Abstract
In experiments using 5 separate groups involving a total of 103 Ss, performance data were obtained on free recall of dichotically presented digits. The purpose was to investigate the incidence and distribution of the “Broadbent effect,” the phenomenon in which lists of dichotic digits are recalled in ear order so that the half-spans presented to each ear remain intact. Although the Broadbent effect occurred more frequently than would be expected by chance in all groups tested, over-all incidence was not high and there were marked differences from one S to another. The fact that incidence proved to be high in some persons but zero in many others suggested that there are 2 different populations of Ss as far as the phenomenon under investigation is concerned. In cases where errors occurred during recall of a list, the half-span reported first was less likely to contain an error than the half-span reported second. An 18-db difference in playback level between the 2 ears appeared to attenuate rather than enhance the Broadbent effect.