# © 2007 Pearson Education

Chapter 3: Random Variables
and Probability Distributions
Part 1: Probability and Probability
Distributions
Experiments and Outcomes
 Experiment – the observation of some activity
or the act of taking a measurement
 Roll dice
 Measure dimension of a manufactured good
 Conduct market survey
 Outcome – a particular result of an
experiment
 Sum of the dice
 Length of the dimension
 Proportion of respondents that favor a product

Events
 Sample space - all possible outcomes of an
experiment
 Dice rolls of 2, 3, …, 12
 All real numbers corresponding to a dimension
 A proportion between 0 and 1
 An event is a collection of one or more
outcomes from
 Obtaining a 7 or 11 on a roll of dice
 Having a manufactured dimension larger or
smaller than the required tolerance
 The proportion of respondents that favor a
product is at least 0.60
Probability
 Probability – a measure (between 0 and 1) of
the likelihood that an event will occur
 Three views:
 Classical definition: based on theory
 Relative frequency: based on empirical data
 Subjective: based on judgment
Classical Definition
 Probability = number of favorable outcomes
divided by the total number of possible
outcomes
 Example: There are six ways of rolling a 7
with a pair of dice, and 36 possible rolls.
Therefore, the probability of rolling a 7 is
6/36 = .167.
Relative Frequency Definition
 Probability = number of times an event has
occurred in the past divided by the total
number of observations
 Example: Of the last 10 days when certain
weather conditions have been observed, it
has rained the next day 8 times. The
probability of rain the next day is 0.80

Subjective Definition
 What is the probability that the New York
Yankees will win the World Series this year?
 What is the probability your school will win its
conference championship this year?
 What is the probability the NASDAQ will go
up 3% next week?
Probability Rules
1. Probability associated with any outcome must be
between 0 and 1
2. Sum of probabilities over all possible outcomes
must be 1.0
• Example: Flip a coin three times
• Outcomes: HHH, HHT, HTH, THH, HTT, THT, TTH,
TTT
• Each has probability of (1/2)
3
= 1/8
Mutually Exclusive Events
 Two events are mutually exclusive if the
occurrence of any one means that none
of the other can occur at the same time.
C
B
B & C
A
B
Probability Calculations
1. Probability of any event is the sum of
the outcomes that compose that event
2. If events A and B are mutually exclusive,
then
P(A or B) = P(A) + P(B)
3. If events A and B are not mutually
exclusive, then
P(A or B) = P(A) + P(B) – P(A and B)

Example
 What is the probability of obtaining exactly
two heads or exactly two tails in 3 flips of a
coin?
 These events are mutually exclusive. Probability =
3/8 + 3/8 = 6/8
 What is the probability of obtaining at least
two tails or at least one head?
 A = {TTT, TTH, THT, HTT}, B = {TTH, THT, THH,
HTT, HTH, HHT, HHH} The events are not
mutually exclusive.
 P(A) = 4/8; P(B) = 7/8; P(A&B) = 3/8. Therefore,
P(A or B) = 4/8 + 7/8 – 3/8 = 1
Conditional Probability
 Conditional probability – the probability
of the occurrence of one event, given
that the other event is known to have
occurred.
P(A|B) = P(A and B)/P(B)
Example
 What is the probability of exactly 2 heads in
three flips, given that the first flip is a head?
HHH, HHT, HTH, THH, HTT, THT, TTH, TTT
 A = exactly 2 heads in 3 flips
 B = first flip is a head, P(B) = 4/8
 A&B = {HHT, HTH} P(A&B) = 2/8
 P(A|B) = P(A&B)/P(B) = (2/8)/(4/8) = ½
 i.e.:
B = {HHH, HHT, HTH, HTT}; A shown in red

Joint Probability
 Joint probability table – probabilities of
outcomes of two different variables
 Variable 1: Number of heads in three flips
of a coin
 Variable 2: First flip (H or T)
Frequencies Joint Probabilities
Marginal Probability
P(A|B) = P(A&B) / P(B) = (1/4)/(1/2) = ½
Marginal
probabilities
Other Uses for Conditional
Probability
 P(A and B) = P(B|A)*P(A)
 P(A) = P(A|B
1
)*P(B
1
) + P(A|B
2
)*P(B
2
) +
… + P(A|B
k
)*P(B
k
)
where B
1
, B
2
, …, B
k
are k mutually
exclusive and exhaustive outcomes.
Statistical Independence
 Two events A and B are statistically
independent if the probability of A given
B equals the probability of A; in other
words P(A|B) = P(A) or P(B|A) = P(B).
 Implications for marketing
Bayes’ Theorem
 Suppose that A
1
, A
2
, …, A
k
is a set of
mutually exclusive and collectively
exhaustive events, and seek the
probability that some event A
i
occurs
given that another event B has occurred.
P(B|A
i
)*P(A
i
)
P(A
i
|B) = --------------------------------------------------------------------
P(B|A
1
)*P(A
1
) + P(B|A
2
)*P(A
2
) + … + P(B|A
k
)*P(A
k
)
Example
 Suppose that a digital camera manufacturer is
developing a new model. Historically, 70 percent of
cameras that have been introduced have been
successful it terms of meeting a required return on
investment, while 30 percent have been unsuccessful.
Analysis of past market research studies, conducted
prior to worldwide market introduction, has found that
90 percent of all successful product introductions had
received favorable consumer response, while 60
percent of all unsuccessful products had also received
favorable consumer response. If the new model
receives a favorable response from a market research
study, what is the probability that it will actually be
successful?
Example (continued)
 Events
 A
1
= new camera successful, P(A
1
) = 0.7
 A
2
= new camera unsuccessful, P(A
2
) = 0.3
 B
1
= marketing response favorable,
 B
2
= marketing response unfavorable
 Other know information:
 P(B
1
|A
1
) = 0.9
 P(B
1
|A
2
) = 0.6

Applying Bayes’ Theorem
P(B
1
|A
1
)*P(A
1
)
P(A
1
|B
1
) = -------------------------------------------------------------------------
P(B
1
|A
1
)*P(A
1
) + P(B
1
|A
2
)*P(A
2
) + … + P(B
1
|A
k
)*P(A
k
)

(0.9)(0.7)
P(A
1
|B
1
) = -------------------------------- = 0.778
(0.9)(0.7) + (0.6)(0.3)

 Although 70 percent of all previous new models have been
successful, knowing that the marketing report is favorable
increases the likelihood to 77.8 percent.
Random Variables
 Random variable – a function that assigns a real
number to each element of a sample space (a
numerical description of the outcome of an
experiment). Random variables are denoted by capital
letters, X, Y, …; specific values by lower case letters,
x, y, …
 Random variables may be discrete, with a finite or
infinitely countable number of values (number of
inaccurate orders or number of imperfections on a
car) or continuous, having any real value possibly
within some limited range (tomorrow’s temperature)
Examples of Random Variables
 Experiment: flip a coin 3 times.
 Outcomes: TTT, TTH, THT, THH, HTT, HTH, HHT,
HHH
 Random variable: X = number of heads.
 X can be either 0, 1, 2, or 3.
 Experiment: observe end-of-week closing
stock price.
 Random variable: Y = closing stock price.
 X can be any nonnegative real number.

Probability Distributions
 Probability distribution – a characterization of
the possible values a random variable may
assume along with the probability of
occurrence.
 Probability distributions may be defined for
both discrete and continuous random
variables.

Discrete Random Variables
 Probability mass function f(x): specifies the
probability of each discrete outcome
 Two properties:
0 s f(x
i
) s 1
¿f(x
i
) = 1

 Cumulative distribution function, F(x): specifies
the probability that the random variable will be
less than or equal to x.

Example: Census Data
Charts for f(x) and F(x)
Example: Roll Two Dice
Sum
of
dice, x
2 3 4 5 6 7 8 9 10 11 12
f(x)
1/36 2/36 3/36 4/36 5/36 6/36 5/36 4/36 3/36 2/36 1/36
F(x)
1/36 3/36 6/36 10/36 15/36 21/36 26/36 30/36 33/36 35/36 1.0
Joint Probability Distributions
 Joint probability distribution of two random variables,
f(x,y): specifies that probability that X = x and Y = y
simultaneously.
Not a High
High School
Some College
No Degree
Associate's
Degree
Bachelor's
Degree
Degree
Age
25-34 0.027 0.073 0.045 0.020 0.049 0.014 0.228
35-44 0.031 0.088 0.047 0.024 0.047 0.021 0.258
45-54 0.026 0.063 0.035 0.017 0.035 0.022 0.198
55-64 0.026 0.048 0.019 0.007 0.017 0.012 0.129
65-74 0.030 0.038 0.014 0.004 0.010 0.007 0.104
75 and older 0.031 0.027 0.011 0.003 0.007 0.004 0.082
0.172 0.338 0.172 0.075 0.164 0.080 1.000
Joint
Marginal
Continuous Random Variables
 Probability density function, f(x), a continuous
function that describes the probability of
outcomes for the random variable X. A
histogram of sample data approximates the
shape of the underlying density function.
Properties of Probability
Density Functions
 f(x) > 0 for all x
 Total area under f(x) = 1
 There are always infinitely many values for X
 P(X = x) = 0
 We can only define probabilities over
intervals: e.g.,
P( c < X < d), P(X < c), or P(X > d)

Cumulative Distribution
Function
 F(x) specifies the probability that the random
variable X will be less than or equal to x; that
is, P(X s x).
 F(x) is equal to the area under f(x) to the left
of x
 The probability that X is between a and b is
the area under f(x) from a to b = F(b) – F(a)

Example
f(x) = 4x – 16 for 4 s x s 4.5
= -4x + 20 for 4.5 s x s 5
F(x) = 2x
2
– 16x + 32 for 4 s x s 4.5
= -2x
2
+ 20x – 49 for 4.5 s x s 5
P(X > 4.7) = F(5) – F(4.7) = 1 - 0.82 = 0.18
Cumulative Distribution Function
Expected Value and Variance
of Random Variables
 Expected value of a random variable X is the theoretical
analogy of the mean, or weighted average of possible
values:

 Variance and standard deviation of a random variable X:
¿
·
1 = i
i i
) f(x x = E[X]
¿
·
1 j =
j
2
j
) f(x E[X]) - (x = Var[X]
¿
·
1 j =
j
2
j X
) f(x E[X]) - (x = o
Example
 You play a lottery in which you buy a ticket
for \$50 and are told you have a 1 in 1000
chance of winning \$25,000. The random
variable X is your net winnings, and its
probability distribution is
x f(x)
-\$50 0.999
\$24,950 0.001
Calculations
 E[X] = -\$50(0.999) + \$24,950(0.001) =
-\$25.00
 Var[X] = (-50 - [-25.00])2(0.999) +
(24,950 - [-25.00])2(0.001) = 624,375
Functions of Random
Variables
 Two useful formulas:
E[aX] = aE[X]
Var[aX] = a
2
Var[X]
where a is any constant
Discrete Probability
Distributions
 Bernoulli
 Binomial
 Poisson
Bernoulli Distribution
 A random variable with two possible
outcomes (x = 0 and x = 1), each with
constant probabilities of occurrence:

f(x) = p if x = 1
f(x) = 1 – p if x = 0

Binomial Distribution
 n independent replications of a Bernoulli trial
with constant probability of success p on each
trial

 Expected value = np
 Variance = np(1-p)
( )
x
n
=
)! ( !
!
x n x
n
÷
( )
otherwise
n x for p p
x
n
x f
x n x
0
,..., 2 , 1 , 0 ) 1 ( ) ( = ÷ =
÷
Example
 If the probability that any individual will react
positively to a new drug is 0.8, what is the probability
distribution that 4 individuals will react positively out
of a sample of 10
Excel Function
 BINOMDIST(number_s, trials, probability_s,
cumulative)
Examples of the Binomial
Distribution
Poisson Distribution
 Models the number of occurrences in some unit of
measure, e.g., events per unit time, number of items
per order
 X = number of events that occur; x = 0, 1, 2, …

 Expected value = ì; variance = ì
 Poisson approximates binomial when n is large and p
small

otherwise
x for
x
e
x f
x
0
,... 2 , 1 , 0
!
) (
=
= =
÷
ì
ì
Example
 Suppose that the average number of customers
arriving at an ATM during lunch hour is l = 12
customers per hour. The probability that exactly x
customers will arrive during the hour is
Excel Function
 POISSON (x, mean, cumulative)

Poisson Distribution (ì = 12)
Properties of Continuous
Distributions
 Continuous distributions have one or
more parameters that characterize the
density function:
 Shape parameter – controls the shape of
the distribution
 Scale parameter – controls the unit of
measurement
 Location parameter – specifies the location
relative to zero on the horizontal axis
Uniform Distribution
 Density function

 Distribution function
f x
b a
if a x b ( ) =
÷
s s
1
0
1
if x a
F x
x a
b a
if a x b
if b x
<
=
÷
÷
s s
<
( )
a b
x
f(x)
µ = (a + b)/2
o
2
= (b – a)
2
/12
a = location
b – a = scale
Normal Distribution
 Familiar bell-shaped curve.
 Symmetric, median = mean = mode; half the area is on
either side of the mean
 Range is unbounded: the curve never touches the x-axis
Parameters
 Mean, µ (location)
 Variance o
2
> 0 (scale)
 Density function:
f(x) =
e
-(x- )
2
µ
o
to
2
2
2
2
Standard Normal Distribution
 Standard normal: mean = 0, variance = 1,
denoted as N(0,1)

 See Appendix Table A.1
Standardized Normal Values
 Transformation from N(µ,o) to N(0,1):

 Standardized z-values are expressed in units
of standard deviations of X.

o
µ ÷ x
= z
Areas Under the Normal
Density
 About 68.3% is within one sigma of the mean
 About 95.4% is within two sigma of the mean
 About 99.7% is within three sigma of the
mean
Normal Probability
Calculations
 Customer demand averages 750
units/month with a standard deviation of
100 units/month.
 Find P(X>900), P(X>700), P(700<X<900)
and the level of demand that will be
exceeded only 10% of the time.
P(X>900)
5 . 1
100
750 900
= z =
÷
Area = 0.9332
Area = 0.0668
P(X>700)
5 . 0
100
750 700
= z ÷ =
÷
P(X > 700) = P(Z > -0.5) = 1 - P(Z < -0.5).
From Table A.1, P(Z < -0.5) = 0.3085.
Thus, P(X > 700) = 1 - 0.3085 = 0.6915.
P(700<X<900)
First standardize the values of 700 and 900, obtaining
z = -0.5 and z = 1.5. Then using Table A.1, we have
P(700 < X < 900) = F(900) – F(700)
= P(Z < 1.5) - P(Z < -0.5) = 0.9332 - 0.3085 = 0.6247
P(X>x)>.10
28 . 1
100
750
= z =
÷ x
First, find the value
of z from Table A.1,
then solve for x
x = 878
Excel Support
 NORMDIST(x, mean, standard_deviation,
cumulative)
 NORMSDIST(z): same as Table A.1
 STANDARDIZE(x, mean, standard_deviation)
computes z-values
Computing Normal Probabilities
Triangular Distribution
 Three parameters:
 Minimum, a
 Maximum, b
 Most likely, c
 a is the location parameter;
 (b – a) the scale parameter,
 c the shape parameter.
2
2
0
( )
( )( )
( )
( )
( )( )
x a
b a c a
if a x c
f x
b x
b a b c
if c x b
otherwise
÷
÷ ÷
s s
=
÷
÷ ÷
< s
Mean = (a + b + c)/3
Variance = (a
2
+ b
2
+ c
2
– ab – ac – bc)/18
Exponential Distribution
 Models events that occur randomly over time
 Customer arrivals, machine failures

Properties of Exponential
 Density function
 Distribution function
 Mean = 1/ì
 Variance = 1/ì
2
 Excel function EXPONDIST(x, lambda, cumulative).
f(x) = ìe
-ìx
, x > 0

F(x) = 1 - e
-ìx
, x > 0

Lognormal Distribution
 X is lognormal if the distribution of lnX is
normal
 Positively skewed, bounded below by 0
 Models task times, stock and real estate
prices
Gamma Distribution
 Family of distributions defined by shape o,
scale |, and location L
 Defined for x > L
 Models task times, time between events,
inventories
Weibull Distribution
 Family of distributions defined by shape o,
scale |, and location L
 When L = 0 and | = 1, same as exponential
with ì = 1/o.
 Models results from life tests, equipment
failures, and task times.
Beta Distribution
 Defined over the range (0, s)
 Two shape parameters o and |; if equal, beta
is symmetric; o < |, positively skewed; if
either equals 1 and the other > 1, “J”
shaped.
Geometric Distribution
 Sequence of Bernoulli trials that describes the
number of trials until the first success.
Negative Binomial
 Distribution of number of trials until the r
th

success.
Logistic Distribution
 Describes the growth of a population over
time.
Pareto Distribution
 Distribution in which a small proportion of
items accounts for a large proportion of some
characteristic.
PHStat Tool: Probability &
Probability Distributions
 PHStat menu > Probability & Probability
Distributions > choice of distribution:
• Normal
• Binomial
• Exponential
• Poisson
• Hypergeometric

PHStat Output

Sign up to vote on this title