An empirical comparison of alternative methodologies for the evaluation of configural displays

Two different methodologies (visual, memory) were used to evaluate alternative versions of the same configural display. One version (composite display) had several graphical design techniques applied, whereas the other version (baseline display) did not. Two types of information probes (high-level property, low-level data) were administered. When the displays were visible during completion of the probes (visual methodology), the display manipulation had the largest impact on performance (composite display associated with better performance); when the displays were not visible (memory methodology) the probe manipulation had the largest impact on performance (high-level probes associated with better performance). These results are interpreted in light of the mutually interacting constraints introduced by factors in display design, task requirements, and the participants' cognitive and perceptual capabilities/limitations. Implications for both the design and the evaluation of displays and interfaces in general are discussed. Actual or potential applications of this research include design techniques for improving the quality of graphic displays and methodological insights for interpreting previous research and guiding future experimentation.