The second grand principle that Edward Tufte espouses is to "show causality" in visuals. Tufte illustrates his point using a poster of Napolean's march in 1812. The poster shows multiple variables interacting and effectively links the military loss to temperature, not his opponents. As temperatures plummet the armies of Napolean are decimated. Tufte argues that this visual effectively shows causality.
For our purposes we need to contrast visuals used for argument and visual displays used for exploratory purposes. In data analysis for data-driven decision making the analyst is not “showing causality”, but instead wants a visual display of data that allows the data to talk back. Student achievement data (when arrayed with a comparative dataset) do not include causality, but are useful in describing reality. In analyzing student achievement data we are not juxtaposing data that result in cause, but rather looking for patterns. For example, a teacher might notice that all her students performed lower than average on vocabulary on the recent district benchmark assessment when compared to the rest of the students in the school. Nothing about that visual shows causality, but it is an elegant and appropriate display to begin exploring the reality of student performance. Thus, for exploratory data analysis “show causality” is not a necessary condition for a high-quality visual display.
When teachers and administrators begin attributing cause for an observed pattern, Tufte’s principle of “showing cause” is a necessary condition for a high-quality visual.
Again, excuse my scribbles. They lack the design and attractiveness that i would prefer, but illustrate the point nonetheless.