The effect of experimentally-induced success and failure in a complex decision-making task on subjective estimates of infor­ mation quality was obtained for information relevant to Ss' own decision-making area and that of a marginal group member. It was found that quality initially is perceived to improve. Estimates of quality for success and failure conditions do not differ until success and failure levels are quite high. Once high levels of success are reached, Ss in the success condition consider information was further improving, while Ss in failure conditions maintain previous perceptions. The implications of this result for complexity theory are considered. Considerable research has been concerned with the effect of information on human behavior. Implicitly some of the work has dealt directly with the effect of information quality. For instance, from the view of information theory (e.g., Attneave, 1959), an amount of information that reduces uncertainty to a particular degree can be viewed as possessing "quality." In other words, information communicating a bite of information would be of higher quality than information communicating a bit of infor­ mation. Similarly, information levels that produce optimal integrative information processing characteristics (Driver & Streufert, 1966; Schroder, Driver, & Streufert, 1967) might be seen as possessing greater quality than information levels that produce less than optimal integrative information processing. Such conceptualizations of information quality, although based on subject performance, are experimenter defined. What, however, is the S's perception of information quality? A partial attempt to answer that question is made in this paper. Streufert & Streufert (1968) have recently demonstrated that groups of Ss exposed to increasing failure conditions attribute as much cause, to their own decisions as they credit to an opponent's decisions when they are asked to explain the basis of their present (failure) situation. These estimates of causality do not change over different failure levels. However, Ss in increasing success condi­ tions tend to attribute more and more causality to their own decisions. If attributions of quality to information would follow similar patterns, then one might expect that under high success condi­ tions, Ss would consider information to be of higher quality than they would under low success, or low and high failure conditions. If, on the other hand, information quality perceptions are structurally determined (Driver & Streufert, 1966; Schroder et ai, 1967), no differences between success and failure conditions should emerge. In this research the effect of increasing failure and increasing success on perceived information quality is investigated. In addition to estimates of information quality concerning the S's own decision area, estimates of quality for information relevant to the decisions of a marginal group member are obtained. Subjects and Task. Thirty-six paid undergraduate male volunteers from an eastern state university were placed into 18 two-man decision-making teams. The Ss were instructed to act as equal rank decision makers in a simulated internation game situation. (The setting is discussed in detail in Streufert, K1iger, Cast ore, & Driver, 1967.) Each team was given the task of making military, economic, intelligence, and negotiation decisions regarding an international conflict situation with some Vietnam characteristics. Teams were told that they were playing a game against another team which supposedly had been instructed to oppose them. All functions of the "enemy team" were pre-programmed and performed by the Es. Consequences of the