Instructor’s Appendix
Excel Project
From Data to Decision
NOTE: This INSTRUCTOR’S APPENDIX offers suggestions and insights into the EXCEL PROJECT
and FROM DATA TO DECISION sections at the end of each chapter. It has been added in response
to requests for such material from users of the previous editions. Since many of the questions
posed in these sections are open-ended and/or subjective, the material that follows represents only
one possible interpretation of the questions posed is not intended to be a definitive answer and/or
the only “correct” solution.
Chapter 1 – Excel Project
1. Data Analysis is one of the options available by opening Tools in the Excel toolbar.
The first five items listed are
Anova: Single Factor
Anova: Two-Factor with Replication
Anova: Two-Factor without Replication
Correlation
Covariance
2. The first five options available by opening DXXL in the Excel toolbar are
Summaries
Tables
Charts and Plots
Regression
ANOVA
3. Printing the CARS.xls worksheet produces the following.
car weight length braking cylinders disp city hway GHG
Acura RL 4035 194 131 6 3.5 18 26 8.7
Acura TSX 3315 183 136 4 2.4 22 31 7.2
Audi A6 4115 194 129 6 3.2 21 29 7.7
BMW 525i 3650 191 127 6 3 21 29 7.7
Buick LaCrosse 3565 198 146 4 3.8 20 30 7.9
Cadillac STS 4030 196 146 6 3.6 18 27 8.7
Chevrolet Impala 3710 200 155 6 3.9 19 27 8.2
Chevrolet Malibu 3135 188 139 4 2.2 24 32 6.8
Chrysler 300 4105 197 133 8 5.7 17 25 9.3
Dodge Charger 4170 200 131 8 5.7 17 25 9.3
Dodge Stratus 3190 191 131 4 2.4 22 30 7.4
Ford Crown Vict 4180 212 140 8 4.6 17 25 9.3
Ford Focus 2760 168 137 4 2 26 32 6.5
Honda Accord 3195 190 144 4 2.4 24 34 6.6
Hyundai Elantra 2980 177 133 4 2 24 32 6.7
…[rows omitted] … … … … … … … …
Volvo S80 3630 190 136 6 2.9 20 27 8.2
, INSTRUCTOR’S APPENDIX 537
Chapter 1 – From Data to Decision
Too many details of the original experiment are missing from the brief summary presented in the
text to allow for definitive answers, but the following answers suggest possible areas for
discussion.
1. The original experiment contains several possible design flaws. If the original 60 persons in
Wichita were not selected at random, they may have been a convenience sample or connected in
some other way such as not to be representative of the general population. Those who chose to
participate are a voluntary response sample, which is not necessarily representative of the
general population.
2. There may be a discrepancy between what “degrees of separation” means in theory and what
was actually measured by the experiment. If one of the original subjects had a friend who
happened to be going to Cambridge, and that friend went out his way to hand deliver the letter
to the otherwise unknown recipient while in Cambridge, there would be only one intermediary –
but it would not be correct to say there was one only one degree of separation in the sense of
people who actually knew people. Determining whether the experiment justifies the concept of
“six degrees of separation” would involve clarification both of exactly what that concept means
and of the directions communicated to each person in the ongoing chain of contacts.
3. Given the present technology and available databases that allow a person to find and contact
almost any other given person, the concept of degrees of separation may no longer be
meaningful – but an experiment to identify the number of legitimate person-to-person links
necessary to connect any two people may still be possible and have some merit. In fact, one
might even employ the present technology to conduct the experiment more quickly and with
better control. With strict rules allowing contact only between persons that already have
legitimate links with each other, the basic scenario used by Stanley Milgram might be re-run
using either the telephone or e-mail.
,538 INSTRUCTOR’S APPENDIX
Chapter 2 – Excel Project
The histograms as produced by Histogram in the Charts and Plots option of DXXL are given
below without modification for the females (on the left) and the males (on the right).
Since the horizontal scales are different, direct visual comparison difficult. Using Histogram in
the Data Analysis option of Tools, figures can be edited and rescaled so that they both have the
same horizontal and vertical scale and better labeling. The pulse rates are similar in that the
frequency for both genders peaks around 70. But the female rates appear to be positively skewed
and more spread out.
Female Pulse Rates Male Pulse Rates
15 15
Frequency
Frequency
10 10
5 5
0 0
60 70 80 90 100 110 120 60 70 80 90 100 110
beats per minute beats per minute
The histogram of the combined pulse rates may be obtained by placing the female and male values
in a single file. While the histogram of combined rates can use the same horizontal scale as the
above gender-specific figures, the vertical scale must be adjusted.
Combined Pulse Rates
25
20
Frequency
15
10
5
0
60 70 80 90 100 110 120
beats per minute
, INSTRUCTOR’S APPENDIX 539
Chapter 2 – From Data to Decision
Two histograms, using the same horizontal and vertical scales should provide a good visual image
of the differences between the groups.
WINNING ACTRESSES
25
20
15
Frequency
10
5
0
20 30 40 50 60 70 80
age (years)
WINNING ACTORS
25
20
15
Frequency
10
5
0
20 30 40 50 60 70 80
age (years)
There seems to be a significant difference between the two distributions. The winning actresses
appear to be younger than the winning actors, and there appears to be more variability among the
ages of the actresses. An age of about 30 seems to be the most frequently occurring winning age
for females, with the frequencies tailing off from that in both directions – but an age of about 30 is
the lowest winning age for males, and with a low frequency for that category. There are at least
three plausible explanations for this discrepancy.
(1) Acting ability peaks at different ages for females and males.
(2) Female and males actors are judged according to different standards.
(3) Movies are made such that the most powerful female parts are for younger (i.e., about 30 years
old) women and the most powerful male parts are for older (i.e., about 40 years old) men.
Excel Project
From Data to Decision
NOTE: This INSTRUCTOR’S APPENDIX offers suggestions and insights into the EXCEL PROJECT
and FROM DATA TO DECISION sections at the end of each chapter. It has been added in response
to requests for such material from users of the previous editions. Since many of the questions
posed in these sections are open-ended and/or subjective, the material that follows represents only
one possible interpretation of the questions posed is not intended to be a definitive answer and/or
the only “correct” solution.
Chapter 1 – Excel Project
1. Data Analysis is one of the options available by opening Tools in the Excel toolbar.
The first five items listed are
Anova: Single Factor
Anova: Two-Factor with Replication
Anova: Two-Factor without Replication
Correlation
Covariance
2. The first five options available by opening DXXL in the Excel toolbar are
Summaries
Tables
Charts and Plots
Regression
ANOVA
3. Printing the CARS.xls worksheet produces the following.
car weight length braking cylinders disp city hway GHG
Acura RL 4035 194 131 6 3.5 18 26 8.7
Acura TSX 3315 183 136 4 2.4 22 31 7.2
Audi A6 4115 194 129 6 3.2 21 29 7.7
BMW 525i 3650 191 127 6 3 21 29 7.7
Buick LaCrosse 3565 198 146 4 3.8 20 30 7.9
Cadillac STS 4030 196 146 6 3.6 18 27 8.7
Chevrolet Impala 3710 200 155 6 3.9 19 27 8.2
Chevrolet Malibu 3135 188 139 4 2.2 24 32 6.8
Chrysler 300 4105 197 133 8 5.7 17 25 9.3
Dodge Charger 4170 200 131 8 5.7 17 25 9.3
Dodge Stratus 3190 191 131 4 2.4 22 30 7.4
Ford Crown Vict 4180 212 140 8 4.6 17 25 9.3
Ford Focus 2760 168 137 4 2 26 32 6.5
Honda Accord 3195 190 144 4 2.4 24 34 6.6
Hyundai Elantra 2980 177 133 4 2 24 32 6.7
…[rows omitted] … … … … … … … …
Volvo S80 3630 190 136 6 2.9 20 27 8.2
, INSTRUCTOR’S APPENDIX 537
Chapter 1 – From Data to Decision
Too many details of the original experiment are missing from the brief summary presented in the
text to allow for definitive answers, but the following answers suggest possible areas for
discussion.
1. The original experiment contains several possible design flaws. If the original 60 persons in
Wichita were not selected at random, they may have been a convenience sample or connected in
some other way such as not to be representative of the general population. Those who chose to
participate are a voluntary response sample, which is not necessarily representative of the
general population.
2. There may be a discrepancy between what “degrees of separation” means in theory and what
was actually measured by the experiment. If one of the original subjects had a friend who
happened to be going to Cambridge, and that friend went out his way to hand deliver the letter
to the otherwise unknown recipient while in Cambridge, there would be only one intermediary –
but it would not be correct to say there was one only one degree of separation in the sense of
people who actually knew people. Determining whether the experiment justifies the concept of
“six degrees of separation” would involve clarification both of exactly what that concept means
and of the directions communicated to each person in the ongoing chain of contacts.
3. Given the present technology and available databases that allow a person to find and contact
almost any other given person, the concept of degrees of separation may no longer be
meaningful – but an experiment to identify the number of legitimate person-to-person links
necessary to connect any two people may still be possible and have some merit. In fact, one
might even employ the present technology to conduct the experiment more quickly and with
better control. With strict rules allowing contact only between persons that already have
legitimate links with each other, the basic scenario used by Stanley Milgram might be re-run
using either the telephone or e-mail.
,538 INSTRUCTOR’S APPENDIX
Chapter 2 – Excel Project
The histograms as produced by Histogram in the Charts and Plots option of DXXL are given
below without modification for the females (on the left) and the males (on the right).
Since the horizontal scales are different, direct visual comparison difficult. Using Histogram in
the Data Analysis option of Tools, figures can be edited and rescaled so that they both have the
same horizontal and vertical scale and better labeling. The pulse rates are similar in that the
frequency for both genders peaks around 70. But the female rates appear to be positively skewed
and more spread out.
Female Pulse Rates Male Pulse Rates
15 15
Frequency
Frequency
10 10
5 5
0 0
60 70 80 90 100 110 120 60 70 80 90 100 110
beats per minute beats per minute
The histogram of the combined pulse rates may be obtained by placing the female and male values
in a single file. While the histogram of combined rates can use the same horizontal scale as the
above gender-specific figures, the vertical scale must be adjusted.
Combined Pulse Rates
25
20
Frequency
15
10
5
0
60 70 80 90 100 110 120
beats per minute
, INSTRUCTOR’S APPENDIX 539
Chapter 2 – From Data to Decision
Two histograms, using the same horizontal and vertical scales should provide a good visual image
of the differences between the groups.
WINNING ACTRESSES
25
20
15
Frequency
10
5
0
20 30 40 50 60 70 80
age (years)
WINNING ACTORS
25
20
15
Frequency
10
5
0
20 30 40 50 60 70 80
age (years)
There seems to be a significant difference between the two distributions. The winning actresses
appear to be younger than the winning actors, and there appears to be more variability among the
ages of the actresses. An age of about 30 seems to be the most frequently occurring winning age
for females, with the frequencies tailing off from that in both directions – but an age of about 30 is
the lowest winning age for males, and with a low frequency for that category. There are at least
three plausible explanations for this discrepancy.
(1) Acting ability peaks at different ages for females and males.
(2) Female and males actors are judged according to different standards.
(3) Movies are made such that the most powerful female parts are for younger (i.e., about 30 years
old) women and the most powerful male parts are for older (i.e., about 40 years old) men.
