ABSTRACT
Data visualization can involve opportunities for graphic abuse, especially with regard to use of axes, data selectivity, and measurement adjustment. Common abuses of axes inviting misinterpretation include (1) not starting the origin at 0, as the viewer normally assumes, but rather presenting only an attenuated range of the data so as to exaggerate effect sizes; or (2) starting the origin at 0, not using equal intervals as tick marks, but rather allowing different scales at different ranges of the axis, so as to exaggerate effect sizes for selected ranges. A third type of graphic abuse is to switch the X and Y axes. The informed viewer is accustomed to the horizontal X axis representing the causal variable and the vertical Y axis representing the effect variable. Reversing these, for example to obtain a steeper and more dramatic line, can promote misinterpretation rather than helpful data visualization.
