AbstractThis paper utilizes the conventional statistical tests associated with the rational expectations hypothesis so as to compare the relative accuracy of individual versus group forecasting within the organization. In order to maintain comparability between forecasting regimens the study employs like information sets for the two prediction methods. Using the rational expectations tests as criteria, the statistical results show group forecasts inferior to individually produced predictions These findings imply that group‐produced forecasting accuracy may be hampered by the psychological interaction associated with consensus behavior. Conversely, we find forecasting accuracy improves when predictions are elicited from individuals in an isolated laboratory‐like sett