Exercises

Clarifying the Concepts

Question 3.1

3.1

What are the five techniques discussed in this chapter for misleading with graphs?

Question 3.2

3.2

What are the steps to create a scatterplot?

Question 3.3

3.3

How can you convert a scatterplot into a range-frame?

Question 3.4

3.4

What does it mean for two variables to be linearly related?

Question 3.5

3.5

How can we tell whether two variables are linearly or nonlinearly related?

Question 3.6

3.6

What is the difference between a line graph and a time plot?

Question 3.7

3.7

What is the difference between a bar graph and a Pareto chart?

Question 3.8

3.8

Bar graphs and histograms look very similar. In your own words, what is the difference between the two?

Question 3.9

3.9

What are pictorial graphs and pie charts?

Question 3.10

3.10

Why are bar graphs preferred over pictorial graphs and pie charts?

Question 3.11

3.11

Why is it important to identify the independent variable and the dependent variable before creating a visual display?

Question 3.12

3.12

Under what circumstances would the x-axis and y-axis not start at 0?

Question 3.13

3.13

Chartjunk comes in many forms. What specifically are moiré vibrations, grids, and ducks?

Question 3.14

3.14

Geographic information systems (GIS), such as those provided by computerized graphing technologies, are particularly powerful tools for answering what kinds of research questions?

Question 3.15

3.15

How is a bubble graph similar to a traditional scatterplot?

Question 3.16

3.16

How does a bubble graph differ from a traditional scatterplot?

Calculating the Statistics

Question 3.17

3.17

Alumni giving rates, calculated as the total dollars donated per year from 2006 to 2016, represent which kind of variable—nominal, ordinal, or scale? What would be an appropriate graph to depict these data?

Question 3.18

3.18

Alumni giving rates for a number of universities, calculated as the number of alumni who donated and the number who did not donate in a given year, represent which kind of variable—nominal, ordinal, or scale? What would be an appropriate graph to depict these data?

Question 3.19

3.19

You are exploring the relation between gender and video game performance, as measured by final scores on a game.

  1. In this study, what are the independent and dependent variables?

  2. Is gender a nominal, ordinal, or scale variable?

  3. Is final score a nominal, ordinal, or scale variable?

  4. Which graph or graphs would be appropriate to depict the data? Explain why.

69

Question 3.20

3.20

Do the data in the graph below show a linear relation, a nonlinear relation, or no relation? Explain.

image

Question 3.21

3.21

Do the data in the graph below show a linear relation, a nonlinear relation, or no relation? Explain.

image

Question 3.22

3.22

What elements are missing from the graphs in Exercises 3.20 and 3.21?

Question 3.23

3.23

The following figure presents the enrollment of graduate students at a university, across six fall terms, as a percentage of the total student population.

image
  1. What kind of visual display is this?

  2. What other type of visual display could have been used?

  3. What is missing from the axes?

  4. What chartjunk is present?

  5. Using this graph, estimate graduate student enrollment, as a percentage of the total student population, in the fall terms of 2007, 2008, and 2010.

  6. How would the comparisons between bars change if the y-axis started at 0?

Question 3.24

3.24

When creating a graph, we need to make a decision about the numbering of the axes. If you had the following range of data for one variable, how might you label the relevant axis?

337 280 279 311 294 301 342 273

Question 3.25

3.25

If you had the following range of data for one variable, how might you label the relevant axis?

0.10 0.31 0.27 0.04 0.09 0.22 0.36 0.18

Question 3.26

3.26

The scatterplot in How It Works 3.1 depicts the relation between fertility and life expectancy. Each dot represents a country (or, as in the case of Hong Kong, a special administrative region).

  1. Approximately, what is the highest life expectancy in years? Approximately, what fertility rate (children per woman) is associated with the highest life expectancy?

  2. Does this seem to be a linear relation? Explain why or why not, and explain the relation in plain English.

Question 3.27

3.27

Based on the data in the bubble graph in Figure 3-18, what is the relation between physical health and positive emotions?

Question 3.28

3.28

The colors and sizes of the bubbles in the bubble graph in Figure 3-18 represent the gross domestic product (GDP) for each country. Using this information, explain what the relation is between positive emotions and GDP.

Applying the Concepts

Question 3.29

3.29

Type of graph for the relation between height and attractiveness: A social psychologist studied the effect of height on perceived overall attractiveness. Students were recruited to come to a research laboratory in pairs. The pairs sat together in the waiting room for several minutes and then were brought to separate rooms, where their heights were measured. They also filled out a questionnaire that asked, among other things, that they rate the attractiveness of the person who had been sitting with them in the waiting room on a scale of 1 to 10.

70

  1. In this study, are the independent and dependent variables nominal, ordinal, or scale?

  2. Which graph or graphs would be most appropriate to depict the data? Explain why.

  3. If height ranged from 58 inches to 71 inches in this study, would the axis start at 0? Explain.

Question 3.30

3.30

Type of graph for the effects of cognitive-behavioral therapy on depression: A social worker tracked the depression levels of clients being treated with cognitive-behavioral therapy for depression. For each client, depression was assessed at weeks 1 to 20 of therapy. She calculated a mean for all of her clients at week 1, week 2, and so on, all the way through week 20.

  1. What are the independent and dependent variables in this study?

  2. Are the variables nominal, ordinal, or scale?

  3. Which graph or graphs would be most appropriate to depict the data? Explain why.

Question 3.31

3.31

Type of graph for comparative suicide rates: The World Health Organization tracks suicide rates by gender across countries. For example, in 2011, the rate of suicide per 100,000 men was 17.3 in Canada, 17.7 in the United States, 44.6 in Sri Lanka, 53.9 in the Russian Federation, 1.4 in South Africa, and 2.5 in the Philippines.

  1. What are the variables in this study?

  2. Are the variables nominal, ordinal, or scale?

  3. Which graph would be most appropriate to depict the data? Explain why.

  4. If you wanted to track the suicide rates for three of these countries over 50 years, what type of graph might you use to show these data?

Question 3.32

3.32

Scatterplot of daily cycling distances and type of climb: Every summer, the touring company America by Bicycle conducts the “Cross Country Challenge,” a 7-week bicycle journey across the United States from San Francisco, California, to Portsmouth, New Hampshire. At some point during the trip, the exhausted cyclists usually start to complain that the organizers are purposely planning for days with lots of hill and mountain climbing to coincide with longer distances. The tour staff counter that no relation exists between climbs and mileage and that the route is organized based on practical issues, such as the location of towns in which riders can stay. The organizers who planned the route (these are the company owners who are not on the tour) say that they actually tried to reduce the mileage on the days with the worst climbs. Here are the approximate daily mileages and climbs (in vertical feet), as estimated from one rider’s bicycle computer.

Mileage Climb Mileage Climb Mileage Climb
83 600 69 2500 102 2600
57 600 63 5100 103 1000
51 2000 66 4200 80 1000
76 8500 96 900 72 900
51 4600 124 600 68 900
91 800 104 600 107 1900
73 1000 52 1300 105 4000
55 2000 85 600 90 1600
72 2500 64 300 87 1100
108 3900 65 300 94 4000
118 300 108 4200 64 1500
65 1800 97 3500 84 1500
76 4100 91 3500 70 1500
66 1200 82 4500 80 5200
97 3200 77 1000 63 5200
92 3900 53 2500
  1. Construct a scatterplot of the cycling data, putting mileage on the x-axis. Be sure to label everything and include a title.

  2. We haven’t yet learned to calculate inferential statistics on these data, so we can’t estimate what’s really going on, but do you think that the amount of vertical climb is related to a day’s mileage? If yes, explain the relation in your own words. If no, explain why you think there is no relation.

  3. It turns out that inferential statistics do not support the existence of a relation between these variables and that the staff seems to be the most accurate in their appraisal. Why do you think the cyclists and organizers are wrong in opposite directions? What does this say about people’s biases and the need for data?

Question 3.33

3.33

Scatterplot of gross domestic product and education levels: The Group of Eight (G8) consists of many of the major world economic powers. It meets annually to discuss pressing world problems. In 2013, for example, the G8 met in the United Kingdom with an agenda that included international security. Decisions made by G8 nations can have a global impact; in fact, the eight nations that make up the membership reportedly account for almost two-thirds of the world’s economic output. Here are data for seven of the eight G8 nations for gross domestic product (GDP) in 2004 and a measure of education. The measure of education is the percentage of the population between the ages of 25 and 64 that had at least one university degree (Sherman, Honegger, & McGivern, 2003). No data point for education in Russia was available, so Russia is not included.

Country GDP (in trillions of $US) Percentage with University Degree
Canada 0.98 19
France 2.00 11
Germany 2.71 13
Italy 1.67 9
Japan 4.62 18
United Kingdom 2.14 17
United States 11.67 27

71

  1. Create a scatterplot of these data, with university degree on the x-axis. Be sure to label everything and to give it a title. Later, we’ll use statistical tools to determine the equation for the line of best fit. For now, draw a line of best fit that represents your best guess as to where it would go.

  2. In your own words, describe the relation between the variables that you see in the scatterplot.

  3. Education is on the x-axis, indicating that education is the independent variable. Explain why it is possible that education predicts GDP. Now reverse your explanation of the direction of prediction, explaining why it is possible that GDP predicts education.

Question 3.34

3.34

Time series plot of organ donations: The Canadian Institute for Health Information (CIHI) is a nonprofit organization that compiles data from a range of institutions—from governmental organizations to hospitals to universities. Among the many topics that interest public health specialists is the problem of low levels of organ donation. Medical advances have led to ever-increasing rates of transplantation, but organ donation has not kept up with doctors’ ability to perform more sophisticated and more complicated surgeries. Data reported by CIHI (https://secure.cihi.ca/free_products/2011_CORR_Annua_Report_EN.pdf, 2015) provide Canadian transplantation and donation rates for 2001–2010. Here are the donor rates per million in the population; these numbers include only deceased donors.

Year Donor Rate per Million People Year Donor Rate per Million People
2001 13.4 2006 14.1
2002 12.9 2007 14.7
2003 13.3 2008 14.4
2004 12.9 2009 14.4
2005 12.7 2010 13.6
  1. Construct a time series plot from these data. Be sure to label and title your graph.

  2. What story are these data telling?

  3. If you worked in public health and were studying the likelihood that families would agree to donate after a loved one’s death, what research question might you ask about the possible reasons for the trend suggested by these data?

Question 3.35

3.35

Bar graph of alumni donation rates at different kinds of universities: U.S. universities are concerned with increasing the percentage of alumni who donate to the school because alumni donation rate is a factor in the U.S. News & World Report’s university rankings. What role might type of university play in alumni donation rates? U.S. News & World Report ranks private national universities, public national universities, and private liberal arts colleges (http://www.usnews.com/education/best-colleges/the-short-list-college/articles/2015/10/20/10-universities-where-the-most-alumni-donate). National universities tend to focus on graduate education and research, whereas liberal arts colleges tend to focus on undergraduate education. Here are recent alumni donation rates for the top 10 schools on this variable. With the exception of Princeton, a private national university, all of these institutions are private liberal arts colleges.

Top 10 Colleges and Universities in Terms of Alumni Donation Alumni Donation Rates (%)
Princeton University (NJ) 62.6
Thomas Aquinas College (CA) 58.9
Carleton College (MN) 58.2
Williams College (MA) 57.5
Amherst College (MA) 57.3
Centre College (KY) 54.9
Middlebury College (VT) 54.7
Davidson College (NC) 53.4
College of the Holy Cross (MA) 51.1
Bowdoin College (ME) 50.2
  1. What is the independent variable in this example? Is it nominal or scale? If nominal, what are the levels? If scale, what are the units and what are the minimum and maximum values?

  2. What is the dependent variable in this example? Is it nominal or scale? If nominal, what are the levels? If scale, what are the units and what are the minimum and maximum values?

  3. Construct a bar graph of these data, with one bar for each of the top 10 schools, using the default options in your computer software.

  4. Construct a second bar graph of these data, but change the defaults to satisfy the guidelines for graphs discussed in this chapter. Aim for simplicity and clarity.

    72

  5. Cite at least one research question that you might want to explore next if you worked for one of these universities. Your research question should grow out of these data.

  6. Explain how these data could be presented as a pictorial graph. (Note that you do not have to construct such a graph.) What kind of picture could you use? What would it look like?

  7. What are the potential pitfalls of a pictorial graph? Why is a bar chart usually a better choice?

Question 3.36

3.36

Bar graph versus Pareto chart of countries’ gross domestic product: In How It Works 3.2, we created a bar graph for the 2012 GDP, in U.S. dollars per capita, for each of the G8 nations. In How It Works 3.3, we created a Pareto chart for these same data.

  1. Explain the difference between a Pareto chart and a standard bar graph in which the countries are in alphabetical order along the x-axis.

  2. What is the benefit of the Pareto chart over the standard bar graph?

Question 3.37

3.37

Bar graph versus time series plot of graduate school mentoring: Johnson, Koch, Fallow, and Huwe (2000) conducted a study of mentoring in two types of psychology doctoral programs: experimental and clinical. Students who graduated from the two types of programs were asked whether they had a faculty mentor while in graduate school. In response, 48.00% of clinical psychology students who graduated between 1945 and 1950 and 62.31% who graduated between 1996 and 1998 reported having had a mentor; 78.26% of experimental psychology students who graduated between 1945 and 1950 and 78.79% who graduated between 1996 and 1998 reported having had a mentor.

  1. What are the two independent variables in this study, and what are their levels?

  2. What is the dependent variable?

  3. Create a bar graph that depicts the percentages for the two independent variables simultaneously.

  4. What story is this graph telling us?

  5. Was this a true experiment? Explain your answer.

  6. Why would a time series plot be inappropriate for these data? What would a time series plot suggest about the mentoring trend for clinical psychology graduate students and for experimental psychology graduate students?

  7. For four time points—1945–1950, 1965, 1985, and 1996–1998—the mentoring rates for clinical psychology graduate students were 48.00, 56.63, 47.50, and 62.31, respectively. For experimental psychology graduate students, the rates were 78.26, 57.14, 57.14, and 78.79, respectively. How does the story we see here conflict with the one that we developed based on just two time points?

Question 3.38

3.38

Bar graph versus pie chart of student participation in community activities: The National Survey on Student Engagement (NSSE) has surveyed more than 400,000 students—freshmen and seniors—at 730 U.S. schools since 1999 (“America’s Best Colleges 2004,” 2003). Among the many questions on the NSSE, students were asked how often they “participated in a community-based project as part of a regular course.” For the students at the 19 institutions classified as national universities that made their data publicly available through the U.S. News & World Report Web site, here are the data: never, 56%; sometimes, 31%; often, 9%; very often, 5%. (The percentages add up to 101% because of rounding decisions.) Explain why a bar graph would be more suitable for these data than a pie chart.

Question 3.39

3.39

Software defaults of graphing programs that portray satisfaction with graduate advisors: The 2000 National Doctoral Program Survey asked 32,000 current and recent Ph.D. students in the United States, across all disciplines, to respond to the statement “I am satisfied with my advisor.” The researchers calculated the percentage of students who responded “agree” or “strongly agree”: current students, 87%; recent graduates, 86%; former students who left without completing the Ph.D., 48%.

  1. Use a software program that produces graphs (e.g., Excel, SPSS, Minitab) to create a bar graph for these data.

  2. Play with the options available to you. List aspects of the bar graph that you are able to change to make your graph meet the guidelines listed in this chapter. Be specific, and include the revised graph.

Question 3.40

3.40

Software defaults of graphing programs for depicting the “world’s deepest” trash bin: The car company Volkswagen has sponsored a “fun theory” campaign in recent years in which ordinary behaviors are given game-like incentives to promote pro-social behaviors such as recycling or obeying the speed limit (http://www.thefuntheory.com/). In one example, Volkswagen created a seemingly super-deep trash bin; when a person would throw something out, a highpitched whistling sound played for 7 seconds, culminating in a far-off-sounding boom, as if the item had fallen for hundreds of feet! The fun-theory people collected data; in a single day, 72 kilograms of trash were thrown out in their “deep” bin, versus just 31 kilograms in a nearby trash bin.

73

  1. Use a software program that produces graphs (e.g., Excel, SPSS, Minitab) to create a bar graph for these data.

  2. Play with the options available to you. List aspects of the bar graph that you are able to change to make your graph meet the guidelines listed in this chapter. Be specific, and include the revised graph.

Question 3.41

3.41

Multivariable graphs and college rankings by academics and sexiness: Buzzfeed.com published a multivariable graph that purported to rank colleges by academics and “hotness.” The data from this graph are represented here.

  1. What kind of graph is this? Explain.

  2. List the variables that are included in this graph.

  3. List at least three ways in which this graph could be redesigned in light of the guidelines in this chapter.

image

Question 3.42

3.42

Types of graph appropriate for behavioral science research: Give an example of a study—real or hypothetical—in the behavioral sciences for which the researchers could use each type of graph. State the independent variable(s) and dependent variable, including levels for any nominal variables.

  1. Frequency polygon

  2. Line graph (line of best fit)

  3. Bar graph (one independent variable)

  4. Scatterplot

  5. Time series plot

  6. Pie chart

  7. Bar graph (two independent variables)

Question 3.43

3.43

Creating the perfect graph: What advice would you give to the creators of each of the following graphs? Consider the basic guidelines for a clear graph, for avoiding chartjunk, and regarding the ways to mislead through statistics. Give three pieces of advice for each graph. Be specific—don’t just say that there is chartjunk; say exactly what you would change.

  1. Business and women:

    image
  2. Workforce participation:

    74

    image

Question 3.44

3.44

Graphs in the popular media: Find an article in the popular media (newspaper, magazine, Web site) that includes a graph in addition to the text.

  1. Briefly summarize the main point of the article and graph.

  2. What are the independent and dependent variables depicted in the graph? What kind of variables are they? If nominal, what are the levels?

  3. What descriptive statistics are included in the article or on the graph?

  4. In one or two sentences, what story is the graph (rather than the article) trying to tell?

  5. How well do the text and graph match up? Are they telling the same story? Are they using the same terms? Explain.

  6. Write a paragraph to the graph’s creator with advice for improving the graph. Be specific, citing the guidelines from this chapter.

  7. Redo the graph, either by hand or by computer, in line with your suggestions.

Question 3.45

3.45

Interpreting a graph about two kinds of career regrets: The Yerkes–Dodson graph demonstrates that graphs can be used to describe theoretical relations that can be tested. In a study that could be applied to the career decisions made during college, Gilovich and Medvec (1995) identified two types of regrets—regrets of action and regrets of inaction—and proposed that their intensity changes over time. You can think of these as Type I regrets—things you have done that you wish you had not done (regrets of action)—and Type II regrets—things you have not done that you wish you had done (regrets of inaction). The researchers suggested a theoretical relation between the variables that might look something like the graph below.

image
  1. Briefly summarize the theoretical relations proposed by the graph.

  2. What are the independent and dependent variables depicted in the graph? What kind of variables are they? If nominal or ordinal, what are the levels?

  3. What descriptive statistics are included in the text or on the graph?

  4. In one or two sentences, what story is the graph trying to tell?

Question 3.46

3.46

Thinking critically about a graph of the frequency of psychology degrees: The American Psychological Association (APA) compiles many statistics about training and careers in the field of psychology. The accompanying graph tracks the number of bachelor’s, master’s, and doctoral degrees conferred between the years 1970 and 2000.

image

75

  1. What kind of graph is this? Why did the researchers choose this type of graph?

  2. Briefly summarize the overall story being told by this graph.

  3. What are the independent and dependent variables depicted in the graph? What kind of variables are they? If nominal or ordinal, what are the levels?

  4. List at least three things that the graph creators did well (that is, in line with the guidelines for graph construction).

  5. List at least one thing that the graph creators should have done differently (that is, not in line with the guidelines for graph construction).

  6. Name at least one variable other than number that might be used to track the prevalence of psychology bachelor’s, master’s, and doctoral degrees over time.

  7. The increase in bachelor’s degrees over the years is not matched by an increase in doctoral degrees. List at least one research question that this finding suggests to you.

Question 3.47

3.47

Thinking critically about a graph about international students: Researchers surveyed Canadian students on their perceptions of the globalization of their campuses (Lambert & Usher, 2013). The 13,000 participants were domestic undergraduate and graduate students—that is, they were not recently from countries outside of Canada. The pie chart here shows the responses to one survey item: “The increasing number of international students attending my institution has led to improvements in the university’s reputation and image.”

image
  1. What is the story that these data are telling?

  2. Why would a bar graph of these data tell this story better than a pie chart does?

  3. Create a bar graph of these data, keeping the bars in the order of the labels here—from “Strongly disagree” on the left to “Not sure/Don’t know/N/A” on the right.

  4. Why would it not make sense to create a Pareto chart in this case?

Question 3.48

3.48

Interpreting a graph about traffic flow: Go to http://maps.google.com/. On a map of your country, click on the traffic button in the upper left corner.

  1. How is the density and flow of traffic represented in this graph?

  2. Describe traffic patterns in different regions of your country.

  3. What are the benefits of this interactive graph?

Question 3.49

3.49

Interpreting a graph about political support: As we learned in this chapter, pie charts are often not the best choice to present data.

  1. What is the primary flaw in this presentation of data as a pie chart?

  2. How would you redesign this graph? Be specific and cite at least three ways in which you would change it.

image

Question 3.50

3.50

Multivariable graphs and shipwrecks: The cliché “women and children first” originated in part because of the Titanic captain’s famous directive as his ship was sinking. Yet, the cliché is not grounded in reality. See the multivariable graph on page 76.

  1. What kind of graph is this?

  2. List the variables that are included in this graph.

  3. Are the data on the top and the bottom necessary to tell this story? Explain. Why do you think both sides are included?

  4. In your own words, tell the story that this graph is visualizing.

76

image

Putting It All Together

Question 3.51

3.51

Type of graph describing the effect of romantic songs on ratings of attractiveness: Guéguen, Jacob, and Lamy (2010) wondered if listening to romantic songs would affect the dating behavior of the French heterosexual women who participated in their study. The women were randomly assigned to listen to either a romantic song (“Je l’aime à Mourir,” or “I Love Her to Death”) or a nonromantic song (“L’heure du Thé,” or “Tea Time”) while waiting for the study to begin. Later in the study, an attractive male researcher asked each participant for her phone number. Of the women who listened to the romantic song, 52.2% gave their phone number to the researcher, whereas only 27.9% of the women who listened to the nonromantic song gave their phone number to the researcher.

  1. What is the independent variable in this study?

  2. What is the dependent variable?

  3. Is this a between-groups or a within-groups study? Explain your answer.

  4. Think back to our discussion of variables in Chapter 1. How did the researcher operationalize “dating behavior” in this study? Do you think this is a valid measure of dating behavior? Explain your answer.

  5. What is the best type of graph to depict these results? Explain your answer.

  6. Create the graph you described in part (e) using software without changing any of the defaults.

  7. Create the graph you described in part (e) a second time, changing the defaults to meet the criteria in the checklist introduced in this chapter.

Question 3.52

3.52

Developing research questions from graphs: Graphs not only answer research questions, but can spur new ones. Figure 3-9 on page 55 depicts the pattern of changing attitudes, as expressed through Twitter.

  1. On what day and at what time is the highest average positive attitude expressed?

  2. On what day and at what time is the lowest average negative attitude expressed?

  3. What research question(s) do these observations suggest to you with respect to weekdays versus weekends? (Remember, 0 on Sunday is midnight—or late Saturday night.) Name the independent and dependent variables.

  4. How do the researchers operationalize mood? Do you think this is a valid measure of mood? Explain your answer.

  5. One of the highest average negative attitudes occurs at midnight on Sunday. How does this fit with the research hypothesis you developed in part (c)? Does this suggest a new research hypothesis?