Article ID: | iaor20072230 |
Country: | Netherlands |
Volume: | 42 |
Issue: | 1 |
Start Page Number: | 186 |
End Page Number: | 203 |
Publication Date: | Oct 2006 |
Journal: | Decision Support Systems |
Authors: | Marsden James R., Pakath Ramakrishnan, Wibowo Kustim |
Keywords: | decision: studies |
In prior reports, we describe the design and implementation of a software platform and a laboratory experiment aimed at a preliminary investigation of the efficacy of an alternative communication language for a time-pressured, incentive-driven decision-making setting. Our experimental findings suggest that, despite its relative novelty, the alternative language is at least as effective as the more traditional communication modes of written and spoken English in the specific setting examined, while succinctly conveying what must be conveyed. That work embodies three central underlying constructs – information communication mode, communication speed, and complexity level. Some analysis focuses solely on mode. In this concluding portion of the research, we analyze the two remaining constructs – speed and complexity level. Taken together with that in an earlier paper, our analysis enables a fuller understanding of the interplay between communication mode, speed, and complexity. We note two particular findings of our research: (1) whereas real-world systems are often designed based on prescriptive tradition, including ‘best practice’ criteria widely disseminated and adopted by practitioners without demonstrated benefits to users, in our study, we found that performance at any level of a particular construct was typically sensitive to the levels set for the remaining two attributes, and (2) despite the fact that many experimental studies continue to emphasize subject perception over more objective performance measures, in our study, we found that subject preexperimental perceptions did not always coincide with actual experimental outcomes. Our work, like most controlled experimental analyses, is subject to caveats related to limited sample size and specific orderings examined.