1 An Introduction to Statistics

and Statistical Inference

Looking Forward
▲

■ Recognize that data and statistics are pervasive and that statistics are used to describe typical
values and variability, and to make decisions that affect everyone.
■ Understand the relationships among a population, a sample, probability, and statistics.
■ Learn the basic steps in a statistical inference procedure.

Is it safe to eat rice?

Arsenic is a naturally occurring element that is found mainly in the Earth’s
crust. Some people are exposed to high levels of arsenic in their jobs, or near
hazardous waste sites, or in some areas of the country in which there are high
levels of arsenic in the surrounding soil, rocks, or even water. Exposure to small
amounts of arsenic can cause skin discoloration, and long-term exposure has
been associated with higher rates of some forms of cancer. Excessive exposure
can cause death.
In 2012, the U.S. Food and Drug Administration (FDA) and Consumer Reports
announced test results that revealed many brands of rice contain more arsenic in
a single serving than is allowed by the Environmental Protection Agency (EPA) in
a quart of drinking water.1 Trace amounts of arsenic may also be found in flour,
juices, and even beer. Earlier in that year, a study conducted at Dartmouth
College detected arsenic in cereal bars and infant formula.
The FDA has established a safe level of arsenic in drinking water, 10 parts per
billion (ppb). However, there is no equivalent safe maximum level for food. Sup-
pose the FDA is conducting an extensive study to determine whether to issue any
warnings about rice consumption. One hundred random samples of rice are
obtained, and each is carefully measured for arsenic.
The methods presented in this chapter will enable us to identify the popula-
tion of interest and the sample, and to understand the definition and impor-
tance of a random sample. Most important, we will characterize the deductive
process used when an an extraordinary event is observed and cannot be attrib-
uted to luck.

CONTENTS
1.1 Statistics Today
1.2 Populations, Samples, Probability, and Statistics
1.3 Experiments and Random Samples
wanphen chawarung/Shutterstock

9
10 CHAP TE R 1 An Introduction to Statistics and Statistical Inference

1.1 Statistics Today

Statistics data are everywhere: in newspapers, magazines, the Internet, the evening
weather forecast, medical studies, and even sports reports. They are used to describe
typical values and variability, and to make decisions that affect every one of us. It is
important to be able to read and understand statistical summaries and arguments with a
critical eye. This chapter presents the basic elements of every statistics problem—a
population and a sample and their connection to probability and statistics. Two common
methods for data collection, observational sampling and experimentation, are also
introduced.
Statistics data are used by professionals in many different disciplines. Actuaries are prob-
ably the biggest users of statistics. They conduct statistical analyses, assess risk, and esti-
mate financial outcomes. An actuary helped compute your last automobile insurance bill.
Statistical analyses are used in a variety of settings. The National Agricultural Statistics
Service publishes statistics on food production and supply, prices, farm labor, and even
the price of land. Pollsters use statistical methods to predict a candidate’s chances of
winning an election. Using complex statistical analyses, companies make decisions about
new products.
Traditional statistical techniques and new sophisticated methods are used every day in
making decisions that affect our lives directly. Pharmaceutical companies use a battery of
standard statistical tests to determine a new drug’s efficacy and possible side effects. Data
mining, a combination of computer science and statistics, is a new technique used for
constructing theoretical models and detecting patterns. This technique is used by many
companies to understand customers better and to respond quickly to their needs. Predic-
tive microbiology is used to ensure that our food is not contaminated and is safe to con-
sume. Given certain food properties and environmental parameters, a mathematical
model is used to predict safety and shelf life.
Statistics is the science of collecting and interpreting data, and drawing logical conclu-
sions from available information to solve real-world problems. This text presents several
numerical and graphical procedures for organizing and summarizing data. The constant
theme throughout the course, however, is statistical inference using a four-step approach:
claim, experiment, likelihood, and conclusion.
Here are some examples of statistics in the news.
1. Statistical inference: As reported in the Archives of Internal Medicine,2 researchers
discovered a decline in the incidence of heart attacks in one Minnesota county follow-
ing the implementation of new smoke-free workplace laws. The incidence of heart
attacks decreased by 33%, from 150.8 to 100.7 per 100,000 people. The authors con-
cluded that second-hand smoke affects the cardiovascular system nearly as much as
active smoking.
2. Summary statistics: In July 2012, Time Newsfeed reported that the average Canadian
is now richer that the average American. The story summarized an article published in
the Toronto-based Globe and Mail and concluded that the average net worth of a
Canadian household was approximately $40,000 more that a typical American house-
hold. The average net worths in this report are summary statistics that suggest the
middle, or central tendency, of a data set.
3. Probability and odds: Every year, approximately 1 in 5 Americans gets the flu.
WebMD suggests that the best way to prevent the flu is to get a flu shot, or influenza
vaccine. Some people have a higher risk of getting the flu. For example, children and
infants, pregnant women, and seniors have a greater chance of getting the flu. Indi-
viduals with disabilities are at higher risk because of their lack of mobility. People with
certain health conditions have a greater chance of getting the flu, and people traveling
to certain areas of the world may be subject to a higher probability of getting the flu. A
solid background in probability is necessary to understand statistical inference.
 1.2 Populations, Samples, Probability, and Statistics 11

4. Likelihood and inference: Recent research suggests that there is a link between the
incidence of lung cancer and cancer-causing chemical pollutants near large industrial
facilities, especially oil refineries.3 The incidence of lung cancer in oil refinery coun-
ties was higher than in non–oil refinery counties. The chance of this happening was so
small that the researchers concluded that oil refinery products significantly affect lung
carcinogenesis.
5. Relative frequency and probability: According to Oddee, some vending machines
mysteriously fall over and crush 13 people per year.4 If all 312 million Americans are
equally likely to be killed by a vending machine, then the probability that a randomly
selected individual will be crushed by a vending machine during a particular year is
0.00000004167 5 13/312,000,000. The relative frequency of occurrence is a good
estimate of probability and is often used to develop statistical models and make
predictions.
There has been an explosion of numerical information, in stories like those above, in
business, in consumer reports, and even in casual conversation. Interpretation of graphs
and evaluation of statistical arguments are no longer reserved for academics and research-
ers. It is essential for all of us to be able to understand arguments based on acquired data.
This numerical, or quantitative, literacy is a vital life-long tool.
No matter how you are employed or where you live, you will have to make decisions
based on available information or data. Here are some questions you may have to consider.
Stephen Finn/Shutterstock
1. Do you have enough information (data) to make a confident decision? How were the
data obtained? If more information is necessary, how will these data be gathered?
2. How are the data summarized? Are the graphical and/or numerical techniques appro-
priate? Does the summary represent the data accurately?
3. What is the appropriate statistical technique for analyzing the data? Are the conclu-
sions reasonable and reliable?

1.2 Populations, Samples, Probability, and Statistics

There are two very general applications of statistics: descriptive statistics and inferential
statistics. Descriptive statistics involve summarizing and organizing the given information,
graphically and/or numerically. The focus of this text is statistical inference. The procedures
of inferential statistics allow us to use the given data to draw conclusions and assess risk.

Definition
Descriptive statistics: Graphical and numerical methods used to describe, organize, and
summarize data.
Here’s a dictionary definition for Inferential statistics: Techniques and methods used to analyze a small, specific set of
inference: a deduction or logical data so as to draw a conclusion about a large, more general collection of data.
conclusion.

Example 1.1 Mishandled Baggage

The U.S. Department of Transportation publishes information concerning automobiles,
public transportation, railroads, and waterways. Much of this information can be sum-
marized or organized—in tables or charts, with a variety of graphs, and numerically—
to describe typical values and variability. These summary descriptive procedures might
be used to indicate preference for a certain airline or to promote safety records.
Figure 1.1 shows a bar graph of the total mishandled baggage reports for certain airlines
in July 2012.
12 CH APT ER 1 An Introduction to Statistics and Statistical Inference

40000

35000

30000

Number of Reports
25000

20000

15000

10000

5000

an

So iian
Ha n
e

Am ta

ka

Sk a
r

t
t

d
tie

es
es
a
lu

es
ite
l
r tr

as
ic
De
tb

yw
hw
wa
on

M
er

Un
Ai

Al
Je

Fr

ut
Airline
Figure 1.1 Mishandled baggage reports. (Source: Office of Aviation
Enforcement and Proceedings, U.S. Department of Transportation.)

Example 1.2 Summary Statistics

The U.S. Census Bureau maintains a huge database of people and households, business
and industry, geography, and other special topics. This information may be neatly orga-
nized using tables, bar charts, pie charts, histograms, or stem-and-leaf plots, or summa-
rized numerically using the mean, median, quartiles, percentiles, variance, or standard
deviation. These simple descriptive statistics reveal characteristics of the entire data set.
Part of a table is shown in Table 1.1.

Example 1.3 Hand Sanitizers that Kill Germs

A department manager for Walgreens claims that Germ-X HandiSani hand sanitizer
kills 99.99% of common harmful germs and bacteria in as little as 15 seconds. For a
hand sanitizer to be effective, the alcohol concentration must be at least 60%. To check
the germ-killing claim, several randomly selected bottles of the hand sanitizer are
obtained and the alcohol concentration in each is carefully measured. These concentra-
tions are used to determine whether there is any evidence to suggest sanitizer is
ineffective. The collected data are used in inferential statistics to draw a conclusion
regarding a claim.

Example 1.4 Magnetic Fuel Savers

Many companies sell magnetic fuel savers for stoves, which are designed to condition
liquef ied propane gas (LPG) prior to combustion to increase power output,
reduce emissions, and save gas. An independent agency tests these devices by record-
ing the amount of gas necessary to boil a specific volume of water. Each test boil is
classified by stove brand, shape and thickness of the pot, and burner size. The data
collected are used to determine whether there is a difference in efficiency. If there is
a difference in the amount of LPG gas used, further inferential statistical techniques
will be used to isolate this difference. It may be due to the stove brand, type of pot, or
burner size.

TRY IT NOW GO TO EXERCISE 1.5

Whether we are summarizing data or making an inference, every statistics problem in-
volves a population and a sample. Consider the definitions on the page that follows.
 1.2 Populations, Samples, Probability, and Statistics 13

Table 1.1 A portion of table summarizing health and safety characteristics in owner-occupied units (national)
Region

Characteristics Total Northeast Midwest South West

Total 76,091 13,480 18,032 29,119 15,460

Safety equipment
Smoke detectors:
Working smoke detector 70,801 12,893 17,015 26,476 14,417
Powered by:
Electricity 4,506 853 954 1,791 908
Batteries 40,094 7,553 10,372 14,221 7,949
Both 25,763 4,409 5,622 10,253 5,479
Not reported 438 78 68 211 81
Carbon monoxide detectors:
Working carbon monoxide detector 35,215 9,311 10,832 9,129 5,944
Powered by:
Electricity 7,248 1,764 2,811 1,647 1,026
Batteries 15,895 4,467 4,533 4,205 2,689
Both 11,810 2,998 3,416 3,205 2,191
Not reported 262 80 72 72 38
Mold
Housing units with mold in last 12 months 2,015 527 524 611 353
Kitchen 213 28 49 88 49
Bathroom(s) 683 99 180 265 139
Bedroom(s) 378 78 69 137 94
Living room 219 45 52 86 35
Basement 611 276 216 94 26
Other room 277 75 65 76 60
Mold not present 72,817 12,762 17,231 27,953 14,870
Not reported 1,259 191 277 554 238
Musty smells
Housing units with musty smells in last 12 months 11,238 2,377 2,999 4,000 1,861
Daily 772 233 216 202 121
Weekly 5,235 808 1,083 2,316 1,028
Monthly 354 94 112 85 63
A few times 4,877 1,242 1,589 1,397 649
Musty smells not present 63,563 10,912 14,760 24,544 13,346
Not reported 1,291 191 273 574 253
All numbers are in thousands.
Source: U.S. Census Bureau.

Definition
A population is the entire collection of individuals or objects to be considered or studied.
A sample is a subset of the entire population, a small selection of individuals or objects
taken from the entire collection.
A variable is a characteristic of an individual or object in a population of interest.

A CLOSER L OK
1. A population consists of all objects of a particular type. There are usually infinitely
many objects in a population, or at least so many that we cannot look at all of them.
14 CHA PT ER 1 An Introduction to Statistics and Statistical Inference

2. A sample is simply a (usually) small part of a population.

3. A variable may be a qualitative (categorical) or a quantitative (numerical) attribute of
each individual in a population.
The Solution Trail is a technique and visual aid for problem solving (illustrated in the
next example). It is a guide to help us plan how to solve a problem. Look at the Solution
Trail before you read the steps of the solution. Start this hike by identifying keywords and
phrases. The four steps to solving each problem are
1. Find the keywords.
2. Correctly translate these words into statistics.
3. Determine the applicable concepts.
4. Develop a vision for the solution.
The keywords lead to a translation into statistics. The statistics question is solved using
specific concepts. The keywords, translation, and concepts are all used to develop a vision
for the solution. This technique is not applicable or necessary in every problem. It is most
appropriate for probability through hypothesis testing, the foundation of most introduc-
tory statistics courses.
The following examples illustrate the relationships among populations, samples, and
variables.

Solution Trail 1.5 Example 1.5 High Anxiety

Various research studies suggest that whole-grain foods may be a natural help for those
KEYW OR DS people who suffer from anxiety, or maybe even statistics anxiety! Whole grains generally
■ Magnesium level contain high levels of magnesium, and magnesium deficiency can lead to anxiety.5 A new
■ Slice of whole-grain bread study is concerned with the magnesium level in a slice of whole-grain bread. One hundred
■ One hundred slices slices of whole grain bread are selected at random from various markets, and the magne-
sium in each is carefully measured and recorded. Describe the population, sample, and
T RA NSL ATI O N
variable in this problem.
■ Characteristic of each slice
■ All slices of whole-grain bread
SOLUTION
■ Subset of all slices
STEP 1 The population consists of all slices of whole-grain bread in the entire world.
CO NCE P TS Although this population is not infinite, we certainly could not examine every
■ Variable single slice.
■ Population STEP 2 The sample is the 100 slices selected at random. This is a subset of or selection
■ Sample from the population.
VIS I ON
STEP 3 The variable in this problem is the magnesium level. This characteristic will be
carefully measured for each slice, and the data will be summarized or used to
Determine the set of all objects
of interest, the subset, and the draw a conclusion.
attribute to be measured.
Example 1.6 Asleep at the Wheel
The Centers for Disease Control and Prevention released results from a study that indicated
approximately 4% of all adults in the United States said they had fallen asleep at least once
while driving in the last month. Nodding off while driving seems to be more common in
men, and some officials claim the percentage of all U.S. adults who have fallen asleep while
driving is greater than 4%. To check this claim, 10 adult drivers were selected from across
the country. Each person was asked if he or she had fallen asleep while driving, and the
results were recorded. Describe the population, sample, and variable in this problem.

SOLUTION
STEP 1 The population consists of all adult drivers in the United States. This population is
not infinite, but it is so large that it would be impossible to contact every adult driver.
 1.2 Populations, Samples, Probability, and Statistics 15

Solution Trail 1.6 STEP 2 The sample consists of the 10 adult drivers selected.
STEP 3 The variable in the problem is whether the driver has fallen asleep at the wheel,
K EY WO R D S a yes/no response.
■ Adult drivers
■ 10 adult drivers
TRY IT NOW GO TO EXERCISE 1.7
■ Fallen asleep while driving

TRA N SLATI O N A CLOSER L OK

■ All adult drivers in the Example 1.6 raises some important issues regarding the sample of 10 adult drivers.
United States
■ Subset of all adult drivers 1. How large a sample is necessary for us to be confident in our conclusion? Ten adult
■ Characteristic of adult drivers drivers may not seem like enough. But how many do we need? 100? 1000? We will
consider the problem of sample size in Chapter 7 and beyond.
CON C EPT S 2. This problem does not say how the sample was obtained. Perhaps the first 10 drivers
■ Population who recently renewed their licenses were selected. Or maybe only those from one state
■ Sample were included. To draw a valid conclusion, we need to be certain the sample is repre-
■ Variable sentative of the entire population. The formal definition of a representative sample is
presented in Section 1.3.
V I SI O N
Determine the set of all objects Statistical inference is based on, and follows from, basic probability concepts.
of interest, the subset, and the Probability and inferential statistics are both related to a population and a sample, but
attribute measured for each from different perspectives. For the rest of this chapter, statistics really means inferen-
object. tial statistics.

Definition
To solve a probability problem, certain characteristics of a population are assumed to be
known. We then answer questions concerning a sample from that population.
In a statistics problem, we assume very little about a population. We use the information
about a sample to answer questions concerning the population.

Figure 1.2 illustrates this definition. Picture an entire population of individuals or

objects. Suppose we know everything about the population and we select a sample from
this population. A probability problem would involve answering a question concerning
the sample. In a typical statistics problem, we assume very little about the population. We
select a sample and analyze it completely. We use this information to draw a conclusion
about the population.

Probability

Population Sample

Statistics
Figure 1.2 Relationships among probability,
statistics, population, and sample.

In Figure 1.2, it may seem like we can start our study anywhere in this circular dia-
gram. However, we need to understand probability before we can learn statistics. A solid
background in probability is necessary before we can actually do statistical inference.
16 CHA PT ER 1 An Introduction to Statistics and Statistical Inference

Example 1.7 Most-Watched Television Finales

According to Koldcast Entertainment Media, the final episode of M*A*S*H, with almost
106 million viewers, is still the most-watched television finale ever.6 Other top television
finales include Cheers, Seinfeld, Breaking Bad, and Friends. Consider the population
consisting of all television viewers and a sample of 20 from this population.
Population: All television viewers at the time M*A*S*H was aired
Sample: The 20 television viewers from this population
Here is a probability question. The final episode of M*A*S*H was watched by 47% of
all television viewers. What is the probability that 10 or more (of the 20) selected viewers
watched this episode? We know something about the population, and try to answer a ques-
tion about the sample.
Here is a statistics question. Suppose we interview the 20 viewers in the sample and find
that 9 of the 20 watched the final episode of M*A*S*H. What can we conclude about the
percentage of all television viewers who watched the final episode of M*A*S*H? We know
about the sample and try to answer a question about the (whole or general) population.

Example 1.8 Trouble Falling Asleep

A recent study of sleep habits by Statistics Canada indicated that 35% of those women
surveyed had difficulty falling asleep and staying asleep, whereas only 25% of men expe-
rienced the same troubles. The study also showed that Canadian women tend to sleep
longer than Canadian men.7 Consider the population consisting of all Canadian women
and a sample of 100 from this population.
A probability question: Suppose 35% of all Canadian women have difficulty falling
asleep. What is the probability that at most 30 (of the 100) women in the sample have
difficulty falling asleep?
A statistics question: All 100 women selected are asked to complete an extensive ques-
tionnaire. The information indicates that 45 of the 100 have difficulty falling asleep. What
does this suggest about the proportion of all Canadian women who have difficulty falling
asleep?

TRY IT NOW GO TO EXERCISE 1.12

SECTION 1.2 EXERCISES

Concept Check a. It has been reported that 62% of all people use a social
media.
1.1 True/False Inferential statistics are used to draw a b. Only 37% of all people in the United States are eligible to
conclusion about a population. donate blood.
1.2 True/False Descriptive statistics are used to indicate how c. Fifty-two percent of all old washing machines are
the data were collected. front-loading.
d. Forty percent of the people in the Washington, DC, area
1.3 Fill in the Blank travel over the holiday season.
a. The entire collection of objects being studied is called the
. Applications
b. A small subset from the set of all 2013 minivans is called
a . 1.5 Descriptive or Inferential Statistics Determine
c. Consider the amount of sugar in breakfast cereals. This whether each of the following is a descriptive or an inferential
characteristic of breakfast cereal (objects) is called a statistics problem.
. a. The Nebraska Department of Transportation maintains
records concerning all trucks stopped for inspection. A
report of these inspections lists the proportion of all
Practice
trucks stopped, by cargo.
1.4 Probability/Statistics In each of the following prob- b. Eric Knudsen, a researcher at Stanford University Medical
lems, write a probability and a statistics question associated Center, obtains a random sample of wild owls and
with the given information. measures how far each can turn its neck. The data are used
 1.2 Populations, Samples, Probability, and Statistics 17

to conclude that an owl can turn its neck more than 300
120 degrees from the forward position.
250
c. A Navy research facility runs several tests to check the
structural integrity of a new submarine. A laboratory 200

Bankruptcies
report states the vessel can withstand pressure at depths
of at most 800 ft. 150
d. A safety inspector in Atlanta selects a sample of
100
apartment buildings and checks the fire ladders on each.
The proportion of broken ladders in the sample is used to 50
estimate the proportion of broken fire ladders in the
entire city. 0

te
n

de

e
sa in g
e. Like most states, a large portion of the New York

Re rad

tio
io

io

nc

ta
ra
ct

at
r

na
ta

es
tu

t
ru

rm
State budget is spent on health care, pensions, and

or
le

il
ac

Fi
t

ta

al
ns

fo
sp
uf

Re
In
Co

le
education. The pie chart in Figure 1.3 shows the

an
an
ho

Tr
M
W
percentage of the budget spent on each item for Fiscal Industry
Year 2013.8
Figure 1.4 Total bankruptcies in Quebec.

Fiscal Year 2013 Spending

e. A report issued by the athletic department at Brigham

Health Care 38% Young University listed each item in a trainer’s bag and
the number of times each was used.
Pensions 12% f. The manager at the Bear Pause Theater in Hackensack,
Minnesota, surveyed patrons and summarized opinions
General 2% associated with seating comfort, movie sound, and
Education 11% Interest 3% snacks.
Protection 6%
1.7 Medicine and Clinical Studies Managers at
Transportation 10%
Cedarcrest Hospital in Newington, Connecticut, are
Welfare 8%
interested in the length of stay (in days) of patients admitted
Other Spending 10%
for open-heart surgery. Hospital managers have decided to
Figure 1.3 Pie chart illustrating New York State Fiscal limit their investigation to open-heart patients who were
Year 2013 spending. operated on within the last year. Thirty open-heart surgery
patients admitted to the hospital within the last year are
selected. What is the population of interest, the sample, and
f. A report from the Louisiana Department of Agriculture the variable in this problem? Write a Solution Trail for this
and Forestry lists the prices paid for raw forest products at problem.
the first point of sale.
1.8 Marketing and Consumer Behavior T-shirt
1.6 Descriptive or Inferential Statistics Determine
labels irritate many people’s skin, so Calico Graphics
whether each of the following is a descriptive or an inferential
of Wolfeboro, New Hampshire, would like to produce shirts
statistics problem.
without a label. The company wants to know whether there
a. The bar chart in Figure 1.4 shows the number of
is an advantage to producing this type of T-shirt. Fifty people
bankruptcies for certain industries in Quebec during
are surveyed about whether they cut the tags off their
a recent year.9
T-shirts. What is the population of interest, the sample, and
b. The resting heart rate was measured for adult males
the variable in this problem? Write a Solution Trail for this
from two separate groups: those who exercise at least
problem.
three days per week, and those who do not exercise
regularly. The resulting data are used to suggest that 1.9 Psychology and Human Behavior Managers at Citi-
regular exercise decreases the resting heart rate in group, Inc., in New York, are concerned about the number of
adult males. employees who eat and/or drink at their desks while working.
c. Interior Exterior Remodeling, in Northridge, California, Some managers believe this is an unnecessary distraction, and
maintains a comprehensive list of each home constructed spills can cause computer failures and ruin documents. Thirty-
by type, size, exterior color, etc. five employees are selected, and each is questioned about eating/
d. Researchers at the Center for Food Safety selected a drinking while working. Describe the population and the sample
sample of frozen toaster apple strudel sold in grocery in this problem.
stores. Measurements indicated the producer was baking
each piece of strudel with less apple than advertised on 1.10 Public Policy and Political Science Senator Marco
the box. Rubio of Florida is unsure of his vote on an emotional and
18 CHA PT ER 1 An Introduction to Statistics and Statistical Inference

controversial issue. Before he votes, he would like to know 1.13 Psychology and Human Behavior During each
what his constituents think about the proposed bill. An aide for summer, many families spend part of their vacation time at a
the senator selects 500 people from Florida and asks them beach along the East or West Coast. Due to the popularity of
whether they believe the bill should become law. Describe the movies like Jaws, and recent shark attacks on surfers, swim-
population and the sample in this problem. mers, snorkelers, and spearfishermen, Americans have become
increasingly concerned about water activities. Research
1.11 Economics and Finance In 2012, Hurricane Sandy suggests that 46% of all shark attacks are on divers.10 One
caused over 2 million households to lose power and damaged thousand records of shark attacks are selected, and each is
over 70,000 homes and businesses. Every family filed some categorized by victim group.
sort of insurance claim for damage to their home and/or car. a. What is the population of interest?
An insurance company serving the area is interested in the b. What is the sample?
typical amount of a claim as a result of this storm. Seventy- c. Describe the variable of interest.
five affected families are selected and their total claims are
recorded. Describe the population and the sample in this 1.14 Manufacturing and Product Development Spray-on
problem. tans, or fake tans, contain several chemicals that have been
linked to allergies, diabetes, and obesity.11 Twenty fake tan
1.12 Probability or Statistics In each of the following products are selected and the amount of dihydroxyacetone (the
problems, identify the population and the sample, and active ingredient) in each is measured. Describe the population,
determine whether the question involves probability or the sample, and variable in this problem.
statistics.
a. Seventy-five percent of all people who buy a dining room 1.15 Medicine and Clinical Studies There is some
table purchase matching chairs. Five people who evidence to suggest people with chronic hepatitis C have a
purchased a dining room table within the last month are liver enzyme level that fluctuates between normal and
selected at random. What is the probability that all five abnormal.12 Fifty patients diagnosed with hepatitis C are
purchased matching chairs? selected and their liver enzyme levels are recorded each day
b. Twenty-five people entering a rest area and food court on for one month. Describe the population, sample, and variable
Highway 59 near Houston are selected at random. Of in this problem.
these 25 people, 20 purchased food from at least one of 1.16 Manufacturing and Product Development Paper
the eateries. Estimate the true proportion of people towel manufacturers constantly advertise their products’
stopping at this rest area who purchase food. strength, amount of stretch, and softness. A consumer group
c. Historical records indicate 1 out of every 500 people is interested in testing the absorption of Bounty paper
using a particularly steep water slide suffer some kind of towels. Thirty-five rolls are selected, and the amount of
injury. Fifty people using the slide are selected at random. absorption for a single paper towel from each roll is
How many do you expect to be injured? recorded.
d. A building inspector in Henderson, Nevada, is checking a. What is the population of interest?
public buildings with doors that open automatically. One b. What is the sample?
hundred doors are randomly selected. Careful inspection c. Describe the variable of interest.
reveals that 12 doors are broken. Use this information to
estimate the percentage of automatic doors in Henderson
that are broken. Extended Applications
e. One thousand people entering Los Angeles International
1.17 Manufacturing and Product Development While
Airport (LAX) are selected at random. Each person is
much of the cheddar cheese consumed around the world is
asked to complete a short survey regarding travel. The
processed, some is still produced in the traditional manner:
survey results show 637 carry a frequent-flier card. Is
made in small batches, wrapped in cloth to breathe, and allowed
there evidence to suggest the true proportion of travelers
to age. Most traditional cheddar is aged one to two years; like
entering LAX who carry frequent flier cards is greater
fine wines, older cheddars assume their own character and
than 0.60?
flavor. Suppose 75% of all cheddars are aged less than two
f. The Risdall Advertising Agency reports 65% of all
years, and a sample of 20 cheddar cheeses from around the
women have purchased perfume within the last three
world is obtained.
months. Thirty-four women are selected at random. Is it
a. Describe the population and the sample in this problem.
likely more than 20 of these women purchased perfume
b. Write a probability question and a statistics question
within the last three months?
involving this population and sample.
g. Representatives from the Occupational Safety and Health
Administration inspected several for-profit and Medicare 1.18 Marketing and Consumer Behavior Magazines,
nursing homes for any violations. The resulting data will newspapers, and books have become more readily available
be used to determine whether there is any evidence to in digital format. In addition, the quality of readers, for
suggest the quality of treatment is different in the two example, the Kindle, Nook, and iPad, has increased. A recent
types of nursing homes. study suggests that 21% of adults read an ebook within the
 1.3 Experiments and Random Samples 19

past year.13 Suppose a sample of 500 adults in the United making certain investments overseas. A recent study suggests
States is obtained. that 4% of large companies have plans to relocate jobs back to
a. Describe the population and the sample in this problem. the United States.14 Seventy-five large companies are selected,
b. Write a probability question and a statistics question and each is surveyed to determine if it plans to move jobs back
involving this population and sample. to the United States.
a. What is the population of interest?
1.19 Business and Management One of the main reasons b. What is the sample?
that U.S. companies shift jobs overseas is labor costs. Although c. Describe the variable of interest.
the compensation gap between the United States and China has d. Write a probability question and a statistics question
decreased recently, the tax code still rewards companies for involving this population and sample.

1.3 Experiments and Random Samples

Statisticians analyze data from two types of experiments: observational studies and
experimental studies. The definitions are given below.

Definition
In an observational study, we observe the response for a specific variable for each indi-
vidual or object.
In an experimental study, we investigate the effects of certain conditions on individuals
or objects in the sample.

The collected data in an observational study may be summarized in a variety of ways,

or used to draw a conclusion about the entire population. The following is an example of
an observational study.

Example 1.9 Is There Time for Breakfast?

Lorraine GaNun, a guidance counselor at Rice School in Marlton, New Jersey, is inter-
ested in the amount of time each student spends in the morning eating breakfast. Some
students wake up an hour before the bus arrives, have a leisurely breakfast, read the news-
paper comics, and complete last-minute homework. Others roll out of bed and onto the
© Lynch Creative Ltd/Alamy
school bus. Mrs. GaNun decides to measure the amount of time from wake-up to school
bus arrival. A random sample of students is selected, and each is asked for the amount of
school-day preparation time. The data are summarized graphically and numerically in this
observational study.

In almost all statistical applications, it is important for the data to be representative

of the relevant population. A representative sample has characteristics similar to those
of the entire population, and therefore can be used to draw a conclusion about the (gen-
eral) population. The following definition describes a method for obtaining data in an
observational study to ensure the resulting sample is representative of the correspond-
ing population.

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.
20 CH APT ER 1 An Introduction to Statistics and Statistical Inference

A CLOSER L OK
How can we be absolutely certain 1. In practice, a random sample may be very difficult to achieve. Statisticians employ
every possible sample of size n is various techniques, including random number tables and random number generators,
equally likely? to select a random sample.
2. If a sample is not random, then it is biased. There are many different kinds of bias and
factors that contribute to a biased sample.
3. Nonresponse bias is very common when data are collected using surveys. The
majority of people who receive a survey in the mail simply discard it. The original
collection of people receiving the survey may be random, but the final sample of
completed surveys is not. Because the sample is biased, it is impossible to draw a
valid conclusion.
4. Self-selection bias occurs when the individuals (or objects) choose to be included in
the sample, as opposed to being selected. For example, a television news program may
ask viewers to respond to a yes/no question by dialing one of two phone numbers to
cast their vote. Viewers choose to participate, and usually those with strong opinions
(either way) vote. There are many more who did not have the opportunity to respond—
every single sample is not equally likely. Certainly this sample is biased, and hence no
valid conclusion is possible.
5. If the population is infinite, then the number of simple random samples is also infinite.
For finite populations, the formula for the number of possible random samples is pre-
sented in Chapter 7.
6. A simple random sample is vital for sound statistical practice. Before doing any analy-
sis, you should always ask how the data were obtained. If there is any evidence of a
pattern in selection, if the observations are associated or linked in some way, or if there
is some connection among the observations, then the sample is not random. There is
simply no way to transform bad data into good statistics.

Example 1.10 Town Facility Master Plan

STATISTICAL APPLET Mark Traeger, a member of the Sandown, New Hampshire, planning board, is conducting
SIMPLE RANDOM a survey of residents concerning proposed changes in facilities over the next 10 years. He
SAMPLE plans to choose 100 residents from the total town population of 5143 and will ask each
selected person to complete a short questionnaire. The results will be summarized and
presented at the next town meeting.
Traeger would like a simple random sample of size 100, a representative sample of the
entire town population. Here is one basic selection procedure. Write each person’s name
on a piece of paper and place all of them in a hat. Thoroughly mix the papers and then
select 100 names.
Although this procedure is clear-cut and uncomplicated, it can be very tedious if the
number of individuals or objects in the population is large. In addition, it is hard to guar-
antee a thorough mixing of the slips of paper.
More practical methods for selecting a simple random sample include the use of a
random number table or a random number generator (available in most statistical software
packages). In this example, we might assign each resident a number, from 1 to 5143, and
use a random number generator to produce a list of 100 numbers in this range. The resi-
dents associated with these 100 numbers would comprise the random sample.

TRY IT NOW GO TO EXERCISE 1.29

Researchers often investigate the effects of certain conditions on individuals or
objects. The data obtained are from an experimental study. Individuals are randomly
assigned to specific groups, and certain factors are systematically controlled, or imposed,
 1.3 Experiments and Random Samples 21

in order to investigate and isolate specific effects. The following example is of an exper-
imental study.

Example 1.11 To Fertilize or Not to Fertilize

The manager of Gardener’s Supply Company claims that a new organic fertilizer, in
comparison with the leading brand, increases the yield and size of tomatoes. To test
this claim, tomato plants are randomly assigned to one of two groups. One group is
grown using the leading fertilizer and the other is cultivated using the new product.
At harvest time the size and weight of each tomato is recorded, along with the total
yield per plant. The collected data from this experiment are used to compare the two
fertilizers.

TRY IT NOW GO TO EXERCISE 1.28

In an experimental study, researchers must be careful to ensure that significant effects

are indeed due to an imposed treatment, or controlled factor. Confounding occurs when
several factors together contribute to an effect, but no single cause can be isolated. Sup-
pose the tomato plants in one of the groups in Example 1.11 are watered more and/or
exposed to more sunlight and warmer temperatures. If the tomato plants that received the
new fertilizer were subject to these different (favorable) growing conditions, a difference
in yield cannot be attributed to the new product.
The focus of this text is statistical inference, most of which is based on determining the
likelihood of an observed experimental outcome. This strategy will be used informally in the
early chapters of this book. Formal procedures will be presented beginning in Chapter 9. For
now, we will follow the four-step process presented below.

Statistical Inference Procedure

The process of checking a claim can be divided into four parts.
Claim: This is a statement of what we assume to be true.
Experiment: To check the claim, we conduct a relevant experiment.
Likelihood: This considers the likelihood of occurrence of the observed experimental
outcome assuming the claim is true. We will use many techniques to determine whether
the experimental outcome is a reasonable observation (subject to reasonable variabil-
ity), or whether it is a rare occurrence.
Conclusion: There are only two possible conclusions. (1) If the outcome is reason-
able, then we cannot doubt the claim. We usually write, “There is no evidence to
suggest the claim is false.’’ (2) If the outcome is rare, we disregard the lucky alter-
native, and question the claim. A rare outcome is a contradiction. It shouldn’t hap-
pen (often) if the claim is true. In this case we write, “There is evidence to suggest
the claim is false.’’

Example 1.12 Cell Phone Chargers

The Wireless Emporium ships a box containing 1000 cell phone chargers and claims 999
are in perfect condition and only 1 is defective. Upon receipt of the shipment, a quality
control inspector reaches into the box, mixes the chargers around a bit, selects one at
random, and it’s defective!
Claim: There were 999 good cell phone chargers and 1 defective charger in the box.
Experiment: The quality control inspector selected one cell phone charger from the box,
tested it, and found it to be defective.
22 CHAP TE R 1 An Introduction to Statistics and Statistical Inference

Likelihood: One of two things has happened.

1. The quality control inspector could be incredibly lucky. Intuitively, the chance
of selecting the one defective charger from among the 1000 total chargers is
very small. It is possible to select the one defective charger, but it is very
unlikely.
2. The claim (999 perfect chargers, 1 defective) is false. Because the chance of
selecting the single defective charger is so small, it is more likely the manufac-
We have found evidence the turer (Wireless Emporium) lied about the number of defective chargers in the
claim is false by showing that the shipment. (Perhaps there are really 999 defective chargers and only one good
observed experimental outcome is charger in the box.)
unreasonable, an outcome so rare Conclusion: Typically, statistical inference discounts the lucky alternative. Selecting
that it should almost never the single defective charger is an extremely rare occurrence. Therefore, there is
happen if the claim is really true. evidence to suggest the manufacturer’s claim is false, because this outcome is
very rare.

We will use this four-step process to check a claim in many different contexts.
The method for determining likelihood is the key to this valuable tool for logical
reasoning.

SECTION 1.3 EXERCISES

Concept Check evening, the volunteers each take a specific dorm wing, knock
on doors, and record student answers to several prepared
1.20 True/False In an observational study, we record the questions.
response for a specific variable for each individual or object. a. Is this an observational or an experimental study?
1.21 True/False In an experimental study, we investigate the b. Describe the sample in this problem.
effects of certain conditions on at least three different groups. c. Is this a random sample? Justify your answer.

1.22 True/False The number of simple random samples is 1.28 Demographics and Population Statistics State Farm
always infinite. Insurance Company would like to estimate the proportion of
volunteer firefighters across the country who are full-time
1.23 True/False A simple random sample is representative teachers. The 25 largest volunteer fire companies in the United
of the entire population of interest. States are identified. Each is contacted and asked to complete a
short survey regarding the number of volunteers and the
1.24 Fill in the Blank
occupation of each volunteer.
a. If a sample is not random, then it is .
a. Is this an observational or an experimental study?
b. It is very common to experience when
b. Describe the sample in this problem.
data is collected using surveys.
c. Is this a random sample? Justify your answer.
c. occurs when individuals ask to be
included in an survey. 1.29 Manufacturing and Product Development The
Visniak Bottling Plant in Cheektowaga, New York, has been
1.25 Statistical Inference Name the four parts of every
accused of systematically underfilling 12-oz bottles of soda. An
statistical inference problem.
inspection team enters the plant one afternoon and selects
1.26 Liar, Liar Suppose an experimental outcome is very bottled soda ready for shipment from various locations within
rare. What two things could have happened? the plant. The contents of each selected bottle are carefully
measured.
a. Describe the population and the sample in this problem.
Applications b. Is this a random sample? Justify your answer.
1.27 Fuel Consumption and Cars The administration at 1.30 Biology and Environment Science Oregon Scientific
the University of Nebraska in Lincoln is interested in student has come under suspicion of purposely shipping defective
reaction to a planned parking garage on campus. A dormitory wireless weather stations. The Attorney General’s office in
near the proposed site is selected and several Student Senate Delaware would like to estimate the proportion of defective
members volunteer to solicit responses. One Thursday products being shipped by this company. Describe a method
 1.3 Experiments and Random Samples 23

for obtaining a simple random sample of shipped wireless 1.37 Fuel Consumption and Cars Electric and plug-in
weather stations. electric cars are designed to save gasoline and help the
environment. In addition, there are certain tax credits for these
1.31 Fuel Consumption and Cars The Massachusetts State
types of hybrid automobiles.15 Although there are certainly
Police union is interested in the number of miles driven by each
benefits to owning a hybrid car, many people complain about
officer during an 8-hour shift. Twelve officers are selected from
the slow acceleration, repair expense, and overall comfort.
the 11:00 P.M. to 7:00 A.M. shift, and the number of miles
Thirty-five passengers are randomly selected. Each is blind-
traveled by each officer is recorded.
folded and taken for a ride in a traditional combustion-engine
a. Is this an observational or an experimental study?
automobile and in a comparably sized hybrid car (over the same
b. Describe the population and the sample in this problem.
route). The passenger is then asked to select the car with the
c. Is this a random sample? Justify your answer.
most comfortable ride.
1.32 Manufacturing and Product Development Gillette a. Is this an observational or an experimental study?
claims a new disposable razor provides a closer shave than any b. What is the variable of interest?
other brand currently on the market. One hundred men who are c. Describe possible sources of bias in these results.
observed buying a disposable razor are selected and asked to
participate in a shaving study. 1.38 Manufacturing and Product Development The
a. Describe the population and the sample in this problem. ceramic tile used to construct the floors in a mall must be
b. Is this a random sample? Justify your answer. sturdy, easy to clean, and long-lasting. Before installing a
specific tile, a construction firm orders a box of 25 tiles and
1.33 Manufacturing and Product Development Midwest uses a standard strength test on each. The results are used to
Pet Supplies claims its K9 Chain Link Dog Kennel can be set determine whether the tiles will be used throughout the
up in less than 30 minutes. An investigative reporter would like new mall.
to check this claim. Describe a method for obtaining a simple a. Describe the population and the sample in this
random sample of customers who set up this kennel. problem.
1.34 Sports and Leisure A National Football League coach b. Is this a random sample? If so, justify your answer.
is permitted to initiate two challenges to referee calls per game If not, describe a technique for obtaining a random
(outside of the final 2 minutes in each half). If both challenges sample.
are successful, then the coach is given a third. During a
1.39 Manufacturing and Product Development Many
challenge, the referee reviews the play in question on a replay
comforters contain both white feathers and down in order to
monitor on the field, and the call is either confirmed or the
provide a warm, soft cover. A bed-and-bath company would
challenge is upheld. The NFL reports the time required to
like to expand its line of products and sell comforters for
resolve a coach’s challenge is less than 5 minutes. A sports
queen- and king-size beds. Before manufacturing begins, a
statistician would like to check this claim. Describe a method
random sample of comforters is obtained from other companies
for obtaining a simple random sample of challenges during
and the proportions of white feathers, down, and other compo-
NFL games.
nents are measured and recorded. These data will be used to
1.35 Travel and Transportation The Department of determine the exact mixture of feathers and down for the new
Public Works in Bismarck, North Dakota, would like to line of comforters.
estimate the number of potholes per mile (after a long, snowy a. Is this an observational or an experimental study?
winter). Each selected mile-long stretch will be thoroughly b. What are the variables of interest?
examined for potholes, and the number in each section will be c. Describe a method for obtaining a random sample of
recorded. comforters from current manufacturers.
a. Describe a method for obtaining a simple random sample
of mile-long road segments. 1.40 Marketing and Consumer Behavior Disney World
b. Is this an observational or experimental study? is going to initiate the use of wireless tracking wristbands
for visitors to the Orlando, Florida, theme park.16 The new
1.36 Biology and Environmental Science The Faber Floral MagicBand has several functions: it serves as a hotel room
Company in Kankakee, Illinois, claims to have developed a key and park entry pass, and is linked to customer credit
special spray for roses that causes the blossom to last longer card information. Visitors wearing these wristbands will
than an untreated flower. Fifty long-stemmed roses are obtained have immediate access to certain rides. Suppose a random
and randomly assigned to one of two groups: treated versus sample of Disney World visitors wearing wristbands is
untreated. The treated roses are sprayed, and the lifetime of obtained and the wait time for Big Thunder Mountain is
each blossom is carefully recorded. recorded for each.
a. Is this an observational or an experimental study? a. Is this an observational or an experimental study?
b. What is the variable of interest? b. What are the variables of interest?
c. Describe a technique to randomly assign each rose to a c. Describe a method for obtaining a random sample of
group. visitors wearing wristbands.
 24 CHAP TE R 1 An Introduction to Statistics and Statistical Inference

CHAPTER 1 SUMMARY
Concept Page Notation / Formula / Description

Descriptive statistics 11 Graphical and numerical methods used to describe, organize, and summarize
data.
Inferential statistics 11 Techniques and methods used to draw a conclusion or make an inference.
Population 13 The entire collection of individuals or objects to be considered or studied.
Sample 13 A subset of the entire population.
Variable 13 A characteristic of an individual or object in a population of interest.
Probability problem 15 Certain properties of a population are assumed known. Questions involve a
sample taken from this population.
Statistics problem 15 Information about a sample is used to answer questions concerning a population.
Observational study 19 We observe the response for a specific variable for each individual or object in
the sample.
Experimental study 19 We investigate the effects of certain conditions on individuals or objects in the
sample.
Simple random sample of size n 19 A sample selected in such a way that every possible sample of size n has the
same chance of being selected.
Statistical inference procedure 21 Four-step process: Claim, Experiment, Likelihood, and Conclusion.

CHAPTER 1 EXERCISES

1 APPLICATIONS b. The Agricultural Research Service obtained samples of

natural cocoa from a variety of sources and measured the
1.41 Descriptive or Inferential Statistics Determine total antioxidant capacity in each sample. The resulting data
whether each of the following is a descriptive or an inferential were used to suggest that eating a moderate amount of
statistics problem. chocolate may help prevent cancer, heart disease, and stroke.
a. The Society of Government Economists conducted a c. The U.S. Patent Office issued a report listing every
salary and working conditions survey of top bank company that was granted a patent in 2013 and the
executives in the United States. A report issued by this number of patents awarded to each company.
group included a table that listed the number of bank d. A researcher at Emory University used brain scans to
executives in each state with salaries above $1 million. conclude that zen meditation may help treat disorders
b. The Flowers Canada Growers obtained a sample of people characterized by distracting thoughts.
who sent roses for Valentine’s Day and recorded the color
of the roses purchased. This information was used to 1.43 Descriptive or Inferential Statistics Determine
construct a table listing the proportion of each color of whether each of the following is a descriptive or an inferential
rose purchased on Valentine’s Day. statistics problem.
c. The Intergovernmental Panel of Climate Change collected a. The Food Channel conducted a blind taste test to determine
data associated with global warming and predicted the the best chocolate for baking. A random sample of adults was
extinction of up to 30% of plant and animal species in the obtained, and each was asked to select the best chocolate from
world. among 10 varieties. The final report listed each chocolate
d. American Express conducted a survey of travelers at Los along with the number of people who rated it the best.
Angeles International Airport. The information was used b. After an extensive survey, the Association of Realtors in
to estimate the proportion of all travelers who make a Chicago concluded that the mean price of a single-family
purchase in an airport duty-free shop. home was less the $500,000.
c. After conducting several measurements, the Beijing
1.42 Descriptive or Inferential Statistics Determine Municipal Environmental Monitoring Center issued a
whether each of the following is a descriptive or an inferential warning that indicated the density of PM2.5 (fine particulate
statistics problem. matter, a measure of air pollution) was over the safe limit.
a. A report by NASA listed each weather satellite orbiting the d. International Living issued a report listing percentages of
Earth and the number of years each has been in service. Americans retired and living in each foreign country.
 Chapter 1 Exercises 25

1.44 Public Health and Nutrition Parents Association EXTENDED APPLICATIONS

would like to determine the proportion of teenagers who have
the ability to prepare an entire meal. A sample of teenagers was 1.50 Technology and Internet [Link] reported
obtained and all were asked if they can cook. Describe the that the most recent software piracy rate in the United States
population of interest, the sample, and the variable of interest in was 20%. They define the piracy rate as the number of units of
this problem. pirated software deployed divided by the total number of units
of software installed. One thousand installed software titles are
1.45 Marketing and Consumer Behavior Hallmark is selected. Each is carefully examined to determine if the
interested in the proportion of adults who sent a greeting card software was pirated.
on Mother-in-Law Day. A sample of 400 adults was obtained a. What is the population of interest?
and all were asked whether they sent a greeting card on this b. What is the sample?
holiday, which started in 2002. Describe the population and the c. Describe the variable of interest.
sample in this problem. d. Write a probability question and a statistic question
involving this population and sample.
1.46 Medicine and Clinical Studies A recent study by the
American Academy of Neurology suggests that soft drinks, iced 1.51 Travel and Transportation The Channel Tunnel, or
tea, and even fruit drinks may lead to depression.17 One thousand Chunnel, is a 31.4-mile railroad tunnel beneath the English
individuals who regularly consume soft drinks were selected and Channel between Folkstone, Kent, in England and Coquelles in
each was evaluated for signs of depression. Describe the France. To ensure passenger safety, engineers selected the
population, the sample, and the variable in this problem. 35 deepest areas in the tunnel and measured the pressure on
each section.
1.47 Public Policy and Political Science The Office of the
a. Is this an observational or an experimental study?
Privacy Commissioner of Canada’s (OPC) Contributions
b. What is the variable of interest?
Program is interested in reaction to a proposal to allow police
c. Is this sample random? If so, justify your answer. If not,
to obtain cell phone records without a subpoena. One thousand
describe a technique to obtain a random sample.
people in British Columbia were called and each was asked to
respond to several questions. 1.52 Manufacturing and Product Development In January
a. Is this an observational or an experimental study? 2013, flaws were discovered in two Boeing 787 Dreamliners
b. Describe the sample in this problem. aircraft. Japan Airlines found cracks in the cockpit window in
c. Is this a random sample? Justify your answer. one jet and a minor oil leak in another.18 To assure the public
that the jet is safe, the FAA selected 20 Dreamliners currently
1.48 Travel and Transportation Amtrak would like to
operated by American Airlines and carefully inspected each for
estimate the proportion of travelers on the Sunset Limited, from
any flaws.
New Orleans to Los Angeles, who utilize the Sightseer Lounge
a. Is this an observational or an experimental study?
en route. At the end of the trip on March 15, an Amtrak
b. What is the variable of interest?
representative stopped every third person getting off the train
c. Is this sample random? If so, justify your answer. If not,
and asked them if they used the Sightseer Lounge to buy food,
describe a technique to obtain a random sample.
a drink, or souvenirs.
a. Is this an observational or an experimental study? 1.53 Public Policy and Political Science The Thirty Bench
b. Describe the sample in this problem. Wine Makers in Beamsville, Ontario, would like to determine if
c. Is this a random sample? If not, suggest a method for the alcohol content of its wine is determined by the grape
obtaining a random sample. variety. Samples from two Riesling wines, made from different
grape varieties, were obtained and the alcohol content in each
1.49 Physical Sciences The Air Liquide Company has
bottle was carefully measured.
developed a new deicing chemical for airplanes, consisting of
a. Is this an observational or an experimental study?
glycol and several proprietary additives. The new chemical was
b. What is the variable of interest?
designed to keep aircraft wings ice-free for a longer period of
time. Ten typical Dehaviland commuter airplanes were obtained
and randomly assigned to one of two groups: new chemical
LAST STEP
versus old chemical. Each plane was subject to constant icing
conditions in a controlled environment and treated with one of 1.54 Is it safe to eat rice? A rice manufacturer claims
the chemicals. The length of time until ice formed on the wings that a single serving contains at most 10 ppb of arsenic.
was recorded for each plane. Suppose a random sample of 100 rice servings is obtained and
a. Is this an observational or an experimental study? each is measured for arsenic.
b. What is the variable of interest? a. Identify the population of interest and the sample.
c. Describe a technique to randomly assign each plane to a b. Apply the statistical inference procedure to draw a
chemical group. conclusion when an extraordinary, rare event is observed.