Instructor’s Solution Manual to Chapter Exercises in
The Essentials of Political Analysis, 7th Edition
By Philip H. Pollock II and Barry C. Edwards
Contents
Introduction: Studying Politics Scientifically (No exercises) ................................................... 2
Chapter 1: The Definition and Measurement of Concepts ........................................................ 2
Chapter 2: Measuring and Describing Variables....................................................................... 6
Chapter 3: Creating and Transforming Variables ................................................................... 11
Chapter 4: Proposing Explanations, Framing Hypotheses & Making Comparisons .............. 13
Chapter 5: Graphing Relationships and Describing Patterns .................................................. 18
Chapter 6: Research Design, Research Ethics & Evidence of Causation ............................... 21
Chapter 7: Making Controlled Comparisons ........................................................................... 26
Chapter 8: Foundations of Statistical Inference ...................................................................... 33
Chapter 9: One and Two Sample Hypothesis Tests ................................................................ 36
Chapter 10: Chi-Square Test and Analysis of Variance .......................................................... 39
Chapter 11: Correlation and Bivariate Regression Analysis ................................................... 42
Chapter 12: Multiple Regression ............................................................................................. 47
Chapter 13: Analyzing Regression Residuals ......................................................................... 50
Chapter 14: Logistic Regression ............................................................................................. 53
Chapter 15: Conducting Your Own Political Analysis ........................................................... 56
Version 1.2. Last updated December 5, 2024
1
, Introduction: Studying Politics Scientifically (No exercises)
There are no end-of-chapter exercises in the Introduction (Studying Politics Scientifically).
Chapter 1: The Definition and Measurement of Concepts
Exercise 1-1.
(A) Some possibilities for completing the inventory of contrasting characteristics:
• Frequently / never attends religious services;
• believes / does not believe in supreme deity;
• says that religion is most important / least important aspect of their lives;
• frequently / never contributes time or money to a religious organization;
• frequently / never attends group discussions of religious texts;
• displays / does not display religious symbols or imagery;
• feels proud / does not feel proud of a faith tradition;
• feels / does not feel joy [sorrow] when something good [bad] happens to a fellow
believer.
(B) Students must justify choices in A. This may be easier to do for some attributes. For
example, when someone says that religion is the most (or least) important aspect of his/her
life, it is plausible to assume that the response is tapping into the concept of religiosity and
not any other concept. However, religious attendance may tap a generalized desire for a
sense of belonging to a group (frequent attendance), or the absence of that desire (no
attendance). In any event, the purpose of this part of the exercise is to encourage students to
think empirically and justify their choices.
(C) Example (using attendance): The concept of religiosity is defined as the extent to which
individuals exhibit the characteristic of frequently attending religious services.
Exercise 1-2.
(A) No, the conclusion is not supported. The ecological fallacy could be at work here. The
ecological fallacy occurs when the researcher analyze aggregate-level units of analysis (in
this case, states) to make cross-level inferences about individuals (low-income and high-
income individuals).
(B) Low-income individuals, who are less likely to vote than are high-income individuals, are
2
, more likely to reside in states with higher percentages of high-income individuals. When
one compares states having high percentages of low-income residents with states having
low percentages of low-income residents, one is also comparing states having higher
percentages of voters with states have lower percentages of voters.
Exercise 1-3.
(A) (i) It is likely that a sizeable number of gun owners would be less willing to tell female
interviewers that they own guns compared to male interviewers.
(ii) The social desirability effect is a type of response bias that occurs when human
subjects alter their responses to survey questions to appear more likeable to the
researcher. Because males tend to favor gun ownership more than females, gun
owners may be more willing to tell a male interviewer they own a gun compared to
a female because some subjects tell interviewers what they want to hear to be liked.
(B) (i) Systematic measurement error.
(ii) The social desirability effect would be a consistent and chronic distortion in the
measurement of gun ownership. The problem would cause researchers to
consistently underestimate the true proportion of gun ownership. It is not a random
measurement error because it doesn’t cause non-gun owners to say they own guns;
there’s a pattern to the measurement error.
Exercise 1-4.
(A) Yolanda’s measure has high validity and high reliability. It is valid because all of the
measurements are very close to the true value of 50. It is reliable because the differences
among the four measurements are very small.
(B) Xavier’s measure has high validity and low reliability. Is it valid because the measurements
are close to the true value of 50: one is on the true value; two are 2 units below; one is 4
units above. Compared with Yolanda’s measurements, Xavier’s are less reliable, because
the differences among the four measurements are larger, ranging from 48 to 54.
(C) Zelda’s measure has low validity and high reliability. It has low validity because it
underestimates true knowledge by 4-6 points. It is reliable because the measurements are
very consistent, within a 2-point range.
3
, (D) Warren’s measure has low validity and low reliability. It has low validity because none of
the measurements is at all close to the true value. It is unreliable because the measures vary
wildly in a 30-point range, from 40 to 70.
Exercise 1-5.
Following are several reasons “number of yard signs” is not a valid measure of “pre-election
support.” Because yard signs require public commitment by the sign displayers, yard signs
measure intensity of support among a handful of people, not breadth of support in the electorate.
Because some people display yard signs at the request of the candidate (not on the initiative of the
displayer), yard signs measure candidate initiative and/or campaign strategy, not voter sentiment.
If the number of yard signs is counted, a large family home will be counted as having the same
electoral support as a single person’s home. Some districts, such as dense urban districts and large
rural districts, don’t have a lot of homes with yards to display signs. Using the yard signs
measurement strategy, individuals who don’t own homes will not be counted at all.
Exercise 1-6.
(A) Reliable but not valid.
(B) Mutt Jeffley’s strategy for weighing his dog is reliable because it will yield consistent
results. Each time he steps on the scale holding his dog, the scale will report the same
weight (or very close to it). The measurements are not valid, however, because Mutt is
stepping on the scale with his dog, thereby adding his weight to the measurement.
(C) Mutt could weigh himself without the dog and subtract his weight from the combined
weight he obtains when he steps on the scale holding his dog.
Exercise 1-7.
(A) (i) This question has high face validity.
(ii) The question is intended to measure individual preferences on the trade-off between
creating jobs and protecting the environment. It does so by requiring respondents
to choose between these two values.1
1
Interestingly, this traditional ANES survey question was revised starting in 2012. Respondents are asked to choose
between: “Regulate business to protect the environment and create jobs” and “No regulation because it will not work
4
The Essentials of Political Analysis, 7th Edition
By Philip H. Pollock II and Barry C. Edwards
Contents
Introduction: Studying Politics Scientifically (No exercises) ................................................... 2
Chapter 1: The Definition and Measurement of Concepts ........................................................ 2
Chapter 2: Measuring and Describing Variables....................................................................... 6
Chapter 3: Creating and Transforming Variables ................................................................... 11
Chapter 4: Proposing Explanations, Framing Hypotheses & Making Comparisons .............. 13
Chapter 5: Graphing Relationships and Describing Patterns .................................................. 18
Chapter 6: Research Design, Research Ethics & Evidence of Causation ............................... 21
Chapter 7: Making Controlled Comparisons ........................................................................... 26
Chapter 8: Foundations of Statistical Inference ...................................................................... 33
Chapter 9: One and Two Sample Hypothesis Tests ................................................................ 36
Chapter 10: Chi-Square Test and Analysis of Variance .......................................................... 39
Chapter 11: Correlation and Bivariate Regression Analysis ................................................... 42
Chapter 12: Multiple Regression ............................................................................................. 47
Chapter 13: Analyzing Regression Residuals ......................................................................... 50
Chapter 14: Logistic Regression ............................................................................................. 53
Chapter 15: Conducting Your Own Political Analysis ........................................................... 56
Version 1.2. Last updated December 5, 2024
1
, Introduction: Studying Politics Scientifically (No exercises)
There are no end-of-chapter exercises in the Introduction (Studying Politics Scientifically).
Chapter 1: The Definition and Measurement of Concepts
Exercise 1-1.
(A) Some possibilities for completing the inventory of contrasting characteristics:
• Frequently / never attends religious services;
• believes / does not believe in supreme deity;
• says that religion is most important / least important aspect of their lives;
• frequently / never contributes time or money to a religious organization;
• frequently / never attends group discussions of religious texts;
• displays / does not display religious symbols or imagery;
• feels proud / does not feel proud of a faith tradition;
• feels / does not feel joy [sorrow] when something good [bad] happens to a fellow
believer.
(B) Students must justify choices in A. This may be easier to do for some attributes. For
example, when someone says that religion is the most (or least) important aspect of his/her
life, it is plausible to assume that the response is tapping into the concept of religiosity and
not any other concept. However, religious attendance may tap a generalized desire for a
sense of belonging to a group (frequent attendance), or the absence of that desire (no
attendance). In any event, the purpose of this part of the exercise is to encourage students to
think empirically and justify their choices.
(C) Example (using attendance): The concept of religiosity is defined as the extent to which
individuals exhibit the characteristic of frequently attending religious services.
Exercise 1-2.
(A) No, the conclusion is not supported. The ecological fallacy could be at work here. The
ecological fallacy occurs when the researcher analyze aggregate-level units of analysis (in
this case, states) to make cross-level inferences about individuals (low-income and high-
income individuals).
(B) Low-income individuals, who are less likely to vote than are high-income individuals, are
2
, more likely to reside in states with higher percentages of high-income individuals. When
one compares states having high percentages of low-income residents with states having
low percentages of low-income residents, one is also comparing states having higher
percentages of voters with states have lower percentages of voters.
Exercise 1-3.
(A) (i) It is likely that a sizeable number of gun owners would be less willing to tell female
interviewers that they own guns compared to male interviewers.
(ii) The social desirability effect is a type of response bias that occurs when human
subjects alter their responses to survey questions to appear more likeable to the
researcher. Because males tend to favor gun ownership more than females, gun
owners may be more willing to tell a male interviewer they own a gun compared to
a female because some subjects tell interviewers what they want to hear to be liked.
(B) (i) Systematic measurement error.
(ii) The social desirability effect would be a consistent and chronic distortion in the
measurement of gun ownership. The problem would cause researchers to
consistently underestimate the true proportion of gun ownership. It is not a random
measurement error because it doesn’t cause non-gun owners to say they own guns;
there’s a pattern to the measurement error.
Exercise 1-4.
(A) Yolanda’s measure has high validity and high reliability. It is valid because all of the
measurements are very close to the true value of 50. It is reliable because the differences
among the four measurements are very small.
(B) Xavier’s measure has high validity and low reliability. Is it valid because the measurements
are close to the true value of 50: one is on the true value; two are 2 units below; one is 4
units above. Compared with Yolanda’s measurements, Xavier’s are less reliable, because
the differences among the four measurements are larger, ranging from 48 to 54.
(C) Zelda’s measure has low validity and high reliability. It has low validity because it
underestimates true knowledge by 4-6 points. It is reliable because the measurements are
very consistent, within a 2-point range.
3
, (D) Warren’s measure has low validity and low reliability. It has low validity because none of
the measurements is at all close to the true value. It is unreliable because the measures vary
wildly in a 30-point range, from 40 to 70.
Exercise 1-5.
Following are several reasons “number of yard signs” is not a valid measure of “pre-election
support.” Because yard signs require public commitment by the sign displayers, yard signs
measure intensity of support among a handful of people, not breadth of support in the electorate.
Because some people display yard signs at the request of the candidate (not on the initiative of the
displayer), yard signs measure candidate initiative and/or campaign strategy, not voter sentiment.
If the number of yard signs is counted, a large family home will be counted as having the same
electoral support as a single person’s home. Some districts, such as dense urban districts and large
rural districts, don’t have a lot of homes with yards to display signs. Using the yard signs
measurement strategy, individuals who don’t own homes will not be counted at all.
Exercise 1-6.
(A) Reliable but not valid.
(B) Mutt Jeffley’s strategy for weighing his dog is reliable because it will yield consistent
results. Each time he steps on the scale holding his dog, the scale will report the same
weight (or very close to it). The measurements are not valid, however, because Mutt is
stepping on the scale with his dog, thereby adding his weight to the measurement.
(C) Mutt could weigh himself without the dog and subtract his weight from the combined
weight he obtains when he steps on the scale holding his dog.
Exercise 1-7.
(A) (i) This question has high face validity.
(ii) The question is intended to measure individual preferences on the trade-off between
creating jobs and protecting the environment. It does so by requiring respondents
to choose between these two values.1
1
Interestingly, this traditional ANES survey question was revised starting in 2012. Respondents are asked to choose
between: “Regulate business to protect the environment and create jobs” and “No regulation because it will not work
4
