The Visual Manipulation and Misrepresentation of Data in Public Media

Visualizations can be an incredibly powerful tool for communicating data. Data visualizations can summarize large data sets into one view, allow for easy comparisons between variables, and show trends or relationships in data that cannot be seen by looking at the raw data. Empirical information and by extension data visualizations are often seen as objective and honest. Unfortunately, data visualizations are susceptible to errors that may make them misleading. When visualizations are made for public audiences that do not have the statistical training or subject matter expertise to identify misleading or misrepresented data, these errors can have very negative effects. There is a good deal of research on how best to create guidelines for creating or systems for evaluating data visualizations. Many of the existing guidelines have contradicting approaches to designing visuals or they stress that best practices depend on the context. The goal of this work is to define the guidelines for making visualizations in the context of a public audience and show how context-specific guidelines can be used to effectively evaluate and critique visualizations. The guidelines created here are a starting point to show that there is a need for best practices that are specific to public media. Data visualization for the public lies at the intersection of statistics, graphic design, journalism, cognitive science, and rhetoric. Because of this, future conversations to create guidelines should include representatives of all these fields.