELATION
1. “If 2 or more quantities vary is sympathy, so that movements is one tend to be accompanied by
corresponding movements in the other(s), then they are said to be correlated”. Connor.
2. “Correlation means that between 2 series or groups of data there exists some casual
correction”. WI King
3. “Analysis of Correlation between 2 or more variables is usually called correlation.”A.M. Turtle
4. “Correlation analysis attempts to determine the degree of relationship between variables.
YaLunchou
TYPES OF CORRELATION

28
 Correlation

Positive Negative Correlation

Simple & Multiple Correlation
Partial & Total Correlation
Liner & Non Linear Correlation

POSITIVE CORRELATION NEGATIVE CORRELATION

1 Value of 2 variables move in the same direction Value of 2 variables move in opposite direction
i.e. when increase/decrease in value of one i.e. when one variable increased, other variable
variable will cause increase or decrease in value decreaseswhen one variable is decreased, other
of other variable. variable increase.
2 E.g. Supply & Price E.g. Demand & Price

So, supply and price are …….correlated So, Demand & Price vely correlated
P = Price/Unit P = Price/Unit
Q = quantity Supplied Q = quantity Supplied

SIMPLE CORRELATION MULTIPLE CORRELATION

1 In simple correlation, the relationship is The relationship between more than 2
confined to 2 variables only, i.e. the effect of variables is studied.
only one variable is studied
2 E.g. Demand& Price E.g. Demand & Price
Demand depends on  Price Demand depends on  Price
Demand on  income
This is case of simple correlation because This is case of multiplecorrelationsbecause
relationship is confined to only one factor (that 2factors (Price & Income) that affects demand
affects demand) i.e. price so we have to find are taken.We have to find correlation between
correlation between demand & price. demand & price.
If, demand = Y Demand & Price
If, demand – X If, demand = Y
Then, Correlation between Y & X Price = X1
Price = X2
Then
Correlation between Y & X1
Correlation between Y & X2

SIMPLE CORRELATION MULTIPLE CORRELATION

In partial correlation though more than 2 factors In total correlation relationship between all
are involved but correlation is studies only the variables is studied i.e., none of item is
between to be constant. assumed to be constant
E.g. E.g.
X1 Y = Demand X1 Y = Demand
Y X1 = Price Y X1 = Price
X2X2 = Income X2X2 = Income
If we study correlation between Y & X1 & assume If we assume that income is not constant i.e.
X2 to be constant it is a case of partial we study the effect of both price & income on
correlation. this is what we do in law of demand – demand, it is a case of total correlation.
assume factors other than price as constant In other words, cataris paribus assumption is
(Ceteris paribus – Keeping other things constant) relaxed in this case.

29
 LINEAR CORRELATION NON-LINEAR CORRELATION
1 In linear correlation, due to unit, change value of In non linearor curvilinear correlation, due to
one variable there is constant change in the unit, change value of one variable, the change in
value of other variable. The graph for such a the value of other variable is not constant. the
relationship is straight line. E.G. – If in a factory graph for such a relationship is a curve. E.G. –
no of workers are doubled, the production The amount spent on advertisement will not
output is also doubled, and correlation would be bring the change in the amount of sales in the
linear. same ratio, it means the variation.
2 If the changed in 2 variables are in the same If thechange in 2 variables is in the same
direction and in the constant ratio, itis linear direction but not in constant ratio, the
positive correlation correlation is non linear positive.
X Y Y X Y
Y
2 3 50 10
4 6 55 12
6 9 60 15
8 12 90 30
X 100 45 X
3 If changes in 2 variables are in the opposite If changes in 2 variables are in opposite
direction but in constant ratio, the correlation is direction and not in constant ratio, the
linear negative. For eg. every 5%  is price of a correlation is non linear negative. For eg: -
good is associated with 10% decrease in every 5% in price of good is associated with
demand the correlation between price and 20% to 10%in demand, the correlation
demand would be linear negative. between price & demand would be non linear
X Y Y negative.
Y
2 21 X Y
4 18 80 50
6 15 55 60
8 12 50 75
10 9 X 90 130 X

30
TYPE – 1 [BASED ON KARL PEARSON’S COFFICIENT OF CORRELATION]
Before use move to numerical, use understand the basic notions & concepts –
dx = Deviations of xi value from mean =(xi -𝑥̅)
x = Mean of x value [Average of X values] =𝑛xi

Deviation of y value from mean = (y - 𝑦̅ )

n = No. of observations
dy =
𝑦̅ Mean of y values = 𝑛yi
Square of deviation of x values = (xi- 𝑥̅ )2
=

Square of deviation of x values = (yi- 𝑦̅ )2

d2x =

Product of deviations = (xi - 𝑥̅ ) (yi - 𝑦̅ )

d2y =
dxdy =

Covariance (x,y) = (xi − 𝑥̅ ) (yi − 𝑦̅ )

𝑛 2
 = Variance of x values =(xi − 𝑥̅)

2
Variance of y values = (yi 𝑛− 𝑦̅ )
x i
 =
𝑛
y i
r or rxy = coefficient of correlation between x 7 y variables.

Direct Method for Karl Pearson’s Coefficient of correlation

31
Deviation from actual mean method

Deviation from assumed mean method (Short Cut Method)

This method is used in the situation where mean of any series (x or y) is not in whole number, i.e. in
decimal value. in this case it is advisable to take deviation from assumed mean rather than actual mean
and then use the above formula.
In the above short cut method
Let, A = Assumed mean of X series
B = Assumed mean of y series
thendx = (xi – A) &dy = (yi – B) &
dx2= (xi – A)2&dy2= (yi – B)2
dxdy= (xi – A)(xi – B)

REGRESSION ANALYSIS
The dictionary meaning of regression is “Stepping Back”. The term was first used by a British
Biometrician” Sir Francis Galton 1822 – 1911) is 1877. He found in his study the relationship between
the heights of father & sons. In this study he described “That son deviated less on the average from the
mean height of the race than their fathers, whether the father’s were above or below the average, son
tended to go back or regress between two or more variables in terms of the original unit of the data.

Meaning
Regression Analysis is a statistical tool to study the nature extent of functional relationship between
two or more variable and to estimate the unknown values of dependent variable from the known
values of independent variable.

Dependent Variables – The variable which is predicted on the basis of another variable is called
dependent or explained variable (usually devoted as y)

Independent variable – The variable which is used to predict another variable called independent
variable (denoted usually as X)

32
Definition
Statistical techniques which attempts to establish the nature of the relationship between variable and
thereby provide a mechanism for prediction and forecasting is known as regression Analysis.
– Ya-lun-Chon”
Importance/uses of Regression Analysis
 Forecasting
 Utility in Economic and business area
 Indispensible for goods planning
 Useful for statistical estimates.
 Study between more than two variable possible
 Determination of the rate of change in variable
 Measurement of degree and direction of correlation
 Applicable in the problems having cause and effect relationship
 Regression Analysis is to estimate errors
 Regression Coefficient (bxy&byx) facilitates to calculate of determination ® & coefficient or
correlation (r)

Regression Lines
The lines of best fit expressing mutual average relationship between two variables are known as
regression lines – there are two lines of regression

Why are two Regression lines –

1. While constructing the lines of regression of x on y is treated as independent variables where
as ‘x’ is treated as treated as dependent variable. This gives most probable values of ‘X’ for gives
values of y. the same will be there for y on x.

1. Whenthereis perfect correlation between two series (r = ± 1) theregression

RELATIONSHIP BETWEEN CORRELATION & REGRESSION

withcoincideandtherewill be only one regression line.

3. Where there is more degree of correction, say (r = ±70 or more the two regression line with be
2. When there is no correction (r = o)>Both the lines will cut each other at point.

next to each other whereas when less degree of correction. Say (r=± 10 on less) the two
regression line will be a parted from each other.

REGRESSION LINES AND DEGREE OF CORRELATION

DIFFERENCE BETWEEN CORRELATION AND REGRESSION ANALYSIS

33
Thecorrelationandregressionanalysis, both, help us in studying the relationship between two variables
yet they differ in their approach and objectives. The choice between the two depends on the purpose of
analysis.
S.NO BASE CORRELATION REGRESSION
1 MEANING Correlation means relationship between Regression means step ping back
two or more variables in which or returning to the average value,
movement in one have corresponding i.e., it express average
movements in other relationship between two or more
variables.
2 RELATIONSHIP Correlation need not imply cause and Regression analysis clearly
effect relationship between the variables indicates the cause and effect
under study relationship. the variable(s)
constituting causes(s) is taken as
independent variables(s) and the
variable constituting the variable
consenting the effect is taken as
dependent variable.
3 OBJECT Correlation is meant for co-variation of Regression tells use about the
the two variables. the degree of their co- relative movement in the variable.
variation is also reflected in correlation. We can predict the value of one
but correlation does not study the variable by taking into account
nature of relationship. the value of the other variable.
4 NATURE There may be nonsense correlation of There is nothing like nonsense
the variable has no practical relevance regression.
5 MEASURE Correlation coefficient is a relative Theregression coefficient is
measure of the linear relationship absolute measure representing
between X and Y. It is a pure number the change in the value of
lying between 1 and +1 variable. We can obtain the value
of the dependent variable.
6 APPLICATION Correlation analysis has limited Regression analysis studies linear
application as it is confined only to the as well asnonlinearrelationship
study of linear relationship between the between variables and therefore,
variables. has much wider application.

Why least square is the Best?

When data are plotted on the diagram there is no limit to the number of straight lines that could be
drawn on any scatter diagram. Obviously many lines would not fit the data and disregarded. If all the
points on the diagram fall on a line, that line certainly would the best fitting line but such a situation is
rare and ideal. Since points are usually scatters, we need a criterion by which the best fitting line can be
determined.

Methods of Drawing Regression Lines –

1. Freecurve –
2. Regressionequationxony,
X = a + by..........................................(1)
3. Regressionequationyonx
Y = a + bx
Where
‘a’ isthatpointwhereregressionlinestouchesyaxis (thevalueofdependentvariablevalue when
value or independent variable is zero)
‘b’ is the slop of the said line (The amount of change in the value of the dependent variable per
unit change)

34
Change in independent variable)
A and b constants can be calculated through –
(x = a + by) (by multiplying ‘’)
x = Na + by (1)

x (y = a + bx) (by multiplying x)

xy = xa + bx2 (2)

KINSDS OF REGRESSION ANALYSIS

1. Linear and Non- Linear Regression
2. Simple and Multiple Regression

FUNCTIONS OF REGRESSION LINES –

1. Tomakethebestestimate –
2. Toindicatethenatureandextentofcorrelation

REGRESSION EQUATIONS –
Theregressionequation’sexpresstheregressionlines, asthere are two regression lines there are two
regression equations –
Explanation is given in formulae –

REGRESSION LINES
1. Regressionequationofxony
X – X = bxy (y – y)
Wherebxy = regressioncoefficientofXonY
2. Regressioneuationofyonx
Y – Y = bxy (x – x) wherebxy = regressioncoefficientofY onX

35
REGRESSION COEFFICIENT –Thereare two regression coefficient like regression equation, they are
(bxy and byx)
Properties of regression coefficients –
 Same sign – Both coefficient have the same either positive on negative
 Both cannot by greater than one – If one Regression is greater than “One” or unity. Other must
be less than one.
 Independent of origin – Regression coefficient are independent of origin but not of scale.
 A.M.> ‘r’ – mean of regression coefficient is greater than ‘r’
 R is G.M. – Correlation coefficient is geometric mean between the regression coefficient
 R, bxyand bxy – They all have same sign

INDEX NUMBERS
Index numbers are devices which measure the change in the level of a phenomenon with respect to
time, geographical location or some other characteristic. The first index number was constructed in the
year 1764 by an Italian named Carli to compare the changes in the price for the year 1750 with the
price level of the year 1500. In present day situation changes in production, consumption, exports,
imports, national income, cost of living, incidence of crimes, number of road accidents, business failures
and a very wide variety of other phenomena are studied with the help of index numbers. Index
numbers are supposed to be barometers which measure the change in the level of a phenomena.
“An index number is a statistical measure designed to show changes in variable or a group of related
variables with respect to time, geographical location or other characteristics.”

CHARACTERISTICS OF INDEX NUMBERS

1. Index number are a specialised type of average. Averages can be used to compare only
those series which are expressed in the same units. However the device of index number helps
us in comparing change in series which are in different units.

36
 2. Index numbers study the effects of such factors which cannot be measured directly. Index
numbers are meant to study the changes in the effects of such factors which cannot be
measured directly.
3. Index numbers being out the common characteristics of a group items.
4. Index number measure only relative changes in the values of a phenomenon.

USES OF INDEX NUMBERS

1. Help in Studying Trends. Index numbers helps to find out the trend of exports, imports,
balance of payments, industrial production, prices, national income and a variety of other
phenomena.
2. Help in policy formulation. Index numbers help us in studying trends of various phenomena
and these trends and tendencies are the bases on which may policy decisions are taken index
number are used by the government in deciding the rates of D.A. and levy of excise duties.
3. Help in measuring the Purchasing Power of Money. Index numbers are helpful in finding out
the intrinsic worth of money as contrasted with its nominal worth.
4. Helps in deflating various value. Index numbers are very helpful in deflating national income
on the basis of constant prices.
5. Act as economic barometers. Index numbers measure the pulse of an economy and act as
barometers to find the ups and down in the general economic condition of a country.

PROBLEMS IN THE CONSTRUCTION OF INDEX NUMBERS

1. The selection of item- The first problem which the marker of an index number of wholesale
prices has to face is that of the selection of items from which the index number is to be
constructed.
2. The selection of the base year- Second problem in the construction of index numbers is the
selection of the base year and the conversion of current prices to price relatives based on the
prices of the base year.
3. The selection of the average- The next step in the construction of wholesale price index
number is to average the prices relatives of the various commodities.
4. Selecting suitable weights. All the items used in the construction of an index number are not
of equal importance and as such if the index number is to be a representative one, weights
should be assigned to various items in relation to their importance.

METHODS OF CONSTRUCTING INDEX NUMBERS

Broadly speaking various methods of constructing index numbers can be classified in two groups viz.
A. Unweighted Index Numbers
B. Weighted Index Numbers
i. Simple Aggregative Method
ii. Simple Average of Relatives Method.

A. Unweighted Index Numbers

Where
P = Index number of the current year
01

= Total of the current year; price of all commodities.

= Total of the base year’s price of all commodities.

37
 Simple Average of Relatives Method.

B. Weighted Index Numbers

1. Laspeyres Method - 2. Passche’s Method

3. Drobish and Bowleys Method

4. Fisher’s Ideal Index.

5. Marshall-Edgeworth formula

or

6. Walsch Formula. 7. Kelly’s Method.

Quantity Index Number

Quantity index number measure the changes in the volume of production, construction or employment
over a period of years.
Formula for simple or unweighted quantity index;

Here

Base Shifting:- Base shifting is generally required due to following reasons

i. The base year is too old to compare the current year.
ii. If different series of index numbers are based on different base years and they are to be
compared from each other.

Deflation of index numbers

Computation of real wages from money income with taking the effect of price level changes is called as
deflating of index numbers.

Splicing: Sometimes series of index number based on a certain year is discontinued and a new series of
index number is prepared by taking another year as base. Thus two series of index number would
result. In this situation index number of these two series are not comparable because both are based

38
on different years. If these are to be compared then new series will be covered on the basis of old series
or vice-versa; this conversion/shifting is called as spicing. Splicing may be taken as another form of
base shifting.
Formula for splicing :-
a. Splicing of new series in old series (Forward splicing):

Spliced Index Number

b. Splicing of old series in new series (backward splicing):

TESTS OF ADEQUACY OF INDEX NUMBER FORMULAE

We have discussed a large number of formulae for the construction of both simple and weighted index
numbers. We formula should be chosen for the construction of an index number is a question which can
not be satisfactorily answered. However some tests have been suggested to determine the adequacy of
an index number formula. These tests are:
1. Unit Test – This test requires that the formula for the construction of index numbers should be
such which is not affected by the unit in which prices or quantities have been quoted. This test is
satisfied by all index number formulae discussed above except the simple (unweighted) aggregative
index formula. In this index as we have discussed earlier the units play an important part in
determining the value of the index. If only the unit is changed (say from kg to quintal) the value of
the index would change.

2. Time Reversal Test- In the worlds of Fisher: “The test is that the formulae for calculating an
index number should be such that it will give the same ratio between one point of comparison and the
other no matter which of the two is taken as base.” This mean that the index number should work both
backwards as well as forwards. Thus, if the index number of the current year is 4000 then the index
number of the base year (based on the current year) should be 25. In other words, the two index
numbers thus calculated (without the figure 100) should be reciprocals of each other. The reciprocal of
4 is .25 and the reciprocal of .25 is 4. The product of these two ratios would always be equal to one.
Thus, if P represents the price change in the current year and P the price change of the base
10 10

year (based on the current year) the following equation should be satisfied:-

3. Factor Reversal Test- In the words of Fisher: “Just as each formula should permit inter-
changing the price and quantities without giving inconsistent result, i.e., the two results multiplied
together should give the trust value ratio.” It means that the changes in the prices multiplied by the
changes in quantity should be equal to the total change in value. Change in value is the result of
changes should represent the total change in value. Thus, if the price of a commodity has doubled
during a certain period and if in this period the quantity has trebled the total change in the value should
be six time the former level. In the other words, if p and p represent the prices and q and q
1 0 1 0

the quantities in the current and the base years respectively, and if p represent the change in price in
01

the current year and q the change in the quantity in the current year then
01

The factor reversal test is satisfied only by the Fisher’s Ideal Index Number.
The proof of it is given below:

39
Circular Test
Another test applied in index number studies is the circular test. It is a short of extension of the time
reversal test. Suppose an index number is constructed for the year 1983 with the base of 1982 and
another index number for 1982 on the base of 1981, then it should be possible for us to directly get an
index number for 1983 on the base of 1981. If the index number calculated directly does not give an
inconsistent value, the circular test is said to be satisfied. If p represent the price change of the current
01

year on the base year and P the price change of the base year on some other base and p the
12 20

price change of the current year on this second base then the following equation should be satisfied.

40