Understanding the effectiveness of computer graphics for decision support: a cumulative experimental approach

Reviewer: Philip Barnard

Under what conditions should information be presented in graphs rather than tables__?__ Answers to this question are of key importance to the designers and users of decision support systems. Experimental work has, as yet, provided little in the way of clear answers. This paper contributes to the debate by reporting three experiments that examined the effectiveness of graphs and tables as aids to decision making. In these experiments, the authors vary the nature of the decision making tasks. With a simple task, graphical and tabular formats for presenting information resulted in similar levels of user performance. With a more complex task, graphical presentation proved superior. For tasks of even greater complexity, users of graphs marginally outperformed users of tables, but only under conditions of high information load. In highlighting the importance of task conditions, this research makes a valuable contribution. The effectiveness of alternative formats for information presentation clearly depends upon what the information is being used for. However, in order to understand how information is being used, it is necessary to relate the precise structure and content of tables and graphs to the nature of the decisions being made. Without a penetrating analysis of this sort, it is difficult to assess whether or not the most appropriate form of graphical presentation was being compared with the most appropriate type of table. Until such analyses are forthcoming, robust answers to the essential question are not going to emerge, and guidelines based upon this sort of evidence should be viewed with caution.