maize mpub9873254 in

    8  Deconstructing data visualizations: What every teen should know

    Emily wakes up to the sound of her cell phone alarm, thinking about what lies ahead for her at school as a high school senior. Her phone screen lights up with today’s celebrity news, texts from friends, and calendar reminders of school assignments. Before she puts her feet on the floor, she posts to her Facebook page and checks the day’s news, laden with images, graphs, and charts. In her first period environmental science class, the textbook requires her to decipher a chart depicting climate change in order to project the consequences of continued warming. During her free period she decides to research her father’s recently diagnosed heart disease, as she’s feeling anxious about the family discussion last evening. After school she’s in charge of reporting her lacrosse team’s stats to the school newspaper, then remembers she needs more acne medicine, and checks online reviews and data to see which one is most effective. Before she leaves school, friends reminded her to register to vote in the upcoming primary, but she is unconvinced that young voters can really make a difference, so she consults her state’s polling data.

    Each of these activities involves data analysis and reading data in visualized form, some of it rather sophisticated. Today’s information universe is vast, and there is too little time to scrutinize data, especially when it is embedded in colorful graphics with a convincing message. Data visualization has long been used by academic researchers to summarize their findings, but more recently publishers and commercial marketing firms have embraced the use of visualization to impart information in new ways that are attractive and space-efficient (Centaur Communications 2010). These graphics draw the reader in, provide an instant takeaway, and are easy to share on social media. In order to read these charts, graphs, and statistics, today’s teen needs a toolkit for data literacy that requires a critical eye and a keen understanding of the creator’s intentions.

    How do we move beyond a simple check of the sponsoring website that published the information? When data and graphics have different creators, how do we evaluate the infographic – holistically or piecemeal?

    As educators, we must model what it means to be a critical consumer of data visualizations and other visual representations of data while at the same time giving our students a few rules of thumb to help simplify the process in their everyday lives. We can promote healthy skepticism by helping students learn how to effectively question what they view. Students love to critique things; we can teach them a framework to do this constructively. By doing so, we can help them grapple with uncertainty and use their questions to spur further investigation. Evaluating data visualizations across the curriculum as well as providing opportunities for students to create their own visual depictions of real world data are important additions to 21st-century information literacy skills.

    You may be thinking, Why can’t I just leave data literacy to math and science teachers? We know post-secondary students will likely contribute to the creation of data in their discipline and must become proficient at content-specific data visualizations. The challenge in high school is to both scaffold for these advanced academic skills as well as to prepare students to intelligently consume visualized data in their everyday lives. To this end, analyzing graphs and charts “in the wild” teaches them to identify misleading or ambiguous representations. Surveys of college faculty reveal that while students master the use of chart- and map-making tools quickly, they do not know what makes a good visualization (Carlson and Johnston 2015).

    In a 2012 Pew study on teens and technology, media expert Sam Punnet observed, “[A]ll communications [in the teen world] must be short, visual, and distracting/entertaining” (in Anderson and Rainie 2012). Data visualizations fit this bill; they tell a story in a compact combination of words and images. Data are often embedded in infographics to lend credibility; after all, numbers do not lie ... do they? Complicated infographics containing data visualizations require students to evaluate graphical techniques. In a world where visual information is preferred by consumers and marketers alike, students need to leave high school able to read data visualizations as powerful sources of information across disciplines, both in and out of the classroom.

    How can you help students gain data literacy skills if the discipline you teach doesn’t typically include data? Keep an eye out for teaching materials, collect subject-relevant examples embedded in media coverage of current events, scrutinize your textbook with a fresh eye for data and charts, offer visuals as an alternative to textual information, and consult resources in this book for professional development in this area. In the meantime, we can identify some simple evaluation tactics to add to our toolkit. This chapter will help you think more like a designer, extend basic information literacy skills, and add some new heuristics so you can de-mystify data visualizations for your students.

    Think like a designer

    Data visualization expert Edward Tufte (2006) suggests that every designer makes moral and intellectual choices in the creation of an infographic, and in consuming them we must hold the creator(s) responsible for those decisions. Every visualization involves a series of intentional choices, and while we can never fully appreciate each, it is helpful to read all graphics with this in mind. Every visualization is a picture with a message; students must think in terms of the designer’s agenda or perspective, e.g., Why was this data, and the story it tells, depicted in a series of bar graphs rather than a pie chart, and why the use of bright, bold colors? As consumers, we must actively engage with each graphic; as teachers we need to model this in each discipline.

    Infographics and data visualizations represent a burgeoning new field of marketing. Companies spend 62% of their content marketing dollars on infographics, up 15% over a year ago (Content Marketing Institute 2016). An estimated 84% of Internet communication will be visual by 2018, according to Reuters (Lopes 2014). In the world of marketing, designers work with their clients to craft a visual message consistent with the company’s mission. Similarly, political campaigns depend on infographics to quickly communicate their track record to voters on issues like job creation. In this case, graphics are often designed to make the reader feel aligned with the candidate, and every detail — from the range and unit of data selected, to the color palette used, to the font type used for text labels — is carefully selected to support the message. A graph of the same data may look very different if produced by BarackObama.com rather than WashingtonPost.com. One site is intended to persuade you to agree with a candidate’s policies, and the other’s goal is primarily to inform from an objective standpoint. As readers, it’s our job to decipher the message, given the perspective of its creator(s), rather than to merely react to the image and accept the data visualization at face value. As teachers, we need to help students to develop this fluency.

    Extending information literacy skills

    Information literate students already learn how to evaluate text sources. They can leverage this knowledge to deconstruct infographic data visualizations. Using techniques similar to those employed for the evaluation of text sources (in which we prompt students to consider who created it, on what authority, over what period of time, and reflecting what point of view) teachers must consider the following when teaching about visual information:

    • Provide students with time to look and think about the message, and begin to generate their own questions. Unlike a purely textual source, there are many layers to visualized data, including text, graphics, data, proportion, and color. What is seen initially, changes with time and focus given to each element. When students generate their own questions about the data, they engage with the graphic.
    • Graphics, and the datasets within, may originate from different (or multiple) sources. Students need to interrogate how and when data was collected, analyzed, and ultimately displayed; the answers are essential to judge the veracity of an infographic.
    • Look for the source of the data, not the URL where the infographic may have been found. Basic website evaluation techniques usually include deconstructing URLs and investigating the domain. Such common practices give rise to black-and-white rules for students such as “All .gov sites can be trusted” or “.edu sites are all created by scholars.” However, where a data visualization is published (or re-published) on the web may be unrelated to who generated the data or the agenda they suggest. For example, while respected universities often produce datasets and infographics, they may be funded by an outside interest or be republished on blogs or online journals. “Following the money” may uncover the story behind the story.

    A few basic rules of thumb

    While mastering all the design and statistical principles necessary to evaluate visualizations is daunting, we can lead students through their deconstruction.

    A key question to ask is, “What is the story, and who are the storytellers?” Every infographic seeks to tell a story using images or icons and sparse text, laid out cleverly to convey a message. It is important to consider both your first impression (the thesis) as well as deeper questions about each element. Question the story: Are the claims clear, reasonable, and accurate?

    What can we ask about the graphs, charts, and images? How can we better understand the choice of visualization?

    What does the data really mean?

    Having examined the story and those telling it, we now turn our attention to the data itself. Americans often accept data at face value; we – and our students! – tend to accept numbers as irrefutable “facts.” When students don’t know much about the claim being made, it seems overwhelming to try to evaluate the data. Like a textual search, there is a need to build context in order to evaluate. The graphs and charts contained in an infographic reflect a constellation of choices, and this is where we must focus our evaluation. Page 199 provides some questions to discuss with students, and Appendix B provides a sample infographic accompanied by a sample student conversation using those questions.

    Numerical information

    When analyzing data, we can ask many questions of the numbers themselves. Here are some useful strategies to keep in mind:

    • Correlation of two sets of data should not be used to imply cause and effect. For example, the fact that autism and ice cream consumption both increase over some period of time does not imply that ice cream consumption causes autism. This may be the simplest yet one of the most valuable data literacy skills.
    • Can you quickly verify some benchmark statistics (Best 2013), e.g., population, distance, birth or death rates that can help provide context for the data and numbers in a visualization? Providing students with a short list of statistics related to the subject of the infographic can make this easier.
    • Do the numbers seem too big or too conveniently rounded? Big, round numbers are often inexact guesses (Best 2013).
    • What is being measured over what period of time? Are the units of measure consistent, or are different units used to describe a single phenomenon? Does it seem implausible based on what you know? Odd units or preposterous claims should be questioned.
    • Check the y-axis on the graph. We are used to seeing graphs in math class where the intersection of the x- and y-axes begin at zero. In the wild, this may not be the case, so students need to consciously check the labels on graphs, particularly the y-axis. A y-axis that does not begin at zero is not inherently incorrect – it could be that truncating the range of numbers is space-efficient or that there are no data points that correspond with numbers close to zero. On the other hand, graphics with non-traditional or unexpected y-axes can raise questions: Was the range of data selectively chosen, perhaps to cherry-pick a short-term trend to support an argument? Would you feel differently about the trend if a longer time period was depicted, or a wider range of numbers? Were the intervals selected to support the creator’s intended conclusion?
    • Be cautious when looking at percentages, percentile, and percent change. As noted in Chapter 1, these terms sound similar but have distinct meanings.

    Methods

    An infographic’s raw data and the organization or researchers that produced it may not always be included, but some additional searching may lead to these kinds of details.

    • Can you locate the sample size (expressed as n=x)?
    • Is the methodology for data collection available, especially information on how the sample was gathered? Was it randomized, meaning the researcher’s data collection accurately depicts a larger population or dataset? Beware of generalizations made using small samples (n<25). For more on sampling, see Chapter 1.
    • Are the intervals regularly or irregularly spaced?Logarithmic graphs are used to depict change over time, and feature an accelerated rate of change using unevenly spaced intervals. They can be useful when a few data points are skewed away from the bulk of the data or when depicting percent change (rather than absolute). Absolute graphs feature regularly-spaced intervals (the same distance is present between each). Beware of charts comparing data graphed on an absolute scale with those using a logarithmic scale.
    • Is the data cherry-picked? Are date ranges or time boundaries specifically chosen because they demonstrate a desired result? If you looked at a different or broader interval, would the data look very different? For more on intervals, see Chapter 6.
    • Beware of the word average. Many people assume that average means the arithmetic mean (adding up all numbers and dividing by the quantity of numbers), but the author may mean median or mode, which can distort the raw data. For more discussion on averages, see Chapter 1.
    • Ask questions of text and labels

      • Does the title convey a thesis or put forth an argument? Does this match your “read” of the data?
      • Can you spot emotionally-laden words: rhetoric (words like freedom-fighter or terrorist) and hyperbole (claims like the most, the largest, or the best) or is the language neutral and purely descriptive?
      • Do font size or color connote importance (size) or caution (red, orange)? Were either used to intentionally mislead? Read the fine print for important details that the creator may have tried to de-emphasize.
      • Are all values clearly labeled using consistent units? Do the labels clarify or confuse? Are the terms defined, e.g., the definition of family versus household when used to collect income data.

    Layout

    Helping students understand the organization and flow of elements in an infographic can aid in improving comprehension. Here are some questions you may ask of your students.

    • Is this the best way to visualize this data? If not, can you find it graphed another way or find the original dataset and lay it out differently? See Figure 2 for some common charts and when to use them (Maguire 2016).
    • What does the color choice tell us? How do the colors used make you feel? Red usually signals trouble, or at least emphasis. Are the colors communicating emotion that is supported by the data? Or distracts from or over-emphasizes it?
    • Is the graphic layout designed to emphasize or de-emphasize some of the data?
    • Does the relative size of elements/icons accurately represent the data? Respective icon size should reflect data not emphasis; the area of two-dimensional shapes should be proportional to values.

    For a sample conversation about layout, see Appendix C.

    Practice, practice, practice

    Short lessons, peppered throughout the high school curriculum, will make evaluation of visualized data a habit of mind. Plan collaborative lessons with other teachers to scaffold learning throughout the year, for example:

    • Conduct a whole-class evaluation of a single infographic; assign groups to evaluate one of the categories.
    • Display a lesson-relevant image as a warm-up exercise, coupled with a checklist to complete or a required post to an online class discussion. Introduce a small subset of the criteria shown in this chapter at a single sitting so that students gain evaluative experience and confidence over time.
    • Assign groups to evaluate one of a set of related infographics on a single topic and have each share what their evaluation revealed. Be sure to include “what’s missing?” as a prompt.
    • Include an infographic requirement to a research paper and require a critical annotation to explain why it was chosen as a source.
    • Post the chart in Appendix D and remind students that the selection of a visualization style can be critical in assisting readers’ comprehension.

    Conclusion

    Every opportunity we provide students to critically evaluate infographics improves their data literacy. Ultimately – and with practice – students will transfer the critical eye we require in the classroom to the myriad graphics they confront in their everyday lives.

    Resources

    Appendix A: Tips for integrating visualization analysis into your classroom

    Figure 1: The 4-Corners Game

    Figure 1: The 4-Corners Game

    Appendix B: Practice evaluation with the 5 W’s

    Appendix C: Unpacking visual decisions in an infographic

    Figure 3: August 2013 infographic by the United States Department of Education
                    showing the anticipated cost of raising a child from birth to age 18. Source:
                    https://www.cnpp.usda.gov/sites/default/files/expenditures_on_children_by_families/CRC2012InfoGraphic.pdf

    Figure 3: August 2013 infographic by the United States Department of Education showing the anticipated cost of raising a child from birth to age 18. Source: https://www.cnpp.usda.gov/sites/default/files/expenditures_on_children_by_families/CRC2012InfoGraphic.pdf

    Appendix D: Choosing effective visualization types

    Type of Chart

    Best for. . .

    Pie Chart

    Showing the relative proportion of variables as parts of a whole. Beware of many small “slices,” especially without labels. Pie charts make comparisons difficult.

    Bar Graph (horizontal or vertical)

    Illustrating discrete quantities of a number of outcomes. Must include labels and legends.

    Line Graph

    Showing change in a variable over time. Labels and units of measure must be clear.

    3D Graphs (any type)

    Nothing! 3D graphics are inherently difficult to interpret without error.