Unit 6

Probability Distributions

Overview
In Unit 5, we assigned probabilities to events. This unit extends on the concept of
probability by discussing probability distributions. The social scientist seldom works
with events arising from an experiment. Instead, focus is placed on the random variable
that arises out of the experiment, or the outcomes and their probabilities.

In preparation for inferential statistics, we now introduce random variables, classify these,
construct the corresponding probability distributions, introduce expected value and
variance of these variables, and focus specifically on the poisson, binomial and normal
probability distributions.

Unit 6 Learning Objectives

After completing this unit, you should be able to:
• recognize and define random variables in a given situation;
• distinguish between discrete and continuous random variables;
• create cumulative probability distributions from a probability
distribution for a discrete random variable;
• select and utilize the binomial and poisson distribution appropriately;
• link the probability distribution function to the probability density
function for a continuous random variable;
• interpret and apply the properties of expected value and variance for
both discrete and continuous random variables;
• select and utilize the normal distribution appropriately.

This Unit is divided into 3 sessions :

Session 1: Understanding Random Variables and Probability Distributions.

Session 2: Binomial and Poisson Probability Distributions.
Session 3: The Normal Distribution.

129
 Note to students
This unit contains several activities and one practice assignment at the end of
the unit. You are to work on the activities on your own. If you have any questions
or concerns please post a message in the unit discussion forum so that your
E-tutor can provide assistance to you. The assignment is to be uploaded in the
practice assignment area.

130
Session 1
Understanding Random Variables and Probability
Distributions.

Introduction
In this session we explain the term ‘random variable’ and distinguish between discrete
and random variables before moving on to a discussion of probability distributions.

Objectives

After completing this session you should be able to:

• define the term ‘random variable’;

• distinguish between discrete and continuous random variables;
• create discete and continous probability distributions;
• apply the rules of expectation for random variables;
• apply the rules of variance for discrete random variables.

A Random Variable

We begin by considering the following experiment that forms part of a game during a
party among some friends. The rules of the game specify that you are required to toss a
‘fair’ coin twice. Each outcome of a ‘head’ entitles the player to a $1.00 prize while each
outcome of a ‘Tail’ results in the player paying out $1.00.
Recall the sample space from tossing a ‘fair’ coin twice is given by
S = { HH, HT, TH, TT}.
The rules of the game allow us to assign a net prize money to each outcome in the
sample space.

Table 6.1
Net Prize Money from the Experiment

Outcome Net Prize Money

HH $1.00 + $1.00 = $2.00

HT $1.00 - $1.00 = $0.00
TH -$1.00 + $1.00 = $0.00
TT -$1.00 - $1.00 = -$2.00

131
 Each time that the experiment is repeated, the player stands to get net prize money of
$2.00 or $0.00 or pay out $2.00. We do not, however, know the precise prize money in
advance as these moneys all depend on outcomes that are random.

The net prize money is seen to vary. Hence, it is said to be a variable .

We typically represent variables by X. Here X takes on the values of $2, $0 and -$2 in a
random manner. X is called a random variable. We can also represent X as a mapping that
assigns a real valued number to each of the outcomes of the sample space.
.
Mapping Diagram

Outcome Pay Off

HH $2

HT
$0
TH

TT -$2

Hence we can state the formal definition of a random variable.

DEFINITION 6.1

A random variable is any function, assignment or mapping that assigns exactly

one real valued numerical value to each element of a sample space.

The nature of the assignment gives the random variable its name. In our experiment
above, X is the net prize money from the game of tossing two ‘fair’ coins.

Random variables form the core of inferential statistics. There are two types of random
variables:
• discrete, and
• continuous.

132
Discrete Random Variables

DEFINITION 6.2

A discrete random variable is a random variable that takes on only countable

values, that is, consecutive values separated by gaps.

The random variable X, arising from our experiment of tossing a ‘fair’ coin twice, takes
on 3 real number dollar values, namely, 2, 0 and -2. These are indeed countable. Hence X
here is a discrete random variable.

Consider an experiment of selecting a sample of 10 Form I male students randomly from

the student body of a city secondary school and recording the age (in years) of each
student in the sample. The ages recorded (in years) were 11, 12, 12, 11, 13, 11, 12, 12, 12,
11. Is there a discrete random variable here?

If so, do you agree that the name of this random variable should be ‘Age of a Form I
male student from the city secondary school’?

Now that you have been presented with two examples of a random variable, you need
to create your own example of a discrete random variable.

ACTIVITY 6.1.

Choose an experiment and create a discrete random variable that arises from
that experiment.

Continuious Random Varibles

Now we turn to continuous random variables. The fundamental difference between

discrete random variables and continuous random variables is that the continuous
random variables can take on every possible value between any two given real numbers.
Suppose we survey the customers arriving at a bank for service from the bank’s tellers
and measured the interval of time between two consecutive arrivals. If the time between
the 15th and 16th arrivals was 45 seconds and the time between the 23rd and 24th
arrivals was 75 seconds, it is possible that the time between two consecutive arrivals
could equal any value between 45 seconds and 75 seconds. Characteristics like time,
speed, temperature, height, length, etc. give rise to continuous random variables.

133
 DEFINITION 6.3

A continuous random variable takes on any values that are not countable. In
fact, these variables can take on every value within an interval of the real line.

ACTIVITY 6.2

Choose an experiment and create a continuous random variable that arises

from that experiment.

Before we leave the discussion on random variables, you are required to complete the
next activity.

ACTIVITY 6.3

Indicate which of the following random variables are discrete and which are
continuous:

a. the number of new accounts opened at a bank during a certain month.

b. the time taken to run a marathon.

c. the price of a concert ticket (in dollars).

d. the number of rotten eggs in a randomly selected box of one dozen

eggs.
e. the runs scored in a cricket game.

f. the weight of a randomly selected courier package.

g. the age of a house.

h. the amount of coke in a 12-ounce can.

i. the winnings from the lottery.

Discrete Probability Distributions

Let us return to our experiment involving the coin where the sample space and the net
prize money for this experiment were presented in Table 6.1.

In Unit 5, we used the ‘classical’ & ‘relative frequency’ approaches to determine

probabilities for the outcomes in a sample space. Given that the coin used in the game

134
was fair, the classical approach allows us to assign a probability of 0.5 for a head and 0.5
for a tail.

In Unit 5, we also introduced the laws of probability. We can use the multiplication law
to determine the probabilities of the outcomes from two tosses of a fair coin as shown in
the table below.

Table 6.2
Probabilities of Outcomes in the Experiment of Tossing of a Fair Coin Twice

Outcome Probability
HH 0.25
HT 0.25
TH 0.25
TT 0.25

Merging Table 6.1 and Table 6.2 we get, Table 6.3.

Table 6.3
Probabilities of Outcomes and Related Random Variables
Outcome Net Prize Money Probability

HH $2 0.25
HT $0 0.25
TH $0 0.25
TT -$2 0.25

Looking again at Table 6.3, we observe that we have listed all the values of net prize
money and their corresponding probabilities. Further the sum of the probabilities is 1.
Remember that net prize money is the discrete random variable in this example. If we
were to create a table in which we present all values of the random variables and their
corresponding probabilities, we would refer to this table as a probability distribution of
the discrete random variable, net prize money.

Table 6.4
Probability Distribution of Net Prize Money

Net Prize Money Probability

$2 0.25
$0 0.50
-$2 0.25
Total 1.00

We can formally state the definition.

135
 DEFINITION 6.4

A probability distribution for a discrete random variable is any table, function

or formula that allows for the following to be presented or discerned:
i. all values taken on by the random variable;
ii. their corresponding probabilities; and
iii. the probabilities sum to one.

If you were asked to develop a probability distribution for a discrete random variable,
you should follow these steps:
1. Define the experiment.
2. Define the related sample space.
3. Perform a sufficiently large number of repetitions of the experiment.
4. Create the frequency table for the outcomes of the repetitions of the experiment.
5. Create the corresponding relative frequency table.
6. In accordance with the relative frequency approach to probability, we can
rename relative frequency as probability. Hence the relative frequency table in
step 5 above becomes the probability distribution.

Note that when the relative frequencies are known for the population, these are seen to
be the theoretical probabilities of the outcomes. Accordingly, the relative frequency table is
seen to be the theoretical probability distribution of the random variable.

ACTIVITY 6.4

Each of the three tables below lists certain values x of a discrete random
variable X. Show that Table 2 is the lone valid probability distribution.

Table 1 Table 2 Table 3

x P(x) x P(x) x P(x)
0 .08 1 .25 7 .70
1 .11 2 .34 8 .50
2 .39 3 .28 9 -.20
3 .27 4 .13

When you studied frequency tables/distributions in Unit 4, you followed through in

Unit 4 with pictorial representations of these tables/distributions. Now that we have
introduced the probability distribution in this unit, we can ask the question whether we
can pictorially represent a probability distribution for a discrete random variable. The
answer is yes. For example, we can proceed to show in Figure 6.1 a bar chart of the only
probability distribution among the three tables in Activity 6.4.

136
 Figure 6.1

P(x)

0.4

0.3
0.2 P(x)

0.1
0
1 2 3 4

We call such a bar chart the graph of the probability distribution.

Recall when you were introduced to frequency distributions in Unit 3, you proceeded
to create corresponding frequency distributions. Can we then create a cumulative
probability distribution from a probability distribution of a discrete random variable?
The answer is in the affirmative.
The discrete probability distribution can be used with the addition law to derive these
cumulative probabilities. As an example, the corresponding cumulative probability
distribution for Table 6.2 is given by:

Table 6.5
Cumulative Probability Distribution

x P(X ≤ x)
1 .25
2 .59
3 .87
4 1.00

How did we get P(X ≤ 2) = 0.59 above? We used the addition law of probability to say
that P(X ≤ 2) = P(X = 1 or 2) = P(X = 1) + P(X=2) = .25 + .34 = .59.
In a similar way, we can justify the other values in Table 6.5 above.
Observe that all the cumulative probabilities are of the type P(X ≤ x ). Alternative
cumulative probabilities are of the form P(X < x), P(X > x ) and P(X ≥ x ) and these give
rise to three other cumulative probability distributions.

ACTIVITY 6.5

Recall Table 6.4. Show that the cumulative probability distribution of Net
Prize Money is given by:
x P(X ≤ x)
-$2 0.25
$0 0.25+0.50 = 0.75
$2 0.75+0.25 = 1.00

137
 Expected Value of a Discrete Random Variable

Recall our game involving the tossing of two ‘fair’ coins.

Recall also the discrete random variable X representing the net prize money and its
probability distribution (as shown in Table 6.4).

Suppose that we played this game repeatedly a large number of times, say 4000 times.
From the probability distribution we can say that:

• Net prize money of -$2 is expected in 25% of the games i.e. in 1000 games
• Net prize money of $0 is expected in 50% of the games i.e. in 2000 games
• Net prize money of $2 is expected in 25% of the games i.e. in 1000 games.

What then would be our average net prize money after these 4000 games?

Average Net Prize Money = Total Net Prize Money

No. of games played

= 1000 (-$2) + 2000($0) + 1000($2)

4000

Simplifying we get

= 1 (-$2) + 2($0) + 1($2)

4 4 4

= $0

If therefore we played the game a large number of times, on average, we can expect to
win nothing (and, by extension, lose nothing). The value of $0 so computed is called the
expected value of the discrete random variable X. It is denoted by E(X). What does this
mean?

First, we can say what it does not mean. It certainly does not mean that a player is
guaranteed to win nothing or lose nothing in one game. Rather, some players will win,
some will break even and some will lose but, overall, the winnings of one player will
cancel out the losses of another as the number of games gets sufficiently large.

Look at the simplified calculation of E(X)

E(X) = 1 (-$2) + 2($0) + 1($2)

4 4 4

= $0

138
In that calculation, we multiplied each value of the random variable X by its
corresponding probability and summed the resulting products. We can rewrite this
calculation as follows:
3
E(X) = ∑ xi P(xi)
i=1

where x1 = -$2; P(x1) = 0.25 or 1/4; x2 = $0; P(x2) = 0.50 or 2/4; x3 = $2, P(x3) = 0.25 or 1/4.

We can generalise this formula as follows.

DEFINITION 6.5

Given a discrete random variable X which takes on values x1 , x2 , x3 , ……..

xn with probabilities P(x1), P(x2), P(x3), ………. P(xn) respectively, the expected
value E(X) of the random variable X is given by the formula:
n
E(X) = ∑ xi P(xi)
i=1

f
Since by the relative frequency approach to probability, P(xi) = Rel. Freq (xi) = i/N for
each of the values of I where N is the total frequency,

n n
E(X) = ∑ xi fi = ∑ xi fi
i = 1 N i = 1 N

which is the formula for finding the mean of a frequency distribution.

Hence we say:
• E(X) is equal to the mean of the discrete random variable.
• The mean of a discrete random variable is the value that is expected to occur per
repetition, on average, if an experiment is repeated a large number of times.
• In short, E(X) is called the long run average value of the random variable.

ACTIVITY 6.6

A random variable X assumes the values of –1, 0, 1 and 2 only. The probability
that X assumes the value x is given by
P(X = x) = ( x – 3)2
30
a. What kind of random variable is X?
b. Show that the information above constitutes a probability distribution.
c. Find the mean value of X.

139
 Properties of Expectation

Let X be any discrete random variable.

Property #1: If a is a constant then E(a) = a

Property #2: If a is a constant then E(aX) = a E(X)
Property #3: If b is a constant then
E(X + b) = E(X) + b
Property #4: If a and b are constants then
E(aX + b) = a E(X) + b
Property #5: If X and Y are two distinct random variables then
E(X + Y) = E(X) + E(Y)
Property #6: If g(X) and h(X) are two distinct functions defined on X then
E[g(X) + h(X)] = E[g(X)] + E[ h(X)]

Property #1 is saying that if a random variable for some reason takes on one value a,
then the expected value (otherwise called its mean) will undoubtedly be the value a.

How do you use the other properties? Let us look at these examples.

EXAMPLE 6.1

X is a discrete random variable representing the return on an investment of

$10,000 in a range of financial institutions. We can create another discrete
random variable Y to represent the return on an investment of $30,000 in the
same financial institutions over the same period. Clearly Y = 3X.
Property #2 is confirming that E(Y) = E(3X) = 3E(X). The average return from an
investment of $30,000 will be three times the average return on an investment
of $10,000.
Suppose that each financial institution decided to add a bonus of $10 to
the amount paid out to each investor of $10,000. The average return on the
investment of $10,000 will now be $10 more than the E(X).
In other words E(X + 10) = E(X) + 10.
Property #3 confirms this by saying that E(X + b) = E(X) + b.
Suppose instead that each financial institution decided to add a bonus of $25
to the amount paid out to each investor of $30,000. The average return on the
investment of $30,000 will now be $25 more than the E(Y).
In other words E(Y + 25) = E(Y) + 25. But Y = 3X.
Hence we write E(3X + 25) = E(3X) + 25.
Property #2 tells us that E(3X) = 3E(X).
Thus E(3X + 25) = E(3X) + 25 = 3E(X) + 25.
Property #4 confirms this by saying that E(aX + b) = a E(X) + b.

EXAMPLE 6.1 cont’d on next page

140
EXAMPLE 6.1 cont’d

Suppose that each investor decided to invest only $20,000. Further, each
investor decided to invest $10,000 in a range of local financial institutions and
invest the remaining $10,000 in a range of extra-regional financial institutions.
Let X be the discrete random variable representing the return on the local
investment. Also let Y be the discrete random variable representing the return
on the extra-regional investment. At the end of any given period, the investor’s
total return will be the sum of the local return and the extra-regional return.
Thus, for each investor the total return on the investment of $20,000 can be
represented by the random variable X + Y. The average total return at the end of
any given period will be the E(X + Y). This value will be the sum of the average
return from the local institutions and the average return from the extra-regional
institutions. Essentially, E(X + Y) = E(X) + E(Y) which is Property #5.

Suppose now that the local financial institution offered a bonus of $10 over and
above the return on the investment of $10,000. In addition, the extra-regional
institutions countered by offering a bonus of $13 over and above the return on
the investment of $10,000. Let X be the discrete random variable representing
the return on the local investment. Also let Y be the discrete random variable
representing the return on the extra-regional investment. At the end of any given
period the local investment will yield an average of E(X) + 10 while the extra-
regional investment will yield an average of E(Y) + 13.
The average total yield on the $20,000 investment will be the sum of E(X) + 10
and E(Y) + 25.
By Property #3, E(X) + 10 = E(X + 10) and E(Y) + 13 = E(Y + 13).
E(X) + 10 + E(Y) + 13 = E(X + 10) + E(Y + 13) = E( X + 10 + Y + 13).
X + 10 is a function of X and Y+ 13 is a function of Y.
If we generalise this we see what Property #6 is saying.

Variance of Discrete Random Variables.

The expected value of the discrete random variable, E(X) has been shown to be
equal to the mean of the random variable. Mean is a measure of central tendency.
It highlights where the probability distribution of X is centred.
There should also be an associated measure of dispersion for the variable since the
mean alone does not give an adequate description of the shape of the distribution.
Variance is that measure. It is a measure of how the values of the variable X are
spread out or dispersed from the mean.
Consider a discrete random variable X taking on values x1 , x2 , x3 , …… xn with
associated probabilities p1 , p2 , p3 , ……, pn

We can write mean μ = E(X).

Recall that the probability distribution is the limiting relative frequency table.

141
 For the relative frequency table

Variance = ∑ fi (xi - μ )2
N

i.e. f1 (x1 - μ )2 + f2 (x2 - μ )2 + …+ fn (xn - μ )2

N N N

But each fi/N approximates to probability Pi

Thus variance can be rewritten as

P1 (x1 - μ )2 + P2 (x2 - μ )2 + ……+ Pn (xn - μ )2

which simplifies to ∑ Pi (xi - μ )2.

DEFINITION 6.6

Given a discrete random variable X which takes on values x1 , x2 , x3 , ……..

xn with probabilities P(x1), P(x2), P(x3), ………. P(xn) respectively, the variance
Var(X) of the random variable X is given by the formula:
n
Var (X) = ∑ Pi (xi - μ )2
i = 1

where μ = E(X).

One short comming of variance is the unit of measure. If we are finding for
example the variance of the ages of 20 boys, the variance will be measured in
years2. This has no interpretation in thereal world. We get around this short
comming by using another measure known as standard deviation.

We define standard deviation as follows:

DEFINITION 6.7

The standard deviation of the discrete random variable X is the square root of
the variance of X.

Let us apply the concepts to the money game.

142
ACTIVITY 6.7

1. Find the variance and standard deviation of the net prize money variable
from Table 6.4

2. Find the variance of the random variable X whose probability distribution is

given by
x P(x)
1 0.25
2 0.34
3 0.28
4 0.13

Properties of Variance for Discrete Random Variables

Let X be any discrete random variable.

Property #1: If a is a constant then Var(a) = 0.

Property #2: If a is a constant then

Var(aX) = a2 Var(X).

Property #3: If b is a constant then

Var (X + b) = Var(X).

Property #4: If a and b are constants then

Var(aX + b) = a2 Var(X).

Property #5: If X and Y are two independent random variables then

Var( X + Y) = Var(X) + Var(Y).

Property #1 is saying that if a random variable for some reason takes on one
value a, then the mean will be the value a. and each observation of this variable
will have no dispersion (no deviation) from the mean. Hence variance, which
measures the average deviation from the mean, will equal zero.

How do you use the other properties? Let us look at these examples.

143
 EXAMPLE 6.2

X is a discrete random variable representing the return on an investment of

$10,000 in a range of financial institutions. We can create another discrete
random variable Y to represent the return on an investment of $30,000 in the
same financial institutions over the same period. Clearly Y = 3X. and E(Y) =
E(3X) = 3E(X). The deviation of each observation in the distribution from its mean
is given by Y – E(Y) which we rewrite as 3X – 3E(X). In the formula for finding
variance we must square this deviation to get 32 (X – E(X))2. This procedure
is followed for all observations thus giving rise to a variance that is 32 times the
variance of X. Hence Property #2 which states Var(aX) = a2 Var(X).
.

Suppose that each financial institution decided to add a bonus of $10 to

the amount paid out to each investor of $10,000. The average return on the
investment of $10,000, will be E(X + 10) = E(X) + 10. Each observation in this
distribution is of the form X + 10. Thus the deviation from the mean for each
observation will be X + 10 – (E(X) + 10) which equals X – E(X). In the formula
for finding variance we must square this deviation to get (X – E(X))2.
This procedure is followed for all observations thus giving rise to a variance
that is the same as the variance of X. Hence Property #3 which states that
Var(X + b) = Var(X).

Suppose instead that each financial institution decided to add a bonus of $25 to
the amount paid out to each investor of $30,000. Let X be the discrete random
variable representing the return on the investment of $10,000.The average
return on the investment of $30,000, E(3X + 25) = E(3X) + 25. Each observation
in this distribution is of the form 3X + 25. Thus the deviation from the mean for
each observation will be 3X + 25 – (3E(X) + 25) which equals 3X – 3E(X) or 3(X
– E(X)). In the formula for finding variance we must square this deviation to get
32 (X – E(X)))2.This procedure is followed for all observations thus giving rise to
a variance that is 32 times the variance of X. Hence Property #4 which states
Var(aX + b) = a2 Var(X).

Suppose that each investor decided to invest only $20,000. Further, each
investor decided to invest $10,000 in a range of local financial institutions and
invest the remaining $10,000 in a range of extra-regional financial institutions.
Let X be the discrete random variable representing the return on the local
investment; and Y represent the discrete random variable for the return on the
extra-regional investment. X and Y are independent of each other given the
differences in the structure of the two markets. For each investor the total yield
will be of the form X + Y. Further E(X + Y) = E(X) + E(Y). Each observation in this
distribution is of the form X + Y. Thus the deviation from the mean for each
observation will be X + Y – (E(X) + E(Y)) which equals (X – E(X)) + (Y – E(Y)). This
procedure is followed for all observations thus giving rise to a variance that is
the same as the sum of the variance of X and variance of Y. Hence Property
#5 which states that Var(X + Y) = Var(X) + Var(Y) when X and Y are mutually
independent.

144
Discrete Probability Distributions to be Covered in this Course

Two distributions will be covered later in this course, namely,

• the binomial distribution,

• the poisson distribution.

We shall also explore the approximation of the binomial distribution by the

poisson distribution.

Continuous Probability Distributions

We begin this section by making the observation that because of the uncountable
nature of the values taken on by continuous random variables, we cannot think of
finding probability by adding probabilities of simple events. Because the values
are uncountable, we will never stop adding. Accordingly, we must modify the
approach used for the discrete random variables.

For the discrete random variables, we defined a hierarchy of 6 steps to create

a probability distribution. For continuous random variables, the following
modification is necessary.

1. Define the experiment.

2. Define the sample space.
3. Generate a sufficiently large number of repetitions of the experiment.
4. Construct the frequency table of the outcomes of the repetitions.
5. Construct the relative frequency table from the frequency table in step 4.
6. Construct the relative frequency histogram from the relative frequency table.
7. Mark out the relative frequency polygon from the relative frequency polygon.

We call the graph formed by the outline of the relative frequency polygon the
probability density curve for the continuous random variable.

145
 Activity 6.8

You are shown a relative frequency histogram below for a variable defined as
the height of a basketball player measured in feet. Being a relative frequency
histogram the sum of the areas of the bars equals 100%.
You must shade the relative frequency polygon on the histogram below.
Remember that the area under the relative frequency polygon approximates the
area under the relative frequency histogram. Thus the area under the polygon is
also 100%.
Highlight the top outline of the polygon. You would agree that the outline makes
a curve. That curve is the probability density function for variable defined as the
height of a basketball player.

35

30
P
25
e
r 20
c
e 15
n
10
t
5

6.0 6.1 6.2 6.3 6.4 6.5

Height

35

30

D 25
e
n 20
s
i 15
t
y 10

6.0 6.1 6.2 6.3 6.4 6.5

Height

146
To recap, we confirmed in Activity 6.8 that

• The heights of the bars in the relative frequency histogram sum to 1. Hence
the area under the histogram is 1.
• The relative frequency polygon has an area that approximates the area of the
histogram i.e. it is equal to 1.
• The smoothed outline of the relative frequency polygon is the probability
density curve. Hence the area under the probability density curve also
equals 1.
• The function whose graph is the probability density curve is called the
probability density function of the continuous random variable.

We therefore encounter a new paradigm in which probability is seen as the area under a
probability density curve.

DEFINITION 6.8

The probability density curve is the graph consisting of the outline of the
relative frequency polygon. The total area under the curve equals 1.

Another definition is necessary.

DEFINITION 6.9

The probability density function is the function whose graph is the probability
density curve. It is usually represented by f(x). This function takes on only non-
negative values and the total area under the graph of this function equals 1.

In this paradigm, the probability that the continuous random variable lies between any
two given values a and b is given by the area under the probability distribution curve
bounded by the lines x = a and x = b.

Accordingly,
• all probabilities will lie in the range of 0 to 1 inclusive.
• the probability that the random variable will assume all possible intervals equals
the entire area under the curve i.e. an area of 1.

Hence the axioms of probability hold.

Further, the probability that the continuous random variable assumes a single value is
seen to be the area of a bar with zero width. i.e. such an area equals zero.

147
 Once we know the probability density function, we can refer to cumulative probabilities
such as
1. P(X < a) which is the area under the density curve from the far left to x = a;

2. P(X > a) which is the area under the density curve from x = a to the far right of the curve;

3. P(a < X < b) which is equal to the area under the density curve from x = a to x = b.

These probabilities are computed from a cumulative probability distribution which is

derived from what is called a probability distribution function for the continuous random
variable. This probability distribution function is represented by F(x).

In the probability distribution function

1. P(X < a) = F(a)
2. P(X > a) = 1 – F(a)
3. P(a < X < b) = F(b) – F(a).

148
DEFINITION 6.10

A probability distribution function of a continuous random variable is given

by F(x) and provides cumulative probabilities for that variable. In particular,
(i) F(x) = P(X < x);
(ii) P(a < X < b) = F(b) – F(a)

Expected Value of Continuous Random Variables

Given the fundamental difference in the characterization of probability for a continuous

variable i.e. it is area under the probability distribution curve, we must now find and
interpret expected value, E(X), for continuous variables.

In the discrete case we added the products of x and P(x). We cannot do this for
continuous random variables.

In the continuous case we find the centre of gravity (so to speak) of the area under the
probability distribution curve. Students of Mathematics can show that this centre of
gravity of the area under the curve is computed by integrating the product of x and f(x).

Properties of Expectation for Continuous RandomVariables

These are the same as for discrete random variables.

Property #1: If a is a constant then E(a) = a

Property #2: If a is a constant then E(aX) = a E(X)

Property #3: If b is a constant then E(X + b) = E(X) + b

Property #4: If a and b are constants then E(aX + b) = a E(X) + b
Property #5: If X and Y are two distinct random variables then
E(X + Y) = E(X) + E(Y)
Property #6: If g(X) and h(X) are two distinct functions defined on X then
E[g(X) + h(X)] = E[g(X)] + E[ h(X)]

Variance of Continuous Random Variables.

As in the case of the discrete random variables, variance is a measure of the dispersion
of the distribution of the variable.

149
 Properties of Variance for Discrete Random Variables

The properties for the discrete random variable also apply..

Property #1: If a is a constant then Var(a) = 0.
Property #2: If a is a constant then Var(aX) = a2 Var(X).
Property #3: If b is a constant then Var (X + b) = Var(X).
Property #4: If a and b are constants then Var(aX + b) = a2 Var(X).
Property #5: If X and Y are two independent random variables then
Var( X + Y) = Var(X) + Var(Y).

Continuous probability distributions to be covered in this Course are:

• normal distribution,

• student-t distribution,

• F distribution,

• chi square distribution.

We shall also explore the normal distribution later in this Unit. The other three
distributions will be presented in subsequent units.

Summary
This unit explored discrete and continous random variables and probability distribution
for a discrete random variable. We also explained the concept of expected value for
both discrete and continous variables, variance of discrete random variables and the
probability density function of the continous random variable.

150
Session 2
Binomial and Poisson Probability Distributions

Introduction

In the previous session we discussed random variables and focussed on discrete

and continous variables. In this session we turn our attention to two special discrete
distributions, namely the binomial and poisson distribution.

Objectives

On completing this session you should be able to:

• identify the conditions of a binomial experiment and a poisson process;

• define binomial and poisson random variables
• compute probabilities under the binomial and poisson probability distribution;
• compute and interpret the expected value of a binomial and poisson random
variable;
• compute and interpret the variance of a binomial and a poisson random variable.

Binomial Experiments

We begin our discussion on the binomial distribution by introducing the concept of a

binomial experiment.

DEFINITION 6.11

A Binomial Experiment is an experiment that satisfies the following four

conditions:

1. The experiment consists of a finite number n of repeated trials.

2. Each trial can result in one of only two possible outcomes; one outcome is
dubbed ‘success’, the other ‘failure’.

3. The probability of success is denoted by p and is constant from trial to trial.

4. The repeated trials are independent of each other.

The 2nd and 3rd conditions imply that the probability of failure q equals 1 – p.

151
 Examples of Binomial Experiments

• Quality control process in which a sample of a finished product is

removed from the assembly line and inspected or tested. The product is
recognised as either satisfactory to go to market or unsatisfactory to go to
market.
• The exposure of a community to a virus. Each resident in the community
will either succumb to the virus or not succumb to the virus.
• Playing the LOTTO some twenty times. Each time you play, you can
either win or lose.
• Doctors writing the exams set by the Royal College of Surgeons. Each
doctor either passes and obtains the FRCS qualification or fails and cannot
use the FRCS qualification.

ACTIVITY 6.9

Review the examples above and verify that the four conditions apply in each
case.

Binomial Random Variable

A random variable arises out of the binomial experiment. We define it in this way.

DEFINITION 6.12

The random variable X which represents the number of successes in the n

repeated trials of a binomial experiment is called the binomial random variable.

• X can therefore take on values of 0, 1, 2,…,n

• X is discrete.

• X has a discrete probability distribution.

In keeping with Definition 6.4, we can derive this distribution as a formula or function
that highlights the following:

• all the values assumed by X,

• their corresponding probabilities,
• the probabilities sum to 1.

152
The formula is developed out of our attempt to put meaning to the question
‘What does P(X = r) mean?’

P(X = r) means P(r successes in n trials) or equivalently P(r successes and n-r failures).

The event of ‘r successes and n-r failures’ can occur in several ways; each way is called a
combination of r successes and n-r failures.

Suppose, for example, that the first year students comprise exactly 1/3 of the student
population in the Faculty of Social Sciences at UWI and a small random sample of 5
students was selected from this Faculty. How do we find the probability that exactly
2 of the 5 students will be first year?

Here we have the following:

• a trial i.e. selecting one student from the Faculty and establishing whether this
student is first year or not;
• the trial is being repeated 5 times;
• each trial can result in one of only two outcomes; student is 1st year ‘Success’ or
student is not 1st year success ‘Failure’;
• the outcomes of the five trials are mutually independent;
• the probability of success p = 1/3 for all trials;
• the probability of failure q = 2/3 for all trials;
• hence, we have a binomial experiment.
• Ten combinations of 2 first year students and 3 non-first year students exist viz.
11000 10100
10010 10001
01100 01010
01001 00110
00101 00011

‘1’ means 1st yr student ; ‘0’ means non-first year student

• By the multiplication law, the probability of each combination = (1/3)2 (2/3)3

• By the addition law the probability of the 10 combinations = 10. (1/3)2 (2/3)3

• Hence P(2 first year students) = 10. (1/3)2 (2/3)3

5
• The number 10 is written C2 and is derived from the computation

5 5! 5x4x3x2x1
C2 = =
2! 3! (2x1)x(3x2x1)
5
• Thus the probability of 2 successes can be written as C2. (1/3)2 (2/3)3 = 80/243

153
 We can find the probabilities for 0, 1, 3, 4, and 5 successes respectively in much the same
manner.
5
• P(0 successes) = C0. (1/3)0 (2/3)5 = 32/243
5
• P(1 success) = C1. (1/3)1 (2/3)4 = 80/243
5
• P(3 successes) = C3. (1/3)3 (2/3)2 = 40/243
5
• P(4 successes) = C4. (1/3)4 (2/3)1 = 10/243
5
• P(5 successes) = C5. (1/3)5 (2/3)0 = 1/243

Hence, the sum of all six probabilities equals

32/ + 80/243 + 80/243 + 40/243 + 10/243 + 1/243 = 1
243

Do we have a probability distribution here? Yes. This distribution is called binomial

probability distribution.

More generally, we define the binomial probability distribution for X as follows:

DEFINITION 6.13

Let X be the random variable arising out of a binomial experiment. The

probability of r successes in n trials is given by the binomial probability
formula
n r n-r
P(r successes in n trials) = Cr (p) (q)
or
n r n-r
P(X = r) = Cr (p) (q)
where r = 0, 1, 2, ………..n.

We can compute any cumulative probability for X with the aid of the binomial
probability formula and the addition law. e.g.
• Probability of less than 2 successes = P(X < 2) = P(X = 0 or 1) = P(X = 0) + P(X = 1)
= 32/243 + 80/243 = 112/243

• Probability of more than 3 successes = P(X > 3) = P(X = 4 or 5) = P(X = 4) +

P(X = 5)
= 10/243 + 1/243 = 11/243

• Probability of between 2 and 4 successes P(2 ≤ X ≤ 4) = P(X = 2, 3 or 4)

= P(X = 2) + P(X = 3) + P(X = 4)
= 80 /243 + 40/243 + 10/243 = 130 /243

154
Note that n and p are called the parameters of the binomial distribution. Without values
for n and p we cannot use the binomial distribution formula to compute P(X = r).

Expectation of the Binomial Variable

E(X) = ∑ xi P(xi) = np

Variance of the Binomial Variable

Var(X) = ∑ Pi (xi - μ )2 = npq

EXAMPLE 6.3:
Return to the example involving first year students from the Faculty of Social
Sciences.

Recall that the binomial random variable X was defined as the number of
successes in the 5 students who were selected.

Here n = 5 p = 1/3 q = 2/3

E(X) = np = 5 x 1/3 = 5/
3 (What is your interpretation of this value?)

Var(X) = npq = 5 x 1/3 x 2/3 = 10/

9

Logic to be used in solving problems that involve a probability distribution

1. read the question

2. identify the random variable present
3. Check whether this random variable is discrete or continuous
4. If discrete then the narrative should point you to the probability distribution of
the random variable. In this course, it can be either Binomial or Poisson . Go to 6.
5. If the variable is continuous then the narrative should point you to the
probability distribution of the random variable. In this course, it can be either
Normal or Student-t . Go to 6.
6. State the relevant probability distribution formula
7. From the statement of the question, identify what probability you are being
asked to compute
8. translate the statement of the required probability from Standard English to the
Jargon of Probability
9. Decide whether the Addition Law or Multiplication Law of Probability is
required. If so, use it
10. Apply the probability distribution formula

155
 EXAMPLE 6.4
Return to the example involving first year students from the Faculty of Social
Sciences.

Recall that the binomial random variable X was defined as the number of
successes in the 5 students who were selected.

Here n = 5 p = 1/3 q = 2/3

Find the probability that between 1 and 5 students selected would be first year
students.

Let us now use this logic in solving Example 6.4

1. We read the question. It relates to a binomial experiment involving 5 trials; each
trial involves choosing a student from the Faculty at random and verifying whether the
student is first year or not. Success occurs when the student is first year.
2. The random variable X represents the number of successes in the 5 trials.
3. The random variable can take on values of 0,1,2,3,4,or 5 – this makes it discrete
4. The question prompts us by revealing that the random variable possesses a binomial
distribution. The parameters of this distribution are n = 5 and p = 1/3.
6. P(r successes) = P(X = r) = 5Cr (1/3.) r (1/3.) 5 - r
7. We are asked to find the probability that between 1 and 5 students selected would be
first year students – this is STANDARD ENGLISH.
8. We must translate this to the PROBABILITY JARGON to read
P( 1 to 5 successes inclusive) or simply P( X = 1 or 2 or 3 or 4 or 5)
9. The conjunction ‘or’ tells us that we must use the Addition Law. Remember that in
using the Addition Law, we must check for mutual exclusive events. X = 1, X = 2, X=3, X
= 4 and X= 5 are five events. Are these mutually exclusive? Yes because if X =1, it cannot
be equal to 2 at the same time or 3 or 4 or 5 at the same time. This means that we will use
the short form of the Addition Law.
P( X = 1or 2 or 3 or 4 or 5) = P(X=1) + P(X= 2) + P(X=3) + P(X = 4) + P(X= 5)
10. Using the distribution formula from Step 6 we get
P(X = 1) = 5C1 (1/3.) 1 (1/3.) 4 = 80/243
P(X = 2) = 5C2 (1/3.) 2 (1/3.) 3 = 80/243
P(X = 3) = 5C3 (1/3.) 3 (1/3.) 2 = 40/243
P(X = 4) = 5C4 (1/3.) 4 (1/3.) 1 = 10/243
P(X = 5) = 5C5 (1/3.) 5 (1/3.) 0 = 1/243
Thus P( X = 1 or 2 or 3 or 4 or 5) = 211/243

156
ACTIVITY 6.10
According to a recently conducted survey, 90% of all male drivers in Trinidad
and Tobago claim to consistently drive over the legal speed limit.
Assume that this result is true for the current male driving population of
Trinidad and Tobago. Find the probability that, in a random sample of 20
male drivers:
(i) none of them drive over the legal speed limit
(ii) at least 2 of them drive over the legal speed limit
(iii) 19 of them DO NOT drive over the legal speed limit

The Poisson Probability Distribution

The Poisson probability distribution arises out of a poisson process.

DEFINITION 6.14

A Poisson process is characterized by the following three conditions:

1. The presence of a discrete random variable (which we shall call X).
2. The occurrences of each value x of the variable X must be random (i.e.
they must be entirely unpredictable).
3. The occurrences must be entirely independent of each other [ the
occurrence (or non-occurrence) of an event must not influence the future
occurrences (or non-occurrences) of another event].

The Poisson process fits events that are randomly scattered over time and/or space (i.e.
you cannot predict when or where an event will occur). e.g.

• power outages,
• equipment failures,
• accidents,
• hurricanes.

DEFINITION 6.15

The Poisson random variable is defined as the number of occurrences of an

event in a given interval of time and/or space.

Such variables can take on values between o and infinity. These are clearly discrete.

The probability of a given number of occurrences is dependent on only one value i.e.
the average number of occurrences in the stated interval of time or space. In fact, in a

157
 Poisson process the only available data is the average number of occurrences in the
stated interval of time or space. This average value is denoted by and is called the
parameter of the poisson distribution.

In keeping with Definition 6.4, we can derive this distribution as a formula or function
that highlights the following:
• all the values assumed by X,
• their corresponding probabilities,
• the probabilities sum to 1.

The formula is developed out of our attempt to put meaning to the question
‘What does P(X = r) mean?’

P(X = r) means P(r occurrences in the stated unit of time or space).

We define the Poisson probability distribution formula as follows:

DEFINITION 6.16

Let X be the discrete random variable arising out of a poisson process with
parameter , the poisson probability formula is given by:
P( X = r) = e – λλr r = 0, 1, 2,……
r!
where e is the value 2.718………..

The formula reinforces the statement that once we know the value of we can compute
any probability under the Poisson Distribution.

Expectation of the Poisson Random Variable

E(X) =

Variance of the Poisson Random Variable

Var(X) =

Example: 6.5

Suppose that the number of accidents occurring at the Barataria Roundabout

in Trinidad in a month averages 3. Find the probability that there will be 5
accidents in March 2002.
let X = no. of accidents at Barataria Roundabout in one month. X possesses a
Poisson distribution with = 3.

P(X = 3) = e– 335 = .101

5!

158
Cumulative probabilities can be computed using the poisson probability distribution
and the addition law.

Example: 6.6

Find the probability that there will be at most 3 accidents at the Roundabout in
March 2002.

P(X ≤ 3) = P( X = 0 or 1 or 2 or 3)

= P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3)

= e –330 + e –331 + e –332 + e –333

0! 1! 2! 3!
= .0498 + .1494 + .2241 + .2241
= .6474

ACTIVITY 6.11
Calls reaching a telephone exchange over the past five years average 10
per minute.
Find the probability of that
a) no calls are recieved during a one minute period.
b) between 5 and 9 calls inclusive are recieved during a one minute
period.

Uses of the Poisson Distribution:

The poisson distribution is widely used as

• a distribution for random events (as we demonstrated in the two examples

above)
• an approximation to the binomial distribution.

The Approximation of the Binomial Distribution by the Poisson

Distribution

If the binomial distribution can be approximated to the Poisson distribution, it follows

that the means of the two distributions should be very close and so too should be the
two variances.

159
 Recall that

Variable Mean Variance

Binomial np npq
Poisson

Notice the relationship ‘mean = variance’ for the poisson distribution.

For such a relationship to apply to the Binomial it follows that np = npq; this is possible
when q is approximately 1 (which implies that p is approximately 0).

When q is approximately equal to 1 and we set equal to np, we can be reasonably

accurate in using the Poisson distribution as an approximation. The Poisson
distribution yields a good approximation to theBinomial distribution when n is large and
p is relatively small (i.E. P close to zero).

To execute the approximation we simply set the parameter = np and substitute into the
Poisson distribution formula.

ACTIVITY 6.12

1. The table below shows the comparative values of P(X = 50) when p = 0.02
for three values of n viz. 50, 100 and 200
Binomial Poisson Error % Error
Probability Approximation
n = 50 0.0027 0.0031 .0004 15
n = 100 0.0353 0.0361 .0008 2
n = 200 0.1579 0.1563 .0016 1

What can you conclude from the table?

2. Repeat these approximations for p =0.01 and construct a similar table.
What more can you conclude about these approximations?

Summary
Binomial and Poisson probability distributions were the focus of this session. We
defined the Binomial experiment and the Binomial probability distributions using the
formula:
n r n-r
P(r successes in n trials) = Cr (p) (q)

We went on to define the conditions for the poisson process, the poisson probability
formula and the uses of the Poisson distribution.

160
Session 3
The Normal Distribution

Introduction
The normal distribution is first and foremost a distribution for continuous random
variables. It is the most widely used continuous distribution and hence it is of great
importance in inferential statistics. In fact, a large number of phenomena in the real
world are either exactly or approximately normally distributed. Variables such as
heights, weights, time intervals, weights of packages, productive life of bulbs etc. all
possess a normal distribution. This session focuses on computation, interpretation and
application of the normal distribution.

Objectives:
At the end of this session, students must be able to:

• identify random variables that are normally distributed;

• identify the parameters of normal distributions;
• use the table of the standard normal distribution;
• compute probabilities under the normal distribution;
• compute and interpret the expected value of a normal random variable;
• compute and interpret the variance of a normal random variable;
• apply the normal distribution as an approximation to the binomial distribution;
• identify when a binomial probability can be approximated by using the normal
distribution.
The Normal Curve
The probability density curve for the normal
distribution is called the normal curve.

• It is bell shaped and symmetric.

• It is centred at the mean value μ.
• Its tails extend indefinitely, that is from
- on the left to + on the right without
touching or crossing the horizontal axis.
• The height of the top of the bell is linked to the standard deviation of the normal
random variable.
• The higher the top, the lower the standard deviation and vice versa.
• The total area under the normal curve equals 1.
• The symmetry about the mean value points to the area under the curve to the
left of the mean equals 0.5 or 50%.
• Similarly, the area under the curve to the right of the mean is also 0.5 or 50%.

161
 Let X be the normal random variable. Clearly X is continuous.

All probabilities in the normal distribution are given as areas under the probability
density curve since that is the paradigm used for probability of continuous random
variables. Typical areas under the normal curve are as follows: .
• P(X is less than μ) = .50;

• P( X lies within 1 standard deviation of μ) ≈ .6845;

• P( X lies within 2 standard deviations of μ) ≈ .95;

• P( X lies within 3 standard deviations of μ) ≈ .9975;

162
 • P(X is greater than μ) = .50;

For a normal random variable X, the expected value E(X) = μ and the variance Var(X) =
2 ; hence the Standard Deviation of X = .

Since 99.75% of the area is trapped within 3 standard deviations of the mean , we say
that the Range of X is approximately 6 .

Each combination of and gives rise to a unique normal curve referred to as N(μ , ).
Hence and are called the parameters of the normal distribution; in fact, no probability
can be computed under the normal distribution without values for and .

Because the combinations of and can take on infinitely many values, we recognise
that there exists an infinitely large family of normal curves. The areas under each normal
curve N( ) can be presented in a cumulative probability table. Does this suggest that
we need to access a book containing infinitely many cumulative normal probability
tables? No. The reality is that we need only one probability table, i.e. the Table of the
Standard Normal Distribution. What is this distribution?

DEFINITION 6.17

The normal distribution with = 0 and = 1 is called the Standard Normal

Distribution.

DEFINITION 6.18

The random variable that possesses the standard normal distribution is called
the Standard Normal Variable and it is denoted by Z.

E(Z) = 0 and Var(Z) = 1 ; hence the standard deviation of Z = 1.

See appendix to this unit for your copy of the table

163

• The values of Z are located on the horizontal axis of the standard normal
curve, and are called Z scores otherwise called standard scores.

• Each Z score gives the distance on the z axis between the mean and the point
denoted by z in terms of standard deviations (the unit of measure of z is
therefore standard deviations)

e.g.Let X be the random variable with a normal distribution in which = 10

and = 5. When X = 15, Z = 1; X = 20, Z = 2; X = 5, Z = -1

Typical areas under the standard normal curve are as follows:

• P(Z is less than 0) = .50
• P( -1 < Z <1) .6845
• P(-2< Z < 2) .95
• P(-3 < Z < 3) .9975
• P(Z is greater than 0) = .50

See related sketches on the previous page

Look at your copy of the Table of the Standard Normal Distribution and
double-check these five areas.

Using the Table of the Standard Normal Distribution to find a

Probability under the Normal Distribution N( , )?

e.g. P(a <X < b)

1. Sketch the curve of N( , ) and highlight the area which relates to
P(a <X < b).

2. Transform the end points X = a and X = b to Z scores using the

formula
Z = X-

164
 3. Let the corresponding Z scores be z1 and z2.

4. Sketch the equivalent area under the Standard Normal Curve

bounded by z1 and z2.

5. Read off the area from the table of the standard normal curve.

ACTIVITY 6.13

Use the Table of the Standard Normal Distribution to find the following
probabilities:
P(Z < 1.9) P(Z > 2.1)
P(1.9 < Z < 2.1) P(Z > -1.9)
P(-1.9 < Z < 1.9) P(Z < -2.1)
P(0 < Z < 0.44)

In the interest of much needed practice you should also attempt the next two activities.

ACTIVITY 6.14

Use the Table of the Standard Normal Distribution to find the following
probabilities:
• P( Z < 1.48)
• P( Z < -0.93)
• P( Z > 0.50)
• P( Z > -1.66)
• P( -2.15 < Z < 1.96)
• P(0.51 < Z < 1.12)
• P( -1.35 < Z < -0.64)

165
 In Activities 6.12 and 6.13, you are required to work from the periphery of the standard
normal table to the interior of the table. You must also be able to work from the interior
to the periphery. This is the objective of the next activity.

ACTIVITY 6.15
Find z for each of the following:

• P( -z < Z < z) = 0.7814

• P( Z < z) = .80

Using the Normal Distribution to Approximate the Binomial

Distribution

The binomial distribution formula requires some ‘clumsy’ computations of the form
n
Cr and becomes very tedious when n is large.

We explored earlier in this unit the Poisson approximation to the Binomial; this
approximation could also be very tedious when n is large.

When n is large, we can explore using a Normal approximation.

Recall
Variable Type Mean Variance

Binomial Discrete np npq

Normal Continuous μ 2

Standard Normal Continuous 0 1

We are seeking to approximate a discrete probability distribution by a continuous one.

What then (we might ask) links the Binomial variable with the Standard Normal
variable? The answer is the Central Limit Theorem (CLT) which we describe as follows:

Let X be a discrete random variable. Let X ~ B( n , p ). Assuming that n is large and we

executed the following:

i. define for each value x of the variable X the quotient

z = x – np
√npq

ii. construct a frequency distribution of the set of values of z and

iii. plot a histogram from the frequency distribution.

166
We will find that the outline of the histogram will approximate the curve of the Standard
Normal distribution i.e. symmetric and bell-shaped, mean = 0, variance = 1.

Hence the Standard Normal Z ≈ X – np

√npq

This is the guarantee provided by the CLT and forms the basis of the approximation.

Mindful of the fact that we are using a distribution for a continuous variable to
approximate the distribution of a discrete variable, we must ‘correct’ the discrete
variable so that it is seen to be continuous; this is done via a Continuity Correction.

DEFINITION 6.19

The subtraction of 0.5 and/or addition of 0.5 from the end points X = r1 and X = r2
respectively allows us to rewrite
P(r1 < X < r2)
i.e. the probability of between a and x2 successes
as
P(r1 - 0.5 < X < r2 + 0.5);
i.e. the probability that the ‘Continuous’ Variable lies between r1 – 0.5 and r2 + 0.5

This is the continuity correction.

When do we get a high degree of accuracy from the normal approximation?

The normal distribution is best used as an approximation to the binomial distribution

when np and nq are both greater than 5 i.e. np > 5 and nq > 5

Summary of the Methodology

Let X be a binomial variable with parameters n and p.

In order to estimate P(r1 ≤ X ≤ r2) by a normal approximation we must:

• Perform the continuity correction i.e.

P(r1 ≤ X ≤ r2) = P(r1 - 0.5 < X < r2 + 0.5)
• Set up the transformation
Z = X – np
√npq
• Transform the left end point r1 – 0.5 to z1
• Transform the right end point r2 + 0.5 to z2

167
 • Sketch a curve of the standard normal distribution and shade the area
that corresponds to P( z1 < Z < z2)
• Read off the area from the standard normal table.

ACTIVITY 6.16

Suppose that 75% of the students in an educational institution are known to be

female. When a sample of 100 students is drawn from the school population,
what is the probability that there will be more than 20 male students in that
sample?

PRACTICE ASSIGNMENT #1

According to a recently conducted survey, 90% of all drivers in Trinidad and Tobago
claim to consistently drive over the legal speed limit. Assume that this result is true for
the current population of a Caribbean country. Find the probability that, in a random
sample of 20 drivers:

(i) none of them drive over the legal speed limit

(ii) at least 2 of them drive over the legal speed limit
(iii) 19 of them DO NOT drive over the legal speed limit

PRACTICE ASSIGNMENT #2

An average of 0.8 accidents occur every day on the roads of Trinidad and Tobago. Find
the probability that

(i) no accident will occur in Trinidad and Tobago on a given day

(ii) at least 2 accidents will occur in Trinidad and Tobago on a given day

PRACTICE ASSIGNMENT #3

In a recent poll, 81% of parents with children under 18 years of age said that it is more
difficult to raise children now than it was when their parents raised them. Assume that
this percentage is true for the current population of all parents with children under 18
years of age. Find the probability that, in a random sample of 1000 such parents, less
than 800 will hold the above view

168
PRACTICE ASSIGNMENT #4

Let X and Y denote two random variables that are normally and independently
distributed with means of 75 and 40 respectively, and standard deviations of 5 and 4
respectively. Define a new random variable U = 3X-2Y.

i. Will U be normally distributed? If yes, give a reason for your answer

ii. Find E(U)
iii. Find var(U)

PRACTICE ASSIGNMENT #5

Let x be the number of cards that a randomly selected auto mechanic repairs on a given
day. The following table lists the probability distribution of x.

X 2 3 4 5 6
P(X=x) 0.05 0.22 * 0.23 0.10

(i) Calculate the missing value *

(ii) Find the mean and standard deviation of x.
(iii) Calculate P(X ≤ 5)

169
 Summary
This final session examined the random variables that are normally distributed,
parameters of the normal distribution and the use of the table of the standard normal
distribution to find probabilities. We also discussed use of the normal distribution to
approximate the binomial distribution.

Wrap Up
In this unit, we have defined random variables and provided a basis for classifying
them as discrete or continuous. We have constructed both the probability distribution
and the cumulative probability distribution for a discrete random variable; defined and
interpreted expected value and variance of a discrete random variable. We also defined
binomial experiments and poisson processes from which the binomial distribution and
the poisson distribution respectively arise. The probability distribution formulae for
both these distributions were presented.
We introduced the paradigm by which we conceptualise probability for continuous
random variables; defined probability density function and probability distribution
function to the probability density function for a continuous random variable; extended
the concepts of expected value and variance to continuous random variables.
We also interpreted the properties of expected value and variance for both discrete and
continuous random variables.
Special focus was given to the normal distribution, the poisson approximation to the
binomial distribution and the normal approximation to the binomial distribution.
This unit has introduced several fundamental issues that will be used in Units 7 to 9.

170