[Cognitive structuring by statistical graphics: effects on understanding and retention].

We investigated how displaying the same set of data using different line graphs affects comprehension and retention of the data. In line graphs, the results of two-factor experiments can be displayed and structured in different ways by locating either one or the other independent variable on the abscissa. The experiment showed that the level of measurement as well as the number of levels of each independent variable affected comprehension and retention. Comprehension of graphs and retention of the information was better, if the graph showed an independent variable on the abscissa which was measured using an interval scale rather than a nominal scale, and when there were four rather than two levels. To account for the results, we propose a process model of graph comprehension.