The Effects of Graphical Versus Numerical Response on the Accuracy of Graph-Based Forecasts

Abstract

This article explores the literature in graphic information presentation and managerial decision making. The literature suggests that task must be controlled in experimental design. The results of an experiment undertaken to determine the effects of collecting numeric versus graphic responses onforecasting accuracy are discussed. Results of the study show forecasts were significantly more accurate when subjects responded graphically than when they responded numerically. The difference in accuracy was found to be unassociated with preference for the two response approaches and to be independent of graph-reading ability.Based on the experimenteal results, it appears that more accurate subjective visual forecasts can, on average, be obtained by collecting the forecasts in graphical format. Therefore, when in comparisons of performance based on graphical versus tabular information, the way in which the response is elicited must be considered along with otherfactors that affect performance and preference for presentation style. This finding has implications for the design of graphical interfaces for managerial decision support systems.