A Theoretical Study of Organizational Performance Under Information Distortion

How should organizations of intelligent agents be designed so that they exhibit high performance despite information distortion? We present a formal information-based network model of organizational performance given a distributed decision making environment in which agents encounter a radar detection task. Using this model, we examine the performance of organizations with various designs in different task environments subject to various types of information distortion. We distinguish five sources of information distortion---missing information, incorrect information, agent unavailability, communication channel breakdowns, and agent turnover. This formal analysis suggests that: (1) regardless of information distortion, performance is enhanced if there is a match between the complexity of organizational design and task environment; (2) task environment characteristics have more effect on performance than information distortion and the organizational design; (3) the effects of information distortion can be combated by training, but only to a limited extent; and (4) technology based information distortion typically is more debilitating than personnel induced information distortion.