In an earlier study we reported the ability of a 'set' to alter the accuracy of reporting about a degraded stimulus-object. It was shown that set could be varied in degree; that its influence depends on the level of ambiguity of the degraded stimulus; and that it may affect discrimination, or interpretation, or it may simply make one response more probable by decreasing the alternatives.l A number of questions raised in the postulate remain unanswered. The following experiments were designed to test whether a quite different method of establishing the set, namely, by ear rather than on a printed page, would change the results. The experiments were also designed to test words as stimulus-objects in place of letters. There are evident differences in the functioning of the visual and auditory senses.2 Because of such differences, set may function differently in the two modalities. A similar comparison can be made of words and letters. For example, words come from a larger population of stimuli than do letters of the alphabet. Thus, there is more inherent uncertainty in words than in letters. On the other hand, it might be argued that words possess more meaning than do letters.