Institute for Software Research
School of Computer Science, Carnegie Mellon University
Impact on Information Loss and Information
Michael K. Martin, Geoffrey P. Morgan,
Center for the Computational Analysis of Social and Organizational
This report describes a preliminary Virtual Experiment (VE) utilizing multi-agent simulation that explores how information loss and information error impact decision accuracy in organizations. Results indicate that information loss and error exert interactive influences on decision accuracy. Moreover, the pattern of interaction suggests that information loss due to break-downs in communication networks does not adversely affect decision-making accuracy in situations characterized by moderate degrees of error in information networks. The results of our VE and the trends observed in our exploratory analyses imply that a line of research concerned with controlling access to information in a manner that depends on its veracity can potentially improve organizational decision-making. Put differently, given estimates of information error, communication networks and socio-cognitive network topology can potentially be used as "throttles" to control access in a manner that improves organizational performance. Along these lines, we include an appendix that describes our initial approach to the representation and measurement of error in communication and information networks.