OLUSEGUN AGAGU UNIVERSITY OF SCIENCE & TECHNOLOGY

(OA (OAUSTECH) OKITIPUPA

SCHOOL OF SCIENCE
DEPARTMENT OF MATHEMATICAL SCIENCES

TOPIC: PROBABILITY DISTRIBUTION

OUTLINE

2.1 - INTRODUCTION

2.2 - BINOMIAL DISTRIBUTION

2.3 - POISSON DISTRIBUTION

2.4 - NORMAL DISTRIBUTION

2.5 - COMPARISON OF THE PROPERTIES OF THE THREE DISTRIBUTIONS

2.6 – EXERCISES

OBJECTIVES: At the end of these lectures, students should be able to:

 Understand the concept of random variables and probability distribution

 Identify and solve problems in Binomial and Poisson distribution
 compare the properties of Normal, Binomial & Poisson distribution

INTRODUCTION

In general terms, possibility is a measure of the likelihood that a particular event will occur in
any one trial, or experiment, carried out in prescribed condition. The origin of probability can be
dated back to the 16th century when gambling and games of chance were the usual pastime for
those with money and leisure time. This lecture focuses on elementary theory of probability
distribution. However, before going into this topic, we initially examine the concept of random
variables. The concept of random variable is important because it is the nature or type of random
variable (discrete or continuous) that determines the kind of probability distribution to be
adopted.

1
RANDOM VARIABLES

A function that assigns a numerical value to each outcome in the sample space is called a random
variable. It is a variable whose value is determined by the outcome of a chance experiment.
Random variables are usually denoted by the capital letters X, Y, Z, and so on, and the values
taken by them are denoted by small letters x, y, z, and so on.

A random variable may be either discrete or continuous. A discrete random variable takes only a
finite or countable number of values. For example, in throwing two dice, each numbered 1 to 6,
if we define the random variable X as the sum of the numbers showing on the dice, then X will
take only one of these values: 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 or 12. A continuous random variable,
on the other hand, is one that can take on any value in some interval of values. Thus, the height
of an individual is a continuous variable. It can take any value, depending on the precision of
measurement.

PROBABILITY DISTRIBUTIONS

If each outcome in a sample space is equally likely, then the probability of X, P(x), is given by:

frequency of x
P(x) =
number of outcomes∈t h e sample space

f (x )
=
n

where n is the number of outcomes in the sample space. If the range of a discrete random
variable consists of m different values of x 1 , x 2, x 3, …….., x m, then the sum of the frequencies of
x i is equal to n. that is

f( x 1) + f( x 2) + f( x 3) + . . . . + f( x m) = n

The collection of all probabilities corresponding to the values of the random variable is called the
probability distribution.

If two coins are tossed, the sample space U is given as:

μ = { (T ,T ) , ( H ,T ) , ( T , H ) , ( H , H ) }

The probability distribution f(x) for the random variable representing the number of heads that
turn up can be written as:

x P(x)

0 ¼
2
1 /4
2 ¼

2
An important observation that we can make about the probability distribution above is that the
sum of the probabilities of the values of the random variable is 1. That is,

P(0) + P(1) + P(2) = ¼ +2/4 +1/4 = 1

This is true of the probability distribution for any discrete random variable.

Example 1: A coin is tossed three times. Show the probability distribution for the random
variable indicating the number of tails that turn up.

Solution: The sample space is given as

U = {(HHH), (HTT), (THT), (TTH), (HHT), (THH), (HTH), (TTT)}

Therefore:

x P(x)
1
0 /8
3
1 /8
3
2 /8
1
3 /8

Example 2: Construct a probability distribution for the following frequency distribution

X 1 2 3 4 5 6
F(x) 2 4 6 6 4 2
Solution:

X P(x)
2
1 /24 = 1/12
4
2 /24 = 1/6
6
3 /24 = ¼
6
4 /24 = ¼
4
5 /24 = 1/6
2
6 /24 = 1/12
THE PROBABILITY FUNCTION

This is obtained by plotting the probability of a random variable, P(x) against the random
variable x. This is illustrated below:

BINOMIAL DISTRIBUTION

The Binomial distribution is a Bernoulli distribution with n trials.It is a discrete probability

distributionwhich makes use of integral values, that is, whole numbers which could be finite or countably
infinite.

3
Properties:

(i) The probability distribution associated with a binomial experiment is called a binomial
distribution. A binomial experiment is characterized by the following properties.
(ii) The overall experiment consists of a sequence of n identical experiments called trials.
(iii) Each trial has exactly two possible outcomes. These outcomes are usually refered to as success
and failure i.e p and q.
(iv) The probability of success (and of failure) does not change from one trial to the next.
(v) The trials are independent. This implies that the outcome of one trial does not affect the outcome
of any other trial.

From the above properties, the probability of a success (P) in a binomial experiment is

p=1–q

Similarly, the probability of a failure (q) is

q=1–p

Therefore:

P+q=1

The following are examples of binomial experiments:

Example 3: A fair coin is tossed five times. Each trial can either result in a success (heads up) or failure
(tails up). For this binomial experiment, n = 5, p = ½, q = ½

Example 4: A six-sided die is tossed ten times. Each trial can result in a success (a six showing) or a
failure (a 6 is not showing). For this binomial experiment, we have:

n = 10, p = 1/6, q = 5/6

The random variable that we usually associate with a binomial experiment is the number of successes that
occur in n different trials. This random variable is discrete because the only values it can have are:

{0, 1, 2, 3 . . . n}

PROBABILITY OF SUCCESSES IN A BINOMIAL EXPERIMENT

Let x be the random variable that represents the number of successes in a binomial experiment that has n
trials. If m is an integer between 0 and n, then the probability that exactly m successes will occur in the n
trials is given by:

4
OR

Example 5: From past records, a clothing store has discovered that 30% of the people who enter the store
will actually make a purchase. Eight people enter the store during a one-hour period. Find the
probability that:

a) Exactly four people

b) At least one person
c) All eight people; will make a purchase.

Solution

a) n = 8, p = 0.3 and x = 4

b) To solve this, we initially find the probability that no one will make a purchase. Therefore,
n = 8, p = 0.3 and x = 0

Since , then
p(0) = (0.7)8
= 0.058

The probability that at least one person will make a purchase is ,

c) n = 8, p = 0.3 and x = 8

5
Example 6: Find the binomial distribution for n = 5 if the experiment has (i) P = 0.5 (ii) P = 0.3

Solution: (i) Binomial Distribution, n = 5, P = 0.5

X P(x)
0

(ii) Binomial distribution, n =5, p = 0.5

X P(x)
0

Example 7: A fair die is tossed 5 times. Find the probability that a 4 occurs twice.

Solution:

1 1
P (4) = since there is only one 4 out of the 6 possibilities. So p =
6 6

1 5
q = 1− =
6 6

So n = 5, x = 2

P (4 occurs twice)

6
Example 8: Let X be the event that a head is obtained when a coin is tossed 3 times. Construct the
probability distribution of the experiment.

P = ½, q = 1 - 1/2 = ½

P(no head) =

P(1 head) =

P(2 heads) =

P(3 heads) =

The obtained distribution is thus:

X 0 1 2 3
P(x) 1/8 3/8 3/8 1/8

Example 9: A coin is tossed three times. What is the probability of getting two heads up?

Solution:

n = 3, p = 0.5, m = 2

7
Example 10: If a ludo die is thrown five times, what is the probability of getting

a) One six?
b) Two sixes?
c) Three sixes

Solution:

a)

b)

c)

Example 11: 5 out of every batch of 20 manufactured articles are defective. If 4 articles are picked at
random, what is the probability that they will contain:

a) Exactly 3 defective articles

b) Not more than 2 defective articles

Since 5 out of every 20 is defective,

P (defective) =

(a) P(3 defective) =

=0.0447

(b) P(not more than 2 defective) = P(0 defective) + P(1 defective) + P(2 defective)

8
MEAN AND STANDARD DEVIATION OF A BINOMIAL DISTRIBUTION

Let X be the random variable that represents the number of successes in a binomial experiment that has n
trials. If the probability of success is P, then the mean or expected value of the random variable is

By definition, , where ‘f’ denotes the frequency of occurrence of the random variable. X and
m is the total number of observation.

μ
From the above, can also be rewritten as

Since f/m = p, then μ=Σ px

The simple formula for calculating the mean of a binomial probability distribution is

μ=np

Moreover, the variance and standard deviation of the random variable are

(Variance)

And

(Standard deviation)

Example 12: Find the mean and variance for a binomial distribution with n = 2, p = ½

Solution: (a)

(b)

Example 13: Find the standard deviation of a binomial distribution involving 50 trials for which p = 0.1

Solution:

9
Example 14:

Determine the mean of the following binomial distribution, n = 8, p = 5

X P(x)
0 0.0039
1 0.0312
2 0.1094
3 0.2188
4 0.2734
5 0.2188
6 0.1094
7 0.0312
8 0.0039

This is easily calculated as

POISSON DISTRIBUTION

So far, we have discussed the binomial distribution gives the probability of x successes in ‘n’ independent
trials of an experiment. This led to the formula:

The Binomial distribution becomes tedious to use when n is large and the probability of success very
small and close to zero so that q = 1-p is close to 1, when this happens we use the Poisson as an
approximation of the Binomial. The conditions for a Poisson distribution to be an approximation to the
Binomial are enumerated below:

(i) The overall experiment consists of a sequence of n identical experiments called trials.
(ii) Each trial has exactly two possible outcomes. These outcomes are usually referred to as
success and failure i.e p and q.
(iii) The probability of success (and of failure) does not change from one trial to the next.
(iv) The trials are independent. This implies that the outcome of one trial does not affect the
outcome of any other trial.
(v) The probability of success p, is small and may not be known
(vi) N is large and may be unknown (N ≥ 50)

10
 (vii) The average number of success μ = Np; λ is known and is not large (λ < 5).

If conditions (i) to (vii) above are satisfied, then the Poisson is a good approximation to the Binomial.
Note that conditions (i) to (iv) hold for the Binomial distribution.

When n becomes large and P is fairly small, the calculation of the various probabilities using the above
formula can be very tedious. In this situation, we use the Poisson probability distribution which
approximates the binomial distribution. The Poisson distribution is given by:

OR

where; μ∨ λ=¿ the expected value/mean of the distribution

m or x = the number of successes (m or x = 0, 1, 2, . . .n)

The Poisson distribution approximates closely to the binomial distribution when, in practice,

and

Example15: A machine produces on average 2% defectives. In a random sample of 60 items, determine

the probability of there being three defectives.

Solution:

Example 16: An electronic company manufactures a specific component for every tuner amplifier. It
finds that out of every 1000 components produced, 8 are defective. If the components are packed in
boxes of 250, find:

a) The probability of obtaining 0, 1, 2, 3 defectives in a box

b) The probability that the box contains at least 3 defectives.

Solution:

a) The probability of a defective = 8/1000 = 0.008

11
 b) The probability that the box contains at least 3 defectives is denoted by P(X ≥ 3)

Example 17: Toilet rolls are produced by a manufacturing company in lengths of 500 metres. On the
average, there is one defect for every 200 metres. Assume that the number of defects is distributed as a
Poisson variable; find the probability that a roll has more than 3 defects.

Solution:

The probability that a roll has more than 3 defects is given by P(x >3).

12
MEAN AND VARIANCE OF POISSON DISTRIBUTION

The Poisson distribution is the binomial distribution when the probability of success in n trials is very
small. When this occurs, 1 – p or q becomes close to unity and the variance nP (1 – P) tends to nP. The
implication of this is that the variance of a Poisson distribution nearly equals its mean.

To give an idea of the way in which the binomial tends to approach the Poisson, the table below shows
the binomial distribution for n = 100, P = 0.01 and the binomial for n = 25, p = 0.04, both of these having
μ=1.

X=m Poisson( μ=1 ¿ Binomial(n = 100, p = Binomial(n = 25, p =

0.04) 0.04)
0 0.3679 0.3660 0.3604
1 0.3679 0.3697 0.3754
2 0.1839 0.1849 0.1877
3 0.0613 0.0610 0.0600
4 0.0153 0.0149 0.0137
5 0.0031 0.0029 0.0024
6 0.0005 0.0005 0.0000
7 0.0001 0.0001 0.0000
Total 1.0000 1.0000 0.9999

In the table above, the smaller the probability, P(x), of the binomial the close (1 – p) or q is closer to 1.
Under this situation, the binomial distribution tends to the Poisson distribution. The closeness is greatest
for p = 0.01 then p = 0.04. Students are to calculate the mean and variance of the Poisson distribution in
the table above and compare them.

THE NORMAL DISTRIBUTION

The binomial and Poisson distributions refer to discrete random variables e.g. the number of successes
probable in a trial. Since in most cases we deal with continuous random variables, where we cannot get
the exact values, although we have the approximated values, it means that binomial probability function
cannot take care of such situations.

The suitable distribution for dealing with continuous random variables is the normal distribution. The
normal probability density function otherwise known as the normal curve is represented by the equation

where μis the mean andδ is the standard deviation of the distribution. Moreover, e = 2.71828 and is the
base for natural logarithm. π=3.14159 .

13
Features of the Normal Distribution (μ, σ2)

(i) It is a continuous distribution

(ii) It is symmetrical about x = μ and bell-shaped
(iii) The mean = median = mode = μ
(iv) It is a unimodal distribution
(v) The total area bounded by the curve and the x-axis is 1
(vi) P(a ≤ X ≤ b) is the area under the curve between ordinals X = a and X = b.

To illustrate the role of the standard deviation δ in the shape of the normal curve, the figure 1 below
shows two curves – a solid curve with μ = 0 and δ = 1 and a dotted curve with μ = 0 and δ = 1.5

It is apparent that the dotted curve with δ = 1.5 is lower at the mean and more spread out, while the solid
curve withδ = 1 is higher at the mean and less spread out. In essence, the higher the standard deviation,
the more frequent the values of X that are far from the mean.

In the light of the above, it becomes clear that the normal distribution is completely determined by the
two constants, μ∧δ . The mean' μ ' is located at the centre of the distribution, while the standard
deviation,δ ,measures the spread or variation of the individual measurement.

From figure 2 above, it is apparent that within one sigma on either side of ‘ μ ' , the frequency is decreasing
ever more rapidly, but beyond that point it decreases at a continually lesser rate. Theoretically, the
frequency of occurrence never vanishes entirely, but it approaches zero as X increases indefinitely.

For all normal distributions, approximately 68% of the total region under the curve lies within one
standard deviation from the mean

THE STANDARD NORMAL DISTRIBUTION

A continuous random variable that has a normal distribution with an expected value of μ = 0 and a
standard deviation of δ=1is said to have a standard normal distribution, and the associated probability
density function is called the standard normal probability density function. For this particular function, it
is customary to represent the random variable by the letter Z.

The equation of the standard normal curve or standard normal probability density function is

14
To find the probabilities involving the standard normal distribution, it is convenient to use the standard
normal table.

Example 18: Find the probability that a random variable takes on a ‘z’ value between 0 and 2.14.

Solution:

Check the standard normal curve, look at 2.1 under 0.04.

Example 19: Use the table of standard normal curve to evaluate the following probabilities

Solution:

a)

0.02
1.0 0.3461

This probability is 0.3461 i.e. =0.3461

b) P(-1.43 ≤ z ≤ 0) = P(0 ≤ z ≤ 1.43) = 0.4236

c) P(0.14≤ z ≤ 1.23 ¿=P ( 0 ≤ z ≤ 1.23 )−¿ P (0≤ z ≤ 0.14 ¿=0.3907−0.0557
= 0.3350

Z SCORES FOR NON-STANDARD NORMAL DISTRIBUTION

To convert measurements from a non-standard normal distribution to the standard one, we convert from
x-values to z-scores using the formula.

Example 20: Given that μ=5∧δ=2, find the probability that the random variable will takeon x-values
in the interval [5, 67]

Solution:

15
Example 21: The life expectancy of a particular brand of tire is normally distributed with a mean of
20,000 miles and a standard deviation of 2,500 miles. Find the percentage of tire that is expected to wear
out before 16,000 miles.

Solution:

THE BERNOULLI DISTRIBUTION FUNCTION

It was discovered by James Bernoulli towards the end of the 17th century, it involves a random
experiment having two dichotomous outcomes e.g. head or tail, male or female, good or bad,
success or failure, true or false, yes or no and so on.

The probability of success = P, while the probability of failure = 1 – p or q. It uses only integral
or whole numbers and so it is a discrete probability mass function of Bernoulli random variable.

Probability Density Function of Bernoulli Distribution

This is given as;

We can also express it as:

16
Example 1:

Given a Bernoulli distribution with p = ¼, give an expression for the probability density
function.

Solution:

Mean = p

Variance = pq = p(1 – p)

Example 2:

Find the mean and variance of the distribution in example 1 above.

Solution:

Mean = p = ¼

Variance =

Coefficient of Variation

Generally, coefficient of variation is given as:

Example 3:

17
Find the coefficient of variance of the Bernoulli distribution with parameter 1/4.

Solution:

Coefficient of Variation

EXERCISES

1. Use the normal distribution for the given binomial experiment to find the indicated
probability.

18
2. Construct a probability distribution for the given binomial experiment.

3. From past experience, a sales representative has found that the probability of success
for a sales call is 0.4. What is the probability that the sales representative will have at
least 10 successful sales out of 30 sales calls?
4. Samples of 10 items are drawn from a manufacturing process in which the chance of
any one item being defective is 0.1. Using Poisson approximation to binomial, what
are the chances of having:
a) Exactly two defectives
b) No defectives
c) 10 defective items

19