How much help do repeated presentations give to recognition processes?

Abstract

A recent experiment by Haber and Hershenson (1965) had shown that in a recognition task one long look at a stimulus was always superior to two or more shorter looks summing to the same total presentation time. In order to explore this more fully and to account for opposite results in a different type of recognition task, as well as in a serial learning task, an improved replication of the earlier study was carried out using very short durations and single letters as stimuli. The same non-reciprocity was found, again strongly favoring duration over repetition as a determinant of clarity of a percept, even though repetition alone was also shown to be a significant independent variable. As a subsidiary finding, an error analysis showed that when a letter was misnamed it was nearly always confused with one that looked like it rather than one that sounded like it. Some discussion was offered as to the role of an auditory information storage in low memory load tasks such as this one, as well as some general implications for information processing analyses of the non-reciprocity of duration and repetition.