Dr.

Modar Shbat
Division of Engineering
[email protected]
2
Variables were defined as characteristics that change or vary over time and/or for
different individuals or objects under consideration. Quantitative variables generate
numerical data, whereas qualitative variables generate categorical data. However, even
qualitative variables can generate numerical data if the categories are numerically coded
to form a scale. For example, if you toss a single coin, the qualitative outcome could be
recorded as “0” if a head and “1” if a tail.
Random Variables
A numerically valued variable x will vary or change depending on the particular outcome
of the experiment being measured. For example, suppose you toss a die and measure
x, the number observed on the upper face. The variable x can take on any of six
values—1, 2, 3, 4, 5, 6—depending on the random outcome of the experiment. For this
reason, we refer to the variable x as a random variable.
Definition: A variable x is a random variable if the value that it assumes,
corresponding to the outcome of an experiment, is a chance or random event.

Prob. & Stat. 3

As we discussed before, quantitative random variables are classified as either
discrete or continuous, according to the values that x can assume. It is important to
distinguish between discrete and continuous random variables because different
techniques are used to describe their distributions. We focus on discrete random
variables in the remainder of this section; and we discuss continuous random variables
later.

Prob. & Stat. 4

In the aforementioned lectures, we learned how to construct the relative frequency
distribution for a set of numerical measurements on a variable x. The distribution gave
this information about x:

We also learned how to use the mean and standard deviation to measure the center and
variability of this data set.
In this section, we defined probability as the limiting value of the relative frequency as
the experiment is repeated over and over again. Now we define the probability
distribution for a random variable x as the relative frequency distribution constructed for
the entire population of measurements.

The values of x represent mutually exclusive numerical events. Summing p(x) over all
values of x is equivalent to adding the probabilities of all simple events and therefore
equals 1.

Prob. & Stat. 5

Example: Toss two fair coins and let x equal the number of heads
observed. Find the probability distribution for x.
Solution:
The simple events for this experiment with their respective probabilities are listed in
related Table. Since E1 =HH results in two heads, this simple event results in the value
x=2. Similarly, the value x=1 is assigned to E2, and so on.
For each value of x, you can calculate p(x) by adding the
probabilities of the simple events in that event. For
example, when x=0,

Prob. & Stat. 6

Example (Cont.):
The values of x and their respective probabilities, p(x), are listed in the following table.
Notice that the probabilities add to 1.
The probability distribution can be graphed to form the
probability histogram.† The three values of the random variable
x are located on the horizontal axis, and the probabilities p(x) are
located on the vertical axis (replacing the relative frequencies).
Since the width of each bar is 1, the area under the bar is the
probability of observing the particular value of x and the total area
equals 1.

Prob. & Stat. 7

The probability distribution for a discrete random variable looks very similar to the
relative frequency distribution. The difference is that the relative frequency distribution
describes a sample of n measurements, whereas the probability distribution is
constructed as a model for the entire population of measurements.
Just as the mean and the standard deviation measured the center and spread of the
sample data, you can calculate similar measures to describe the center and spread of
the population.
The population mean, which measures the average value of x in the population, is also
called the expected value of the random variable x. It is the value that you would expect
to observe on average if the experiment is repeated over and over again.
The formula for calculating the population mean is easier to understand by example.
Example: Toss those two fair coins again, and let x be the number of heads observed.
We constructed this probability distribution for x:
Suppose the experiment is repeated a large number of
times—say, n=4,000,000 times. Intuitively, you would expect
to observe approximately 1 million zeros, 2 million ones, and
1 million twos. Then the average value of x would equal:

Prob. & Stat. 8

We could use a similar argument to justify the formulas for the population variance 𝝈𝟐
and the population standard deviation 𝝈.

Prob. & Stat. 9

These numerical measures describe the spread or variability of the random variable
using the “average” or “expected value” of the squared deviations of the x-values from
their mean μ.

Definition: The standard deviation 𝝈 of a random variable x is equal to the

positive square root of its variance.
Example:
An electronics store sells a particular model of computer notebook. There are only four notebooks
in stock, and the manager wonders what today’s demand for this particular model will be. She
learns from the marketing department that the probability distribution for x, the daily demand for the
laptop, is as shown in the table. Find the mean, variance, and standard deviation of x. Is it likely
that five or more customers will want to buy a laptop today?

Prob. & Stat. 10

Example, Solution:

Prob. & Stat. 11

Example (Cont.):
The graph of the probability distribution is shown in the figure. Since the distribution is
approximately mound-shaped, approximately 95% of all measurements should lie within
two standard deviations of the mean, that is:

Since x=5 lies outside this interval, you

can say it is unlikely that five or more
customers will want to buy a laptop today.
In fact, P(x≥5) is exactly .05, or 1 time in
20.

Prob. & Stat. 12

Example:
In a lottery conducted to benefit a local charity, 8000 tickets are to be sold at $10 each.
The prize is a $24,000 automobile. If you purchase two tickets, what is your expected
gain?
Solution:
Your gain x may take one of two values. You will either lose $20 (your “gain” will be -$20)
or win $23,980. The probability distribution for the gain x is shown in the table:

Recall that the expected value of x is the average of the theoretical population that
would result if the lottery were repeated an infinitely large number of times. If this were
done, your average or expected gain would be a loss of $14.

Prob. & Stat. 13

Example:
Determine the yearly premium for a $10,000 insurance policy covering an event that,
over a long period of time, has occurred at the rate of 2 times in 100. Let x equal the
yearly financial gain to the insurance company resulting from the sale of the policy, and
let C equal the unknown yearly premium.
Calculate the value of C such that the expected gain E(x) will equal zero. Then C is the
premium required to break even. To this, the company would add administrative costs
and profit.
Solution:
The first step in the solution is to determine the values that the gain x may take and then
to determine p(x). If the event does not occur during the year, the company will gain the
premium of x=C dollars. If the event does occur, the gain will be negative; that is, the
company will lose $10,000 less the premium of C already collected. Then x=-(10,000-C).
The probabilities of these two values of x are 98/100 and 2/100, respectively.

Prob. & Stat. 14

Example (Cont.):
Solving this equation for C, you obtain:

Therefore, if the insurance company charged a yearly premium of $200, the average
gain calculated for a large number of similar policies would equal zero. The actual
premium would equal $200 plus administrative costs and profit.

Note: The method for calculating the expected value of x for a continuous random
variable is similar to what we have done, but in practice it involves the use of calculus.
Nevertheless, the basic results concerning expectations are the same for continuous
and discrete random variables.
Regardless of whether x is continuous or discrete:

Prob. & Stat. 15

Discrete random variables are used in many practical applications. Three important
discrete random variables—the binomial, the Poisson, and the hypergeometric—are
presented in this section (three discrete probability distributions that serve as models for
a large number of applications). These random variables are often used to describe the
number of occurrences of a specified event in a fixed number of trials or a fixed unit of
time or space.
BINOMIAL PROBABILITY DISTRIBUTION
A coin-tossing experiment is a simple example of an important discrete random variable
called the binomial random variable. Many practical experiments result in data similar
to the head or tail outcomes of the coin toss.

Prob. & Stat. 16

BINOMIAL PROBABILITY DISTRIBUTION

Example: Suppose there are approximately 1,000,000 adults in a county and an

unknown proportion p favor term limits for politicians. A sample of 1000 adults will be
chosen in such a way that every one of the 1,000,000 adults has an equal chance of
being selected, and each adult is asked whether he or she favors term limits. (The
ultimate objective of this survey is to estimate the unknown proportion p). Is this a
binomial experiment?
Prob. & Stat. 17
Example (Cont.):
1. A “trial” is the choice of a single adult from the 1,000,000 adults in the county. This
sample consists of n=1000 identical trials.
2. Since each adult will either favor or not favor term limits, there are two outcomes that
represent the “successes” and “failures” in the binomial experiment.
3. The probability of success, p, is the probability that an adult favors term limits. Does
this probability remain the same for each adult in the sample? For all practical purposes,
the answer is yes.
4. The independence of the trials is guaranteed because of the large group of adults
from which the sample is chosen.
5. The random variable x is the number of adults in the sample who favor term limits.
Because the survey satisfies the five characteristics ⇒ binomial experiment.

Before, we tossed two fair coins and constructed the probability distribution for x, the
number of heads—a binomial experiment with n=2 and p=.5. The general binomial
probability distribution is constructed in the same way.
Prob. & Stat. 18
BINOMIAL PROBABILITY DISTRIBUTION
Fortunately, the probabilities p(x) follow a general pattern. This allows us to use a single
formula to find p(x) for any given value of x.

Prob. & Stat. 19

BINOMIAL PROBABILITY DISTRIBUTION
Example: Find P(x=2) for a binomial random variable with n=10 and p=0.1
Solution: P(x=2) is the probability of observing 2 successes and 8 failures in a sequence
of 10 trials. You might observe the 2 successes first, followed by 8 consecutive failures:

However, many other sequences also result in x=2 successes. The binomial formula
uses to count the number of sequences and gives the exact probability when you
use the binomial formula with k=2:

We could repeat the procedure in this example for each value of x—0, 1, 2, . . . ,10—
and find all the values of p(x) necessary to construct a probability histogram for x.
The following figures show the results for different values of p.

Prob. & Stat. 20

BINOMIAL PROBABILITY DISTRIBUTION
Example (Cont.):

The distribution is exactly symmetric about

the mean.

The graph is skewed right; that is, most

of the time we will observe small values
of x.

The distribution is the “mirror image” of the

distribution for p=.1 and is skewed to the left.

Prob. & Stat. 21

Example: Over a long period of time, it has been observed that a basketball player can
make a free throw on a given trial with probability equal to 0.8. Suppose he shoots four
free throws: 1. What is the probability that he will make exactly two free throws?
2. What is the probability that he will make at least one free throw?
Solution: A “trial” is a single free throw, and you can define a “success” as a basket and
a “failure” as a miss, so that n=4 and p=.8. If you assume that the player’s chance of
making the free throw does not change from shot to shot, then the number x of times
that he makes the free throw is a binomial random variable.

Although we could calculate P(x=1), P(x=2), P(x=3) and P(x=4) to find this probability,
using the complement of the event makes your job easier; that is,

Prob. & Stat. 22

THE POISSON PROBABILITY DISTRIBUTION
Another discrete random variable that has numerous practical applications is the
Poisson random variable. Its probability distribution provides a good model for data
that represent the number of occurrences of a specified event in a given unit of time or
space. Here are some examples of experiments for which the random variable x can be
modeled by the Poisson random variable:

In each example, x represents the number of events that occur in a period of time or
space during which an average of m such events can be expected to occur.
The only assumptions needed when one uses the Poisson distribution to model
experiments such as these are that the counts or events occur randomly and
independently of one another.

Prob. & Stat. 23

For each value of k, you can obtain the individual probabilities for the Poisson random
variable, just as you did for the binomial random variable.

Prob. & Stat. 24

THE POISSON PROBABILITY DISTRIBUTION
Example: The average number of traffic accidents on a certain section of highway is two
per week. Assume that the number of accidents follows a Poisson distribution with μ=2.
1. Find the probability of no accidents on this section of highway during a week period.
2. Find the probability of at most three accidents on this section of highway during 2-
week period.
Solution:
1. The average number of accidents per week is μ=2. Therefore, the probability of no
accidents on this section of highway during a given week is:

2. During a 2-week period, the average number of accidents on this section of highway
is 2(2)=4. The probability of at most three accidents during a week period is:

Prob. & Stat. 25

THE HYPERGEOMETRIC PROBABILITY DISTRIBUTION
If the number of elements in the population is large relative to the number in the sample,
the probability of selecting a success on a single trial is equal to the proportion p of
successes in the population. Because the population is large in relation to the sample
size, this probability will remain constant (for all practical purposes) from trial to trial, and
the number x of successes in the sample will follow a binomial probability distribution.
However, if the number of elements in the population is small in relation to the sample
size (n/N≥.05), the probability of a success for a given trial is dependent on the
outcomes of preceding trials. Then the number x of successes follows what is known as
a hypergeometric probability distribution.
It is easy to visualize the hypergeometric random variable x by thinking of a bowl
containing M red balls and N-M white balls, for a total of N balls in the bowl. You select n
balls from the bowl and record x, the number of red balls that you see. If we now define
a “success” to be a red ball, we have an example of the hypergeometric variable x.

26
THE HYPERGEOMETRIC PROBABILITY DISTRIBUTION

Example:
A case of wine has 12 bottles, 3 of which contain spoiled wine. A sample of 4 bottles is
randomly selected from the case.
1. Find the probability distribution for x, the number of bottles of spoiled wine in the
sample.
2. What are the mean and variance of x?
Solution:

27
THE HYPERGEOMETRIC PROBABILITY DISTRIBUTION
Example (Cont.):

28
THE HYPERGEOMETRIC PROBABILITY DISTRIBUTION
Example:
A particular industrial product is shipped in lots of 20. Testing to determine whether an
item is defective is costly; hence, the manufacturer samples production rather than using
a 100% inspection plan. A sampling plan constructed to minimize the number of
defectives shipped to customers calls for sampling five items from each lot and rejecting
the lot if more than one defective is observed. (If the lot is rejected, each item in the lot is
then tested.) If a lot contains four defectives, what is the probability that it will be
accepted?
Solution:
Let x be the number of defectives in the sample. Then N=20, M=4, (N-M)=16, and n=5.
The lot will be rejected if x=2, 3, or 4. Then:

29