746 C H A P T E R 1 3 Categorical Data Analysis

MINITAB output for Exercise 13.29 SPSS output for Exercise 13.31

(p. 536). Recall that a sample of 96 college students was

shown a video of a simulated theft, then asked to select the a. Find the sample proportion of negative tone news sto-
mug shot that most closely resembled the thief. The stu- ries that is deceptive.
dents were randomly assigned to view either 3, 6, or 12 mug b. Find the sample proportion of neutral news stories that
shots at a time, with 32 students in each group. The number is deceptive.
of students in the 3-, 6-, or 12-photos-per-page groups who c. Find the sample proportion of positive news stories that
selected the target mugshot were 19, 19, and 15, respectively. is deceptive.
a. For each photo group, compute the proportion of stu- d. Compare the sample proportions, parts a–c. Does it
dents who selected the target mug shot. Which group appear that the proportion of news stories that is decep-
yielded the lowest proportion? tive depends on story tone?
b. Create a contingency table for these data, with photo e. Give the null hypothesis for testing whether the authen-
group in the rows and whether or not the target mug ticity of a news story depends on tone.
shot was selected in the columns. f. Use the SPSS printout above to conduct the test, part e.
c. Refer to, part b. Are there differences in the proportions Test at a = .05.
who selected the target mug shot among the three photo 13.32 Healing heart patients with music, imagery, touch, and
groups? Test using a = .10. prayer. “Frontier medicine” is a term used to describe
13.31 Stereotyping deceptive and authentic news stories. Major medical therapies (e.g., energy healing, therapeutic prayer,
newspapers lose their credibility (and subscribers) when they spiritual healing) for which there is no plausible explana-
are found to have published deceptive or misleading news tion. The Lancet (July 16, 2005) published the results of
stories. In Journalism and Mass Communication Quarterly a study designed to test the effectiveness of two types
(Summer 2007), University of Texas researchers investigated of frontier medicine—music, imagery, and touch (MIT)
whether certain stereotypes (e.g., negative references to cer- therapy and therapeutic prayer—in healing cardiac care
tain nationalities) occur more often in deceptive news stories patients. Patients were randomly assigned to receive one
than in authentic news stories. The researchers analyzed 183 of four types of treatment: (1) prayer, (2) MIT, (3) prayer
news stories that were proven to be deceptive in nature and and MIT, and (4) standard care (no prayer and no MIT).
128 news stories that were considered authentic. Specifically, Six months after therapy, the patients were evaluated for
the researchers determined whether each story was negative, a major adverse cardiovascular event (e.g., a heart attack).
neutral, or positive in tone. The accompanying table (saved The results of the study are summarized in the accompany-
in the NEWSSTORY file) gives the number of news stories ing table and saved in the HEALING file.
found in each tone category.
Number of
Authentic News Deceptive News Patients with Major Number of
Stories Stories Cardiovascular Patients with
Therapy Events No Events Total
Negative Tone 59 111
Neutral Tone 49 61 Prayer 43 139 182
Positive Tone 20 11 MIT 47 138 185
Prayer and MIT 39 150 189
Total 128 183 Standard 50 142 192
Based on Lasorsa, D., and Dai, J. “When news reporters deceive: The production Based on Krucoff, M. W., et al. “Music, imagery, touch, and prayer as adjuncts to
of stereotypes.” Journalism and Mass Communication Quarterly, Vol. 84, No. 2, interventional cardiac care: The Monitoring and Actualization of Noetic Trainings
Summer 2007 (Table 2). (MANTRA) II randomized study.” The Lancet, Vol. 366, July 16, 2005 (Table 4).
748 C H A P T E R 1 3 Categorical Data Analysis
were then retested one year later with one of the IQ tests.
The accompanying table gives the number of students
diagnosed with mental deficiency on the basis of the retest.
These data are saved in the MDIQ file. Conduct a chi-
square test for independence to determine whether the
proportion of students diagnosed with MD depends on the
IQ test/retest method. Use a = .01.
13.36 Creating menus to influence others. Refer to the Journal of
Consumer Research (Mar. 2003) study on influencing the
choices of others by offering undesirable alternatives, pre-
sented in Exercise 8.157 (p. 404). In another experiment con-
ducted by the researcher, 96 subjects were asked to imagine
that they had just moved to an apartment with two others
and that they were shopping for a new appliance (e.g., a tele-
vision, a microwave oven). Each subject was asked to create
a menu of three brand choices for his or her roommates;
then subjects were randomly assigned (in equal numbers) to
one of three different “goal” conditions: (1) Create the menu
in order to influence roommates to buy a preselected brand,
(2) create the menu in order to influence roommates to buy
a brand of your choice, and (3) create the menu with no
intent to influence roommates. The researcher theorized that
the menus created to influence others would likely include
undesirable alternative brands. Consequently, the number
of menus in each goal condition that was consistent with
the theory was determined. The data are summarized in the
accompanying table and saved in the MENU3 file. Analyze
the data for the purpose of determining whether the propor-
tion of subjects who select menus consistent with the theory
depends on the goal condition. Use a = .01.
Number Number Not
Consistent Consistent
Goal Condition with Theory with Theory
Influence/preselected 15 17
brand
Influence/own brand 14 18
No influence 3 29
Based on Hamilton, R. W. “Why do people suggest what they do not want?
Using context effects to influence others’ choices.” Journal of Consumer
Research, Vol. 29, March 2003 (Table 2).
13.37 Detecting Alzheimer’s disease at an early age. Refer to the
Neuropsychology (Jan. 2007) study of whether the cogni-
tive effects of Alzheimer’s disease can be detected at an
early age, Exercise 13.17 (p. 735). Recall that a particular
strand of DNA was classified into one of three genotypes:
E4 + /E4 + , E4 + /E4 - , and E4 - /E4 - . In addition to a sample
of 2,097 young adults (20–24 years), two other age groups
were studied: a sample of 2,182 middle-aged adults (40–44
years) and a sample of 2,281 elderly adults (60–64 years).
The accompanying table gives a breakdown of the number
of adults with the three genotypes in each age category for
Age E4 + /E4 + E4 + /E4 - E4 - /E4 -
Group Genotype Genotype Genotype
20–24 56 517 1,524
40–44 45 566 1,571
60–64 48 564 1,669
Source: Jorm, A. F., et al. “APOE genotype and cognitive functioning in a large
age-stratified population sample.” Neuropsychology, Vol. 21, No. 1, January
2007 (Table 1). Copyright © 2007 by the American Psychological Association.
Reprinted with permission.
the total sample of 6,560 adults. (These data are saved in
the E4E4ALL file.) The researchers concluded that “there
were no significant genotype differences across the three
age groups” using a = .05. Do you agree?
13.38 Trapping grain moths. In an experiment described in the
Journal of Agricultural, Biological, and Environmental
Statistics (Dec. 2000), bins of corn were stocked with
various parasites (e.g., grain moths) in late winter. In early
summer (June), three bowl-shaped traps were placed on
the surface of the grain in order to capture the moths.
All three traps were baited with a sex pheromone lure;
however, one trap used an unmarked sticky adhesive, one
was marked with a fluorescent red powder, and one was
marked with a fluorescent blue powder. The traps were set
on a Wednesday, and the catch was collected the following
Thursday and Friday. The accompanying table (saved in
the MOTHTRAP file) shows the number of moths cap-
tured in each trap on each day. Conduct a test (at a = .10)
to determine whether the percentages of moths caught by
the three traps depends on the day of the week.
Adhesive, No Mark Red Mark Blue Mark
Thursday 136 41 17
Friday 101 50 18
Based on Wileyto, E. P. et al. “Self-marking recapture models for estimating
closed insect populations.” Journal of Agricultural, Biological, and Environ-
mental Statistics, Vol. 5, No. 4, December 2000 (Table 5A).
13.39 Classifying air threats with heuristics. The Journal of
Behavioral Decision Making (Jan. 2007) published a study
on the use of heuristics to classify the threat level of
approaching aircraft. Of special interest was the use of a
Totals fast and frugal heuristic—a computationally simple pro-
cedure for making judgments with limited information—
32 named “Take-the-Best-for-Classification” (TTB-C). The
subjects were 48 men and women, some from a Canadian
32
32 Forces reserve unit, others university students. Each sub-
ject was presented with a radar screen on which simulated
approaching aircraft were identified with asterisks. By
using the computer mouse to click on the asterisk, one
could receive further information about the aircraft. The
goal was to identify the aircraft as “friend” or “foe” as fast
as possible. Half the subjects were given cue-based instruc-
tions for determining the type of aircraft, while the other
half were given pattern-based instructions. The researcher
also classified the heuristic strategy used by the subject
Instruction Strategy
Pattern Other
Pattern Other
Pattern Other
Cue TTBC
Cue TTBC
Sample f f
Size
Pattern TTBC
2,097 Cue Guess
2,182 Cue TTBC
2,281 Cue Guess
Pattern Guess
Based on Bryant, D. J. “Classifying simulated air threats
with fast and frugal heuristics.” Journal of Behavioral
Decision Making. Vol. 20, January 2007 (Appendix C).

Supplementary Exercises 13.45–13.69
Understanding the Principles
13.45 True or False. Rejecting the null hypothesis in a chi-square
test for independence implies that a causal relationship
exists between the two categorical variables.
13.46 What is the difference between a one-way chi-square
analysis and a two-way chi-square analysis?
Learning the Mechanics
13.47 A random sample of 250 observations was classified
according to the row and column categories shown in the
following table:
Column
1 2 3 the 113 sports facilities in 1997 would you expect to be
1 20 20 10
Row 2 10 20 70
3 20 50 30
a. Do the data provide sufficient evidence to conclude
that the rows and columns are dependent? Test, using
a = .05.
b. Would the analysis change if the row totals were fixed
before the data were collected?
c. Do the assumptions required for the analysis to be valid
differ according to whether the row (or column) totals
are fixed? Explain.
d. Convert the table entries to percentages by using
each column total as a base and calculating each row
response as a percentage of the corresponding column
total. In addition, calculate the row totals and convert
them to percentages of all 250 observations.
e. Create a bar graph with the percentages from row 1 on
the vertical axis and the column number on the hori-
zontal axis. Draw a horizontal line corresponding to the
total percentage for row 1. Does the graph support the
result of the test conducted in part a?
f. Repeat part e for the percentages from row 2.
g. Repeat part e for the percentages from row 3.
13.48 A random sample of 150 observations was classified into
the categories shown in the following table:
Category
1 2 3 4 5 Medical researchers reported finding a common genetic
ni 28 35 33 25 29
a. Do the data provide sufficient evidence that the catego-
ries are not equally likely? Use a = .10.
b. Form a 90% confidence interval for p2, the probability
that an observation will fall into category 2.
Applying the Concepts—Basic
13.49 Location of major sports venues. There has been a recent
trend for professional sports franchises in Major League
Baseball (MLB), the National Football League (NFL),
the National Basketball Association (NBA), and the
National Hockey League (NHL) to build new stadiums
Supplementary Exercises 13.45–13.69 753
and ballparks in urban, downtown venues. An article in
Professional Geographer (Feb. 2000) investigated whether
there has been a significant suburban-to-urban shift in the
location of major sport facilities. In 1985, 40% of all major
sport facilities were located downtown, 30% in central
cities, and 30% in suburban areas. In contrast, of the 113
major sports franchises that existed in 1997, 58 were built
downtown, 26 in central cities, and 29 in suburban areas.
a. Describe the qualitative variable of interest in the study.
Give the levels (categories) associated with the variable.
b. Give the null hypothesis for a test to determine whether
the proportions of major sports facilities in downtown,
central city, and suburban areas in 1997 are the same
as in 1985.
c. If the null hypothesis of part b is true, how many of
located in downtown, central city, and suburban areas,
respectively?
d. Find the value of the chi-square statistic for testing the
null hypothesis of part b.
e. Find the (approximate) p-value of the test, and give the
appropriate conclusion in the words of the problem.
Assume that a = .05.
13.50 Scanning Internet messages. Inc. Technology (Mar. 18, 1997)
reported the results of an Equifax/Harris Consumer Privacy
Survey in which 328 Internet users indicated their level of
agreement with the following statement: “The government
needs to be able to scan Internet messages and user commu-
nications to prevent fraud and other crimes.” The number of
users in each response category is summarized as follows:
Agree Agree Disagree Disagree
Strongly Somewhat Somewhat Strongly
59 108 82 79
a. Specify the null and alternative hypotheses you would
use to determine whether the opinions of Internet users
are evenly divided among the four categories.
b. Conduct the test of part a, using a = .05.
c. In the context of this exercise, what is a Type I error? A
Type II error?
d. What assumptions must hold in order to ensure the
validity of the test you conducted in part b?
13.51 Risk factor for lumbar disease. One of the most common
musculoskeletal disorders is lumbar disk disease (LDD).
risk factor for LDD (Journal of the American Medical
Association, Apr. 11, 2001). The study included 171 Finnish
patients diagnosed with LDD (the patient group) and
321 without LDD (the control group). Of the 171 LDD
patients, 21 were discovered to have the genetic trait. Of
the 321 people in the control group, 15 had the genetic trait.
a. Consider the two categorical variables group and pres-
ence/absence of genetic trait. Form a 2 * 2 contingency
table for these variables.
b. Conduct a test to determine whether the genetic trait
occurs at a higher rate in LDD patients than in the
controls. Use a = .01.
c. Construct a bar graph that will visually support your
conclusion in part b.
754 C H A P T E R 1 3 Categorical Data Analysis

13.52 Late-emerging reading disabilities. Studies of children with
reading disabilities typically focus on “early-emerging” dif-
ficulties identified prior to the fourth grade. Psychologists
at Haskins Laboratories recently studied children with
“late-emerging” reading difficulties (i.e., children who
appeared to undergo a fourth-grade “slump” in reading
achievement) and published their findings in the Journal
of Educational Psychology (June 2003). A sample of 161
children was selected from fourth and fifth graders at
elementary schools in Philadelphia. In addition to record-
ing the grade level, the researchers determined whether
each child had a previously undetected reading disability.
Sixty-six children were diagnosed with a reading disability.
Of these children, 32 were fourth graders and 34 were fifth
graders. Similarly, of the 95 children with normal reading
achievement, 55 were fourth graders and 40 were fifth
graders.
a. Identify the two qualitative variables (and correspond-
ing levels) measured in the study.
b. From the information provided, form a contingency
table.
c. Assuming that the two variables are independent, cal-
culate the expected cell counts.
d. Find the test statistic for determining whether the pro-
portions of fourth and fifth graders with reading dis-
abilities differs from the proportions of fourth and fifth
graders with normal reading skills.
e. Find the rejection region for the test if a = .10.
f. Is there a link between reading disability and grade
level? Give the appropriate conclusion of the test.
materials in four categories, are summarized as fol-
lows: 64 shoppers responded 100%; 20 shoppers stated
75 to 99%; 18 shoppers stated 50 to 74%; and 4 shop-
pers said less than 50%. Suppose a consumer advocate
group claims that half of all consumers believe that
“Made in the USA” means “100%” of labor and materi-
als are produced in the United States, one-fourth believe
that “75 to 99%” are produced in the United States, one-
fifth believe that “50 to 74%” are produced in the United
States, and 5 percent believe that “less than 50%” are
produced in the United States.
a. Describe the qualitative variable of interest in the
study. Give the levels (categories) associated with the
variable.
b. What are the values of p1, p2, p3, and p4, the prob-
abilities associated with the four response categories
hypothesized by the consumer advocate group?
c. Give the null and alternative hypotheses for testing the
consumer advocate group’s claim.
d. Compute the test statistic for testing the hypotheses
stated in part c.
e. Find the rejection region of the test at a = .10.
f. State the conclusion in the words of the problem.
g. Find and interpret a 90% confidence interval for the
true proportion of consumers who believe that “Made
in the USA” means that “100%” of labor and materials
are produced in the United States.
Based on “ ‘Made in the USA’: Consumer perceptions, deception and
policy alternatives.” Journal of Global Business, Vol. 13, No. 24, Spring
2002 (Table 3).

13.53 Politics and religion. University of Maryland professor 13.55 Hearing impairment study. The Journal of Intellectual
Ted R. Gurr examined the political strategies used by Disability Research (Feb. 1995) published a longitudinal
ethnic groups worldwide in their fight for minority rights. study of hearing impairment in a group of elderly patients
(Political Science & Politics, June 2000). Each in a sample with intellectual disability. The hearing function of each
of 275 ethnic groups was classified according to world patient was screened each year over a 10-year period. At
region and highest level of political action reported. The the study’s conclusion, the hearing loss of each patient was
data are summarized in the contingency table below and categorized as severe, moderate, mild, or none. The classifi-
saved in the ETHNIC file. Conduct a test at a = .10 to cations of the 28 surviving patients are summarized in the
determine whether political strategy of ethnic groups table below and saved in the HEARIMP file.
depends on world region. Support your answer with a
graph. Hearing Loss Number of Patients
13.54 “Made in the USA” survey. Refer to the Journal of None 7
Global Business (Spring 2002) study of what “Made Mild 7
in the USA” on product labels means to the typical Moderate 9
consumer, presented in Exercise 2.179 (p. 99). Recall Severe 5
that 106 shoppers participated in the survey. Their
Total 28
responses, given as a percentage of U.S. labor and

Table for Exercise 13.53

Political Strategy

No Political Mobilization, Terrorism, Rebellion,

Action Mass Action Civil War
Latin American 24 31 7
Post-Communist 32 23 4
World Region South, Southeast, East Asia 11 22 26
Africa/Middle East 39 36 20
Table 1 from “Nonviolence ethnopolitics: Strategies for the attainment of group rights and autonomy” by Ted Robert Gurr. Political Science & Politics,
Vol. 33, No. 2, June 2000. Copyright © 2000 The American Political Science Association. Reprinted with the permission of Cambridge University Press.

a. Conduct a test to determine whether the true proportions
of intellectually disabled elderly patients in each of the
hearing-loss categories differ. Use a = .05.
b. Use a 90% confidence interval to estimate the propor-
tion of disabled elderly patients with severe hearing
loss.
13.56 Butterfly hot spots. Nature (Sept. 1993) reported on a study
of animal and plant species “hot spots” in Great Britain.
A hot spot is defined as a 10@km2 area that is species
rich—that is, heavily populated by a species of interest.
Analogously, a cold spot is a 10@km2 area that is species
poor. The accompanying table gives the number of but-
terfly hot spots and the number of butterfly cold spots in
a sample of 2,588 10@km2 areas. In theory, 5% of the areas
should be butterfly hot spots and 5% should be butterfly
cold spots, while the remaining areas (90%) are neutral.
Test the theory, using a = .01.
Butterfly hot spots 123
Butterfly cold spots 147
Neutral areas 2,318
Total 2,588
Source: Prendergast, J. R., et al. “Rare species, the coincidence
of diversity hotspots and conservation strategies.” Nature, Vol.
365, No. 6444, Sept. 23, 1993, p. 335 (Table 1), copyright 1993.
Adapted by permission from Macmillan Publishers Ltd.
Applying the Concepts—Intermediate
13.57 Iraq War survey. The Pew Internet & American Life Project
commissioned Princeton Survey Research Associates to
develop and carry out a survey of what Americans think
about the War in Iraq. Some of the results of the March
2003 survey of over 1,400 American adults are saved in the
IRAQWAR file. Responses to the following questions were
recorded:
1. Do you support or oppose the Iraq War?
(1 = Support, 2 = Oppose)
2. Do you ever go online to access the Internet or World
Wide Web? (1 = Yes, 2 = No)
3. Do you consider yourself a Republican, Democrat, or
Independent? (1 = Rep., 2 = Dem., 3 = Ind.)
4. Have you or anyone in your household served in
the U.S. military? (1 = Yes, I have; 2 = Yes, other;
3 = Yes, both; 4 = No)
5. In general, would you describe your political views as
very conservative, conservative, moderate, liberal, or very
liberal? (1 = Very conservative, 2 = Conservative,
3 = Moderate, 4 = Liberal, 5 = Very liberal)
6. What is your race? (1 = White, 2 = African@American,
3 = Asian, 4 = Mixed, 5 = Native American, 6 =
Other)
7. What is your income range? (1 = 6 10K,
2 = 10920K, 3 = 20930K, 4 = 30940K, 5 = 40950K,
6 = 50975K, 7 = 759100K, 8 = 7 100K)
8. Do you live in a suburban, rural, or urban community?
(1 = urban, 2 = suburban, 3 = rural)
Conduct a series of contingency table analyses to deter-
mine whether support for the Iraq War depends on one
or more of the other categorical variables measured in
the March 2003 survey.
13.58 Pig farmer study. An article in Sociological Methods &
Research (May 2001) analyzed the data presented in the
Supplementary Exercises 13.45–13.69 755
accompanying table. A sample of 262 Kansas pig farmers
was classified according to their education level (college or
not) and size of their pig farm (number of pigs). The data
are saved in the PIGFARM file. Conduct a test to deter-
mine whether a pig farmer’s education level has an impact
on the size of the pig farm. Use a = .05 and support your
answer with a graph.
Education Level
No College College Totals
*1,000 pigs 42 53 95
1,000–2,000 pigs 27 42 69
Farm Size 2,001–5,000 pigs 22 20 42
+5,000 pigs 27 29 56
Totals 118 144 262
Based on Agresti, A., and Liu, I. “Strategies for modeling a categorical
variable allowing multiple category choices.” Sociological Methods &
Research, Vol. 29, No. 4, May 2001 (Table I).
13.59 Multiple-sclerosis drug. Interferons are proteins produced
naturally by the human body that help fight infections
and regulate the immune system. A drug developed from
interferons, called Avonex, is now available for treating
patients with multiple sclerosis (MS). In a clinical study, 85
MS patients received weekly injections of Avonex over a
two-year period. The number of exacerbations (i.e., flare-
ups of symptoms) was recorded for each patient and is
summarized in the accompanying table. For MS patients
who take a placebo (no drug) over a similar two-year
period, it is known from previous studies that 26% will
experience no exacerbations, 30% one exacerbation, 11%
two exacerbations, 14% three exacerbations, and 19% four
or more exacerbations.
Number of Exacerbations Number of Patients
0 32
1 26
2 15
3 6
4 or more 6
Based on data from Biogen, Inc.
a. Conduct a test to determine whether the exacerbation
distribution of MS patients who take Avonex differs
from the percentages reported for placebo patients. Use
a = .05.
b. Find a 95% confidence interval for the true percentage
of Avonex MS patients who remain free of exacerba-
tions during a two-year period.
c. Refer to part b. Is there evidence that Avonex patients
are more likely to have no exacerbations than placebo
patients? Explain.
13.60 Flight response of geese to helicopter traffic. Offshore
oil drilling near an Alaskan estuary has led to increased
air traffic—mostly large helicopters—in the area. The
U.S. Fish and Wildlife Service commissioned a study
to investigate the impact these helicopters have on the
flocks of Pacific brant geese that inhabit the estuary
in the fall before migrating (Statistical Case Studies: A
Collaboration between Academe and Industry, 1998). Two
large helicopters were flown repeatedly over the estuary
756 C H A P T E R 1 3 Categorical Data Analysis
at different altitudes and lateral distances from the flock.
The flight responses of the geese (recorded as “low” or
“high”), the altitude (in hundreds of meters), and the lat-
eral distance (also in hundreds of meters) for each of 464
helicopter overflights were recorded and are saved in the
PACGEESE file. The data for the first 10 overflights are
shown in the following table:
Lateral Flight
Overflight Altitude Distance Response
1 0.91 4.99 HIGH
2 0.91 8.21 HIGH
3 0.91 3.38 HIGH
4 9.14 21.08 LOW
5 1.52 6.60 HIGH
6 0.91 3.38 HIGH
7 3.05 0.16 HIGH
8 6.10 3.38 HIGH
9 3.05 6.60 HIGH
10 12.19 6.60 HIGH
Source: From Erickson, W., Nick, T., and Ward, D. “Investigating Flight
Response of Pacific Brant to Helicopters at lzembek Lagoon, Alaska
by Using Logistic Regression.” Statistical Case Studies: A Collaboration
between Academe and Industry, ASA-SIAM Series on Statistics and
Applied Probability, 1998. Copyright © 1998 Society for Industrial and
Applied Mathematics. Reprinted with permission. All rights reserved.
a. The researchers categorized altitude as follows: less
than 300 meters, 300–600 meters, and 600 or more
meters. Summarize the data in the PACGEESE file by
creating a contingency table for altitude category and
flight response.
b. Conduct a test to determine whether flight response of
the geese depends on altitude of the helicopter. Test,
using a = .01.
c. The researchers categorized lateral distance as fol-
lows: less than 1,000 meters, 1,000–2,000 meters,
2,000–3,000 meters, and 3,000 or more meters.
Summarize the data in the PACGEESE file by creat-
ing a contingency table for lateral distance category
and flight response.
d. Conduct a test to determine whether flight response of
the geese depends on lateral distance of helicopter from
the flock. Test, using a = .01.
e. The current Federal Aviation Authority (FAA) mini-
mum altitude standard for flying over the estuary is
2,000 feet (approximately 610 meters). On the basis of
the results obtained in parts a–d, what changes to the
FAA regulations do you recommend in order to mini-
mize the effects to Pacific brant geese?
13.61 Birds feeding on gypsy moths. A field study was con-
ducted to identify the natural predators of the gypsy moth.
(Environmental Entomology, June 1995). For one part of
the study, 24 black-capped chickadees (common wintering
birds) were captured in mist nets and individually caged.
Each bird was offered a mass of gypsy moth eggs attached
to a piece of bark. Half the birds were offered no other
food (no choice), and half were offered a variety of other
naturally occurring foods such as spruce and pine seeds
(choice). The numbers of birds that did and did not feed
on the gypsy moth egg mass are given in the next table
and saved in the MOTH file. Analyze the data in the table
to determine whether a relationship exists between food
choice and feeding or not feeding on gypsy moth eggs. Use
a = .10.
Fed on Egg Mass
Yes No
Choice of foods 2 10
No choice 8 4
13.62 Gangs and homemade weapons. The National Gang Crime
Research Center (NGCRC) has developed a six-level gang
classification system for both adults and juveniles. The six
categories are shown in the accompanying table. The clas-
sification system was developed as a potential predictor
of a gang member’s propensity for violence in prison, jail,
or a correctional facility. To test the system, the NGCRC
collected data on approximately 10,000 confined offend-
ers and assigned each a score from the gang classification
system. (Journal of Gang Research, Winter 1997). One
of several other variables measured by the NGCRC was
whether or not the offender had ever carried a homemade
weapon (e.g., knife) while in custody. The data on gang
score and homemade weapon are summarized in the table
below and saved in the GANGS file. Conduct a test to
determine whether carrying a homemade weapon in cus-
tody depends on gang classification score. (Use a = .01.)
Support your conclusion with a graph.
Homemade
Weapon Carried
Gang Classification Score Yes No
0 (Never joined a gang, no close friends 255 2,551
in a gang)
1 (Never joined a gang, 1–4 close friends 110 560
in a gang)
2 (Never joined a gang, 5 or more friends 151 636
in a gang)
3 (Inactive gang member) 271 959
4 (Active gang member, no position of 175 513
rank)
5 (Active gang member, holds position 476 831
of rank)
Source: From Knox, G. W., et al. “A gang classification system for corrections.”
Journal of Gang Research, Vol. 4, No. 2, Winter 1997, p. 54 (Table 4). Reprinted
with permission from National Gang Crime Research Center.
13.63 Top Internet search engines. Nielsen/NetRatings is a
global leader in Internet media and market research. In
May 2006, the firm reported on the “search” shares (i.e.,
the percentage of all Internet searches) for the most popu-
lar search engines available on the Web. Google Search
accounted for 50% of all searches, Yahoo! Search for
22%, MSN Search for 11%, and all other search engines
for 17%. Suppose that, in a random sample of 1,000
recent Internet searches, 487 used Google Search, 245
used Yahoo! Search, 121 used MSN Search, and 147 used
another search engine.
a. Do the sample data disagree with the percentages
reported by Nielsen/NetRatings? Test, using a = .05.
b. Find and interpret a 95% confidence interval for the
percentage of all Internet searches that use the Google
Search engine.
 Supplementary Exercises 13.45–13.69 757

13.64 Orientation clue experiment. Human Factors (Dec. 1988)
published a study of color brightness as a body orientation
clue. Ninety college students reclining on their backs in
the dark were disoriented when positioned on a rotating
platform under a slowly rotating disk that blocked their
field of vision. The subjects were asked to say “Stop” when
they felt as if they were right-side up. The position of the
brightness pattern on the disk in relation to each student’s
body orientation was then recorded. Subjects selected only
three disk brightness patterns as subjective vertical clues:
(1) brighter side up, (2) darker side up, and (3) brighter and
darker sides aligned on either side of the subjects’ heads.
The frequency counts for the experiment are given in the
accompanying table and saved in the BODYCLUE file.
Conduct a test to compare the proportions of subjects who
fall into the three disk-orientation categories. Assume that
you want to determine whether the three proportions dif-
fer. Use a = .05.
Disk Orientation
Brighter Darker Bright and Dark
Side Up Side Up Sides Aligned
58 15 17
13.65 Coupon usage study. A hot topic in marketing research is
the exploration of a technology-based self-service (TBSS)
encounter, in which various technologies (e.g., ATMs,
online banking, self-scanning at retail stores) allow the
customer to perform all or part of the service. Marketing
professor Dan Ladik of the University of Suffolk investi-
gated whether there were differences in customer char-
acteristics and customer satisfaction between users of
discount coupons distributed through the mail (nontech-
nology users) and users of coupons distributed via the
Internet (TBSS users). A questionnaire measured several
qualitative variables (defined in the accompanying table)
for each of 440 coupon users. The data are saved in the
COUPONS file.
Variable Name Levels (Possible Values)
Coupon User Type Mail, Internet, or Both
Gender Male or Female
Education High School, Vo-Tech/
College, 4-year College
Degree, or Graduate
School
Work Status Full Time, Part Time, Not
Working, Retired
Coupon Satisfaction Satisfied, Unsatisfied,
Indifferent
types of coupon users. For each characteristic, conduct
a contingency table analysis (at a = .05) to determine
whether Coupon User Type is related to that characteris-
tic. Illustrate your results with graphs.
13.66 Battle simulation trials. In order to evaluate their situ-
ational awareness, fighter aircraft pilots participate in
battle simulations. At a random point in the trial, the
simulator is frozen and data on situational awareness are
immediately collected. The simulation is then continued
until, ultimately, performance (e.g., number of kills) is
measured. A study reported in Human Factors (Mar.
1995) investigated whether temporarily stopping the
simulation results in any change in pilot performance.
Trials were designed so that some simulations were
stopped to collect situational awareness data while oth-
ers were not. Each trial was then classified according
to the number of kills made by the pilot. The data for
180 trials are summarized in the accompanying contin-
gency table and saved in the SIMKILLS file. Conduct
a contingency table analysis and interpret the results
fully.
Number of Kills
0 1 2 3 4 Totals
Stops 32 33 19 5 2 91
No Stops 24 36 18 8 3 89
Totals 56 69 37 13 5 180
Applying the Concepts—Advanced
13.67 Goodness-of-fit test. A statistical analysis is to be done
on a set of data consisting of 1,000 monthly salaries.
The analysis requires the assumption that the sample
was drawn from a normal distribution. A preliminary
test, called the x2 goodness-of-fit test, can be used to
help determine whether it is reasonable to assume that
the sample is from a normal distribution. Suppose the
mean and standard deviation of the 1,000 salaries are
hypothesized to be $1,200 and $200, respectively. Using
the standard normal table, we can approximate the
probability of a salary being in the intervals listed in
the accompanying table. The third column represents
the expected number of the 1,000 salaries to be found
in each interval if the sample was drawn from a normal
distribution with m = +1,200 and s = +200. Suppose
the last column contains the actual observed frequencies
in the sample. Large differences between the observed
and expected frequencies cast doubt on the normality
assumption.

a. Consider the variable Coupon User Type. Conduct a Expected Observed

Interval Probability Frequency Frequency
test (at a = .05) to determine whether the proportions
of mail-only users, Internet-only users, and users of both Less than $800 .023 23 26
media are statistically different. Illustrate the results +800 6 +1,000 .136 136 146
with a graph. +1,000 6 +1,200 .341 341 361
b. The researcher wants to know whether there are differ- +1,200 6 +1,400 .341 341 311
ences in customer characteristics (i.e., Gender, Education, +1,400 6 +1,600 .136 136 143
$1,600 or above .023 23 13
Work Status, and Coupon Satisfaction) among the three