7 Sampling Distributions

Looking Back
■ Recall that parameters such as m, s2 , and p completely characterize, or describe, a
population.
■ These parameters are constant and usually unknown, but we would like to estimate these
values, or draw a conclusion about these parameters.

Looking Forward
=
■ Understand and utilize methods to find a sampling distribution (for X and for P ) and use
it in statistical inference.
■ Discover the central limit theorem, and solve probability and inference problems using
this result.

How much was the typical

foreclosure check?
In May 2013, millions of Americans received checks from the federal government
as part of a program to resolve the housing foreclosure crisis.1 Check recipients
had contested foreclosure for a variety of reasons. For example, some claimed
that the mortgage company did not follow the proper procedure and others
asserted that a mistake had been made in the foreclosure process. Thirteen
mortgage servicing companies had been cited for errors, misrepresentations,
and “robo-signing” of documents.
The checks were in amounts from $300 to $125,000. Suppose a spokesperson
from the U.S. Treasury claimed that the mean check amount was $800 with
standard deviation $125.
A random sample of 40 Americans compensated under this program was
obtained. The amount of each check was recorded, and the sample mean check
amount for these Americans was $755. The concepts presented in this chapter
will allow us to find the distribution of the sample mean, to answer probability
questions about this random variable, and to make inferences about a popula-
tion mean. These ideas all rely on the central limit theorem, the most important
result in probability and statistics!

CONTENTS
7.1 Statistics, Parameters, and Sampling Distributions
7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem
7.3 The Distribution of the Sample Proportion
© Kristoffer Tripplaar/Alamy

295
296 CHAP T ER 7 Sampling Distributions

7.1 Statistics, Parameters, and Sampling Distributions

The terms parameter and statistic have been used in previous chapters in intuitive con-
texts. The following definitions distinguish between measures represented by symbols
such as m, s2, and p, and the quantities used to estimate these values.

Definition
A parameter is a numerical descriptive measure of a population.
A statistic is any quantity computed from values in a sample.

A CLOSER L OK
1. A parameter is a population quantity. It is used to describe some characteristic of a
population. Usually we cannot measure a parameter; it is an unknown constant that we
would like to estimate.
2. A statistic is any sample quantity. There are infinitely many quantities one could com-
pute using the data in a sample. For example, x and ~x are statistics, as is the sum of the
smallest and the largest values divided by 2.

The difference between a parameter and a statistic is key: Parameters describe popula-
tions, and statistics describe samples. We use statistics to make inferences about parameters.
Therefore, the properties of a statistic are important.

Example 7.1 Parameter Versus Statistic

In each of the following statements, identify the boldface number as the value of a popu-
lation parameter or a sample statistic.
a. In a recent survey of Americans, 52% of Republicans say global warming is happening.
b. A spokesman for a large insurance agency reported that the proportion of all women
with some form of life insurance is 0.32.
c. The U.S. Department of Transportation recently reported that the mean age of all high-
way bridges in the United States is 42 years.
d. The manager of a large hotel located near Disney World indicated that 20 selected
guests had a mean length of stay equal to 5.6 days.
e. In a recent survey about the Canadian economy, 7.2% of the individuals who responded
were unemployed.

SOLUTION
a. 52% is a statistic. This number describes a characteristic of a sample of Republicans.
b. 0.32 is a parameter. This number describes a characteristic of the entire population of
women.
c. 42 is a parameter. This number is a characteristic of all the highway bridges in the
United States.
d. 5.6 is a statistic. This number describes a characteristic of the sample of 20 guests.
e. 7.2 is a statistic. This number describes a characteristic of the sample of individuals
who responded to the survey.

TRY IT NOW GO TO EXERCISE 7.9

Suppose the mean is computed for a sample of size n from a population of interest.
Denote this value x1 . In a second sample of size n from the same population, let x2 be the
 7.1 Statistics, Parameters, and Sampling Distributions 297

mean. It is reasonable to expect x1 and x2 to be close to each other but not equal. The
important realization is that x1 will be different from x2. In fact, the sample mean will dif-
fer from sample to sample. This statistic is subject to sampling variability. Therefore, the
sample mean, X , is a random variable, and X has a mean, a variance, a standard devia-
tion, and a probability distribution. This distribution is called a sampling distribution.
Statistics are random variables! Any statistic is a random variable, because the value of the statistic differs from
sample to sample. One cannot predict the value of a statistic with absolute certainty.
To make a reliable inference based on a specific statistic, we need to know the proper-
ties of the distribution of the statistic.

Definition
The sampling distribution of a statistic is the probability distribution of the statistic.

A CLOSER L OK
1. A sampling distribution (like any random variable) describes the long-run behavior of
the statistic.
2. We can use either of two techniques to obtain, or find, a sampling distribution.
a. Recall that to approximate the distribution of a population, we construct a histo-
gram (or stem-and-leaf plot) using values from the population. If the sample is
representative, then the histogram should be similar in shape, center, and spread to
the population distribution.
Similarly, to approximate the distribution of a statistic, we obtain (many) values of
the statistic and construct a histogram. The resulting graph approximates the sam-
pling distribution of the statistic.
For example, to approximate the sampling distribution of the mean of a sample of
size n 5 10 from a population: (i) obtain several samples of size 10 from the popu-
lation; (ii) compute the sample mean for each sample; (iii) construct a histogram
using all the sample means. The histogram approximates the sampling distribution
of the sample mean.
b. In some cases, the exact sampling distribution of a statistic can be obtained. If the
statistic is a discrete random variable, the sampling distribution includes all the
values the statistic assumes and the associated probabilities. If the statistic is a
continuous random variable, the sampling distribution consists of a probability
density curve.

Example 7.2 Stuffed Animals

DATA SET A young girl has started a collection of stuffed animals and currently has five koala bears
TOYS on her dresser. The plush toys vary in color, design, and size. These factors, along with
different stuffing materials, contribute to the various weights of the five koalas: 10.9, 14.5,
17.1, 18.1, and 17.6 ounces.
Suppose a sample of three koalas is selected and the sample median weight is computed.
Find the sampling distribution for the sample median.

SOLUTION
5
The order of selection does not STEP 1 There are a b 5 10 ways to select three koalas from the five in the population.
3
matter here. Therefore, this is a
List all the possible samples, the computed value of the statistic for each sample,
combination.
and the probability of selecting each sample. Use this table to construct the sam-
pling distribution.
298 CHA P T E R 7 Sampling Distributions

STEP 2 Here is the resulting table.

Sample ~x Probability Sample ~x Probability

10.9, 14.5, 17.1 14.5 0.1 10.9, 18.1, 17.6 17.6 0.1
10.9, 14.5, 18.1 14.5 0.1 14.5, 17.1, 18.1 17.1 0.1
10.9, 14.5, 17.6 14.5 0.1 14.5, 17.1, 17.6 17.1 0.1
10.9, 17.1, 18.1 17.1 0.1 14.5, 18.1, 17.6 17.6 0.1
10.9, 17.1, 17.6 17.1 0.1 17.1, 18.1, 17.6 17.6 0.1

The median for each sample is the middle value. The probability of each sample
is 1/10 5 0.1, because we assume that each sample is equally likely.
STEP 3 Of all the possible samples, only three values are possible for the sample median
in this case. Sum the probabilities associated with each value. The probability
~ ~
distribution for the random variable X lists the values (of X ) and the associated
probabilities.
~x 14.5 17.1 17.6
p( ~x ) 0.3 0.4 0.3

TRY IT NOW GO TO EXERCISE 7.12

A CLOSER L OK
~
1. In Example 7.2, X is a discrete random variable. Using Equations 5.1, 5.3, and 5.4, the
Challenge: Find the sampling ~
~ mean, variance, and standard deviation for X are m 5 16.47, s2 5 1.71, and s 5 1.31.
distribution for X if sampling
is done with replacement. 2. The koalas were selected without replacement. Suppose three bears are selected with
Hint: There are 5 3 5 3 5 5 125 replacement. That is, select a stuffed animal, record its weight, and place it back on the
possible samples. dresser. The same koala could be selected two or three times. This sampling scheme
~
changes the probability distribution for X .

DATA SET Example 7.3 Sick Days

SICKDAY Employees at the Hilmar Cheese Company are allowed up to three sick days during a
calendar year.2 Suppose the probability distribution for X, the number of sick days used
by an employee during a year, is given in the following table.
x 0 1 2 3
p(x) 0.40 0.35 0.20 0.05
Two employees are independently selected at random, and the number of sick days is
recorded for each. Consider the statistic M, the maximum number of sick days taken by
either employee. Find the probability distribution for M.

SOLUTION
STEP 1 A sample consists of two numbers: The first represents the sick days for employee
1, and the second denotes the sick days for employee 2. There are 16 possible
samples. Using the multiplication rule, there are two slots to fill: 4 3 4 5 16.
List all the possible samples, the computed value of the statistic for each sample,
and the probability of selecting each sample.
STEP 2 The probability associated with each sample is computed using the indepen-
Anyra/[Link] dence assumption. For example, the probability the first employee used one sick
day and the second employee used two sick days is given by

Each event is given in terms of a P(X 5 1 d X 5 2) Independent events.

random variable. 5 P(X 5 1) # P(X 5 2) Use the probability distribution for X.

5 ( 0.35 ) ( 0.20 ) 5 0.07

 7.1 Statistics, Parameters, and Sampling Distributions 299

STEP 3 Use the table below to construct the sampling distribution.

Sample m Probability Sample m Probability

0, 0 0 0.1600 2, 0 2 0.0800
0, 1 1 0.1400 2, 1 2 0.0700
0, 2 2 0.0800 2, 2 2 0.0400
0, 3 3 0.0200 2, 3 3 0.0100
1, 0 1 0.1400 3, 0 3 0.0200
1, 1 1 0.1225 3, 1 3 0.0175
1, 2 2 0.0700 3, 2 3 0.0100
1, 3 3 0.0175 3, 3 3 0.0025

The maximum for each sample is the largest of the two values.
Remember, a capital letter, such STEP 4 There are four possible values for the discrete random variable M. Sum the prob-
as M, represents a random abilities associated with each value. The probability distribution is given in the
variable. The corresponding table below.
lowercase letter m denotes a
m 0 1 2 3
specific value the random variable
M can assume. p(m) 0.1600 0.4025 0.3400 0.0975

Challenge: Find the mean, TRY IT NOW GO TO EXERCISE 7.14

variance, and standard deviation
for the random variable M. In almost all observational studies it is assumed that the data are obtained from a
simple random sample. Usually the sampling is done without replacement. Consider an
exit poll or a study of the time spent each week on lawn care by homeowners. An
individual is selected from the population and an observation is recorded. The individual
is not placed back into the population. There is no chance the individual will be selected
again.
If sampling is done without replacement, individual responses are dependent. How-
ever, if the population is large enough and the sample is small relative to the size of the
population, then the responses are almost independent. Calculating probabilities assum-
ing independence results in little loss of accuracy. As a rule of thumb, if the sample size
is at most 5% of the total population, then successive observations can be considered
independent. Even though sampling is done without replacement, the data are assumed to
be part of a simple random sample.
STEPPED
STEPPED TUTORIAL
TUTORIALS Recall the following definition from Chapter 1.
SIMPLE RANDOM
BOX PLOTS
SAMPLES
Definition
A (simple) random sample (SRS) of size n is a sample selected in such a way that every
possible sample of size n has the same chance of being selected.

A CLOSER L OK
1. Suppose the population is finite of size N and the sample is of size n. The number of
N
possible simple random samples is a b.
n
2. A random sample consists of individuals or objects, and a variable is a characteristic
of an individual or object. A value of the variable is obtained for each member of the
random sample. We often refer to the values of the variable as the random sample
rather than the individuals. For example, consider a study in which the amount of the
trace element chromium is measured in coal from around the world. The random
300 CHAP TER 7 Sampling Distributions

sample consists of pieces of coal (objects) and the values are the chromium measure-
ments. However, it is common practice to say, “Consider the random sample of chro-
mium measurements.”
3. Unless stated otherwise, all data presented in this text are obtained from a simple ran-
dom sample.

DATA SET Example 7.4 Cattle Feeds

FEED The following table lists the percentage of protein for the entire population of 20 different
cattle feeds.3
Note: The population mean is the
sum of all the observations 18 19 34 26 11 3 20 9 12 11
divided by n 5 20: m 5 15.8. 13 24 4 10 24 10 25 15 12 16

VIDEO TECH MANUALS Find an approximate sampling distribution for the sample mean of five observations from
EXEL DISCRIPTIVE
SAMPLING FROM this population.
A DATA SET
SOLUTION
20
STEP 1 There are a b 5 15,504 possible samples of size 5. Instead of considering
5
every one of these samples, select some (say, 100) samples of size 5, compute
the mean for each sample, and construct a histogram of the sample means.
STEP 2 The table below lists the first few samples of size 5 and the mean for each sample.

Sample x Sample x
25 13 18 10 12 15.6 24 3 13 20 10 14.0
3 10 13 12 11 9.8 4 24 10 9 19 13.2
19 9 10 11 16 13.0 20 16 26 24 15 20.2
( ( ( (

STEP 3 Figure 7.1 shows a histogram of the sample means for 100 samples of size 5.
STEP 4 There are some very interesting, curious results here. Even though the population is
not normally distributed (the population is finite; and consider a graph of the 20
cattle feeds in Figure 7.2), the shape of the sampling distribution appears to be
approximately normal! In addition, the center of the sampling distribution of the
mean is approximately the population mean ( m 5 15.8 ) . Although the relationship
between the original population parameters and the sampling distribution parame-
ters is not clear, there is certainly less variability in the sampling distribution than in
the population distribution ( s 5 8.11 ) . These three observations suggest the exact
distribution of the sample mean, as discussed in the next section.

8
25

20 6
Frequency
Frequency

15
4
10
2
5

6 8 10 12 14 16 18 20 22 24 26 5 10 15 20 25 30 35
Mean Percentage

Figure 7.1 Histogram of the sample Figure 7.2 Histogram of the

means. original population.
 7.1 Statistics, Parameters, and Sampling Distributions 301

SECTION 7.1 EXERCISES

Concept Check weighed the contents in the truck bed. The mean weight
was 1037 pounds.
7.1 Fill in the Blank A parameter describes a . b. A toy manufacturer issued a recall on a small wooden
A statistic describes a . toy car because the wheels could break off and pose a
7.2 True/False Statistics are used to estimate parameters. choking hazard for small children. The proportion of
buyers who took advantage of the recall was reported to
7.3 True/False Any statistic is a random variable. be 0.45.
7.4 True/False An exact sampling distribution can never c. A recent study indicated that more people today are rising
be obtained. before 6:00 A.M. each weekday to prepare for work and to
help children get ready for school. In a random sample of
7.5 Short Answer Describe a method to approximate a sam- 500 adults, 42% said they get up each weekday before
pling distribution. 6:00 A.M.
7.6 Fill in the Blank If sampling is done without replacement, d. In a random sample of dentists, 80% recommended a
individual responses are . certain product to help whiten teeth.
e. During a recent winter, the month of January was
7.7 Short Answer Suppose a population is finite of size N particularly cold. A power company in Pennsylvania
and the sample is of size n . How many possible simple random reported that the mean number of kilowatts used by each
samples are there? customer during January was 1346.
7.8 Short Answer Describe three possible relationships 7.11. Parameter Versus Statistic In each of the following
between the distribution of the sample mean and the distribu- statements, identify the boldface number as the value of a pop-
tion of the original population. ulation parameter or a sample statistic.
a. The maker of a new, enhanced water product conducted a
Applications survey to determine if people are drinking enough liquids
each day. In a random sample of Americans, 90% were
7.9 Parameter Versus Statistic In each of the following chronically dehydrated.
statements, identify the boldface number as the value of a pop- b. Last year, the proportion of all robberies in Washington,
ulation parameter or a sample statistic. D.C., that involved a cellphone was 0.42.
a. A researcher in Boston conducted a study to investigate c. NASA reported that the mean time for all spacewalks at
the prevalence of Alzheimer’s disease in people over the International Space Station was 45 minutes.
the age of 85. The results indicated that 47.2% of the d. The International Energy Agency reported that 93% of
patients studied experienced symptoms consistent with Canada’s renewable energy comes from hydroelectric
the disease. sources.
b. A brokerage firm reported that the mean dividend paid e. In a random sample of publishing companies, 23% of
by all Fortune 500 companies for the year 2003 was sales were ebooks.
$0.62.
c. A consumer magazine tested a random sample of 19 7.12 Sports and Leisure The Sports Authority sells five
green teas for taste and health benefits. The mean cost per different medicine balls with weights (in kilograms) given in
cup of tea was reported to be 14 cents. the table below. MEDBALL

d. A computer manufacturer claims that the mean lifetime 10 12 15 18 25

of laptop batteries is 6.7 hours.
e. A new law requires that all propane tanks (used, for a. Find the population mean and the population median.
example, on barbecue grills) must have new, safer valves b. Suppose a random sample of size 3 is selected from this
before they can be refilled. A gas company that refills population without replacement. Find the sampling
tanks took a random sample of customers on a recent distribution of the sample mean. Find the mean, variance,
Saturday morning and reported the mean age of the and standard deviation for the sample mean.
propane tanks to be 3.45 years. c. Suppose a random sample of size 3 is selected from this
population without replacement. Find the sampling
7.10 Parameter Versus Statistic In each of the following
distribution of the sample median. Find the mean,
statements, identify the boldface number as the value of a
variance, and standard deviation for the sample median.
population parameter or a sample statistic.
d. Compare the mean of the sample mean with the
a. A half-ton pickup truck is designed to safely carry a
population mean. Compare the mean of the sample
maximum of 1000 pounds. The Department of
median with the population median.
Transportation is concerned that drivers are hauling much
heavier loads. The State Police randomly stopped 50 7.13 Manufacturing and Product Development The
pickup trucks on an interstate highway and carefully coverage of a gallon of paint depends on the surface, the type
302 CHA PT ER 7 Sampling Distributions

and quality of the paint, and the applicator. A local hardware 7.17 Education and Child Development The number of
store stocks 30 different paints. PAINT copies sold (in millions) for the all-time best-selling children’s
a. Use a computer or calculator to draw 50 random samples books are given in the table below. BOOKS
of size 5 without replacement, and compute the mean for
each sample. 9.9 9.7 7.1 7.0 6.8
b. Construct a histogram of the sample means.
c. Use the histogram to approximate the sampling
Suppose two of these books are selected at random without
distribution of the mean. What is the approximate shape
replacement.
of the distribution? What is the approximate value of the
a. Find the sampling distribution of the sample mean.
mean of the sample mean?
b. Find the sampling distribution of the total number of
d. Find the population mean. How does this compare with
books sold.
the approximate mean of the sampling distribution?
7.14 Demographics and Population Statistics For plan- Extended Applications
ning purposes, U.S. Bancorp has determined the probability
distribution for the retirement age, X, of employees in the mort- 7.18 Fuel Consumption and Cars An automobile manu-
gage work group. The probability distribution for X is given in facturer lists several specifications for every one of its cars.
the table below. RETIRE For example, the length, width, wheelbase, turning circle,
curb weight, and interior room measurements are readily
x 64 65 66 available to customers. The acceleration time from 0 to 60
p(x) 0.1 0.7 0.2 miles per hour (in seconds) for 10 cars is given in the table
below.6 CARS
a. Find the mean of X.
b. Suppose two employees from this work group are selected
at random. Find the exact probability distribution for the 2.5 2.8 3.0 3.5 3.4 3.7 3.9 4.1 4.7 5.1
sample mean, X .
c. Find the mean of X . How does this compare with your a. Use a computer or calculator to draw 30 random samples
answer in part (a)? of size 3 without replacement, and compute the standard
deviation for each sample.
7.15 Public Health and Nutrition The U.S. Food & Drug b. Construct a histogram of the sample standard deviations.
Administration regulates the use of certain terms used on food c. Use the histogram to describe the shape of the sampling
labels—for example, reduced fat, low fat, and light. Lean meat distribution.
must contain less than 4.5 grams of saturated fat per serving.4 d. Compute the population standard deviation. Find an
Suppose the probability distribution for the amount of saturated approximate mean of the sampling distribution. How
fat in a randomly selected serving of lean meat is given in the do these two numbers compare?
table below. USFDA
7.19 Travel and Transportation American Airlines offers
x 0 1 2 3 4 limited first-class seating to passengers on a flight from Newark
p(x) 0.50 0.25 0.10 0.10 0.05 to Los Angeles. The probability distribution for the number of
passengers in first class on a randomly selected flight is a ran-
a. Suppose two servings of lean meat are selected at dom variable, X. The probability distribution for X is given in
random. Find the exact probability distribution for the the table below. FLIGHT
~
sample median amount of saturated fat, X .
~
b. Find the mean, variance, and standard deviation of X . x 5 6 7 8
7.16 Biology and Environmental Science In April 2013, p(x) 0.50 0.30 0.15 0.05
Eddie (“Flathead Ed”) Wilcoxson caught a flathead catfish that
weighed 76.52 pounds on Bartlett Lake in Arizona.5 The cat- a. Find the variance of X.
fish was 53.5 inches long and its girth measured 34.75 inches. b. Suppose two American Airlines cross-country flights are
Suppose five catfish were caught on the same lake with randomly selected. Find the exact probability distribution
weights (in pounds) given in the following table. CATFISH for the sample variance, S 2.
c. Find the mean of S 2. How does this compare with your
11 15 32 21 27 answer in part (a)?
7.20 Sports and Leisure The times (in minutes) for
A random sample of three of these catfish is selected without three men in the 3M Half Marathon are given in the
replacement. following table.7 RUNTIME
a. Find the sampling distribution of the sample mean X .
b. Find the sampling distribution of the total length for all
three catfish, T. 66.08 68.58 77.10
 7.1 Statistics, Parameters, and Sampling Distributions 303

Suppose a random sample of two of these times is selected c. Use technology to select 100 random samples of size 10
with replacement. from this population. Find the sample mean for each sample.
a. Find the sampling distribution of the minimum time. d. Construct a histogram of the 100 sample means. Where is
b. Find the sampling distribution of the total time. the histogram centered? How does this value compare to m?
e. Find the standard deviation for the 100 sample means.
7.21 Economics and Finance The Kentucky Derby is
How does this value compare to s?
run every May at Churchill Downs. This extravagant affair
features mint juleps and unusual woman’s hats. The jockeys
Challenge
with the most Kentucky Derby mounts are given in the fol-
lowing table.8 DERBY 7.25 Which Estimator Is Better? Each of the following
graphs shows the probability distribution for two statistics that
Jockey Mounts
could be used to estimate a parameter u (describing an underly-
Bill Shoemaker 26 ing population). Select the statistic that would be a better esti-
Pat Day 22 mator of u and justify your answer.
Eddie Arcaro 21 a.
Laffit Pincay Jr. 21
S2
Suppose a random sample of two of these jockeys is selected
at random without replacement.
a. Find the sampling distribution for the maximum number
of mounts. S1
b. Find the sampling distribution for the total number of
mounts.
7.22 Psychology and Human Behavior Five people own !
and operate the Victrola Coffee Shop in Seattle. Each person is b.
married, and the number of years each has been married is
given below. COFFEE
S1 S2
5 3 7 2 12

Suppose a random sample of two of the owners is selected with

replacement. Let D be a statistic defined to be the absolute
value of the difference in the number of years each has been
married. For example, if 2 and 7 were selected, the value of D
!
would be 0 2 2 7 0 5 0 25 0 5 5. Find the sampling distribution
c.
of D.
7.23 Simple Random Sample Consider the population con-
sisting of the numbers 1 through 50.
a. Find the population mean m. S2
b. Use technology to select 50 random samples of size 10
from this population. Find the sample mean for each
S1
sample.
c. Construct a histogram of the 50 sample means. Where is
the histogram centered?
!
7.24 Sports and Leisure The Toronto Rock are in the
d.
National Lacrosse League. The total number of shots on goal
for each player during the 2013 season is given in the following
table.9 LACROSSE

215 129 138 133 114 68 70 33 S2

45 6 8 15 9 18 4 4
9 15 0 12 5 4 0
S1
a. Find the population mean number of shots on goal, m.
b. Find the population standard deviation of shots on goal, s. !
304 CHA P TE R 7 Sampling Distributions

7.2 The Sampling Distribution of the Sample Mean

and the Central Limit Theorem
It seems reasonable to use a value of the sample mean, x, to make an inference about the
population mean m. However, as discussed in the previous section, the sample mean var-
ies from sample to sample. This sampling variability makes it difficult to know how far
a specific x is from m, or even whether the sample mean is an overestimate or underesti-
mate. In this section, the sampling variability of the sample mean is completely character-
ized. The probability distribution of the sample mean, X , can be used to make a sensible
guess about the true value of the population mean.
To make a reliable estimate, we need to know the exact probability distribution of the
sample mean. As the next few examples will show, the distribution of X is related to n,
the sample size, and to the parameters of the original, or underlying, population.

Example 7.5 Approximate Distribution of X

Consider a population consisting of the numbers 1, 2, 3, . . . , 20. The population mean
is m 5 ( 1 1 2 1 3 1 c1 20 ) /20 5 10.5. Use frequency histograms to approximate
the distribution of the mean, X .

SOLUTION
STEP 1 Consider a random sample of n observations selected with replacement. For
n 5 5, five numbers are selected at random from the population, and the sam-
ple mean is computed. This procedure is repeated 500 times. A histogram of the
resulting 500 sample means is shown in Figure 7.3. For n 5 10, a similar pro-
cedure is followed. The resulting histogram of the sample means is shown in
Figure 7.4.

90 140
80 120
70 100
Frequency

Frequency

60
50 80
40 60
30 40
20
10 20
0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
Sample mean Sample mean

Figure 7.3 Histogram of sample means for Figure 7.4 Histogram of sample means for
n 5 5. n 5 10.

STEP 2 These histograms suggest some very surprising results. Each distribution appears
to be centered near the population mean, 10.5. In addition, the shape of each
The original population is certainly distribution is approximately normal! Also notice that the variability of the sam-
not normally distributed; it is pling distribution decreases as n increases. The sampling distribution for X for
finite, and a histogram is a n 5 10 is more compact than that for n 5 5.
horizontal line. It is amazing and STEP 3 This example implies that the distribution of the sample mean is approximately
perhaps counterintuitive that X normal, with mean equal to the underlying, original, population mean and vari-
has a normal distribution. ance related to the sample size, n.

TRY IT NOW GO TO EXERCISE 7.59

In the next example, the exact sampling distribution of the mean is obtained.
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 305

DATA SET Example 7.6 Essay Sources

ESSAY Mary Dunn, a teacher at The Hill School, has assigned a short research paper in which
each student must write about an endangered species. Ms. Dunn requires each student
to correctly reference at least two sources. Past experience indicates that the number
of sources cited is a random variable, X, with probability distribution given in the table
below.
x 2 3 4
p(x) 0.5 0.3 0.2
Suppose a random sample of three student papers is selected.
a. Find the sampling distribution of X , the sample mean number of sources cited.
b. Find the mean and the variance of the random variable X.
c. Find the mean and the variance of the random variable X .
d. How do the results from (b) compare with the results from (c)?

SOLUTION
a. A sample consists of three numbers: The first represents the number of sources for
student 1, the second for student 2, and the third for student 3. By the multiplication
rule, there are 3 3 3 3 3 5 27 possible samples. For each sample, list the sample
mean and the probability of selecting that sample.
The probability associated with each sample is computed using independence. For
example, the probability of observing the sample 2, 2, 4 is
P(X 5 2 d X 5 2 d X 5 4) Intersection of three events.

5 P(X 5 2) # P(X 5 2) # P(X 5 4) Independence.

5 ( 0.5 ) ( 0.5 ) ( 0.2 ) 5 0.05 Given probabilities.

Use the table below to construct the sampling distribution.

Sample x Probability Sample x Probability
2, 2, 2 2 0.125 3, 3, 4 10 / 3 0.018
2, 2, 3 7/3 0.075 3, 4, 2 3 0.030
2, 2, 4 8/3 0.050 3, 4, 3 10 / 3 0.018
2, 3, 2 7/3 0.075 3, 4, 4 11 / 3 0.012
2, 3, 3 8/3 0.045 4, 2, 2 8/3 0.050
2, 3, 4 3 0.030 4, 2, 3 3 0.030
2, 4, 2 8/3 0.050 4, 2, 4 10 / 3 0.020
2, 4, 3 3 0.030 4, 3, 2 3 0.030
2, 4, 4 10 / 3 0.020 4, 3, 3 10 / 3 0.018
3, 2, 2 7/3 0.075 4, 3, 4 11 / 3 0.012
3, 2, 3 8/3 0.045 4, 4, 2 10 / 3 0.020
3, 2, 4 3 0.030 4, 4, 3 11 / 3 0.012
3, 3, 2 8/3 0.045 4, 4, 4 4 0.008
3, 3, 3 3 0.027

The probability distribution for X is given in the table below. It lists all the values X
can assume and the corresponding probabilities.
x 2 7/3 8/3 3 10 / 3 11 / 3 4
p( x ) 0.125 0.225 0.285 0.207 0.114 0.036 0.008
306 CHA P T ER 7 Sampling Distributions

b. The mean and variance for X are

E ( X ) 5 ( 2 )( 0.5 ) 1 ( 3 ) ( 0.3 ) 1 ( 4 ) ( 0.2 ) 5 2.7
Var ( X ) 5 ( 2 2 2.7 ) 2 ( 0.5 ) 1 ( 3 2 2.7 ) 2 ( 0.3 ) 1 ( 4 2 2.7 ) 2 ( 0.2 ) 5 0.61
c. The mean and variance for X are
E ( X ) 5 ( 2 )( 0.125 ) 1 ( 7/3 ) ( 0.225 ) 1 c1 ( 4 ) ( 0.008 ) 5 2.7
Var ( X ) 5 ( 2 2 2.7 ) 2 ( 0.125 ) 1 c1 ( 4 2 2.7 ) 2 ( 0.008 ) 5 0.2033
d. Notice the following extraordinary relationship between the means and the variances.
E ( X ) 5 2.7 5 E ( X )
The mean of the sample mean is equal to the original population mean!
0.61 Var ( X )
Var ( X ) 5 0.2033 5 5
3 n
The variance of X is equal to the original variance divided by the sample size.

Here is one more approach to illustrate the very important connections between the
distribution of the sample mean and the distribution of the original population.
Consider three original, or underlying, distributions: (1) the standard normal distribu-
tion, a normal distribution with mean 0 and standard deviation 1; (2) a uniform distribu-
tion with parameters a 5 0 and b 5 1; and (3) an exponential distribution with parameter
l 5 0.5. The probability density function and the mean for each distribution are shown
in the first row of Figure 7.5.
Consider the following process for each distribution. Select 500 samples of size n 5 2
and compute the mean for each sample. Construct a histogram of the sample means. The
resulting smoothed histogram is shown in the second row of Figure 7.5. Repeat this pro-
cedure for n 5 5, 10, and 20, for each underlying distribution. These smoothed histo-
grams are also given in Figure 7.5.
Notice the following incredible patterns.
1. If the underlying population is normal, the distribution of the sample mean appears to
be normal, regardless of the sample size. See Figure 7.5(a).
2. Even if the underlying population is not normal, the distribution of the sample mean
becomes more normal as n increases. See Figures 7.5(b) and 7.5(c).
3. The sampling distribution of the mean is centered at the mean of the underlying popu-
lation. See Figures 7.5(a), 7.5(b), and 7.5(c).
4. As the sample size, n, increases, the variance of the distribution of the sample mean
decreases. See Figures 7.5(a), 7.5(b), and 7.5(c).
The previous examples and observations lead to the following properties concerning
the sample mean, X .

Properties of the Sample Mean

Let X be the mean of observations in a random sample of size n drawn from a population
with mean m and variance s2.
1. The mean of X is equal to the mean of the underlying population.

In symbols: mX 5 m.
2. The variance of X is equal to the variance of the underlying population divided by the
sample size.
s2
In symbols: sX2 5 .
n
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 307

s2 s
The standard deviation of X is sX 5 5 .
Ån !n
3. If the underlying population is distributed normally, then the distribution of X is also
exactly normal for any sample size.
In symbols: X , N ( m, s2 /n)

Normal distribution Uniform distribution Exponential distribution

! ! 0, " 2 ! 1 a ! 0, b ! 1, ! ! 0.5, " 2 ! 0.0833 # ! 0.5, ! ! 2, " 2 ! 4
f(x ) f(x)
f(x )
0.8
1.5 5
4 0.6
1.0
3 0.4
0.5 2
0.2
1
0.0 0 0.0
"3 "2 "1 0 1 2 3x 0.00 0.25 0.50 0.75 1.00 x 0 1 2 3 4 x

n ! 2 f(x ) f(x) f(x )

0.8
1.5 5
4 0.6
1.0
3 0.4
2
0.5 0.2
1
0.0 0 0.0
"3 "2 "1 0 1 2 3x 0.00 0.25 0.50 0.75 1.00 x 0 1 2 3 4 x

n ! 5 f(x) f(x) f(x)

0.8
1.5 5
4 0.6
1.0
3 0.4
0.5 2
0.2
1
0.0 0 0.0
"3 "2 "1 0 1 2 3x 0.00 0.25 0.50 0.75 1.00 x 0 1 2 3 4 x

n ! 10 f(x ) f(x)
f(x) 0.8
1.5 5
4 0.6
1.0
3 0.4
0.5 2
0.2
1
0.0 0 0.0
"3 "2 "1 0 1 2 3x 0.00 0.25 0.50 0.75 1.00 x 0 1 2 3 4 x

n ! 20 f(x ) f(x) f(x)

0.8
1.5 5
4 0.6
1.0
3 0.4
2
0.5 0.2
1
0.0 0 0.0
"3 "2 "1 0 1 2 3x 0.00 0.25 0.50 0.75 1.00 x 0 1 2 3 4 x
(a) (b) (c)

Figure 7.5 Three different original, or underlying, populations, and approximations to the
distribution of the sample mean for various sample sizes n.
308 CHA P T ER 7 Sampling Distributions

The mean of the sampling distribution of X is the same as the mean of the underlying
distribution that we want to estimate, so the sample mean is an unbiased estimator of the
population mean m.
Even if the underlying population is not normal, the previous examples suggest the
distribution of X becomes more normal as the sample size increases. This amazing result
is the most important idea in all of statistics, the central limit theorem.

Central Limit Theorem (CLT)

Let X be the mean of observations in a random sample of size n drawn from a popula-
tion with mean m and finite variance s2 . As the sample size n increases, the sampling
distribution of
X2m
Z5
s / !n
approaches the standard normal distribution. In practice, or informally, this means
the sampling distribution of X will increasingly approximate a normal distribution,
with mean m and variance s2 /n, regardless of the shape of the underlying population
The symbol , means “is
d distribution.
In symbols: X , N ( m, s2 /n ) .
d
approximately distributed as.”

STATISTICAL APPLET The central limit theorem is really a remarkable result. No matter what the shape of the
THE CENTRAL LIMIT underlying population (skewed, bimodal, lots of variability, etc.), the CLT says the distri-
THEOREM bution of X is approximately normal, as long as n is large enough!

A CLOSER L OK
1. A better name for this result might be the normal convergence theorem. The distri-
bution of X converges, or gets closer and closer, to a normal distribution.
2. If the original population is normally distributed, then the distribution of X is normal,
no matter how large or small the sample size (n) is.
3. If the original population is not normal, the central limit theorem says the distribution
of X approaches a normal distribution as n increases, and the approximation improves
as n increases: The approximation gets better and better as the sample size n gets big-
ger and bigger.
There is no magical threshold value for n. However, in most cases, if n $ 30, then the
approximation is pretty good. Even for severely skewed populations, as long as n is at
least 30, the approximation is reasonable.
There is a tendency to think that if n 5 29 the distribution of X will not be approxi-
mately normal, but if n 5 31 it will be. This is simply not true. The CLT should be
interpreted to mean that the distribution of X approaches a normal distribution, looks
more and more like a normal distribution, as n increases. In some cases, for n as little
as 5 the approximation will be excellent. In others, n might have to be at least 26 before
the approximation is good.
4. To compute a probability involving the sample mean, we treat X just like any other
normal random variable: Standardize and use cumulative probability where appropri-
ate. We use the same method even if X is only approximately normal.
5. The expression for the variance of X mathematically confirms our observations regard-
ing the variability of the sampling distribution. As n increases, the distribution of the
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 309

Solution Trail 7.7 sample mean becomes more compact. sX2 5 s2 /n and s2 is constant. This fraction
becomes smaller as n (the denominator) increases.
K EY WO R D S
6. A more general version of the CLT includes a statement about the sum of independent
■ Normally distributed observations, T. If n is sufficiently large, the distribution of T approaches a normal
■ Mean 50, standard deviation distribution with mean nm and variance ns2.
4, 25 trips
In symbols: T , N ( nm, ns2 ) .
d

■ Sample mean

TR A NSL ATI O N
■ Normal An even more general version of the CLT concludes, essentially, that any statistic that
■
2
m 5 50, s 5 16, n 5 25 is a sum or a mean tends toward a normal distribution as n increases. This is useful in
■ X many inference problems because the appropriate statistic is often a sum or a mean. It is
easy to compute probabilities associated with a normal statistic (random variable). There-
CON C EPT S fore, the likelihood, or tail probability, associated with an observed value of the statistic is
■ Properties of X a straightforward calculation.
V I SI O N
In addition, the central limit theorem helps to explain why so many real-world distri-
butions are approximately normal. Almost any statistic can be decomposed into a sum of
The underlying distribution is
normal, so the distribution of X
other variables. For example, the height of a tomato plant after six weeks might be
is exactly normal. Find the mean, directly related to, or is the sum of, the effects of a large number of independent variables
variance, and standard deviation including the amount of water, fertilizer, and sunlight it has received. Therefore, the
of X , translate each question into distribution of the height should be approximately normal. This single theorem explains
a probability statement, empirical evidence that suggests almost every measurement distribution is approxi-
standardize, and use cumulative mately normal.
probability where appropriate. The following examples illustrate the properties of X and the central limit theorem.

Standardization illustrated: Example 7.7 Green Line Time

P(X ! 48)
The Massachusetts Bay Transportation Authority (MBTA) Green Line from Eliot to
f (x )
Lechmere in Boston has trolleys leaving regularly throughout the day beginning at 5:00 A.M.
Although the length of the trip and the number of stops are constant, the time taken for each
trip varies due to weather, traffic, and time of day. According to the MBTA, the mean time
taken for the trip is 50 minutes. Suppose the travel time is normally distributed with stan-
dard deviation s 5 4 minutes. A random sample of 25 trips is obtained, and the time for
each is recorded.
48 50 x
a. Find the probability that the sample mean time will be less than 48 minutes.
P(Z " !2.5) b. Find the probability that the sample mean will be within one minute of the population
f(z) mean (50 minutes).

SOLUTION
a. The underlying distribution is normal with m 5 50 and s2 5 16. The sample size is
n 5 25. The sample mean is (exactly) normally distributed:
!2.5 0 z
X , N ( m, s2 /n ) 5 N ( 50, 16/25 ) ; sX 5 !16/25 5 4/5 5 0.80.
To solve part (a), begin with a probability statement involving the random variable X .
We need the probability that X will be less than 48 minutes.
X 2 50 48 2 50
P (X , 48 ) 5 Pa , b Standardize.
0.80 0.80
5 P ( Z , 2 2.5 ) Equation 6.8; simplify.

5 0.0062 Use Table III in the Appendix.

The probability that the sample mean time for the 25 trips will be less than 48 minutes is
Figure 7.6 P(X , 48). 0.0062. Figure 7.6 shows a technology solution.
310 CHA P T E R 7 Sampling Distributions

Remember, one point in a b. Within 1 minute of the population mean means 49 # X # 51. Write the probability
continuous world (Z) contributes statement, and standardize again.
no probability:
P ( 49 # X # 51 )
P(Z , 2 1.25) 5 P(Z # 2 1.25).
49 2 50 X 2 50 51 2 50
5 Pa # # b Standardize.
0.8 0.8 0.8
5 P ( 2 1.25 # Z # 1.25 ) Equation 6.8; simplify.

5 P ( Z # 1.25 ) 2 P ( Z , 2 1.25 ) Use cumulative probability.

5 0.8944 2 0.1056 Use Table III in the Appendix.

5 0.7888
Standardization illustrated:
P(49 ! X " 51)
f (x) P(!1.25 " Z " 1.25)
f (z)

49 50 51 x !1.25 0 1.25 z

The probability that the sample mean time for the 25 trips will be between 49 and 51
Figure 7.7 P(49 # X # 51). minutes is 0.7888. Figure 7.7 shows a technology solution.

TRY IT NOW GO TO EXERCISE 7.39

Solution Trail 7.8 Example 7.8 Milk Deliveries

KEYW OR DS In upstate New York, milk tanker trucks follow a daily routine, stopping at the same dairy
■ Mean 7750, standard devia- farms every day. Farm output, however, varies because of weather, time of year, number
tion 150, 36 trucks of cows, and other factors. From years of recorded data, the mean amount of milk picked
■ Sample mean up by a truck for processing is 7750 liters, with a standard deviation of 150 liters. Suppose
36 trucks are randomly selected.
T RA NSL ATI O N
a. Find the probability that the sample mean amount of milk picked up by the 36 trucks
■ m 5 7750, s 5 150, n 5 36
is more than 7800 liters.
■ X
b. Find a value m such that the probability that the sample mean is less than m is 0.1.
CO NCE P TS
■ Central limit theorem SOLUTION
VIS I ON a. The underlying distribution has m 5 7750 and s 5 150. The sample size is n 5 36.
The exact distribution of the By the central limit theorem, the distribution of X is approximately normal:
underlying population is not
X , N ( m, s2 /n ) 5 N ( 7750, 1502 /36 ) 5 N ( 7750, 625 )
d

known. However, the sample size

is large: n 5 36 ($30). The standard deviation of X is
Therefore, the CLT can be applied.
sX 5 "s2 /n 5 !625 5 25 ( or sX 5 s / !n 5 150/6 5 25 )
To solve part (a), start with a probability statement involving X . We need the probability
that the sample mean will be more than 7800 liters.
X 2 7750 7800 2 7750
P (X . 7800 ) 5 Pa . b Standardize.
25 25
5 P(Z . 2) Equation 6.8; simplify.

5 1 2 P(Z # 2) Use cumulative probability.

5 1 2 0.9772 Use Table III in the Appendix.

5 0.0228
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 311

The probability that the sample mean amount of milk picked up by the 36 trucks will
be greater than 7800 is 0.0228.
Standardization illustrated:
P(X ! 7800) P(Z ! 2)
f (x) f (z)

7750 7800 x 0 2 z

b. To solve part (b), translate the question into a probability statement. Convert the
expression into cumulative probability involving Z. Because the probability is
already given, this is an inverse cumulative probability problem. Work backward in
Table III.
X 2 7750 m 2 7750
P (X , m ) 5 Pa , b Standardize.
25 25
m 2 7750
5 PaZ , b 5 0.1 Equation 6.8.
25

There is no further simplification within the probability statement. However, the result-
ing probability statement involves Z and is cumulative probability. Find a value in the
body of Table III, as close to 0.1 as possible. Set the corresponding z value equal to
m 2 7750
a b , and solve for m.
25

m 2 7750
5 21.28 Table III in the Appendix.
25
m 2 7750 5 232 Multiply both sides by 25.

m 5 7718 Add 7750 to both sides.

The probability that the sample mean will be less than m 5 7718 is (approximately) 0.1.
Standardization illustrated:

P(X ! m) " 0.1 P(Z " m!7750) # 0.1

25
f (x) f (z)

m 7750 x !1.28 0 z

Figure 7.8 shows a technology solution.

Figure 7.8 JMP

calculations for P(X , 7800)
and P(X , m) 5 0.1.

TRY IT NOW GO TO EXERCISE 7.56

The last example in this section involves an inference using the random variable X .
312 CHA PT E R 7 Sampling Distributions

Example 7.9 Supermarket Checkout Lines

Kroger Supermarkets are now using infrared cameras similar to those used by the military
and law enforcement personnel to determine customer waiting time in checkout lines.10
The cameras detect body heat, and special software is used to control the number of
checkout lanes that need to be open. As a result of this new system, the company claims
that mean customer waiting time in a checkout line is 26 seconds. Suppose s 5 9 sec-
onds, a random sample of 32 Kroger customers is selected, and the sample mean waiting
time is x 5 28.0 seconds. Is there any evidence to suggest that the population mean cus-
tomer waiting time in a checkout line is more than 26 seconds?
Cheryl A. Guerrero/Los Angeles Times/MCT

SOLUTION
Solution Trail 7.9 STEP 1 Assume the underlying distribution has m 5 26 and s 5 9. We do not know the
shape of the underlying distribution of customer waiting times, but the sample
KEYW OR DS size is large: n 5 32 ($30). By the central limit theorem, the distribution of X
■ Is there any evidence? is approximately normal.
T RA NSL ATI O N STEP 2 Claim: m 5 26 S X , N ( 26, 92 / 32 ) ,
d
sX 5 9/ !32.
■ Statistical inference
Experiment: x 5 28.0.
CO NCE P TS Likelihood: We are looking for evidence that the mean customer waiting time is
■ Central limit theorem greater than 26 seconds, so find the right-tail probability (to be conservative).
VIS I ON X 2 15 28.0 2 26.0
This inference question concerns P (X $ 28.0 ) 5 Pa $ b Standardize.
5/ !40 9/ !32
a population mean. Use the CLT
and the distribution of X to 5 P ( Z $ 1.26 ) Equation 6.8; simplify.

determine the likelihood of an 5 1 2 P ( Z , 1.26 ) Use cumulative probability.

observed sample mean of 5 1 2 0.8962 Table III in the Appendix.
x 5 28.0 or greater.
5 0.1038
Conclusion: This probability is large (.0.05 ) , so there is no evidence to sug-
gest the mean customer waiting time in a checkout line is greater than 26. Even
if the mean waiting time is 26, an observation of x 5 28.0 is not very unlikely,
and so we cannot doubt that the mean is really 26. Figure 7.9 illustrates the
distribution of X , and the right-tail probability; Figure 7.10 shows a technology
solution.

f (x )

22 24 26 28 30 x

Figure 7.9 The distribution of X . The shaded region Figure 7.10 Use calculator
represents the right-tail probability P(X $ 28). variables to make the input easier.

TRY IT NOW GO TO EXERCISE 7.42

 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 313

Technology Corner
Procedure: Solve probability questions involving the sample mean X .
Reconsider: Example 7.7, solution, and interpretations.
Note: In these probability questions X is either normally distributed or approximately normally distributed. In either case,
the technology procedures are similar to those described in Section 6.2.

CrunchIt!
CrunchIt! has a built-in function to compute probabilities associated with a normal random variable. Select Distribution
Calculator; Normal. To find cumulative probability or right-tail probability, select the Probability tab. Choose an appropri-
ate inequality symbol and enter a value for the endpoint. To solve an inverse cumulative probability problem, select the
Quantile tab and enter a cumulative probability.
1. Select Distribution Calculator; Normal. Enter the mean, 50, and the standard deviation, 0.8. Under the Probability tab,
select # and enter 48 for the endpoint. Click the Calculate Tab. See Figure 7.11.
2. To find the probability that X takes on a value in an interval, use cumulative probability.
P ( 49 # X # 51 ) 5 P (X # 51 ) 2 P (X # 49 ) .

Figure 7.11 CrunchIt! returns a graph of the

corresponding probability density function. The shaded
area corresponds to P(X , 48).

TI-84 Plus C
The built-in function normalcdf is used to find (calculate) cumulative probability: the probability that X takes on a value
between a and b. This function takes four arguments: a (lower), b (upper), m, and s. The built-in function invNorm takes
three arguments: p (area), m, and s. This function returns a value x such that P ( X # x ) 5 p. The default values for m and
s are 0 and 1, respectively.
1. Select DISTR ; DISTR; normalcdf. Enter the left endpoint (lower), 21E99 (negative calculator infinity), the
right endpoint (upper), 48, the mean ( m ) , 50, and the standard deviation ( s ) , 0.8. Highlight Paste and tap
ENTER . Refer to Figure 7.6.
2. Select DISTR ; DISTR; normalcdf. Enter the left endpoint (lower), 49, the right endpoint (upper), 51, the
mean ( m ) , 50, and the standard deviation ( s ) , 0.8. Highlight Paste and tap ENTER . Refer to Figure 7.7.
314 CHA P TER 7 Sampling Distributions

Minitab
There are several built-in functions to compute cumulative probability, tail probability, and inverse cumulative probability.
These functions may be accessed through a graphical input window or by using the command language.
1. In a session window, use the function CDF to find P (X , 48 ) . See Figure 7.12.
2. Select Graph; Probability Distribution Plot; View Probability. In the Distribution menu, select Normal. Enter the Mean
and Standard deviation. Under the Shaded Area tab, choose X Value and Middle. Enter the X value 1 (49) and X value 2
(51). Click OK. Minitab displays a distribution plot with the shaded area corresponding to probability. See Figure 7.13.

Figure 7.12 P(X , 48) using the Figure 7.13 P(49 # X # 51) using
command language. Probability Distribution Plot.

Excel
There are built-in functions to compute cumulative probability associated with a standard normal random variable
([Link]) and a normal random variable with arbitrary mean and standard deviation ([Link]). The functions
[Link] and [Link] are the corresponding inverse cumulative probability functions.
1. Use the function [Link] to find P (X , 48 ) . See Figure 7.14.
2. Use the function [Link] to find P (X # 49 ) and P (X # 51 ) . Compute the difference to find P ( 49 # X # 51 ) .
See Figure 7.15.

Figure 7.14 P(X , 48): Figure 7.15 Calculations for

cumulative probability. P(49 # X # 51).

SECTION 7.2 EXERCISES

Concept Check 7.28 True/False The distribution of X is never exactly normal.

7.26 True/False The sample mean varies from sample to 7.29 True/False If the underlying population is not normal,
sample. the central limit theorem says the distribution of X approaches
a normal distribution as n increases.
7.27 Fill in the Blank Let X be the mean of observations in
7.30 True/False As n increases, the distribution of the sum
a random sample of size n from a population with mean m and
of the observations approaches a normal distribution.
variance s2.
a. The mean of X is . 7.31 True/False As n increases, the variance of X also
b. The variance of X is . increases.
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 315

Note: To find the distribution of X , or any random variable, 7.37 The figure below shows the graphs of the probability den-
you need to describe the distribution, by name if possible, and sity functions for the random variable X, the random variable X
provide the values of the parameters that characterize the for n 5 5, and the random variable X for n 5 15.
distribution. For example, X is normally distributed with mean
m 5 27 and variance s2X 5 0.35, or X , N ( 27, 0.35 ) , com-
pletely describes the distribution of X .

Practice
7.32 Consider a normally distributed population with mean
m 5 10 and standard deviation s 5 2.5. Suppose a random sam-
ple of size n is selected from this population. Find the distribution
of X and the indicated probability in each of the following cases.
a. n 5 7, P (X # 9 ) .
b. n 5 12, P (X . 11.5 ) .
c. n 5 15, P ( 9.5 # X # 10.5 ) .
d. n 5 25, P (X $ 10.25 ) . 30 40 50 60 70
e. n 5 100, P (X # 9.8 c X $ 10.2 ) .
Identify each probability density function.
7.33 Suppose X is a normal random variable with mean
m 5 17.5 and standard deviation s 5 6 . A random sample 7.38. The figure below shows graphs of the probability density
of size n 5 24 is selected from this population. function for the random variable X and the approximate density
a. Find the distribution of X . functions for the random variable X for n 5 5 and the random
b. Carefully sketch a graph of the probability density variable X for n 5 15.
functions for X and X on the same coordinate axes.
c. Find P ( X # 14 ) and P (X , 14 ) .
d. Find P ( 15 , X , 19 ) and P ( 15 , X , 19 ) .

7.34 Suppose X is a random variable with mean m 5 50 and

variance s2 5 49. A random sample of size n 5 38 is selected
from this population.
a. Find the approximate distribution of X . Why is the
central limit theorem necessary here?
b. Find P (X , 49 ) .
c. Find P (X $ 52 ) .
d. Find P ( 49.5 # X # 51.5 ) .
e. Find a value c such that P (X . c ) 5 0.15.
0 2 4 6 8 10 12 14 16 18
7.35 Suppose X is a random variable with mean m 5 1000 and
standard deviation s 5 100. A random sample of size n 5 36 Identify each probability density function.
is selected from this population.
a. Find the approximate distribution of X . Carefully sketch Applications
a graph of the probability density function.
b. Find P (X . 975 ) . 7.39 Public Health and Nutrition One measure of general
c. Find P (X # 1030 ) . health is your body mass index (BMI). Adults with a BMI
d. Find P ( mX 2 sX # X # mX 1 sX ) . between 18.5 and 24.9 are generally considered to have an ideal
e. Find a symmetric interval about the mean mX such that body weight for their height. Despite a reputation as a food-
P ( mX 2 c # X # mX 1 c ) 5 0.95. loving nation, the mean BMI for adults in France is 24.5.11
Suppose the BMI for adults in France is normally distributed
7.36 Suppose X is a random variable with mean m 5 30 and with standard deviation 1.3.
standard deviation s 5 50. A random sample of size n 5 40 is a. Suppose one adult from France is selected at random.
selected from this population. What is the probability that the person’s BMI is more
a. Find the approximate distribution of X . Carefully sketch than 25.5?
a graph of the probability density function. b. Suppose 10 adults from France are selected at random.
b. Find P (X $ 38 ) . What is the probability that the sample mean BMI is
c. Find P ( 20 # X # 40 ) . greater than 25.5?
d. Find P (X , 15 ) . c. What is the probability that the sample mean BMI is less
e. Find a value of c such that P (X # c ) 5 0.001. than 24?
316 CHAP T E R 7 Sampling Distributions

d. A certain health spa claims its clients have BMI between b. How can this probability in part (a) be so small when 11.9
23.5 and 24. Suppose eight health spa clients are selected is so close to the population mean m 5 12?
at random. What is the probability that the sample mean c. From your answer to part (a), is there any evidence to
will be in this interval? suggest that the mean weight of bags of potato chips is
less than 12 ounces?
7.40 Public Policy and Political Science In certain
hurricane-prone areas of the United States, concrete columns 7.44 Psychology and Human Behavior In attempting to
used in construction must meet specific building codes. The flee from police, criminals often barricade themselves inside a
minimum diameter for a cylindrical column is 8 inches. building and create a police standoff. The criminal is usually
Suppose the mean diameter for all columns is 8.25 inches with armed with a dangerous weapon and may hold hostages. Sup-
standard deviation 0.1 inch. A building inspector randomly pose the mean length of a police standoff, ending with some
selects 35 columns and measures the diameter of each. sort of resolution, is 6.5 hours with standard deviation 4 hours.
a. Find the approximate distribution of X . Carefully sketch A random sample of 35 police standoffs is selected.
a graph of the probability density function. a. Find the distribution of the sample mean.
b. What is the probability that the sample mean diameter for b. What is the probability that the sample mean for the 35
the 35 columns will be greater than 8 inches? police standoffs will be greater than 7 hours?
c. What is the probability that the sample mean diameter for c. A new psychological technique was used in negotiations
the 35 columns will be between 8.2 and 8.4 inches? with the criminals involved in the 35 police standoffs.
d. Suppose the standard deviation is 0.15 inch. Answer parts Suppose the sample mean for the 35 police standoffs is
(a), (b), and (c) using this value of s. x 5 5.1 hours. Is there any evidence to suggest that the
mean police standoff time is lower when this new
7.41 Manufacturing and Prodcut Development A large
technique is used?
part of the luggage market is made up of overnight bags. These
bags vary by weight, exterior appearance, material, and size. 7.45 Biology and Environmental Science During monsoon
Suppose the volume of overnight bags is normally distributed season in India, the mean amount of rainfall is 89 centimeters.12
with mean m 5 1750 cubic inches and standard deviation Suppose the standard deviation is 5 centimeters and 30 monsoon
s 5 250 cubic inches. A random sample of 15 overnight bags seasons are selected at random.
is selected, and the volume of each is found. a. What is the probability that the sample mean rainfall is
a. Find the distribution of X . less than 87 centimeters?
b. What is the probability that the sample mean volume is b. What is the probability that the sample mean rainfall is
more than 1800 cubic inches? greater than 91.5 centimeters?
c. What is the probability that the sample mean volume is c. Find a symmetric interval about the mean, 89, such that the
within 100 cubic inches of 1750? probability that the sample mean lies in this interval is 0.90.
d. Find a symmetric interval about 1750 such that 95% of all
7.46 Biology and Environmental Science Carbon dioxide
values of the sample mean volume lie in this interval.
(CO2) is one of the primary gases contributing to the green-
7.42 Biology and Environmental Science The U.S. house effect and global warming. The mean amount of CO2 in
Environmental Protection Agency (EPA) is concerned about the atmosphere for March 2013 was 397.34 parts per million
pollution caused by factories that burn sulfur-rich fuel. To (ppm).13 Suppose 40 atmospheric samples are selected at ran-
decrease the impact on the environment, factory chimneys must dom in May 2013 and the standard deviation for CO2 in the
be high enough to allow pollutants to dissipate over a larger atmosphere is s 5 20 ppm.
area. Assume the mean height of chimneys in these factories is a. Find the probability that the sample mean CO2 level is
100 meters (an EPA-acceptable height) with standard deviation less than 393 ppm.
12 meters. A random sample of 40 chimney heights is obtained. b. Find the probability that the sample mean CO2 level is
a. What is the probability that the sample mean height for between 395 and 403 ppm.
the 40 chimneys is greater than 102 meters? c. Suppose the sample mean CO2 level is 400. Is there any
b. What is the probability that the sample mean height is evidence to suggest that the population mean CO2 level
between 101 and 103 meters? has increased? Justify your answer.
c. Suppose the sample mean is 98.5 meters. Is there any
7.47 Public Health and Nutrition Typhoid fever is more
evidence to suggest that the true mean height for
common in developing countries with poor sanitation and
chimneys is less than 100 meters? Justify your answer.
greater incidence of food contamination. The mean duration of
7.43 Manufacturing and Product Development The man- this illness is approximately 4 weeks.14 Periodically, a team of
ager at an HEB grocery store in Corpus Christi, Texas, suspects a doctors randomly selects 10 patients with typhoid fever and
supplier is systematically underfilling 12-ounce bags of potato carefully measures the duration of the illness. If the mean
chips. To check the manufacturer’s claim, a random sample of 100 duration is greater than 4.5 weeks, then a health alert is issued.
bags of potato chips is obtained, and each bag is carefully weighed. Suppose the duration of this illness is normally distributed with
The sample mean is 11.9 ounces. Assume s 5 0.3 ounces. standard deviation one week.
a. Find the probability that the sample mean is 11.9 ounces a. Find the probability that the sample mean duration is less
or less. than 4.5 weeks.
 7.2 The Sampling Distribution of the Sample Mean and the Central Limit Theorem 317

b. Find the probability that a health alert is issued. Extended Applications

c. Suppose the true mean duration has increased to
4.3 weeks. What is the probability that a health alert will 7.51 Public Health and Nutrition Recent research studies
be issued? suggest that lycopene, found in tomatoes, may reduce the risk
of certain cancers. According to the USDA/NCC Carotenoid
7.48 Biology and Environmental Science There are many Database for U.S. Foods, the mean amount of lycopene in a
regulations for catching lobsters off the coast of New England, medium-sized fresh tomato is 3.7 milligrams (mg). Suppose the
including required permits, allowable gear, and size prohibi- amount of lycopene is a normal random variable with standard
tions. The Massachusetts Division of Marine Fisheries requires deviation 1 mg. Consider the sample mean based on 5 observa-
a minimum carapace length measured from a rear eye socket to tions, X 5, and the sample mean based on 20 observations, X 20.
the center line of the body shell. Any lobster measuring less a. Find the distributions for X 5 and X 20.
than 314 inches must be returned to the ocean. For all lobster b. Carefully sketch the probability density functions for X,
caught, suppose the carapace length is normally distributed X 5, and X 20 on the same coordinate axes.
with mean 418 inches and standard deviation 1 inch. A random c. Find P ( X , 3 ) , P (X 5 , 3 ) , and P (X 20 , 3 ) . Explain why
sample of 15 lobsters is obtained. these values are different.
a. Find the distribution of the sample mean carapace length. d. Find P ( 3.6 # X # 3.8 ) , P ( 3.6 # X 5 # 3.8 ) , and
b. What is the probability that the sample mean carapace P ( 3.6 # X 20 # 3.8 ) . Explain why these values are
length is more than 412 inches? different.
c. What is the probability that the sample mean carapace 7.52 Sports and Leisure A health club recently added tan-
length is between 378 and 418 inches? ning booths in addition to a weight room, pool, and racquetball
d. If the sample mean carapace length is less than 312 inches, courts. For insurance purposes, an employee maintains careful
a lobsterman will look for other places to set his traps. records of the length of time each member spends in a tanning
What is the probability that a lobsterman will be looking booth. The mean length of time spent tanning is 15 minutes
for a different location? with standard deviation 2 minutes. Consider a random sample
of 35 members who use a tanning booth.
7.49 Sports and Leisure One measure of an athlete’s abil-
a. Find the distribution of the total time spent tanning by the
ity is the height of his or her vertical leap. Many professional
35 members, T.
basketball players are known for their remarkable vertical
b. Find the probability that the total time spent tanning is
leaps, which lead to amazing dunks. DJ Stephens is the cur-
between 8 and 9 hours.
rent vertical-leap record holder at 46 inches. However, the
c. If the total time spent tanning is more than 9.2 hours, an
mean vertical leap of all NBA players is 28 inches.15 Suppose
employee must work overtime. What is the probability
the standard deviation is 7 inches and 36 NBA players are
that the employee works overtime on a day when 35
selected at random.
members tan?
a. What is the probability that the mean vertical leap for the
d. Find a value t such that P ( T $ t ) 5 0.01.
36 players will be less than 26 inches?
b. What is the probability that the mean vertical leap for the 7.53 Sports and Leisure The Tonight Show with Johnny
36 players will be between 27.5 and 28.5 inches? Carson lasted for 30 years and featured movie stars, comedians,
c. A high-priced athletic trainer has been hired to work with animal acts, and Carnac the Magnificent. The first part of the
a group of NBA players to improve their hip flexibility, show was reserved for Johnny’s monologue. The mean length of
which should improve their vertical leap. After one month a monologue was 12 minutes with standard deviation 45 sec-
of training, the mean vertical leap for 50 of these players onds. Suppose 40 shows are selected at random and the length
selected at random was 29.75 inches. Is there any of each monologue is recorded.
evidence to suggest this flexibility program has increased a. Find the distribution of the total monologue time for the
the mean vertical leap (from m 5 28 inches)? 40 shows, T.
b. Find P ( T . 500 ) .
7.50 Physical Sciences The ozone hole is a region in the
c. Find P ( 470 # T # 490 ) .
Southern Hemisphere that has increased since 1979. The mean
d. After his retirement, Johnny Carson’s company started to
ozone hole area (in million km2) in 2012 was 17.916 with stan-
sell DVDs of his monologues. Suppose one DVD holds
dard deviation 5.6. Suppose 33 days are selected at random and
7.9 hours of recording. What is the probability that the
the ozone hole is measured each day.
company will be able to fit 40 randomly selected
a. What is the probability that the sample mean ozone hole
monologues onto one DVD?
area is less than 15.5 million km2?
b. What is the probability that the sample mean ozone hole 7.54 Manufacturing and Product Development A (destruc-
area is between 16 and 18 million km2? tive) tensile test is standard procedure for testing a cross-wire
c. Some researchers suggest that increased use of chemicals weld. Suppose a certain weld is designed to withstand a force with
has caused the ozone hole area to increase. Suppose the mean 0.8 kilonewtons (kN). To check the quality of welds, a man-
sample mean for the 33 days is 20.7 million km2. Is there ufacturer randomly selects 25 welds (every hour) and performs a
any evidence to suggest that the mean ozone hole area has tensile test on each weld. If the mean force required (for the 25
increased? Justify your answer. welds) is between 0.75 and 0.85 kN, the process is allowed to
318 CHA P T E R 7 Sampling Distributions

continue. Otherwise, it is shut down. Suppose the force distribu- Standard clip-on weights for steel rims on automobiles (used
tion is normal with standard deviation 0.1 kN. when tires are balanced) are available in 14 -ounce to 6-ounce
a. If the true mean is 0.8 kN, what is the probability that the sizes. Suppose an automobile garage receives a large shipment
process will be shut down? This probability represents the of 2-ounce weights. A garage mechanic will select a random
chance of making one kind of error. sample of 30 weights and weigh each one on a precise scale.
b. If the true mean is 0.82 kN, what is the probability that If the sample mean is within 0.05 ounce of the printed weight
the process will be allowed to continue? What if (of 2 ounces), then the shipment is accepted. Otherwise, the
m 5 0.84 kN? These probabilities represent the chance of entire shipment is rejected and returned to the manufacturer.
making a different kind of error. Suppose the population standard deviation is 0.13 ounce.
c. Suppose the process is allowed to continue if the mean a. Find the probability of accepting the entire shipment if
force required is between 0.76 and 0.84 kN. Answer parts the true population mean weight is 1.86, 1.88, 1.90, 1.92,
(a) and (b) with this new interval. 1.94, 1.96, 1.98, 2.00, 2.02, 2.04, 2.06, 2.08, 2.10, 2.12,
of 2.14 ounces.
7.55 Manufacturing and Product Development Shetland
b. Plot the probability of accepting the entire shipment versus
wool is considered some of the finest in the world because it is
the true population mean weight. The resulting graph is the
soft, durable, and easy to spin. The mean fleece weight from a
OC curve for the given acceptance sampling plan.
typical sheep is 3.25 pounds with a standard deviation of 0.4
pound. Suppose a farmer has 100 sheep ready to be sheared. 7.58 Normal Approximation to the Binomial
a. Find the distribution for the total weight of fleece from Distribution Consider a binomial experiment with n
the 100 sheep, T. trials and probability of success p. If we assign a 0 to each fail-
b. What is the probability that the total fleece weight is less ure and a 1 to each success, then the binomial random variable
than 323 pounds? can be defined as a sum. By the central limit theorem, as n
c. What is the probability that the total fleece weight is increases the distribution of X (the sum) approaches a normal
between 330 and 340 pounds? distribution with mean np and variance np ( 1 2 p ) . Suppose X
d. Find a value t such that P ( T $ t ) 5 0.15. is a binomial random variable with n 5 30 and probability of
success p 5 0.5.
7.56 Sports and Leisure A wingsuit is a specially designed
a. Construct a probability histogram for the binomial
jumpsuit with added surface area to provide more lift. Extreme-
random variable X. Find P ( 12 # X # 16 ) using the
sports enthusiasts wear this suit and jump from planes or a fixed
binomial distribution.
object (BASE jumping, where BASE stands for Building,
b. Find the approximate normal distribution for X. Find
Antenna, Span, Earth). One of the most popular BASE environ-
P ( 12 # X # 16 ) using the normal distribution for X.
ments is Troll Wall in Norway. Suppose the mean flight time for
c. Compare the probabilities found in parts (a) and (b).
a wingsuit jump at Troll Wall is 22.5 seconds with standard
d. Find P ( 11.5 # X # 16.5 ) using the normal distribution
deviation 2.3 seconds. Thirty-two jumpers are selected at ran-
for X. Compare this answer with the probability in part (a).
dom to test a new wingsuit at Troll Wall.
Why do you think this is a much better approximation?
a. What is the probability that the sample mean flight time is
less than 22 seconds? 7.59 Manufacturing and Product Development A hard-
b. Suppose the sample mean flight time for these jumpers is ware store has 20 interior plantation shutter sets in stock. The
23.15 seconds. Is there any evidence to suggest that this width (in inches) of each set is given in the table below.
new wingsuit increases the mean flight time? SHUTTER
c. Find a value w such that P (X # w ) 5 0.005.
30 30 28 34 36 28 34 36 24 35
Challenge 28 30 32 30 44 34 22 32 20 30
7.57 Fuel Consumption and Cars Companies receiving large
shipments of raw materials of any product often use a specific a. Find the (population) mean width of the 20 shutters.
plan for accepting the entire shipment. An acceptance sampling b. Use technology to generate at least 100 random samples
plan usually includes the sample size for close inspection, the of size five (without replacement). Find the sample mean
acceptance criterion, and the rejection criterion. For a given plan, width for each sample.
the operating characteristic (OC) curve shows the probability of c. Construct a histogram of the sample means found in part
accepting the entire lot as a function of the actual quality level. (a). Describe the distribution.

7.3 The Distribution of the Sample Proportion

In Section 7.2 the sampling distribution of the sample mean was introduced. The central
limit theorem helps if the underlying population is not normal. Knowing the distribution
of X , we can use the sample mean to make an inference about the population mean m.
Similarly, we are often interested in drawing a conclusion about the population proportion
 7.3 The Distribution of the Sample Proportion 319

p (the probability of a success). For example, a politician might want to estimate the pro-
portion of voters in a district in favor of a certain highway bill, or a quality-control super-
visor might need to estimate the true proportion of defective parts in a large shipment.
=
It seems reasonable to use a value of the sample proportion, p, to make an inference
concerning the population proportion p. Therefore, knowledge of the sampling distribu-
tion of the sample proportion is necessary. We need to completely characterize the vari-
ability of this statistic.
Consider a sample of n individuals or objects (or trials) and let X be the number of suc-
cesses in the sample. The sample proportion (introduced in Section 3.1) is defined to be
= X the number of successes in the sample
P5 5 (7.1)
n the sample size
The sample proportion is simply the proportion of successes in the sample, or a relative
frequency.
An approach
= similar to the one in Section 7.2 can be used to approximate the distri-
bution of P: Generate lots of sample proportions, construct a histogram, and try to char-
acterize the distribution in terms of shape, center, and variability. The sampling distribution
is summarized below.
=
The Sampling Distribution of P
=
Let P be the sample proportion of successes in a sample of size n from a population with
true proportion of success p.
=
1. The mean of P is the true population proportion.
In symbols: mP= 5 p.
= p( 1 2 p)
2. The variance of P is sP2= 5 .
n
= p( 1 2 p)
As n increases, the distribution The standard deviation of P is sP= 5 .
= Å n
of P approaches a normal =
3. If n is large and both np $ 5 and n ( 1 2 p) $ 5, then the distribution of P is approxi-
distribution. There is no threshold
mately normal.
value for n. The larger the value =
In symbols: P , N ( p, p( 1 2 p) /n ) .
d
of n and the closer p is to 0.5, the
better is the approximation.

A CLOSER L OK
=
1. It may be a little surprising to learn that P is approximately normal. However, the
sample proportion can be written as a sample mean. Assign 0 to each failure and 1 to
each success. X, the total number of successes, is a sum (of 0s and 1s). Therefore, the
sample proportion, X/n, is really a sample mean. Remember, the central limit theorem
says that the sample mean is approximately normal for n sufficiently large.
= =
2. Because the mean of P is the true population proportion, P is an unbiased estimator
for p.
3. A large sample isn’t enough for normality. The two products np and n ( 1 2 p) must
both be greater than or equal to 5. This is called the nonskewness criterion. It guaran-
=
tees that the distribution of P is approximately symmetric, that is, centered far enough
away from 0 or 1.
=
4. To compute a probability involving the sample proportion, treat P like any other normal
random variable: Standardize and use cumulative probability where appropriate.
=
The following example illustrates the properties of P and the technique for computing
probabilities associated with this random variable.
320 CHA PT ER 7 Sampling Distributions

Solution Trail 7.10 Example 7.10 Discount Double Check

KEYW OR DS In early 2013, the number of people looking to change auto insurance companies
■ 110 people reached a record low. However, for the population of those people who were looking for
■ 45%
a better deal, 45% did change companies.17 Suppose 110 people looking for a better
deal on auto insurance are selected at random and the number who actually switch
■ Sample proportion
policies is determined.
T RA NSL ATI O N =
a. Find the distribution of the sample proportion of people who switch policies, P. Care-
■ n 5 110 trials
fully sketch the probability density function for this random variable.
■ Probability of a success
(changing policies) is b. What is the probability that the sample proportion (for the 110 people selected) is
p 5 0.45 greater than 0.50?
=
■ P c. Find the probability that the sample proportion will be between 0.37 and 0.47.

CO NCE P TS
= SOLUTION
■ Distribution of P
a. For n 5 110 and p 5 0.45, check the nonskewness criteria.
VIS I ON
This question involves the np 5 ( 110 )( 0.45 ) 5 49.5 $ 5 and n ( 1 2 p) 5 ( 110 ) ( 0.55 ) 5 60.5 $ 5
= =
random variable P . Use the Both inequalities are satisfied. The distribution of P is approximately normal with
=
properties of P to find the
distribution, translate each p( 1 2 p) ( 0.45 ) ( 0.55 )
mP= 5 p 5 0.45 and s2P= 5 5 5 0.00225
question into a probabiity n 120
statement, and write one =
In symbols: P , N ( 0.45, 0.00225 ) sP= 5 !0.00225.
d

equivalent statement involving Z.

Figure 7.16 shows a graph of the probability density function and the associated
probability.

f(p)
!

P(P ! 0.50)

0.35 0.40 0.45 0.50 0.55 p

!

=
Figure 7.16 The probability density function for P . The shaded
=
area represents P(P . 0.50).

=
b. P (P . 0.50 ) Probability that the sample proportion is greater than 0.50.
=
P 2 0.45 0.50 2 0.45
5 Pa . b Standardize.
!0.00225 !0.00225
5 P ( Z . 1.05 ) Equation 6.8; simplify.

5 1 2 P ( Z # 1.05 ) Use cumulative probability.

5 1 2 0.8531 Use Table III in the Appendix.

5 0.1469
The probability that the sample proportion is greater than 0.50 is 0.1469.
 7.3 The Distribution of the Sample Proportion 321

=
Standardization illustrated c. P ( 0.37 # P # 0.47 ) Probability the sample proportion is between 0.37 and 0.47.
part (c): =
0.37 2 0.45 P 2 0.45 0.47 2 0.45
5 Pa # # b Standardize.
!0.00225 !0.00225 !0.00225
!

P(0.37 ! P ! 0.47)
f (p ) 5 P ( 21.69 # Z # 0.42 )
!

Equation 6.8; simplify.

5 P ( Z # 0.42 ) 2 P ( Z # 21.69 ) Use cumulative probability.

5 0.6628 2 0.0455 Use Table III in the Appendix.

5 0.6173
Figure 7.17 and 7.18 show technology solutions to parts (b) and (c).
p
!

0.37 0.45 0.47

P(!1.69 " Z " 0.42)

f (z)

!1.69 0 0.42 z = =
Figure 7.17 P(P . 0.50): Figure 7.18 P(0.37 # P # 0.47).
right-tail probability.
Remember, it is very important for
the sample to be selected randomly TRY IT NOW GO TO EXERCISE 7.70
from the underlying population.
Otherwise, the results are not valid. The following example illustrates an inverse cumulative probability problem and an
=
inference associated with the random variable P .
Solution Trail 7.11a
Example 7.11 Too Many Regulations
K EY WO RD S
Company executives often complain that there are too many government regulations. Sti-
■ 150 CEOs
fling rules and endless bureaucracy may limit creativity, new-product research, and cor-
■ 60% of all CEOs
porate profits. Suppose 60% of all CEOs believe there are too many government
■ Sample proportion regulations for business. A random sample of 150 CEOs is obtained, and each is asked
TRA N SLATI O N whether he or she believes there are too many government regulations.
■ n 5 150 trials a. Find a value r such that the probability the sample proportion is greater than r is
■ Probability of a success (a 0.25.
belief in too many regulations)
b. In recent years, big business has lobbied politicians to relax regulations to stimulate
is p 5 0.60
= the economy. Suppose the sample proportion for the 150 CEOs is 0.56. Is there any
■ P evidence to suggest that the true proportion of CEOs who believe there are too many
CON C EPT S regulations has decreased?
=
■ Distribution of P
SOLUTION
V I SI O N
= For n 5 150 and p 5 0.60, check the nonskewness criterion.
Knowing the distribution of P ,
we can work backward to find np 5 ( 150 ) ( 0.60 ) 5 90 $ 5 and n ( 1 2 p) 5 ( 150 ) ( 0.40 ) 5 60 $ 5
the value r. Draw a picture to =
illustrate the probability Both inequalities are satisfied. The distribution of P is approximately normal with
statement, standardize, and use p( 1 2 p) ( 0.60 ) ( 0.40 )
(inverse) cumulative probability. mP= 5 p 5 0.60 and s2P= 5 5 5 0.0016
n 150
=
In symbols: P , N ( 0.60, 0.0016 ) sP= 5 !0.0016 5 0.04.
d

Another way to ask this question:

= =
Find the third quartile of the P a. Find a value r such that P (P . r ) 5 0.25. Figure 7.19 illustrates this probability
distribution. statement.
322 CHAPT ER 7 Sampling Distributions

f(p)

!
P(P ! r ) " 0.25

0.70 p

!
0.50 0.55 0.60 r 0.65
=
Figure 7.19 The shaded area represents P(P . r).

To find the value r, write an equivalent expression involving cumulative probability.

Standardize and work backward in Table III.
=
= P 2 0.60 r 2 0.60
P (P # r ) 5 Pa # b Standardize.
0.04 0.04
r 2 0.60
5 PaZ # b 5 0.75 Equation 6.8.
0.04
There is no further simplification, but the resulting probability statement involves
Z and is cumulative probability. Find a value in the body of Table III in the Appen-
dix that is as close to 0.75 as possible. Set the corresponding z value equal to
r 2 0.60
a b , and solve for r.
Figure 7.20 A technology 0.04
solution: Use invNorm to solve
an inverse cumulative probability r 2 0.60
5 0.675 Table III in the Appendix; interpolation.
problem. 0.04
r 2 0.60 5 0.027 Multiply both sides by 0.04.

r 5 0.627 Add 0.60 to both sides.

Solution Trail 7.11b
The probability that the sample proportion is greater than 0.627 is 0.25.
KEYW ORD S
See Figure 7.20 for a technology solution.
■ Is there any evidence?
b. Follow the usual four-step inference procedure and consider a tail probability as a
T RA NSL ATI O N measure of likelihood.
■ Use available evidence to draw
Claim: 60% of all CEOs believe there are too many government regulations.
a reasonable conclusion.
=
p 5 0.60 S P , N ( 0.60, 0.04 ) .
d

CO NCE P TS
■ Inference procedure Experiment: The proportion of CEOs who believe there are too many government
=
VIS I ON regulations is p 5 0.56.
=
To decide whether the observed Likelihood: Because p 5 0.56 is to the left of the mean ( p 5 0.60 ) , consider a left-
sample proportion is reasonable, tail probability as a measure of likelihood.
we need to follow the four-step =
inference procedure. Consider = P 2 0.60 0.56 2 0.60
P (P # 0.56 ) 5 Pa # b Standardize.
the claim, the experiment, and 0.04 0.04
the likelihood of the
5 P ( Z # 21.00 ) Equation 6.8; simplify.
experimental outcome.
5 0.1587 Table III in the Appendix.
 7.3 The Distribution of the Sample Proportion 323

Standardization illustrated part (b):

!
P(P ! 0.56) P(Z " !1.00)
f(p) f(z)

!
= p

!
0.56 0 !1.00 0 z
Figure 7.21 P is approximately
normal, so use normalcdf.
In statistical inference problems, any probability less than or equal to 0.05 is consid-
Note: Even though the sample ered small.
proportion can not be less than 0,
Conclusion: This probability is larger than 0.05, so it is reasonable to observe a sam-
21E99 was used in this
ple proportion of 0.56 or smaller. There is no evidence to suggest that the claim of
technology solution because a
p 5 0.60 is wrong.
normal distribution is defined for
all real numbers. Most of the time, Figure 7.21 shows a technology solution.
when the approximation is good,
using 0 (or 1 as a right bound) will TRY IT NOW GO TO EXERCISE 7.75
not change the probability.

Technology Corner
=
Procedure: Solve probability questions involving the sample proportion P.
Reconsider: Example 7.11, solutions, and interpretations.

CrunchIt!
=
If P is approximately normal, use the Distribution Calculator as described in Section 6.2 to find probabilities and to solve
inverse cumulative probability problems.
1. Select Distribution Calculator; Normal. Enter the mean, 0.60, and the standard deviation, 0.04. Under the Quantile tab,
enter the cumulative probability and click Calculate. See Figure 7.22.
2. Select Distribution Calculator; Normal. Enter the mean, 0.60, and the standard deviation, 0.04. Under the Probability
tab, select # and enter 0.56 for the endpoint. See Figure 7.23.

=
Figure 7.22 A solution to Figure 7.23 P(P # 0.56).
=
P(P # r) 5 0.75.
324 CHA P T ER 7 Sampling Distributions

TI-84 Plus C
=
If P is approximately normal, use the built-in function normalcdf to find (calculate) cumulative probability and the
built-in function invNorm to solve inverse cumulative probability problems.
1. Select DISTR ; DISTR; invNorm. Enter the area (cumulative probability), 0.75, the mean, 0.6, and the standard
deviation, 0.04. Refer to Figure 7.20.
2. Select DISTR ; DISTR; normalcdf. Enter the left endpoint, 21E99, the right endpoint, 0.56, the mean, 0.6, and the
standard deviation, 0.04. Refer to Figure 7.21.

Minitab
As in Chapter 6 for normal random variables, use the built-in functions to find cumulative probability or inverse cumula-
tive probability. These functions may be accessed through a graphical input window: Calc; Probability Distributions;
Normal, Graph; Probability Distribution Plot; or by using the command language.
1. Use the function INVCDF to solve inverse cumulative probability problems. See Figure 7.24.
2. Use the function CDF to find cumulative probability. See Figure 7.25.

Figure 7.24 Use INVCDF to solve inverse Figure 7.25 Use CDF to find cumulative
cumulative probability problems. probability.

Excel
As in Chapter 6 for normal random variables, use the built-in function [Link] to compute cumulative probability
associated with a normal random variable, and the function [Link] to solve inverse cumulative probability problems.
=
1. Use the function [Link] to find the value r such that P (P # r ) 5 0.75. See Figure 7.26.
2. Use the function [Link] to find cumulative probability. See Figure 7.27.

Figure 7.26 Use [Link] to Figure 7.27 Use [Link] to

solve inverse cumulative find cumulative probability.
probability problems.

SECTION 7.3 EXERCISES

Concept Check 7.64 Fill in the Blank The inequalities np $ 5 and
= n ( 1 2 p) $ 5 are called the .
7.60 Fill in the Blank The mean of P is and
the variance is .
= Practice
7.61 True/False The distribution of P is approximately
normal for any sample size n. 7.65 Suppose a random sample of size n is obtained. In each
problem below, check the nonskewness criterion, and find the
7.62 True/False The sample proportion can be characterized =
distribution of the sample proportion P .
as a sample mean.
a. n 5 100, p 5 0.25. b. n 5 150, p 5 0.90.
7.63 True/False The sample proportion is an unbiased esti- c. n 5 100, p 5 0.75. d. n 5 1000, p 5 0.85.
mator for p. e. n 5 5000, p 5 0.006.
 7.3 The Distribution of the Sample Proportion 325

7.66 Suppose a random sample of size n 5 200 is obtained 7.72 Marketing and Consumer Behavior Movie trailers
from a population with probability of success p 5 0.40. Find are designed to attract large audiences. However, according to a
each of the following probabilities. recent survey, 32% of Americans believe movie trailers give
=
a. P (P # 0.37 ) . away too many of a movie’s best scenes, that is, they reveal too
= much.20 Suppose 250 Americans are selected at random and
b. P (P . 0.45 ) .
= asked if they believe movie trailers reveal too much.
c. P ( 0.38 # P # 0.42 ) .
= = a. Find the distribution of the sample proportion of
d. P (P , 0.33 ) or P (P . 0.47 ) .
Americans who believe movie trailers give away too much.
7.67 Suppose a random sample of size n 5 500 is obtained b. Find the probability that the sample proportion is more
from a population with probability of success p 5 0.50. Find than 0.35.
each of the following probabilities. c. Find the probability that the sample proportion is less
= =
a. P ( 0.52 . P) . b. P (P $ 0.47 ) . than 0.25.
= = d. Find a symmetric interval about the mean ( p 5 0.32 )
c. P ( 0.44 # P , 0.49 ) . d. P ( 0.45 # P , 0.55 ) .
such that the probability that the sample proportion is in
7.68 Suppose a random sample of size n 5 80 is obtained this interval is 0.90.
from a population with probability of success p 5 0.35.
= 7.73 Sports and Leisure Anyone who bets money in a
a. Find a value a such that P (P # a) 5 0.10.
= casino is classified as a winner if he or she wins more than he
b. Find a value b such that P (P . b) 5 0.01.
= or she loses. Casino operators in Atlantic City, New Jersey,
c. Find a value c such that P ( 0.35 2 c # P # 0.35 1 c ) 5 believe the proportion of all players who go home a winner is
0.95. 0.46. Suppose 75 Atlantic City gamblers are selected at random.
a. Find the sampling distribution of the proportion of
7.69 Suppose a random sample of size n 5 1000 is obtained
from a population with probability of success p 5 0.25. gamblers who go home winners.
= b. Find the probability that the sample proportion is less
a. Find a value a such that P (P # a) 5 0.05.
= than 0.40.
b. Find a value b such that P (P . b) 5 0.005.
= c. Find the probability that the sample proportion is more
c. Find a value c such that P ( 0.25 2 c # P # 0.25 1 c ) 5 than 0.45.
0.99. d. Find a symmetric interval about the mean ( p 5 0.46 )
=
d. Find the quartiles of the distribution of P . such that the probability that the sample proportion is in
this interval is 0.99.
7.74 Marketing and Consumer Behavior According to a
Applications
recent study, approximately 69% of consumers order food
7.70 Public Health and Nutrition Approximately 70% of online using a mobile device, for example, a smartphone or
American adults monitor a specific health indicator, for exam- tablet.21 Consumers use mobile applications to check restaurant
ple weight, diet, or exercise.18 Suppose a random sample of 250 menus, consider fast-food options, and even to order a pizza.
American adults is obtained, and all were asked if they monitor Suppose 220 consumers are randomly selected.
their health in some way. a. Find the probability that the sample proportion of
a. Find the distribution of the sample proportion of consumers who order food online using a mobile device
American adults who monitor their health. is less than 0.65.
b. Find the probability that the sample proportion is less b. Find the probability that the sample proportion of
than 0.66. consumers who order food online using a mobile device
c. Find the probability that the sample proportion is more is more than 0.72.
=
than 0.71. c. Find a value t such that P (P , t ) 5 0.95.
d. Find the probability that the sample proportion is between
7.75 Public Health and Nutrition The U.S. Centers for
0.68 and 0.78.
Disease Control announced that skin allergies have risen among
7.71 Technology and the Internet According to a report U.S. children to 12.5%.22 Serious allergies can affect a child’s
from Microsoft, 24% of PCs worldwide are not adequately pro- education, sleep, and ordinary daily activities. Suppose 320 U.S.
tected by antivirus software.19 Suppose 200 PCs from around children are randomly selected and tested for skin allergies.
the world are selected at random. a. Find the probability that the sample proportion of
a. Find the distribution of the sample proportion of PCs that children with skin allergies is less than 0.10.
are not adequately protected. b. Find the probability that the sample proportion of
b. Find the probability that the sample proportion is less children with skin allergies is between 0.08 and 0.16.
than 0.20. c. Suppose 150 U.S. children aged 5 to 9 are randomly
c. Find the probability that the sample proportion is more selected and each is tested for skin allergies. Nine children
than 0.29. are found to have skin allergies. Is there any evidence to
d. Find a value v such that the probability the sample suggest that children in this age group have a lower
proportion is less than v is 0.01. incidence rate of skin allergies? Justify your answer.
326 CH APT ER 7 Sampling Distributions

7.76 Education and Child Development Admissions 7.80 Biology and Environmental Science Ships dumping
offices at universities and colleges keep careful records of garbage and ordinary beachgoers contribute to the increasing
acceptance and yield rates. The yield rate is the percentage of amount of trash that washes onto the shore and collects in the
admitted students who decide to accept an offer of admission. oceans all over the world. In 2012 the International Coastal
Historical data help them to plan for the incoming classes and Cleanup collected more than 9 million pounds of garbage from
guide some admission decisions. The yield rate for Yale the world’s oceans. Common debris items collected included
University is 64.1%.23 Suppose 100 students who were food wrappers, bottle caps, and beverage bottles. For the United
accepted to Yale are randomly selected. States, cigarettes accounted for 28.1% of all debris items.26 A
a. Find the probability that the sample proportion of those random sample of 430 debris items from a clean-up along a
who enroll is less than 0.55. Texas beach was obtained.
b. Find the probability that the sample proportion of those a. Find the probability that the proportion of cigarette debris
who enroll is between 0.60 and 0.65. items is greater than 0.31.
c. Suppose the actual sample proportion of those who enroll b. Find the probability that the proportion of cigarette debris
is 0.70. Is there any evidence to suggest that the yield has items is between 0.25 and 0.30.
increased? Justify your answer. c. Suppose 94 of the debris items are cigarettes. Is there any
evidence to suggest that the true proportion of cigarette
7.77 Business and Management A certain philanthropic debris items is different from 0.281? Justify your answer.
organization funds one out of every ten grant proposals. Sup-
pose a random sample of 300 grant proposals is obtained.
a. Find the probability that the sample proportion of funded Extended Applications
grant proposals is less than 0.075.
7.81 Manufacturing and Product Development Low-
b. Find the probability that the sample proportion of funded
quality coffee shipped from various locations in Europe tends
grant proposals is between 0.11 and 0.15.
to contain a high proportion of defective beans (beans com-
c. If the funding rate increases, the board of directors may
posed of foreign matter; moldy, black, unripe, or fermented
become concerned about resources being depleted.
beans; or those known as stinkers). Suppose a shipper claims
Suppose the actual sample proportion of funded grant
the proportion of defective beans is 0.07. A U.S. packaging
proposals is 0.16. Is there any evidence to suggest that
company receives a huge shipment of coffee beans and ran-
the funding rate has increased? Justify your answer.
domly select 1000 beans. If the sample proportion of defective
7.78 Sports and Leisure The fielding percentage for a beans is more than 0.09, then the entire shipment will be
major league baseball team is a measure of how well defen- returned to the supplier in Europe.
sive players handle a batted ball (without error) and is based a. If the true proportion of defective beans is 0.07 (as
on total chances ( putouts 1 assists 1 errors ) . One quarter claimed), what is the probability that the shipment will be
into the 2013 baseball season, the Boston Red Sox fielding sent back?
percentage was 0.984.24 Suppose 150 chances are randomly b. If the true proportion of defective beans is 0.08, what is
selected. the probability that the shipment will be accepted?
a. Find the distribution of the sample proportion of chances 7.82 Manufacturing and Product Development McGuckin
handled without error. Hardware in Boulder, Colorado, routinely receives shipments of
b. Find the probability that the sample proportion is less
4 3 8 foot sheets of 12 -inch-thick plywood. A plywood sheet
than 0.97.
= may contain defects, for example, a knot, a split, or a deviation
c. Find a value f such that P (P # f ) 5 0.05. in wood structure. Defective sheets reduce profits because they
d. If the team fielding percentage for these random chances are sold at a lower price. The manufacturer claims the propor-
is greater than 0.99, then everyone on the team receives a tion of all plywood sheets that are defective is 0.05. Suppose a
bonus. What is the probability of a team bonus? large shipment of plywood sheets is received and 200 are ran-
7.79 Psychology and Human Behavior Many Americans domly selected for inspection. If the sample proportion of
own a DVR or subscribe to a special TV recording service. For defective sheets is more than 0.09, the entire shipment will be
example, the Hopper from DISH network allows users to record sent back to the supplier.
up to six HD programs at once. Despite this increased record- a. If the true proportion of defective plywood sheets is 0.05,
ing flexibility, Motorola Mobility recently announced that for what is the probability that the entire shipment will be
U.S. TV viewers, 41% of the recorded content on DVRs is sent back?
never watched.25 Suppose 425 programs recorded on DVRs in b. If the true proportion of defective plywood sheets is 0.03,
the United States are randomly selected. what is the probability that the entire shipment will be
a. Find the probability that the sample proportion of sent back?
programs never watched is less than 0.37. c. If the true proportion of defective plywood sheets is 0.10,
b. Find the probability that the sample proportion is between what is the probability that the shipment will be accepted?
0.42 and 0.43.
= 7.83 Demographics and Population Statistics According
c. Find a value w such that P (P $ w ) 5 0.01. to National Geographic, 28% of Canadians are of British
 Chapter 7 Exercises 327

descent.27 Suppose 150 Canadians in Alberta are randomly is defective. The acceptance sampling plan states: Accept the
selected and 185 Canadians in Ontario are randomly selected. entire shipment if the sample proportion of defective tiles is
a. What is the probability that the sample proportion of 0.08 or less; otherwise, reject the entire shipment.
those Canadians from Alberta of British descent is less a. Find the probability of accepting the entire shipment if
than 0.23? the true proportion of defective floor tiles is 0.01, 0.02,
b. What is the probability that the sample proportion of 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.15.
those Canadians from Ontario of British descent is less b. Plot the probability of accepting the entire shipment (on
than 0.23? the y axis) versus the true proportion of defective floor
c. What is the probability that the sample proportion of tiles (on the x axis). The resulting graph is the OC curve
those Canadians from both provinces is greater than 0.35? for the given acceptance sampling plan.
7.85 Maximum Variance For a fixed sample size n, find
Challenge the value of p that maximizes the variance of the sample
p( 1 2 p)
7.84 Manufacturing and Product Development Suppose proportion, s2P= 5 .
a company is receiving a large shipment of peel-and-stick vinyl n
floor tiles. The manufacturer claims the proportion of defective Hint: Compute the value of the variance for several different
floor tiles is 0.05. The company will select a random sample of values of p. Plot the variance (on the y axis) versus the value
200 floor tiles, carefully inspect each, and determine whether it of p (on the x axis).

CHAPTER 7 SUMMARY
Concept Page Notation / Formula / Description

Parameter 296 A numerical descriptive measure of a population.

Statistic 296 Any quantity computed from values in a sample.
Sampling distribution 297 The probability distribution of a statistic.
s2
Properties of the 306 mX 5 m, sX2 5 . If the underlying population is normal, then X is normal.
n
sample mean X
Central limit theorem 308 As the sample size n increases, the sampling distribution of X will increasingly
approximate a normal distribution, with mean m and variance s2 /n, regardless
of the shape of the underlying population distribution.
p( 1 2 p)
Distribution of the sample
= 319 mP= 5 p, sP2= 5 . If n is large and both np $ 5 and n ( 1 2 p) $ 5,
= n
proportion of successes P then P is approximately normal.

CHAPTER 7 EXERCISES

APPLICATIONS b. Is there any evidence to suggest the manufacturer is 7

including too much hydrogen peroxide in the product?
7.86 Manufacturing and Product Development Up until Justify your answer.
the early 1900s, people who colored their hair used only herbs
and natural dyes. Today, hair-coloring products contain two 7.87 Physical Sciences A floor slip tester is used to measure
main ingredients: hydrogen peroxide and ammonia. The makers the safety of a floor by comparing the measured coefficient of
of a Clairol hair-color product claim the mean amount of hydro- static friction with accepted standards and guidelines. Several
gen peroxide in each bottle is 0.10 mg/m3. Assume the standard factors can affect floor safety, for example, dampness, polishes,
deviation is 0.05 mg/m3. A random sample of these hair-color and maintenance chemicals. A marble floor is consider safe if
products was obtained, and the amount of hydrogen peroxide in the coefficient of static friction is no greater than 0.5. A random
each bottle was measured. HAIRDYE sample of 50 rainy days was selected, and the coefficient of
a. Suppose the manufacturer’s claim is true. Find the distri- static friction for the marble floor was measured on each day.
bution of the sample mean. The resulting sample mean was 0.6. Is there any evidence to
328 CH A PT ER 7 Sampling Distributions

suggest the marble floor is unsafe on rainy days? Assume the b. What is the probability that the sample mean oxygen
underlying population standard deviation is 0.2 and justify your produced is between 7.25 and 7.5 ml/h?
answer. c. Suppose each plant is exposed to a new high-intensity
grow light and the sample mean oxygen produced for the
7.88 Marketing and Consumer Behavior In the fresh 35 plants is 8.1 ml/h. Is there any evidence that the new
fruits and vegetables section of a Kroger grocery store, lamp has increased oxygen output?
customers can purchase any desired amount (by placing the d. Answer parts (a), (b), and (c) if s 5 3.75 ml/h.
food in a plastic bag to be weighed and/or priced at the
checkout line). For those people who purchase cucumbers, the 7.92 Marketing and Consumer Behavior Seafood
probability distribution for the number purchased is given in restaurants along the coast of New England offer a variety of
the table below. CUCUMBER entrees, but lobster is the most popular meal. At Newick’s
Seafood Restaurant in Dover, New Hampshire, 37% of all
x 1 2 3 4 5 diners order lobster. Suppose 120 customers are selected at
p(x) 0.10 0.50 0.20 0.15 0.05 random, and the meal ordered by each is recorded.
a. Find the distribution of the sample proportion of diners
a. Suppose two customers who purchase cucumbers are
who order lobster, and carefully sketch the probability
selected at random. Find the exact probability distribution distribution.
for the sample mean number of cucumbers purchased, X . b. Find the probability that the sample proportion of diners
b. Find the mean, variance, and standard deviation of X . who order lobster is less than 0.30.
7.89 Marketing and Consumer Behavior A drive-in movie c. Find the probability that the sample proportion of diners
theater charges viewers by the carload, but keeps careful who order lobster is between 0.35 and 0.40.
records of the number of people in each car. The probability d. The manager of Newick’s is concerned that more cus-
distribution for the number of people in each car entering the tomers might be ordering lobster. This would require
drive-in is given in the table below. DRIVEIN a change in restaurant ordering and a shift in kitchen
staff. Suppose the actual proportion of diners who order
x 1 2 3 4 5 6 lobster is 0.42. Is there any evidence to suggest that the
p(x) 0.02 0.30 0.10 0.30 0.20 0.08 proportion of diners who order lobster has increased?
Justify your answer.
a. Suppose two cars entering the drive-in are selected at
random. Find the exact probability distribution for the 7.93 Physical Sciences Tropical rainforests are located in
maximum number of people in either one of the cars, M. areas near the Equator. The temperature in a typical rainforest
b. Find the mean, variance, and standard deviation of M. ranges from 20 to 30°C, and the humidity is usually between
77% and 88%.29 The mean amount of rain per year in any one
7.90 Demographics and Population Statistics Many clubs rainforest is m 5 155 inches with standard deviation s 5 35
and companies around the country offer hot-air balloon rides. inches. Suppose 30 rainforests are selected at random and the
Most impose strict safety regulations and take at most four amount of rain per year is recorded for each.
adults at one time, plus a pilot. The mean weight for an adult a. What is the probability that the mean rainfall per year for
male in the United States is 195.5 pounds.28 Suppose the the 30 rainforests is more than 170 inches?
distribution is normal with a standard deviation of 30 pounds. b. What is the probability that the mean rainfall is between
a. If a hot-air balloon pilot (an adult male) takes three adult 140 and 150 inches?
males for a ride, what is the distribution for the total c. Find a value r such that the probability that the mean
weight aboard, T ? Carefully sketch the probability distri- rainfall is less than r is 0.001.
bution for T.
b. If the total weight is less than 900 pounds, the pilot will 7.94 Technology and the Internet An office manager
have enough fuel to extend the ride by a few minutes. has several computers running distributed programs.
What is the probability of an extended ride? Because of the demands on the system, the machines may
c. If the total weight is over 975 pounds, then the balloon crash at various times during the day and require a hard
will not be able to take off. What is the probability that reset. The probability distribution for the number of times a
the balloon will not be able to take off? randomly selected machine crashes during a day, X, is given
in the table below. COMPUTER
7.91 Biology and Environmental Science A typical
houseplant produces oxygen (from carbon dioxide) in varying x 0 1 2 3 4 5
amounts, depending on the amount of light and water. Suppose
p(x) 0.50 0.30 0.10 0.07 0.02 0.01
a medium Norfolk Island pine produces 7.5 ml per hour of
oxygen when exposed to normal sunlight. Thirty-five Norfolk a. Find the mean, variance, and standard deviation for the
Island pines are selected at random, and the oxygen output is number of crashes by a single machine during a day.
carefully measured for each. Assume s 5 1.75 ml/h. b. Suppose n 5 2 machines are selected at random. Find the
a. What is the probability that the sample mean oxygen sampling distribution of the statistic T, the total number
produced is less than 7 ml/h? of crashes for the two machines.
 Chapter 7 Exercises 329

c. Find the mean, variance, and standard deviation of T. b.

d. Verify the relationships mT 5 2mX and s2T 5 2s2X .
7.95 Parameter Versus Statistic In each of the following
statements, identify the boldface number as the value of a
population parameter or a sample statistic.
a. Some political observers claim that state senators spend
f2
too much time addressing colleagues about pending legis-
lation. The mean length for a random sample of speeches f1
was 23.7 minutes.
b. A consulting firm prepared a report from census data and
concluded that the proportion of single-family homes in a
certain county is 0.52. !
c. In a random sample of adults, the mean number of blinks c.
per day was 22,037.
d. A ballet instructor found the variance of the number of
round-trip miles (to practice) traveled by all students was
150.76.
e. A random sample of people who snowboard at least five f2 f1
times per year found the mean number of injuries per
person per year to be 3.4.
f. In a survey conducted by the U.S. Postal Service, 80% of
respondents supported the proposed new six-day package,
five-day delivery schedule.
7.96 Medicine and Clinical Studies The stirrup bone, !
located in the middle ear, is the smallest bone in the human
d.
body, with a mean length of 3 mm.30 Suppose the standard
deviation of the length is 0.16 mm. This bone, located in the
ear, is part of a leverage system that can affect hearing.
f2
There is some speculation that high noise levels can affect
the development of this bone and inhibit hearing. In a
random sample of adults who have lived in a large city their
entire lives, the length of the right-ear stirrup bone was
obtained for each. STIRRUP
a. Find the distribution of the sample mean. Carefully sketch f1
the probability density function.
b. Is there any evidence to suggest the length of the stirrup
bone for lifetime city adults is different from 3 mm? !
Justify your answer.
7.98 Manufacturing and Product Development Above-
7.97 Sampling Distribution Each of the following graphs ground swimming pools are very popular in some parts of the
shows the probability distribution for an underlying distribution country. The vinyl liners used in these pools vary in thickness,
and for the sampling distribution of the sample mean of n but most pools are sold with a 15- or 16-mil-thick liner.31 A
observations (drawn from the underlying distribution). Identify certain company sells special liners and claims the mean thick-
each probability distribution function. ness is 20 mil. Suppose the thickness of these liners is normally
a. distributed with standard deviation 0.63. A random sample of
these vinyl liners was obtained and the thickness of each was
carefully measured. The resulting data are given in the follow-
ing table. LINERS

f2 20.73 18.31 20.19 19.55 19.35

20.42 19.59 20.42 20.76 20.03
f1
a. Suppose the company’s claim is true. Find the distribution
of the mean thickness of these liners.
b. Use the data to determine whether there is any evidence
! to suggest that the company’s claim is false.
330 CH A PT ER 7 Sampling Distributions

7.99 Travel and Transportation The Greenline Taxi recyclable products once a week, sorted into barrels of paper,
Company in New York City keeps careful records of items glass, and plastic. The amount of glass recycled per week by a
left behind by riders. Most items are claimed, but many single household is normally distributed with mean m 5 27
remain in a lost-and-found area at the company’s headquar- pounds and standard deviation s 5 7 pounds. Suppose 12
ters. Records indicate that the proportion of riders who leave households are randomly selected.
an item in a taxi is 0.12. Suppose a random sample of 250 a. What is the probability that the total glass collected for
riders is obtained. the 12 homes, T, will be less than 350 pounds?
a. Find the sampling distribution of the sample proportion b. Find a value g such that P ( T $ g) 5 0.05.
of riders who leave an item behind. Carefully sketch the c. If the total glass collected for the 12 homes is more than
probability density function. 400 pounds, the recycling plant will make a profit. Find
b. What is the probability that the sample proportion of rid- a value of m such that the probability of making a profit
ers who leave an item behind is more than 0.15? is 0.10.
c. What is the probability that the sample proportion
of riders who leave an item behind is between 0.11 7.103 Sports and Leisure The Hotel Association of Canada
and 0.115? reported that, because of the high cost of airline tickets and
d. Some people speculate that in bad economic times, rid- gasoline, more people are taking a staycation, a vacation close
ers (and people in general) are more careful with their to their hometown. According to a survey, 31% of travelers
belongings. Suppose this sample was obtained during a stayed in a hotel near home, visited local attractions, restau-
recession, and the sample proportion of riders who left rants, and shopped.32 Suppose 250 Canadians planning to take a
an item behind was 0.09. Is there any evidence to suggest vacation are selected at random and asked whether they will
that the true proportion is less than 0.12? take a staycation.
a. What is the probability that the sample proportion who
7.100 Manufacturing and Product Development Eye- will take a staycation is between 0.30 and 0.35?
drops are used by many people to soothe and relieve irritation b. What is the probability that the sample proportion is more
and to lubricate their eyes. A manufacturer claims that the than 0.37?
amount of dextran, which helps to prolong the effect of c. Find a value of the sample size n such that the probability
eyedrops, contained in its eyedrops product is 70%. Thirty-six that the sample proportion is less than 0.40 is 0.95.
randomly selected bottles of these eye drops were obtained and
the sample mean amount of dextran was 68.25%. Assume the 7.104 Manufacturing and Product Development A
population standard deviation is s 5 5%. manufacturer of ice pops fills each plastic container with a
a. Find the sampling distribution of the sample mean. fruity liquid, leaving enough room so that consumers can freeze
b. Find the probability that the amount of dextran is more the product. A filling machine is set so that the amount of liquid
than 71%. in each ice pop is normally distributed with mean 8.00 ounces
c. Does the sample mean found suggest that the manufac- and standard deviation 0.25 ounce. Suppose 16 ice pops are
turer’s claim is wrong? Justify your answer. randomly selected.
a. Find the probability distribution of the sample mean num-
7.101 Public Health and Nutrition Individuals who belong
ber of ounces in each container, X .
to a health maintenance organization (HMO) usually share the
b. Find the probability that the sample mean is less than
cost of certain medical services, for example, paying $10 for an
7.9 ounces.
office visit. An insurance company study indicates that the
c. Find the probability that the sample mean is more than
proportion of all people who belong to an HMO who have a
8.15 ounces.
copayment of more than $10 for an office visit is 0.65. Suppose
d. Suppose the filling machine operator can fine-tune the
1000 people who belong to an HMO are randomly selected, and
process by controlling the standard deviation of the
the office copayment amount is recorded.
fill. Find a value for s such that the probability that the
a. Find the distribution of the sample proportion of people
sample mean is more than 8.05 ounces is 0.05.
who have a copayment of more than $10 for an office
visit. Verify the nonskewness criterion. 7.105 Biology and Environmental Science During
b. Find the probability that the sample proportion is more Spring 2013, a huge population of 17-year cicadas emerged
than 0.66. from the ground on the East Coast of the United States.
c. Find the probability that the sample proportion is between These prehistoric-looking bugs have red eyes, huge wings,
0.64 and 0.67. and make noises as loud as a lawnmower.33 Suppose the
d. Find a value h such that the probability that the sample mean noise level is 90 decibels and the standard deviation is
proportion is less than h is 0.01. 15.6 decibels. During the cicada emergence in Virginia and
other Southern states, a random sample of 40 locations was
selected and the noise level (in decibels) was carefully
EXTENDED APPLICATIONS
measured.
7.102 Physical Sciences Carlinville, Illinois, has just started a. Find the probability distribution of the sample mean
an ambitious recycling program. Special trucks collect noise level.
 Chapter 7 Exercises 331

b. Find the probability that the sample mean is less than b. Suppose the true mean diameter of the table tennis balls
60 decibels, the level of moderate conversational speech. is 40.4 mm. What is the probability that the shipment will
c. Find the probability that the sample mean is more than be accepted by the sporting goods store?
110 decibels, the level of a passing train. c. Suppose the true mean diameter of the table tennis balls
d. Find a value of n such that the probability that the sample is 39.4 mm. What is the probability that the shipment will
mean is less than 93 decibels with probability 0.95. be accepted by the sporting goods store?
7.106 Sports and Leisure A large sporting goods company
LAST STEP
has just received a shipment of 100,000 table tennis balls. USA
Table Tennis tournament regulations specify that the diameter 7.107 How much was the typical foreclosure check?
of the ball must be 40 millimeters (mm). Suppose the distribu- The checks sent to Americans in 2013 were in the amount
tion of the diameter is normal with standard deviation 0.4 mm. of $300 to $125,000. Suppose a spokesperson from the U.S.
Twenty-five table tennis balls will be selected at random, and Treasury claimed that the mean check amount was $800 with
the diameter of each will be carefully measured. If the mean standard deviation $125. A random sample of 40 Americans
diameter is within 0.2 mm of 40 mm, then the shipment will be compensated under this program was obtained. The amount of
accepted. Otherwise the entire shipment will be returned to the each check was recorded and the sample mean check amount
manufacturer. for these Americans was $755.
a. Suppose the true mean diameter of the table tennis balls a. Find the distribution of the sample mean.
is 40 mm. What is the probability that the entire shipment b. Is there any evidence to suggest that the true mean check
will be sent back to the manufacturer? amount was less than $800?