A key factor for effective system management and evaluation of decision processes is accurate and timely feedback. It is through feedback one detects deviation from reference points thereby allowing for more effective system maintenance and performance. This paper demonstrates, in living systems at the organization level, including the total system and all subsystems, that even with accurate and timely information there are problems that can strongly bias feedback. Two models, a Partial Information Model and a Full Information Model, are developed and compared to predict the bias effects. A final section of the paper considers how bias in the feedback process may be corrected.