Beruflich Dokumente
Kultur Dokumente
MB0024-Unit-01-Understand the
usefulness of Statistics
Introduction
Learning Objective 1
Our day-to-day activities interact with personnel, public, social, political, economic,
business and other environments. Decision making encompasses all these activities.
Suppose we wish to purchase a television we would like to know the price, quality,
durability, maintainability etc. Therefore there is a need for collecting data and making an
optimum decision. Again suppose a company wishes to introduce a new product, it has to
collect data on market potential, consumer likings, availability of raw materials,
feasibility of producing the product etc. In other words data collection is the back-bone of
any decision making process. Many organizations find themselves data-rich but poor in
drawing information from it. Therefore it is important to develop the ability to extract
meaningful information from raw data to make better decisions. Statistics play an
important role in this aspect.
Learning Objective 2
Learning Objective 3
Characteristic of Statistics
Statistics Deals with aggregate of facts: Single figure cannot be analyzed. Thus, the fact
‘Mr Lee is 170 cms. tall’ cannot be statistically analyzed. On the other had, if we know
the heights of 60 students of a class, we can comment upon the average height, variation,
etc.
1. Statistics are affected to a marked extent by multiplicity of causes: The statistics
of yield of paddy is the result of factors such as fertility of soil, amount of rainfall,
quality of seed used, quality and quantity of fertilizer used, etc.
2. Statistics are numerically expressed: Only numerical facts can be statistically
analyzed. Therefore, facts as ‘price decreases with increasing production’ cannot
be called statistics.
3. Statistics are enumerated or estimated according to reasonable standards of
accuracy: The facts should be enumerated (collected from the field) or estimated
(computed) with required degree of accuracy. The degree of accuracy differs from
purpose to purpose. In measuring the length of screws, an accuracy upto a
millimeter may be required, whereas, while measuring the heights of students in a
class, accuracy upto a centimeter is enough.
4. Statistics are collected in a systematic manner: The facts should be collected
according to planned and scientific methods. Otherwise, they are likely to be
wrong and misleading.
5. Statistics are collected for a pre-determined
purpose: There must be a definite purpose for collecting facts. Eg. Movement of
wholesale price of a commodity.
6. Statistics are placed in relation to each other: The facts must be placed in such a
way that a comparative and analytical study becomes possible. Thus, only related
facts which are arranged in logical order can be called statistics.
Learning Objective 4
Learning Objective 5
1. Statistics does not deal with qualitative data. It deals only with quantitative data.
2. Statistics does not deal with individual fact: Statistical methods can be applied
only to aggregate to facts.
3. Statistical inferences (conclusions) are not exact: Statistical inferences are true
only on an average. They are probabilistic statements.
4. Statistics can be misused and misinterpreted: Increasing misuse of Statistics has
led to increasing distrust in statistics.
5. Common men cannot handle Statistics properly: Only statisticians can handle
statistics properly.
With the advent of computers lot of Statistical programmes are available in the market.
They help us in summarizing, presenting and analyzing the mass data in short time. Some
of them are Minitab, SPSS, Texto & Contexto, Excel, E-View etc.
SUMMARY
Decision making process become more efficient with the help of Statistics. It deals with
aggregate of facts. It is applied in all fields of our activities more efficiently. Its
interpretation requires skilled and experienced statistician
MB0024-Unit-02
Introduction
Learning Objective 1
A. Planning B. Execution
The relevance and accuracy of data obtained in a survey depends upon the care exercised
in planning. A properly planned investigation can lead to best results with least cost and
time. The planning stage consists of the following sequence of activities.
Learning Objective 2
Examples:
Learning Objective 3
a. Collection of data is the first and most important stage in any statistical
survey
c. Data collected for the first time keeping in view the objective of the
survey is known as primary data. They are likely to be more reliable. However
cost of collection of such data are much higher.
Advantages are, we get i) original data ii) more accurate and reliable iii)
Satisfactory information can be extracted by the investigator through indirect
questions iv) Data are homogeneous and comparable v) additional information
can be gathered and vi) Misinterpretation of questions can be avoided.
However it consumes more time and cost.
f. Indirect oral interview is used when area to be covered is large. The data
is collected from a third party or witness or head of institution. This method is
generally used by police department.
Advantages are i) economical in terms of time, cost and man power, ii)
confidential information can be collected, iii) information is likely to be
unbiased and reliable. However the degree of accuracy of information is less.
It is used where the area to be covered is very large and periodic information
is required. The information is likely to be affected by the bias of the
correspondents or agencies.
This method is generally adopted by research workers and other official and non-
official agencies. It covers large area of investigation. It is more economical and
free from investigator’s bias. However it results in many “non-response”
situations. The respondent may be illiterate. They can provide wrong information
due to wrong interpretation of questions.
v. It should be unambiguous.
k. The information used for the investigation of the current problem and
obtained from the data collected and used by some other agency or person
before for his investigation is known a secondary data.
SUMMARY
A Statistical survey is a search for knowledge. There are two main stages in any
statistical survey, namely, planning and execution. Planning encompasses i)
nature of problem, ii) the objectives, iii) the scope, iv) statistical units, v) degree
of accuracy, vi) period, vii) source of information and viii) organization.
MB0024-Unit-03
Introduction
Collected data in the raw form would be voluminous and non-comprehensible. Therefore
it should be condensed and simplified for better understanding and usefulness.
Classification is first stage in simplification.
Each of the group is called class. For example in survey of Industrial workers of a
particular industry, workers can be classified as unskilled, semi-skilled and skilled each
of which form a class.
Learning Objective 1
Functions of classification
ii. Exhaustive: every unit should be allotted to one and only one class
viii. Revealing: Should bring out essential features of the collected data.
Types of classification
Methods of Classification
iii. Classification done according to more than two attributes or variables is known as
manifold classification.
iv. Examples:
1. One-way classification
No. of students who secured more than 60 % in various sections of same course
3. Manifold classification.
Classification of employees according to skill, sex and education.
Note: G: Graduate
NG: Non-Graduate
Tabulation
In spite of the fact that they are closely related, the differences are as follows.
Learning Objective 2
Parts of a Table.
ii. Title: It indicates the scope and the nature of contents in concise form.
vi. Ruling and Spacing: They separate columns and rows. However totals are
separated from main body by thick lines.
vii. Head Note: It is given below the title of the table to indicate the units of
measurement of the data and enclosed in brackets.
viii. Source Note: It indicates the source from which data is taken.
Types of Table
Tables are classified on the basis of
ii. Specific purpose table or text table or summary table deals with specific
problems. They are smaller in size and they highlight relationship between
characteristics. Example: Cost of living indices.
i. Primary Table: They contain data in the form in which it were originally collected
Ref table No.1.
ii. Derived Table: They represents figures like totals, averages, ratios etc. derived
from original data. Ref : table – 2
Table – 1
Department Age
Total
s 20 – 40 40 and Above
A A
C C
B B
under under
Graduate Post Graduate Post
Graduate Graduate
graduate graduate
Accounts 10 40 10 10 15 5 90
Finance 10 30 10 12 14 7 83
Personal 15 25 10 10 14 5 79
Production 10 30 10 8 12 6 76
Marketing 5 25 10 0 15 7 62
Total 50 150 50 40 70 30 390
Table – 2
Age
Department
s 20 –
40 & above
40
Accounts 2.564 1.282
Finance 2.564 1.795
Personal 3.846 1.282
Production 2.564 2.051
Marketing 1.282 1.795
12.920 8.205
iii. The cross – classified Table: entries are classified in both directions. Ref table
5
i. Simple Table
Table No.3
Table No.4
Distribution of Defectives according to Batch and Nature of defective
Batc
Defects
h
Major Minor
I 8 7
II 15 5
III 25 15
Total 40 27
Table No.5
Population of a city according to age, sex and education during 2003 to 2005
Learning Objective 3
a. The number of units associated with each value of the variable is called
frequency of that value. Suppose the variable takes the value 15 and the value 15
occurs 3 times then 3 is called the frequency of the value 15.
Table – 6
Table – 7
The class interval that does not include upper class limit is called Exclusive type of class
interval. The class-interval that includes the upper class limits is called Inclusive – type
of class interval.
Inclusive Type
Marks
0 – 9 15
10 –
20
19
Exclusive Type
The class 0 – 10 does not include the value 10. If the value of 10 occurs, it is included in
the class 10 – 20.
e. From a given frequency distribution we can form five derived frequency distributions.
They are i) Relative frequency distribution, ii) Percentage frequency distribution, iii)
Frequency Density, Distribution, iv) Less than cumulative frequency distribution, v)
More – than cumulative frequency distribution.
If “f” is the class frequency and “N” is the total frequency, the relative frequency
distribution is formed by calculating f/N. Total will always be one.
The percentage frequency distribution is formed by multiplying the ratio f/N by 100.
If “c” is the width of the class-interval and “f” is the frequency of the class, then
frequency density distribution is formed by calculating f/c.
The less than cumulative frequency distribution is formed with number of observations
which are less than a given value.
The more – than cumulative distribution is formed with number of observations which
are more than a given value.
Table – 8
Table 8 (a)
Table 8 (b)
Frequency distribution of more than two variables is known as multi – variate frequency
distribution. If the number of variables is only two then it is called bivariate frequency
distribution. A bivariate frequency distribution will have two Marginal Distributions and
“m+n” conditional distribution.
Table 9
Numbers in last row & column represents Marginal distribution of Age. Any row or
column number represents conditional distribution of salary.
Represents conditional distribution of Age for given salary.
ii. No. of class intervals is given by the Sturge’s Rule viz. K = 1+3.2 log N. where N is
the total number of observations.
Note: In Practice divide the range either by 2 or 5 or 10 or multiples of 10 such that the
number of class intervals will be between 7 and 15. Avoid open-end class interval. Make
sure that class-intervals do not overlap. Tally marks are used to construct frequency
Table. Tally Mark is a small vertical line drawn against a class as soon as we observe a
value belonging to the class. The fifth tally mark is crossed for easy counting purposes.
If the class interval that does not prescribe lower limit for 1st class or upper limit for last
class is known as open-end class interval.
Learning Objective 4
Table 10
They are drawn when we have two or more sets of comparable values.
Example:
Simple Bar Diagram: It is drawn when items are to be compared with respect to a single
characteristic. A rectangular bar is constructed with height proportional to the magnitude
of the items.
Example:
Represent the following data regarding the yield / acre of paddy in Karnataka over the
last five years.
Example:
Product A
a. Component Pie Diagram: It is drawn when data have magnitudes for two or more
components. Circles with area proportional to magnitudes are drawn to represent the total
magnitude. Then circles are divided sector-wise according to the magnitude of the
components.
If T is the total magnitude and R is the magnitude of a component, then the angle at the
centre is given by.
Example:
Draw pie – diagram for the following data regarding expenses of two families.
Monthly Expenses of
Items
Family A Family B
Food 2000 4000
Rent 1000 1500
Fuel 500 1000
Misc 500 1500
Total 4000 8000
We draw two circles with radii 1.3 cms and 1.8. Where 1 cm = 50 units. The angle at the
center are determined as follows.
Graphical Presentation
i. Histogram
If the class intervals have equal width then the variable is taken along X-axis and
frequency along Y-axis and a rectangle is constructed.
Example: For the following distribution of Age the histogram is drawn as follows.
We join the upper left corner of highest rectangle to the right adjacent rectangle’s left
corner and right upper corner of highest rectangle to left adjacent rectangle’s right corner.
From the intersecting point of these lines we draw a perpendicular to the X-axis. The X-
reading at that point gives the mode of the distribution.
If the widths of the rectangles are not equal then we make areas of rectangles
proportional and draw the histogram as follows
Ogives
1. Less than-ogive: Variables are taken along X-axis and less than cumulative
frequencies are taken along Y-axis. Less than cumulative frequencies are plotted
against upper limit of class interval and joined by a smooth-curve.
2. More than Ogive: More than cumulative frequencies are plotted against lower
limit of the class-interval and joined by a smooth-curve.
From the meeting point of these two ogives if we draw a perpendicular to X-axis,
the point where it meets X-axis gives Median of the distribution.
Example: Draw ogive for the following data and hence determine Median
SUMMARY
Classified data is tabulated in rows and columns for presentation, using various
types of classification. Frequency distribution is a special type of tabulation. In
more concise form it brings out the salient features of the distribution.
Data is presented in Diagram or graph form is more appealing and gives rough
idea of the situation for busy executives.
MB0024-Unit-04
Introduction
Mass data, which are collected, classified, tabulated and presented systematically, is
analyzed further to bring its size to a single representative figure.
The tendency of data to cluster around a figure is known as central tendency. Measures of
central tendency or averages of first order describes the concentration of large numbers
around a value. It is a single value which represents all units.
Learning Objective 1
Understand the concepts of Central tendency
Statistical Averages
Example: Arithmetic mean of 15, 17, 22, 21, 19, 26, 20 is given by
i. For discrete data with frequency it is given by
Example:
Example
No. of students: 50 65 80 55
Solution:
=
i. Algebraic sum of deviations of a set of values taken from their Mean is always Zero
i,e.
ii. Sum of squares of deviations of a set of values from their mean is always minimum.
iii. It is capable of further algebraic treatment. Suppose if X1, X2….. Xn are the means
of n1, n2…….nn sets of values, then their combined arithmetic mean value is given by
Example: If average height of 30 men is 158 cm and average height of another group of
40 men is 162 cm. Find average height of combined group
Given n1 = 30 X1 = 158
n2 = 40 X2 = 162
Note: In the above example given any 4 values, we can find the fifth value.
n2 = 40 = 162
= 160.28
Then,
Or,
or,
∴
= 4739.6 = 157.98
Example: In an office there are 84 employees. The distribution of their salaries are
follows
Solution
Example: The following data is related to the marks scored by students of a class in an
examination. Calculate the mean.
Percentag
e Less than Less than Less than Less than Less than Less than Less than
10 20 30 40 50 60 70
marks
No. of
4 16 20 65 85 97 100
students
Example 8: Average weight of 100 screws in box “A” is 10.4 gms. It is mixed with 150
screws of box “B”. Average weight of mixed screws is 10.9 gms. Find the average weight
of screws of box “B”
Solution:
=? n2=150
We know
Solving we get,
Example: Find the missing frequency for the following distribution given the mean value
as 129.
Class
80 – 100 100 – 120 120 – 140 140 – 160 160 – 180 Total
Interval
Frequency 8 - 26 14 10 80
Mid X f fx
90 8 720
110 f 110f
130 26 3380
150 14 2100
170 10 1700
58+f 7900+110f
Since
Missing frequency is 22
v. It is more stable.
Learning Objective 2
Median
Median of a set value is the value is the middle most value when they are arranged in the
ascending order of magnitude and is denoted by M.
In case of Discrete series without or with frequency it is given by M= value
Note: To solve problems on Median, arrange Data in ascending order or descending order
(2) Make class-interval as exclusive type.
Set values 45, 32, 31, 46, 40, 28, 27, 37, 36, 41, 47, 50
27, 28, 31, 32, 36, 37, 40, 41, 45, 46, 47, 50
f: 4, 9, 3, 5, 4, 2, 10
Solution:
n = 37
M = 15
In case of continuous series, it is given by
fc = frequency of class
Frequency: 10 15 40 27 8
∴ M = 43.125
Example: Find the missing frequency for the following data given that its median is 34
Solution: Since Median is 34, it falls in the class-interval 30 – 40. Let “f” be the missing
frequency. Therefore we have
Or, f = 35 – 16 = 19
∴ Missing frequency is 19
Merits of median
Demerits of Median
Mode:
Mode is the value which has the highest frequency and is denoted by Z.
Modal value is most useful for business people. For example shoe and Ready made
Garment manufacturers will like to know the modal size of the people to plan their
operations.
For discrete data with or without frequency it is that value corresponding to highest
frequency
Example: The following data relate to size of shoes. Find the mode.
6, 7, 6, 8, 9, 9, 9, 10, 8, 7, 7, 9, 10, 9, 9, 9, 8, 8, 11
Where
Example: Praveen apartment builders found the number of customers who wishes to
have plinth area of their apartments as follows:
Find the modal plinth area
Solution: we note that the intervals are exclusive type. Highest frequency is 25.
Therefore corresponding interval is 1200 – 1400, which is called Modal class.
Therefore,
Merits of Mode
Demerits of Mode
Geometric Mean
The geometric mean of a series of “n” positive numbers is given by
GM =
GM =
Where n = f1 + f2 + ………….. + fn
GM =
Where n = f1 + f2 + …………. + fn and x1, x2 are the mid points of class intervals.
Example: The growth in bad-debt expense for Das office supply company over the last
few years is as follows. Calculate the average percentage increase in bad-debt expense
over this time period.
Solution:
G.M =
= 1.09675
Whenever data deal with rates, ratios, growth rate, etc Geometric mean is the best
measure
Note: Geometric mean is not defined even if one of the value is zero or negative.
Harmonic Mean
If x1, x2, …………xn are “n” values for discrete series without frequency then their
Harmonic Mean
X f/λ
9.7 0.10.31
9.8 0.10.20
9.5 0.1053
9.4 0.1064
9.7 0.1031
Total 0.5199
∴ HM = 5 /(0.5199) = 9.6172
Solution:
X f f/X
121 5 0.04132
122 25 0.20492
123 36 0.29268
124 37 0.29839
125 20 0.16000
Total 123 0.99731
Example : In a locality the distribution of average speed of birds in the evening were
observed to be as follows. Find average speed of birds using harmonic mean.
Class –
80 – 82 82 – 84 84 – 86 86 – 88
Interval
Frequency 5 7 3 2
Solution:
Mid x f f/x
81 5 0.06173
83 7 0.08434
85 3 0.03520
87 2 0.02299
Total 17 0.20435
→ Median is the midvalue of series of data. It divides the distribution into two equal
portions. Similarly we can divide a given distribution into four, ten or hundred or any
other number of equal portions.
Quartiles: When distribution is divided into four equal portions, then we get first quartile
(Q1), second Quartile (Q2 = Median) and third quartile (Q3) as the positional averages.
For discrete series with or without frequency Q1 is given by (N+1 / 4)th value and Q3 is
given by
3 (N + 1)/4th value
Example : Weekly sales of a product on 8 different shops are as follows. Calculate the
Quartiles.
Sales in units: 309, 312, 305, 307, 310, 308, 308, 306, 308
Solution:
Q1 = 306.25
Weighted Averages: Suppose the values x1, x2, ……. xn are assigned the weights w1,
w2………wn then their weighted average is given by
and their weighted Geometric Mean is given by
Learning Objective 3
Dispersion:
Machine A B
Sample size 1000 1000
Average wt 80 80
Minimum wt 20 40
Maximum wt 140 100
Machine B produces products with weights much closer to the average than Machine A.
As a manufacturer or customer we would choose Machine B. In other words we choose
that machine whose spread is smaller.
The property of deviations of values from the average is called Dispersion or Variations.
The degree of variations is found by the measures of variations.
1. Range (R)
2. Quartile Deviations ( Q.D)
3. Mean Deviations (M.D)
4. Standard Deviations (S.D)
They have units of measurement attached to them. Therefore they are known as absolute
measures of variations. However we may want to compare two different distributions
whose measurements are one in terms of Kg and another in terms of cm. Then we use the
following relative measures that do not have any units attached to them. They are
1. Coefficient of Range
2. Coefficient of Quartile Deviations
3. Coefficient of Mean Deviations
4. Coefficient of Variations.
They are known as Relative measures.
Range
Range is the difference between highest and lowest value of the data.
L: Lowest value
Merits
Demerits
Use:
Solution: R = 30-24 = 6
Solution: R = 25-0 = 25
Note: If the class interval’s are open then Range is not defined.
Quartile Deviations
2. Q3-Q1 gives the middle 50% of reading. Q3 and Q1 are also known as upper and lower
limit of middle 50% of readings.
Example: Find the inter quartile Range, Q.D and coefficient of Q.D for the demand
distribution of toothpaste packs for various price categories.
Q1=9.5+{(25-15)X5}/25=10.5 Rs
Q3 = 75th value
Q3 =14.5+{(75-40)X5}/38=19.11 Rs
Q.D.=(19.11-10.5)/2=8.61/2=4.305 Rs
Coefficient of Q.D.=(19.11-10.5)/(19.11+10.5)=(8.61)/(29.61)
Merits:
Demerits:
Mean Deviation:
It is defined as the mean of absolute deviations of the values from central value.
The Mean deviation from Mean for discrete series without frequency is given by
M.D ( ) = Σ f (X – X)/Σf
In case of continuous series “X” represents Mid value of class-interval.
Similarly we can have Mean Deviation from Median or Mode. X is replaced by Median
or Mode in the above formulae.
However Mean Deviation from Median is the least. It is known as Minimal property of
Mean Deviation.
Coefficient of M.D.( )=
Coefficient of M.D.(Median)=M.D.(Median)/Median
Example: Calculate mean deviation and also coefficient of Mean deviation using i) Mean
and ii) Median. Compare the results.
Solution:
Merits
Demerits
Uses of M.D
Standard deviation
Measures of dispersion Range and Q.D are not based on all values. Mean deviation based
on all values does not take into consideration the sign. Therefore a measure that removes
both drawbacks is given by standard Deviation (S.D).
The standard deviation of a set of values is the positive square root of mean of the
squared deviations of the values from their arithmetic mean. It is denoted by σ (sigma).
σ = √Variance
σ = √Variance
For (A)
Variance =
σ = √Variance
For (B)
σ = √Variance
by σ 2
Example: Calculate the S.D for Variation in temperature observed during two months at
Bangalore.
σ
Temp0c 18 19 20 21 22 23 24 25 Total
Frequency 3 5 8 16 12 8 5 3 60
Solution:
X f d = x-21 fd fd2
1
3 -3 -9 27
8
1
5 -2 -10 20
9
2
8 -1 -8 8
0
2
16 0 0 0
1
2
12 1 12 12
2
2
8 2 16 32
3
2
5 3 15 45
4
2
3 4 12 48
5
60 28 192
Example 37: The diastolic blood pressure of men is distributed as fallows. Find the
standard deviation and variance.
Pressure(men) 78-80 80-82 82-84 84-86 86-88 88-90
No of Men 3 15 26 23 9 4
Solution:
d = x-83
Class Interval Mid value X Frequency ‘f’ fd fd2
2
78-80 79 3 -2 -6 12
80-82 81 15 -1 -15 15
82-84 83 26 0 0 0
84-86 85 23 1 23 23
86-88 87 9 2 18 36
88-90 89 4 3 12 36
80 32 122
Example: The average weight of 100 apples from area “A” is 150gms with standard
deviation of 10gms. Similarly the average weight of 200 apples from area “B” is 200gms
with standard deviation of 15gms. Find the combine standard deviation.
Solution:
= 150 = 200
σ 1 = 10 σ 2 = 15
Therefore,
Merits
1. It is rigidly defined.
2. It is based on all values.
3. It is capable of further algebraic treatment.
4. It is not very much affected by sampling fluctuations.
Demerits
1. It is difficult to understand.
2. It gives undue weight age for extreme values.
3. It cannot be calculated for classes with open end interval.
2.
Example: Find Standard deviations of the following two series and state which is
more stable.
Series A: 192, 288, 236, 229, 184, 160, 384, 291, 330, 243
Series B: 31, 48, 13, 51, 38, 43, 50, 36, 47, 82
Solution:
Series A
d = x-260 d2
X
192 -68 4624
288 28 784
236 -24 576
229 -31 961
184 -76 5776
160 0 0
384 124 15376
291 31 961
330 70 4900
243 -17 289
+37 34247
Therefore,
And
Series B
X2
X
31 961
48 2304
13 169
51 2601
38 1444
43 1849
50 2500
36 1296
47 2209
82 6724
Total 439 22057
Since c.v for series A (22.15) is less than c.v for series B (38.02), series A is more
stable.
Example: The bursting and tensile strength of a type of paper showed the
following results
SUMMARY
MB0024-Unit-05
Introduction
Every human activity has an element of uncertainty. Uncertainty affects the decision
making process. We use the word “Probably” every often, like, probably it may rain
today, probably the share price may go up in the next week. Therefore there is a need to
handle uncertainty systematically and scientifically. Probability theory helps us to make
wiser decisions.
Learning Objective 1
Definition
a. Experiment:
b. Random Experiment:
c. Sample space or total number of outcomes of an experiment is the set of all possible
outcomes of a random experiment and is denoted by S.
In tossing two coins S = {HH, HT, TH, TT}. The number of out comes is denoted by n(s)
= 4.
If the number of outcomes is finite then it is called Finite Sample Space otherwise it is
called Infinite Sample Space.
d. Event:
P (A) = ½
Two or more events are said to be equally likely if they have equal chance of occurrence.
In tossing an unbiased coin getting head and tail are equally likely.
Two or more events are said to be mutually exclusive if the occurrence of one prevents
the occurrence of other events.
In tossing a coin if head falls, it prevents the occurrence of tail and vice versa.
i. Independent Events Two events said to be independent of each other if the occurrence
of one is not affected by the occurrence of other or does not affect the occurrence of the
other.
Illustration
Event A, B, C and D are mutually exclusive and exhaustive but not equally likely.
Approaches to probability:
There are four approaches to probability They are i) Classical / Mathematical / Priori
approach ii) Statistical / Relative frequency / Empirical / posteriori approach iii)
Subjective approach and iv) Axiomatic approach.
Under this approach the probability of an event is known before conducting the
experiment
Examples:
However it is not possible to give probability to all events of our life. We cannot attach a
definite probability to the event “that it will rain today”.
In real life it is not possible to conduct experiments because of high cost or of destructive
type experiments or of vast area to be covered.
Under this approach the investigator or Researcher assigns probability to the events either
from his experience or from past records. It is more suitable when the sample size is ten
or less than ten. The investigator has full knowledge about the characteristics of each and
every individual. However there is a chance of personal bias being introduced in such
probability.
such that
a. 0 ≤ P (Ai) ≤ 1
b. ∑ P (Ai) = 1 for i = 1 to n
Rules of Probability
a. Addition Rule
i. If A and B are any two events then the probability of the occurrence of either A or B is
given by
P (A U B) = P (A) + P (B) – P (A ∩ B)
P (A U B) = P (A) + P (B)
iii. If A, B and C are any three events then the probability of occurrence of either A or B
or C is given by
iv. If A1, A2, A3………An are “n” mutually exclusive and exhaustive events then the
probability of occurrence of at least one of them is given by
Managers are very often come across with situations where they have to take decision
about implementing either course of action A or course of action B or the course of action
C. Sometimes they have to take decisions regarding the implementation of both A and B.
For example a sales manger may like to know the probability that he will exceed the
target for product A or product B. Sometimes he would like to know the probability that
sales of product A and B will exceed the target. The first type of probability is answered
by addition rule. The second type of probability is answered by multiplication rule.
b. Multiplication Rule
i. If A and B are two independent events then the probability of occurrence of A and B is
given by
Conditional Probability:
Sometimes we wish to know the probability that the price of a particular petroleum
product will rise, given that the finance minister has risen the petrol price. Such
probabilities are known as conditional probability.
Thus the conditional probability of occurrence of an event “A” given that the event “B”
has already occurred is denoted by P (A / B). Here A and B are dependent events.
Therefore we have the following rules.
If A and B are dependent events then the probability of occurrence of A and B is given by
For any Bivariate distribution there exists two marginal distributions and “m + n”
conditional distributions, where m and n are the number of classifications / characteristics
studied on two variables. Consider the following example.
A librarian analyzed the type of visitors and their choice of library section as follows:
i.
ii.
News
Magazine Novels Subjects Total
paper
220 250 200 300 970
iii.
Level of
News paper Magazine Novels Subjects Total
education
Under graduate 50 100 120 50 320
This represents the distribution of people in sections given that they are under graduate.
Therefore it is a conditional distribution. Thus for any Bivariate distributions having m
and n classifications there exits two marginal distributions and m + n conditional
distributions. In this case there are 3 + 4 = 7 conditional distributions.
iv. If the words “either, or” is used check whether the events are mutually exclusive or
not, to apply addition rule.
v. If the words “both or and” used, check whether the events are independent or
dependent, to apply proper multiplication rules.
vi. To find the total outcome of the experiment use 2n or 6n in the case of coin or die
respectively, where “n” is the number of coins or dice thrown at a time or a coin or die
thrown “n” times. In all other cases use nCr.
Example 1:
1. 0! = 1
Solution:
S = {H, T}
∴ n(S) = 2
n(A) = 1
Therefore,
Example 4: What is probability of getting two heads when 3 coins are tossed and what is
the probability of getting at least one head?
Note: n(S) = 23 = 8
Therefore,
Therefore,
Example 5: What is the probability of getting a sum “Nine” when two dice are thrown?
n(S) = 62 = 36
∴ n(A) = 4
Therefore,
Example 6: A number is selected at random from the numbers 1 to 30. What is the
probability that
n(S) = 30C1 = 30
n(A) =10
n(B) = 4
n(S) = 30 C1 = 30
Solution: Let “A” be the event of selecting 2 Scientists, 1 Engineer and 2 Accountants.
N(S)=15C5=
Therefore, P(A)=600/3003
Example 9: The odds favouring the event of a person hitting a target are 3 to 5. The odds
against the event of another person hitting the target are 3 to 2. If each of them fire once
at the target, find the probability that i) both of them hit it ii) at least one of them hit ‘d’.
Solution: i) Let “A” be event of first person hitting a target. Odds in favour means
Therefore,
Therefore
Therefore,
Given
Example 11: The probabilities that “A” and “B” will tell truth are 2/3 and 4/5
respectively. What is the probability that i) they agree with each other ii) they contradict
each other while giving a witness in the court.
i. Both will agree if they say truth or they together lie i,e A ∩ B or Ac
∩ Bc. They are mutually exclusive.
∴ P(A ∩ B) + P (Ac
∩ Bc)
ii) They will contradict if A tells truth and B tells lies or B tells truth and A tells lie
Example 12: A bag contains 5 red and 4 blue similar balls. Two balls are drawn at
random from the bag. Find the probability that both of them are red if i) the balls are
drawn together ii) the balls are drawn one after the other, with replacement iii) the balls
are drawn one after the other, without replacement.
n(S)=9C2=(9X1)/(1/2)=36
n(A)=5C2=(9X8)/(1X2)=36
Therefore, P(G)=10/36=5/18
ii) Let “A” be the event of drawing a red ball in the first draw
Let “B” be the event of drawing a red ball in the second draw
iii) Let “A” be the event of drawing red ball in the first draw
Let “B” be the event of drawing red ball in the second draw
Since the first ball is not replaced the sample space changes for second draw
Example 13: Box I contains 5 Red and 6 Blue balls. Box II contains 6 Red and 4 Blue
balls. A ball is drawn at random from box I and is transferred to box II. Now from Box II
a ball is drawn at random. What is the probability that it is red?
Solution: A ball drawn from Box I and transferred to Box II could be either red or blue.
Let “A” be the event of drawing a red ball from Box I.
Let “C” be the event of drawing red ball from Box II.
Example 14: The probabilities that component A and component B of a machine will fail
are 0.09 and 0.06 respectively. The machine will fail if any one of them fails. Find the
probability that it will fail?
P(B) = 0.06
= 0.1446
It has 52 Mondays. For one more Monday we select from the following combination of
the remaining 2 days.
Baye’s Probability
Let A1, A2, A3, A4 be mutually exclusive and exhaustive events of a random
experiment. Let “B” be a common event. In Venn Diagram it is presented as
follows.
= [From (3)]
In general Baye’s Theorem states that if A1, A2…………..An are “n” mutually exclusive
and exhaustive events and B is a common event to all theorems then probability of
occurrence of A1 given that “B” has already occurred is given by
P(A1/B)=
Whenever there are two probabilities connected with an event then we have to apply
Baye’s approach to solve it.
Example 16: The probabilities that Mr.Aravind, Mr.Anand and Mr.Akil will become
vice-president of a company are 0.40, 0.35 and 0.25 respectively. The probabilities that
they will introduce new product are 0.10, 0.15 and 0.20 respectively. What is the
probability that Mr.Anand introduced a new product by becoming vice-president?
Solution:
Even
Joint Prob
t Prior Probability Conditional prob
Posterior Probability
P(Ai) P(B/Ai)
P(Ai n B)
Ai
A1 0.4 0.10 0.0400 0.0400/0.1425=0.2807
A2 0.35 0.15 0.0525 0.0525/0.1425=0.3684
A3 0.5 0.20 0.000 0.0500/0.1425=0.3509
Total 1.00 P(B) 0.1425 1.0000
Example 17: A factory has three Machines M1, M2 and M3. They produce 4000, 10,000
and 6,000 products per day. From past records it is known that M1, M2, and M3 produce
5%, 4%, and 8% defectives. A product is selected at random from the day’s production.
What is the probability that it was not produced by Machine M3.
Solution:
Let “B” be the event that the product is defective. Then we are given
P(A1)=4000/20000=0.2
P(A2)=10000/2000=0.5
P(A3)=6000/20000=0.3
P(B/A1) = 0.05 P(B/A2) = 0.04 P(B/A3) = 0.08
Even
Joint Prob
t Prior Probability Conditional prob
Posterior Probability
P(Ai) P(B/Ai)
P(Ai n B)
Ai
A1 0.2 0.05 0.010 0.010/0.054=0.1852
A2 0.5 0.04 0.020 0.020/0.054=0.3704
A3 0.3 0.08 0.024 0.024/0.054=0.4444
1.00 P(B) 0.054 1.0000
= 1 – 0.4444 = 0.5556
Learning Objective 2
Random Variable
i. P (Xi) = P [X = Xi] V i
ii. P (Xi) ≥ 0 V i
1. For example let us consider the tossing of three coin. The resulting events are:
No. of Heads
P(Xi)
(Xi)
3 ⅛
2 ⅜
1 ⅜
0 ⅛
Total 1
For every Xi we are able to assign a P(Xi) such that ∑ P(Xi) = 1. No. of heads probability
form a probability distribution.
A systematic presentation of random variable with its value and probabilities is called a
probability distribution of that random variable. The distribution will have its mean and
standard deviation.
Examples
Example 18:
A Random variable takes the values -3, -2, 1, 0, 4, 6 with probabilities 1/12, 2/12, 3/12,
4/12, 1/12, 1/12 respectively find its mean or expected value and variance.
Solution: Given
= 6 – ¼ = 23/4
Example 19: Mr. A and B play a game. If “A” picks up an even number from 1 to 6, B
will pay him double the amount equal to picked up number. If “A” picks up an odd
number then he has to pay amount equal to double the picked up number. What is A’s
expectation?
Solution: Let Xi be the random variable and P(Xi) be its probability, then we have
Example 20: If Xi is a random variable with the following distribution find i) P(Xi) ≥ 3
ii) P(Xi = 0) iii) P(1 ≤ Xi
≤ 3) iv) P(Xi) ≥ 4
Xi -3 -2 0 1 2 3 4 5
P(Xi) K 2K 2K 3K 3K 2K K K
Solution: Since Xi is a random variable ∑ P(Xi) = 1
i) P(Xi
≥ 3) = P(Xi = 3) + P(Xi =4) + (P(Xi) = 5)
= 2K + K + K = 4K = 4/15
iii) P(1 ≤ Xi
≤ 3)
= 3K + 3K + 2K = 8K = 8/15
iv) P(Xi
≥ 4) = P(Xi = 4) + P(Xi = 5)
= K + K = 2K = 2/15
SUMMARY
Probability plays an important role in decision making process. The basic definitions and
approaches were explained with examples. The environments where to use the different
rules are also explained with examples.
MB0024-Unit-06
Introduction
Individuals and corporates generate several data that resembles certain theoretical
distributions. Since mathematically we have many derived characteristics of the
theoretical distributions, we can make use of them for a quick analysis of the observed
distributions. Examples of observed distributions are:-
i. Number of male children in a family.
Learning Objective 1
Bernoulli Distributions:
A variable which assumes values 1 and 0 with probabilities p and q=1-p, is called
Bernoulli variable. It has only one parameter p. For different values of p (0≤ p≤ 1), we
get different Bernoulli distributions.
Example: When a fair coin is tossed the outcome is either head or tail. The variable “X”
assumes 1 or 0.
An experiment which results in two mutually exclusive and exhaustive outcomes is called
a Bernoulli experiment. Let a Bernoulli experiment be repeated “n” times under identical
conditions, Let Xi, for i=1 to n, assume the values 1 or 0. Then Xi is a Bernoulli Variate
with probability p. Let X = X1 + X2 +……..+Xn denote the number of success in the “n”
repetition. Then X forms Bernoulli distribution. Its mean is p an variance is pq.
Learning Objective 2
ii. The probability of success should remain the same from experiment to
experiment.
Example 1: An unbiased coin is tossed 6 times. What is the probability that the
tosses will result in i) Exactly two head ii) At least 5 head iii) at most two heads
iv) not greater than one v) not less than five heads vi) at least one head.
. (1/2)2=
i. P(X = 0) = (0.8)5
= 10 x 0.512 x 0.04
= 0.2048
= 0.00032
Example 3: The probability that a bomb dropped on a bridge hits it is 0.5. Eight
bombs are dropped on the bridge. The bridge will be destroyed if any two bombs
fall on it. i) Find the probability that all bombs hit it ii) the bridge is destroyed.
Solution:
Let the probability that the bomb will hit the bridge be p.
Type ii)
Example 4: A random sample of 5 sachets of coconut oil were examined and two
were found to be leaking. A wholesaler receives six hundred and twenty five
packets, each containing 5 sachets. Find the expected number of packets to
contain exactly one packet leaking.
Solution:
Given n = 5
Example 5: Bring out the fallacy, if any, in the following statement on Binominal
Distribution. “The mean of a B.D is 4 and variance is 5″.
Solution:
Given np = 4 (Mean)……………..(1)
npq = 5 (Variance)………….(2)
Therefore, npq/np=5/4
Solution:
and since np = 3
n . 1/3 = 3 or n = 9
Poisson Distribution:
P(X)=e-m.mx/X!
X varies from 0 to infinity. The mean and variance of the distribution is m. Its
Assumption
Examples 7: Suppose 2 house in thousand catches fire in a year and there are
2000 houses in a village. What is the probability that
i) none ii) at least one iii) Not more than 2 houses catches fire.
Solution:
P=2/1000=0.002 n=2000
∴ m = np = 2000 x 0.002 = 4
i.
= = =0.2366
Solution:
n = 200
∴ m = np = 200 x 0.01 = 2
P(X ≥ 3) = 1 – [P(X = 0) + (P(X = 1) + P(X + 2)]= =1-e-
2
[1+2+2]
= 1 – 0.67670 = 0.3333
Solution:
Given m = 3
P(X=1)=e-3.3/1!=0.04979×3=0.14937
Example 10:
In example 9, how many pages would you expect to be free from mistakes.
Solution:
= 9.978 10 pages.
Type iii)
Example 11: X is a Poisson Variate such that P(X = 1) = P(X = 2). Find P(X = 0)
Solution:
P(X = 1) = P(X = 2)
Or,
Or, m/1=m2/2
Learning Objective 4
Normal Distribution
Examples
Examples 11:
The weight of bournvita packs packed by the filling machine follow a normal
distribution with mean weight of 500 gms and standard deviation of 10 gms. A
pack is selected at random. What is the probability that i) its weight will exceed
515 gms ii) packs weight lie within 480 to 520 gms. iii) What proportion of packs
will have less than 480 and greater than 520 gms. If 10,000 packs are supplied
how many will be rejected gms, if 480 and 520 are upper and lower limit for
acceptance.
Solution: To solve such problem will draw the normal curve and represent the
information’s given in the problem as follows
i.
P(X ≥ 515) = 0.5 – P (500 ≤ x ≤ 515)=
= 0.5 – P [0 ≤ Z ≤ 1.5]
= 0.5 – 0.4332
= 0.0668
ii)
P[480 ≤ x ≤ 520]
= P[ - 2 ≤ = ≤ 0] + P [0 ≤ = ≤ 2]
= 0.4772 + 0.4772
= 0.9544
iii)
Probability of acceptance is as found in (ii),
Type iii)
Example 12:
The sales volume of 1000 retail outlets of a soap company follow Normal
Distribution. 20 % of retail outlets sells less than 50 units per day and 15 % of
them sells 200 unit and above a) find Mean and Standard Deviation of the sales
volume b) find the expected number of retail outlets that sells units between 50
and 118 units.
Solution:
Quick analysis of observed data can be done if it is identified with the theoretical
distribution. The probabilities associated with random Variate of the distribution
help us to know the chances of occurrence of several events within specified
values. We can extend the solution to the cost aspects also.
MB0024-Unit-07
Introduction
In different fields of human activity, in doing the ordinary actions of our daily life, the
decision making process is based on the observations of few units which forms a portion
of the total population. This process of studying only a portion of the population and
making decisions involves risk, the risk of making wrong decisions. Evaluation of risk
will be discussed in Testing of hypothesis chapter. This unit deals with the various
techniques of drawing samples from the population.
Learning Objective 1
a. Universe or Population:
Statistical Survey or enquiries deal with studying various characteristics of unit belonging
to a group. The group consisting of all the units is called Universe or Population.
Example: In the statistical survey aimed at determining average per capita income of the
people in the city, all earning individuals in the city form the population.
Sampling Theory
The law lays down that a group of units chosen at random from a large group tends to
posses the characteristics of that large group.
Suppose a particular characteristic of the population has the following shape, then the
same characteristics will also follow the same shape in the Sample.
This principle states that “other things being equal, as the sample size increases, the
results tend to be more reliable and accurate”.
Suppose the population mean is 25 units. If a sample size of 50 results in average of 24.5
units, then larger sample size of 100 will result in 24.8 units. In other words larger the
sample size, the more accurate will be the result.
c. Principle of persistence of small numbers:
If some of the units in a population possess markedly distinct characteristics, then it will
be reflected in the sample values also.
For example, if there are 300 Blind persons in a population of 10,000 persons, then a
sample of hundred will have more or less same proportion of Blind persons in it.
d. Principle of Validity:
A sampling design is said to be valid if it enables us to obtain tests and estimation about
population parameters.
e. Principle of Optimization:
a. Parameter: Any statistics, like mean, median, etc calculated from population
values are known as parameters of the population and denoted by Greek letters (µ ,
σ etc).
b.
Statistics: Any statistics calculated from the sample are known as statistic and are
denoted by English letters (X, S, etc)
This gives the means of sample size 2. We form a distribution of sample means.
X f
fx fx2
Mean Frequency
1.5 1 1.5 2.25
2 1 2.0 4.00
2.5 2 5.0 12.50
3 2 6.0 18.00
3.5 2 7.0 24.5
4 1 4.0 16.0
4.5 1 4.5 20.25
N 10 30 97.50
The above table represents the sampling distributions of Means. We observe that mean of
sample means is equal to population mean.
1. d.Standard error
In other words the standard deviation of sampling distribution of any statistic is called
standard error of that statistic. Standard error helps us in
i) Testing of hypothesis.
iii) Giving reliability measure for the statistic by its reciprocal value.
Errors in Statistics
The term error denotes the difference between population value and its estimate provided
by sampling technique. Therefore the term is not referred in its ordinary sense in
statistics.
a) Sampling error
c) Biased errors
d) Unbiased errors
a.Sampling error: The sample results are bound to differ from population results, since
sample is only a small portion of the population. It is also known as inherent error and
cannot be avoided. It is not worth to eliminate them completely.
They are attributed to factors that can be controlled and eliminated by suitable actions. It
is worth to eliminate these errors. They are due to the following factors.
c.Biased errors.
It arises in both census and sampling method. They are due to personal bias of the
investigator and the instruments used for measuring. They are also due to faculty
collection of data, Respondent’s bias and bias due to non-response.
Biased errors have a tendency to grow with sample size. Therefore they are also known
as cumulative errors. The magnitude of biased errors is directly proportional to sample
size.
d.Unbiased errors.
The errors that are due to over-estimate and underestimate such that they are equal are
known as unbiased errors. They are known as compensatory errors. They do not increase
with sample size.
ii. Relative Error: is the ratio of the Absolute Error to the actual value symbolically.
It provides a degree of error for comparison purposes between different sets of data.
Learning Objective 2
i. Probability Sampling
Probability Sampling.
Under this technique sample units are drawn in such a way that each and every unit in the
population has an equal and independent chance of being included in the sample. If
sample unit is replaced before drawing next unit, then it is known as Simple Random
Sampling with replacement [SRSWR]. If the sample unit is not replaced before drawing
next unit, then it is called Simple Random Sampling without replacement [SRSWOR]. In
first case probability of drawing a unit is 1/N, where N is the population size. In the
second case probability of drawing a unit is 1/Nn.
Selection of Simple Random Sampling can be done by a) Lottery Method b) the use of
table of random numbers.
a) In lottery Method we identify each and every unit with distinct numbers by allotting an
identical card. The cards are put in a drum and thoroughly shuffled before each unit is
drawn.
b) There are several Random Numbers Tables. They are Tippet’s Random Number Table,
Fisher’s and Yate’s Tables, Kendall and Babington Smiths random tables, Rand
Corporation random number etc Specimen of Random Numbers by Tippetts is given
below.
Suppose we want to select 10 units from a population size of 100. we number the
population units from 00 to 99. Then we start taking 2 digits. Suppose we start with 41
(second row) then the other numbers selected will be 67, 95, 24, 15, 45, 13, 96, 72, 03.
We subdivide the population into several groups or strata such that i) units within
each stratum is more homogeneous ii) units between stratum are heterogeneous
and iii ) Strata do not overlap, in other words every unit of population belongs to
one and only one stratum.
The criterion used for stratification are geographical, sociological, age, sex,
income etc. The population of size N is divided into ‘K’ strata relatively
homogenous of size N1, N2………….Nk such that N1 + N2 +……… + Nk = N.
Then we draw a simple random sample from each stratum either proportional to
size of stratum OR equal units from each stratum.
Demerits
b. Appropriate sample sizes are not drawn from each of the stratum.
Example
Suppose 200, 300 and 500 items are produced by Factories located at three cities
X, Y and Z. We wish to draw a sample of 20 items under proportional stratified
sampling. We number the unit from 0 to 999. Then refer to Random Table and
select the numbers as
854, 772, 733, 741, 822, 853, 570, 802, 629, 525
Systematic Sampling
Suppose the population size is “N”. The population units are serially numbered 1
to N in some systematic order and we wish to draw a sample of “n” units, then we
divide units from 1 to N into “K” groups such that each group has n units. This
implies nK = N or K = N/n. From the first group we select a unit at random.
Suppose the unit selected is 6th unit, thereafter we select every 6 + Kth units. If K =
20, n = 5 and N = 100 then units selected are 6, 26, 46, 66, 86.
Cluster Sampling
The total population is divided into recognizable sub-divisions, known as clusters
such that within each cluster units are more heterogeneous and between clusters
they are homogenous. The units are selected from each cluster by suitable
sampling techniques.
Multi-stage Sampling
The total population is divided into several stages. The sampling process is
carried out through several stages. For example we want to select 1000 colleges
from southern states. In the first stages we may select any three state. In the
second stage we may select some districts in that state. In the 3rd stage, we may
select the colleges in each district. We may adopt any sampling technique at each
stage.
Demerits are
2.
Non-Probability Sampling
Judgment Sampling
Convenience Sampling
Quota Sampling
It is a type of judgment sampling. Under this design Quotas are set up according
to some specified characteristic such as age group, income groups etc. From each
group a specified number of units are sampled according to the Quota allotted to
the group. Within the group the selection of sample units depends on personal
judgment. It has a risk of personal prejudice and bias entering the process. This
method is often used in public opinion studies.
Learning Objective 3
Sample size depends upon the size of the population; the resources available, the
degree of accuracy desired, homogeneity of the population, nature of study,
Methods of sampling used and nature of respondents.
N is population size.
iii.
iv.
v.
If X1, X2…………Xn is a random sample of size “n” from any population, then
the sample mean (X) is normally distributed with mean µ and variance σ 2 / n
provided “n” is sufficiently large.
From the theorem we infer i) the mean of the sampling distribution of mean will
be equal to the population mean ii) the sampling distribution of the mean
approaches normal distribution as the sample size increases iii) it permits us to
use sample statistics to make inference about population parameters irrespective
of the shape of frequency distribution of the population.
SUMMARY
There are two methods of studying the characteristics of population, census and
sampling. The various advantages of sampling and the various errors that could
prop up in using these methods were explained. Mainly there are two methods of
sampling namely i) Probability Sampling ii) non-probability sampling. The merits
and demerits of each sampling method were explained. We discussed the
procedure for determining sample size. We concluded the chapter with the
importance of central limit theorem.
MB0024-Unit-08
Introduction
Everyone makes estimates. When you are ready to cross a street, you estimate the speed
of any car that is approaching, the distance between you and that car, and your own
speed. Having made these quick estimates, you decide whether to wait, walk, or run.
Learning Objective 1
All mangers must make quick estimates too. The outcome of these estimates can affect
their organizations as seriously as the outcome of your decision as to whether to cross the
street. Credit managers estimate whether a purchaser will eventually pay his bills.
Prospective home buyers make estimates concerning the behaviour of interest rates in the
mortgage market. All these people make estimates without worry about whether they are
scientific but with the hope that the estimates bear a reasonable resemblance to the
outcome.
Managers use estimates because in all but the most trivial decisions, they must make
rational decisions without complete information and with a great deal of uncertainty
about what the future will bring. As educated citizens and professionals, you will be able
to make more useful estimates by applying the techniques described in this and
subsequent chapters.
Statistical inference is based on estimation, and hypothesis testing. In both estimation and
hypothesis testing, we shall be making inferences about characteristics of populations
from information contained in samples. Here we infer something about a population from
information taken from a sample.
Here we try to estimate with reasonable accuracy the population proportion (the
proportion of the population that possesses a given characteristic) and the population
mean. To calculate the exact proportion or the exact mean would be an impossible goal.
Even so, we will be able to make an estimate, and implement some controls to avoid as
much of the error as possible.
Types of estimates
Learning Objective 2
Learning Objective 3
Point estimates:
Results of a samples of 35 Box of bolts (bolts per box)
101 103 112 102 98 97 93
105 100 97 107 93 94 97
97 100 110 106 110 103 99
93 98 106 100 112 105 100
114 97 110 102 98 112 99
Consider the table above, we have taken a sample of 35 boxes of bolts from a
manufacturing line and have counted the bolts per box. We can arrive at the
population mean i.e. mean number of bolts by taking the mean for the 35 boxes
we have sampled. i.e. adding all the bolts and dividing by the number of boxes.
Thus using the sample mean x as the estimator we have a point estimate of the
population mean µ.
Similarly we can use the sample variance s2 and estimate the population variance,
where the sample variance s2 is given by the formula.
Interval Estimates
The marketing research director needs an estimate of the average life in months of
car batteries his company manufactures. We select a random sample of 200
batteries with a mean life of 36 months. If we use the point estimate of the sample
mean x as the best estimator of the population mean µ, we would report that the
mean life of the company’s batteries is 36 months.
The director also asks for a statement about the uncertainty that will be likely to
accompany this estimate, that is, a statement about the range within which the
unknown population mean is likely to lie. To provide such a statement, we need to
find the standard error of the mean.
If we select and plot a large number of sample means from a population, the
distribution of these means will approximate to normal curve. Furthermore, the
mean of the sample means will be the same as the population mean. Our sample
size of 200 is large enough that we can apply the central limit theorem. Suppose
we have already estimated the standard deviation of the population of the batteries
and reported that it is 10 months. Using this standard deviation we can calculate
the standard error of the mean: so using the formula
We can tell to the director that our estimate of the life of the company’s batteries
is 36 months, and the standard error that accompanies this estimate is 0.707. In
other words, the actual mean life for all the batteries may lie somewhere in the
interval estimate of 35.293 to 36.707 months. This is helpful but insufficient
information for the director. Next, we need to calculate the chance that the actual
life will lie in this interval or in other intervals of different widths that we might
choose, ± 2σ (2 x 0.707), ± 3σ (3 x 0.707), and so on.
The probability is 0.955 that the mean of a sample size of 200 will be within ±2
standard errors of the population mean. Stated differently, 95.5 percent of all the
sample means are within ±2 standard errors from µ . “The population mean µ will
be located within ±2 standard errors from the sample mean 95.5 percent of the
time.”
Hence from the above example we can now report to the director, that the best
estimate of the life of the company’s batteries is 36 months, and we are 68.3
percent confident that the life lies in the interval from 35.293 to 36.707 months
(36 ± 1 σ x ). Similarly, we are 95.5 percent confident that the life falls within the
interval of 34.586 to 37.414 months (36 ± 2 σ x), and we are 99.7 percent
confident that battery life falls within the interval of 33.879 to 38.121 months (36
± 3 σ x).
In using interval estimates, we are not confined to ±1,2 and 3 standard errors; for
example, ± 1.64 standard errors includes about 90 percent of the area under the
curve; it includes 0.4495 of the area on either side of the mean in a normal
distribution. Similarly, ±2.58 standard error includes about 99 percent of the area,
or 49.51 percent on each side of the mean.
The confidence interval is the range of the estimate we are making. If we report
that we are 90 percent confident that the mean of the population of incomes of
people in a certain community will lie between Rs. 8,000 and Rs. 24,000, then the
range Rs. 8,000-Rs. 24,000 is our confidence interval. Often, however, we will
express the confidence interval in standard errors rather than in numerical values.
Thus, confidence limits are the upper and lower limits of the confidence interval.
In this case, X + 1.64 is called the upper confidence limit (UCL) and X – 1.64
If the samples are large then we use the finite population multiplier to calculate
the standard error. This is given from the previous unit as
Mean µ = np
q = probability of failure = 1- p
Since we are taking the mean of the sample to be the mean of the population we
actually mean that µ -p = p
Similarly, we can modify the formula for the standard deviation of the binomial
distribution, √(npq), which measures the standard deviation in the number of
successes. To change the number of successes to the proportion of successes, we
divide √npq by n and get √(pq )/ √(n)
Example: In a very large organization the director wanted to find out what
proportions of the employees prefer to provide their own retirement benefits in
lieu of a company – sponsored plan. A simple random sample of 75 employees
was taken and found that 40%, i.e. 0.4 of them are interested in providing their
own retirement plans. The management requests that we use this sample to find an
interval about which they can be 99 percent confident that it contains the true
population proportion.
There the interval estimate for 99% level of confidence is 0.4 ± 2.58 (0.057) =
0.253 and 0.547.
So far, the sample sizes we were examining were all larger than 30. This is not
always the case. Questions like how can we handle estimates where the normal
distribution is not the appropriate sampling distribution, that is, when we are
estimating the population standard deviation and the sample size is 30 or less?
Suppose we have data only form let us say 10 weeks or sample sizes less than 30,
then fortunately, another distribution exists that is appropriate in these cases. It is
called the t distribution.
Because it is used when the sample size is 30 or less, statisticians often associate
the t distribution with small sample statistics. This is misleading because the size
of the sample is only one of the conditions that lead us to use the t distribution.
The second condition is that the population standard deviation must be unknown.
Use of the t distributions for estimating is required whenever the sample size is 30
or less and the population standard deviation is not known. Furthermore, in using
the t distribution, we assume that the population is normal or approximately
normal.
Degrees of freedom
With two sample values, we have one degree of freedom (2-1 = 1), and with
seven sample values, we have six degrees of freedom (7-1 = 6). In each of these
two examples, then, we had n-1 degrees of freedom, assuming n is the sample
size. Similarly, a sample of 23 would give us 22 degrees of freedom.
The table of t distribution values differs in construction from the z table or normal
distribution table used previously. The t table is more compact and shows areas
and t values for only a few percentages (10, 5, 2, and 1 Percent). Because there is
a different t distribution for each number of degrees of freedom, a more complete
table would be quite lengthy. Although we can conceive of the need for a more
complete table
A second difference in the t table is that it does not focus on the chance that the
population parameter being estimated will fall with our confidence interval.
Instead, it measures the chance that the population parameter we are estimating
will not be within our confidence interval (that is, that it will lie outside it). If we
are making an estimate at the 90 percent confidence level, we would look in the t
table under the 0.10 column (100 percent – 90 percent = 10 percent). This is 0.10
chance of error is symbolized by the Greek letter alpha α. We would find the
appropriate t values for confidence intervals of 95 percent, 98 percent, and 99
percent under the columns headed 0.05, 0.02, and 0.01, respectively.
A third difference in using the t table is that we must specify the degrees of
freedom with which we are dealing. Suppose we make an estimate at the 90
percent confidence level with a sample size of 14, which is 13 degrees of
freedom. Look under the 0.10 column until you encounter the row labelled 13.
Like a z value the t value there of 1.771 shows that if we mark off plus and minus
1.7716 (estimated standard errors of ) on either side of the mean, the area
under the curve between these two limits will be 90 percent, and the area outside
these limits(the chance of error) will be 10 percent.
Remember that in any estimation problem in which the sample size is 30 or less
and the standard deviation of the population is unknown and the underlying
population can be assumed to be normal or approximately normal, we use the t
distribution.
In all the examples above we have used, the sample size was known. Now we are
trying to estimate the sample size n. if it is too small we may fail to achieve the
objective, if it is too large we will be wasting resources. However, let’s try to
examine some of the methods that are useful in determining what sample is
necessary for any specified level of precision.
If you look at the problem above: it is stated that variation of $ 500 on either side
of the populations mean.
Meaning n should be greater than 34.6 or 35 if the university want to estimate the
precision with which it wants to conduct the survey.
SUMMARY
In this chapter we have seen point estimates and interval estimates. These are the
foundation for inferential statistics in estimation and hypothesis testing which we
will be discussing in the next unit. Also we have seen the concept of confidence
levels and make estimations when the sample sizes are small and large. Also we
have gone in reverse to estimate a sample size provided we know the level of
accuracy we want to construct the estimate. Also we have seen that if the sample
size is less than 30 and the populations standard deviation is not known, we use
the student’s t distribution for estimations.
MB0024-Unit-09
Introduction
Hypothesis testing begins with an assumption, called a hypothesis, that we make about a
population parameter. We assume a certain value for a population mean. To test the
validity of our assumption, we gather sample data and determine the difference between
the hypothesized value and the actual value of the sample mean. Then we judge whether
the difference is significant. The smaller the difference, the greater the likelihood that our
hypothesized value for the mean is correct. The larger the difference, the smaller the
likelihood.
Unfortunately, the difference between the hypothesized population parameter and the
actual statistic is more often neither so large that we automatically reject our hypothesis
nor so small that we just as quickly accept it. So in hypothesis testing, as in most
significant real-life decisions, clear-cut solutions are the exception, not the rule.
Learning Objective 1
Assumptions: Although hypothesis testing sounds like some formal statistical term
completely unrelated to business decision making, in fact managers propose and test
hypothesis all the time. “if we drop the price of this car model by Rs.1,500, we’ll sell
50,000 cars this year” is a hypothesis. To test this hypothesis, we have to wait until the
end of the year and count sales. Managerial hypothesis are based on intuition; the
marketplace decides whether the manager’s intuitions were correct. Hint: Hypothesis
testing is about making inferences about a population from only a small sample. The
bottom line in hypothesis testing is when we ask ourselves (and then decide) whether a
population like we think this one is would be likely to produce a sample like the one we
are looking at.
1. Testing Hypothesis
2. Null and Alternate hypothesis
In hypothesis testing, we must state the assumed or hypothesized value of the
population parameter before we begin sampling. The assumption we wish to test
is called the null hypothesis and is symbolized Ho.
Suppose we want to test the hypothesis that the population mean is equal to 500.
We would symbolize it as follows and read it, “The null hypothesis is that the
population mean = 500 written as Ho: µ = 500. The term null hypothesis arises
from earlier agricultural and medical applications of statistics. In order to test the
effectiveness of a new fertilizer or drug, the tested hypothesis (the null
hypothesis) was that it had no effect, that is, there was no difference between
treated and untreated samples.
If our sample results fail to support the null hypothesis, we must conclude that
something else is true. Whenever we reject the hypothesis, the conclusion we do
accept is called the alternative hypothesis and is symbolized H1 (”H sub-one”).
For the null hypothesis H0: µ = 200
H1: µ
≠ 200 (population mean is not equal to 200)
The purpose of hypothesis testing is not to question the computed value of the
sample statistic but to make a judgment about the difference between that sample
statistic and a hypothesized population parameter. The next step after stating the
null and alternative hypotheses, then, is to decide what criterion to use for
deciding whether to accept or reject the null hypothesis. If we assume the
hypothesis is correct, then the significance level will indicate the percentage of
sample means that is outside certain limits. (In estimation, please remember, the
confidence level indicated the percentage of sample means that fell within the
defined confidence limits.
Hypothesis are accepted and not proved.
Even if our sample statistic does fall in the non-shaded region (the region that
makes up 95 percent of the area under the curve), this does not prove that our null
hypothesis (H0) is true; it simply does not provide statistical evidence to reject it.
Why? Because the only way in which the hypothesis can be accepted with
certainty is for us to know the population parameter; unfortunately, this is not
possible. Therefore, whenever we say that we accept the null hypothesis, we
actually mean that there is not sufficient statistical evidence to reject it. Use of the
term accept, instead of do not reject, has become standard. It means simply that
when sample data do not cause us to reject a null hypothesis, we behave as if that
hypothesis is true.
If the significance level is high then we would rarely accept the null hypothesis
when it is not true but, at the same time, often reject it when it is true.
When testing a hypothesis we come across with four possible situations depicted
as follows.
Suppose that making a Type I error (rejecting a null hypothesis when it is true)
involves the time and trouble of reworking a batch of chemicals that should have
been accepted. At the same time, making a Type II error (accepting a null
hypothesis when it is false) means taking a chance that an entire group of users of
this chemical compound will be poisoned. Obviously, the management of this
company will prefer a Type I error to a Type II error and, as a result, will set very
high levels of significance in its testing to get low β s.
Suppose, on the other hand, that making a Type I error involves disassembling an
entire engine at the factory, but making a Type II error involves relatively
inexpensive warranty repairs by the dealers. Then the manufacturer is more likely
to prefer a Type II error and will set lower significance levels in its testing.
After deciding what level of significance to use, our next task in hypothesis
testing is to determine the appropriate probability distribution. We have a choice
between the normal distribution, and the t distribution. The rules for choosing the
appropriate distribution are similar to those we encountered in the unit on
estimation. The Table below summarizes when to use the normal and t
distributions in making tests of means. Later in this unit, we shall examine the
distributions appropriate for testing hypotheses about proportions.
Remember one more rule when testing the hypothesized values of a mean. As in
estimation, use the finite population multiplier whenever the population is finite in
size, sampling is done without replacement, and the sample is more than 5 percent
of the population.
Learning Objective 2
A two-tailed test of a hypothesis will reject the null hypothesis if the sample mean
is significantly higher than or lower than the hypothesized population mean. Thus,
in a two-tailed test, there are two rejection regions. This is shown in figure 1 of
9.12.
However, there are situations in which a two-tailed test is not appropriate, and we
must use a one-tailed test.
Example 2: Consider the case of a wholesaler that buys light bulbs from the
manufacturer discussed earlier. The wholesaler buys bulbs in large lots and does
not want to accept a lot of bulbs unless their mean life is at least 1,000 hours or a
minimum of 1,000 hours. As each shipment arrives, the wholesaler tests a sample
to decide whether it should accept the shipment. The company will reject the
shipment only if it feels that the mean life is below 1,000 hours. If it fells that the
bulbs are better than expected (with a mean life above, 1,000 hours), it certainly
will not reject the shipment because the longer life comes at no extra cost. So the
wholesaler’s hypotheses are Ho: µ = 1,000 and H1: µ < 1,000 hours. It rejects
Ho only if the mean life of the sampled bulbs is significantly below 1,000 hours.
This situation is illustrated in the figure below. From this figure, we can see why
this test is called a left-tailed test (or a lower-tailed test).
In general, a left tailed (lower-tailed) test is used if the hypotheses are Ho: µ =
µ Ho. In such a situation, it is sample evidence with the sample mean significantly
below the hypothesized population mean that leads us to reject the null hypothesis
in favor of the alternative hypothesis. Stated differently, the rejection region is in
the lower tail (left tail) of the distribution of the sample mean, and that is why we
call this a lower-tailed test.
A left-tailed test is one of two kinds of one-tailed tests. As you have probably
guessed by now, the other kind of one-tailed test is a right-tailed test (or an upper-
tailed test). An upper-tailed test is used when the hypotheses are Ho: µ > µ Ho.
Only values of the sample mean that are significantly above the hypothesized
population mean will cause us to reject the null hypothesis in favor of the
alternative hypothesis. This is called an upper-tailed test because the rejection
region is in the upper tail of the distribution of the sample mean.
This is to remind you again that, in each example of hypothesis testing, when we
accept a null hypothesis on the basis of sample information, we are really saying
that there is no statistical evidence to reject it. We are not saying that the null
hypothesis is true. The only way to prove a null hypothesis is to know the
population parameter, and that is not possible with sampling. Thus, we accept the
null hypothesis and behave as if it is true simply because we can find no evidence
to reject it.
Test
Description of Test Test Statistics Notes
No.
P – Population
proportion
Q = 1 – P, n sample
size
2 Test for specified proportion – P = Population
Finite Population proportion
Ps = Sample
Q = 1 –P, n – Sample
size
N – Population size
P1 -first sample
proportion
P2 -second sample
proportion
Test between proportions –
3
different Population Q1 = 1 – P, Q2 = 1-P2
n2 – second sample
size
P1 -first sample
proportion
P2 -second sample
proportion
Test between proportion –
4
same population Q1 = 1 – P, Q2 = 1-P2
n2 – second sample
size
σ = Population S.D
µ s = Sample mean
P2 -second sample
proportion
Test between means –
7
different Population
Q1 = 1 – P, Q2 = 1-P2
Test Procedure
Step 2: State the level of significance. This gives you the tabulated normal / t –
value
Step 3: Select the appropriate test from the list given in 9.2 and next chapter 10
Step 4: Calculate the required values for the test
Learning Objective 3
Step 2: Check whether the data is attribute or variable. If the words mean and S.D
are used, then it is test for variable, other wise it is test for attribute.
Step 3: Check whether it is a test for specified value or between values. If two
sample sizes are given, then it is between values, otherwise it is for specified
value.
Step 6: If the words improved, more, higher, less, lower, effective, efficient,
superior, inferior etc used then it is one-tailed test, otherwise it is two tailed test.
Examples:
Example 1: Thompson press hypotheses that the average life of its latest web-
offset press is 14,500 hours. They know the SD of the press life is 2,100 hours.
From a sample of 25 presses, the company finds a sample mean of 13,000 hours.
At 0.01 significance level, should the company conclude that the average life of
the presses is less than the hypothesized 14,500 hours?
Therefore
9. Test
10. Conclusion
Example 2: Theater owners in India know that a hit movie ran for an average of
84 days with a standard deviation of 10 days in each city the movie was screened.
A particular movie distributor was interested in comparing the popularity of
movie in his region with that of the population. He randomly chose 75 theatres at
random in the region and found a popular movie ran for 81.5 days.
11. State appropriate hypotheses for testing whether there was significant
difference between theatres in the distributor’s region and the population.
12. At a 1% significance level, test these hypotheses.
Therefore,
17. Test
18. Conclusion
23. Test
24. Conclusion
Therefore,
29. Test
30. Conclusion
Therefore,
35. Test
36. Conclusion
Since Zcal (2.60) > Ztab (1.96) Ho is rejected ⇒ The machine is not
functioning properly.
SUMMARY
In this unit we discussed the four tests available for small samples. These tests can
be used for sample size (n ≤ 30) and samples whose population S.D are not
known. The different tests are illustrated with examples.
MB0024-Unit-10
Introduction
In the previous units we learned how to test hypotheses using data from either one or two
samples. We used one-sample tests to determine whether a mean or a proportion was
significantly different from a hypothesized value. In the two-sample tests, we examined
the difference between either two means or two proportions, and we tried to learn
whether this difference was significant.
Suppose we have proportions from five populations instead of only two. In this case, the
methods for comparing proportions described in for testing hypothesis for two-samples
do not apply; we must use the chi-square X2 test. chi-square X2 tests enable us to test
whether more than two population proportions can be considered equal.
Actually, chi-square X2 tests allow us to do a lot more than just test for the quality of
several proportions. If we classify a population into several categories with respect to two
attributes (such as age and job performance), we can then use a chi-square X2 test to
determine whether the two attributes are independent of each other.
Learning Objective 1
Characteristics of X2 test
The sample observations should be independently and normally distributed. For this
either the parent population should be infinitely large (say, greater than 50) or
sampling should be done with replacement.
Application of X2 test
X2 is used in testing: (i) the significance of sample variances, (ii) the goodness of fit
of a theoretical distribution, (iii) the independence in a contingency table and (iv)
whether the observed results are consistent with the expected segregations in breeding
experiments of Genetics.
Levels of significance
Tables have been prepared for the values of P, the probability of getting a value of X 2
≥ X02 where X2 is an observed value. From these tables, we can find the value of P
corresponding to an observed value if X2 and then proceed to test whether the
difference between observed and theoretical frequencies is significant or not. Smaller
the values of P, greater the divergence between fact and theory so that small values
lead us to suspect the hypothesis. Not only small values of P lead us to suspect the
hypothesis but a value of P very near to unity may also lead to a similar result. Thus if
P = 1, X2 = 0, showing that there is perfect agreement between fact and theory which
is a very improbable event. There are two conventional levels of significance.
fe is expected frequency
1. Calculate the expected frequencies. In general the expected frequency for any cell
can be calculated from the following expression:
2. Take the difference between observed and expected frequencies and obtain the
squares of these differences (O – E)2
3. Divide the values obtained in step 2 by the respective expected frequency and add
all the values to get the value according to the formula Σ (f0 – fe)2 / fe.
Interpretation
After ascertaining the X2 value, the X2 table comprises of columns headed with
symbols ψ 0.05 for 5% level of significance, X20.01
for 1% level of significance and so on. The left hand side indicates the degrees of
freedom. If the calculated value of X2 falls in the acceptance region, the null
hypothesis HO is accepted and vice-versa.
Learning Objective 2
Which is the sum of the squares of n independent standard normal variates, following the
X2 distribution with n degrees of freedom.
Properties of X2 distribution
Conditions
1. The frequencies used in chi-square test must be absolute and not in relative terms.
2. The total no. of observations collected for this test must be large.
3. Each of the observations which make up the sample of this test must be
independent of each other.
4. As X2 test is based wholly on sample data, no assumption is made concerning the
population distribution. In other words it is a non parametric-test.
5. X2 test is wholly dependent on degrees of freedom.
6. The expected frequency of any item or cell must not be less than 5, the
frequencies of adjacent items or cells should be polled together in order to make it
more than 5.
7. The data should be expressed in original units for convenience of comparison and
the given distribution should not be replaced by relative frequencies or
proportions.
8. This test is used only for drawing inferences through test of the hypothesis, so it
cannot be used for estimation of parameter value.
Uses of X2 test
• Test goodness of fit for one way classification or for one variable only.
• Test of independence or interaction for more than one row or column in the form
of a contingency table concerning several attributes
• Test of population Variance σ 2 through confidence intervals suggested by X2
test.
Application of X2 – test
Example 1: The following table gives the production in three shifts and the number of
defective goods that turned out in three weeks. Test at 5% level of significance whether
weeks and shifts are independent.
Solution:
Example 2: Out of 1000 people surveyed 600 belonged to urban area and rest to rural
area. Among 500 who visited other states 400 belonged to urban area. Test at 5% level of
significance whether area and visiting other states are dependent.
(O – E)2
2
Observed Value (O) Expected Value (E) (O – E)
E
400 300 10000 33.33
200 300 10000 33.33
100 200 10000 50.00
300 200 10000 50.00
X2cal 1.66.66
Solution: If the absenteeism is uniformly distributed over the week, then expected No. of
absenteeism per day should be
E = 66 + 57 + 54 + 48 + 75 /5 = 60
2. Level
3. Test Statistics
Solution:
Suppose we want to test whether the population has a given variance σ 02, then
and
If the calculated value lie between K1 and K2 then H0 is accepted K1 and K2 values are
read from the table.
Solution:
X d = X – 160 d2
17
12 144
2
15
-4 16
6
15
-6 36
4
16
3 9
3
17
10 100
0
16
9 81
9
17
10 100
0
16
4 16
4
38 502
S2=40.1875
∴ nS2 = 321.5
SUMMARY
MB0024-Unit-11
Introduction
and
are given by these formulae
It is also known as variance Ratio test. It has two degrees of freedom, one for numerator
and another for denominator of the ratio. They are represented by
ν 1 = n1 – 1 and ν 2 = n2 – 1.
Learning Objective 1
Examples
Can we conclude that variance of time distribution for method I and method II are
same?.
Solution:
Method I Method II
X d = X – 22 d2
27 5 25
23 1 1
16 -6 36
20 -2 4
26 4 11
22 0 0
Total 2 82
X d = X – 35 d2
33 -2 4
35 0 0
34 -1 1
27 -8 64
42 7 49
32 -3 1
38 3 9
Total -4 136
= 22.286
3. Test
FCal=22.286/16.266=1.37
1. Conclusion
Learning Objective 2
ANOVA: ANOVA will enable us to test for the significance of the differences of
variances among more than two sample means. Using analysis of variance we will be
able to make inferences about whether our samples are drawn from populations having
the same mean.
In statistical terms the difference between two statistical data is known as variance. When
two data are compared for any practical purpose, their difference is studied through the
techniques of Analysis of Variance. Initially the technique was applied in the field of
Zoology and Agriculture but in a later stage it was applied in other fields also. In analysis
of variance the degree of variance between two or more data as well as the factor
contributing towards the variance is studied.
The analysis of variance is a method of splitting the total variation of data into constituent
parts which measure different sources of variations. The total variation is split up into the
following two-components.
Total variance = Variance between the samples + Variance within the samples.
After obtaining the above two variations, these two variations are tested for their
significance by F-test which is also known as Variance Ration Test.
ii. To find a measure of variation between or among the components. Then the
significance of difference between the variations in two series or more may be
measured.
In other words, with the help of the technique of analysis of variance we can test the
hypothesis that the means of all the components constituting a population are equal to the
mean of the population or that the samples have come from the same population.
ANOVA Table
Assumptions
ii. Population from which the samples are selected are normally distributed.
iv. Each of the population has the same variations and identical means.
The analysis of variance is mainly carried on under the following two classifications:
ii. Two way analysis of variance or two way classified data or manifold classification.
‘ANOVA’ table presents the various results obtained while carrying out the analysis of
variance. A specimen of ANOVA table is given below.
Example 1: Below are given the yield (in Kg) per acre for 5 trial plots of varieties of
treatment.
2. 50 66 52 94
3. 62 68 76 78
4. 34 78 64 82
5. 52 70 70 66
Treatment
Plot No. (X1) (X2) (X3) (X4)
1 2 3 4
1. 42 48 68 80
2. 50 66 52 94
3. 62 68 76 78
4. 34 78 64 82
5. 52 70 70 66
Total 240 330 330 400
T2 / N = 13002 / 20 = 84500
MSC=SSC/(K-1)=2580/(4-1)=2580/3=860
MSE=SSE/(N-K)=1656/(20-4)=103.5
F=MSC/MSE=860/103.5=8.3
The table value of F at 5% level of significance for (3, 16) df is 3.24 is less than the
calculated value of F. Therefore the null hypothesis is rejected. Therefore the treatments
do not have the same effect.
In the two way classification, observations are classified into groups on the basis of two
criteria.
Steps:
1. (a) Assume the means of all columns are equal. That is the effects of all factors in
one kind of treatment are equal.
i,e., α 1 = α 2 = α 3 =………α c
(b) Assume the means of all rows are equal. That is, the effects of all factors in
the second kind of treatment are equal.
i,e., β 1 = β 2 = β 3 =………β r
4. Find
2.
Example
2:
Three varieties of crops A, B, C are tested in a randomized block design with four
replications – The yields are given below:
Variet Replications
y 1 2 3 4
A 6 4 8 6
B 7 6 6 9
C 8 5 10 9
Test whether there is difference between replications. Test also whether varieties differ
significantly
= 62+72+82+42+62+52+82+62+102+62+92+92 – 588 = 36
SST = 36
=18
MSC=SSC/(c-1)=18/3=6
SSR = 8
MSR=SSR/(r-1)=8/2=4
MSE=SSE/{(r-1)(c-1)}=10/6=1.667
Source of
Sum of Squares d.f. Mean Square F.Ratio
Variation
Between Columns SSC = 18 c–1=3 MSC = 6
Fc = 6/1.667 = 3.6
Within Rows SSR = 18 r–1=2 MSR = 4
Fr = 4/1.667 = 2.4
Residual SSE = 10 (c-1) x (r -1) = 6 MSE = 1.667
Total SST = 36 N – 1 = 11
Between columns:
Between rows:
Therefore we accept the hypothesis that there is no significant difference between the
varieties.
SUMMARY
F-test is used to test the equality of two variance. ANOVA is used to test the
equality of several means using the relation σ x = σ / √n.
MB0024-Unit-12
Introduction
Both correlation and regression are used to measure the strength of relationships between
variables.
The following statistical tools measure the relationship between the variable analyzed in
social science research.
1. Correlation
a. Simple correlation – Here the relationship between two variables are studied.
b. Partial correlation – Here the relationship of any two variables are studied,
keeping all others constant.
c. Multiple correlation – Here the relationship between variables are studied
simultaneously.
2. Regression
a. Simple regression
b. Multiple regression
3. Association of Attributes
Correlation measures the relationship (positive or negative, perfect) between the two
variables. Regression analysis considers relationship between variables and estimates the
value of another variable, having the value of one variable. Association of Attributes
attempts to ascertain the extent of association between two variables.
Learning Objective 1
Correlation
When two or more variables move in sympathy with other, then they are said to be
correlated. If both variables move in the same direction then they are said to be positively
correlated. If the variables move in opposite direction then they are said to be negatively
correlated. If they move haphazardly then there is no correlation between them.
The correlation between two variables may be due to the following causes,
i) Due to small sample sizes. Correlation may be present in sample and not in
population.
ii) Due to a third factor. Correlation between yield of rice and tea may be due to a
third factor “rain”
Types of Correlation
Types of correlation are given below
a. Positive or Negative
Positive correlation: Both the variables (X and Y) will vary in the same direction. If
variable X increases, variable Y also will increase; if variable X decreases, variable Y
also will decrease. Negative Correlation: The given variables will vary in opposite
direction. If one variable increases, other variable will decrease.
Simple, Partial and Multiple correlations: In simple correlation, relationship between two
variables are studied. In partial and multiple correlations three or more variables are
studied. Three or more variables are simultaneously studied in multiple correlations. In
partial correlation more than two variables are studied, but the effect on one variable is
kept constant and relationship between other two variables is studied.
Linear and Non-Linear correlation: It depends upon the constancy of the ratio of change
between the variables. In linear correlation the percentage change in one variable will be
equal to the percentage change in another variable. It is not so in non linear correlation.
Measures of correlation
i) Scatter Diagram.
Scatter Diagram
The ordered pair of observed values are plotted on x y plane as dots. Therefore it is also
known as Dot Diagram. It is diagrammatic representation of relationship.
If the dots lie exactly on a straight line that runs form left bottom to right top, then the
variables are said to be perfectly positively correlated (fig.i).
If the dots lie close to a straight line that runs from left bottom to right top, then the
variables are said to be positively correlated (fig.ii).
If the dots lie exactly on a straight line that runs from left top to right bottom then the
variables are said to be perfectly negatively correlated (fig iii).
If the dots lie very close to a straight line that runs from left top to right bottom then the
variables are said to be negatively correlated (fig iv).
If the dots lie all over the graph paper then the variables have zero correlation (fig v).
Scatter diagram tells us the direction in which they are related and does not give any
quantitative measures for comparison between sets of data.
It is defined as
1. i. …………………………….(A)
Where
n – number of paired observations
∑xy / N is called covariance of x and y. The other forms of this formula are
ii. ii.
For all practical purpose we can conveniently use form D. Whenever summary
information is given choose proper form from A to C.
The size of r is very much dependent upon the variability of measured values in the
correlation sample. The greater the variability, the higher will be the correlation,
everything else being equal.
The size of r is altered when researchers select extreme groups of subjects in order to
compare these groups with respect to certain behaviors. Selecting extreme groups on one
variable increases the size of r over what would be obtained with more random sampling.
Combining two groups which differ in their mean values on one of the variables is not
likely to faithfully represent the true situation as far as the correlation is concerned.
Addition of an extreme case (and conversely dropping of an extreme case) can lead to
changes in the amount of correlation. Dropping of such a case leads to reduction in the
correlation while the converse is also true. (Source: Aggarwal.Y.P, Statistical Methods,
Sterling Publishers Pvt Ltd., New Delhi, 1998, p.131).
Problems
X 20 16 12 8 4
Y 22 14 4 12 8
X Y X2 Y2 XY
20 22 400 484 440
16 14 256 196 224
12 4 144 16 48
8 12 64 144 96
4 8 16 64 32
∑X = 60 ∑Y = 60 ∑X = 880 ∑Y = 904 ∑XY = 840
2 2
Applying the formula for r and substituting the respective values from the above
table we get r as:
Example 2: Calculate Karl Pearson Coefficient of Correlation from the following data:
Solution:
= 104 = 15
X 50 60 58 47 49 33 65 43 46 68
Y 48 65 50 48 55 58 63 48 50 70
Solution:
Using the formula for calculating r as
R=-17.5/(7X3)=-0.833
There is a high negative correlation.
Example 5: Ten observation in Weight (x) and Height (y) of a particular age group gave
the following data.
Find “r”
Solution: we know
Probable Error
It is used to
c) If P.E < r < 6 P.E, we can not say anything about the significance of “r”
ii) Construct confidence limits within which population “P” is expected to lie.
Conditions under which P.E can be used.
Example 6
If r = 0.6 and N = 64, a) Interpret ‘r’ b) find the limits within which ‘ρ ‘ is suppose to lie.
Solution:
It is highly significant
= 0.6 ± 0.054
= 0.546 – 0.654
ii) The variables under study are affected by a large number of independent
causes so as to form a normal distribution. When we do not know the shape of
population distribution and when the data is qualitative type Spearman’s Ranks
correlation coefficient is used to measure relationship.
It is defined as
Example 7: In a singing competition, two judges assigned the following ranks for 7
candidates. Find Spearman’s rank correlation coefficient.
Competitor 1 2 3 4 5 6 7
Judge I 5 6 4 3 2 7 1
Judge II 6 4 5 1 2 7 3
Solution:
Difference
Score on Score on Rank Of Rank on Difference
Student
Test I Test II Test I Test II between squared
Ranks
X Y R1 R2 D D2
A 16 8 2 5 -3 9
B 14 14 3 3 0 0
C 18 12 1 4 -3 9
D 10 16 4 2 2 4
E 2 20 5 1 4 16
N=5 ∑D2 = 38
Example 9: The sales statistics of 6 sales representatives in two different localities. Find
whether there is a relationship between buying habits of the people in the localities.
Representative 1 2 3 4 5 6
Locality I 70 40 65 110 60 20
Locality II 70 30 80 100 90 20
Solution:
There is high positive correlation between buying habits of the locality people.
iii When Ranks are repeated
Example 10
Student A B C D E F G H I J
Score on Test I 20 30 22 28 32 40 20 16 14 18
Score on Test II 32 32 48 36 44 48 28 20 24 28
Difference
Score on Score on Rank Of Rank on Difference
Student
Test I Test II Test I Test II between squared
Ranks
X Y R1 R2 D D2
A 20 32 6.5 5.5 0 1.00
B 30 32 3 5.5 - 2.5 6.25
C 22 48 5 1.5 3.5 12.25
D 28 36 4 4 0 0
E 32 44 2 3 - 1.0 1.00
F 40 48 1 1.5 - 0.5 0.25
G 20 28 6.5 7.5 - 1.0 1.00
H 16 20 9 10 - 1.0 1.00
I 14 24 10 9 1.0 1.00
J 18 28 8 7.5 0.5 0.25
N = 10 ∑D2 = 24
“t” test is used to test correlation coefficient. Height and weight of a random sample of
six adults
It is reasonable to assume that these variables are normally distributed, so the Karl
Pearson Correlation coefficient is the appropriate measure of the degree of association
between height and weight. R = 0.875
H1: ρ > 0 This implies that there is positive correlation in the population (increasing
height is associated with increasing weight) 5% significance level is taken
Since the calculated value is more than the table value. Null hypothesis is rejected. There
is significant positive correlation between height and weight.
Partial Correlation
Partial Correlation is used in a situation where three and four variables involved. Three
variables such as age, height and weight. Correlation between height and weight can be
computed by keeping age constant. Age may be the important factor influencing the
strength of relationship between height and weight. Partial Correlation is used to keep
constant the effect of age. The effect of one variable is partialled out from the correlation
between other two variables. This statistical technique is known as partial correlation.
Partial Correlation is denoted by the symbol r12.3. Here correlation between variable 1 and
2 keeping 3rd variable constant.
r12.3 = Partial correlation between variables 1 and 2 keeping 3rd constant
Similarly,
and
Multiple Correlation
Three or more variables are involved in multiple correlations. The dependent variable is
denoted by X1 and other variables are denoted by X2, X3 etc. Gupta S.P, has expressed
that “the coefficient of multiple linear correlation is represented by R1 and it is common
to add subscripts designating the variables involved. Thus R1.234 would represent the
coefficient of multiple linear correlations between X1 on the one hand X2, X3 and X4 on
the other. The subscript of the dependent variable is always to the left of the point:
The coefficient of multiple correlations for r12, r13 and r23 can be expressed
The following formula can be used to determine a multiple correlation coefficient with
three independent variables.
Multiple correlation analysis measures the relationship between the given variables. In
this analysis the degree of association between one variable considered as the dependent
variable and a group of other variables considered as the independent variables.
Example 11: The following zero order correlation coefficients are given
Calculate multiple correlation coefficient treating first variable as dependent and second
and third variables as independent. (source: Gupta S.P, Statistical Method)
Solution:
Using the formula for multiple correlation coefficient for R1.23 we get:
Learning Objective 2
Regression
Regression is defined as, “the measure of the average relationship between two or more
variables in terms of the original units of the data.”
Correlation analysis attempts to study the relationship between the two variables x and y.
Regression analysis attempts to predict the average x for a given y. In Regression it is
attempted to quantify the dependence of one variable on the other. Example: There are
two variables x and y. y depends on x. The dependence is expressed in the form of the
equations.
Regression Analysis
Regression Analysis used to estimate the values of the dependent variables from the
values of the independent variables.
Regression analysis is used to get a measure of the error involved while using the
regression line as a basis for estimation.
Regression Lines
For a set of paired observations there exist two straight lines. The line drawn such that
sum of vertical deviation is zero and sum of their squares is minimum, is called
Regression line of y on x. It is used to estimate y – values for given x – values. The line
drawn such that sum of horizontal deviation is zero and sum of their squares is minimum,
is called Regression line of x on y. it is used to estimate x – values for given y – values.
The smaller angle between these lines, higher is the correlation between the variables.
Y – = byx (X –)
X– = bxy (Y – )
Where
And
The regression equations found by the above conditions is said to fitted by method of
least squares. byx and bxy are called regression coefficients.
• byx . bxy ≤ 1
• If byx is –ve, then bxy is also –ve and r is –ve.
• They can also be expressed as
•
• It is an absolute measure.
Age of Husband 18 19 20 21 22 23 24 25 26 27
Age of Wife 17 17 18 18 19 19 19 20 21 22
Solution:
Age of husband (x) dx = x-22 dx2 Age of wife (y) dy = y-19 dy2 dx dy
18 -4 16 17 -2 4 8
19 -3 9 17 -2 4 6
20 -2 4 18 -1 1 2
21 -1 1 18 -1 1 1
22 0 0 19 0 0 0
23 1 1 19 0 0 0
24 2 4 19 0 0 0
25 3 9 20 1 1 3
26 4 16 21 2 4 8
27 5 25 22 3 9 15
Total 225 5 85 190 0 24 43
Regression equation of Y on X is
Y – = byx (X – )
⇒ Y – 19 = 0.521 (X – 22.5)
⇒ Y = 0.521X + 7.2775
Series X Series Y
Mean S.D 65 67
S.D 2.5 3.5
Correlation
0.8
coefficient
Solution:
also
also
Example 13
1. The following results were worked out from scores in Statistics and Mathematics
in a certain examination.
Scores in Statistics (X) Scores in Mathematics (Y)
Mean 40 48
Standard Deviation 10 15
2.
Find the regression lines x on y and y on x. Use the regression lines to find the value of y
when x = 50 and value of x when y = 30.
Solution:
Is (Y – ) = r σ y / σ x (X – )………….(2)
Y = 0.63 x + 22.80…………(4)
X 12 4 20 8 16
Y 18 22 10 16 14
Solution
X– Y–
X Y (X – )2 (Y – )2 (X – ) (Y – )
X – 12 Y – 16
1
8 0 2 0 4 0
2
4 22 -8 6 64 36 - 48
2
10 8 -6 64 36 - 48
0
8 16 -4 0 16 0 0
1
14 4 -2 16 4 -8
6
160 80 - 104
And
X – 12 = – 1.3 (Y – 16)
Y – 16 = – 0.65 (X – 12)
Example 14: Estimating dependent variable X1 from the independent variables X2,
X3………….. It is known as regression equation of X1 on X2, X3…………..
a1.23 = (Constant) the intercept made by the regression plan. It gives the value
X1 = (X1 – )
X2 = (X2 – )
X3 = (X3 – )
Reliability of Estimates
Reliability of estimates test the estimated value obtained by applying regression equation,
whether the estimated value is very close to actual observed value. Standard error uses to
measure the closeness of estimate derived from the regression equation to actual observed
values. The measure of reliability is an average of the deviations of the actual value of
non-dependent variable from the estimate from the regression equation. Determining the
accuracy of estimates from the multiple regression is reliability of estimates. It is also
known as standard error of estimate.
Where
Summary
In this unit we studied the concept of correlation and regression and the different types of
correlation and regression. We saw how regression helps us to study unknown variables
with the help of known variables. It also establishes reliability measure for estimated
values.
MB0024-Unit-13
Introduction
The growing competition, rapidity of change in circumstances and the trend towards
automation demand that decisions in business are not based purely on guesses and
hunches rather on a careful analysis of data concerning the future course of events. The
future is unknown to us. Yet every day we are forced to make decisions involving future
and therefore uncertainty. Great risk is associated with business affairs. All businessmen
are forced to make forecast regarding business activities.
One of the most important task before businessmen and economists these days are to
make estimates for the future. For example, a business man is interested in finding out his
likely sales next year or as long term planning in next five or ten years so that he could
adjust his production accordingly and avoid the possibility of either inadequate
production to meet the demand or unsold stocks. Similarly, an economist is interested in
estimating the likely population in the coming years so that proper planning can be
carried out with regard to jobs for the people, food supply etc. First step in making
estimates for the future consists of gathering information from the past. In this connection
we usually deal with statistical data which are collected, observed or recorded at
successive intervals of time. Such data are generally referred to as Time series. Thus
when we observe numerical data at different points of time the set of observations is
known as time series.
Learning Objective 1
Business Forecasting
Business forecasting refers to the analysis of past and present economic conditions with
the object of drawing inferences about probable future business conditions. The process
of making definite estimates of future course of events is referred to as forecasting and
the figure or statements obtained from the process is known as ‘forecast’ future course of
events is rarely known. In order to be assured of coming course of events, help is taken of
an organized system of forecasting. These are two aspects of scientific business
forecasting.
i. Analysis of past economic conditions: For this purpose, the components of active series
are to studied. The secular trend will show how the series has been moving in the past
and what its future course is likely to be over a long period. The cyclic fluctuations would
reveal whether the business activity is subjected to boom or depression. The seasonal
fluctuations would indicate the seasonal changes in the business activity.
ii. Analysis of present economic conditions: The object of analyzing present economic
conditions is to study those factors which affect the sequential changes expected on the
basis of the past conditions. Such factors are new inventions, changes in fashion, changes
in economic and political spheres, economic and monetary policies of the Government,
war etc. These factors may affect and alter the duration of trade cycle. Therefore it is
essential to keep in mind the present economic conditions since they have an important
bearing on the probable future tendency.
Forecasting is a part of human conduct. Businessmen have also to look to the future.
Success in business depends on correct predictions. In fact when a man enters business,
he automatically takes with it the responsibility for attempting to forecast the future and
to a very large extent his success or failure would depend upon the ability to forecast
successfully the future course of events. Since without same element of continuity
between past, present and future, there would be little possibility of successful prediction.
But history is not likely to repeat itself and we would hardly expect economic conditions
next year or over the next ten years to follows a clear cut prediction. Yet, frequently past
patterns prevail sufficiently to justify using the past as a basis for predicting the future.
While forecasting, we should know that it is impossible to forecast the future precisely –
these always time must be same range of error allowed in the forecast. Statistical
forecasts are those in which we can use the mathematical theory of probability to measure
the risks of errors in predictions.
A great amount of confusion seem to have grown up in the use of words ‘forecast’,
‘prediction’ and ‘projection’. A prediction is an estimate based solely in past data of the
series under investigation. It is purely mechanical extrapolation. A projection is a
prediction where the extrapolated values are subjects to a certain numerical assumptions.
A forecast is an estimate which relates the series in which we are interested to external
factors. Forecasts are made by estimating future values of the external factors by means
of prediction, projection or forecast and from these values calculating the estimate of the
dependent variable.
i. Based on past and present conditions: The business forecasting is based on past and
present economic condition of the business. To forecast the future, various data,
information and facts concerning to economic condition of business for past and present
are analyzed.
ii. Based on mathematical and statistical methods: The process of forecasting includes
the use of statistical and mathematical methods. By using these methods the actual trend
which may take place in future can forecasted.
iii. Period: The forecasting can be made for long term, short term, medium term or any
specific term.
iv. Estimation of future: The business forecasting is to forecast the future regarding
probable economic conditions.
i. Understanding why changes in the past have occurred: One of the basic principles of
statistical forecasting is that the forecaster should use the data on past performance. The
current rate and changes in the rate constitute the basis of forecasting. Once they are
known various mathematical techniques can develop projections from them. If an attempt
is made to forecast business fluctuations without understanding why past changes have
taken place, the forecast will be purely mechanical based solely upon the application of
mathematical formulae and subject to series error.
ii. Determining which phases of business activity must be measured: After it is knowing
why business fluctuations have occurred, it is necessary to measure certain phase of
business activity in order to predict what changes will probably follow the present level
of activity.
iv. Analyzing the data: Lastly, the data are analyzed in the light of understanding of the
reason why change occurs. For example, if it is reasoned that a certain combination of
forces will result in a given change, the statistical part of the problem is to measure these
forces, from the data available, to draw conclusions on the future course of action. The
methods of drawing conclusions may be called forecasting techniques.
Learning Objective 2
Almost all the businessmen make forecasting about the business conditions related to
their business. In recent years scientific methods of forecasting have been developed. The
base of scientific forecasting is statistics. To handle the increasing variety of managerial
forecasting problems, several forecasting techniques have been developed in recent years.
Forecasting techniques vary from simple expert guesses to complex analysis of mass
data. Each techniques has its special use, and care must be taken to select the correct
technique for a particular situation. Before applying a method of forecasting the
following questions should be answered:
i. What is the purpose of the forecast how is it to be used?
ii. What are the dynamics and components of the system for which the forecast will be
made?
1. Business Barometers
Business indices are constructed to study and analyze the business activities on
the basis of which future conditions are predetermined. As business indices are
the indicators of future conditions, so they are also known as “Business
Barometers” or ‘Economic Barometers’. With the help of these business
barometers the trend of fluctuations in business conditions are made known and
by forecasting a decision can be taken relating to the problem. The construction of
business barometer consists of gross national product, wholesale prices, consumer
prices, industrial production, stock prices, bank deposits etc. These quantities may
be concerted into relatives on a certain base. The relatives so obtained may be
weighted and their average be computed. The index thus arrived at in the business
barometer.
ii. Business barometers for specific business or industry: These barometers are
used as the supplement of general index of business activity and these are
constructed to measure the future variations in a specific business or industry.
Advantages:
i. The business barometer method is scientific and reliable and used by
management for the purpose of various business decisions at different levels.
iii. The business barometers are the indicators of future business trends and help
to forecast the speed of fluctuations.
iv. This method helps to find solution of various business problems such as
development of market, capital investment, exploration of new consumer market
etc.
Disadvantages:
ii. In most of the cases, the business barometers provide inaccurate, incomplete
and conclusive forecasting due to index numbers prepared on the basis of
incorrect and inadequate data.
iii. The business barometers are the indicators of past conditions and the
forecasting based on these conditions may be erroneous.
iv. Separate indices are calculated for individual industry and firm which are
entirely different from general indices.
Time series analysis is also used for the purpose of making business forecasting.
The forecasting through time series analysis is possible only when the business
data of various years are available which reflects a definite trend and seasonal
variation. By time series analysis the long term trend, secular trend, seasonal and
cyclical variations are ascertained, analyzed and separated from the data of
various years.
Merits
Demerits
iii. This method can only be used when the data for several years are available.
1. Extrapolation
ii. Knowledge about the course of events relating to the problem under
consideration.
In extrapolation we assume that the variable will follow the established pattern of
growth. For the purpose of business forecasting it is to determine accurately the
appropriate trend curve and the values of its parameters. some of these curves are:
i. Arithmetic trend: The straight line arithmetic trend assumes that growth will be
a constant amount each year.
ii. Semi log trend: It assumes a constant percentage increase each year. As the
annual increment is constant in logarithm, this line will become a straight line
when drawn on semi log paper.
iii. Modified exponential curve: The curve is given by y = abx. This relationship is
referred to as an exponential function. It assumes that each increment of growth
will be a constant per cent of the previous one.
iv. Logistic curve:
Yc = abcx
Merits:
Demerits:
2. Regression Analysis
The regression approach offers many valuable contribution to the solution of the
forecasting problem. It is the means by which we select from among the many
possible relationships between variables in a complex economy those which will
be useful for forecasting. Regression relationship may involve one predicted or
dependent and one independent variables simple regression, or it may involve
relationships between the variable to be forecast and several independent
variables under multiple regressions. Statistical techniques to estimate the
regression equations are often fairly complex and time-consuming but there are
many computer programs now available that estimate simple and multiple
regressions quickly.
At the present time, most short-term forecasting uses only statistical methods with
little qualitative information. However, in the years to come when most large
companies develop and refine econometric models of their major business, this
tool of forecasting will become more popular.
Merits:
i. Accurate and reliable results are obtained under this method because it is
a scientific method where computer is used.
ii. This method explains in detail and in quantitative terms the way in
which various aspects of the economy are interrelated.
Demerits:
ii. This method can be used only when adequate series of data is available.
The forecaster should use a technique that makes the best use of available
data. Furthermore, where a company wishes to forecast with reference to a
particular product, it must consider the stage of the products life cycle for
which he is making the forecasts.
There are a few theories that are followed while making business forecast. Some
of them are:
iv. Government can use this technique for the purpose of economic
stability of the economy by exercising control over possible losses.
Demerits
ii. This method can not be regarded accurate because by using statistical
techniques the results can be up to the truth but not accurate one.
This theory is based on two assumptions: every action has a reaction, and the
magnitude of the original action influences the reaction. Thus if the price of rice
has gone up above a certain level in a certain period, there is likelihood that after
some time it will go down below the normal level. Thus, according to this theory
a certain level of business activity is normal or abnormal, conditions cannot
remain so for ever. Thus, we find four phases of a business cycle.
i.. Prosperity
ii. Decline
iv. Improvement
Merits
ii. By this theory more reliable results can be obtained because this theory
gives attention to action and reaction of event.
Demerits
The basic assumption of this theory is that history repeats itself and hence assume
that all economic and business events behave in a rhythmic order. According to
this theory, the speed and time of all business cycles are more or less same and by
using statistical and mathematical methods a trend is obtained which will
represent a long term tendency of growth or decline.
It is done on the basis of the assumption that the trend line denotes the normal
growth or decline of business events.
Merits
Demerits
i.The business events are not strictly periodic and prediction of business
cycle on the basis of statistical method is not satisfactory.
ii.Past conditions are given more weight-age than the present conditions
History repeats itself is the main foundation of this theory. Whatever happened in
the past under a set of circumstances is likely to happen in future also if
conditions are the same. A time series relating to the data in question is
thoroughly scrutinized and from it such period is selected in which conditions
were similar to those prevailing at the time of making the forecast but it is largely
dependent on past data.
Merits
i. It is an easy method
ii. As the future is forecasted on the basis of past business conditions, the
forecasting will be more reliable.
Demerits
ii. It is very difficult to select the past period with the same business
conditions like present.
Merits
Demerits
ii. Past facts are equally important for the purpose of forecasting but in this
method no weight-age in given to past.
iii. The forecasting made on the basis of this technique cannot be regarded
reliable.
Utility of Business Forecasting
iv. Basis for capital market: The business forecasting helps in estimating the
requirement of capital, position of stock exchange and the nature of investors etc.
v. Useful in controlling the business cycles: The trade cycles cause various
depressions in the business such as sudden change in price level, increase in the
risk of business, increase in unemployment etc. By adopting a systematic business
forecasting the businessman and government can handle and control the
depression of trade cycles.
vi. To achieve the goals: The business forecasting help to achieve the objective of
business through proper planning of business activities.
viii. Utility to society: With the help of business forecasting the entire society is
also benefited because the adverse effects of fluctuations in the conditions of
business are kept under control.
iv. The forecasting cannot guarantee the elimination of errors and mistakes. The
managerial decision will be wrong if the forecasting is wrong.
v. Factors responsible for economic changes are often difficult to discover and to
measure. Hence business forecasting becomes an unnecessary exercise.
vii. The forecasting is made on the basis of past information and data and relies
that economic events are repeated under the same conditions. But there may be
circumstances where these conditions are not repeated
Summary
MB0024-Unit-14
Introduction
A time series is a set of numerical values of a given variable listed at successive intervals
of time. That is, the data regarding the variable is listed in chronological order. Usually
the interval of time is taken as uniform.
Example: Yearly production of wheat in the country, hourly temperature of a city,
bimonthly electricity bills etc. Almost all the data like industrial production, agricultural
production, exports, imports, diary products can be arranged in chronological order.
Learning Objective 1
i. Study the forces that influence the variations in time series, and
ii. Study the behaviour of phenomenon over the given period of time.
For example, consider the sale of T.V sets (in thousands) by a producing company
We would like to analyse the above data and give some trends about the sales. For
example, the company would like to know as to why the sales dropped in 1998 and 1999,
and then why the sales increased. That is, the company would like to analyse the various
forces that affect the sales.
There can be changes in the values of the variable recorder over different points of time
due to various forces. Analysing the effect of all such forces on the values of the variable
is generally known as the analysis of time series. Broadly there can be four types of
changes in the values of the variable as discussed below:
i. Changes which generally occur due to general tendency of the data to increase
or decrease.
ii. Changes which occur due to change in climate, weather conditions, festivals
etc.
iii. Changes which occur due to booms and depressions.
iv. Changes which occur due to some unpredictable forces like floods, famines,
earthquakes etc.
Learning Objective 2
The behaviour of a time series over periods of time is called the movement of the
time series. The time series is classified into the following four components:
Examples:
i. The prices of vegetables drop down after rainy season or in winter months
and they go up during summer, every year.
ii. The prices of cooking oils reduce after the harvesting of oil seeds and go up
after some time.
Learning Objective 3
We shall be studying the following methods of measuring the trend of a time series:
ii. It cannot be used for any predictions of trends, as drawing the trend curve is
arbitrary.
Example: Find trend with the help of freehand curve method for the data given below:
Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Productio
n (in lakh 15 18 16 22 19 24 20 28 22 30 26
ton)
Solution:
1. Semi-Average Method:
The methods of fitting a linear trend with the help of semi average method are as follows:
i. The number of years in even: The data of the time series are divided into
two equal parts. The total of the items in each of the part is done and it is
then divided by the number of items to obtain arithmetic means of the two
parts. Each average is then centred in the period of time from which it has
been computed and plotted on the graph paper. A straight line is drawn
passing through these points. This is the required trend line.
ii. The number of years is odd: When the number of years is odd, the value
of the middle year is omitted to divide the time series into two equal parts.
Then the procedure (i) is followed.
A trend value of any future year may be predicted by multiplying the periodic increment
by the number of years into the future that is desired and adding the result to the best
trend value listed in the series.
Merits:
ii. The trend line can be extended on either side in order to obtain past or future
estimates.
iii. This is an objective method, as any one applying this method get the same
trend line.
Demerits:
i. The method of semi average assumes a straight line relationship between the
plotted points, regardless of the fact whether such relationship exists or not.
ii. This method has an in built limitation of arithmetic mean. This method is not
suitable is case of very low or very large extreme values.
This method is used for smoothing the time series. That is, it smoothens the
fluctuations of the data by the method of moving averages.
iii. Compute moving totals according to the length of the period of moving
average.
If the length of the period of moving average is 3 i,e., 3-yearly moving average is
to be calculated, compute moving totals as follows:
a + b + c, b + c + d, c + d + e, d + e + f…..
a + b + c + d + e, b + c + d + e + f, c + d + e + f + g…..
Placing the moving totals at the centre of the time span from which they are
computed.
iv. Compute moving averages by moving totals in step (3) by the length of the
period of moving average and place them at the centre of the time span from
which the moving totals are computed. These moving averages are also called the
trend values.
By plotting these trend values (if desired) one can obtain the trend curve with the help of
which we can determine the trend whether it is increasing or decreasing.
If needed, one can also compute short-term fluctuations by subtracting the trend values
from the actual values.
Illustrative Example:
Year 1998 1989 1990 1991 1992 1993 1994 1995 1996 1997
Production (in
15 18 16 22 19 24 20 28 22 30
lakh ton)
ii. Obtain the length of the period of moving average. Let the length of the
moving averages period be 4-years.
iii. Compute 4 yearly moving totals and place them at the centre of time span.
The four – yearly moving totals are computed as follows:
a + b + c + d, b + c + d + e, c + d + e + f,
iv. Compute 4 – yearly moving average and place them at the centre of the
time span. Note that this placement is inconvenient, because the moving
average so placed would not coincide with original time period.
v. Take two – period moving average of moving averages and place them at
the middle of the periods. This process is called centring of moving averages.
ii. This method is objective in the sense that any body working on a problem with
this method will get the same results.
iii. This method is used for determining seasonal, cyclic and irregular variations
besides the trend values.
iv. This method is flexible enough to add more figures to the data because the
entire calculations are not changed.
Limitations:
i. There is no functional relationship between the values and the time. Thus, this
method is not helpful in forecasting and predicting the values on the basis of time.
ii. There are no trend values for some year in the beginning and some in the end.
For example, for 5 – yearly moving average there will be no trend values for the
first two years and the last three years.
iii. In case of non – linear trend the values obtained by this method are biased in
one or the other direction.
iv. The selection of the period of moving average is a difficult task. Therefore
great care has to be taken in selecting the period, particularly, when there is no
business cycle during that time.
Under this method the trend curve is determined by fitting a mathematical equation. This
method is more accurate and precise and can be used even for forecasting. We can fit
either a straight line or a parabolic curve from the given data by this method. Let y be the
actual values of y and yc be the computed values of y for a given value of x.
Let y = a + bx be a straight line to be fitted for trend. To find the values of a and b such
that the sum of squares of differences of the actual and computed values of y is least, i.e
∑ (y – yc)2 is least, where the condition ∑ (y – yc) = 0 is satisfied, is known as method of
least squares. The line obtained by the method is known as the ‘line of best fit.’
For a given time series data, to find a linear trend, the values of a and b are obtained by
the normal equations.
Here a is intercept of the line on they y – axis and b is the slope of the line, b is
also known as growth rate (if b > 0) or decline rate (if b<0), b gives the change in
the value of y, for per unit change in the value of x.
Direct method
i. Convert the years into natural numbers ( 1, 2, 3……) and denote by x and find
∑ x.
iii. Multiply the x – values with corresponding y – values and obtain ∑ xy.
v. Put these values in the two normal equations and solve for a and b.
vi. Substitute these values of a and b in y = a + bx and then find trend values for
various values of x.
Measure the variables x from any point of time in origin as the first year, but the
calculations are simplified when the mid – point in time is taken as origin so that ∑ x = 0,
when ∑ x = 0, then normal equations reduce to:
Merits
ii. This method gives the trend values for the entire time period.
iii. This method can be used to forecast future trend because trend line establishes a
functional relationship between the value and the time.
Demerits
ii. If even a single item is added to the series a new equation has to be formed.
iii. Future forecasts made by this method are based only on trend values. Seasonal,
cyclical or irregular variations are ignored.
Fitting a Parabolic Curve or Non – linear Curve by the Method of Least Squares.
When the time series data do not confirm with the linear trend then we obtain Non –
linear trend. For this we use the equation of the form.
If we can change the origin at a suitable point such that Σ x = 0, then the normal
equations reduce to:
Mathematical Models for Time series
The following are the two models commonly used for the decomposition of a time series
into its components.
1. Additive Model: This model assumes that the observed value is the sum of
four components of time series, i,e.,
Y = T + S + C + 1,
The additive model for decomposition of time series assumes that all the four
components of the time series operate independently of one another. It also
assumes that the behaviour of components is of an additive character. It is to
be noted that only absolute values are added or deducted from the trend value
to arrive at the observed value.
Y = T x S x C x 1.
i. Time Variation : When data are available on monthly basis, the effect of
time variation needs to be adjusted because all months of the year do not
have the same number of days. This adjustment of time variation is done by
dividing each monthly total by daily average, it is then multiplied by 365 /
12 which is the average number of days in a month.
iv. Comparability: In order to have valid conclusion the data which are
being analysed should be comparable. When we are dealing with the
analysis of time series it involves the data relating to past which must be
homogeneous and comparable. When we are dealing with the analysis of
time series it involves the data relating to past which must be homogenous
and comparable. Therefore, effects should be there to make the data as
homogeneous and comparable as possible.
Symbolically
L1=(S1/S)X100
j = 1, 2, 3, 4……..k
Merits
ii. This method is useful where no definite, trend exists in the time series.
Demerits / Limitations
i. Most economic time – series have trends and therefore, the seasonal index
computed by this method is really an index of trends and seasons.
ii. The simple averages method of isolating seasonal fluctuations in time – series is
based on the assumption that the series contains only the seasonal and irregular
fluctuations.
iii. This method does not give a true reflection of the normal seasonal variation
because it is obtained from the original data which are affected by not only seasonal
movements but also by remaining three components.
iv. The effects of cycles of the original data are not eliminated by the process of
averaging.
This method is also known as Percentage of Moving Average Method. The steps
involved in the computation of seasonal indices by this method are as follows:
i. The moving averages of the data are computed. If the data are monthly then
12 – monthly moving average, if they are quarterly, then 4 – quarterly moving
averages will be computed. In both the cases time periods of moving averages are
even, hence these moving averages are to be centred.
ii. Under additive model, from each original value, the corresponding moving
average is deducted to find out short time fluctuations:
iii. Y–T=S+C+1
iv. By preparing a separate table, monthly (or quarterly) short time fluctuations
are added for each month (or quarter) over all the years and their average is
obtained. these averages are known as seasonal variations for each month or
quarter.
ii. The mean of the Link Relatives for each season is computed over all the
years. Median can also be taken instead of mean of the Link Relatives.
iii. These average link, relatives are converted into chain relatives. The chain
relative of first is taken as 100.
iv. The second chain relative of first is computed on the basis of the chain
relative for the last:
Chain relative of the first quarter = (Average Link Relative of the first x Chain
Relative of the last)/100
This chain relative may or may not be 100. It is not equal to 100 due to secular
trend. If it is 100 go to step (vi), if it is not 100 go to step (v) and then step (vi).
v. Compute the difference d between the new chain relatives of first obtained in
step (iv) and chain relative assumed as 100. d is divided by the number of seasons
and the resulting figure is multiplied by 1, 2, 3 and the product is deducted
respectively from the chain relatives of 2nd, 3rd, and 4th quarters. These are called
corrected relatives.
vi.The chain relatives obtained is step (iv), if correction is not necessary for the
corrected chain relatives obtained in step are expressed as percentages of the
average to have adjusted chain relatives.
ii. To find ratio to trend, divide the original data by the corresponding trend
values and multiply these ratios by 100,i,e
iii. Calculate the Arithmetic Mean of the Trend Ratios obtained in step (ii).
iv. Finally all the trend ratios will be converted into seasonal indices. For this
add all averages obtained in (iii) and find their General Average. Seasonal indices
are calculated by using the following formula:
In this method the trend effect and cyclic effects do not come into account.
1. Naïve Forecast: In this method we forecast the value, for the time period t,
to be equal to the actual value observed in the previous period, i,e, time period
(t-1). This is given as
Y t = yt – ν
Where x-value will be calculated from the value of t and the constants a.
Learning Objective 5
ii. Forecasting can be done using the time series. By studying the variations and
other behaviour of the variables over a sufficiently long period of time, it may be
possible to forecast the future behaviour of the variables. However, such a
forecast has meaning only if the period of forecast is a normal period. For
example, various five-year plans by the Government of India are formulated by
studying the time series and forecasting.
iii. Study of the time series helps in analysing the post behaviour of the
variables. This helps in identifying the various forces that effect its behaviour.
Summary
In this unit we studied about the business forecasting. The different step involved in
forecasting is discussed in a simple manner. The concept of time series analysis is
discussed next with good examples. Action and reaction theory is explained with its
merits and demerits in a simple manner. Lastly in this unit we discuss about the method
of least squares with merits and demerits discussed in detail.
MB0024-Unit-15
Introduction
We know the most values change and therefore may want to know-how much changes
has taken place over a period of time. For example, we may want to know how much the
prices of different times essential to a household have increased or decreased so that
necessary adjustments can be made in the monthly budget. However, while price of a few
items may have increased, others may have decreased over a given period of time.
Consequently, in all such situations, an average measure needs to be defined to compare
such difference over a time period. Index numbers are yardsticks for describing such
differences.
Learning Objective 1
Learning Objective 2
Relative
The value of a variable in a given year (or place) divided by the value of the same
variable in a specified year (or place) is called a relative and is generally, expressed in
percentage.
a.Price Relative: The price of commodity in a given year expressed as a percentage of the
price of the same commodity in a specified year is called price relative.
Suppose the price of a commodity in India in 2001 was Rs.95 per kg and in 2000 it was
Rs.80 per kg
Then the price relative for 2001, (using 2000 as base) is: 95 / 80 x 100 = 118.75%
b.Production Relative: If the wheat production in India in 2002 was 5,82,000 metric tons
and in 2004 it was 6,96,000 metric tons, then assuming the production of 2002 as 100, the
production relative for 2004 is equal to (696000/582000)x100=119.6%
c.Quantity Relative: The quantity (q1) of a commodity consumed in a given year
expressed as a percentage of the quantity (q0) of the same commodity consumed in a
specified year is called Quantity Relative.
d.Value Relative: If p1 and q1 are the price and quantity respectively for a commodity in a
given and p0 and q0 are the specified price and quantity respectively of the same
commodity, in a specified year, then V1 = p1q1 is the value of given year and V0 = p0q0 is
the value of the specified year.
The ratio V1 / V0 x 100 = p1q1 / p0q0 x 100 is called the value relative of the specified year
with respect to the given year.
The overall change in price, production, quantity or value etc. is represented by these
typical summaries which are known as relatives.
1. Based on Variables.
iii.Simple aggregate
iv.Weighted aggregate.
In the computation of an index number we require two years (or places). The given year
whose values are to be compared is called a current year (or current period) and the
specified year whose values are taken as standard (say 100) is called a Base year {Base
Period}. For example, if the prices of 2005, are compared with the prices 0f 2004, then
2005 is the current year and 2004 is the base year. The index number of 2005 based on
2004, in general, denoted by O01or P01, where 0 stands for 2004, and 1 stands for 2005.
iii.Relative measure: Index numbers measure changes which are not capable of
direct measurement.
v.Basis of Comparison: Index numbers by their very nature are comparative. They
compare changes overtime or between places or like categories.
Learning Objective 3
To follow the steps involved in the construction of index numbers many problems are
encountered which are to be discussed carefully:
1. Purpose of Index Number: The steps which are taken in the construction of index
numbers generally depend on the purpose of the index number. Hence the purpose
of an index numbers must be defined clearly and precisely. For example, the
purpose of the general index number of wholesale price index number is to know
the general price level, while that of consumer price index number is to give an
idea of the effect of the change in retail prices on the cost of living of classes of
people.
1. Selection of Base Period: The base period of an index number is the period of
time against which the comparisons are made. There are three types of based
period.
iii.Chain base
While selecting the base a decision has to be made so as to whether we have fixed
base or chain base in a fixed base (a single period):
iv.The base period must be a normal period. By normal period we mean that
period which is free from all sorts of abnormalities or random causes such as
financial crisis, floods, famines, earth quakes, strikes of labourers, wars etc.
v.The base period should be a period for which reliable figures are available.
When it is difficult to choose just one single period as the normal, then a better
choice will be an average of several periods.
If the comparisons are required form year to year a system of chain base is used.
In this method, there is 10 fixed base for comparing the values of subsequent
years, but the value of each year is compared with the value of the preceding year.
2. Selection of Commodities:
i.It represents the real tastes, habits and the customs of the people,
ii.It should be of a standard quality and there must be no significant
variation in the quality,
ii.Explicit method: In this method, the weights are laid down on the basis
of one outward evidence of importance of commodities. One fo the
problems in the selection of appropriate weight is to decide this evidence.
Another problem with regard to the system of weighting is whether
weights should be fixed or fluctuating.
3. Selection of the Average: To find composite index number we can use any
average such as arithmetic mean, geometric mean, harmonic mean, median and
mode. The use of an average depends on the relative merits and demerits of the
various averages. The average may be weighted or unweighted.
4. Selection of Suitable Formula: There are various formulae for computing index
numbers so the selection of a suitable formula also possess some problem. A
particular formula is suitable in a particular situation.
Learning Objective 4
In unweighted index numbers each item is supposed to have the same weight but
in weighted index numbers the weights are assigned to various items in
accordance with their importance.
i. Add the prices of all commodities in the current year, i,e., find
∑p1
ii. Add the prices of all commodities in the base year, i,e., ∑p0
iii. Divide the total of current year prices by the total of base year
prices and multiply the quotient by 100, i,e
iv.Here I01is the simple price index number of current year (1)
based on based year (0).
Merits: This is the simplest method of constructing index numbers because it is
simple to understand and requires simple calculations.
Demerits:
ii. Since weights are not used, this method does not give any
consideration to the relative importance of commodities.
Example 1: Find the simple aggregative price index from the following data:
Solution:
Therefore
This implies that the prices had increased by 7.7% in year 2004 as compare to the
year 2000.
R=(p1/p0)X100
Merits / Advantages:
iii. It gives equal importance to all items and extreme items do not unduly
affect the index number.
iv. The index number calculated by this method satisfies the unit test.
Demerits / Limitations
i. As it is an unweighted average the importance of all items is assumed to
be the same.
ii. The index number constructed by this method does not satisfy all the
criterion laid down for an ideal index.
iii. The index number is unduly influenced by high or low prices when
arithmetic mean is used.
To meet the weakness of the simple or unweighted method, we weight the price
of each commodity by a suitable factor often take as the quantity or the volume of
the commodity sold during the base year. In other words, in this method
appropriate weights are assigned to various commodities to reflect their relative
importance in the group. The weight can be production figures, consumption
figures or distributive figures. For the construction of the price index number
quantity weights are used. If w is the weight attached to a commodity, then the
price index is given by
Laspeyre’s Price Index: Laspeyre’s method is based on fixed weights of the base
year. Base year’s quantities are used as weights. The formula given by Laspeyre
is given below:
Laspeyre’s Price Index:
This index number has an upward bias i,e when prices increase, there is a
tendency to reduce the consumption of higher priced goods. The index
number is very widely used in practical work.
This index number has downward bias. This formula is not used frequently in
practice where the number of commodities in large.
Merits
ii.This formula takes into account both current years as well as base year
prices and quantities.
iii.It satisfies both ‘time several test’ as well as the ‘factor reversal test’.
This is why it is called an ideal index number.
Demerits:
iii. It requires the prices and quantities for base year and current yar.
The quantity index numbers measure the average storage in quantities and enable
us to compare changes in physical quantity of goods produced or sold. These
index numbers can also be simple or weighted. Weights in quantity index number
in price. Therefore quantity index numbers can be easily obtained from price
index numbers just by interchanging p’s and q’s in the above formulae.
The value index numbers are very easy to calculate. Value is the product of price
and quantity. A simple value index number is equal to the value of the current
year divided by the value of the base year. If this value is multiplied by 100 we
get the value index number. The required formula is:
Such Index numbers are not weighted because they do not take into account either
the price or the quantity. These index numbers are not very popular because the
situation revealed by price and quantities are not fully revealed by the values.
The ‘cost of living index’, also known as “consumer price index’ or ‘cost of living
price index’ is the country’s principal measure of price change. It measures
average change over time in the prices paid by the consumer of specific baskets of
goods and services.
The consumer price index numbers are designed to measure the average change in
the price index numbers are designed to measure the average change in the price
paid by the ultimate consumers for specified quantities of goods and services over
a period of time. The consumer price index helps us in determining the effect of
rise and fall in prices on different classes of consumers living in different areas.
iii. Market price for a particular kind of goods and services are
analysed by consumer price index.
iv. The salaries and wages are fixed on the basis of consumer price
index. So, it is very helpful to revise wage of dearness allowance.
Assumptions: Cost of living index number is based on same assumptions which
are as follows:
i.Similar needs: The needs of the people for which this index number
is constructed are same.
ii.Same goods: The goods consumed in the base year and the current
year unchanged.
v.True on the average: Cost of living index numbers are true on the
average.
These are three methods for constructing consumer price index number:
P
ii.Family budget method: [or the method of weighted relatives
where weights are the value (P0Q0) in the base year often denoted
by V]
sane as (i)
Then
Learning Objective 5
There is no doubt that the technique of index numbers is a very useful tool. But
these are certain limitations of index numbers which should be borne in mind.
Learning Objective 6
Now-a-days, index number is not only used for measuring price changes alone. The
factors like wages, employment, production, trade, demand, supply, business condition,
industrial activity, financial problems etc., are also studied through this statistical device.
As a Barometer measures the pressure of atmosphere or gases so the index numbers
measure the pressure of economic behaviour, and thus the index numbers are called
economic Barometers.
• Comparative Study
• Simplifies data
• Provide guidelines to economic policy and in formulating decisions
• Measures purchasing power of money
• Change in cost of living
• National income
• It is used as control by government
• Reveal trends and tendencies
• Useful in deflating
• Comparative study is made possible
• Universal utility
Summary
In this unit we studied about the concept of index numbers, and classification of index
numbers into different types. The different index numbers formally available, the utility
and importance of index numbers are explained in a simple way.