Multidimensional data portrayal in executive information systems: Schematic faces

Decision making in executive information systems frequently involves the evaluation of complex sets of multidimensional data. Evaluation of tabular displays can be tedious and time consuming for executives. Furthermore, relationships among variables may be hard to determine and a comprehensive view of the data may be hard to establish. Simple graphical displays are typically adopted to help overcome these limitations. However, the graphical methods available to executives portray only two or three dimensions of data effectively.;A technique for effectively portraying up to 20 dimensions of data was developed by Chernoff in 1973. It consists of representing multidimensional data as a schematic of a face whose facial features (nose length, mouth curvature, eye size, etc.) correspond to the variables of interest. The popularity of the human face as a method of portraying multidimensional data lies in the fact that it has natural aggregating properties that facilitate a global judgement regarding the emotion in the face. For example, by aligning polar emotional states (happiness, sadness) with polar financial outcomes (profitable, not profitable), schematic faces can be used to summarize the overall financial state of a firm.;The purpose of this study is to determine if there is any significant difference in decision making speed and/or accuracy when viewing multidimensional data portrayed by schematic faces as opposed to tables or bar graphs. A model is developed which includes the treatment (table, bar graphs, schematic faces), cognitive style of subjects (high- or low-analytic), and gender (male, female). Experiments are conducted with student decision makers ranking three sets of 10 cities based on nine livability attributes. Decision time and accuracy are measured. Multiple analysis of variance is used to test the hypotheses.;Results indicate that for decisions involving multidimensional data, both decision speed and accuracy are better when decisions are based on schematic faces as opposed to tables or bar graphs. High-analytic subjects made more accurate decisions with tables and bar graphs than low-analytic subjects. However, when viewing schematic faces, decision accuracy is similar for both high- and low analytic subjects. Gender had no effect.