June 14, 2009
June 14, 2009
June 17, 2009
Educational Research and Methods
14.1009.1 - 14.1009.17
Reasoning About Categorical Data: Multiway Plots as Useful Research Tools
Key words: categorical data display, multiway plot, research methods
In this paper, we help our audience learn to create and interpret “multiway plots”—a powerful tool for exploring and presenting categorical data. We use eighth-semester undergraduate persistence data as a case study of how multiway plots are used, but do not explore a particular research question. Instead, our goal is to disseminate a powerful, yet underutilized, research tool that facilitates an iterative process of reasoning about one’s categorical data: design the display/reason about the data/redesign the display/reason about the data etc., until the logic of one’s display is consistent with the logic of one’s analysis. Our case study begins with familiar column graphs and bar graphs typically used in engineering education journals. We then show how the same information is transformed into a multiway plot. Each step in the transformation is illustrated and explained. The results are two very different visualizations of the same data: clustered-column or bar graph (“before”) and multiway plot (“after”). Specific elements of the before and after graphs are highlighted to let the reader experience the perceptual advantages and to assess the utility of multiway plots in drawing meaningful conclusions from categorical data. We also alert our audience to the technical issues involved in creating multiway plots including software resources. Through this work, we hope to raise the awareness of the engineering education community of the benefits of multiway plots for visualizing, exploring, and presenting categorical data. In doing so, we hope to contribute to the continued enhancement of research quality in our discipline.
Multiway plots are powerful tools for exploring and presenting categorical data. Developed by statistician William Cleveland1 based on work on human perception of quantitative data, multiway plots are respected by experts in data presentation such as Naomi Robbins2 and are consistent with graphical design principles advocated by Edward Tufte.3 However, a quick review of journals in the field of engineering education reveals that column and bar charts are the dominant form of data display. Our goal in this paper is to bring the form and function of the multiway plot to the attention of the engineering education community. We illustrate the advantages of this type of display, compared to clustered-column charts, for interpreting and presenting categorical data.
Categorical data is regularly encountered in engineering education research. For example, in studying the effects of gender and race on undergraduate student pathways, we might categorize students by sex, race, and their academic major in the eighth semester. Each of these categories has two or more levels that are mutually exclusive. For example, the category “Sex” has the levels Female and Male; the category “Race” might include the levels Asian, Black, Hispanic, and others; the category “Eighth-Semester Major” might have the levels Arts and Humanities, Business, Computer Science, Engineering, and others.
Layton, R., & Lord, S., & Ohland, M. (2009, June), Reasoning About Categorical Data: Multiway Plots As Useful Research Tools Paper presented at 2009 Annual Conference & Exposition, Austin, Texas. https://peer.asee.org/4683
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2009 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015