MODULE-1

PROBABILITY DISTRIBUTIONS
Introduction
The three main components of statistics are collection of data, analysis of data and
inference. The scope of statistics is very vast and finds its applications in almost all spheres.
Statistical methods guides us in any scientific enquiry and leans heavily on probability
theory.
Basic Definitions and concepts
Experiment is a process by which measurement or observation is obtained.
Random experiment : An experiment which does not result in the same out come when
performed under the same conditions.
Eg : 1. Tossing of a fair coin.
2. Throwing of a fair die.
Sample space : The set of all possible outcomes of a random experiment is called a sample
space usually denoted by S.
Eg 1. S = H,T is the sample space in tossing of a single coin.
2. S = HH,TH,HT,TT is the sample space obtained in tossing of two coins.
3. S = 1,2,3,4,5,6 in throwing of a die.

Event : Any event is a subset of an appropriate sample space.

Exhaustive event : An event consisting of all the various possibilities is called an exhaustive
event.
Mutually exclusive events : Events are said to be mutually exclusive if and only if one of
them can take place at a time, in other words happening of one event prevent the
simultaneous happening of the others.
Eg: In the tossing of a coin, either heads or tails may turn up but not both.
Independent events: Two or more events are said to be independent if the happening or non
happening of one event does not prevent the happening or non happening of the others.

Eg: 1.When two coins are tossed the event of getting head is an independent event as both
the coins can turn out heads.
Probability is the measure of the chance with which we can expect the event to occur. If an
event A can occur in m different ways out of number of n possible ways, all of which are
m
equally likely then the probability of the event A denoted by P(A) is given by P(A) = n .

Note: 1) Since 0  m  n  P (A) is non negative and 0  P(A)  1

1
 2) If P(A) = 0 then A is called a null event.
3) If P(A) = 1 then A is called a sure event.
̅̅̅ = 1 , the set of all unfavourable events is denoted by A
4) P(A) + P(A) ̅.

Addition theorem of probability

The probability of the happening of one or the other mutually exclusive events is equal to
the sum of the probabilities of the two events.
That is, if A, B are two mutually exclusive events then,
𝐏(𝐀 𝐨𝐫 𝐁) = 𝐏(𝐀) + 𝐏(𝐁)
Product theorem of probability
If a compound event is made up of a number of independent events, the probability of the
happening of the compound event is equal to the product of the probabilities of the
independent events .
That is, if A, B are independent events then,
𝐏(𝐀 𝐚𝐧𝐝 𝐁) = 𝐏(𝐀). 𝐏(𝐁)

RANDOM VARIABLE & PROBABILITY DISTRIBUTION

Introduction
When we consider an experiment of tossing two coins. The sample points or out comes are
qualitative and are described by their attributes though the probabilities attached to each of
these outcomes are quantitative. We now associate to each sample point a real number called
the random variable.
Random variable
Definition: Let S be a sample space then X : S  R defined over the elements of S is
called a random variable. A random variable is a function that assigns a real number to
every sample point in the sample space of a random experiment. Random variables are
usually denoted by X, Y, Z … and it may be noted that different random variables may be
associated with the sample space S. The set of all real numbers of a random variable X is
called the range of X.
Eg:
1. While tossing a coin, suppose that the value 1 is associated for the outcome ‘head’ and 0
for the outcome tail. We have the sample space S = {H, T} and if X is the random
variable , then X(H) = 1 & X(T) = 0.
Range of X = {0,1}

2. Suppose a coin is tossed twice, we shall associate two different random variables X, Y as
follows, where we have the sample space
S = {HH, HT, TH, TT}
X = Number of ‘heads’ in the outcome.
The association of the elements in S to X is as follows.

2
 Out come HH HT TH TT

Random variable 2 1 1 0
X

Range of X = {0,1,2}
Suppose Y = Number of ‘tails’ in the outcome, we have

Out come HH HT TH TT

Random variable Y 0 1 1 2

Range of Y = {0,1,2}

Discrete and Continuous random variables

If a random variable takes finite or countably infinite number of values then it is called a
discrete random variable.
Discrete random variable will have finite or countably infinite range.
Eg: 1. Tossing of a coin and observing the outcome.
2. Throwing a ‘die’ and observing the numbers on the face.
If a random variable takes non countable number of values then it is called a non discrete or
continuous random variable.
Eg:
1. Weight of articles
2. Length of nails produced by machine.

Generally counting problems correspond to discrete random variables and measuring

problems lead to continuous random variables.
Probability function and Discrete probability distribution
If for each value xi of a discrete random variable X, we assign a real number p(xi ) such
that
i) p(xi ) ≥ 0 ii) ∑i p(xi ) = 1 then the function p(x) is called a probability function.
If the probability that X take the values xi is pi , then P(X = xi ) = pi or p(xi ).
The set of values [xi , p(xi )] is called a discrete (finite) probability distribution of the
discrete random variable X. The function P(X) is called the probability density function
(p.d.f) or the probability mass function (p.m.f)
The distribution function f(x) defined by f(x) = P(X ≤ x) = ∑ni=1 P(xi ) , x being an integer,
is called the cumulative distribution function (c.d.f).
The mean and variance of the discrete probability distribution are defined as follows

3
 Mean μ = ∑i xi . p(xi )
Variance V = ∑i(xi − μ)2 . p(xi )
V = ∑i xi2 . p(xi ) − [∑i xi . p(xi )]2
V = ∑ xi2 . p(xi ) − μ2
Standard deviation σ = √V

Problems
1. A coin is tossed twice. A random variable X represent the number of heads turning up.
Find the discrete probability distribution for X. Also find its mean and variance.
Sol: Here S = {HH, HT, TH, TT}
X = {0,1,2}
1 1 1 1
Now, P(HH) = 4 P(HT) = 4 P(TH) = 4 P(TT) = 4

1
P(X = 0, i. e. , no head turning up) = P(TT) =
4
1 1 1
P(X = 1, i. e. , one head turning up) = P(HT ∪ TH) = P(HT) + P(TH) = + =
4 4 2
1
P(X = 2, i. e. , two heads turning up) = P(HH) =
4
The discrete probability distribution for X is as follows

X=𝐱 𝐢 0 1 2

p(xi ) 1 1 1
4 2 4

We observe that, p(xi ) > 0 and ∑i p(xi ) = 1.

1 1 1 1 1
Mean , μ = ∑i xi . p(xi ) = (0). (4) + (1). (2) + (2). (4) = 2 + 2

𝛍=𝟏

Variance, V = ∑i(xi − μ)2 . p(xi )

1 1 1
V = (0 − 1)2 . 4 + (1 − 1)2 . 2 + (2 − 1)2 . 4
1 1
V= 4+0+4

𝟏
𝐕=𝟐

4
2. A random experiment of tossing a ‘die’ twice is performed. Random variable X on this
sample space is defined to be the sum of the two numbers turning up on the toss. Find the
discrete probability distribution for the random variable X and compute the corresponding mean
and standard deviation.
Sol: Here S = {(x, y) where x = 1,2, … ,6 ; y = 1,2, … ,6}
i.e., S = {(1,1), (1,2), (1,3), (1,4), (1,5), (1,6), (2,1), (2,2), (2,3), … , (6,5), (6,6)}
Number of elements in S=36.
The set of values of the random variable X defined as sum of two numbers on the face of the
‘die’ are {2,3,4,5, … ,10,11,12}
n(E) 1 2
Now, p(x1 ) = = 36 ; p(x2 ) = 36 ; …
n(S)

The discrete probability distribution for X is as follows

𝐗 2 3 4 5 6 7 8 9 10 11 12
= 𝐱𝐢

p(xi ) 1 2 3 4 5 6 5 4 3 2 1
36 36 36 36 36 36 36 36 36 36 36

Here p(xi ) > 0 and ∑i p(xi ) = 1.

Mean μ = ∑i xi . p(xi )
1 2 2 1
μ = (2). (36) + (3). (36) + ⋯ + (11). (36) + (12). (36)
𝟐𝟓𝟐
𝛍= =𝟕
𝟑𝟔

Variance, V = ∑i(xi − μ)2 . p(xi )

1 2 1
V = (2 − 7)2 . 36 + (3 − 7)2 . 36 + ⋯ + (12 − 7)2 . 36

𝟐𝟏𝟎 𝟑𝟓
𝐕= =
𝟑𝟔 𝟔

Std. deviation , σ = √V

𝟑𝟓
𝛔 = √ 𝟔 = 𝟐. 𝟒𝟐

5
3. Show that the following distribution represents a discrete probability distribution. Find
the mean and variance

𝐱 10 20 30 40

p(x) 1 3 3 1
8 8 8 8

Sol: We observe that p(x) > 0 and ∑ p(x) = 1.

Mean μ = ∑ x . p(x)

1 3 3 1
μ = (10). + (20). + (30). + (40).
8 8 8 8

10+60+90+40 200
μ= =
8 8

𝛍 = 𝟐𝟓
Variance, V = ∑(x − μ)2 . p(x)

1 3 3 1
V = (10 − 25)2 . 8 + (20 − 25)2 . 8 + (30 − 25)2 . 8 + (40 − 25)2 . 8

𝟔𝟎𝟎
𝐕= = 𝟕𝟓
𝟖

4. Find the value of k such that the following distribution represents a finite probability
distribution. Hence find its mean and standard deviation. Also find p(x ≤ 1) , p(x > 1) and
p(−1 < x ≤ 2)
𝐱 -3 -2 -1 0 1 2 3

p(x) k 2k 3k 4k 3k 2k k

Sol: We must have p(x) ≥ 0 for all x and ∑ p(x) = 1. The first condition is satisfied if
k ≥ 0 . Since ∑ p(x) = 1

k + 2k + 3k + 4k + 3k + 2k + k = 1

16k = 1

𝟏
∴ 𝐤 = 𝟏𝟔

6
The discrete probability distribution is as follows

𝐱 -3 -2 -1 0 1 2 3

p(x) 1 2 3 4 3 2 1
16 16 16 16 16 16 16

Mean μ = ∑ x . p(x)

1 2 3 4 3 2 1
μ = (−3). 16 + (−2). 16 + (−1). 16 + (0). 16 + (1). 16 + (2). 16 + (3). 16

𝛍=𝟎
Variance, V = ∑(x − μ)2 . p(x)
1 2 3 4 3
V = (−3 − 0)2 . 16 + (−2 − 0)2 . 16 + (−1 − 0)2 . 16 + (0 − 0)2 . 16 + (1 − 0)2 . 16 +
2 1
(2 − 0)2 . 16 + (3 − 0)2 . 16
𝟏 𝟒𝟎 𝟓
𝐕 = 𝟏𝟔 (𝟗 + 𝟖 + 𝟑 + 𝟎 + 𝟑 + 𝟖 + 𝟗) = 𝟏𝟔 = 𝟐

Standard deviation σ = √V

𝟓
𝛔 = √𝟐 = 𝟏. 𝟓𝟖
Now, p(x ≤ 1) = p(−3) + p(−2) + p(−1) + p(0) + p(1)

1 2 3 4 3
= 16 + 16 + 16 + 16 + 16

𝟏𝟑
𝐩(𝐱 ≤ 𝟏) = 𝟏𝟔

p(x > 1) = p(2) + p(3) p(−1 < x ≤ 2) = p(0) + p(1) + p(2)

2 1 4 3 2
= 16 + 16 = 16 + 16 + 16
𝟑 𝟗
𝐩(𝐱 > 𝟏) = 𝟏𝟔 𝐩(−𝟏 < 𝐱 ≤ 𝟐) = 𝟏𝟔

5. A random variable X has the following probability function for various values of x.

𝐱 0 1 2 3 4 5 6 7

P(x) 0 k 2k 2k 3k k2 2k 2 7k 2 + k

i) Find k ii) Evaluate P(x < 6), P(x ≥ 6) and P(3 < x ≤ 6). Also find the probability
distribution and the distribution function of X.

7
Sol: We must have P(x) ≥ 0 and ∑ P(x) = 1.
The first condition is satisfied for k ≥ 0 and we have to find k such that
∑ P(x) = 1
i.e, 0 + k + 2k + 2k + 3k + k 2 + 2k 2 +(7k 2 + k) = 1

10k 2 + 9k − 1 = 0

(10k − 1)(k + 1) = 0

𝟏
𝐤 = 𝟏𝟎 and 𝐤 = −𝟏 ( Neglect)

Hence we have the following table

𝐱 0 1 2 3 4 5 6 7

P(x) 0 1 1 1 3 1 1 17
10 5 5 10 100 50 100

Now, P(x < 6) = P(0) + P(1) + P(2) + P(3) + P(4) + P(5)

1 1 1 3 1 81
= 0 + 10 + 5 + 5 + 10 + 100 = 100

𝐏(𝐱 < 𝟔) = 𝟎. 𝟖𝟏

P(x ≥ 6) = P(6) + P(7)

1 17 19
P(x ≥ 6) = 50 + 100 = 100

𝐏(𝐱 ≥ 𝟔) = 𝟎. 𝟏𝟗
P(3 < x ≤ 6) = P(4) + P(5) + P(6)

3 1 1 33
= 10 + 100 + 50 = 100

𝐏(𝟑 < 𝐱 ≤ 𝟔) = 𝟎. 𝟑𝟑
The probability distribution is as follows

𝐱 0 1 2 3 4 5 6 7

P(x) 0 0.1 0.2 0.2 0.3 0.01 0.02 0.17

The distribution function of X is f(x) = P(X ≤ x) = ∑ni=1 p(xi ) is called cumulative

distribution function and same is as follows

8
 𝐱 0 1 2 3 4 5 6 7

f(x) 0 0+0.1 0.1+0.2 0.3+0.2 0.5+0.3 0.8+0.01 0.81+0.02=0.83 0.83+0.17=1

=0.1 =0.3 =0.5 =0.8 =0.81

6. A random variable X take the values −3, −2, −1,0,1,2,3 such hat P(X = 0) = P(X < 0)
and P(X = −3) = P(X = −2) = P(X = −1) = P(X = 1) = P(X = 2) = P(X = 3) .Find the
probability distribution.
Sol: Let the distribution be as follows

𝐗 -3 -2 -1 0 1 2 3

P(X) p1 p2 p3 p4 p5 p6 p7

By data P(X = 0) = P(X < 0)

⇒ P(X = 0) = P(X = −1) + P(X = −2) + P(X = −3)
i.e., p4 = p3 + p2 + p1 --- (1)
Also we have
P(X = −3) = P(X = −2) = P(X = −1) = P(X = 1) = P(X = 2) = P(X = 3)
i.e., p1 = p2 = p3 = p5 = p6 = p7 --- (2)
Further we must have
p1 + p2 + p3 + p4 + p5 + p6 + p7 = 1 --- (3)

Using (2) in (1), we get p4 = p1 + p1 + p1

𝐩𝟒 = 𝟑𝐩𝟏
Using (2) in (3), we get p1 + p1 + p1 + p4 + p1 + p1 + p1 = 1
6p1 + p4 = 1
6p1 + 3p1 = 1 (since p4 = 3p1 )
9p1 = 1
𝟏
𝐩𝟏 = 𝟗
𝟑
Hence 𝐩𝟒 = 𝟗
1
Thus , p2 = p3 = p5 = p6 = p7 = 9
Thus the probability distribution is as follows
𝐗 -3 -2 -1 0 1 2 3

P(X) 1 1 1 3 1 1 1
9 9 9 9 9 9 9

9
7. The p.d.f. of variate X is given by the following table.

𝐱 0 1 2 3 4 5 6

P(x) K 3K 5K 7K 9K 11K 13K

For what value of K, this represents a valid probability distribution? Also find P(x ≥ 5) and
P(3 < x ≤ 6).
Sol: The probability distribution is valid if P(x) ≥ 0 and ∑ P(x) = 1
Hence we must have K ≥ 0 and K + 3K + 5K + 7K + 9K + 11K + 13K = 1
𝟏
49K = 1𝐊 =
𝟒𝟗

Also, P(x ≥ 5) = P(5) + P(6)

P(x ≥ 5) = 11K + 13K = 24K
𝟐𝟒
𝐏(𝐱 ≥ 𝟓) =
𝟒𝟗
Now, P(3 < x ≤ 6) = P(4) + P(5) + P(6)
P(3 < x ≤ 6) = 9K + 11K + 13K
P(3 < x ≤ 6) = 33K
𝟑𝟑
𝐏(𝟑 < 𝐱 ≤ 𝟔) =
𝟒𝟗

Binomial Distribution
If 𝐩 is the probability of success and 𝐪 is the probability of failure, the probability of 𝐱
success out of 𝐧 trials is given by 𝐏(𝐱) = 𝐧𝐂𝐱 𝐩𝐱 𝐪𝐧−𝐱
Mean and Standard deviation of the Binomial Distribution
Mean, 𝛍 = 𝐧𝐩
Variance, 𝐕 = 𝐧𝐩𝐪

Std. Deviation, 𝛔 = √𝐧𝐩𝐪

Problems
4
1. Find the binomial probability which has mean 2 and variance 3.
4
Sol: Given Mean, μ = 2 and Variance, V = 3

For binomial distribution, μ = np

10
 np = 2
npq = V
4
2q =
3
𝟐
𝐪=
𝟑
2
Since p = 1 − q = 1 − 3

𝟏
𝐩=
𝟑
Since np = 2
n
=2
3

𝐧=𝟔
The binomial probability function is 𝐏(𝐱) = 𝐧𝐂𝐱 𝐩𝐱 𝐪𝐧−𝐱 becomes

𝟏 𝐱 𝟐 𝟔−𝐱
𝐏(𝐱) = 𝟔𝐂𝐱 ( ) ( )
𝟑 𝟑
The distribution of probability is as follows

𝐱 0 1 2 3 4 5 6

P(x) 𝟐 𝟔 𝟔 𝟏 𝟏 𝟐 𝟓 𝟔 𝟏 𝟐 𝟐 𝟒 𝟔 𝟏 𝟑 𝟐 𝟑 𝟔 𝟏 𝟒 𝟐 𝟐 𝟔 𝟏 𝟓 𝟐 𝟏 𝟏 𝟔
( ) 𝐂𝟏 ( ) ( ) 𝐂𝟐 ( ) ( ) 𝐂𝟑 ( ) ( ) 𝐂𝟒 ( ) ( ) 𝐂𝟓 ( ) ( ) ( )
𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑 𝟑

2.When a coin is tossed 4 times, find the probability of getting i) exactly one head
ii) atmost 3 heads iii) atleast two heads.
𝟏 𝟏
Sol: Given 𝐧 = 𝟒 , 𝐩 = 𝐏(𝐇) = 𝟐 , 𝐪 = 𝐏(𝐓) = 𝟐 (since p + q = 1)

From binomial distribution ,the probability of x successes out of n trials is given by

𝐏(𝐱) = 𝐧𝐂𝐱 𝐩𝐱 𝐪𝐧−𝐱
1 x 1 4−x
P(x) = 4Cx (2) (2)

1 4
P(x) = 4Cx (2)
𝟏
𝐏(𝐱) = 𝟏𝟔 . 𝟒𝐂𝐱

11
 1 1
i) P(1 head) = P(x = 1) = . 4C1 = (4)
16 16

𝟏
𝐏(𝐱 = 𝟏) = 𝟒

𝐏(𝐱 = 𝟏) = 𝟎. 𝟐𝟓
ii) P(atmost 3 heads) = P(x ≤ 3)
P(x ≤ 3) = P(x = 0) + P(x = 1) + P(x = 2) + P(x = 3)
1 1 1 1
P(x ≤ 3) = 16 . 4C0 + 16 . 4C1 + 16 . 4C2 + 16 . 4C3
1
P(x ≤ 3) = 16 [1 + 4 + 6 + 4]
15
P(x ≤ 3) = 16

𝐏(𝐱 ≤ 𝟑) = 𝟎. 𝟗𝟑𝟕𝟓
iii) P(atleast 2 heads) = P(x ≥ 2)
P(x ≥ 2) = 1 − P(x < 3)
P(x ≥ 2) = 1 − [P(x = 0) + P(x = 1)]
1 4 1
P(x ≥ 2) = 1 − [ . C0 + . 4C1 ]
16 16
1
P(x ≥ 2) = 1 − 16 [1 + 4]
5
P(x ≥ 2) = 1 − 16
11
P(x ≥ 2) = 16

𝐏(𝐱 ≥ 𝟐) = 𝟎. 𝟔𝟖𝟕𝟓
3. In a consignment of electric lamps 5% are defective. If a random sample of 8 lamps are
inspected. What is the probability that one or more lamps are defective?
𝟓
Sol: Given n = 8, Probability of defective lamp =𝐩 = 𝟓% = 𝟏𝟎𝟎 = 𝟎. 𝟎𝟓

∴ q = 1 − p = 1 − 0.05
𝐪 = 𝟎. 𝟗𝟓
From binomial distribution,
P(x) = ncx px qn−x
P(x) = 8cx (0.05)x (0.95)8−x
Probability that one or more lamps are defective
P(x ≥ 1) = 1 − P(x < 1)
P(x ≥ 1) = 1 − P(x = 0)

12
 P(x ≥ 1) = 1 − 8c0 (0.05)0 (0.95)8−0
𝐏(𝐱 ≥ 𝟏) = 𝟎. 𝟑𝟑𝟔𝟔.
Thus the required probability is 0.3366
4. The probability that a person aged 60 years will live upto 70 is 0.65. What is the
probability that out of 10 person aged 60 atleast 7 of them will live upto 70.
Sol: Given n = 10, 𝐩 = 𝟎. 𝟔𝟓
∴ q = 1 − p = 1 − 0.65
𝐪 = 𝟎. 𝟑𝟓
Let x denote the number of person aged 60 years living upto 70 years.
From binomial distribution,
P(x) = ncx px qn−x
P(x) = 10cx (0.65)x (0.35)10−x
Probability that out of 10 person aged 60 atleast 7 of them will live upto 70
P(x ≥ 7) = P(7) + P(8) + P(9) + P(10)
P(x ≥ 7) = 10c7 (0.65)7 (0.35)3 + 10c8 (0.65)8 (0.35)2 + 10c9 (0.65)9 (0.35)1 +
10c10 (0.65)10 (0.35)0
P(𝐱 ≥ 𝟕) = 𝟎. 𝟓𝟏𝟑𝟖
5.The number of telephone lines busy at an instant of time is a binomial variate with
probability 0.1 that a line is busy. If 10 such lines are chosen at random, what is the
probability that i) no line is busy ii) all lines are busy iii) atleast one line is busy
iv) atmost 2 lines are busy.
Sol: Given n = 10, 𝐩 = 𝟎. 𝟏
∴ q = 1 − p = 1 − 0.1
𝐪 = 𝟎. 𝟗
Let x denote the number of telephone lines busy.
From binomial distribution,
P(x) = ncx px qn−x
𝐏(𝐱) = 𝟏𝟎𝐜𝐱 (𝟎. 𝟏)𝐱 (𝟎. 𝟗)𝟏𝟎−𝐱
i) Probability that no line is busy
P(x = 0) = 10c0 (0.1)0 (0.9)10−0 = (0.9)10 = 𝟎. 𝟑𝟒𝟖𝟕
ii) Probability that all lines are busy
P(x = 10) = 10c10 (0.1)10 (0.9)0 = (𝟎. 𝟏)𝟏𝟎

13
iii) Probability that atleast one line is busy
P(x ≥ 1) = 1 − P(x < 1)
P(x ≥ 1) = 1 − P(x = 0)
P(x ≥ 1) = 1 − 10c0 (0.1)0 (0.9)10
P(x ≥ 1) = 1 − 0.3487
P(𝐱 ≥ 𝟏) = 𝟎. 𝟔𝟓𝟏𝟑
iv) Probability that atmost 2 lines are busy
P(x ≤ 2) = P(x = 0) + P(x = 1) + P(x = 2)
P(x ≤ 2) = (0.9)10 + 10c1 (0.1)1 (0.9)2 + 10c2 (0.1)2 (0.9)8
P(𝐱 ≤ 𝟐) = 𝟎. 𝟗𝟐𝟗𝟖
6. In a quiz contest of answering ‘Yes’ or ‘No’ what is the probability of guessing atleast 6
answers correctly out of 10 questions asked ? Also find the probability of the same if
there are 4 options for a correct answer.
Sol: Given n = 10
Let x denote the correct answer .
𝟏
i) 𝐩 = 𝟐 = 𝟎. 𝟓

∴ q = 1 − p = 1 − 0.5
𝐪 = 𝟎. 𝟓
From binomial distribution,
P(x) = ncx px qn−x
P(x) = 10cx (0.5)x (0.5)10−x
𝐏(𝐱) = 𝟏𝟎𝐜𝐱 (𝟎. 𝟓)𝟏𝟎
Probability of guessing atleast 6 answers correctly out of 10 questions
P(x ≥ 6) = P(x = 6)+P(x = 7) + P(x = 8) + P(x = 9) + P(x = 10)

P(x ≥ 6) = 10c6 (0.5)10 + 10c7 (0.5)10 + 10c8 (0.5)10 + 10c9 (0.5)10 + 10c10 (0.5)10
P(x ≥ 6) = (0.5)10 [10c6 + 10c7 + 10c8 + 10c9 + 10c10 ]
P(x ≥ 6) = (0.5)10 [210 + 120 + 45 + 10 + 1]
P(𝐱 ≥ 𝟔) = 𝟎. 𝟑𝟕𝟕𝟎
𝟏
ii) 𝐩 = 𝟒
1
∴q =1−p= 1−4
𝟑
𝐪=𝟒

14
From binomial distribution,
P(x) = ncx px qn−x

1 x 3 10−x
P(x) = 10cx ( ) ( )
4 4

310−x
P(x) = 10cx x+10−x
4
310−x
P(x) = 10cx
410
𝟏
𝐏(𝐱) = [𝟑𝟏𝟎−𝐱 . 𝟏𝟎𝐜𝐱 ]
𝟒𝟏𝟎
Probability of guessing atleast 6 answers correctly out of 10 questions
P(x ≥ 6) = P(x = 6)+P(x = 7) + P(x = 8) + P(x = 9) + P(x = 10)

1
P(x ≥ 6) = 410 [34 . 10c6 + 33 . 10c7 + 32 . 10c8 + 3.10c9 + 10c10 ]
1
P(x ≥ 6) = 410 [81.210 + 27.120 + 9.45 + 3.10 + 1]

P(𝐱 ≥ 𝟔) = 𝟎. 𝟎𝟏𝟗𝟕
7. In sampling a large number of parts manufactured by a company, the mean number of
defectives in samples of 20 is 2. Out of 1000 such samples how many would be expected
to contain atleast 3 defective parts.
Sol: Given Mean, μ = 2 and n = 20
np = 2
20p = 2
𝟏
𝐩 = 𝟏𝟎 = 𝟎. 𝟏

q= 1−p
1 𝟗
q = 1 − 10 𝐪 = 𝟏𝟎 = 𝟎. 𝟗

Let x denote the defective part.

From binomial distribution,
P(x) = ncx px qn−x
𝐏(𝐱) = 𝟐𝟎𝐜𝐱 (𝟎. 𝟏)𝐱 (𝟎. 𝟗)𝟐𝟎−𝐱
Probability of atleast 3 defective parts
P(x ≥ 3) = 1 − P(x < 3)

15
 P(x ≥ 3) = 1 − [P(0) + P(1) + P(2)]
P(x ≥ 3) = 1 − [20c0 (0.1)0 (0.9)20 + 20c1 (0.1)1 (0.9)19 + 20c2 (0.1)2 (0.9)18 ]
P(x ≥ 3) = 1 − 0.6769
P(𝐱 ≥ 𝟑) = 𝟎. 𝟑𝟐𝟑𝟏
Thus, the number of defectives in 1000 such samples= 1000*0.3231=323.1≅323
8. If the mean and standard deviation of the number of correctly answered questions in a
test given to 4096 students are 2.5 and √1.875 . Find an estimate the number of
candidates answering correctly i) 5 questions ii) 2 or less iii) 8 or more.

Sol: Given Mean = 2.5 σ = √1.875

Wkt, μ = np and σ = √V
Thus, np = 2.5 and V = npq = 1.875
npq 1.875
Now, =
np 2.5

We get q= 0.75, and p = 1-q = 0.25

n=10
Let x denote the number of correctly answered questions
P(x) = ncx px qn−x = 10cx (0.25)x (0.75)10−x
i) Probability that 5 questions are answered correctly
P(x = 5) = 10c5 (0.25)5 (0.75)10−5
𝐏(𝐱 = 𝟓) = 𝟎. 𝟎𝟓𝟖𝟒
Estimation for 4096 students is =4096*0.0584 ≅ 𝟐𝟑𝟗
Number of students correctly answering 5 questions is 239
ii) Probability that 2 or less questions are answered correctly
P(x ≤ 2) = P(x = 0) + P(x = 1) + P(x = 2)
= 10c0 (0.25)0 (0.75)10 + 10c1 (0.25)1 (0.75)9 + 10c2 (0.25)2 (0.75)8
𝐏(𝐱 ≤ 𝟐) = 𝟎. 𝟓𝟐𝟓𝟔
Estimation for 4096 students is =4096*0.5256 ≅ 𝟐𝟏𝟓𝟑
Number of students correctly answering 2 or less than 2 questions is 2153
iii) Probability that 8 or more questions are answered correctly

P(x ≥ 8) = P(x = 8) + P(x = 9) + P(x = 10)

= 10c8 (0.25)8 (0.75)2 + 10c9 (0.25)9 (0.75)1 + 10c10 (0.25)10 (0.75)0
𝐏(𝐱 ≥ 𝟖) = 𝟎. 𝟎𝟎𝟎𝟓

16
Estimation for 4096 students is =4096*0.0005 ≅2
Number of students correctly answering 8 or more than 8 questions is 2
9. In 800 families with 5 children each how many families would be expected to have
i) 3 boys ii)5 girls iii) either 2 or 3 boys iv) atmost 2 girls by assuming probabilities for
boys and girls to be equal.
10. An airline knows that 5% of the people making reservations on a certain flight will not
turn up. Consequently, their policy is to sell 52 tickets for a flight that can only hold 50
people. What is the probability that there will be a seat for every passenger who turns up?
11. 4 coins are tossed 100 times and the following results were obtained. Fit a binomial
distribution for the data and calculate the theoretical frequencies.

Number of 0 1 2 3 4
heads

Frequency 5 29 36 25 5

Sol: Let x denote the number of heads and f the corresponding frequency. Since the data is in
the form of a frequency distribution we shall first calculate the mean.
∑ fx (0)(5)+(1)(29)+(2)(36)+(3)(25)+(4)(5) 196
Mean, μ = ∑f
= = 100 = 1.96
5+29+36+25+5

From binomial distribution, μ = np. Here n=4

Hence 4p = 1.96
𝐩 = 𝟎. 𝟒𝟗
𝐪 = 𝟎. 𝟓𝟏
From binomial distribution,
P(x) = ncx px qn−x 𝐏(𝐱) = 𝟒𝐜𝐱 (𝟎. 𝟒𝟗)𝐱 (𝟎. 𝟓𝟏)𝟒−𝐱
Since 4 coins were tossed , expected(theoretical) frequencies are obtained from
F(x) = 100P(x)
F(x) = 100.4cx (0.49)x (0.51)4−x where x = 0,1,2,3,4.
F(0) = 100.4c0 (0.49)0 (0.51)4 = 6.765 ≅7
F(1) = 100.4c1 (0.49)1 (0.51)3 = 25.999 ≅ 26
F(2) = 100.4c2 (0.49)2 (0.51)2 = 37.47 ≅ 37
F(3) = 100.4c3 (0.49)3 (0.51)1 = 24.0004 ≅ 24
F(4) = 100.4c4 (0.49)4 (0.51)0 = 5.765 ≅ 6
Thus the required theoretical frequencies are 7,26,37,24,6.

17
Poisson Distribution
Poisson distribution is regarded as the limiting form of the binomial distribution when 𝐧 is
very large (n → ∞) and 𝐩 the probability of success is very small (p → 0) so that 𝐧𝐩 tends
to a fixed finite constant say 𝐦.
Let m be a positive constant. Consider the infinite discrete probability distribution of a
𝐦𝐱
random variable x is given by P(𝐱) = 𝐞−𝐦
𝐱!

The above probability density function is called the poisson probability function.
Mean and Standard deviation of the Poisson Distribution
Mean, 𝛍 = 𝐦
Variance, 𝐕 = 𝐦

Std. Deviation, 𝛔 = √𝐦

Problems
1. The probabilities of a poisson variate taking the values 3 and 4 are equal. Find the PDF.
Also, calculate the probabilities of the variate taking the values 0 and 1.
Sol: Given P(x = 3) = P(x = 4)
m3 −m m4 −m
e = e
3! 4!
4! m4
=
3! m3
∴ 𝐦 = 𝟒.
From poisson distribution,
mx −m 4x −4
P(x) = e = e
x! x!
40 −4 41 −4
P(x = 0) = e = 𝟎. 𝟎𝟏𝟖𝟑 P(x = 1) = e = 𝟎. 𝟎𝟕𝟑𝟑
0! 1!
1
2. In a certain factory turning out razor blades there is a small probability of for any
500
blade to be defective. The blades are supplied in packets of 10. Use poisson distribution
to calculate the approximate number of packets containing i) no defective ii) one
defective iii) two defective blades in a consignment of 10,000 packets.
1
Sol: Probability of a defective blade= p = 500 = 0.002

In a packet of 10, the mean number of defective blades is

Mean, m = np
m = 10(0.002)

18
 𝐦 = 𝟎. 𝟎𝟐
From poisson distribution,
mx −m
P(x) = e
x!
(0.02)x −(0.02)
P(x) = e
x!
(0.02)x
P(x) = (0.9802) ( Since e−0.02 = 0.9802)
x!

i) Probability of no defective
(0.02)0
P(x = 0) = (0.9802) = 0.9802
0!

For consignment of 10,000 packets= 𝟏𝟎, 𝟎𝟎𝟎 ∗ 𝟎. 𝟗𝟖𝟎𝟐 = 𝟗𝟖𝟎𝟐

ii) Probability of one defective
(0.02)1
P(x = 1) = (0.9802) = 0.0196
1!

For consignment of 10,000 packets= 𝟏𝟎, 𝟎𝟎𝟎 ∗ 𝟎. 𝟎𝟏𝟗𝟔 = 𝟏𝟗𝟔

iii) Probability of one defective
(0.02)2
P(x = 2) = (0.9802) = 0.0002
2!

For consignment of 10,000 packets= 𝟏𝟎, 𝟎𝟎𝟎 ∗ 𝟎. 𝟎𝟎𝟎𝟐 =2

3. The number of accidents in a year to taxi drivers in a city follows a Poisson distribution
with mean 3. Out of 1000 taxi drivers find approximately the number of the drivers with
i) no accident in a year ii) more than 3 accidents in a year.
Sol: Given Mean=3
Mean, μ = m
𝐦=𝟑
From poisson distribution,
mx −m
P(x) = e
x!
(3)x −3
P(x) = e
x!
(3)x
P(x) = (0.05) ( Since e−3 = 0.05)
x!

i) Probability of number of drivers with no accident in a year

(3)0
P(x = 0) = (0.05) = 0.05
0!
Number of drivers out of 1𝟎𝟎𝟎 with no accident in a year = 𝟏, 𝟎𝟎𝟎 ∗ 𝟎. 𝟎𝟓 = 𝟓𝟎

19
ii) Probability of more than 3 accidents in a year
P(x > 3) = 1 − P(x ≤ 3)
P(x > 3) = 1 − [P(0) + P(1) + P(2) + P(3)]
(3)0 (3)1 (3)2 (3)3
P(x > 3) = 1 − [(0.05) + (0.05) + (0.05) + (0.05) ]
0! 1! 2! 3!

P(x > 3) = 1 − (0.05)[1 + 3 + 4.5 + 4.5]

P(x > 3) = 0.35
Number of drivers out of 1𝟎𝟎𝟎 with more than 3 accidents in a year = 𝟏, 𝟎𝟎𝟎 ∗ 𝟎. 𝟑𝟓 =
𝟑𝟓𝟎
4. Given that 2% of the fuses manufactured by a firm are defective, find, by using poisson
distribution, the probability that a box containing 200 fuses has i) no defective fuse
ii) at least one defective fuse iii) exactly 3 defective fuses iv) 3 or more defective fuses.
Sol: The probability that a fuses is defective p = 2/100=0.02.
And n=200
Mean, μ = np = 200 ∗ 0.02 ; 𝐦 = 𝟒
mx −m
From poisson probability function P(x) = e
x!

4x −4
P(x) = e
x!
Let x denotes the number of defective fuses in a box containing 200 fuses
i) Probability that the box contains no defective fuse
40 −4
P(x = 0) = e = 𝟎. 𝟎𝟏𝟖𝟑
0!

ii) Probability that the box contains at least one defective fuse is
P(x ≥ 1) = 1 − P(x < 1)
= 1 − P(x = 0)
= 1 − 0.013
𝐏(𝐱 ≥ 𝟏) = 𝟎. 𝟗𝟖𝟏𝟕
iii) Probability that the box contains exactly 3 defective fuse
43 −4
P(x = 3) = e
3!

𝐏(𝐱 = 𝟑) = 𝟎. 𝟏𝟗𝟓𝟐
iv) Probability that the box contains 3 or more defective fuses
P(x ≥ 3) = 1 − P(x < 3)
40 −4 41 −4 42 −4
P(x ≥ 3) = 1 − [ e + e + e ]
0! 1! 2!

20
 𝐏(𝐱 ≥ 𝟑) = 𝟎. 𝟕𝟔𝟐𝟏
5. A certain screw making machine has a chance of producing 2 defectives out of 1000. The
screws are packed in boxes of 100. Using poisson distribution , find the approximate
number of boxes containing i) no defective screw ii)one defective screw iii) two defective
screw, in a consignment of 5000 boxes.
Sol: The probability that a screw is defective is given as p=2/1000 = 0.002.
Here n=100
Mean = np=100X0.002=0.2 ; 𝐦 = 𝟎. 𝟐
mx −m
From poisson probability function P(x) = e
x!

(0.2)x −0.2
P(x) = e
x!
Let x denotes the number of defective fuses in a box containing 100 screws
i) Probability that the box contains no defective screw
𝟎.𝟐𝟎
𝐏(𝐱 = 𝟎) = 𝐞−𝟎.𝟐 = 𝟎. 𝟖𝟏𝟖𝟕
𝟎!

Number of boxes containing no defective screw in 5000 boxes =5000*0.8187=4094

ii) Probability that the box contains one defective screw is

(0.2)1 −0.2
P(x = 1) = e
1!

= 0.2X0.8187

𝐏(𝐱 = 𝟏) = 𝟎. 𝟏𝟔𝟑𝟕
Number of boxes containing one defective screw in 5000 boxes =5000*0.1637=819
iii) Probability that the box contains two defective screw is
(0.2)2 −0.2
P(x = 2) = e
2!

= 0.02X0.8187

𝐏(𝐱 = 𝟐) = 𝟎. 𝟎𝟏𝟔𝟑
Number of boxes containing two defective screw in 5000 boxes =5000*0.0163=82
6. A communication channel receives independent pulses at the rate of 12 pulses per micro
second. The probability of transmission error is 0.001 for each micro second. Compute
the probabilities of i) no error during a micro second ii) one error per micro second
iii) atleast one per micro second iv) two errors v) atmost two errors.
Sol: Given n = 12 and p = 0.001
Mean number of errors in one micro second

21
 μ = np
m = 12 ∗ 0.001 (Since μ = m, from poisson distribution)
𝐦 = 𝟎. 𝟎𝟏𝟐
From Poisson distribution
mx −m
P(x) = .e
x!
(0.012)x −(0.012)
P(x) = .e
x!
(0.012)x
P(x) = (0.9880) (Sincee−(0.012) = 0.9880)
x!

i) Probability of no error during a micro second

(0.012)0
P(x = 0) = (0.9880) 0!

𝐏(𝐱 = 𝟎) = 𝟎. 𝟗𝟖𝟖𝟎
ii) Probability of one error during a micro second
(0.012)1
P(x = 1) = (0.9880) 1!

𝐏(𝐱 = 𝟏) = 𝟎. 𝟎𝟏𝟏𝟗
iii) Probability of atleast one error during a micro second
P(x ≥ 1) = 1 − P(x < 1)
P(x ≥ 1) = 1 − P(x = 0)
P(x ≥ 1) = 1 − P(x = 0)
P(x ≥ 1) = 1 − 0.9880
𝐏(𝐱 ≥ 𝟏) = 𝟎. 𝟎𝟏𝟐
iv) Probability of two error during a micro second
(0.012)2
P(x = 2) = (0.9880) 2!

𝐏(𝐱 = 𝟐) = 𝟎. 𝟎0007
v) Probability of atmost two error during a micro second
𝐏(x ≤ 2) = P(x = 0) + P(x = 1) + P(x = 2)
P(x ≤ 2) = 0.9880 + 0.0119 + 0.00007
𝐏(𝐱 ≤ 𝟐) = 𝟎.99997

22
Continuous Probability Distribution
If for every x belonging to the range of a continuous random variable X, we assign a real
number f(x) satisfying the conditions,
∞
i) f(x) ≥ 0 ii) ∫−∞ f(x)dx = 1 , then f(x) is called a Continuous probability function or
probability density function ( p.d.f )

Cumulative distribution function

If X is a continuous random variable with probability density function f(x) then the
x
function F(x) defined by F(x) = P(X ≤ x) = ∫−∞ f(x)dx is called the Cumulative distribution
function (c.d.f) of X.
NOTE: If 𝐫 is any real number, then
∞
1. P(x ≥ r) = ∫r f(x)dx
∞

P(x < r) = 1 − P(x ≥ r) = 1 − ∫ f(x)dx

r

Mean and Variance

If X is a continuous random variable with probability density function f(x) where
−∞ < x < ∞ , the mean (μ) or expectation E(X) and the variance (σ2 ) of X is defined as
follows.
∞
Mean, 𝛍 = ∫−∞ 𝐱. 𝐟(𝐱)𝐝𝐱
∞ ∞
Variance, 𝛔𝟐 = ∫−∞(𝐱 − 𝛍)𝟐 . 𝐟(𝐱) 𝐝𝐱 or 𝐕 = ∫−∞ 𝐱 𝟐 . 𝐟(𝐱) 𝐝𝐱 − 𝛍𝟐

Problems
1. Find which of the following is a probability density function
2x, 0 < x < 1 |x|, 0 < |x| ≤ 1
a) f1 (x) = { b) f2 (x) = {
0, otherwise 0, otherwise
2x , 0 < x ≤ 1
c) f3 (x) = { 4 − 4x , 1 < x < 2
0 , otherwise
∞
Sol: Conditions for p.d.f are f(x) ≥ 0 and ∫−∞ f(x)dx = 1

a) Clearly f1 (x) ≥ 0
1
∞ 1 1 x2
∫−∞ f1 (x)dx = ∫0 f1 (x) dx = ∫0 2x dx = 2. [ 2 ] = [1 − 0] = 1
0

∴ f1 (x) is a p.d.f
b) Evidently f2 (x) = |x| ≥ 0

23
 ∞ 1 1
∫−∞ f2 (x)dx = ∫−1 f2 (x) dx = ∫−1|x| dx
−x, − 1 < x < 0
But,|x| = {
+x, 0 < x < 1
∞ 0 1
∴ ∫−∞ f2 (x)dx = ∫−1(−x) dx + ∫0 x dx
∞ 0 1
x2 x2
∫ f2 (x)dx = − [ ] + [ ]
2 −1 2 0
−∞
∞
1 1 1 1
∫ f2 (x)dx = − [0 − ] + [ − 0] = + = 1
2 2 2 2
−∞

∴ f2 (x) is a p.d.f
c) Given f3 (x) = 2x > 0 in 0 < x ≤ 1
But f3 (x) = 4 − 4x is negative in 1 < x < 2
The first condition is not satisfied.
∴ f3 (x) is not a p.d.f
x
+ c, 0 ≤ x ≤ 3
2. Find the value of c such that f(x) = { 6 is a p.d.f . Also find
0, elsewhere
P(1 ≤ x ≤ 2).
∞
Sol: Here f(x) ≥ 0 if c ≥ 0 ; Also we must have ∫−∞ f(x)dx = 1
3 x
i.e., ∫0 (6 + c) dx = 1
3
x2
i.e., [12 + cx] = 1
0
1
i.e., 12 (9 − 0) + c(3 − 0) = 1
9
i.e., 12 + 3c = 1
9
i.e., 3c = 1 − 12

3
3c =
12
𝟏
∴ 𝐜 = 𝟏𝟐
2
Now, P(1 ≤ x ≤ 2) = ∫1 f(x)dx
2 x 1
= ∫1 (6 + 12) dx

24
 2
x2 x
= [12 + 12]
1
1 1
= 12 (4 − 1) + 12 (2 − 1)
3 1 4
= 12 + 12 = 12
𝟏
𝐏(𝟏 ≤ 𝐱 ≤ 𝟐) = 𝟑

kx 2 , 0 < x < 3
3. Find the constant k such that f(x) = { is a p.d.f . Also compute
0, otherwise
i) P(1 < x < 2) ii) P(x ≤ 1) iii) P(x > 1) iv) Mean v) Variance
∞
Sol: : Here f(x) ≥ 0 if k ≥ 0 ; Also we must have ∫−∞ f(x)dx = 1
3
i.e., ∫0 kx 2 dx = 1
3
kx3
i.e., [ ] =1
3 0
k
i.e., 3 (27 − 0) = 1

i.e., 9k = 1
𝟏
i.e., 𝐤 = 𝟗

x2
, 0<x<3
∴ f(x) = { 9
0, otherwise
2
i) P(1 < x < 2) = ∫1 f(x)dx
2 x2
= ∫1 dx
9
2
x3
= [27]
1
1
= 27 (8 − 1)
𝟕
𝐏(𝟏 < 𝐱 < 𝟐) = 𝟐𝟕
1
ii) P(x ≤ 1) = ∫0 f(x)dx
1
x2
P(x ≤ 1) = ∫ dx
9
0
1
x3
= [27]
0
1
= 27 (1 − 0)

25
 𝟏
𝐏(𝐱 ≤ 𝟐) =
𝟐𝟕
3
iii) P(x > 1) = ∫1 f(x)dx
3 x2
= ∫1 dx
9
3
x3
P(x > 1) = [27]
1
1
= 27 (27 − 1)
𝟐𝟔
𝐏(𝐱 > 𝟏) = 𝟐𝟕
∞
iv) Mean, μ = ∫−∞ x. f(x)dx
3 x2
= ∫0 x. dx
9

3 x3
= ∫0 dx
9
3
x4
μ = [36]
0
1
μ = 36 (81 − 0)
81
μ = 36
𝟗
𝛍=𝟒
∞
v)Variance, V = ∫−∞ x 2 . f(x) dx − μ2

3 x2 9 2
= ∫0 x 2 . dx − (4)
9

3 x4 81
= ∫0 dx − 16
9
3
x5 81
= [45] − 16
0
1 81
V = 45 [35 − 0] − 16
243 81
V= − 16
45
81
V = 240
𝟐𝟕
𝐕 = 𝟖𝟎

26
Exponential Distribution
The continuous probability distribution having the probability density function f(x) given by
αe−αx , for x > 0
f(x) = { , where α > 0 is known as the exponential distribution.
0, otherwise
∞
Evidently f(x) ≥ 0 and we have ∫−∞ f(x)dx = 1.

f(x) satisfy both the conditions required for a continuous probability function.
Mean and standard deviation of the Exponential Distribution
𝟏
Mean , 𝛍 = 𝛂
𝟏
Std. deviation , 𝛔 =
𝛂

Problems
1. If x is an exponential variate with mean 3 find i) P(x > 1) ii) P(x < 3)
Sol: The p.d.f of the exponential distribution is given by
αe−αx , for 0 < x < ∞
f(x) = {
0, otherwise
1
The mean of the distribution is given by α.
1
By data, μ = α = 3
𝟏
∴ 𝛂=𝟑

1 −x
e 3, for 0 < x < ∞
Hence , f(x) = {3
0, otherwise
i) P(x > 1) = 1 − P(x ≤ 1)
P(x > 1) = 1 − P(x ≤ 1)
1
=1− ∫0 f(x)dx
x
11
= 1 − ∫0 3 e−3 dx
x
1 1
= 1 − 3 ∫0 e−3 dx
x 1
−
1 e 3
= 1 − 3[ 1 ]
−
3 0
1
= 1 + [e−3 − e−0 ]
1
= 1 + e− 3 − 1

27
 1
= e− 3
𝐏(𝐱 > 𝟏) = 0.7165
3

ii) P(x < 3) = ∫ f(x)dx

0
x
31
= ∫0 3 e−3 dx
x
1 3
= 3 ∫0 e−3 dx
x 3
−
1 e 3
= 3[ 1 ]
−
3 0

= −[e−1 − e−0 ]
1
= −e−1 + 1 = 1 − e

𝐏(𝐱 < 𝟑) = 𝟎. 𝟔𝟑𝟐𝟏

2. The length of telephone conversation in a booth has been an exponential distribution and
found on an average to be 5 minutes. Find the probability that a random call made from
this booth i) ends less than 5 minutes ii) between 5 and 10 minutes.
Sol: We have, f(x) = αe−αx , 0 < x < ∞
1
By data, μ = α = 5
𝟏
∴ 𝛂=𝟓

1 −x
f(x) = e 5, 0 < x < ∞
5
5

i) P(x < 5) = ∫ f(x)dx

0

51 x
= ∫0 5 e−5 dx
x
1 5
= 5 ∫0 e−5 dx

x 5
−
1 e 5
P(x < 5) = [ ]
5 −1
5 0
= −[e−1 − e−0 ]
1
= −e−1 + 1 = 1 − e

28
 𝐏(𝐱 < 𝟓) = 𝟎. 𝟔𝟑𝟐𝟏
10

ii) P(5 < x < 10) = ∫ f(x)dx

5

10 1 −x
= ∫5 e 5 dx
5

10 x
1
= 5 ∫5 e−5 dx
x 10
−
1 e 5
P(5 < x < 10) = 5 [ 1 ]
−
5 5

= −[e−2 − e−1 ]
1 1
= −e−2 + e−1 = e − e2

𝐏(𝟓 < 𝐱 < 𝟏𝟎) = 𝟎. 𝟐𝟑𝟐𝟓

3. In a certain town the duration of a shower is exponentially distributed with mean 5
minutes. What is the probability that a shower will last for: (i) 10 minutes or more
(ii) less than 10 minutes (iii) between 10 and 12 minutes.
Sol: We have, f(x) = αe−αx , 0 < x < ∞
1
By data, μ = α = 5
𝟏
∴ 𝛂=𝟓

1 x
f(x) = e−5 , 0 < x < ∞
5
∞

i) P(x ≥ 10) = ∫ f(x)dx

10
x
∞1
= ∫10 5 e−5 dx
x
1 ∞
= 5 ∫10 e−5 dx
x ∞
−
1 e 5
= 5[ 1 ]
−
5 10

= −[e−∞ − e−2 ]
= −e−∞ + e−2 = −0 + e−2
𝐏(𝐱 ≥ 𝟏𝟎) = 𝟎. 𝟏𝟑𝟓𝟑

29
ii) P(x < 10) = 1 − P(x ≥ 10)
P(x < 10) = 1 − 0.1353
𝐏(𝐱 < 𝟏𝟎) = 𝟎. 𝟖𝟔𝟒𝟕
12

iii) P(10 < x < 12) = ∫ f(x)dx

10
x
12 1
= ∫10 5 e−5 dx
x
1 12
= 5 ∫10 e−5 dx
x 12
−
1 e 5
= 5[ 1 ]
−
5 10

= −[e−2.4 − e−2 ]
P(10 < x < 12) = e−2 − e−2.4
𝐏(𝟏𝟎 < 𝐱 < 𝟏𝟐) = 𝟎. 𝟎𝟒𝟒𝟔
4.The life of a compressor manufactured by a company is known to be 200 months on an
average following an exponential distribution. Find the probability that the life of a
compressor of that company is i) less than 200 months ii) between 100 months and 25
years.
Sol: We have, f(x) = αe−αx , 0 < x < ∞
1
By data, μ = α = 200
𝟏
∴ 𝛂 = 𝟐𝟎𝟎

1 − x
f(x) = e 200 , 0 < x < ∞
200
200

i) P(x < 200) = ∫ f(x)dx

0
x
200 1
= ∫0 e−200 dx
200

1 200 − x
= 200 ∫0 e 200 dx
x 200
−
1 e 200
= 200 [ 1 ]
−
200 0

= −[e−1 − e−0 ]

30
 1
= −e−1 + 1 = 1 −
e

𝐏(𝐱 < 𝟐𝟎𝟎) = 𝟎. 𝟔𝟑𝟐𝟏

Now, 25 years=25*12 months=300 months
300

ii) P(100 < x < 300) = ∫ f(x)dx

100

300 1 x
= ∫100 e−200 dx
200
x
1 300
= 200 ∫100 e−200 dx
x 300
−
1 e 200
= 200 [ 1 ]
−
200 100

= −[e−1.5 − e−2 ]
𝐏(𝟏𝟎𝟎 < 𝐱 < 𝟑𝟎𝟎) = 𝟎. 𝟑𝟖𝟑𝟒
k
5. Determine k if random variable x has the density function f(x) = 1+x2 , − ∞ < x < ∞
hence find i)P(x ≥ 0) ii) P(0 < x < 1) iii) c. d. f
Sol: Given f(x) is a probability density function.
∞
∴ f(x) ≥ 0 and ∫−∞ f(x)dx = 1
∞ k
i.e., ∫−∞ 1+x2 dx = 1
∞
1
2k ∫ dx = 1
1 + x2
0

2k[tan−1 x]∞
0 =1 ; 2k[tan−1 (∞) − tan−1 (0)] = 1
π
2k [ − 0] = 1
2
𝟏
k. π = 1𝐤 =
𝛑
1
f(x) = π 2
1+x
∞

i)P(x ≥ 0) = ∫ f(x)dx
0
1
∞ π
= ∫0 dx
1+x2

31
 1 ∞ 1
= ∫0 dx
π 1+x2
1
= π [tan−1 x]∞
0

1
P(x ≥ 0) = [tan−1 (∞) − tan−1 (0)]
π
1 π
P(x ≥ 0) = [ − 0]
π 2
𝟏
𝐏(𝐱 ≥ 𝟎) =
𝟐
1

ii)P(0 < x < 1) = ∫ f(x)dx

0
1
1 π
= ∫0 dx
1+x2
1 1 1
= π ∫0 dx
1+x2

1
P(0 < x < 1) = [tan−1 x]10
π
1
= [tan−1 (1) − tan−1 (0)]
π
1 π
= π [ 4 − 0]

𝟏
𝐏(𝟎 < 𝐱 < 𝟏) =
𝟒
x

iii)c. d. f = ∫ f(x)dx
−∞
1
x π
= ∫−∞ 1+x 2
dx
1 x 1
= π ∫−∞ 1+x2 dx

1
c. d. f = [tan−1 x]x−∞
π
1
= [tan−1 (x) − tan−1 (−∞)]
π
1 π
= π [tan−1 x − (− 2 )]
𝟏 𝛑
𝐜. 𝐝. 𝐟 = 𝛑 [𝐭𝐚𝐧−𝟏 𝐱 + 𝟐] if −∞ < 𝐱 < ∞.

32
6. The daily turn over in a medical shop is exponentially distributed with Rs.6000 as the
average with a net profit of 8%. Find the probability that the net profit exceeds Rs.500
on a randomly chosen day.
Sol: Let ‘x’ denote the turn over per day.
∴ By exponential distribution
f(x) = αe−αx ; x > 0.
1
Given, By data, μ = α = 6000
𝟏
∴ 𝛂 = 𝟔𝟎𝟎𝟎

1 − x
f(x) = e 6000 , 0 < x < ∞
6000
Let ‘A’ be the turn over for which the net profit is Rs.500.
Since the net profit is 8% of the turn over,
8
A∗ = 500
100
⇒ 𝐀 = 𝟔𝟐𝟓𝟎
∴ The probability of profit exceeding Rs.500 is
∞

P(x > 6250) = ∫ f(x)dx

6250
∞
1 − x
P(x > 6250) = ∫ e 6000 dx
6000
6250
x
1 ∞
= 6000 ∫6250 e−6000 dx
x ∞
−
1 e 6000
= 6000 [ 1 ]
−
6000 6250
6250
= − [e−∞ − e−6000 ]
6250
= e− 6000
𝐏(𝐱 > 𝟔𝟐𝟓𝟎) = 𝟎. 𝟑𝟓𝟐𝟗

33
Normal Distribution
The continuous probability distribution having the probability density function f(x) given
(x−μ)2
1 −
by f(x) = σ√2π e 2σ2 where −∞ < x < ∞ , − ∞ < μ < ∞ and σ > 0 is known as the
normal distribution.
∞
Evidently f(x) ≥ 0 and we have ∫−∞ f(x)dx = 1.

f(x) satisfy both the conditions required for a continuous probability function.
Mean and standard deviation of the Normal Distribution
𝐌𝐞𝐚𝐧 = 𝛍
𝐕𝐚𝐫𝐢𝐚𝐧𝐜𝐞 = 𝛔𝟐
Geometrically we can write
∞ 𝟎 ∞ 𝟏
𝐢) ∫−∞ 𝛟(𝐳)𝐝𝐳 = 𝟏 𝐢𝐢) ∫−∞ 𝛟(𝐳)𝐝𝐳 = ∫𝟎 𝛟(𝐳)𝐝𝐳 = 𝟐

Note
1. P(−∞ ≤ z ≤ ∞) = 1
1
2. P(−∞ ≤ z ≤ 0) = 2
1
3. P(0 ≤ z ≤ ∞) 𝐨𝐫 P(z ≥ 0) = 2

Also, P(−∞ < z < z1 ) = P(−∞ < z ≤ 0) + P(0 ≤ z < z1 )

i.e., 𝐏(𝐳 < 𝐳𝟏 ) = 𝟎. 𝟓 + 𝛟(𝐳𝟏 )
Also, P(z > z2 ) = P(z ≥ 0) − P(0 ≤ z < z2 )
i.e., 𝐏(𝐳 > 𝐳𝟐 ) = 𝟎. 𝟓 − 𝛟(𝐳𝟐 )

Problems
1. Evaluate the following probabilities with the help of normal probability tables
i) P(z ≥ 0.85) ii) P(−1.64 ≤ z ≤ −0.88) iii) P(z ≤ −2.43) iv) P(|z| ≤ 1.94)
Sol: i) P(z ≥ 0.85) = P(z ≥ 0) − P(0 ≤ z ≤ 0.85)
= 0.5 − ϕ(0.85)
= 0.5 − 0.3023
𝐏(𝐳 ≥ 𝟎. 𝟖𝟓) = 𝟎. 𝟏𝟗𝟕𝟕

34
ii) P(−1.64 ≤ z ≤ −0.88)
By Symmetry, P(−1.64 ≤ z ≤ −0.88) = P(0.88 ≤ z ≤ 1.64)
= P(0 ≤ z ≤ 1.64) − P(0 ≤ z ≤ 0.88)
= ϕ(1.64) − ϕ(0.88)
= 0.4495 − 0.3106
𝐏(−𝟏. 𝟔𝟒 ≤ 𝐳 ≤ −𝟎. 𝟖𝟖) = 𝟎. 𝟏𝟑𝟖𝟗
iii) P(z ≤ −2.43) = P(z ≥ 2.43)
P(z ≤ −2.43) = P(z ≥ 0) − P(0 ≤ z ≤ 2.43)
= 0.5 − ϕ(2.43)
= 0.5 − 0.4925
𝐏(𝐳 ≥ −𝟐. 𝟒𝟑) = 𝟎. 𝟎𝟎𝟕𝟓
iv) P(|z| ≤ 1.94) = P(−1.94 ≤ z ≤ 1.94)
= 2 P(0 ≤ z ≤ 1.94)
= 2. ϕ(1.94)
= 2. (0.4738)
𝐏(|𝐳| ≤ 𝟏. 𝟗𝟒) = 𝟎. 𝟗𝟒𝟕𝟔
2. If x is a normal variate with mean 30 and standard deviation 5 find the probability that
i) 26 ≤ x ≤ 40 ii) x ≥ 45
x−μ x−30
Sol: We have standard normal variate, z = = [By data μ = 30, σ = 5 ]
σ 5

i) To find 𝐏(𝟐𝟔 ≤ 𝐱 ≤ 𝟒𝟎)

26−30 40−30
If x = 26 , z = = −0.8 ; If x = 40 , z = =2
5 5

Hence we need to find , 𝐏(−𝟎. 𝟖 ≤ 𝐳 ≤ 𝟐)

P(−0.8 ≤ z ≤ 2) = P(−0.8 ≤ z ≤ 0) + P(0 ≤ z ≤ 2)
= P(0 ≤ z ≤ 0.8) + P(0 ≤ z ≤ 2)
= ϕ(0.8) + ϕ(2)
= 0.2881 + 0.4772
𝐏(−𝟎. 𝟖 ≤ 𝐳 ≤ 𝟐) = 𝟎. 𝟕𝟔𝟓𝟑
ii) To find 𝐏(𝐱 ≥ 𝟒𝟓)
45−30
If x = 45 , z = 5 = 3

Hence we need to find , 𝐏(𝐳 ≥ 𝟑)

35
P(z ≥ 3) = P(z ≥ 0) − P(0 ≤ z ≤ 3)
= 0.5 − ϕ(3)
= 0.5 − 0.4987
𝐏(𝐳 ≥ 𝟑) = 𝟎. 𝟎𝟎𝟏𝟑
3. The marks of 1000 students in an examination follows a normal distribution with mean
70 and standard deviation 5. Find the number of students whose marks will be
i) less than 65 ii) more than 75 iii) 65 to 75
Sol: Let x represent the marks of students
x−μ x−70
We have standard normal variate, z = = [By data μ = 70, σ = 5 ]
σ 5

i) To find 𝐏(𝐱 < 𝟔𝟓)

65−70
If x = 65 , z = = −1
5

Hence we need to find , 𝐏(𝐳 < −𝟏)

P(z < −1) = P(z > 1)
= P(z ≥ 0) − P(0 ≤ z ≤ 1)
= 0.5 − ϕ(1)
= 0.5 − 0.3413
𝐏(𝐳 < −𝟏) = 𝟎. 𝟏𝟓𝟖𝟕
Number of students scoring less than 65 marks=1000*0.1587= 𝟏𝟓𝟖. 𝟕 ≅ 𝟏𝟓𝟗
ii) To find 𝐏(𝐱 > 𝟕𝟓)
75−70
If x = 75 , z = 5 = 1

Hence we need to find , 𝐏(𝐳 > 𝟏)

P(z > 1) = P(z ≥ 0) − P(0 ≤ z ≤ 1)
= 0.5 − ϕ(1)
= 0.5 − 0.3413
𝐏(𝐳 > 𝟏) = 𝟎. 𝟏𝟓𝟖𝟕
Number of students scoring more than 75 marks=1000*0.1587= 𝟏𝟓𝟖. 𝟕 ≅ 𝟏𝟓𝟗

36
iii) To find 𝐏(𝟔𝟓 < 𝐱 < 𝟕𝟓)
Hence we need to find , 𝐏(−𝟏 < 𝐳 < 𝟏)
𝐏(−𝟏 < 𝐳 < 𝟏) = 𝟐𝐏(𝟎 < 𝐳 < 𝟏)
= 2ϕ(1)
= 2(0.3413)
𝐏(−𝟏 < 𝐳 < 𝟏) = 𝟎. 𝟔𝟖𝟐𝟔
Number of students scoring marks between 65 to 75 marks=1000*0.6826= 𝟔𝟖𝟐. 𝟔 ≅
𝟔𝟖𝟑
4. In a test on electric bulbs, it was found that the life time of a particular brand was
distributed normally with an average life of 2000 hours and S.D of 60 hours. If a firm
purchases 2500 bulbs, find the number of bulbs that are likely to last for i) more than 2100
hours ii) less than 1950 hours iii) between 1900 to 2100 hours.
Sol: Given μ = 2000, σ = 60
𝐱−𝛍 𝐱−𝟐𝟎𝟎𝟎
We have standard normal variate, 𝐳 = =
𝛔 𝟔𝟎

i) To find 𝐏(𝐱 > 𝟐𝟏𝟎𝟎)

2100−2000
If x = 2100 , z = = 1.67
60

Hence we need to find , 𝐏(𝐳 > 𝟏. 𝟔𝟕)

P(z > 1.67) = P(z ≥ 0) − P(0 < z < 1.67)
= 0.5 − ϕ(1.67)
= 0.5 − 0.4525
𝐏(𝐳 > 𝟏. 𝟔𝟕) = 𝟎. 𝟎𝟒𝟕𝟓
∴ Number of bulbs that are likely to last for more than 2100 hours = 𝟐𝟓𝟎𝟎 ∗ 𝟎. 𝟎𝟒𝟕𝟓
= 𝟏𝟏𝟖. 𝟕𝟓 ≅ 𝟏𝟏𝟗
ii) To find 𝐏(𝐱 < 𝟏𝟗𝟓𝟎)

1950−2000
If x = 1950 , z = = −0.83
60

Hence we need to find , 𝐏(𝐳 < −𝟎. 𝟖𝟑)

P(z < −0.83) = P(z > 0.83)
= P(z ≥ 0) − P(0 < z < 0.83)
= 0.5 − ϕ(0.83)

37
 = 0.5 − 0.2967
𝐏(𝐳 < −𝟎. 𝟖𝟑) = 𝟎. 𝟐𝟎𝟑𝟑
∴ Number of bulbs that are likely to last for more than 1950 hours = 𝟐𝟓𝟎𝟎 ∗ 𝟎. 𝟐𝟎𝟑𝟑
= 𝟓𝟎𝟖. 𝟐𝟓 ≅ 𝟓𝟎𝟖
iii) To find 𝐏(𝟏𝟗𝟎𝟎 < 𝐱 < 𝟐𝟏𝟎𝟎)

1900−2000 2100−2000
If x = 1900 , z = = −1.67 ; If x = 2100 , z = = 1.67
60 60

Hence we need to find , 𝐏(−𝟏. 𝟔𝟕 < 𝐳 < 𝟏. 𝟔𝟕)

P(−1.67 < z < 1.67) = 2P(0 < z < 1.67)
= 2 ∗ ϕ(1.67)
= 2 ∗ 0.4525
𝐏(−𝟏. 𝟔𝟕 < 𝐳 < 𝟏. 𝟔𝟕) = 𝟎. 𝟗𝟎𝟓
∴ Number of bulbs that are likely to last between1900 and 2100 hours
= 𝟐𝟓𝟎𝟎 ∗ 𝟎. 𝟐𝟎𝟑𝟑
= 𝟐𝟐𝟔𝟐. 𝟓 ≅ 𝟐𝟐𝟔𝟑
5. In a normal distribution 31% of the items are under 45 and 8% of the items are over 64.
Find the mean and S.D of the distribution.
Sol: Let μ and σ be the mean and S.D of the normal distribution.
By data, P(x < 45) = 0.31 and P(x > 64) = 0.08
𝐱−𝛍
We have standard normal variate, 𝐳 =
𝛔
45−μ
When x = 45 , z = = z1 (Say)
σ
64−μ
x = 64 , z = = z2 (Say)
σ

So we have ,
P(z < z1 ) = 0.31 and P(z > z2 ) = 0.08
i.e., 0.5 + φ(z1 ) = 0.31 and 0.5 − φ(z2 ) = 0.08
φ(z1 ) = 0.31 − 0.5 and φ(z2 ) = 0.5 − 0.08

φ(z1 ) = −0.19 and φ(z2 ) = 0.42

Referring to the normal probability tables, we have
0.1915(≅ 0.19) = φ(0.5) and 0.4192(≅ 0.42) = φ(1.4)
φ(z1 ) = −φ(0.5) and φ(z2 ) = φ(1.4)

38
 ⇒ 𝐳𝟏 = −𝟎. 𝟓 and 𝐳𝟐 = 𝟏. 𝟒
45−μ 64−μ
i.e., = −0.5 and = 1.4
σ σ

45 − μ = −0.5σ and 64 − μ = 1.4σ

μ − 0.5σ = 45 and μ + 1.4σ = 64
Solving above equations, we get
𝛍 = 𝟓𝟎 and 𝛔 = 𝟏𝟎
Thus, Mean=50 and S.D=10
6. The mean weight of 500 students during a medical examination was found to be 50 kgs and
S.D weight 6 kgs. Assuming that the weights are normally distributed, find the number of
student having weight i) between 40 and 50 kgs ii) more than 60 kgs. [φ(1.67) = 0.4525]
Sol: Given μ = 50, σ=6
𝐱−𝛍 𝐱−𝟓𝟎
We have standard normal variate, 𝐳 = =
𝛔 𝟔

i) To find 𝐏(𝟒𝟎 < 𝐱 < 𝟓𝟎)

40−50 50−50
If x = 40 , z = = −1.67 ; If x = 50 , z = =0
6 6

Hence we need to find , 𝐏(−𝟏. 𝟔𝟕 < 𝐳 < 𝟎)

P(−1.67 < z < 0) = P(0 < z < 1.67)
= ϕ(1.67)
𝐏(−𝟏. 𝟔𝟕 < 𝐳 < 𝟎) = 0.4525
∴ Number of students having weight between 40 and 50 kgs= 𝟓𝟎𝟎 ∗ 𝟎. 𝟒𝟓𝟐𝟓 = 𝟐𝟐𝟔. 𝟐𝟓
≅ 𝟐𝟐𝟔
ii) To find 𝐏(𝐱 > 𝟔𝟎)
60−50
If x = 60 , z = 6 = 1.67

Hence we need to find , 𝐏(𝐳 > 𝟏. 𝟔𝟕)

P(z > 1.67) = P(z ≥ 0) − P(0 < z < 1.67)
= 0.5 − ϕ(1.67)
= 0.5 − 0.4525
𝐏(𝐳 > 𝟏. 𝟔𝟕) = 𝟎. 𝟎𝟒𝟕𝟓
∴ Number of students having weight more than 60 kgs = 𝟓𝟎𝟎 ∗ 𝟎. 𝟎𝟒𝟕𝟓
= 𝟐𝟑. 𝟕𝟓 ≅ 𝟐𝟒

39
7. In an examination 7% of students score less than 35% marks and 89% of students score
less than 60% of marks. Find the mean and S.D of the distribution if the marks are
normally distributed. It is given that P(1.2263)=0.39 and P(1.4757)=0.43.
Sol: Let μ and σ be the mean and S.D of the normal distribution.
By data, P(x < 35) = 0.07 and P(x < 60) = 0.89
𝐱−𝛍
We have standard normal variate, 𝐳 = 𝛔
35−μ
When x = 35 , z = = z1 (Say)
σ
60−μ
x = 60 , z = = z2 (Say)
σ

So we have ,
P(z < z1 ) = 0.07 and P(z < z2 ) = 0.89
i.e., 0.5 + φ(z1 ) = 0.07 and 0.5 + φ(z2 ) = 0.89

φ(z1 ) = 0.07 − 0.5 and φ(z2 ) = 0.89 − 0.5

φ(z1 ) = −0.43 and φ(z2 ) = 0.39

Using the given data in the RHS of these, we have
φ(z1 ) = −φ(1.4757) and φ(z2 ) = φ(1.2263)
⇒ 𝐳𝟏 = −𝟏. 𝟒𝟕𝟓𝟕 and 𝐳𝟐 = 𝟏. 𝟐𝟐𝟔𝟑
35−μ 60−μ
i.e., = −1.4757 and = 1.2263
σ σ

35 − μ = −1.4757σ and 60 − μ = 1.2263σ

μ − 1.4757σ = 35 and μ + 1.2263σ = 60
Solving above equations, we get
𝛍 = 𝟒𝟖. 𝟔𝟓 and 𝛔 = 𝟗. 𝟐𝟓
Thus, Mean = 𝟒𝟖. 𝟔𝟓 and S.D = 𝟗. 𝟐𝟓

40
 MODULE-2
JOINT PROBABILITY DISTRIBUTIONS AND MARKOV CHAIN
Joint Probability Distribution
If 𝑋 and 𝑌 are two discrete random variables, we define the joint probability
function of 𝑋 and 𝑌 by P(𝑋 = 𝑥, 𝑌 = 𝑦) = 𝑓(𝑥, 𝑦) where 𝑓(𝑥, 𝑦) satisfy the conditions
𝑓(𝑥, 𝑦) ≥ 0 and ∑𝑥 ∑𝑦 𝑓(𝑥, 𝑦) = 1.

Suppose 𝑋 = {𝑥1 , 𝑥2 , … . , 𝑥𝑚 } and 𝑌 = {𝑦1 , 𝑦2 , … . , 𝑦𝑛 } then,

P(𝑋 = 𝑥𝑖 , 𝑌 = 𝑦𝑗 ) = 𝑓(𝑥𝑖 , 𝑦𝑗 ) is denoted by 𝐽𝑖𝑗 .

Expectation, Variance, Covariance and Correlation

If 𝑋 is a discrete random variable taking values 𝑥1 , 𝑥2 , … . , 𝑥𝑛 having probability
function 𝑓(𝑥) then,
The expectation of 𝑋 is denoted by 𝜇𝑋 = 𝐸(𝑋) = ∑𝑛𝑖=1 𝑥𝑖 𝑓(𝑥𝑖 )

The Variance of 𝑋 is denoted by 𝑉(𝑋) = ∑𝑛𝑖=1 𝑥𝑖2 𝑓(𝑥𝑖 ) − 𝜇 2 , where 𝜇 is the mean of 𝑋.
𝑉 = 𝐸(𝑋 2 ) − 𝜇𝑋2

The Standard deviation of 𝑋 is denoted by 𝜎𝑋 = √𝑉(𝑋)

If X and Y are two discrete random variables having the joint probability distribution
f(x, y) then the expectation of X and Y are defined as

𝜇𝑋 = 𝐸(𝑋) = ∑ 𝑥𝑖 𝑓(𝑥𝑖 )
𝑖

𝜇𝑌 = 𝐸(𝑌) = ∑ 𝑦𝑗 𝑔(𝑦𝑗 )
𝑗

and 𝐸(𝑋𝑌) = ∑𝑖,𝑗 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗

If X and Y are random variables having mean μX [orE(X)] and μY [orE(Y)] respectively,
then the covariance of X and Y are defined as
Cov(X, Y) = E(XY) − E(X)E(Y)
𝐶𝑜𝑣(𝑋,𝑌)
The correlation of 𝑋 and 𝑌 denoted by 𝜌(𝑋, 𝑌) is given by 𝜌(𝑋, 𝑌) = where
𝜎𝑋 𝜎𝑌

𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 , 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2

Note: If 𝑋 and 𝑌 are independent random variables then
i) 𝐸(𝑋𝑌) = 𝐸(𝑋)𝐸(𝑌)
ii) 𝐶𝑜𝑣(𝑋, 𝑌) = 0 𝑎𝑛𝑑 𝜌(𝑋, 𝑌) = 0

41
Problems
1. The joint distribution of two random variables 𝑋 and 𝑌 is as follows

𝑌→ -4 2 7
𝑋↓

1 1 1 1
8 4 8

5 1 1 1
4 8 8

Compute the following

i) 𝐸(𝑋) and 𝐸(𝑌) ii) 𝐸(𝑋𝑌) iii) 𝜎𝑋 and 𝜎𝑌 iv) 𝐶𝑜𝑣(𝑋, 𝑌) v) 𝜌(𝑋, 𝑌)
Sol: The distribution ( marginal distribution ) of 𝑋 and 𝑌 is as follows.

𝑌→ -4 2 7 𝑓(𝑥𝑖 )
𝑋↓

1 1 1 1 1 1 1 1
+ + =
8 4 8 8 4 8 2

5 1 1 1 1 1 1 1
+ + =
4 8 8 4 8 8 2

𝑔(𝑦𝑗 ) 1 1 3 1 1 3 1 1 1 1
+ = + = + =
8 4 8 4 8 8 8 8 4

Distribution of 𝑋

𝑥𝑖 1 5

𝑓(𝑥𝑖 ) 1 1
2 2

Distribution of 𝑌

𝑦𝑗 -4 2 7

𝑔(𝑦𝑗 ) 3 3 1
8 8 4

42
i) 𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )
1 1 3 3 1
𝐸(𝑋) = (1) (2) + (5) (2) 𝐸(𝑌) = (−4) (8) + (2) (8) + (7) (4)

1 5 3 3 7
𝐸(𝑋) = + 𝐸(𝑌) = − + +
2 2 2 4 4
Thus, 𝐸(𝑋) = 𝜇𝑋 = 3 𝐸(𝑌) = 𝜇𝑌 = 1
ii) 𝐸(𝑋𝑌) = ∑𝑖,𝑗 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗
1 1 1 1 1 1
𝐸(𝑋𝑌) = (1)(−4) ( ) + (1)(2) ( ) + (1)(7) ( ) + (5)(−4) ( ) + (5)(2) ( ) + (5)(7) ( )
8 4 8 4 8 8
1 1 7 5 35
𝐸 (𝑋𝑌) = − + + − 5 + +
4 2 8 4 8
3
Thus, 𝐸(𝑋𝑌) = 2

iii) 𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 ; 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2

Now, 𝐸(𝑋 2 ) = ∑𝑖 𝑥𝑖2 𝑓(𝑥𝑖 ) 𝐸(𝑌 2 ) = ∑𝑗 𝑦𝑗2 𝑓(𝑦𝑗 )
1 1 3 3 1
𝐸(𝑋 2 ) = (12 ) (2) + (52 ) (2) 𝐸(𝑌 2 ) = ((−4)2 ) (8) + (22 ) (8) + (72 ) (4)
79
𝐸(𝑋 2 ) = 13 𝐸(𝑌 2 ) = 4
79
∴ 𝜎𝑋2 = 13 − 32 ; 𝜎𝑌2 = − 12
4
75
𝜎𝑋2 = 4 ; 𝜎𝑌2 = 4

𝜎𝑋 = 2 ; 𝜎𝑌 = 4.33
iv) 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)
3
𝐶𝑜𝑣(𝑋, 𝑌) = 2 − (3)(1)

3
𝐶𝑜𝑣(𝑋, 𝑌) = −
2
𝐶𝑜𝑣(𝑋,𝑌)
v) 𝜌(𝑋, 𝑌) = 𝜎𝑋 𝜎𝑌
3
−
2
𝜌(𝑋, 𝑌) = (2)(4.33)

𝜌(𝑋, 𝑌) = −0.1732

43
2. The joint probability distribution table for two random variables 𝑋 and 𝑌 is as follows

𝑌→ -2 -1 4 5
𝑋↓

1 0.1 0.2 0 0.3

2 0.2 0.1 0.1 0

Determine the marginal distributions of 𝑋 and 𝑌 . Also compute i) Expectations of 𝑋, 𝑌 and

X𝑌 ii) Standard deviations of 𝑋, 𝑌 iii) Covariance of 𝑋 and 𝑌 iv) correlation of 𝑋 and 𝑌.
Further verify that 𝑋 and 𝑌 are dependent random variables. Also find 𝑃(𝑋 + 𝑌 > 0).
Sol: Marginal distribution of 𝑋

𝑥𝑖 1 2

𝑓(𝑥𝑖 ) 0.6 0.4

Marginal distribution of 𝑌

𝑦𝑗 -2 -1 4 5

𝑔(𝑦𝑗 ) 0.3 0.3 0.1 0.3

i) 𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )

𝐸(𝑋) = (1)(0.6) + (2)(0.4) 𝐸(𝑌) = (−2)(0.3) + (−1)(0.3) + (4)(0.1) + (5)(0.3)

𝐸(𝑋) = 0.6 + 0.8 𝐸(𝑌) = −0.6 − 0.3 + 0.4 + 1.5
Thus, 𝐸(𝑋) = 𝜇𝑋 = 1.4 𝐸(𝑌) = 𝜇𝑌 = 1

𝐸(𝑋𝑌) = ∑ 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗
𝑖,𝑗

𝐸(𝑋𝑌) = (1)(−2)(0.1) + (1)(−1)(0.2) + (1)(4)(0) + (1)(5)(0.3) + (2)(−2)(0.2)

+ (2)(−1)(0.1) + (2)(4)(0.1)
+(2)(5)(0)
𝐸(𝑋𝑌) = −0.2 − 0.2 + 0 + 1.5 − 0.8 − 0.2 + 0.8 + 0
Thus, 𝐸(𝑋𝑌) = 0.9
ii) 𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 ; 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2
Now, 𝐸(𝑋 2 ) = ∑𝑖 𝑥𝑖2 𝑓(𝑥𝑖 ) 𝐸(𝑌 2 ) = ∑𝑗 𝑦𝑗2 𝑔(𝑦𝑗 )

𝐸(𝑋 2 ) = (12 )(0.6) + (22 )(0.4) 𝐸(𝑌 2 ) = ((−2)2 )(0.3) + ((−1)2 )(0.3) + (42 )(0.1) + (52 )(0.3)

44
 𝐸(𝑋 2 ) = 2.2 𝐸(𝑌 2 ) = 10.6
∴ 𝜎𝑋2 = 2.2 − (1.4)2 ; 𝜎𝑌2 = 10.6 − 12
𝜎𝑋2 = 0.24 ; 𝜎𝑌2 = 9.6
𝜎𝑋 = 0.49 ; 𝜎𝑌 = 3.1
iii) 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)
𝐶𝑜𝑣(𝑋, 𝑌) = 0.9 − (1.4)(1)
𝐶𝑜𝑣(𝑋, 𝑌) = −0.5
𝐶𝑜𝑣(𝑋,𝑌)
iv) 𝜌(𝑋, 𝑌) = 𝜎𝑋 𝜎𝑌

−0.5
𝜌(𝑋, 𝑌) = (0.49)(3.1)

𝜌(𝑋, 𝑌) = −0.3

If 𝑋 and 𝑌 are independent random variables we must have 𝑓(𝑥𝑖 ). 𝑔(𝑦𝑗 ) = 𝐽𝑖𝑗

It can be seen that 𝑓(𝑥1 ). 𝑔(𝑦1 ) = (0.6)(0.3) = 0.18 and 𝐽11 = 0.1 i.e., 𝑓(𝑥𝑖 ). 𝑔(𝑦𝑗 ) ≠ 𝐽𝑖𝑗

Similarly for others also the condition is not satisfied.

Hence we conclude that 𝑋 and 𝑌 are dependent random variables.
We have, 𝑋 = {𝑥𝑖 } = {𝑥1 , 𝑥2 } = {1,2} respectively.

𝑌 = {𝑦𝑗 } = {𝑦1 , 𝑦2 , 𝑦3 , 𝑦4 } = {−2, −1,4,5} respectively.

Also, 𝐽11 = 0.1 , 𝐽12 = 0.2 , 𝐽13 = 0 , 𝐽14 = 0.3

𝐽21 = 0.2 , 𝐽22 = 0.1 , 𝐽23 = 0.1 , 𝐽24 = 0
𝑋 + 𝑌 > 0 is possible when (𝑋, 𝑌) take the values
(𝑥1 , 𝑦3 ) = (1,4) ; (𝑥1 , 𝑦4 ) = (1,5) ; (𝑥2 , 𝑦2 ) = (2, −1) ; (𝑥2 , 𝑦3 ) = (2,4) and
(𝑥2 , 𝑦4 ) = (2,5)
Hence , P(𝑋 + 𝑌 > 0) = 𝐽13 + 𝐽14 + 𝐽22 + 𝐽23 + 𝐽24
P(𝑋 + 𝑌 > 0) = 0 + 0.3 + 0.1 + 0.1 + 0
Thus, 𝑃(𝑋 + 𝑌 > 0) = 0.5

45
3. Suppose 𝑋 and 𝑌 are independent random variables with the following respective
distribution, find the joint distribution of 𝑋 and 𝑌. Also verify 𝐶𝑜𝑣(𝑋, 𝑌) = 0.

𝑥𝑖 1 2

𝑓(𝑥𝑖 ) 0.7 0.3

𝑦𝑗 -2 5 8

𝑔(𝑦𝑗 ) 0.3 0.5 0.2

Sol: Since 𝑋 and 𝑌 are independent, the joint distribution J(𝑥, 𝑦) is obtained by using the
definition 𝑓(𝑥𝑖 ). 𝑔(𝑦𝑗 ) = 𝐽𝑖𝑗 .

𝐽𝑖𝑗 are obtained on multiplication of the marginal entries.

From the given data , the required 𝐽𝑖𝑗 is found

𝐽11 = 𝑓(𝑥1 ). 𝑔(𝑦1 ) = (0.7)(0.3) = 0.21

𝐽12 = 𝑓(𝑥1 ). 𝑔(𝑦2 ) = (0.7)(0.5) = 0.35
𝐽13 = 𝑓(𝑥1 ). 𝑔(𝑦3 ) = (0.7)(0.2) = 0.14
𝐽21 = 𝑓(𝑥2 ). 𝑔(𝑦1 ) = (0.3)(0.3) = 0.09
𝐽22 = 𝑓(𝑥2 ). 𝑔(𝑦2 ) = (0.3)(0.5) = 0.15
𝐽23 = 𝑓(𝑥2 ). 𝑔(𝑦3 ) = (0.3)(0.2) = 0.06
The joint distribution table is as follows

𝑌→ -2 5 8 𝑓(𝑥𝑖 )
𝑋↓

1 0.21 0.35 0.14 0.7

2 0.09 0.15 0.06 0.3

𝑔(𝑦𝑗 ) 0.3 0.5 0.2 1

We have 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)

𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )

𝐸(𝑋) = (1)(0.7) + (2)(0.3) 𝐸(𝑌) = (−2)(0.3) + (5)(0.5) + (8)(0.2)

𝐸(𝑋) = 0.7 + 0.6 𝐸(𝑌) = −0.6 + 2.5 + 1.6

46
Thus, 𝐸(𝑋) = 𝜇𝑋 = 1.3 𝐸(𝑌) = 𝜇𝑌 = 3.5

𝐸(𝑋𝑌) = ∑ 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗
𝑖,𝑗

𝐸(𝑋𝑌) = (1)(−2)(0.21) + (1)(5)(0.35) + (1)(8)(0.14) + (2)(−2)(0.09) + (2)(5)(0.15) + (2)(8)(0.06)

𝐸(𝑋𝑌) = −0.42 + 1.75 + 1.12 − 0.36 + 1.5 + 0.96

Thus, 𝐸(𝑋𝑌) = 4.55
Hence 𝐶𝑜𝑣(𝑋, 𝑌) = 4.55 − (1.3)(3.5) = 0
Thus the result 𝐶𝑜𝑣(𝑋, 𝑌) = 0 for independent random variables 𝑋 and 𝑌 is verified.
4. 𝑋 and 𝑌 are independent random variables. 𝑋 take values 2,5,7 with probability
1⁄ , 1⁄ , 1⁄ respectively. 𝑌 take values 3,4,5 with the probability 1⁄ , 1⁄ , 1⁄ .
2 4 4 3 3 3
i) Find the joint probability distribution of 𝑋 and 𝑌
ii) Show that the covariance of 𝑋 and 𝑌 is equal to zero.
iii) Find the probability distribution of Z = 𝑋 + 𝑌
Sol: Given

𝑥𝑖 2 5 7

𝑓(𝑥𝑖 ) 1⁄ 1⁄ 1⁄
2 4 4

𝑦𝑗 3 4 5

𝑔(𝑦𝑗 ) 1⁄ 1⁄ 1⁄
3 3 3

i) We have 𝐽𝑖𝑗 = 𝑓(𝑥𝑖 ). 𝑔(𝑦𝑗 ) where 𝑖, 𝑗 = 1,2,3

𝐽11 = 𝑓(𝑥1 ). 𝑔(𝑦1 ) = 1⁄2 . 1⁄3 = 1⁄6 𝐽12 = 𝑓(𝑥1 ). 𝑔(𝑦2 ) = 1⁄2 . 1⁄3 = 1⁄6

𝐽13 = 𝑓(𝑥1 ). 𝑔(𝑦3 ) = 1⁄2 . 1⁄3 = 1⁄6 𝐽21 = 𝑓(𝑥2 ). 𝑔(𝑦1 ) = 1⁄4 . 1⁄3 = 1⁄12

𝐽22 = 𝑓(𝑥2 ). 𝑔(𝑦2 ) = 1⁄4 . 1⁄3 = 1⁄12 𝐽23 = 𝑓(𝑥2 ). 𝑔(𝑦3 ) = 1⁄4 . 1⁄3 = 1⁄12

𝐽31 = 𝑓(𝑥3 ). 𝑔(𝑦1 ) = 1⁄4 . 1⁄3 = 1⁄12 𝐽32 = 𝑓(𝑥3 ). 𝑔(𝑦2 ) = 1⁄4 . 1⁄3 = 1⁄12

𝐽33 = 𝑓(𝑥3 ). 𝑔(𝑦3 ) = 1⁄4 . 1⁄3 = 1⁄12

47
The joint distribution table is as follows

𝑌→ 3 4 5 𝑓(𝑥𝑖 )
𝑋↓

2 1⁄ 1⁄ 1⁄ 1⁄
6 6 6 2

5 1⁄ 1⁄ 1⁄ 1⁄
12 12 12 4

7 1⁄ 1⁄ 1⁄ 1⁄
12 12 12 4

𝑔(𝑦𝑗 ) 1⁄ 1⁄ 1⁄ 1
3 3 3

ii) We have 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)

𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )

𝐸(𝑋) = (2)(1⁄2) + (5)(1⁄4) + (7)(1⁄4) 𝐸(𝑌) = (3)(1⁄3) + (4)(1⁄3) + (5)(1⁄3)

𝐸(𝑋) = 4 𝐸(𝑌) = 4
Thus, 𝐸(𝑋) = 𝜇𝑋 = 4 𝐸(𝑌) = 𝜇𝑌 = 4

𝐸(𝑋𝑌) = ∑ 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗
𝑖,𝑗

𝐸(𝑋𝑌) = (2)(3)(1⁄6) + (2)(4)(1⁄6) + (2)(5)(1⁄6) + (5)(3)(1⁄12) + (5)(4)(1⁄12) + (5)(5)(1⁄12) + (7)(3)(1⁄12) + (7)(4)(1⁄12) + (7)(5)(1⁄12)

𝐸(𝑋𝑌) = 16
Thus, 𝐸(𝑋𝑌) = 16
Thus, 𝐶𝑜𝑣(𝑋, 𝑌) = 16 − (4)(4) = 0
iii) Z = 𝑋 + 𝑌
Let 𝑧𝑖 = 𝑥𝑖 + 𝑦𝑖 and hence {𝑧𝑖 } = {5,6,7,8,9,10,11,12}
The corresponding probabilities are ,
1⁄ , 1⁄ , 1⁄ , 1⁄ , 1⁄ , 1⁄ + 1⁄ , 1⁄ , 1⁄
6 6 6 12 12 12 12 12 12
The probability distribution of Z = 𝑋 + 𝑌 is as follows

Z 5 6 7 8 9 10 11 12

P(Z) 1⁄ 1⁄ 1⁄ 1⁄ 1⁄ 1⁄ + 1⁄ = 1⁄ 1⁄ 1⁄
6 6 6 12 12 12 12 6 12 12

We note that ∑ 𝑃(𝑍) = 1.

48
5. The joint probability distribution of two discrete random variables 𝑋 and 𝑌 is given by
𝑓(𝑥, 𝑦) = 𝑘(2𝑥 + 𝑦) where 𝑥 and 𝑦 are integers such that 0 ≤ 𝑥 ≤ 2 0 ≤ 𝑦 ≤ 3.
i) Find the value of the constant k.
ii) Find the marginal probability distributions of 𝑋 and 𝑌.
iii) Show that the random variables 𝑋 and 𝑌 are dependent.
iv) Compute E(𝑋)𝐸(𝑌), 𝐸(𝑋𝑌)
v) Compute E(𝑋 2 ), 𝐸(𝑌 2 )
vi) Compute 𝜎𝑋 , 𝜎𝑌
vii) Find P(𝑋 = 1, 𝑌 = 2) , P(𝑋 = 2, 𝑌 = 1)
viii) Find P(𝑋 ≥ 1, 𝑌 ≤ 2) , P(𝑋 + 𝑌 > 2)

Sol: 𝑋 = {𝑥𝑖 } = {0,1,2} and 𝑌 = {𝑦𝑗 } = {0,1,2,3}

𝑓(𝑥, 𝑦) = 𝑘(2𝑥 + 𝑦) and the joint probability distribution table is formed as follows

𝑌→ 0 1 2 3 Sum
𝑋↓

0 0 𝑘 2𝑘 3𝑘 6𝑘

1 2𝑘 3𝑘 4𝑘 5𝑘 14𝑘

2 4𝑘 5𝑘 6𝑘 7𝑘 22𝑘

Sum 6𝑘 9𝑘 12𝑘 15𝑘 42𝑘

i) We must have , 42𝑘 = 1

𝟏
𝒌 = 𝟒𝟐

ii) Marginal probability distribution is as follows

𝒙𝒊 0 1 2

𝑓(𝑥𝑖 ) 6⁄ 14⁄ 22⁄

42 42 42

𝒚𝒋 0 1 2 3

𝑔(𝑦𝑗 ) 6⁄ 9⁄ 12⁄ 15⁄

42 42 42 42

iii) It can be easily seen that 𝑓(𝑥𝑖 ). 𝑔(𝑦𝑗 ) ≠ 𝐽𝑖𝑗

Hence the random variables are dependent.

49
iv) 𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )
𝐸(𝑋) = (0)(6⁄42) + (1)(14⁄42) + (2)(22⁄42) 𝐸(𝑌) = (0)(6⁄42) + (1)(9⁄42) + (2)(12⁄42) + (3)(15⁄42)

𝐸(𝑋) = 58⁄42 𝐸(𝑌) = 39⁄21

Thus, 𝑬(𝑿) = 𝝁𝑿 = 𝟐𝟗⁄𝟐𝟏 𝑬(𝒀) = 𝝁𝒀 = 𝟏𝟑⁄𝟕

𝐸(𝑋𝑌) = ∑ 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗
𝑖,𝑗

𝐸(𝑋𝑌) = (0)(0 + 𝑘 + 4𝑘 + 9𝑘) + (1)(0 + 3𝑘 + 8𝑘 + 15𝑘) + (2)(0 + 5𝑘 + 12𝑘 + 21𝑘)

𝐸(𝑋𝑌) = 0 + 26𝑘 + 76𝑘
102
𝐸(𝑋𝑌) = 102𝑘 =
42
𝟏𝟕
Thus, 𝑬(𝑿𝒀) = 𝟕

v) 𝐸(𝑋 2 ) = ∑𝑖 𝑥𝑖2 𝑓(𝑥𝑖 ) 𝐸(𝑌 2 ) = ∑𝑗 𝑦𝑗2 𝑔(𝑦𝑗 )

14 22 9 12 15
𝐸(𝑋 2 ) = (0) + (12 ) ( ) + (22 ) ( ) 𝐸(𝑌 2 ) = (0) + (12 ) ( ) + (22 ) ( ) + (32 ) ( )
42 42 42 42 42

102 192
𝐸(𝑋 2 ) = 𝐸(𝑌 2 ) =
42 42
𝟏𝟕 𝟑𝟐
Thus, 𝑬(𝑿𝟐 ) = 𝑬(𝒀𝟐 ) =
𝟕 𝟕

vi) 𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2

17 29 2 32 13 2
𝜎𝑋2 = −( ) 𝜎𝑌2 = −( )
7 21 7 7
230 495
𝜎𝑋2 = 441 𝜎𝑌2 = 441

√230 √495
𝜎𝑋 = 𝜎𝑌 =
21 21

Thus, 𝝈𝑿 = 𝟎. 𝟕𝟐 and 𝝈𝒀 = 𝟏. 𝟎𝟔
vii) Let 𝑋 = {𝑥𝑖 } = {𝑥1 , 𝑥2 , 𝑥3 } = {0,1,2} respectively.

𝑌 = {𝑦𝑗 } = {𝑦1 , 𝑦2 , 𝑦3 , 𝑦4 } = {0,1,2,3} respectively.

Also, 𝐽11 = 0 , 𝐽12 = 1⁄42 , 𝐽13 = 2⁄42, 𝐽14 = 3⁄42

𝐽21 = 2⁄42, 𝐽22 = 3⁄42, 𝐽23 = 4⁄42, 𝐽24 = 5⁄42

𝐽31 = 4⁄42, 𝐽32 = 5⁄42, 𝐽33 = 6⁄42, 𝐽34 = 7⁄42

Now, P(𝑋 = 1, 𝑌 = 2) = 𝑓(𝑥2 , 𝑦3 ) = 𝐽23 = 4⁄42 = 2⁄21

50
P(𝑋 = 2, 𝑌 = 1) = 𝑓(𝑥3 , 𝑦2 ) = 𝐽32 = 5⁄42

viii) (𝑋, 𝑌) = {(1,0), (1,1), (1,2), (2,0), (2,1), (2,2)} = {(𝑥2 , 𝑦1 ), (𝑥2 , 𝑦2 ), (𝑥2 , 𝑦3 ), (𝑥3 , 𝑦1 ), (𝑥3 , 𝑦2 ), (𝑥3 , 𝑦3 )}
P(𝑋 ≥ 1, 𝑌 ≤ 2) = 𝐽21 +𝐽22 +𝐽23 +𝐽31 +𝐽32 +𝐽33
2 3 4 4 5 6 24
P(𝑋 ≥ 1, 𝑌 ≤ 2) = 42 + 42 + 42 + 42 + 42 + 42 = 42
𝟒
𝑷(𝑿 ≥ 𝟏, 𝒀 ≤ 𝟐) = 𝟕

To find 𝑷(𝑿 + 𝒀 > 𝟐)

(𝑋, 𝑌) = {(𝟎, 𝟑)(𝟏, 𝟐)(𝟏, 𝟑)(𝟐, 𝟏)(𝟐, 𝟐)(𝟐, 𝟑) }
= {(𝑥1 , 𝑦4 ), (𝑥2 , 𝑦3 ), (𝑥2 , 𝑦4 ), (𝑥3 , 𝑦2 ), (𝑥3 , 𝑦3 ), (𝑥3 , 𝑦4 )}
P(𝑋 + 𝑌 > 2) = 𝐽14 +𝐽23 +𝐽24 +𝐽32 +𝐽33 +𝐽34
3 4 5 5 6 7 30
P(𝑋 + 𝑌 > 2) = 42 + 42 + 42 + 42 + 42 + 42 = 42
𝟓
𝑷(𝑿 + 𝒀 > 𝟐) = 𝟕

6. A fair coin is tossed thrice. The random variables 𝑋 𝑎𝑛𝑑 𝑌 are defined as follows.
𝑋 = 0 𝑜𝑟 1 according as head or tail occurs on the first toss. 𝑌 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 ℎ𝑒𝑎𝑑𝑠
i) Determine the distributions of 𝑋 𝑎𝑛𝑑 𝑌.
ii) Determine the joint distribution of 𝑋 𝑎𝑛𝑑 𝑌.
iii) Obtain the expectations of 𝑋, 𝑌 𝑎𝑛𝑑 𝑋𝑌. Also find S.D’s of 𝑋 𝑎𝑛𝑑 𝑌.
iv) Compute the covariance and correlation of 𝑋 𝑎𝑛𝑑 𝑌.
Sol: The sample space 𝑆 and the association of random variables 𝑋 𝑎𝑛𝑑 𝑌 is given by the
following table

𝑺 𝑯𝑯𝑯 𝑯𝑯𝑻 𝑯𝑻𝑯 𝑯𝑻𝑻 𝑻𝑯𝑯 𝑻𝑯𝑻 𝑻𝑻𝑯 𝑻𝑻𝑻

𝑋 0 0 0 0 1 1 1 1

𝑌 3 2 2 1 2 1 1 0

i) The probability distribution of 𝑋 𝑎𝑛𝑑 𝑌 is found as follows.

𝑋 = {𝑥𝑖 } = {0,1} and 𝑌 = {𝑦𝑗 } = {0,1,2,3}

4 1 4 1
𝑃(𝑋 = 0) = 8 = 2 ; 𝑃(𝑋 = 1) = 8 = 2

1 3 3 1
𝑃(𝑌 = 0) = 8 ; 𝑃(𝑌 = 1) = 8 ; 𝑃(𝑌 = 2) = 8 ; 𝑃(𝑌 = 3) = 8

Thus, we have the following probability distribution of 𝑋 𝑎𝑛𝑑 𝑌.

51
 𝑥𝑖 0 2

𝑓(𝑥𝑖 ) 1 1
2 2

𝑦𝑗 0 1 2 3

𝑔(𝑦𝑗 ) 1 3 3 1
8 8 8 8

ii) The joint distribution of 𝑋 𝑎𝑛𝑑 𝑌 is found by computing 𝐽𝑖𝑗 = 𝑃(𝑋 = 𝑥𝑖 , 𝑌 = 𝑦𝑗 ) where

we have 𝑥1 = 0, 𝑥2 = 1 and 𝑦1 = 0, 𝑦2 = 1, 𝑦3 = 2, 𝑦4 = 3
𝐽11 = 𝑃(𝑋 = 0, 𝑌 = 0) = 0 (𝑋 = 0 implies that there is a head turn out and 𝑌 the total
number of heads is 0, outcome is impossible)
1 2 1 1
𝐽12 = 𝑃(𝑋 = 0, 𝑌 = 1) = 𝐽13 = 𝑃(𝑋 = 0, 𝑌 = 2) = = 𝐽14 = 𝑃(𝑋 = 0, 𝑌 = 3) =
8 8 4 8
(Corresponding to the out come HTT) (out comes are HHT and HTH) (out come is HHH)

1
𝐽21 = 𝑃(𝑋 = 1, 𝑌 = 0) = 8 ; out come is TTT
2 1
𝐽22 = 𝑃(𝑋 = 1, 𝑌 = 1) = 8 = 4 ; out come is THT, TTH
1
𝐽23 = 𝑃(𝑋 = 1, 𝑌 = 2) = 8 ; out come is THH

𝐽24 = 𝑃(𝑋 = 1, 𝑌 = 3) = 0 ; since the out come is impossible.

The required joint probability distribution of 𝑋 𝑎𝑛𝑑 𝑌 is as follows.

𝒀→ 0 1 2 3 Sum
𝑿↓

0 0 1 1 1 1
8 4 8 2

1 1 1 1 0 1
8 4 8 2

Sum 1 3 3 1 1
8 8 8 8

iii) 𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )

𝐸(𝑋) = (0)(1⁄2) + (1)(1⁄2) 𝐸(𝑌) = (0)(1⁄8) + (1)(3⁄8) + (2)(3⁄8) + (3)(1⁄8)

52
𝐸(𝑋) = 1⁄2 𝐸(𝑌) = 12⁄8

Thus, 𝑬(𝑿) = 𝝁𝑿 = 𝟏⁄𝟐 𝑬(𝒀) = 𝝁𝒀 = 𝟑⁄𝟐

Now, 𝐸(𝑋𝑌) = ∑𝑖,𝑗 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗

𝐸(𝑋𝑌) = 0 + (1)(1)(1⁄4) + (1)(2)(1⁄8)+0

𝐸(𝑋𝑌) = 1⁄4 + 2⁄8

Thus, 𝑬(𝑿𝒀) = 𝟏⁄𝟐

Also, 𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2

Now, 𝐸(𝑋 2 ) = ∑𝑖 𝑥𝑖2 𝑓(𝑥𝑖 ) 𝐸(𝑌 2 ) = ∑𝑗 𝑦𝑗2 𝑓(𝑦𝑗 )
1 1 1 3 3 1
𝐸(𝑋 2 ) = (02 ) (2) + (12 ) (2) 𝐸(𝑌 2 ) = (0) ( ) + (12 ) ( ) + (22 ) ( ) + (32 ) ( )
8 8 8 8

𝟏
𝑬(𝑿𝟐 ) = 𝟐 𝑬(𝒀𝟐 ) = 𝟑

1 1 2 3 2
𝜎𝑋2 = 2 − (2) 𝜎𝑌2 = 3 − (2)
1 3
𝜎𝑋2 = 4 𝜎𝑌2 = 4

𝟏 √𝟑
𝝈𝑿 = 𝟐 𝝈𝒀 = 𝟐

iv) We have 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)

1 1 3
𝐶𝑜𝑣(𝑋, 𝑌) = 2 − 2. 2
𝟏
𝑪𝒐𝒗(𝑿, 𝒀) = − 𝟒
𝐶𝑜𝑣(𝑋,𝑌)
Also, 𝜌(𝑋, 𝑌) = 𝜎𝑋 𝜎𝑌

𝟏
−
𝟒
𝜌(𝑋, 𝑌) = 𝟏 √𝟑
( )( )
𝟐 𝟐

𝟏
𝝆(𝑿, 𝒀) = −
√𝟑

7. Let X be a random variable with the following distribution and Y is defined to be X 2

𝑥𝑖 -1 -2 1 2

𝑓(𝑥𝑖 ) 1 1 1 1
4 4 4 4

53
Determine i) the distribution g of Y ii) the joint distribution of 𝑋 𝑎𝑛𝑑 𝑌
iii) E(𝑋), 𝐸(𝑌), 𝐸(𝑋𝑌) iv) Cov(𝑋, 𝑌) v) 𝜌(𝑋, 𝑌)
Sol: Here 𝑌 is defined to be 𝑋 2 .
The distribution g of Y is given by

𝑦𝑗 1 4

𝑔(𝑦𝑗 ) 1 1
2 2

ii) The joint distribution of 𝑋 𝑎𝑛𝑑 𝑌 is

𝒀→ 1 4
𝑿↓

-2 0 1
4

-1 1 0
4

1 1 0
4

2 0 1
4

iii) 𝐸(𝑋) = ∑𝑖 𝑥𝑖 𝑓(𝑥𝑖 ) 𝐸(𝑌) = ∑𝑗 𝑦𝑗 𝑔(𝑦𝑗 )

𝐸(𝑋) = (−1)(1⁄4) + (−2)(1⁄4) + (1)(1⁄4) + (2)(1⁄4) 𝐸(𝑌) = (1)(1⁄2) + (4)(1⁄2)

𝐸(𝑋) = 0 𝐸(𝑌) = 5⁄2

Thus, 𝑬(𝑿) = 𝝁𝑿 = 𝟎 𝑬(𝒀) = 𝝁𝒀 = 𝟓⁄𝟐

Now, 𝐸(𝑋𝑌) = ∑𝑖,𝑗 𝑥𝑖 𝑦𝑗 𝐽𝑖𝑗

𝐸(𝑋𝑌) = (−2)(0) + (−2)(1) + (−1)(1⁄4) + (−1)(0) + (1)(1⁄4) + (1)(0) + (2)(0) + (2)(1)

𝐸(𝑋𝑌) = −2 − 1⁄4 + 1⁄4 + 2

Thus, 𝑬(𝑿𝒀) = 𝟎
Also, 𝜎𝑋2 = 𝐸(𝑋 2 ) − 𝜇𝑋2 𝜎𝑌2 = 𝐸(𝑌 2 ) − 𝜇𝑌2
Now, 𝐸(𝑋 2 ) = ∑𝑖 𝑥𝑖2 𝑓(𝑥𝑖 ) 𝐸(𝑌 2 ) = ∑𝑗 𝑦𝑗2 𝑓(𝑦𝑗 )

54
 1 1 1 1 1 1
𝐸(𝑋 2 ) = ((−1)2 ) ( ) + ((−2)2 ) ( ) + ((1)2 ) ( ) + ((2)2 ) ( ) 𝐸(𝑌 2 ) = (12 ) ( ) + (42 ) ( )
4 4 4 4 2 2

𝟓 17
𝑬(𝑿𝟐 ) = 𝟐 𝑬(𝒀𝟐 ) = 2

iv) Cov(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋)𝐸(𝑌)

𝐶𝑜𝑣(𝑋, 𝑌) = 0 − 0(5⁄2)

𝑪𝒐𝒗(𝑿, 𝒀) = 𝟎
𝐶𝑜𝑣(𝑋,𝑌)
v) 𝜌(𝑋, 𝑌) = 𝜎𝑋 𝜎𝑌

0
𝜌(𝑋, 𝑌) = (1.581)(1.658)

𝝆(𝑿, 𝒀) = 𝟎
8. The joint probability function of two discrete random variables 𝑋 and 𝑌 is given by
𝑓(𝑥, 𝑦) = 𝑐𝑥𝑦 for 𝑥 = 1,2,3 and 𝑦 = 1,2,3. Find the following
i) Constant 𝑐 ii) 𝑃(𝑋 = 2, 𝑌 = 3) iii) 𝑃(1 ≤ 𝑋 ≤ 2, 𝑌 ≤ 2) iv) 𝑃(𝑋 ≥ 2) v) 𝑃(𝑌 < 2)
vi) 𝑃(𝑋 = 1) v) 𝑃(𝑌 = 3)

55
 MODULE-3&4
STATISTICAL INFERENCE-1&2

Sampling Theory
Random Sampling
A large collection individuals or attributes or numerical data can be understood as a
population or universe.
A finite subset of the universe is called a sample. The number of individuals in a sample is
called a sample size(n).
If 𝑛 ≤ 30 then the sample is said to be small otherwise it is a large sample.
The selection of an individual or item from the population in such a way that each has the
same chance of being selected is called as random sampling.
Sampling where a member of the population may be selected more than once is called as
sampling with replacement.
If a member cannot be chosen more than once is called as sampling without replacement.
Parameter and Statistics
The statistical constants of population is mean(𝜇) , standard deviation (𝜎), correlation (𝜌)
are called the parameter.
The statistical constants drawn from the given population is mean(𝑥̅ ) , standard deviation
(𝑠), correlation (𝑟) are called statistic.
Testing of Hypothesis
To reach the decisions about the populations and the basis of sample information, we
make certain assumptions about the population involved. Such assumptions which may or
may not be true are called hypothesis. OR Quantitative statement about the population.
Explanation on the basis of limited evidences.
Null Hypothesis
The hypothesis formulated for the purpose of its rejection under the assumption that it is
true is called the Null Hypothesis denoted by 𝐻0 .
Any hypothesis which is complimentary to the null hypothesis is called Alternative
Hypothesis denoted by 𝐻1 .
Type-I and Type-II Errors
Type-I Error: Wrong decision to reject the null hypothesis when it is actually true.
Type-II Error: Wrong decision to accept the null hypothesis when it is actually false.

56
 Accepting the Rejecting the
hypothesis hypothesis

Hypothesis True Correct decision Wrong decision

( Type-I Error )

Hypothesis False Wrong decision Correct decision

( Type-II Error )

Significance level
The probability level, below which leads to the rejection of the hypothesis is known as the
significance level. The probability levels are fixed at 0.05 or 0.01 being 5% or 1%. These are
called Significance level.

Test of Significance
The process which enables us to decide about the acceptance or rejection of the hypothesis
is called the test of significance.

Confidence Intervals
Let us suppose that we have a normal population with mean 𝜇 and S.D 𝜎 . If 𝑥̅ is the
sample
𝑥̅ −𝜇
mean of a random sample size 𝑛 the quantity 𝑧 defined by 𝑧 = 𝜎 is called the Standard
( )
√𝑛
normal variate.
From the normal distribution table we find the 95% of the area lies between 𝑧 = −1.96
and 𝑧 = +1.96 i.e., 95% confidence interval lies between −1.96 and +1.96 .
Further 5% level of significance is denoted by 𝑧0.05 .Mathematically we can write
𝑥̅ − 𝜇
−1.96 ≤ ≤ 1.96
𝜎
( )
√𝑛
𝜎 𝜎
− (1.96) ≤ 𝑥̅ − 𝜇 ≤ (1.96)
√𝑛 √𝑛
𝜎 𝜎
𝜇 ≤ 𝑥̅ + (1.96) and 𝑥̅ − (1.96) ≤ 𝜇
√𝑛 √𝑛

Also we can write combining the two results in the form,

𝜎 𝜎
𝑥̅ − 1.96 ( 𝑛) ≤ 𝜇 ≤ 𝑥̅ + 1.96 ( 𝑛) --- (1)
√ √

57
 Similarly, from the table of normal areas 99% of the area lies between -2.58 and 2.58.
This is equivalent to the form,
𝜎 𝜎
𝑥̅ − 2.58 ( 𝑛) ≤ 𝜇 ≤ 𝑥̅ + 2.58 ( 𝑛) --- (2)
√ √

We can say that (1) is the 95% confidence interval and (2) is the 99% confidence interval.

One tailed and two tailed tests

In our test of acceptance or non acceptance of a hypothesis, we concentrated on the value
of 𝑧 on both sides of the mean. This can be categorically stated that, the focus of the attention
lies in the two “tails” of the distribution and hence such a test is called a two tailed test.
Sometimes we will be interested in the extreme values to only one side of the mean in
which the region of significance will be a region to one side of the distribution. Obviously the
area of such a region will be equal to the level of significance itself. Such a tail is called a one
tailed test.

One tailed test Two tailed test

The following table will be useful for working problems

Critical Value of 𝒛
Test
5% level 1% level

One-tailed test -1.645 or 1.645 -2.33 or 2.33

58
 Two-tailed test -1.96 or 1.96 -2.58 or 2.58

Tests of significance of Proportions

1. Let 𝑥 be the observed number of successes in a sample size of 𝑛 and 𝜇 = 𝑛𝑝 be the
expected number of successes let the associated standard normal variate 𝑍 be defined by 𝑧 =
𝑥−𝜇 𝑥−𝑛𝑝
=
𝜎 √𝑛𝑝𝑞

2. If |𝑍| > 2.58, we conclude that the differences is highly significant and reject the
𝑝𝑞
hypothesis .Probable limits = 𝑝 ± 2.58√ 𝑛

3. If Z < 1.96, difference between the observed and expected number of successes is not
significant.

4. If Z >1.96, difference is significant at 5% level of significance.

Test of hypothesis for means

Let 𝜇1 and 𝜇2 be the mean of two populations. Let (𝑥̅1 , 𝜎1 ); (𝑥̅2 , 𝜎2 ) be the mean and
standard deviation of two large samples of size 𝑛1 and 𝑛2 respectively. To test the null
hypothesis 𝐻0 that there is no difference between the population means. i.e., 𝐻0 : 𝜇1 = 𝜇2
(𝒙 ̅𝟐 )
̅𝟏 −𝒙
The static for this is given by 𝒁 =
𝝈 𝟐𝝈 𝟐
√ 𝟏+ 𝟐
𝒏𝟏 𝒏𝟐

Test of significance of proportions

1.A coin is tossed 1000 times and head turns up 540 times. Decide on the hypothesis that the
coin in unbiased.
Sol: Let us suppose that the coin is unbiased
p = Probability of getting a head in one toss = ½
Since p + q = 1, q= ½
Expected number of heads in 1000 tosses = np
= 1000 × ½ = 500
Actual Number of heads = 540 = 𝑥
Now 𝑥 − np = 540 − 500 = 40
𝑥−𝜇 𝑥−𝑛𝑝 540−500
Consider z = = =
√𝑛𝑝𝑞 √𝑛𝑝𝑞 √1000𝑋0.5𝑋0.5

59
 𝑧 =2.53<2.58
Thus, we can say that the coin is unbiased.
2. A sample of 900 days was taken in a coastal town and it was found that on 100 days the
weather was very hot. Obtain the probable limits of the percentage of very hot weather.
100 1
Sol: Probability of very hot weather , 𝑝 = 900 = 9
8
∴ 𝑞=9 , 𝑛 = 900

𝑝𝑞
Probable limits = 𝑝 ± (2.58)√ 𝑛

1 1 8 1
= 9 ± (2.58)√9 . 9 . 900

= 0.111 ± 0.027
= 0.084 𝑎𝑛𝑑 0.138
Probability limits of very hot weather is 8.4% to 13.8%
3. In a sample of 500 men it was found that 60% of them had over weight. What can we
infer about the proportion of people having over weight in the population?
60
Sol: Probability of persons having over weight is , 𝑝 = 100 = 0.6

∴ 𝑞 = 1 − 𝑝 = 0.4 , 𝑛 = 500

𝑝𝑞
Probable limits = 𝑝 ± (2.58)√ 𝑛

(0.6)(0.4)
= 0.6 ± (2.58)√ 500

= 0.6 ± 0.0565
Probable limits = 0. 5435 𝑎𝑛𝑑 0.6565
Thus, the probable limits of people having over weight is 54.35% and 65.65%
4. A survey was conducted in a slum locality of 2000 families by selecting a sample size
800. It was revealed that 180 families were illiterates. Find the probable limits of the
illiterate families in the population of 2000.
180
Sol: Probability of illiterate families is , 𝑝 = 800 = 0.225

∴ 𝑞 = 1 − 𝑝 = 0.775 , 𝑛 = 800

𝑝𝑞
Probable limits = 𝑝 ± (2.58)√ 𝑛

60
 (0.225)(0.775)
= 0.225 ± (2.58)√ 800

= 0.225 ± 0.038
Probable limits = 0. 187 𝑎𝑛𝑑 0.263
The probable limits of the illiterate families in the population of 2000
= 0. 187 ∗ 2000 𝑎𝑛𝑑 0.263 ∗ 2000
= 374 𝑎𝑛𝑑 526
Thus, 374 to 526 are probably illiterate families.
5. In 324 throws of a six faced ‘die’, an odd number turned up 181 times. Is it reasonable to
think that the ‘die’ is an unbiased one?
Sol: Probability of the turn up of an odd number , p = 3/6= ½
Since p + q = 1, q= ½
Expected number of successes = np
= 324 × ½ = 162
Actual Number of heads = 181 = 𝑥
Now 𝑥 − np = 181 − 162 = 19
𝑥−𝜇 𝑥−𝑛𝑝 19 19
Consider z = = = =
√ 𝑛𝑝𝑞 √𝑛𝑝𝑞 √324𝑋0.5𝑋0.5 9

𝑧 =2.11<2.58
Thus, we can say that the die is unbiased.

6. A sample of 100 days is taken from meteorological records of a certain district and 10 of
them are found to be foggy. What are the probable limits of the percentage of foggy days
in the district.
10
Sol: Proportion of foggy days in a sample of 100 days is , 𝑝 = 100 = 0.1

∴ 𝑞 = 1 − 𝑝 = 0.9 , 𝑛 = 100

𝑝𝑞
Probable limits = 𝑝 ± (2.58)√ 𝑛

(0.1)(0.9)
= 0.1 ± (2.58)√ 100

= 0.1 ± 0.0774
Probable limits = 0. 0226 𝑎𝑛𝑑 0.1774

61
Thus, the percentage of foggy days lies between 2.26% and 17.74%
7. A random sample of 500 apples was taken from a large consignment and 65 were found
to be bad. Estimate the proportion of bad apples in the consignment as well as the
standard error of the estimate. Also find the percentage of bad apples in the consignment.
65
Sol: Proportion of bad apples in the sample is given by , 𝑝 = 500 = 0.13

∴ 𝑞 = 1 − 𝑝 = 0.87 , 𝑛 = 500

𝑝𝑞 (0.13)(0.87)
S.E proportion of bad apples = √ 𝑛 = √ = 0.015
500

𝑝𝑞
Probable limits = 𝑝 ± (2.58)√
𝑛

= 0.13 ± (2.58)(0.015)
= 0.13 ± 0.0387
Probable limits= 0.0913 𝑎𝑛𝑑 0.1687
= 9.13% 𝑎𝑛𝑑 16.87%
Thus, the required percentage of bad apples in the consignment lies between 9.13 and 16.87.
8. The mean and standard deviation of the maximum loads supported by 60 cables are 11.09
tonnes and 0.73 tonnes respectively. Find a) 95% b) 99% confidence limits for mean of
the maximum loads of all cables produced by the company.
Sol: By data 𝑥̅ = 11.09 , 𝜎 = 0.73 , n = 60
a) 95% confidence limits for mean of the maximum loads are given by
𝜎 0.73
𝑥̅ ± 1.96 ( 𝑛) = 11.09 ± (1.96) ( )
√ √60

= 11.09 ± 0.18
= 10.91 𝑎𝑛𝑑 11.27
Thus,10.91 tonnes to 11.27 tonnes are the 95% confidence limits for the mean of the
maximum loads.
b) 99% confidence limits for mean of the maximum loads are given by
𝜎 0.73
𝑥̅ ± 2.58 ( 𝑛) = 11.09 ± (2.58) ( )
√ √60

= 11.09 ± 0.24
= 10.85 𝑎𝑛𝑑 11.33
Thus, 10.85 tonnes to 11.33 tonnes are the 95% confidence limits for the mean of the
maximum loads.

62
9. An unbiased coin is thrown n times. It is desired that the relative frequency of the
appearance of heads should lie between 0.49 and 0.51. Find the smallest value of n that
will ensure this result with a) 95% confidence b) 99% confidence
1
Sol: Probability of getting a head, 𝑝 = 2
1
𝑞=2

𝑝𝑞
S.E proportion of heads = √ 𝑛

11
. 1
S.E proportion of heads = √2 2 = 𝑛 2√𝑛

𝑝𝑞
a) Probable limits for 95% confidence level is given by : 𝑝 ± (1.96)√ 𝑛 , which should lie

between 0.51 and 0.49

𝑝𝑞
𝑝 ± (1.96)√ 𝑛 = 0.51 𝑜𝑟 0.49

1
0.5 ± (1.96) 2 = 0.51 𝑜𝑟 0.49
√𝑛
1.96
0.5 ± 2 = 0.51 𝑜𝑟 0.49
√𝑛
1.96 1.96
0.5 + 2 = 0.51 0.5 − 2 = 0.49
√ 𝑛 √𝑛
1.96 1.96
= 0.51 − 0.5 = 0.5 − 0.49
2√𝑛 2√𝑛

1.96 1.96
= 0.01 = 0.01
2√𝑛 2 √𝑛

1.96
√𝑛 = 0.02

√𝑛 = 98
Thus, 𝐧 = 𝟗𝟔𝟎𝟒
b) Taking the confidence coefficient equal to 1.645 for 90% confidence level we have as
before
1.645
= 0.01
2√𝑛
1.645
√𝑛 =
0.02
√𝑛 = 82.25
Thus, 𝒏 = 𝟔𝟕𝟔𝟓

63
Test of significance of a sample mean
1. A manufacturer claimed that atleast 95% of the equipment which he supplied to a factory
conformed to specifications. An examination of a sample of 200 pieces of equipment
revealed that 18 of them were faulty. Test his claim at a significance level of 1% and 5%.
Sol: Let 𝑝 be the probability of success which being the probability of the equipment
supplied to the factory conformal to the specifications.
∴ By data 𝑝 = 0.95 and hence 𝑞 = 0.05
𝐻0 : 𝑝 = 0.95 and the claim is correct.
𝐻1 : 𝑝 < 0.95 and the claim is false.
We choose the one tailed test to determine whether the supply is conformal to the
specification.
𝜇 = 𝑛𝑝 = 200 ∗ 0.95 = 190

𝜎 = √𝑛𝑝𝑞 = √200 ∗ 0.95 ∗ 0.05 = 3.082

Expected number of equipment's according to the specification =190

Actual number =182 , since 18 out of 200 were faulty.
𝑥−𝑛𝑝
Now, 𝑍 = 𝜎
190−182
𝑍= 3.082

𝒁 = 𝟐. 𝟔
The value of 𝑍 is greater than the critical value 1.645 at 5% level and 2.33 at 1% level of
significance.
The claim of the manufacturer (null hypothesis that claim is correct) is rejected at 5% as
well as 1% level of significance in accordance with the one tailed test.
2. It has been found from experience that the mean breaking strength of a particular brand
of thread is 275.6gms with standard deviation of 39.7gms. Recently a sample of 36 pieces
of thread showed a mean breaking strength 0f 253.2gms. Can one conclude at a
significance level of a) 0.05 b) 0.01 that the thread has become inferior ?
Sol: We have to decide between the two hypothesis,
𝐻0 : 𝜇 = 275.6 𝑔𝑚𝑠 , mean breaking strength.
𝐻1 : 𝑝 < 275.6 𝑔𝑚𝑠 , inferior in breaking strength.
We choose the one tailed test.
Mean breaking strength of a sample of 36 pieces=253.2

64
∴ Difference =275.6 – 253.2=22.4 ; n = 36
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
Now, 𝑍 =
(𝜎⁄ )
√𝑛

22.4
𝑍=
(39.7⁄ )
√36
𝑍 = 3.38
The value of Z is greater than the critical value of Z=1.645 at 5% level and 2.33 at 1% level
of significance.
Under the hypothesis 𝐻1 that the thread has become inferior is accepted at both 0.05 and 0.01
levels in accordance with one tailed test.

Test of significance of difference between means

1.In an elementary school examination the mean grade of 32 boys was 72 with a standard
deviation of 8, while the mean grade of 36 girls was 75 with a standard deviation of 6.
Test the hypothesis that the performance of girls is better than boys.
Sol: We have , 𝑥̅𝐵 = 72 , 𝜎𝐵 = 8, 𝑛𝐵 = 32 [Boys]
𝑥̅𝐺 = 75 , 𝜎𝐺 = 6, 𝑛𝐺 = 36 [Girls]
𝑥̅ 𝐺 −𝑥̅ 𝐵
Consider, 𝑍 =
𝝈 𝟐
𝝈 𝟐
√ 𝑮+ 𝑩
𝒏𝑮 𝒏𝑩

75−72
𝑍= 𝟑𝟔 𝟔𝟒
√ +
𝟑𝟔 𝟑𝟐

𝑍 = √3
𝑍 = 1.73
> 𝑍0.05 = 1.645 (𝑜𝑛𝑒 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
𝑍 = 1.73 {
< 𝑍0.01 = 2.33 (𝑜𝑛𝑒 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
The difference in the performance of girls and boys in the examination is significant at 5%
level but not at 1% level.
2. A sample of 100 bulbs produced by a company A showed a mean life of 1190 hours and a
standard deviation of 90 hours. Also a sample of 75 bulbs produced by a company B
showed a mean life of 1230 hours and a standard deviation of 120 hours. Is there a
difference between the mean life time of the bulbs produced by the two companies at
a) 5% level of significance b) 1% level of significance.
Sol: By data 𝑥̅𝐴 = 1190 , 𝜎𝐴 = 90, 𝑛𝐴 = 100 [Company A]

65
 𝑥̅𝐵 = 1230 , 𝜎𝐵 = 120, 𝑛𝐵 = 75 [Company B]
𝑥̅ 𝐵 −𝑥̅ 𝐴
Consider, 𝑍 =
𝝈 𝟐
𝝈 𝟐
√ 𝑩+ 𝑨
𝒏𝑩 𝒏𝑨

1230 − 1190
𝑍=
𝟐 𝟐
√(𝟏𝟐𝟎) + (𝟗𝟎)
𝟕𝟓 𝟏𝟎𝟎
𝑍 = 2.42
> 𝑍0.05 = 1.96 (𝑇𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
𝑍 = 2.42 {
< 𝑍0.01 = 2.58 (𝑇𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
The null hypothesis that there is no difference between the mean life time of bulbs is rejected
at 5% level but not at 1% level of significance.
The null hypothesis that there is no difference between the mean life time of bulbs is rejected
at both levels of significance in a one tailed test as the respective critical values are 1.645 and
2.33.
3. The mean life of two large samples of 1000 and 2000 members are 168.75 cms and 170
cms respectively. Can this be regarded as drawn from the sample population of standard
deviation 6.25 cms ?
Sol: By data 𝑥̅1 = 168.75 , 𝑛1 = 1000
𝑥̅2 = 170 , 𝑛2 = 2000 , 𝜎 = 6.25
𝑥̅ 2 −𝑥̅ 1
Consider, 𝑍 = 𝟏 𝟏
𝜎√ +
𝒏𝟏 𝒏𝟐

170−168.75
𝑍= 𝟏 𝟏
6.25√ +
𝟏𝟎𝟎𝟎 𝟐𝟎𝟎𝟎

𝑍 = 5.16
𝑍 = 5.16 is very much greater than 𝑍0.05 = 1.96 and 𝑍0.01 = 2.58. Thus we say that the
difference between the sample means is significant and we conclude that the samples cannot
be regarded as drawn from the same population.
4. A random sample for 1000 workers in company has mean wage of Rs.50 per day and S.D
of Rs.15.Another sample of 1500 workers from another company has mean wage of Rs.45
per day and S.D of Rs.20.Does the mean rate of wages varies between the two companies?
Find the 95% confidence limits for the difference of the wages of the population of the two
companies?
Sol: By data 𝑥̅1 = 50 , 𝜎1 = 15, 𝑛1 = 1000 [Company 1]
𝑥̅2 = 45 , 𝜎2 = 20, 𝑛2 = 1500 [Company 2]

66
 𝑥̅ 1 −𝑥̅ 2
Consider, 𝑍 =
𝝈 𝟐𝝈 𝟐
√ 𝟏+ 𝟐
𝒏𝟏 𝒏𝟐

50−45
𝑍= 𝟐 𝟐
√ 𝟏𝟓 + 𝟐𝟎
𝟏𝟎𝟎𝟎 𝟏𝟓𝟎𝟎

𝑍 = 7.1307
The value of 𝑍 = 7.1307 is greater than 𝑍0.05 = 1.96 and 𝑍0.01 = 2.58. Hence we can say
that the difference between the mean wages is significant both at 5% and 1% level of
significance.
Also 95% confidence limits for the difference of mean wages is given by

𝜎2 𝜎2
(𝑥̅1 − 𝑥̅2 ) ± 1.96√𝑛1 + 𝑛2 = 5 ± 1.96(0.7012)
1 2

= 5 ± 1.374
= 3.626 𝑎𝑛𝑑 6.374
Thus we can say with 95% confidence that the difference of population mean of wages
between the two companies lies between Rs.3.63 and Rs.6.37.

Student’s t distribution/test
We need to test the hypothesis, whether the sample mean x̅ differs significantly from the
population mean or hypothetical value µ.
x̅−μ
We compute, t = s
√n and consider |t|.
∑x 1
Here , x̅ = and s2 = (n−1) ∑ni=1(xi − x̅)2
n

v = n − 1 , denote the number of degrees of freedom ‘t’.

If |t| > t 0.05 , the difference between x̅ and μ is said to be significant at 5% level of
significance.
If |t| > t 0.01 , the difference between x̅ and μ is said to be significant at 1% level of
significance.
If |t| is less than the table value at a certain level of significance, the data is said to be
conformal/consistent with the hypothesis that μ is the mean of the population.
𝒔
̅±
We have 95% confidence limits for 𝜇 is given by 𝒙 𝒕
√𝒏 𝟎.𝟎𝟓

67
 𝒔
̅±
Similarly 99% confidence limits for 𝜇 is given by 𝒙 𝒕
√𝒏 𝟎.𝟎𝟏

Note: 1. Confidence limits are also called Fiducial limits.

2. Test of significance for difference between sample means is
𝑥̅ −𝑦̅
𝑡= 1 1
𝑠.√ +
𝑛1 𝑛2

1 𝑛1 2
Where, 𝑠 2 = 𝑛 {∑𝑖=1(𝑥𝑖 − 𝑥̅ )2 + ∑𝑛𝑗=1
2
(𝑦𝑗 − 𝑦̅) }
1 +𝑛2 −2

And degrees of freedom, 𝑣 = 𝑛1 + 𝑛2 − 2

Problems
1. Find the student’s t for the following variable values in a sample of eight:
-4, -2, -2, 0, 2, 2, 3, 3 taking the mean of the universe to be zero.
x̅−μ
Sol: Wkt t = s
√n

By data μ = 0 and we have n = 8

∑x
x̅ =
n
1
x̅ = (−4 − 2 − 2 + 0 + 2 + 2 + 3 + 3)
8
1
x̅ = = 0.25
4
𝑛
2
1
𝑠 = ∑(𝑥𝑖 − 𝑥̅ )2
(𝑛 − 1)
𝑖=1

1
𝑠2 = {(−4.25)2 + (−2.25)2 + (−2.25)2 + (−0.25)2 + (1.75)2 + (1.75)2 + (2.75)2 + (2.75)2 }
(8 − 1)

1
𝑠2 = (49.5)
(7)
𝑠 2 = 7.07
∴ 𝒔 = 𝟐. 𝟔𝟔
0.25−0
Thus, 𝑡 = √8
2.66

𝒕 = 𝟎. 𝟐𝟔𝟔
2. A machine is expected to produce nails of length 3 inches. A random sample of 25 nails
gave an average length of 3.1 inch with standard deviation 0.3. Can it be said that the
machine is producing nails as per specification ? (𝑡0.05 𝑓𝑜𝑟 24 𝑑. 𝑓 𝑖𝑠 2.064)
Sol: By data , we have 𝜇 = 3 , 𝑥̅ = 3.1, 𝑛 = 25, 𝑠 = 0.3

68
 𝑥̅ −𝜇
𝑡= √𝑛
𝑠
3.1−3
𝑡= √25
0.3

𝒕 = 𝟏. 𝟔𝟕 < 𝟐. 𝟎𝟔𝟒
Thus, the hypothesis that the machine is producing nails as per specification is accepted at
5% level of significance.
3. Ten individuals are chosen at random from a population and their heights in inches are
found to be 63, 63, 66, 67, 68, 69, 70, 70, 71, 71. Test the hypothesis that the mean height
of the universe is 66 inches.(𝑡0.05 = 2.262 𝑓𝑜𝑟 9 𝑑. 𝑓).
Sol: We have 𝜇 = 66, 𝑛 = 10
∑𝑥
𝑥̅ =
𝑛
678
𝑥̅ =
10
̅ = 𝟔𝟕. 𝟖
𝒙
𝑛
2
1
𝑠 = ∑(𝑥𝑖 − 𝑥̅ )2
(𝑛 − 1)
𝑖=1

1
𝑠 2 = [(63 − 67.8)2 + (63 − 67.8)2 + ⋯ + (71 − 67.8)2 ]
9
𝑠 2 = 9.067
𝒔 = 𝟑. 𝟎𝟏𝟏
𝑥̅ −𝜇
Wkt 𝑡 = √𝑛
𝑠
(67.8−66)
𝑡= √10
3.011

𝒕 = 𝟏. 𝟖𝟗 < 𝟐. 𝟐𝟔𝟐
Thus the hypothesis is accepted at 5% level of significance.
4. A sample of 10 measurements of the diameter of a sphere gave a mean of 12 cm and a
standard deviation 0.15 cm. Find the 95% confidence limits for the actual diameter.
Sol: By data, 𝑛 = 10 , 𝑥̅ = 12, 𝑠 = 0.15
Also 𝑡0.05 = 2.262 𝑓𝑜𝑟 9 𝑑. 𝑓
𝒔
̅ ± [ ] 𝒕𝟎.𝟎𝟓
Confidence limits for the actual diameter is given by 𝒙 𝒏 √

69
 𝑠 0.15
𝑥̅ ± [ ] 𝑡0.05 = 12 + [ ] (2.262) = 12 ± 0.1073
√𝑛 √10
Thus 11.893 cm to 12.107 cm is the confidence limits for the actual diameter.
5. A certain stimulus administered to each of the 12 patients resulted in the following
change in blood pressure 5, 2, 8, -1,3,0, 6, -2, 1, 5, 0, 4. Can it be concluded that the
stimulus will increase the blood pressure ? (𝑡0.05 𝑓𝑜𝑟 11 𝑑. 𝑓 𝑖𝑠 2.201)
∑𝑥
Sol: 𝑥̅ = 𝑛
31
𝑥̅ = 12

̅ = 𝟐. 𝟓𝟖𝟑𝟑
𝒙
𝑛
1
𝑠2 = ∑(𝑥𝑖 − 𝑥̅ )2
(𝑛 − 1)
𝑖=1

1
𝑠2 = [(5 − 2.58)2 + (2 − 2.58)2 + ⋯ + (4 − 2.58)2 ]
11
𝑠 2 = 9.538
𝒔 = 𝟑. 𝟎𝟖𝟖
𝑥̅ −𝜇
Wkt 𝑡 = √𝑛
𝑠

Let us suppose that the stimulus administration is not accompanied with increase in blood
pressure, we can take 𝜇 = 0.
(2.5833−0)
𝑡= √12
3.088

𝒕 = 𝟐. 𝟖𝟗𝟕𝟗 > 𝟐. 𝟐𝟎𝟏

Hence the hypothesis is rejected at 5% level of significance. We conclude with 95% confidence
that the stimulus in general is accompanied with increase in blood pressure.
6. A group of boys and girls were given an intelligence test. The mean score, S.D score and
numbers in each group are as follows.

Boys Girls

Mean 74 70

SD 8 10

n 12 10

70
Is the difference between the means of the two groups significant at 5% level of significance
(𝑡0.05 𝑓𝑜𝑟 20 𝑑. 𝑓 𝑖𝑠 2.086).
Sol: By data , 𝑥̅ = 74, 𝑠1 = 8, 𝑛1 = 12 [Boys]
𝑦̅ = 70, 𝑠2 = 10, 𝑛2 = 10 [Girls]
𝑥̅ −𝑦̅
We have 𝑡 = 1 1
𝑠.√ +
𝑛1 𝑛2

1 𝑛1 2
Where, 𝑠 2 = 𝑛 {∑𝑖=1(𝑥𝑖 − 𝑥̅ )2 + ∑𝑛𝑗=1
2
(𝑦𝑗 − 𝑦̅) }
1 +𝑛2 −2

𝑛1 𝑠12 +𝑛2 𝑠12

Or 𝑠 2 = 𝑛1 +𝑛2 −2

12(64)+10(100)
Now 𝑠 2 = 12+10−2

1768
𝑠2 = 20

𝑠 2 = 88.4
𝑠 = 9.402
74−70
Hence 𝑡 = 1 1
(9.402)√ +
12 10

𝒕 = 𝟎. 𝟗𝟗𝟑𝟔 < 𝟐. 𝟎𝟖𝟔

Thus the hypothesis that there is a difference between the means of the two groups is
accepted at 5% level of significance.
7. A sample of 11 rats from a central population had an average blood viscosity of 3.92 with
standard deviation of 0.61. On the basis of this sample, establish 95% fiducial limits for
𝜇, the mean blood viscosity of the central population. (𝑡0.05 = 2.228 𝑓𝑜𝑟 10 𝑑. 𝑓).
Sol: By data 𝑥̅ = 3.92, 𝑠 = 0.61, 𝑛1 = 11
𝒔
̅ ± [ ] 𝒕𝟎.𝟎𝟓
95% fiducial limits for 𝜇 = 𝒙 𝒏 √

𝑠 0.61
𝑥̅ ± [ 𝑛] 𝑡0.05 = 3.92 ± [ ] (2.228)
√ √11

= 3.92 ± 0.41
= 3.51 𝑎𝑛𝑑 4.33
Thus 95% confidence limits for 𝜇 are 3.51 𝑎𝑛𝑑 4.33.
8. Two types of batteries are tested for their length of life and the following results were
obtained.

71
 Battery A : 𝑛1 = 10, 𝑥̅1 = 500 ℎ𝑟𝑠, 𝑠12 = 100
Battery B : 𝑛2 = 10, 𝑥̅2 = 500 ℎ𝑟𝑠, 𝑠12 = 121
Compute student’s t and test whether there is a significant difference in the two means.
Sol: By data Battery A : 𝑛1 = 10, 𝑥̅1 = 500 ℎ𝑟𝑠, 𝑠12 = 100
Battery B : 𝑛2 = 10, 𝑥̅2 = 560 ℎ𝑟𝑠, 𝑠12 = 121
𝑛1 𝑠12 +𝑛2 𝑠12
Now, 𝑠 2 = 𝑛1 +𝑛2 −2

(10∗100) +(10∗121)
𝑠2 = 10+10−2

𝑠 2 = 122.78
𝒔 = 𝟏𝟏. 𝟎𝟖𝟎𝟔
𝑥̅ 2 −𝑥̅ 1
We have 𝑡 = 1 1
𝑠.√ +
𝑛1 𝑛2

60
𝑡= 1 1
(11.0806)√ +
10 10

𝑡 = 12.1081 > 2.101 (𝑡0.05 = 2.101 𝑓𝑜𝑟 18 𝑑. 𝑓)

This value of t is greater then the table value of t for 18 d.f at all levels of significance.
The null hypothesis that there is no significant difference in the two means rejected at all
significance levels.

9. A group of 10 boys fed on a diet A and another group of 8 boys fed on a different diet B
for a period of 6 months recorded the following increase in weight (lbs.)
Diet A : 5 6 8 1 12 4 3 9 6 10
Diet B : 2 3 6 8 10 1 2 8
Test whether diets A and B differ significantly regarding their effect on increase in weight.
Sol: Let the variable 𝑥 correspond to the diet A and 𝑦 to the diet B.
∑𝑥 ∑𝑦
𝑥̅ = 𝑦̅ =
𝑛1 𝑛2

64 40
𝑥̅ = 10 𝑦̅ = 8

̅ = 𝟔. 𝟒
𝒙 ̅=𝟓
𝒚
2
∑𝑛𝑖=1
1
(𝑥𝑖 − 𝑥̅ )2 = 102.4 ; ∑𝑛𝑗=1
2
(𝑦𝑗 − 𝑦̅) = 82

72
 𝑛1 𝑛2
1 2
𝑠2 = {∑(𝑥𝑖 − 𝑥̅ )2 + ∑(𝑦𝑗 − 𝑦̅) }
𝑛1 + 𝑛2 − 2
𝑖=1 𝑗=1

1
𝑠2 = {102.4 + 82}
16
184.4
𝑠2 = = 11.525
16
𝒔 = 𝟑. 𝟑𝟗𝟓
𝑥̅ −𝑦̅
We have 𝑡 = 1 1
𝑠.√ +
𝑛1 𝑛2

1.4
𝑡=
1 1
(3.395)√10 + 8

𝑡 = 0.869 < 2.12 for 16 d.f

Thus we conclude that the two diets do not differ significantly regarding their effect on
increase in weight.
10. Two horses A and B were tested according to the time (in seconds ) to run a particular
race with the following results.
Horse A : 28 30 32 33 33 29 34
Horse B : 29 30 30 24 27 29
Test whether you can discriminate between the two horses.
Sol: Let the variables 𝑥 and 𝑦 respectively correspond to horse A and horse B.
∑𝑥 219 ∑𝑦 169
𝑥̅ = = = 31.3 𝑦̅ = = = 28.2
𝑛1 7 𝑛2 6

2
∑𝑛𝑖=1
1
(𝑥𝑖 − 𝑥̅ )2 = 31.43 ∑𝑛𝑗=1
2
(𝑦𝑗 − 𝑦̅) = 26.84
1 𝑛1 2
𝑠2 = 𝑛 {∑𝑖=1(𝑥𝑖 − 𝑥̅ )2 + ∑𝑛𝑗=1
2
(𝑦𝑗 − 𝑦̅) }
1 +𝑛2 −2

1
𝑠2 = (31.43 + 26.84)
11
𝑠 2 = 5.2973
∴ 𝒔 = 𝟐. 𝟑𝟎𝟏𝟔
𝑥̅ − 𝑦̅
We have 𝑡 = 1 1
𝑠.√ +
𝑛1 𝑛2

31.3−28.2
𝑡= 1 1
(2.3016)√ +
7 6

73
 > 𝑡0.05 = 2.2
𝑡 = 2.42 = {
< 𝑡0.02 = 2.72
The discrimination between the horses is significant at 5% level but not at 2% level of
significance.

Chi-Square distribution
Chi-square distribution provides a measure of correspondence between the theoretical
frequencies and observed frequencies.
If 𝑂𝑖 (𝑖 = 1,2, … , 𝑛) and 𝐸𝑖 (𝑖 = 1,2, … , 𝑛) respectively denotes a set of observed and
estimated frequencies, the quantity chi-square denoted by 𝜒 2 is defined as follows.
(𝑶𝒊 −𝑬𝒊 )𝟐
𝝌𝟐 = ∑𝒏𝒊=𝟏 ; degrees of freedom = 𝑛 − 1
𝑬𝒊

Note: If the expected frequencies are less than 10, we group them suitably for computing
the value of chi. Square.
Chi-square test as a test of goodness of fit
It is possible to test the hypothesis about the association of two attributes. We already
discussed the fitting of Binomial distribution, Normal distribution, Poisson distribution to a
given data. It is easily possible to find the theoretical frequencies from the distribution of fit.
Chi-square test helps us to test the goodness of fit of these distributions.
If the calculated value of 𝝌𝟐 is less than the value of 𝝌𝟐 at a specified level of significance,
the hypothesis is accepted. Otherwise the hypothesis is rejected.

Problems
1.A dice is thrown 264 times and the number appearing on the face (𝒙) follows the following
frequency distribution.

𝒙 1 2 3 4 5 6

𝑓 40 32 28 58 54 60

Calculate the value of 𝝌𝟐 .

Sol: The frequencies in the given data are the observed frequencies. Assuming that the dice
is unbiased, the expected number of frequencies for the numbers 1, 2, 3, 4, 5, 6 to
264
appear on the face is = 44 𝑒𝑎𝑐ℎ.
6

74
Now the data is as follows:

No. on the dice 1 2 3 4 5 6

Observed frequency (𝑶𝒊 ) 40 32 28 58 54 60

Expected frequency (𝑬𝒊 ) 44 44 44 44 44 44

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏

𝝌𝟐 = ∑ 𝑬𝒊
𝒊=𝟏

2
(40 − 44)2 (32 − 44)2 (60 − 44)2
𝜒 = + + ⋯+
44 44 44
1
𝜒2 = [16 + 144 + 256 + 196 + 100 + 256]
44
968
𝜒2 =
44
Thus, 𝝌𝟐 = 𝟐𝟐
2. Five dice were thrown 96 times and the numbers 1, 2 or 3 appearing on the face of the
dice follows the frequency distribution as below.

No. of dice showing 1,2 or 3 5 4 3 2 1 0

Frequency 7 19 35 24 8 3

Test the hypothesis that the data follows a binomial distribution. (𝝌20.05 = 11.07 𝑓𝑜𝑟 5 𝑑. 𝑓)

Sol: The data gives the observed frequencies and we need to calculate the expected
frequencies.
3 1
Probability of a single dice throwing 1, 2 or 3 is 𝑝 = 6 = 2
1
∴𝑞 =1−𝑝 =2

From the binomial distribution, 𝑃(𝑥) = 𝑛𝐶𝑥 𝑝 𝑥 𝑞 (𝑛−𝑥)

The theoretical frequencies of getting 5, 4, 3, 2, 1, 0 successes with 5 dice are respectively
the successive terms of the binomial expansion.
1 1 1 1
They are respectively 96 ∗ 5𝐶0 ∗ 25 , 96 ∗ 5𝐶1 ∗ 25 , 96 ∗ 5𝐶2 ∗ 25, 96 ∗ 5𝐶3 ∗ 25, 96 ∗
5 1 1
𝐶4 ∗ 25 , 96 ∗ 5𝐶5 ∗ 25 and they are 3,15,30,30,15,3.

75
We have the table of observed and expected frequencies.

𝑂𝑖 7 19 35 24 8 3

𝐸𝑖 3 15 30 30 15 3

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏

𝝌𝟐 = ∑ 𝑬𝒊
𝒊=𝟏

16 16 25 36 49 0
𝜒2 = + + + + +
3 15 30 30 15 3
𝜒 2 = 11.7 > 𝝌20.05 = 11.07
Thus the hypothesis that the data follows a binomial distribution is rejected.
3. A sample analysis of examination results of 500 students was made. It was found that 220
students had failed, 170 had secured third class 90 had secured second class and 20 had
secured first class. Do these figures support the general examination result which is in the
ratio 4:3:2:1 for the respective categories (𝝌20.05 = 7.81 𝑓𝑜𝑟 3 𝑑. 𝑓).
Sol: Let us take the hypothesis that these figures support to the general result in the ratio
4:3:2:1
The expected frequencies in the respective category are
4 3 2 1
∗ 500, 10 ∗ 500, 10 ∗ 500, 10 ∗ 500 or 200, 150, 100, 50
10

We have the following table.

𝑂𝑖 220 170 90 20

𝐸𝑖 200 150 100 50

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏
𝟐
𝝌 =∑ 𝑬𝒊
𝒊=𝟏

400 400 100 900

𝜒2 = + + +
200 150 100 50
2
𝜒 2 = 23.67 > 𝜒0.05 = 7.81
Thus the hypothesis is rejected.

76
4. 4 coins are tossed 100 times and the following results were obtained. Fit a binomial
distribution for the data and calculate the theoretical frequencies(𝝌20.05 = 9.49 𝑓𝑜𝑟 4 𝑑. 𝑓).

Number 0 1 2 3 4
of heads

Frequency 5 29 36 25 5

Sol: Let 𝑥 denote the number of heads and 𝑓 the corresponding frequency. Since the data is
in the form of a frequency distribution we shall first calculate the mean.
∑ 𝑓𝑥 (0)(5)+(1)(29)+(2)(36)+(3)(25)+(4)(5) 196
Mean, 𝜇 = ∑𝑓
= = 100 = 1.96
5+29+36+25+5

From binomial distribution, 𝜇 = 𝑛𝑝. Here n=4

Hence 4𝑝 = 1.96
𝒑 = 𝟎. 𝟒𝟗
𝒒 = 𝟎. 𝟓𝟏
From binomial distribution,
𝑃(𝑥) = 𝑛𝑐𝑥 𝑝 𝑥 𝑞 𝑛−𝑥
𝑷(𝒙) = 𝟒𝒄𝒙 (𝟎. 𝟒𝟗)𝒙 (𝟎. 𝟓𝟏)𝟒−𝒙
Since 4 coins were tossed , expected(theoretical) frequencies are obtained from
F(𝑥) = 100𝑃(𝑥)
F(𝑥) = 100.4𝑐𝑥 (0.49)𝑥 (0.51)4−𝑥 where 𝑥 = 0,1,2,3,4.
F(0) = 100.4𝑐0 (0.49)0 (0.51)4 = 6.765 ≅7
F(1) = 100.4𝑐1 (0.49)1 (0.51)3 = 25.999 ≅ 26
F(2) = 100.4𝑐2 (0.49)2 (0.51)2 = 37.47 ≅ 37
F(3) = 100.4𝑐3 (0.49)3 (0.51)1 = 24.0004 ≅ 24
F(4) = 100.4𝑐4 (0.49)4 (0.51)0 = 5.765 ≅ 6
Thus the required theoretical frequencies are 7,26,37,24,6.
We have the following table

𝑂𝑖 5 29 36 25 5

𝐸𝑖 7 26 37 24 6

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏

𝝌𝟐 = ∑ 𝑬𝒊
𝒊=𝟏

77
 4 9 1 1 1
𝜒2 = + + + +
7 26 37 24 6
2
𝜒 2 = 1.15 < 𝜒0.05 = 9.49
Thus the hypothesis that the fitness is good can be accepted.
5. Fit a poisson distribution for the following data and test the goodness of fit given that
𝝌20.05 = 7.815 𝑓𝑜𝑟 3 𝑑. 𝑓

𝑥 0 1 2 3 4

𝑓 122 60 15 2 1

Sol: Let 𝑥 denote the number of heads and 𝑓 the corresponding frequency. Since the data is
in the form of a frequency distribution we shall first calculate the mean.
∑ 𝑓𝑥 (0)(122)+(1)(60)+(2)(15)+(3)(2)+(4)(1) 100
Mean, 𝜇 = ∑𝑓
= = 200 = 0.2
122+60+15+2+1

From poisson distribution, 𝜇 = 𝑚.

∴ 𝒎 = 𝟎. 𝟐
𝑚𝑥 𝑒 −𝑚
Wkt, 𝑃(𝑥) = 𝑥!

Let 𝐹(𝑥) = 200 ∗ 𝑃(𝑥)

𝑚𝑥 𝑒 −𝑚
𝐹(𝑥) = 200 ∗ 𝑥!

(0.5)𝑥 𝑒 −(0.5)
𝐹(𝑥) = 200 ∗ But 𝑒 −(0.5) = 0.6065
𝑥!

(0.5)𝑥
𝐹(𝑥) = 121.3
𝑥!

𝐹(0) = 121 , 𝐹(1) = 61, 𝐹(2) = 15, 𝐹(3) = 3, 𝐹(4) = 0

Thus the required theoretical frequencies are 121,61,15,3,0.
Since the last of the expected frequency is 0 we shall club it with the previous one.
We have the following table

𝑂𝑖 122 60 15 2+1=3

𝐸𝑖 121 61 15 3+0=3

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏
𝟐
𝝌 =∑ 𝑬𝒊
𝒊=𝟏

1 1
𝜒 2 = 121 + 61 + 0 + 0

78
 2
𝜒 2 = 0.025 < 𝜒0.05 = 7.815
Thus the hypothesis that the fitness is good can be accepted.
6. The number of accidents per day (𝑥) as recorded in a textile industry over a period of 400
days is given below. Test the goodness of fit in respect of Poisson distribution of fit to
the given data(𝝌20.05 = 9.49 𝑓𝑜𝑟 4 𝑑. 𝑓).

x 0 1 2 3 4 5

f 173 168 37 18 3 1

Sol: Let 𝑥 denote the number of heads and 𝑓 the corresponding frequency. Since the data is
in the form of a frequency distribution we shall first calculate the mean.
∑ 𝑓𝑥 (0)(173)+(1)(168)+(2)(37)+(3)(18)+(4)(3)+(5)(1) 313
Mean, 𝜇 = ∑𝑓
= = 400 = 0.7825
173+168+37+18+3+1

From poisson distribution, 𝜇 = 𝑚.

∴ 𝒎 = 𝟎. 𝟕𝟖𝟐𝟓
𝑚𝑥 𝑒 −𝑚
Wkt , 𝑃(𝑥) = 𝑥!

Let 𝐹(𝑥) = 400 ∗ 𝑃(𝑥)

𝑚𝑥 𝑒 −𝑚
𝐹(𝑥) = 400 ∗ 𝑥!

(0.7825)𝑥 𝑒 −(0.7825)
𝐹(𝑥) = 400 ∗ But 𝑒 −(0.7825) = 0.4573
𝑥!

(0.7825)𝑥
𝐹(𝑥) = 182.92 𝑥!

𝐹(0) = 183 , 𝐹(1) = 143, 𝐹(2) = 56, 𝐹(3) = 15, 𝐹(4) = 3, 𝐹(5) = 0
Thus the required theoretical frequencies are 183,143,56,15,3,0.

Since the last of the expected frequency is 0 we shall club it with the previous one.
We have the following table

𝑶𝒊 173 168 37 18 3+1=4

𝐸𝑖 183 143 56 15 3+0=3

(𝑶𝒊 − 𝑬𝒊 )𝟐
𝒏
𝟐
𝝌 =∑ 𝑬𝒊
𝒊=𝟏

100 625 361 9 1

𝜒2 = + + + +
183 143 56 15 3

79
𝜒 2 = 12.297 > 𝝌20.05 = 9.49

Thus the hypothesis that the fitness is good can be rejected.

80
 MODULE-5
DESIGN OF EXPERIMENTS &ANOVA

ANALYSIS OF VARIANCE
The analysis of variance is a powerful statistical tool for tests of significance.

We consider the following types of ANOVA

I) One- way classification (One factor ANOVA) –Completely Randomized Design.

II) Two-Way classification (Two factors)-Randomized Block Design.
III) Latin Square Design-Three Way ANOVA

ONE - WAY ANOVA

One- way classification observations are classified according to one factor

WORKING PROCEDURE:
Null Hypothesis 𝐻0 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … …

Alternative Hypothesis 𝐻1 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … … ..

STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠

𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 + ⋯
𝑇2
Correction factor: 𝐶. 𝐹 = 𝑁

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 + ⋯ − 𝐶. 𝐹
(∑ 𝑋1 )2 (∑ 𝑋2 )2
𝑆𝑆𝐶 = 𝑛1
+ 𝑛2
+ ⋯ − 𝐶. 𝐹

𝑆𝑆𝐸 = 𝑇𝑆𝑆 − 𝑆𝑆𝐶

STEP3: O NE- WAY ANOVA TABLE

SOURCE OF SUM OF SQUARES DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION FREEDOM SQUARES

BETWEEN SSC = C–1= 𝑆𝑆𝐶 𝑀𝑆𝐸

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐶−1 𝑀𝑆𝐶

ERROR SSE = N–C= 𝑆𝑆𝐸

𝑀𝑆𝐸 =
𝑁−𝐶
TOTAL
NOTE: IF CALCULATED VALUE OF F < TABULATED VALUE OF F, THEN 𝐻0 IS ACCEPTED.

IF CALCULATED VALUE OF F > TABULAED VALUE OF F, THEN 𝐻0 𝐼𝑆 𝑅𝐸𝐽𝐸𝐶𝑇𝐸𝐷.

PROBLEM 1:

The following figures relate to production in kgs. of three variables A, B, C of wheat sown on 12 plots

A 14 16 18

B 14 13 15 22

C 18 16 19 19 22

Is there any significant difference in the production of the Varieties.

SOLUTION:

Null Hypothesis:

𝐻0 ∶ 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝑣𝑎𝑟𝑖𝑒𝑡𝑖𝑒𝑠.

Alternative Hypothesis:

𝑯 𝟏 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑣𝑎𝑟𝑖𝑒𝑡𝑖𝑒𝑠.

𝑋1 𝑋2 𝑋3 𝑋12 𝑋22 𝑋32

14 14 18 196 196 324

16 13 1 256 169 256

18 15 19 324 225 341

22 19 484 361

20 400

TOTAL ∑𝑋2 = 64 ∑𝑋3 = 92 ∑𝑋12 = 776 ∑𝑋22 = 1074 ∑𝑋32 = 1702

∑𝑋1 = 48
STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠 = 12
𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 = 48 + 64 + 92 = 204
𝑇2 (204)2
Correction factor: 𝐶. 𝐹 = 𝑁
=
12
= 3468

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 − 𝐶. 𝐹

= 776 + 1074 + 1702 − 3468 = 84

(∑ 𝑋1 )2 (∑ 𝑋2 )2 (∑ 𝑋3 )2
𝑆𝑆𝐶 = + + – 𝐶. 𝐹
𝑛1 𝑛2 𝑛3

(48) 2 (64)2 (92)2

= + + − 3468 = 16.8
3 4 5

𝑆𝑆𝐸 = 𝑇𝑆𝑆 − 𝑆𝑆𝐶 = 84 − 16.8 = 67.2

STEP 3: ANOVA TABLE

SOURCE OF SUM OF SQUARES DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION FREEDOM SQUARES

BETWEEN SSC = 16.8 C – 1 = 3- 1 = 2 𝑆𝑆𝐶 𝑀𝑆𝐶

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐶−1 𝑀𝑆𝐸

16.8 8.4
= =
2 7.47
= 𝟖. 𝟒 = 1.13

ERROR SSE = 67.2 N – C = 12 – 3 = 9 𝑆𝑆𝐸

𝑀𝑆𝐸 =
𝑁−𝐶
67.2
= = 𝟕. 𝟒𝟕
9

TOTAL
RESULT:

Table value of 𝐹𝑐 at 5% level in ( 2 , 9 ) degrees of freedom (d.f) is 4.26

Since calculated value of 𝐹𝑐 < 𝑡𝑎𝑏𝑢𝑙𝑎𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝐹𝑐

Which implies 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑.

TWO WAY ANOVA

Two - way classification observations are classified according to two factors

WORKING PROCEDURE:

Null Hypothesis 𝐻0 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … …

Alternative Hypothesis 𝐻1 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … … ..

STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠
𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 + ⋯
𝑇2
Correction factor: 𝐶. 𝐹 =
𝑁

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 + ⋯ − 𝐶. 𝐹

(∑ 𝑋1 )2 (∑ 𝑋2 )2
𝑆𝑆𝐶 = 𝑛1
+ 𝑛2
+ ⋯ − 𝐶. 𝐹

(∑ 𝑌1 )2 (∑ 𝑌2 )2
𝑆𝑆𝐶 = 𝑛1
+ 𝑛2
+ ⋯ − 𝐶. 𝐹

𝑆𝑆𝐸 = 𝑇𝑆𝑆 − 𝑆𝑆𝐶 − 𝑆𝑆𝑅

STEP 3: TWO – WAY ANOVA TABLE:

SOURCE OF SUM OF SQUARES DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION FREEDOM SQUARES

BETWEEN SSC = C–1= 𝑆𝑆𝐶 𝑀𝑆𝐸

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐶−1 𝑀𝑆𝐶

BETWEEN ROWS SSR = R–1= 𝑆𝑆𝑅 𝑀𝑆𝐸

𝑀𝑆𝑅 = 𝐹𝑟 =
𝑅−1 𝑀𝑆𝑅

ERROR SSE = (C – 1) X (R – 1) = 𝑆𝑆𝐸

𝑀𝑆𝐸 =
𝑁−𝐶
PROBLEM 2:

Four doctors each test four treatments for a certain disease and observe the number of days each
patient takes to recover. The results are as follows (recovery time in days)

_____________________________________________________________________________________
Treatment

Doctor T1 T2 T3 T4

D1 10 14 19 20

D2 11 15 17 21

D3 9 12 16 19

D4 8 13 17 20

Discuss the difference between (a) doctors and (b) treatments.

SOLUTION:

Null Hypothesis: 𝑯0 ∶
𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑒𝑟𝑓𝑜𝑚𝑒𝑛𝑐𝑒 𝑜𝑓 𝑑𝑜𝑐𝑡𝑜𝑟𝑠 𝑎𝑛𝑑 𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡𝑠.
Alternative Hypothesis:

𝑯 𝟏 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑒𝑟𝑓𝑜𝑚𝑒𝑛𝑐𝑒 𝑜𝑓 𝑑𝑜𝑐𝑡𝑜𝑟𝑠 𝑎𝑛𝑑 𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡𝑠
USE CODE: 15

𝑋1 𝑋2 𝑋3 𝑋4 TOTAL 𝑋12 𝑋22 𝑋32 𝑋42

−5 −1 4 5 ∑𝑌1 = 3 25 1 16 25

−4 0 2 6 ∑𝑌2 = 4 16 0 4 36

−6 −3 1 4 ∑𝑌3 36 9 1 16
= −4

−7 −2 2 5 ∑𝑌4 49 4 4 25
= −2

TOTAL ∑𝑋2 ∑𝑋3 = 9 ∑𝑋4 ∑𝑋12 = ∑𝑋22 = ∑𝑋32 = ∑𝑋42 =

=−6 = 20 126 14 25 102
∑𝑋1=
− 22
STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠 = 16
𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 + ∑ 𝑋4
= −22 − 6 + 9 + 20 = 1
𝑇2 ( 1) 2
Correction factor: 𝐶. 𝐹 = 𝑁
=
16
= 𝟎. 𝟎𝟔𝟐𝟓

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 + ∑𝑋42 − 𝐶. 𝐹

= 126 + 14 + 25 + 102 − 0.0625

= 266.94
(∑ 𝑋1 )2 (∑ 𝑋2 )2 (∑ 𝑋3 )2 (∑ 𝑋4 )2
𝑆𝑆𝐶 = + + + – 𝐶. 𝐹
𝑛1 𝑛2 𝑛3 𝑛4

(−22) 2 (−6)2 (9)2 (20)2

= + + + − 0.0625
4 4 4 4
= 250.19

(∑ 𝑌1 )2 (∑ 𝑌2 )2 (∑ 𝑌3 )2 (∑ 𝑌4 )2
𝑆𝑆𝑅 = + + + – 𝐶. 𝐹
𝑛1 𝑛2 𝑛3 𝑛4
(3) 2 (4)2 (−4)2 (−2)2
= + + + − 0.0625
4 4 4 4
= 11.19

𝑺𝑺𝑬 = 𝑻𝑺𝑺 − 𝑺𝑺𝑪 − 𝑺𝑺𝑹

= 266.94 − 250.19 − 11.19
= 5.56
STEP 3: ANOVA TABLE

SOURCE OF SUM OF SQUARES DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION FREEDOM SQUARES

BETWEEN SSC = 250.19 C–1=4-1=3 𝑆𝑆𝐶 𝑀𝑆𝐶

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐶−1 𝑀𝑆𝑅
250.19 83.40
= 3 = 0.62
= 83.40 = 134.52

BETWEEN ROWS SSR = 11.19 R–1=4–1=3 𝑆𝑆𝑅 𝑀𝑆𝑅

𝑀𝑆𝑅 = 𝐹𝑟 =
𝑅−1 𝑀𝑆𝐸

=
11.19 3.73
3 =
= 3.73 0.62

= 𝟔. 𝟎𝟐
ERROR SSE = 5.56 (C – 1) X(R – 1) = 9 𝑆𝑆𝐸
𝑀𝑆𝐸 =
𝑁−𝐶
5.56
=
9
= 0.62
TOTAL

From the F- table, the value of F at 5% level in ( 3, 9 ) d.f is 3.86

RESULT:

Since both calculated values of 𝐹𝑐 𝑎𝑛𝑑 𝐹𝑟 > 𝑡𝑎𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝐹

Which implies 𝐻0 𝑖𝑠 𝑅𝑒𝑗𝑒𝑐𝑡𝑒𝑑.

Hence the difference between doctors is significant and that between the treatments is highly
significant.
LATIN SQUARE DESIGN:

Three -way classification observations are classified according to three factors

WORKING PROCEDURE:

Null Hypothesis 𝐻0 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … …

Alternative Hypothesis 𝐻1 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 … … … ..

STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠

𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 + ⋯

𝑇2
Correction factor: 𝐶. 𝐹 =
𝑁

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 + ⋯ − 𝐶. 𝐹

(∑ 𝑋1 )2 (∑ 𝑋2 )2
𝑆𝑆𝐶 = + + ⋯ − 𝐶. 𝐹
𝑛1 𝑛2

(∑ 𝑌1 )2 (∑ 𝑌2 )2
𝑆𝑆𝑅 = 𝑛1
+ 𝑛2
+ ⋯ − 𝐶. 𝐹

(∑ 𝑍1 )2 (∑ 𝑍2 )2
𝑆𝑆𝑇 = + + ⋯ − 𝐶. 𝐹
𝑛1 𝑛2

𝑆𝑆𝐸 = 𝑇𝑆𝑆 − 𝑆𝑆𝐶 − 𝑆𝑆𝑅 − 𝑆𝑆𝑇

STEP 3: ANOVA TABLE – LATIN SQUARE DESIGN

SOURCE OF SUM OF DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION SQUARES FREEDOM SQUARES

BETWEEN SSC = K–1= 𝑆𝑆𝐶 𝑀𝑆𝐸

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐶−1 𝑀𝑆𝐶

BETWEEN ROWS SSR = K–1= 𝑆𝑆𝑅 𝑀𝑆𝐸

𝑀𝑆𝑅 = 𝐹𝑟 =
𝑅−1 𝑀𝑆𝑅

BETWEEN SST = K–1= 𝑆𝑆𝑇 𝑀𝑆𝐸

𝑀𝑆𝑇 = 𝐹𝑇 =
TREATMENTS 𝑁−𝐶 𝑀𝑆𝑇

ERROR SSE = (K – 1) X (K – 2) = 𝑀𝑆𝐸

𝑆𝑆𝐸
=
(K – 1)X (K – 2)

TOTAL
PROBLEM 3:

Analysis the variance in the following Latin square of yields (in kgs ) of wheat where A, B,C,D denote
different methods of cultivation

D122 A121 C123 B122

B124 C123 A122 D125

A120 B119 D120 C121

C122 D123 B121 A122

Examine whether the different methods of cultivation have given significantly different yields.

SOLUTION:

Null Hypothesis: 𝐻0 ∶
𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝐿𝑒𝑡𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡 𝑚𝑒𝑡ℎ𝑜𝑑 𝑜𝑓 𝑐𝑢𝑙𝑡𝑖𝑣𝑎𝑡𝑖𝑜𝑛.
Alternative Hypothesis:

𝐻 1 : 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑖𝑛 𝑡ℎ𝑒 𝑖𝑛 𝑡ℎ𝑒 𝐿𝑒𝑡𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡 𝑚𝑒𝑡ℎ𝑜𝑑 𝑜𝑓 𝑐𝑢𝑙𝑡𝑖𝑣𝑎𝑡𝑖𝑜𝑛.
USE CODE: 120

𝑋1 𝑋2 𝑋3 𝑋4 TOTAL 𝑋12 𝑋22 𝑋32 𝑋42

2 1 3 2 ∑𝑌1 = 8 4 1 9 4

4 3 2 5 ∑𝑌2 16 4 4 25
= 14
0 -1 0 1 ∑𝑌3 = 0 0 1 0 1

2 3 1 2 ∑𝑌4 = 8 4 9 1 4

TOTAL ∑𝑋2 = 6 ∑𝑋3 = 6 ∑𝑋4 ∑𝑋12 = ∑𝑋22 = ∑𝑋32 = ∑𝑋42 =

= 10 24 15 14 34
∑𝑋1= 8

A B C D

1 2 3 2

2 4 3 5

0 -1 1 0

2 1 2 3

∑Z1= 5 ∑Z2=6 ∑Z3=9 ∑Z4=10

STEP1: To find the following

𝑁 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠 = 16
𝑇 = ∑ 𝑋1 + ∑ 𝑋2 + ∑ 𝑋3 + ∑ 𝑋4 = 8 + 6 + 6 + 10 = 30
𝑇2 (30)2
Correction factor: 𝐶. 𝐹 = 𝑁
=
16
= 56.25

STEP2: To Calculate the following

𝑇𝑆𝑆 = ∑𝑋12 + ∑𝑋22 + ∑𝑋32 + ∑𝑋42 − 𝐶. 𝐹

= 24 + 15 + 14 + 34 − 56.25
= 35.75
(∑ 𝑋1 )2 (∑ 𝑋2 )2 (∑ 𝑋3 )2 (∑ 𝑋4 )2
𝑆𝑆𝐶 = 𝑛1
+ 𝑛2
+ 𝑛3
+ 𝑛4
– 𝐶. 𝐹

(8) 2 (6)2 (6)2 (10)2

= + + + − 56.25
4 4 4 4
= 2.75

(∑ 𝑌1 )2 (∑ 𝑌2 )2 (∑ 𝑌3 )2 (∑ 𝑌4 )2
𝑆𝑆𝑅 = + + + – 𝐶. 𝐹
𝑛1 𝑛2 𝑛3 𝑛4
(8) 2 (14)2 (0)2 (8)2
= + + + − 56.25
4 4 4 4
= 24.75

(∑ 𝑍1 )2 (∑ 𝑍2 )2 (∑ 𝑍3 )2 (∑ 𝑍4 )2
𝑆𝑆𝑇 = + + + – 𝐶. 𝐹
𝑛1 𝑛2 𝑛3 𝑛4
(5) 2 (6)2 (9)2 (10)2
= + + + − 56.25
4 4 4 4
= 4.25
𝑆𝑆𝐸 = 𝑇𝑆𝑆 − 𝑆𝑆𝐶 − 𝑆𝑆𝑅 − 𝑆𝑆𝑇
= 35.75 − 2.75 − 24.75 − 4.25
= 4.0
STEP 3: ANOVA TABLE

SOURCE OF SUM OF SQUARES DEGREES OF MEAN SUM OF VARIATION RATIO

VARIATION FREEDOM SQUARES

BETWEEN SSC = 2.75 K–1=4-1=3 𝑆𝑆𝐶 𝑀𝑆𝐶

𝑀𝑆𝐶 = 𝐹𝑐 =
COLUMNS 𝐾−1 𝑀𝑆𝑅
2.75 0.92
= = 0.67
3
= 0.92 = 1.37313

BETWEEN ROWS SSR = 24.75 K–1=4–1=3 𝑆𝑆𝑅 𝑀𝑆𝑅

𝑀𝑆𝑅 = 𝐹𝑟 =
𝐾−1 𝑀𝑆𝐸
24.75 8.25
= 3 =
= 8.25 0.67

= 𝟔. 𝟎𝟐
BETWEEN SST = 4.25 K–1=4–1=3 𝑆𝑆𝑇 𝑀𝑆𝑇
𝑀𝑆𝑇 = 𝐹𝑇 =
TREATMENTS 𝐾−1 𝑀𝑆𝐸
4.25
= 3
1.42
= 1.42 =
0.67

= 𝟐. 𝟏𝟏𝟗𝟒
ERROR SSE = 4.0 (K – 1) X(K – 2) = 6 𝑆𝑆𝐸
𝑀𝑆𝐸 =
𝐷. 𝐹
4.0
=
6
= 0.67
TOTAL

From the F- table, the value of F at 5% level in ( 3, 6 ) d.f is 4.76

RESULT:

Since both calculated values of 𝐹𝑐 𝑎𝑛𝑑 𝐹𝑇 < 𝑡𝑎𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝐹

Since with respect to letters, the difference between the method of cultivation is not significant.

.
THANK YOU