Common Discrete

Distributions

5.1 Discrete Uniform Distribution

If the random variable X assumes the values x1 , x2 , ..., xk , with equal probabilities,
then the discrete uniform distribution is given by:

1
f (x; k) = , for x = x1 , x2 , ..., xk .
k
Note: We have used the notation f (x; k) instead of f (x) to indicate that the
uniform discrete distribution depends on the parameter k.
Example: When a die is tossed, each element of the sample space S =
{1, 2, 3, 4, 5, 6} occurs with probability 16 . Therefore, we have a discrete uni-
form distribution with:
1
f (x; 6) = , for x = 1, 2, 3, 4, 5, 6.
6
Example: Suppose that a student is selected at random from a class of 25 to
1
participate in a certain seminar. Each student has the same probability of 25 of
being selected. If we assume that the students have been numbered in some
way from 1 to 25, the distribution is uniform with:

1
f (x; 25) = , for x = 1, 2, ..., 25.
25

5.2 The Bernoulli Distribution

Suppose that a trial, or an experiment, whose outcome can be classified as either
a ’success’ or as a ’failure’ is performed. If we let X = 1 when the outcome is a
success and X = 0 when it is a failure, then the probability mass function of X is
given by:

P(X = 0) = 1 − p and P(X = 1) = p (5.1)

1
2 Elements of Probability & Statistics– MTH1202

where p is the probability that the trial is a "success".

A random variable X is said to be a Bernoulli random variable (after the
Swiss mathematician James Bernoulli) if its probability mass function is given
by Equation (5.1) for some p ∈ (0, 1). If we let q = 1 − p, the probability mass
function of X can as well be written as:
x = 0, 1
( x 1−x
pq for
f (x) = (5.2)
0 otherwise

Its expected value is

E(X) = 1 · P(X = 1) + 0 · P(X = 0) = 1 · p + 0 · q = p

That is, the expectation of a Bernoulli random variable is the probability that the
random variable assumes a value of 1 (success). Also,

E(X2 ) = 12 · p + 02 · q = p

and

Var(X) = E(X2 ) − [E(X)]2 = pq

If the random variables in an infinite sequence X1 , X2 , . . . are independent and

identically distributed (i.i.d), and if each random variable Xi has a Bernoulli
distribution with parameter p, then it is said that the random variables X1 , X2 , . . .
form an infinite sequence of Bernoulli trials with parameter p. Similarly, if n
random variables X1 , X2 , . . . , Xn are i.i.d and each has a Bernoulli distribution
with parameter p, then it is said that the variables X1 , X2 , . . . , Xn form n Bernoulli
trials with parameter p.
For example, suppose that a fair coin is tossed repeatedly. Let Xi = 1 if a
head is obtained on the ith toss, and let Xi = 0 if a tail is obtained (i = 1, 2, . . .).
Then the random variables X1 , X2 , . . . form an infinite sequence of Bernoulli
trials with parameter p = 12 . Similarly, suppose that 10% of the items produced
by a certain machine are defective, and that n items are selected at random and
inspected. In this case, let Xi = 1 if the ith item is defective, and let Xi = 0 if it
is nondefective (i=1,. . . , n). Then the variables X1 , X2 , . . . , Xn form n Bernoulli
trials with parameter p = 10 1
.

5.3 The Binomial Distribution

Suppose now that n independent trials, each of which results in a ’success’ with
probability p and in a ’failure’ with probability 1 - p, are to be performed. If X
represents the number of successes that occur in the n trials, then X is said to be
a binomial random variable with parameters (n, p). This result can be restated
as follows:
If the random variables X1 , X2 , . . . , Xn form n Bernoulli trials with parameter p and
if X = X1 + X2 + · · · + Xn , then X has a binomial distribution with parameters n and p.
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 3

The probability mass function of a binomial random variable with parameters

n and p is given by:
!
n
P(X = x) = px (1 − p)n−x
x

for x = 0, 1, 2,
! ..., n.
n n!
where = =n Cx with n being the number of trials, x the
x x!(n − x)!
number of successes in the n trials and p the probability of success on a single
trial.
If X is a random variable that is binomially distributed with parameters
n and p, we write X ∼ binomial(n, p). Note that, by the binomial theorem, the
probabilities sum to 1, that is,
n n !
X X n
P(X = x) = px qn−x = (p + q)n = 1
x
x=0 x=0

Example 1 A coin is tossed 5 times. Find the probability that exactly 3 heads will
appear.
Soln: Let X be the random variable representing the number of heads that will appear.
X ∼ binomial(n, p), where n = 5 and p = 12 .
!
n
P(X = x) = px (1 − p)n−x , x = 0, 1, 2, 3, 4, 5
x

Thus
!  3  2
5 1 1 5
P(exactly three heads) = P(X = 3) = =
3 2 2 16

5.3.1 Reading Binomial Tables

Frequently, you will be interested in problems where you need to find probabili-
ties such as P(X < a), P(a ≤ X ≤ b) or P(X > b). To obtain desired probabilities
from tables, use the following definitions.
a
X
1. P(X ≤ a) = P(X = x)
x=0

a−1
X
2. P(X < a) = P(X = x)
x=0

3. P(X > a) = 1 − P(X ≤ a)

4. P(a ≤ X ≤ b) = P(X ≤ b) − P(X < a)

4 Elements of Probability & Statistics– MTH1202

5. P(a < X ≤ b) = P(X ≤ b) − P(X ≤ a)

Example 2 The probability that a patient recovers from a rare blood disease is 0.4. If
15 people are known to have contracted the disease, what is the probability that:

(i) Exactly 5 survive,

(ii) At least 10 survive the disease,

(iii) From 3 to 8 survive the disease.

Soln:
Let X be the number of people that survive. X ∼ binomial(n, p), where n = 15 and
p = 0.4 and
!
n
P(X = x) = px (1 − p)n−x , x = 0, 1, . . . , 15
x

(i) P(X = 5) =15 C5 (0.4)5 (0.6)10 = 0.1859

(ii) P(X ≥ 10) = 1 − P(X < 10) = 1 − P(X ≤ 9) =

1 − (P(X = 0) + P(X = 1) + P(X = 2) + ... + P(X = 9))

(iii) P(3 ≤ X ≤ 8) = P(X ≤ 8) − P(X < 3) = P(X ≤ 8) − P(X ≤ 2)

We can use tables to answer this question as well, look at the Cumulative Binomial
table at the back.
5 4
(i) P(X = 5) = b(x; 15, 0.4) = 0.4032 − 0.2173 = 0.1859
P P
b(x; 15, 0.4) −
x=0 x=0

9
(ii) P(X ≥ 10) = 1 − P(X < 10) = 1 − b(x; 15, 0.4) = 1 − 0.9662 = 0.0338
P
x=0

(iii) P(3 ≤ X ≤ 8) = P(X ≤ 8) − P(X < 3) = P(X ≤ 8) − P(X ≤ 2) =

8 2
b(x; 15, 0.4) = 0.9050 − 0.0271 = 0.8779
P P
b(x; 15, 0.4) −
x=0 x=0

Example 3 A fair coin is tossed 14 times. What is the probability that the head will
appear:

(a) at least 8 times,

(b) at most 6 times,

(c) exactly 5 times,

(d) less than 12 times,

(e) less than 9 but greater than 4 times.

 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 5

Example 4 A new car dealer knows from past experience that on the average, she will
make sale to about 20% of her customers. What is the probability that in five randomly
selected presentations, she makes a sale to:

(a) exactly three customers,

(b) at most one customer,

(c) at least 2 customers.

You have p = 20% = 0.2, n = 5. Therefore,

(a) P(X = 3) = P(X ≤ 3) − P(X ≤ 2) = 0.9933 − 0.9421 = 0.0512

(b) P(X ≤ 1) = 0.7373

(c) P(X ≥ 2) = 1 − P(X < 2) = 1 − P(X ≤ 1) = 1 − 0.7373 = 0.2627

5.3.2 Expectation of Binomial Distribution

Definition 1 The mean or expectation of a binomially distributed random variable X is

E(X) = np

where n is the number of repeated trials and p, (0 ≤ p ≤ 1) is the probability of success.

Proof: The first method is by using the meaning of a Bernoulli trial: If the
random variables X1 , X2 , . . . , Xn form n Bernoulli trials with parameter p and if
X = X1 + X2 + · · · + Xn , then X has a binomial distribution with parameters n
and p.
Recall that E(Xi ) = p for i = 1, 2, . . . , n. Thus

E(X) =E(X1 + X2 + · · · + Xn )
=E(X1 ) + E(X2 ) + · · · + E(Xn )
= p + p + · · · + p = np
| {z }
n times
6 Elements of Probability & Statistics– MTH1202

The second method is by use of the definition of E(X).

n
X
E(X) = xP(X = x)
x=0
n !
X n
= x px qn−x
x
x=0
n
X n!
= x px qn−x
x=0
x!(n − x)!
n
X n!
= px qn−x
(x − 1)!(n − x)!
x=1
n
X (n − 1)!
=np px−1 qn−x
(x − 1)!(n − x)!
x=1
n !
X n−1
=np px−1 qn−x
x−1
x=1

=np(p + q)n−1 = np

5.3.3 The Variance of Binomial Distribution

Definition 2 The variance of a binomially distributed random variable X is given by
Var(X) = npq.

This can be proved in two different ways. First from the definition of a Bernoulli
trial. Recall that for a Bernoulli random variable Xi , Var(Xi ) = pq. Since a
binomial random variable X = X1 + X2 + · · · + Xn , then

Var(X) =Var(X1 + X2 + · · · + Xn )
=Var(X1 ) + Var(X2 ) + · · · + Var(Xn )
= pq + pq + · · · + pq = npq
| {z }
n times

√
Thus Var(X) = npq and the standard deviation of X is npq.
In the second method, you use the definition of variance. Recall that E(X) = np
and from Var(X) = E(X2 ) − [E(X)]2 , we seek to find E(X2 ). Also recall from the
Binomial expansion that

n !
X n
(r + s)n = rx sn−x
x
x=0
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 7

Now
n
X
E(X(X − 1)) = x(x − 1)P(X = x)
x=0
n !
X n
= x(x − 1) px qn−x
x
x=0
n
X n!
= x(x − 1) px qn−x
x=0
x!(n − x)!
n
X n!
= px qn−x
x=2
(x − 2)!(n − x)!
n
X(n − 2)!
=n(n − 1)p2 px−2 qn−x
x=2
(x − 2)!(n − x)!
n !
X n−2
=n(n − 1)p2 px−2 qn−x
x−2
x=2
=n(n − 1)p (p + q)n−2
2

=n(n − 1)p2

But
E(X(X − 1)) = E(X2 − X) = E(X2 ) − E(X)
Thus
E(X2 ) = E(X(X − 1)) + E(X) = n(n − 1)p2 + np
Therefore, 7

Var(X) =E(X2 ) − [E(X)]2

=n(n − 1)p2 + np − (np)2
=n2 p2 − np2 + np − n2 p2
=np − np2
=np(1 − p) = npq

Example 5 In a family the probability of having a boy is 0.8. If there are seven children
in the family determine the:
(a) expected number of boys,
(b) expected number of girls,
(c) probability that at least four children are girls,
(d) probability that they are all boys.
Solution:
8 Elements of Probability & Statistics– MTH1202

(a) let X be the random variable representing the number of boys.

Then X ∼ binomial(n, p) with n = 7 and p = 0.8. E(X) = np = 7 ∗ 0.8 = 5.6

(b) let Ybe the random variable representing the number of girls.
Then Y ∼ binomial(n, p) with n = 7 and p = 0.2. E(Y) = 7 ∗ 0.2 = 1.4

(c) P(Y ≥ 4) = 1 − P(Y < 4) = 1 − P(Y ≤ 3) = 1 − 0.9667 = 0.0333

(d) P(all are boys)=P(X = 7) = 0.2097

Example 6 A multiple choice quiz has ten questions each with four alternative answers
of which one is correct. Determine the expected number of correct answers and the
variance, if someone does the quiz by sheer guesswork.
Solution:
Let X be the random variable representing the number of correct answers. Then
X ∼ binomial(n, p) with n = 10 and p = 14 .

1
E(X) = np = 10 ∗ = 2.5
4
and
Var(X) = npq = 10 ∗ 1/4 ∗ 3/4 = 1.875

Example 7 A multiple-choice exam consist of twenty questions each with five alter-
native answers of which only one is correct. Five marks are awarded for each answer
and one mark is subtracted for each incorrect or un attempted question. If the exam is
answered by sheer guesswork, find the:

(i) expected number of correct and incorrect answers,

(ii) expected overall mark,

(iii) variance of the mark in (ii)

Solution:
Let X be the random variable representing the number of correct answers and Y the
number of incorrect answers. Then, with n = 20, p = 51 = 0.2, q = 45 = 0.8

(i) E(X) = np = 20 ∗ 51 = 4 correct answers

E(Y) = nq = 20 ∗ 45 = 16 wrong answers

(ii) The expected mark for correct answers =4∗5 = 20. The expected mark for incorrect
answers =16 ∗ 1 = 16. Therefore, the expected overall mark is 20 − 16 = 4 marks.

(iii) Var(X) = npq = 4 ∗ 1

4 ∗ 3
4 = 0.75

Example 8 If X is a binomially distributed with p = 0.45 and n = 10, find the most
likely value of X to occur.
Solution:
First compute the expected value of X, that is E(X) = np = 10 ∗ 0.45 = 4.5. When this
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 9

is done, try to find the probabilities of values around the mean, in this case 3, 4, 5, 6 as
below.
!
10
P(X = 3) = (0.45)3 (0.55)7 = 0.1664
3
!
10
P(X = 4) = (0.45)4 (0.55)6 = 0.2383
4
!
10
P(X = 5) = (0.45)5 (0.55)5 = 0.2340
5
!
10
P(X = 6) = (0.45)6 (0.55)4 = 0.1595
6

From this you note that X = 4 has the highest probability and as such the most occuring
value, or the mode is 4.

Example 9 In Iganga, 80% of the inhabitants have a particular eye disorder. If 12

people are waiting to see the doctor, what is the most likely number of them to have the
eye disorder?
Solution
This is a binomial distribution with X=the number of people with eye disorder, n = 12
and p = 0.8. First compute E(X) = np = 12 ∗ 0.8 = 9.6. This implies that you find
probabilities of values of X around the mean, say 8, 9, 10, 11.
!
12
P(X = 8) = (0.80)8 (0.20)4 = 0.1328
8
!
12
P(X = 9) = (0.80)9 (0.20)3 = 0.2362
9
!
12
P(X = 10) = (0.80)10 (0.20)2 = 0.2834
10
!
12
P(X = 11) = (0.45)11 (0.20)1 = 0.2061
11

Thus the most likely number of people with eye disorder is 10 since it has the highest
probability.

Example 10 Among companies doing highway or bridge construction, 80% test

employees for substance abuse. A study involves the random selection of 10 such
companies. What is the mean and standard deviation for the number of companies in
this sample that test for substance abuse.
Solution:
This is binomial, since there are 2 outcomes (test or not test), a fixed number
√ of trials (10)
and independence. So µ = 10∗0.8 = 8, σ2 = 10∗0.8∗0.2 = 1.6, and σ = 1.6 = 1.2649
10 Elements of Probability & Statistics– MTH1202

Example 11 30% of college students own video cassette recorders. The Telektronic
Company produced a videotape and sent pilot copies to 20 college students. What is the
mean and standard deviation of this sample?
Solution:
Once again -this is binomial, so µ = 20 ∗ 0.3 = 6, σ2 = 20 ∗ 0.3 ∗ 0.7 = 4.2, and
√
σ = 4.2 = 2.04939.

Example 12 In the testing of a product from a production line, the probability that the
product survives the test is 34 . What is the probability that exactly two of the next four
products tested will also survive the test?
Solution
Assuming that the tests are independent, then X ∼ binomial(4, 43 ). Therefore
!  2  2
4 3 1 27
P(X = 2) = = .
2 4 4 128

5.4 The Poisson Distribution

A random variable X, taking on one of the values 0, 1, 2, . . . , is said to be a
Poisson random variable with mean λ, λ > 0, if its probability mass function is
given by

e−λ λx
f (x, λ) = P(X = x) = , x = 0, 1, 2, . . . (5.3)
x!
The symbol e stands for a constant approximately equal to 2.7183. It is a famous
constant in mathematics, named after the Swiss mathematician L. Euler, and it
is also the base of the so-called natural logarithm.
It is clear that f (x) ≥ 0 for each value of x. In order to verify that the function
f (x, λ) defined in equation (5.3) satisfies the requirements of every p.m.f , it must
∞
X
be shown that f (x, λ) = 1 . It is known from elementary calculus that for any
x=0
real number λ,
∞
X λx
eλ =
x=0
x!

Therefore,
∞ ∞
X X λx
f (x, λ) = e −λ
= e−λ eλ = 1
x=0 x=0
x!

Examples of Poisson random variables are:

• The number of misprints on a page (or a group of pages) of a book.

• The number of people in a community living to 100 years of age.

 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 11

• The number of wrong telephone numbers that are dialed in a day.

• The number of transistors that fail on their first day of use.
• The number of customers entering a post office on a given day.
• The number of α-particles discharged in a fixed period of time from some
radioactive particle.

Mean and Variance

We have stated that the distribution for which the p.m.f is given by Eq. (5.3)
is defined to be the Poisson distribution with mean λ. In order to justify this
definition, we must show that λ is, in fact, the mean of this distribution. The
mean E(X) is specified by the following infinite series:
∞
X
E(X) = x f (x, λ)
x=0

since the term corresponding to x = 0 in this series is 0, we can omit this term
and can begin the summation with the term x = 1. Therefore,
∞ ∞ ∞
X X e−λ λx X e−λ λx−1
E(X) = x f (x, λ) = x =λ
x! (x − 1)!
x=1 x=1 x=1

Now letting y = x − 1 in this summation, we obtain

∞ ∞
X e−λ λ y X λy
E(X) = λ = λe−λ =λ·1=λ
y=0
y! y=0
y!

The variance of a Poisson distribution can be found by a technique similar

to the one that has just been given. We begin by considering the following
expectation:
∞
X ∞
X
E[X(X − 1)] = x(x − 1) f (x, λ) = x(x − 1) f (x, λ)
x=0 x=2
∞ ∞
X e−λ λx X e−λ λx−2
= x(x − 1) = λ2
x=2
x! x=2
(x − 2)!

If we let y = x − 2, we obtain
∞
X e−λ λ y
E[X(X − 1)] = λ2 = λ2 (5.4)
y=0
y!

Since E[X(X − 1)] = E(X2 ) − E(X) = E(X2 ) − λ, it follows from Eq. (5.4) that
E(X2 ) = λ2 + λ. Therefore
Var(X) = E(X2 ) − [E(X)]2 = λ2 + λ − λ2 = λ
12 Elements of Probability & Statistics– MTH1202

Thus, for a Poisson distribution for which the p.m.f is defined by Eq. (5.3), we
have established the fact that both the mean and variance are equal to λ.
Example 13 Suppose that the average number of accidents occurring weekly on a
particular stretch of a highway equals 3. Calculate the probability that there is at least
one accident this week.
Soln: Let X denote the number of accidents occurring on the stretch of highway in
question during this week. Now, X ∼ Poisson(λ), where λ = 3.

e−3 30
P(X ≥ 1) = 1 − P(X = 0) = 1 − = 1 − e−3 = 0.9502
0!
Example 14 Consider an experiment that consists of counting the number of α particles
given off in a one-second interval by one gram of radioactive material. If we know from
past experience that, on the average, 3.2 such α-particles are given off, what is a good
approximation to the probability that no more than 2 α-particles will appear?
Soln: The number of α-particles given off will be a Poisson random variable with
parameter λ = 3.2. Hence the desired probability is

(3.2)2 −3.2
P(X ≤ 2) = e−3.2 + 3.2e−3.2 + e = 0.380
2
Example 15 If the average number of claims handled daily by an insurance company
is 5, what proportion of days have less than 3 claims?
Soln: Suppose that the number of claims handled daily, call it X, is a Poisson random
variable. Since E(X ) = 5, the probability that there will be fewer than 3 claims on any
given day is

51 52
P(X < 3) = P(X = 0) + P(X = 1) + P(X = 2) = e−5 + e−5 + e−5 = 0.1247
1! 2!
Example 16 The average number of field mice per acre of a field is estimated to be 10.
Find the probability that a given acre will contain more than 15 mice.
Soln: Let X be the number of field mice, X ∼ Poisson(10).
P(X > 15) = 1 − P(X ≤ 15)
Note: We can use cumulative Poisson table to get the answer easily. Look at the
table at end of this Topic.
P(X > 15) = 1 − P(X ≤ 15) = 1 − 0.9513 = 0.0487.
Example 17 Suppose that radioactive particles strike a certain target in accordance
with a Poisson process at an average rate of 3 particles per minute. Determine the
probability that 10 or more particles will strike the target in a particular 2-minute
period.
Soln: In a Poisson process, the number of particles striking the target in any particular
one-minute period has a Poisson distribution with mean λ. since the mean number of
strikes in any one-minute period is 3, it follows that λ = 3. Therefore, the number of
strikes X in any 2-minute period will have a Poisson distribution with mean 6. It can
be found from the table of the Poisson distribution that P(X ≥ 10) = 0.0838.
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 13

Example 18 Telephone calls enter a school switchboard on average of two every three
minutes. What is the probability of at least 5 calls arriving in a 9-minute period?
Soln: Let X be the number of calls arriving in a 9-minute period. X ∼ Poisson(λ),
where λ = 6.
P(X ≥ 5) = 1 − P(X ≤ 4) = 1 − 0.285 = 0.715

The Distribution of two Independent Poisson Random Variables

Theorem 1 The sum of two independent Poisson variables with parameters λ1 and λ2
respectively is a Poisson random variable with parameter λ1 + λ2 .

e−λ1 λx1
Proof 1 If X ∼ Poisson(λ1 ), then P(X = x) = so that
x!

P(X = 0) =e−λ1
P(X = 1) =e−λ1 · λ1
e−λ1 · λ21
P(X = 2) =
2!
y
e−λ2 λ2
Also if Y ∼ Poisson(λ2 ), then P(Y = y) = so that
y!

P(Y = 0) =e−λ2
P(Y = 1) =e−λ2 · λ2
e−λ2 · λ22
P(Y = 2) =
2!
If X and Y are two independent Poisson random variables, then you will have

P(X + Y = 0) =P(X = 0) · P(Y = 0) = e−λ1 · e−λ2 = e−(λ1 +λ2 )

P(X + Y = 1) =P(X = 0) · P(Y = 1) + P(X = 1) · P(Y = 0)
=e−λ1 · e−λ2 · λ2 + e−λ1 · e−λ2 · λ1 = e−(λ1 +λ2 ) · (λ1 + λ2 )
P(X + Y = 2) =P(X = 0) · P(Y = 2) + P(X = 2) · P(Y = 0) + P(X = 1) · P(Y = 1)
(λ1 + λ2 )2
=e−(λ1 +λ2 ) ·
2!
And the distribution of X + Y is as in the table below which shows that X + Y is a
Poisson distribution with parameters λ1 + λ2 .

x+y 0 1 2 ···
−(λ1 +λ2 ) −(λ1 +λ2 ) −(λ1 +λ2 ) (λ1 +λ2 )2
P(X + Y = x + y) e e · (λ1 + λ2 ) e · 2! ···

Example 19 A company has 2 machines A and B. On average there are 0.8 breakdowns
per week on machine A and 1.2 breakdowns on machine B. What is the probability of
14 Elements of Probability & Statistics– MTH1202

there being a total of 2 breakdowns on these two machines in a given week?

Soln: Let X= Breakdowns on A per week and Y= Breakdowns on B per week. Then
X ∼ Poisson(0.8) and Y ∼ Poisson(1.2). Let T= Total breakdowns per week. T ∼
Poisson(0.8 + 1.2).
22
P(T = 2) = e−2 = 0.2707
2!
Example 20 The center page of the Monitor magazine consists of 1 page of local news
and 1 page of adverts. The number of misprints in the local news page has a Poisson
distribution with mean 2.3 and the number of misprints in the adverts pages has a
Poisson distribution with mean 1.7.
(a) Find the probability that on the center page, there will be
(i) no misprints
(ii) more than 5 misprints
(b) Find the smallest integer n such that the probability that there are more than n
misprints on the center page is less than 0.1.
Solution Let X be the number of misprints on the local page and Y the number of
misprints on the adverts page. Then both X and Y follow a Poisson distribution. So, you
have X ∼ Poisson(2.3) and Y ∼ Poisson(1.7). If T is the number of misprints on the
center page, then T = X + Y is a Poisson distribution with parameter λ = 2.3 + 1.7 = 4.
Thus,
(a) (i) P(no misprints)= P(T = 0) = e−4 = 0.018
(ii) P(T > 5) = 1 − P(T ≤ 5) = 1 − 0.7851 = 0.215(ans 0.7851 is from tables)
(b) To find the smallest integer n you compute the following probabilities

P(T > 5) = 0.21 > 0.1

P(T > 6) = 0.111 > 0.1
P(T > 7) = 0.00051 < 0.1

Thus the smallest integer n such that the probability that there are more than n
misprints on the center page is less than 0.1 is n = 7.

Example 21 An insurance company offers policies covering houses, the contents of

houses and cars. The number of claims received per day of these policies may be assumed
to have Poisson distributions with means 2, 3 and 5 respectively. Find the probability
that in a day the company will receive more than 12 claims in total.

Example 22 On average a certain intersection result in 3 traffic accidents per month.

What is the probability that in any given month at this intersection (a) exactly 5
accidents will occur (b) less than 3 accidents will occur (c) at least 2 accidents will
occur. Solution: Let X be the number of accidents that occur at this intersection. Then
−λ 5
X ∼ Poisson(3) P(X = 5) = e 5!3
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 15

Example 23 A certain kind of sheet metal has on average five defects per 10 square feet.
If we assume a Poisson distribution, what is the probability that a 15 square foot sheet
of metal will have at least six defects? Solution: Let X denote the number of defects
in a 15 square foot sheet of the metal. Then, since the unit of area is 10 square feet,
t=15/10=1.5 this implies that λ0 = λt = 5 ∗ 1.5 = 7.5. Thus P(X ≥ 6) = 1 − P(X ≤
5) = 1 − 0.2414 = 0.7586.

The Poisson Approximation to the Binomial Distribution

We shall now show that when the value of n is large and the value of p is close
to 0, the binomial distribution with parameters n and p can be approximated by
a Poisson distribution with mean np. This approximation is most adequate for
practical purposes and quite accurate if n ≥ 20 and p ≤ 0.05 and it is very good
if n ≥ 100 and np ≤ 10.
Suppose that a random variable X has a binomial distribution with param-
eters n and p, and let P(X = x) = f (x, n, p), for any given value of x. Then for
x = 1, 2, . . . , n,

n(n − 1) · · · (n − x + 1) x
!
n
f (x, n, p) = px (1 − p)n−x = p (1 − p)n−x
x x!

If we let λ = np, then f (x, n, p) can be rewritten in the following form:

λx n n − 1 n−x+1 λ n λ −x
   
f (x, n, p) = · ··· 1− 1−
x! n n n n n
We shall now let n → ∞ and p → 0 in such a way that the value of the product
np remains equal to the fixed value λ throughout this limiting process. Since
the values of λ and x are held as fixed as n → ∞, then

n n−1 n−x+1 λ −x
 
lim · ··· 1− =1
n→∞ n n n n
Furthermore, it is known from elementary calculus that

λ n
 
lim 1 − = e−λ
n→∞ n
It now follows that for any nonnegative integer x,

e−λ λx
f (x, n, p) →
x!
The expression on the right is the p.m.f f (x, λ) of the Poisson distribution with
mean λ. Therefore when n is large and p is close to zero, the value of the p.m.f
f (x, n, p) of the binomial distribution can be approximated, x = 0, 1, 2, . . ., by the
value of the p.m.f f (x, λ) of the Poisson distribution for which λ = np.
16 Elements of Probability & Statistics– MTH1202

Example 24 A factory packs bolts in boxes of 500. The probability that a bolt is
defective is 0.002. Find the probability that a box contains 2 defective bolts. Solution
Let X be the number of defective bolts in a box. Then X is binomially distributed with
n = 500 and p = 0.002. This problem solved as a pure binomial distribution has the
following solution.
!
500
P(X = 2) = (0.002)2 (0.988)498 = 0.184
2

By Poisson approximation, note that n = 500 is large and p = 0.002 is small. Also
λ = np = 500 ∗ 0.002 = 1. This

e−λ λx e−1
P(X = x) = ⇒ P(X = 2) = = 0.184
x! 2!

Both answers agree to 3 decimal places, but may seem much easier in Poisson approxi-
mation.

Example 25 A manufacturer of Christmas tree light bulbs knows that 2% of his are
defective. Find the probability that a box of 100 bulbs contains at most 3 defective bulbs.
Solution Since n = 100 (large) and p = 2% = 0.02 (small) then you approximate with
Poisson distribution, with λ = np = 100 ∗ 0.02 = 2. Thus

3
X
P(X ≤ 3) = p(x; 2) = 0.857
x=0

*from tables.
Using the binomial distribution we have

3
X
P(X ≤ 3) = b(x; 100, 0.02)
x=0

compute!!

Example 26 Suppose that on average 1 person in 1000 is HIV positive. Find the
probability that a random sample of 8000 people will yield less than 7 HIV-positive
cases. Solution This is a binomial distribution problem where n = 8000 and p = 0.001.
Since n is quite large and p is close to zero, you approximate it with a Poisson,
λ = np = 8000 ∗ 0.001 = 8 and so

6
X 6
X
P(X < 7) = b(x; 8000, 0.001) = p(x; 8) = 0.3134
x=0 x=0

from tables.
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 17

5.5 The Geometric Distribution

Imagine a situation where an "experiment" is repeated until a "success" occurs
for the first time. If X counts the number of repetitions needed until the first
success, then X has a geometric distribution, e.g. the number of tosses of a coin
until a first tail is obtained or a process where components from a production
line are tested, in turn until the first defective item is found. The geometric
distribution is another important discrete distribution. It is important to note
that if a success occurs at the nth trial then the first n − 1 trials are failures.
A random experiment is said to be a geometric experiment if:

1. the outcome of each trial can be classified into one of the two categories,
success or failure,

2. each trial has the same probability p of success,

3. the trials are mutually independent,

4. the trials are repeated until one success is obtained.

If X is the number of trials required until a success is obtained, then X

is said to be a geometric random variable. Consider performing a series of
independent trials each with a constant probability p of success and q of failure,
where q = 1 − p. Let X be a random variable "the number of trials up to and
including the first success". Then:

• P(X = 1) = P(success on first trial)=p

• P(X = 2) = P(success on second trial)=qp

• P(X = 3) = P(success on third trial)=q2 p

• P(X = 4) = P(success on fourth trial)=q3 p

Thus if an experiment is performed x times and success occurs on the x − th

time, then the probability distribution of the random variable X, the number of
the trials on which the first success occurs is given by:

P(X = x) = (1 − p)x−1 p = qx−1 p, x = 1, 2, . . .

Thus a random variable X is said to have a geometric distribution with parameter

p, 0 < p < 1, if its p.m.f is given by:

P(X = x) = g(x, p) = pqx−1 , x = 1, 2, . . .

where q = 1 − p and we write X ∼ Geo(p).

It is always desirable to show that such a distribution is indeed a probability
∞
X
mass function. That is g(x, p) ≥ 0 for all x and g(x, p) = 1.
x=1
18 Elements of Probability & Statistics– MTH1202

Now
∞
X ∞
X ∞
X
g(x, p) = pqx−1 = p qx−1
x=1 x=1 x=1

The expression under summation is a sum of geometric series and is equal to

1
Thus,
1−q
∞
X p
pqx−1 = =1
1−q
x=1

Hence g(x, p) = pqx−1 , x = 1, 2, 3, . . ., is a probability mass function.

Note: If we define X to be the number of failures before we get a success,

then
P(X = x) = (1 − p)x p, x = 0, 1, 2, . . .

If X is a geometrically distributed random variable with parameter p and

probability distribution P(X = x) = pqx−1 , x = 1, 2, 3, . . . ; then:

1. P(X = r) = pqr−1

2. P(X ≤ r) = P(success at some trial in the first r trials)

=1 − p(no success in first r trials)=1 − qr

3. P(X > r) = 1 − P(X ≤ r) = 1 − (1 − qr ) = qr

4. P[(X > a + b)|(X > a)] = P(X > b) because

qa+b
P[(X > a + b)|X > a] = a = qb = P(X > b).
q

e.g. P[(X > 10)|X > 7)] = P(X > 3) = q3

The geometric distribution is related to the binomial distribution in that both
are based on independent trials in which the probability of success is constant
and equal to p. However, a geometric random variable is the number of trials
until the first success ( or up to a first success), whereas a Binomial random
variable is the number of successes in n trials.

Mean and Variance of a Geometric Distribution

If X is a geometric random variable with probability of success p, then the mean

or expectation of X is given by

1
E(X) =
p
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 19

Proof:
X
E(X) = xP(X = x)
allx
X∞
= xpqx−1
x=1
X∞
=p xqx−1
x=1
∞ ∞
X d x d X x
=p (q ) = p q
dq dq
x=1 x=1
!
d q p p 1
=p = 2
= 2 =
dq 1 − q (1 − q) p p

Thus, if independent trials, having a common probability p of being success-

ful, are performed until the first success occurs, then the expected number of
1
required trials is equal to . For example, we could expect to roll a fair die 6
p
times in order to obtain the value 1.

Definition 3 If X is a geometric random variable with probability of success p, then

the variance of X is given by
q
Var(X) = 2
p
.

Proof: First derive an expression for E[X(X − 1)].

X
E[X(X − 1)] = x(x − 1)P(X = x)
allx
∞
X ∞
X
= x(x − 1)pqx−1 = pq x(x − 1)qx−2
x=1 x=2
∞ ∞
q2
!
X d2 x d2 X x d2
=pq (q ) = pq 2 (q ) = pq 2
x=2
dq2 dq x=2 dq 1 − q
2
! !
d 2q − q 2
=pq 2
= pq
dq (1 − q) (1 − q)3
2pq 2pq 2q
= 3
= 3 = 2
(1 − q) p p

Thus,

2q 2q 1 2 1
E[X(X − 1)] = E(X2 ) − E(X) = 2
⇒ E(X2 ) = 2 + = 2 −
p p p p p
20 Elements of Probability & Statistics– MTH1202

Using the definition of variance, you get

Var(X) =E(X2 ) − [E(X)]2

2 1 1
= 2− − 2
p p p
1 1 1−p
= 2− =
p p p2
q
= 2
p

Example 27 Each time a player plays a certain card game called "patience", there is a
probability of 0.15 that the game will work out successfully. The probability of success
in any game is independent of the outcome of any other game. A player plays games of
patience until a game works out successfully.

a) Find the probability that she plays 4 games altogether.

b) Find the probability that at most 6 games are played.

c) Write down the expected number of games that are played.

Solution:
The number of games played, X, has a geometric distribution.
X ∼ Geo(0.15).

a) If she plays 4 games altogether she must lose the first 3 games and be successful
on the 4th game. So

P(X = 4) = 0.853 × 0.15 = 0.0921

b) P(at most 6 games) = P(X = 1) + P(X = 2) + · · · + P(X = 6).

This would give the correct answer but involves quite a lot of working out.
Instead, P(at most 6 games) = 1 − P(more than 6 games).
But if more than 6 games are needed then the she must have been unsuccessful at
the first 6 goes.
So P(at most 6 games) = 1 − P(more than 6 games) = 1 − 0.856 = 0.623

1 1
c) E(X) = = = 6.67
p 0.15

Example 28 If 40% of people in a certain population have blood type A, find the
probability that in a sequential testing of randomly selected people, the first person with
type A blood is found on the fourth test.
Solution:
Let X be the number of tests up to the first person of type A blood. Then X ∼ Ge0(0.4).
We wish to find P(X = 4) = (0.6)3 (0.4)
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 21

Example 29 In a certain manufacturing process it is known that, on average, 1 in

every 100 items is defective. What is the probability that 5 items are inspected before a
defective item is found?
Solution
Using the geometric distribution with p=1/100=0.01, and P(X = 5) = (0.01)(0.99)4 =
0.0096.

Example 30 A coin is biased so that the probability of obtaining a head is 0.6. If X is

the random variable "X=the number of tosses up to and including the first head", find:
(a) P(X ≤ 4),
(b) P(X > 5),
(c) P[(X > 8)|(X > 5).
Solution
Using the definition, P(X = x) = g(x; p) = pqx−1 , x=1,2,... together with p = 0.6, q =
0.4 and P(X > a) = qa you get:
(a) P(X ≤ 4) = 1 − q4 = 1 − (0.4)4 = 0.9744
(b) P(X > 5) = q5 = (0.4)5 = 0.01024
(c) P[(X > 8)|(X > 5) = P(X > 3) = q3 = (0.4)3 = 0.064

Example 31 In a particular game, a player can get out of check only if she gets two
heads when she tosses two coins.
(a) Find the probability that more than 6 attempts are needed to get out of check.
(b) What is the smallest value of n if there is to be at least a 90% chance of getting
out of check on or before the n − th attempt.
Solution:
P(2heads when two coins are tossed)= 14 . So P(success) = 1
4 and q = 34 . If X= the
number of attempts required to get out check, then
(a) P(X > 6) = q6 = ( 34 )6 = 0.17

(b) P(X > n) = ( 34 )n . So P(X ≤ n) = 1 − ( 34 )n . To be at least 90% sure that one is

out of check, ⇒ 1 − ( 34 )n ≥ 0.9 ⇒ ( 34 )n ≤ 0.1. If you take logs to base 10 on both
sides you get

log10 0.1
n log10 0.75 ≥ log10 0.1 ⇒ n ≥ = 8.0039
log10 0.75

Therefore the smallest n is 9.

Example 32 The probability that a marksman hits a target eye is 0.4 for each shot,and
each shot is independent of the other. Find the:
22 Elements of Probability & Statistics– MTH1202

(a) probability that he hits the target for the first time on his fourth attempt,

(b) mean number of shots needed to hit the target and standard deviation,

(c) most likely number of shots until he hits the target.

Solution: Let X be the number of attempts up to and including the first target. Then
P(X = x) = pqx−1 , x = 1, 2, 3, ...; with p = 0.4 and q = 0.6 .

(a) P(X = 4) = q3 p = (0.6)3 (0.4) = 0.0864

q
(b) E(X) = 1
p =
0.4 = 2.5 and Var(X) =
1
p2
= 0.6
(0.4)2
= 3.75. Therefore the standard
√
deviation = 3.75 = 1.94.

(c) P(X = 1) = (0.4)(0.6)0 = 0.4, P(X = 2) = (0.4)(0.6) = 0.24 and

P(X = 3) = (0.4)(0.6)2 = 0.144.

Therefore, the most likely number of shots is 1, since the probabilities are decreasing and it
is the one with the highest probability. Generally it is noted that p > qp > q2 p > q3 p > ...
so the mode of the geometric distribution (or the most likely value) occurs at X = 1 with
probability P(X = 1) = p.

5.6 The Negative Binomial

This is a generalization of the geometric distribution. It gives the probability
distribution of the kth success at the xth trial in a series of independent Bernoulli
trials. Let X denote the trial at which the kth success occurs, where k is a fixed
integer. Then
!
x−1
P(X = x|k, p) = pk qx−k , x = k, k + 1, . . .
k−1

and we say that X has a negative binomial(k, p) distribution. In most of the

literature, it is denoted as b∗ (x; k, p).
The negative binomial distribution is sometimes defined in terms of the
random variable Y =number of failures before kth success. This formulation is
statistically equivalent to the one given above in terms of X =trial at which the
kth success occurs, since Y = X -k. The alternative form of the negative binomial
distribution is:

k+y−1
!
P(Y = y) = pk q y , y = 0, 1, . . .
y

Example: Find the probability that a person tossing 3 coins will get either
all heads or all tails for the second time on the fifth toss.
Solution
Using the Negative Binomial with x = 5, k = 2 and p = 14 , we have:
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 23

!
5−1
b∗ (5; 2, 0.25) = 0.252 0.753 = 0.1445
2−1
Example: An oil company conducts a geological study that indicates that
an exploratory oil well should have a 20% chance of striking oil. What is the
probability that the first strike comes on the third well drilled?
Solution:

This a Negative Binomial problem with k = 1, p = 0.2, q = 1 − p = 0.8, x = 3.

This gives: !
3−1
b (3; 1, 0.2) =
∗
0.21 0.82 = 0.128
1−1
Note: This problem can as well be solved using a Geometric distribution. It
should be observed that when k = 1, the Negative Binomial distribution reduces
to a Geometric distribution.

5.7 The Hypergeometric Distribution

Suppose we are sampling without replacement from a batch of items containing
a variable number of defectives. We are essentially assuming that we know
the probability p that a given item is defective but not the actual number of
defective items contained in the batch. The number of defective items in the
batch is a random variable in this case. When we sample from the batch, we are
left with:

1. a smaller batch:

2. a (possibly) smaller (but still variable) number of defective items. The

number of defective items is still a random variable.

While the probability of finding a given number of defectives in a sample drawn

from the second batch will, (in general) be different from the probability of
finding a given number of defectives in a sample drawn from the first batch,
sampling from both batches may be described by the Binomial distribution for
which:
P(X = r) =n Cr pr (1 − p)n−r
Sampling in this case varies the values of n and p in general but not the
underlying distribution describing the sampling process.

Example 33 A batch of 100 piston rings is known to contain 10 defective rings. If two
piston rings are drawn from the batch, write down the probabilities that:

1. the first ring is defective;

2. the second ring is defective given that the first one is defective.

Soln:
24 Elements of Probability & Statistics– MTH1202

10 1
1. The probability that the first ring is defective is clearly = .
100 10
2. Assuming that the first ring selected is defective and we do not replace it, the
9
probability that the second ring is defective is equally clearly = 1
99 11
The Hypergeometric distribution may be thought of as arising from sam-
pling from a batch of items where the number of defective items contained in
the batch is known.
Essentially the number of defectives contained in the batch is not a random variable, it
is fixed.
The calculations involved when using the Hypergeometric distribution are
usually more complex than their Binomial counterparts. If we sample without
replacement we may proceed in general as follows:

• we may select n items from a population of N items in N Cn ways;

• we may select r defective items from M defective items in M Cr ways;

• we may select n − r non-defective items from N − M non-defective items

in N−M Cn−r ways;
M
• hence we may select n items containing r defectives in Cr ×N−M Cn−r
ways.

• hence the probability that we select a sample of size n containing r defective

items from a population of N items known to contain M defective items is
M
Cr ×N−M Cn−r
NC
n

Definition 1 The distribution given by

M
Cr ×N−M Cn−r
P(X = r) = NC
, r = 0, 1, . . . , n
n

which describes the probability of obtaining a sample of size n containing r defective

items from a population of size N known to contain M defective items is known as the
Hypergeometric distribution.

Example 34 A batch of 10 rocker cover gaskets contains 4 defective gaskets. If we draw

samples of size 3 without replacement, from the batch of 10, find the probability that a
sample contains 2 defective gaskets.
M
Cr ×N−M Cn−r
Soln: Using P(X = r) = NC
, we know that N = 10, M = 4, n = 3 and
n
r = 2.
4
C2 ×6 C1 6×6
Hence P(X = 2) = 10 C
= = 0.3
3 120
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 25

Example 35 A committee of size 5 is to be selected at random from three men and five
women. Find the probability distribution for the number of:

(a) men on the committee,

(b) women on the committee.

Soln:
(i) Let the variable X be the number of men on the committee. X has a hypergeometric
distribution in which N = 8, n = 5, M = 3 and r = 0, 1, 2, 3. Thus,
3
C0 ×5 C5 1
P(X = 0) = 8C
=
5 56
3
C1 ×5 C4 15
P(X = 1) = 8 =
C5 56
3
C2 ×5 C3 30
P(X = 2) = 8 =
C5 56
3
C3 ×5 C2 10
P(X = 3) = 8 =
C5 56

In tabular form we have a hypergeometric distribution of X as

x 0 1 2 3
1 15 30 10
P(X=x) 56 56 56 56

(ii) If we let Y to be the number of women on the committee, the distribution can
easily be seen from (i) as:

y 2 3 4 5
10 30 15 1
P(Y=y) 56 56 56 56

Note: The committee cannot have one or no woman as this will require to have more
than three men given that the committee must be composed of five individuals.

Example 36 Boxes of 40 light bulbs are called acceptable if they contain no more than
3 defectives. The procedure of sampling the box is to select 5 bulbs at random and reject
a box if a defective bulb is found. What is the probability that exactly one defective bulb
will be found in a sample if there are three defective bulbs in the entire box?
M
Cr ×N−M Cn−r
Soln: Using P(X = r) = NC
, we know that N = 40, M = 3, n = 5 and
n
r = 1.
3
C1 ×37 C4
Hence P(X = 1) = 40 C
= 0.3011
5

It is possible to derive formulae for the mean and variance of the Hyper-
geometric distribution. However, the calculations are more difficult than their
Binomial counterparts.
26 Elements of Probability & Statistics– MTH1202

5.7.1 The Mean and Variance of a Hypergeometric Distribution

Let X be a hyper geometric random variable with p.m.f

M
Cx ×N−M Cn−x
P(X = x) = NC
, x = 0, 1, . . . , n
n

The expectation (mean) and variance of of X are given by

N−n M
E(X) = np and Var(X) = np(1 − p) where p =
N−1 N
Proof: From the definition of expectation of a discrete random variable
n M
X X Cx ×N−M Cn−x
E(X) = xP(X = x) = x NC
x=0 n
allx
n N−M
X M! Cn−x
= x NC
x!(M − x)! n
x=1
n N−M
X (M − 1)! Cn−x
=M N
(x − 1)!(M − x)! Cn
x=1
n M−1 N−M
X C x−1 × Cn−x
=M NC
x=1 n

Letting y = x − 1, we have

n−1 M−1
X C y ×N−M Cn−1−y
E(X) = M NC
y=0 n

Writing
N N−1
N−M
Cn−1−y =(N−1)−(M−1) Cn−1−y and N Cn = × Cn−1
n
we have
n−1 M−1
nM X C y ×(N−1)−(M−1) Cn−1−y
E(X) = N−1 C
N y=0 n−1

The expression inside the summation represents the total of all probabilities of a
hypergeometric experiment when a sample of n − 1 items is selected at random
from N − 1 items, of which M − 1 are defective, and hence equals to 1. Therefore,
the expectation is
nM
E(X) =
N
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 27

To compute the variance, we first derive the expression for E(X(X − 1)). Thus
from the definition of the expectation, we have
n M
X X Cx ×N−M Cn−x
E[X(X − 1)] = x(x − 1)P(X = x) = x(x − 1) NC
x=0 n
allx
n N−M
X (M − 2)! Cn−x
=M(M − 1)
x=2
(x − 2)!(M − x)! N Cn
n−2
n(n − 1)M(M − 1) X M−2 C y ×(N−2)−(M−2) Cn−2−y
= N−2 C
N(N − 1) y=0 n−2

The expression inside the summation is the sum of (n − 2) terms of the hyperge-
ometric distribution and is equal to 1. Hence

M(M − 1)
E[X(X − 1)] = E(X2 ) − E(X) = n(n − 1)
N(N − 1)

Thus
M(M − 1) nM
E(X2 ) = n(n − 1) +
N(N − 1) N

Using the definition of variance we obtain

Var(X) =E(X2 ) − [E(X)]2

M(M − 1) nM n2 M2
=n(n − 1) + −
N(N − 1) N N2
" #
nM (M − 1) nM
= (n − 1) +1−
N (N − 1) N
" #
nM (N − M)(N − n)
=
N N(N − 1)
nM M N−n
 
= 1−
N N N−1
M
If we set p = , then the mean of the hypergeometric distribution coincides
N
with the mean of the binomial distribution, and the variance of the hyper
N−n
geometric distribution is times the variance of the binomial distribution.
N−1
Therefore, we can approximate the hypergeometric distribution by using bino-
nM nM M
mial distribution with mean µ = np = and variance σ2 = npq = (1 − ).
N N N
This is so when n is small relative to N.
We note

• the mean of a hypergeometric distribution is the same as that of a binomial.

28 Elements of Probability & Statistics– MTH1202

• The variance of a hypergeometric distribution is equal to the variance of a

N−n
binomial distribution multiplied by a factor f = . Since f < 1, the
N−1
variance of a hypergeometric distribution is smaller than that of a binomial
distribution. The factor f is called the finite population correction.

5.7.2 Applications
The hypergeometric distribution is used for calculating probabilities for samples
drawn from relatively small populations and without replication, which means
that an item’s choice of being selected increases on each trial.
For sampling with replication, in which case an item has an equal chance of
being selected, the binomial distribution should be used. If N >> n, in which
case the population size of interest is too large compared to the sample size,
f ≈ 1 (the finite population correct can be neglected). The probabilities from
a hypergeometric distribution and those of a binomial distributions are no
different.
Example 37 In the manufacture of car tyres, a particular production process is known
to yield 10 tyres with defective walls in every batch of 100 tyres produced. From a
production batch of 100 tyres, a sample of 4 is selected for testing to destruction. Find:

1. the probability that the sample contains 1 defective tyre

2. the expectation of the number of defectives in samples of size 4
3. the variance of the number of defectives in samples of size 4.
Soln:Sampling is clearly without replacement and we use the Hypergeometric distribu-
tion with N = 100, M = 10, n = 4, x = 1 and p = 0.1. Hence:
M
Cx ×N−M Cn−x
1. P(X = x) = NC
gives
n
10 100−10
C1 × C4−1 10 × 117480
P(X = 1) = 100 C
= ≈ 0.3
4 3921225
2. E(X) = np = 4 × 0.1 = 0.4
N−n 96
3. Var(X) = np(1 − p) = 0.4 × 0.9 × = 0.349
N−1 99
 c
B.K. Nannyonga, H.W. Kayondo, N. Muyinda, & B.N. Kirenga 29

Table A.1 Binomial Probability Table 729

!
r
Table A.1 (continued) Binomial Probability Sums b(x; n, p)
x=0

p
n r 0.10 0.20 0.25 0.30 0.40 0.50 0.60 0.70 0.80 0.90
15 0 0.2059 0.0352 0.0134 0.0047 0.0005 0.0000
1 0.5490 0.1671 0.0802 0.0353 0.0052 0.0005 0.0000
2 0.8159 0.3980 0.2361 0.1268 0.0271 0.0037 0.0003 0.0000
3 0.9444 0.6482 0.4613 0.2969 0.0905 0.0176 0.0019 0.0001
4 0.9873 0.8358 0.6865 0.5155 0.2173 0.0592 0.0093 0.0007 0.0000
5 0.9978 0.9389 0.8516 0.7216 0.4032 0.1509 0.0338 0.0037 0.0001
6 0.9997 0.9819 0.9434 0.8689 0.6098 0.3036 0.0950 0.0152 0.0008
7 1.0000 0.9958 0.9827 0.9500 0.7869 0.5000 0.2131 0.0500 0.0042 0.0000
8 0.9992 0.9958 0.9848 0.9050 0.6964 0.3902 0.1311 0.0181 0.0003
9 0.9999 0.9992 0.9963 0.9662 0.8491 0.5968 0.2784 0.0611 0.0022
10 1.0000 0.9999 0.9993 0.9907 0.9408 0.7827 0.4845 0.1642 0.0127
11 1.0000 0.9999 0.9981 0.9824 0.9095 0.7031 0.3518 0.0556
12 1.0000 0.9997 0.9963 0.9729 0.8732 0.6020 0.1841
13 1.0000 0.9995 0.9948 0.9647 0.8329 0.4510
14 1.0000 0.9995 0.9953 0.9648 0.7941
15 1.0000 1.0000 1.0000 1.0000
16 0 0.1853 0.0281 0.0100 0.0033 0.0003 0.0000
1 0.5147 0.1407 0.0635 0.0261 0.0033 0.0003 0.0000
2 0.7892 0.3518 0.1971 0.0994 0.0183 0.0021 0.0001
3 0.9316 0.5981 0.4050 0.2459 0.0651 0.0106 0.0009 0.0000
4 0.9830 0.7982 0.6302 0.4499 0.1666 0.0384 0.0049 0.0003
5 0.9967 0.9183 0.8103 0.6598 0.3288 0.1051 0.0191 0.0016 0.0000
6 0.9995 0.9733 0.9204 0.8247 0.5272 0.2272 0.0583 0.0071 0.0002
7 0.9999 0.9930 0.9729 0.9256 0.7161 0.4018 0.1423 0.0257 0.0015 0.0000
8 1.0000 0.9985 0.9925 0.9743 0.8577 0.5982 0.2839 0.0744 0.0070 0.0001
9 0.9998 0.9984 0.9929 0.9417 0.7728 0.4728 0.1753 0.0267 0.0005
10 1.0000 0.9997 0.9984 0.9809 0.8949 0.6712 0.3402 0.0817 0.0033
11 1.0000 0.9997 0.9951 0.9616 0.8334 0.5501 0.2018 0.0170
12 1.0000 0.9991 0.9894 0.9349 0.7541 0.4019 0.0684
13 0.9999 0.9979 0.9817 0.9006 0.6482 0.2108
14 1.0000 0.9997 0.9967 0.9739 0.8593 0.4853
15 1.0000 0.9997 0.9967 0.9719 0.8147
16 1.0000 1.0000 1.0000 1.0000
30 Elements of Probability & Statistics– MTH1202

734 Appendix A Statistical Tables and Proofs

!
r
Table A.2 (continued) Poisson Probability Sums p(x; µ)
x=0

µ
r 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0
0 0.0000 0.0000 0.0000
1 0.0005 0.0002 0.0001 0.0000 0.0000
2 0.0028 0.0012 0.0005 0.0002 0.0001 0.0000 0.0000
3 0.0103 0.0049 0.0023 0.0011 0.0005 0.0002 0.0001 0.0000 0.0000
4 0.0293 0.0151 0.0076 0.0037 0.0018 0.0009 0.0004 0.0002 0.0001
5 0.0671 0.0375 0.0203 0.0107 0.0055 0.0028 0.0014 0.0007 0.0003
6 0.1301 0.0786 0.0458 0.0259 0.0142 0.0076 0.0040 0.0021 0.0010
7 0.2202 0.1432 0.0895 0.0540 0.0316 0.0180 0.0100 0.0054 0.0029
8 0.3328 0.2320 0.1550 0.0998 0.0621 0.0374 0.0220 0.0126 0.0071
9 0.4579 0.3405 0.2424 0.1658 0.1094 0.0699 0.0433 0.0261 0.0154
10 0.5830 0.4599 0.3472 0.2517 0.1757 0.1185 0.0774 0.0491 0.0304
11 0.6968 0.5793 0.4616 0.3532 0.2600 0.1848 0.1270 0.0847 0.0549
12 0.7916 0.6887 0.5760 0.4631 0.3585 0.2676 0.1931 0.1350 0.0917
13 0.8645 0.7813 0.6815 0.5730 0.4644 0.3632 0.2745 0.2009 0.1426
14 0.9165 0.8540 0.7720 0.6751 0.5704 0.4657 0.3675 0.2808 0.2081
15 0.9513 0.9074 0.8444 0.7636 0.6694 0.5681 0.4667 0.3715 0.2867
16 0.9730 0.9441 0.8987 0.8355 0.7559 0.6641 0.5660 0.4677 0.3751
17 0.9857 0.9678 0.9370 0.8905 0.8272 0.7489 0.6593 0.5640 0.4686
18 0.9928 0.9823 0.9626 0.9302 0.8826 0.8195 0.7423 0.6550 0.5622
19 0.9965 0.9907 0.9787 0.9573 0.9235 0.8752 0.8122 0.7363 0.6509
20 0.9984 0.9953 0.9884 0.9750 0.9521 0.9170 0.8682 0.8055 0.7307
21 0.9993 0.9977 0.9939 0.9859 0.9712 0.9469 0.9108 0.8615 0.7991
22 0.9997 0.9990 0.9970 0.9924 0.9833 0.9673 0.9418 0.9047 0.8551
23 0.9999 0.9995 0.9985 0.9960 0.9907 0.9805 0.9633 0.9367 0.8989
24 1.0000 0.9998 0.9993 0.9980 0.9950 0.9888 0.9777 0.9594 0.9317
25 0.9999 0.9997 0.9990 0.9974 0.9938 0.9869 0.9748 0.9554
26 1.0000 0.9999 0.9995 0.9987 0.9967 0.9925 0.9848 0.9718
27 0.9999 0.9998 0.9994 0.9983 0.9959 0.9912 0.9827
28 1.0000 0.9999 0.9997 0.9991 0.9978 0.9950 0.9897
29 1.0000 0.9999 0.9996 0.9989 0.9973 0.9941
30 0.9999 0.9998 0.9994 0.9986 0.9967
31 1.0000 0.9999 0.9997 0.9993 0.9982
32 1.0000 0.9999 0.9996 0.9990
33 0.9999 0.9998 0.9995
34 1.0000 0.9999 0.9998
35 1.0000 0.9999
36 0.9999
37 1.0000