Common Errors in Statistics (and How to Avoid Them), 4th Edition

Chapter 10

Graphics

KISS—Keep It Simple, but Scientific.

—Emanuel Parzen [1990]

Getting information from a table is like extracting sunbeams from a cucumber.

—Farquhar and Farquhar [1891]

IS A GRAPH REALLY NECESSARY? Is it a better vehicle than a table for communicating information to the reader? How many dimensions do you really need to illustrate? Do you need to illustrate repeated information for several groups? How do you select from a list of competing choices? How do you know whether the graph is effectively communicating the desired information? Does your graph answer a particular question, and are the elements of the graph chosen for your audience?

Graphics should emphasize and highlight salient features of the underlying data, and should coherently summarize large quantities of information. Although graphics provide a break from dense prose, authors must not forget that these illustrations should be scientifically informative rather than decorative. In this chapter, we outline mistakes in selection, creation, and execution of graphics and then discuss improvements.

Graphical illustrations should be simple and pleasing to the eye, but motivation for their inclusion must remain scientific. In other words, we avoid having too many graphical features that are purely decorative while keeping a critical eye open for opportunities to enhance the scientific implications for the reader. Good graphical designs utilize a large proportion of the ink to communicate scientific information ...