Note: Release from Proactive Interference with Television News Items

Abstract

An adaptation of the Brown-Peterson short-term memory procedure outlined by Wickens (1, 2) for studying the encoded attributes of simple verbal items has been devised to identify the taxonomic dimensions into which individuals categorize and possibly encode television news materials. Instead of word triads, subjects are presented on each trial, with three short television news items recorded on videotape from actual network newscasts. These are carefully edited to control for length and visual format effects and shown to the subjects over a small monochrome television monitor. Following presentation of each triad of news items, the subject performs a distracting task (crossword puzzle) for 1 mia, and then is asked to free recall the news irems. During recall the subject is required to write short accounts of each item in 2 min. and is scored for the number and accuracy of these responses on each trial. All subjects had four trials: for Group 4-0 the taxonomy of the news items is held constant over all trials, but for Group 3-1 the taxonomy is constant for Trials 1-3 only, then on Trial 4 is switched to a different category. In an initial experiment, 48 subjects were shown items about political news events or items about sport. Half Group 4-0 subjects received political items across all trials and half received sports items. Likewise in Group 3-1, critical fourth trial news-category switches from political to sports items or vice versa were balanced across subiects. Responses were scored out of two points for accuracy. Raw scores were converted to proportions and transformed for computation. A two-way (Groups x Trials) splitplot factorial analysis gave a significant decrement in recall across Trials I to 4 occurring in Group 4-0 (P3.1~ = 22.84. P < 0.001), indicating a buildup of proactive inhibition over a series of news items in which taxonomy was held constant.' In Group 3-1, a decrement in performance occurred over Trials 1-3. followed by substantial improvement on Trial 4 after the critical news-category switch. Analysis of simple main effects showed that recall on Trial 4 in Group 3-1 was significantly better than that on the equivalent trial in Group 4-0 (F1.]% = 24.02, fi < 0.001), indicating a significant release from proactive interference. These early findings are consistent with prior ones (1,2) and suggest that the Brown-Peterson technique may be usefully generalized to television news as a means of identifying taxonomic categories into which individuals classify news items. The broad theoretical and practical implications of these and other findings will be discussed later.