Abstract
Textbooks offer instructors an opportunity to promote data visualization and statistical literacy throughout the sociology curriculum. In this study, we examined 463 data visualizations from 27 textbooks for Introduction to Sociology, Social Problems, and intermediate elective courses to illuminate the range of data visualizations and their use of statistical data and numerical variables. We find that textbooks rely on a narrow range of data visualizations (e.g., bar/column charts and tables), statistical data (e.g., percentages), and numerical variables (e.g., age). Introduction to Sociology textbooks used more data visualizations, and Social Problems textbooks incorporated more types of statistics. In contrast, intermediate-level textbooks presented more measures of central tendency. Overall, our results suggest textbooks could better integrate a broader range of figures, statistical data, and numerical variables. We conclude with resources for integrating data visualization in instruction.
The sociology curriculum should support student learning of research consumption skills, which include understanding how scientific knowledge is produced (Murtonen 2015), what makes sociological research different from journalistic accounts or commonsense explanations (Schutt, Blalock, and Wagenaar 1984), the differences between qualitative and quantitative approaches (Medley-Rath 2023), how to use scientific principles (Ferguson 2016; Ferguson and Carbonaro 2016), and how to calculate and interpret descriptive statistics (Ferguson 2016; Ferguson and Carbonaro 2016; Pike et al. 2017). Research consumption skills matter for everyday life (Beuving and de Vries 2020; Markham 1991; Small 2018) and careers (Ciabattari et al. 2018; Nind, Kilburn, and Luff 2015; Senter, Spalter-Roth, and Van Vooren 2015).
Teaching students how to read and interpret figures and tables (i.e., data visualization literacy) promotes research consumption skills, which depend on understanding basic statistics and variables. Therefore, we analyze 463 complete data visualizations presented in 27 textbooks for Introduction to Sociology, Social Problems, and intermediate elective sociology courses. This article describes the frequency and variety of data visualizations and the kinds of statistics and numerical variables these textbooks used. We conclude with resources for instructors to integrate data visualization literacy into their courses.
Data Visualization Literacy
Data visualization literacy is increasingly important (Friel, Curcio, and Bright 2001; Gray, Gerlitz, and Bounegru 2018; Maltese, Harsh, and Svetina 2015; Shreiner 2017). The sheer volume of data (i.e., big data) has grown because of open data (i.e., data that are freely available) and “datafication” (i.e., the growth of quantifying more aspects of everyday life, such as data from a fitness tracker; D’Ignazio 2019). Data visualization literacy helps people comprehend huge volumes of data (Berinato 2016) and research findings using small and large data sets.
Few studies in the scholarship of teaching and learning in sociology have evaluated data visualization literacy. For example, Wills and Atkinson (2007:256) argued that reading figures and tables empowers students “to make conclusions about data independently of another’s summary or commentary about those data” (see also Burdette and McLoughlin 2010). Wilder (2010) found that sociology alums who had read tables or graphs in their sociology coursework were more confident in their ability to interpret them when seen in the media. Others found that using data visualizations made sociological concepts “more powerful” to students (Seguin, Nierobisz, and Kozlowski 2017:148) and made understanding social patterns easier for students (Ploch and Hastings 1992).
To our knowledge, there is minimal research on textbook data visualizations in sociology or social science. Hall (2000) analyzed the placement of tables about poverty in 45 Introduction to Sociology textbooks and whether the tables were undifferentiated (i.e., only about poverty) or differentiated (i.e., about poverty and race or gender). In an analysis of high school social studies textbooks, Shreiner (2017:22) found that the most common data visualizations were “timelines, political or physical maps with temporal overlays, bar graphs, tables, and choropleth maps.” Moreover, these data visualizations were scaffolded: They increased in frequency, variety, and complexity at the high school level compared to middle- and elementary-level social studies textbooks. Given these findings, our research examined whether college sociology textbooks followed this pattern: Were data visualizations frequent and varied, and were they more frequent and varied in intermediate texts compared to introductory texts? 1
Data visualization literacy involves statistical literacy. Research, both within and outside of sociology (Condron, Becker, and Bzhetaj 2018; DeCesare 2007; Liu 2021; Lovekamp, Soboroff, and Gillespie 2017; Mistima Maat et al., 2022), demonstrates the barrier that statistics anxiety presents to the engagement and success of students in undergraduate courses. Previous exposure to statistical information, among other factors like self-confidence and perceived math skills, also relate to student anxiety (Condron et al. 2018). Thus, statistical information found in sociology textbooks, especially in introductory courses, is critical to promoting student learning and success with statistical information central to the profession (e.g., professional journal articles, applied sociology work).
Furthermore, Linneman (2021) analyzed the statistical techniques used in 15 sociology journals and found an expanding gap between these techniques and the skills students learn in undergraduate statistics courses. Linneman (2021:52) categorized articles as Level 1 when they could be read by “pretty much anyone [i.e., introductory students] . . . because they involve simple descriptive statistics, frequencies, and percentages.” Ten percent of quantitative articles published in 1990 fit this category, whereas only 4 percent did in 2019 (Linneman 2021). Sociology textbooks could help bridge this gap between statistical anxiety, research skills, and data literacy.
Hypotheses
Data Visualization Variety
With little scholarship on data visualizations in college-level sociology texts, our hypotheses follow the findings from Shreiner (2017):
Hypothesis 1: Textbooks will rely on a few types of data visualizations (e.g., bar charts).
Hypothesis 2: Intermediate elective textbooks will use more frequent and varied data visualizations than introductory textbooks.
Overall, we expected that textbooks would skew toward using fewer types of figures and that intermediate elective books would use more frequent (e.g., average number of visualizations per chapter) and varied figures (e.g., average number of types of figures per chapter) because they were for more advanced courses.
Statistical Data and Numerical Variables
Data visualizations in sociology textbooks use statistical data and numerical variables about social phenomena.
Hypothesis 3: Textbooks will rely on a few types of statistical data (e.g., mean, percentages) and numerical variables (e.g., age, income).
Hypothesis 4: Intermediate elective course textbooks will use a wider variety of statistical data and numerical variables in their data visualizations than introductory textbooks.
We expected to find percentages widely used across textbook data visualizations and a broader range of statistical data (i.e., measures of central tendency) and numerical variables used in intermediate elective textbooks.
Methods
Sample
Our research analyzed the data visualizations in 27 textbooks. The lead author used purposive convenience sampling to select books for Introduction to Sociology (N = 14), Social Problems (N = 5), and Race, Class, Gender, and Family courses (N = 8). The lead author used the textbooks’ Amazon sales ranking while ensuring that a range of publishers and authors were represented in the sample (for more information on sampling design, see Medley-Rath 2022). Therefore, we selected no more than two texts per publisher for each course. An author was included only once in the sample to increase the breadth of voices. Furthermore, two additional books were included that would not appear in the Amazon sales ranking: the open education resource from OpenStax and the self-published, The Sociology Experiment. Studies that rely solely on best-selling lists miss widely used books using different distribution models.
The lead author identified three chapters for each text to include in the sample. First, the introductory chapter was selected because this chapter should provide some evidence of what to expect in the rest of the text. For introductory textbooks, the marriage and family chapter and one chapter related to inequality (e.g., race, class, or gender) were included. This way, the sampled chapters from introductory texts covered the same subfields as the intermediate elective textbooks. Because the marriage and family books do not have specific chapters on marriage and family, two chapters on inequality were randomly selected for inclusion (for a list of sampled chapters, see Medley-Rath 2022).
We coded the sample’s nonphotographic visualizations (i.e., figures and tables). We coded the following information for each data visualization: type (e.g., column chart, multiline graph) and type of statistical or numerical information (e.g., percentage, rate, time). The lead author and undergraduate student coauthors independently coded a few of the same chapters to establish intercoder reliability. Then, each student coauthor coded a portion of the sample. The lead author reviewed their coding.
Results
There were 463 data visualizations in the sample using 22 distinct formats (see Table 1) and approximately 20 types of statistics or numerical variables (see Table 2). Table 1 shows that over one-quarter of the data visualizations were tables, including frequencies, vital statistics such as fertility rates, and summaries of key concepts or processes. When present, the types of statistics or numbers used in these visualizations also showed a clear modal category—percentages. Over one-half of the data in these visualizations were presented as percentages (33.50 percent) or time (17.04 percent; see Table 2). Of the eight remaining types of statistical data and numerical variables, the next highest types were income (7.93 percent) and frequencies (6.56 percent).
Types of Data Visualizations.
Tables included frequency tables, tables with multiple kinds of demographic statistics (e.g., fertility rate, life expectancy), and organizational (i.e., three paradigms summarized, stages of assimilation listed with major characteristics).
Other data visualizations included unclear visuals (a combined donut and pie chart, photos with explanations); a combination of multiple types of figures (line, columns, and dual axis); timelines; multiple multiline, bar, or column charts; distribution chart (i.e., bell curve); or occurred only once: bubble chart, pictograph, population pyramid, scatterplot, and sunburst.
Types of Statistical Data and/or Numerical Variables.
Not applicable included words, maps, pictures, or other nonnumerical information.
Other included dollars for monthly and annual budget, Consumer Price Index, household size, ranked order, poverty threshold odds likelihood ratio, life satisfaction score, median height in inches, median net worth, occupation prestige scores, ratio, and index of dissimilarity.
Rates were coded as rates if they were named as rates. We did not verify if the term rate was misused (e.g., referring to a poverty rate that is the percentage of people experiencing poverty).
Data Visualizations
Our first hypothesis, that textbooks will rely on a few types of data visualizations (e.g., bar charts), was supported. The most frequently used data visualizations were tables (27.43 percent; see Table 1). Tables present numerical information typically and summarized information written elsewhere in the text (e.g., the three paradigms). Bar graphs (7.56 percent), column charts (15.12 percent), stacked bar charts (1.08 percent), and stacked column charts (3.89 percent) made up just over one-quarter of visualizations (27.65 percent, accumulated). Textbooks ranged from using no bar/column charts to having 66.67 percent of their data visualizations as bar/column charts (see Figure 1). Figure 1 shows that over one-half of data visualizations across all three textbook types were bar/column charts and tables, revealing a narrow range of types across these texts.
Percentage of figures that are bar/column (including stacked) charts and tables per textbook.
Our second hypothesis, that intermediate textbooks will use more frequent and varied figures and tables than introductory textbooks, was partially supported. Examining the median and mode suggests that Introduction to Sociology textbooks use data visualizations more frequently than Social Problems and intermediate elective textbooks (see Table 3). Moreover, the means for Introduction to Sociology and intermediate elective texts are quite similar (means of 16.02 and 16.73 for frequency and 7.87 and 7.85 for variety). Without tests of statistical significance in these observed differences, they are more or less the same, although with rounding, the means differ by a full point.
Frequency and Variety of Data Visualizations by Textbooks.
Variety refers to the number of distinct types of data visualizations used.
Textbooks ranged from 5 to 40 data visualizations in the sampled chapters. All but three texts had 11 or more data visualizations. The three books with fewer than 11 included OpenStax and The Sociology Experiment texts, suggesting that having publisher backing may influence the frequency of figures and tables in textbooks.
Table 3 also shows variety. If Introduction to Sociology and Social Problems textbooks are combined, our second hypothesis is supported. However, Introduction to Sociology textbooks used a broader range of data visualizations than Social Problems or intermediate elective books. Introduction to Sociology textbooks used more figures and tables and drew on a wider range of figures; however, this hypothesis is no longer supported when Introduction to Sociology and Social Problems texts are considered separately. Introduction to Sociology textbooks emerge as having more frequent and more varied data visualizations.
Statistical Data and Numerical Variables
Our third hypothesis, that textbooks will rely on a few types of statistics and numerical variables, was supported. Overall, textbooks used a mean of 6.88 (SD = 1.27) statistics and numerical variables across their data visualizations (see Table 2 for types of statistical data and numerical variables used). As expected, a large share of statistical data in figures and tables were reported as percentages (33.50 percent). One unexpected finding was that textbooks reported frequencies without percentages, making comparison problematic because frequencies are unstandardized counts, whereas percentages consider the size of the category or sample.
Our fourth hypothesis, which stated that intermediate elective course textbooks would use a wider variety of statistics and numerical variables in their data visualizations than introductory textbooks, was not supported. Textbooks used a median of seven statistics or numerical variables (i.e., types of numbers). However, the mean number of statistics or numerical variables was lowest for Introductory Sociology textbooks (mean = 6.74) and highest for Social Problem textbooks (mean = 7.27); intermediate elective textbooks (mean = 7.09) were in between. Because measures of central tendency are important to sociology and introductory statistical concepts, we also looked at the use of measures of central tendency. Introduction to Sociology textbooks (mean = 0.86) used measures of central tendency the least, and intermediate elective courses used them more often (mean = 3.11) than Social Problems textbooks (mean = 1.50).
Discussion
Overall, our first hypothesis was supported. Textbooks relied on a few types of data visualizations (i.e., bar/column charts). Our second hypothesis had mixed support.
Intermediate elective textbooks used more frequent and varied data visualizations than introductory textbooks. Still, when Introduction to Sociology and Social Problems textbooks were disaggregated, Introduction to Sociology textbooks emerged as having more frequent and more varied data visualizations. Our third hypothesis was also supported. Textbooks relied on a few types of statistics and numerical variables. Our fourth hypothesis was not supported. Social Problems textbooks used a broader range of statistics and numerical variables. However, intermediate elective books used more averages, means, and medians.
We anticipated textbooks to rely on a limited range of data visualizations, statistics, and numerical variables, and the results confirm this. However, we also expected to find evidence of scaffolding of data visualization literacies in textbooks because this pattern exists in high school social studies texts (Shreiner 2017). If scaffolding existed, we should have observed evidence of this through differences in figures in tables from the introductory chapter with later chapters.
We did not expect to find that textbooks would report frequencies without percentages. Percentages with frequencies make comparison possible. In the absence of percentages, instructors should teach students how to calculate and explain why percentages are necessary. Frequencies usually do not provide enough information to make sense of the data.
We were surprised to find that few statistics reported in the sample were rates (6.76 percent) because rates are frequently reported in the news media. For example, the news media reported COVID-19 cases as rates. Rates make sense from a public health perspective but are less meaningful to the average citizen using this information—regardless of appropriateness—to make decisions about their own risk.
Textbooks should strive to use the most appropriate data visualization while presenting students with various formats. In our study, texts used tables, stacked and unstacked bar/column charts, and line and multiline graphs most frequently. In contrast, Shreiner’s (2017) study found that high school social science textbooks more frequently used timelines and maps. Shreiner’s (2017) sample did not include textbooks for sociology, which high schools offer infrequently. But her results suggest that students should be familiar with a wide range of data visualizations used by other social scientists, especially geospatial, temporal, and spatiotemporal visualizations.
Whereas social science high school textbooks (see Shreiner 2017) are standards-driven, there are no state or national standards to which college sociology courses must adhere. Therefore, sociology textbooks are market driven. If there is no demand for data visualization, textbook publishers and authors likely would not include them. Therefore, curriculum leaders should consider what kinds of data visualization, statistical data, and numerical variables students should be able to understand and use after completing introductory and intermediate-level sociology courses and before enrolling in statistics or research methods courses. Consider, for example, the limited use of maps in our sample (see Table 1). Maps can promote a global focus, enhance geospatial literacy, and potentially prepare students to learn how to use Geographic Information System Mapping software. In our sample, about half of the textbooks (48.15 percent) used a map, but more than half (58.82 percent) of the maps were found in three Introduction to Sociology textbooks. The textbooks that used the most maps were legacy textbooks, first published between 1977 and 1992, 2 suggesting that maps may be viewed as less pedagogically necessary by publishers, authors, and instructors compared to the past. The lack of maps in more recently established textbooks could be attributed to the accessibility of online maps, increasing expenses of printing multicolored textbooks images or licensing maps, or the U.S.-centric focus of American sociology textbooks.
Recommendations for Instructors
Instructors can use what is in the textbooks as a starting point for supporting data visualization skills among students. We recommend that instructors select figures and tables from their texts to explain, interpret, criticize, and discuss with students. Instructors could ask students to calculate means, medians, and modes from the data in figures and tables. Students could create new figures and tables using the data reported by their textbooks using Microsoft Excel or Google Sheets (see Kunicki et al. 2019). Transforming figures within this software is straightforward, so students can decide which data visualization communicates data most effectively.
Several resources exist to help instructors use data visualizations as part of their instruction. For example, The New York Times (2022) and American Statistical Association’s (2022) “What’s Going on in This Graph?” share a different figure weekly from the newspaper and questions to guide instruction. Instructors could assign data visualization articles from Socius for students to read and interpret. More advanced students could use the General Social Survey’s Data Explorer to create data visualizations.
Lastly, there are numerous resources in TRAILS: Teaching Resources and Innovation Library for Sociology. For example, one resource uses Google Maps (Ellis 2021), and another is on table reading (Medley-Rath 2014). Medley-Rath (2023) includes a compilation of relevant resources from TRAILS and Teaching Sociology.
Strengths, Limitations, and Future Research
A strength of our study is that it addresses an important gap in the teaching sociology literature, that of the pedagogy of data visualization. Sociological research uses data visualizations and information sources from everyday life (i.e., journalistic accounts). Directing attention toward data visualization proficiency will help students become better consumers of research and prepare them for upper-level coursework in sociology.
Another strength of this study is that we include textbooks across the undergraduate sociology curriculum (see also Medley-Rath 2022). Our sampling strategy, however, was also a limitation. Because of how textbooks were selected, research using a different sampling strategy could produce different results. However, we strove to have a sample that brought in more authors and publishers compared with a sample that relied on sales rank alone to reduce the chance of author or publisher nuances that could skew the results. A second limitation of our sampling strategy was our selection of chapters. Some topics may better lend themselves to data visualizations. Future research could examine how data availability for a topic (e.g., Does the U.S. census ask a question on the topic?) is related to the frequency of data visualizations.
Furthermore, we did not measure data visualization literacy among students and can only speculate that it likely has room for improvement. Future research should examine data visualization competency among students. Scholars could use examples from sociology textbooks and ask students to locate information and interpret textbook figures and tables.
Conclusion
Data visualization literacy is essential for sociology majors, careers, and everyday life. Sociology is poised to be a foundational discipline for enhancing this skill. Our analysis of 463 data visualizations from 27 textbooks for Introduction to Sociology, Social Problems, and intermediate elective courses suggests room for improvement in textbook figures and tables. Textbooks are a resource for instructors to promote data visualization and statistical proficiency. However, they were limited in their range of types of data visualizations (e.g., bar/column charts and tables) and statistical data (e.g., percentages) and were not scaffolded (i.e., overall, data visualizations, statistical data, and numerical variables were not more frequent and varied at the intermediate compared to introductory level). We call on curriculum leaders, textbook publishers and authors, and instructors to advance students’ data visualization skills more intentionally.
Footnotes
Acknowledgements
We thank undergraduate research assistant Rebekah Morgan, who helped prepare the data for analysis. We thank Kathleen S. Lowney and Catherine White Berheide for providing feedback on this manuscript. This research has been presented at the 2021 North Central Sociological Association, 2022 Midwest Sociological Society annual meetings, the 2022 Indiana University SoTL Conference, and the 2022 Illinois Sociological Association conference.
Editor’s Note
Reviewers for this manuscript were, in alphabetical order, Charles Seguin, Mary Senter, and Esther Wilder.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: An Indiana University Kokomo Grant in Aid funded this research.
