Recall of Messages Presented Simultaneously to Auditory and Visual Channels with Varying Uncertainty and Delay of Message

Numerous authors have speculated about how resources are divided when attention has to be paid to two simultaneously occurring messages ( 1 , 2 , 3 ) . Our experiment measured percentage recall of one message as a function of the uncertainty of both the target message and a competing other. Messages were single bisyllabic words having about the same frequency of usage in English (25/million). These were presented auditorally and visually simultaneously by video-tape to groups seated at a uniform distance from the monitor screen and speaker (n = 186). The difficulty level for both modalities was equated by presenting the auditory messages against an 80-db spl white noise and by recording the visual ones slightly out of focus. The auditory words were spoken so as to last and occupy the same time interval as the exposure of the visual word, 500 msec. Ss responded by attempting to identify both words from two lists of which 30-sec. inspection had been allowed before presentation of the word. New lists were used with each trial and uncertainty was varied within lists by varying the number of words per list (2, 4, 8, and 16 words). All pair-combinations of list-lengths were used and used only once. Counterbalancing was carried out for position of modality-list on the response-booklet page and for modality-order of recall. Ten-, 20-, and 30-sec. delay conditions between presentation and recall were created by the insertion of a backwards counting task to test for differential short-term memory effects between modalities. Results showed thac recall from either modaliry was not affected by list-length when the list for the competing modality contained only two words, recall rate for both being 78% when corrected for chance. As list-length increased in one modality, however, recall rate for the other decreased for the longer lists in that modality, visual recall rate being 57% and auditory recall rate being 33% when both lists contained 16 words (P = 7.17, d f = 2/18, p < .01). Auditory recall rate declined sharply for the 16/16 list combination whereas visual recall improved over rate values observed for medium list-length combinations. W e suggest thac "task shedding" occurred when the task demand exceeded S's total processing capacity by the abandonment of the auditory task. While there was a pronounced decline in rate of recall for both modalities as delay of recall increased ( F = 126.8, dJ = 2/18, p < .01), delay did not affect the interaction between modalities (F < 1.00), nor did it affect the modalities differentially (P = 1.76, df = 18/18, p > .01). One possible explanation for the absence of a differential effect of delay is chat, although the modality of message input differed, the same form of short-term memory storage was used for all messages. Finally, list uncertainties were additive in a complementary way in that, for any listlength pairing, the sum over the two modalities of the loss-of-recall percentages was proportional to the sum, expressed in bits, of the uncertainties of the two lists of a pair. REFERENCES 1. BROADBENT, D. E. Decision aed stsess. London: Academic Press, 1971. 2. MOREY, N. Where is capacity limited? a survey and a model. Acta Psychologicu, 1967, 27, 84-92. 3. TREISMAN, A. M. Strategies and models of selective attention. Psychological Review, 1969, 76, 282-299.