2B.

Probability Distributions
2B.1. Introduction
We know that the result of a random experiment is called an outcome. The outcome of a trial is
being determined by the chance factor. The all possible outcomes of an experiment are, therefore,
chance outcomes. Usually we are not interested in the outcomes of an experiment as such. It is
often more useful to describe a particular property or an attribute of an outcome in numerical terms.
Probability distributions, as the name suggests, is the listing of all possible outcomes of an
experiment together with their probabilities.

2B.2. Random Variable

A random variable is a particular characteristic of outcomes of a random experiment, which is

used to express all outcomes of the experiment in numerical values.
For example, consider the random experiment tossing two coins and observing the faces
turns up. A random variable X is defined as ‘the number of heads happens’. The possible outcomes
are (H,H), (H,T), (T,H), and (T,T).
The value of X for the outcome (H,H) is 2
The value of X for the outcome (H,T) is 1
The value of X for the outcome (T,H) is 1
The value of X for the outcome (T,T) is 0
So, the random variable X has values 0, 1 and 2
A random variable is said to be discrete if it takes only a finite or countable number of distinct
values. The random variable X defined above is discrete.

2B.3. Discrete probability distribution

We can see that for each value of the random variable X defined in the two-coins tossing
experiment, has a certain probability because X gets values based on the outcome of the random
experiment. When X=2, its probability is 1 because X gets the value 2 when the outcome is
4
(H,H). When X=1, its probability is 1 2 because X gets the value 1 when the outcome is either
(H,T) or (T,H). Similarly, the probability that X=0 is 1 . We may tabulate all these possible
4
values of X along with their corresponding probabilities as below.

x: 0 1 2

P(x): 1 1
2
1
4 4
This presentation is called a discrete probability distribution (probability mass function).

Thus probability mass function (pmf) a random variable X takes values x1, x2, …, xn is defined
as f(x) = P({X=x}), x = x1, x2, …, xn.
Properties of pmf
1. f(x)  0 for all x (probabilities non-negative)
2.  f ( x) =1 ( Total probability is 1)
x

2B. 4. Mathematical Expectation of a discrete distribution

Mathematical expectation of a random variable X takes values x 1, x2, …, xn with probabilities p1,
p2, …, pn is denoted by E(X) and is defined as:
E(X) = x1p1 + x2 p2 + …. + xnpn

Consider the following frequency distribution

Observation(x) : 1 2 3 4 5
Frequency(f) : 2 5 10 2 1
2 5 10 2 1
Relative frequency(p) :
20 20 20 20 20

Here the total frequency is 20. When divide the frequencies by total frequency, we get
relative frequencies. These relative frequencies are the probabilities of the observations. So, it is
possible to find the mean of a probability distribution in the same way as the simple mean can be
found from a frequency distribution.

We know that mean of the values x1, x2, …, xn with frequencies f1, f2, …, fn is:

x 1f1 + x 2 f 2 + . + x n f n
x = , where N is the total frequency.
N
f1 f f
= x1 + x2 2 + …. + xn n
N N N
= x1p1 + x2 p2 + …. + xnpn
= E(X)
Thus E(X) is the arithmetic mean of the distribution.
2 5 10 2 1
The mean of the above distribution is E(X) = 1 + 2 + 3 + 4 + 5
20 20 20 20 20
1  2 + 2  5 + 3  10 + 4  2 + 5  1
=
20
= 2.75
Similarly, E(X2) = x12 p1 + x22 p2 + …. + xn2 pn and Variance of X = E(X2) – E(X)2
 2 5 10 2 1
For the above distribution, E(X2) = 12 + 22 + 32 + 42 + 52
20 20 20 20 20
= 8.45
So, Variance of X = 8.45 – (2.75)2
= 0.8875
2B.5. Standard Distributions
Some populations may have common characteristics and they have the same probability structure
or probability distribution. There are several probability distributions are defined to describe
different populations. We shall, therefore, now describe some discrete probability distributions
applicable to many real life problems.

1. Binomial Distribution
The binomial distribution was derived by the Swiss Mathematician Jacob Bernoulli. It
deals with populations whose members can be divided into two categories with reference to
the presence or absence of a particular characteristic.

For example, the items produced by a manufacturing process can either defective or non-
defective; the sex of new born child is either male or female; a factory worker may be
educated or uneducated, and so on.

Any random experiments that satisfy the following properties are called Bernoulli trials.
1. There are only two mutually exclusive and collectively exhaustive outcomes in the
experiment.
2. In repeated trials, the probabilities of occurrence of the events remain constant.
3. The trials are independent.
Consider a random experiment (trial), which has only two outcomes, say, success and failure.
Let p be the probability of success. Then 1-p = q be the probability of failure. Let the trial be
repeated n times and the trials are independent. Let X be the number of successes happens
out of n trials. The X follows the binomial distribution with parameters n and p. The
probability mass function of X is defined as:

f(r) = P{X=r} = nCr pr qn-r , r = 0, 1, 2, …, n

Illustration:
We have to find out the probability that r successes out of n trials. Then n-r trials are failures.
Since the trials are independent, probability of r successes is p r and probability of n-r failures
is qn-r. Out of n trials r successes can be happened in nCr ways. Thus P{X=r}= nCr pr qn-r.

Example 1: A machining process produces 90% good items. Write the probability mass
function of the number of good items in 10 units produced from this process.
Solution: Here the items produced may be either good or defective. Given that probability of
a good item produced (p) is 0.9 and hence q is 1- .9 = 0.1. Also n= 10. So, the pmf is f(r) = 10Cr
(0.9)r(0.1)10-r

Characteristics of a binomial experiment

A binomial experiment possesses the following characteristics:
1. It consists of n trials and each trial has only two possible outcomes, either of which may
be designated as a success so that the other outcome is a failure.
2. The probability of a success remains the same in n trials.
3. All the trials are independent.

Mean and Variance of binomial distribution

Mean of binomial distribution with parameters n and p = np
Variance of binomial distribution with parameters n and p = npq
Clearly, Mean > Variance.

Example 2: Assuming the probability of a male birth is 1 , find the probability that a family
2
of 3 children will have a) at least one girl, b) two boys and one girl, and c) at the most two
girls. Also find the expected number of families with a), b), and c) out of 300 families with 3
children.

Solution: Let X be the number of girls out of three children. Then X follows binomial
distribution with parameters n=3 and p= 1 .
2
3 r -r
So, f(r) = Cr (0.5) (0.5)3
= 3Cr (0.5)3, r= 0,1,2,3.
a) P(at least one girl) = P(X  1)
= 1 - P(X=0)
= 1 – f(0)
= 1 - 3C0 (0.5)3
= 1 – 0.125 = 0.875
b) P(two boys and one girl) = P(X=1)
= 3C1 (0.5)3
= 0.375
c) P(at most two girls) = P(X  2)
= 1 – P(X=3)
 = 1 - 3C3 (0.5)3
= 1 – 0.125
= 0.875
Expected number of families with at least one girl = 300  P(at least one girl)
= 300  0.875
 263
Similarly, expected number of families with two boys and one girl = 300  0.375
 113
And expected number of families with at most two girls = 300  0.875
 263

Example 3: The incidence of occupational disease in an industry is such that the workers have a 20% chance of suffering from it. What is
the probability that out of six workers, 4 or more will contact the disease?

Solution: Given that, n= 6, p = 0.2 and q = 0.8

Let X denotes the number of people getting the disease.
f(r) = 6Cr (0.2)r (0.8)6-r, r = 0, 1, 2, 3, 4, 5, 6.
P( 4 or more will contact the disease) = P(X  4)
= f(4) + f(5) + f(6)
= 6C4(0.2)4(0.8)2 + 6C5(0.2)5(0.8)+6C6 (0.2)6
= 0.0169

2. Poisson Distribution

The Poisson distribution was derived by Simeon Poisson in 1837. This distribution is useful in
cases of rare events ( i.e, probability is very small). Rare events are common in all fields. For
example, the number of defective articles produced by a high quality machine; in quality control,
the number of defects per item is to be counted; in insurance, the number of occurrence of deaths
and accidents in a specific time period or region is to be considered and so on. Poisson distribution
is a limiting form of binomial distribution as n moves towards infinity and p moves towards zero
but mean=np remains constant. That is, when p is very small and the value of n is very large, the
Poisson distribution is more appropriate than the binomial distribution.
 The probability distribution of the Poisson distribution with parameter ‘m’ is given by f(x) =
e −m m x
, x = 0, 1, 2, …
x!
Mean and Variance of Poisson distribution
Mean = m
Variance = m
Example 4: An inspection of a random sample of 100 pages printed by a press revealed 20 printing
errors. Find the probability that a page contains:
a) less than 3 errors
b) at the most 3 errors
c) at least 3 errors
d) exactly 3 errors

Solution: Let X be the number of printing mistakes per page. Then X follows Poisson
20 e −0.2 (0.2) x
distribution with parameter m = = 0.2. So, f(x) = , x = 0, 1, 2, …
100 x!

a) P(less than 3 errors) = P(X<3) = f(0) + f(1) + f(2)

e −0.2 (0.2) e −0.2 (0.2) 2
= e-0.2 + +
1! 2!
= e-0.2 ( 1+0.2+0.02)
= e-0.2 (1.22)
= 0.9989
b) P(at most 3 errors) = P(X  3) = f(0)+f(1)+f(2)+f(3)

-0.2 e −0.2 (0.2) e −0.2 (0.2) 2 e −0.2 (0.2) 3

=e + + +
1! 2! 3!
= 0.9999
c) P(at least 3 errors) = P(X  3) = 1 – P(X<3)
= 1 – 0.9989
= 0.0011
e −0.2 (0.2) 3
d) P(exactly 3 errors) = P(X=3) =
3!
= 0.0010
Example 5: If 3% of electric bulbs manufactured by a company are defective, find the probability
that in a sample of 100 bulbs exactly 5 bulbs are defective.

Solution: Given that n=100 and probability of a defective bulb, p = 0.03

Let X be the number of defectives in a sample of 100 bulbs. The random variable X can be
approximated to Poisson with m = np = 1000.03 = 3
e -3 3 x
Then f(x) = , x = 0,1, 2, …
x!
e -3 3 5
Probability that exactly 5 are defective = P(X=5) = = 0.10082
5!
Example 5: Between the hours 2 and 4 PM the average number of phone calls per minute coming
into the switch board of a company is 2.5. Find the probability that during a particular minute there
will be no phone call at all.

Solution: Let X be the number of phone calls per minute.

e -2.5 2.5 x
Then X follows Poisson with m = 2.5. So, f(x) = , x = 0, 1, 2, ….
x!
P(no calls in a particular minute) = P(X=0) = e-2.5 = 0.08208

3. Normal distribution
Normal distribution is continuous distribution and is most important of all the theoretical
distribution. Discrete distributions relate to discrete random variables involving count data
expressed in discrete numbers such as 0, 1, 2, …. But in continuous distribution, the variable of
interest may take any value within a given range. That is, measured data like, height, weight,
temperature etc. are follows continuous distributions. In a continuous distribution, the probability
of random variable take a particular value is zero. For example, the probability that a student has
weight exactly 55.5 kg is negligible because the weights close to 55.5 kg in the specified range are
so innumerable that it is difficult to distinguish this particular weight from other weights in close
proximity. That is, if X follows a continuous distribution, then P(X=x) = 0.
The probability density function of the normal distribution with parameters  and σ is given by
( x− )2
1 −
f(x)= e 2 2
,-  < x<  ; -  < <  ; and is denoted by N(  ,  ).
2 

Mean & Variance of normal distribution

Mean = μ
Variance =  2 ; Standard deviation = σ

Properties of normal distribution

The following are the properties of a normal distribution.
 1. The normal curve is symmetrical about the mean
2. The mean, median and mode have the same value. i.e, Mean = Median = Mode
3. The height of the normal curve is maximum at the mean value.
4. The normal curve has a single peak; thus it is unimodal.
5. The mean deviation is 0.7979 times of the standard deviation.
6. The normal curve is asymptotic to the base line. That is the two tails of the normal curve
extend indefinitely and never touch the horizontal axis.
7. The points of inflection ( the points at which the curve changes its direction) are each at a
distance of standard deviation from mean.
8. For a normal distribution, 68.27% of the observations will be in the interval (  −  ,
 +  ); 95.45% observations will be in (  − 2 ,  + 2 ); and 99.73% observations in
the interval (  − 3 ,  + 3  ).

μ-3σ μ-2σ μ-σ 68.27% μ+σ μ+2σ μ+3σ

95.45%

99.73%

The Standard Normal Distribution

The standard normal distribution is a normal distribution with mean μ=0 and standard deviation σ
=1. The standard normal distribution is very useful because normal distributions with mean μ and
standard deviation σ can be converted into a standard normal by change of origin and scale. This
becomes necessary as otherwise in different normal distributions with different values of mean
and standard deviation, it would be very difficult to find out the area between various ordinates of
the normal curve.

Suppose X is a normal variable with mean μ and standard deviation σ. Then the variable Z =
X −
is standard normal. The values obtained by this transformation are called Z scores or

standard scores.

Standard Normal Area Tables

There are two types of standard normal area tables. Firstly, tables which give the area towards the
tail-end of the standard normal curve beyond the ordinate at any particular Z value.

0 z

The other type of tables give the area between z = 0 and and any other positive z-value as given
below.

0 z
As the normal curve is perfectly symmetrical with 50% of the area on either side of the
maximum ordinate at its mean, the areas given by first type tables when subtracted from 0.5 will
give areas offered by second type.

Example 6: A normal variable has a mean of 10 and a standard deviation of 5. What is the
probability that the normal variable will take a value in the interval 0.2 to 19.8?

Solution: The required probability is P(0.2 < X < 19.8). Given mean = 10 Standard deviation = 5

0.2 − 10 X − 10 19.8 − 10
P(0.2 < X < 19.8) = P( < < )
5 5 5
= P(-1.96<Z<1.96)
= 0.95
( From the normal table, we find that area between z = 0 and z = 1.96 is 0.475. By symmetry
property of normal distribution, area in between z = -1.96 and z = 0 is same as the area in between
z = 0 and z = 1.96. Thus area in between z = -1.96 and z = 1.96 is 0.95 )
 0.475 0.475

-1.96 0 1.96

Example 7: A production engineer finds that on an average mechanics working in a machine

shop complete a certain task in 15 minutes. The times required to complete the task is normally
distributed with a standard deviation of 3 minutes. Find the probabilities that the task is
completed a) in less than 8 minutes b) in more than 9 minutes c) in between 10 and 12 minutes.

Solution: Let X be the time required to complete the task. Given X is normally distributed with
mean 15 and standard deviation 3.
a) The required probability = P(X<8)
8 − 15
= P(Z< )
3
= P(Z< -2.33)
= 0.5 – P(0<Z<2.33)
= 0.5 – 0.4901 = 0.0099
b) The required probability = P(X>9)
9 − 15
= P(Z> )
3
= P(Z> -2)
= 0.5+ P(0<Z<2)
= 0.5 + 0.4772 = 0.9772
c) The required probability = P(10<X<12)
= P(-1.67<Z<-1)
= P(0<Z<1.67)-P(0<Z<1)
= 0.4525 –0.3413 = 0.1112

Example 8: An air lines company has the policy of employing only Indian women whose height
is between 62 inches and 69 inches. If the height of Indian women is approximately normal with
mean 64 inches and standard deviation 3 inches. Out of 1000 applicants find the number of
applicants that would be a) too tall b) too short c) of acceptable height.

Solution: Let X be the height of Indian women. Given mean = 64 and standard deviation is 3.
 62 − 64
When X = 62, Z = = -0.67
3
69 − 64
When Z= 69, Z = = 1.67
3
a) P(an applicant is too tall ) = P(Z>1.67)
= 0.5 – P(0<Z<1.67)
= 0.5-.4525 = 0.0475
So, the number of too tall applicants = 1000 0.0475 = 48

b) P(an applicant is too short) = P(Z< -0.67 )

= 0.5 –P(0<Z<0.67)
= 0.5 – 0.2486 = 0.2514
The number of too short applicants = 10000.2514 = 251
c) P( an applicant is acceptable height) = P(-0.67<Z<1.67)
= P(0<Z<0.67) + P(0<Z<1.67)
= 0.2486+0.4525 = 0.7011
The number of acceptable applicants = 10000.7011 = 701

2B.6. Exercises

1. A manufacturer of laptop computer has determined that on an average 3% of screens

produced are defective. A sample of one dozen monitors from a production lot was taken
at random. What is the probability that in this sample fewer than 2 defectives will be
found?
2. A newly married couple plans to have 5 children. An astrologist tells them that based on
astrological reading, they have an 80% chance of having a baby boy on any particular
birth. The couple would like to have 3 boys and 2 girls. Find the probability of this event?
3. An automatic machine makes paper clips from coils of wire. On an average, one in 400
paper clips is defective. If the paper clips are packed in small boxes of 100 clips each,
what is the probability that any given box of clips contains:
a) No defectives
b) One or more defectives
c) Less than two defectives
d) Two or less defectives
4. The customers arrive at a drive- in window of Apple bank at an average rate of one
customer per minute.
a) What is the probability that exactly two customers arrive in a given minute?
b) What is the probability that no arrivals in a particular minute?
5. The local telephone switchboard in a small village receives on an average five calls per
hour for information. Assuming that these calls follow Poisson distribution, what is the
probability that:
a) More than half an hour elapses between two successive calls
b) In a particular hour, five calls are received.
 6. The demand for fax machines per day from a mail order distributor is normally distributed
with a mean of 80 and a variance of 25. Find the percentage of days when the demand will
be:
a) Greater than 70
b) Less than or equal to 70
c) Between 70 and 90
d) Greater than 85 and less than 92
7. The life of a battery is normally distributed with an average life of 5 years and a standard
deviation of 60 days. How long should the guarantee period be if the company wishes to
replace no more than 15% of the batteries?

8. The diameter of ball bearings are normally distributed with a mean of 2.42inches and a
standard deviation of 0.01 inches, Determine the percentage of ball bearings with
diameter:
a) between 2.4 and 2.43 inches
b) greater than 2.43 inches
c) less than 2.39 inches

Reference

1.. Nabendu Pal, Sahadeb Sarkar(2005), Statistics: Concepts and Applications, Prentice- Hall of
India, New Delhi.

2. Jit S Chandan (2003), Statistics for Business and Economics, Vikas Publishing House, New
Delhi.

3. R.P. Hooda (1994), Statistics for Business and Economics, Macmillan India Ltd.