Memorial Representations as a Product of Feedback and Text Variants

Abstract

This study examined the effects of feedback on several memorial representations of a text. Subjects were given either of two variant texts describing a mythical town. One version (route) describes the town as a driver might encounter it, while the other (survey) describes the town using spatial referents. Each version is 25 sentences long and is presented one sentence at a time to control the reading process. For each text version, after reading the text subjects were given 50 seconds to read the entire text in paragraph form (textual feedback) or see a map of the town (map feedback). They were then asked to recall as much of the text as possible and to verify the accuracy of inferential statements. The subjects then read the same version again. Following the reading subjects were given either the same or the alternate type of feedback. Recall and inferential reasoning protocols were obtained again. The recall protocols were scored for number of propositions, and the scores for accuracy of verification of estimated inferential reasoning using a d′ analysis. Recall yielded a doubling of propositions from the first to the second reading. Analysis yielded several findings: (1) a gain in d′ scores between trials, (2) a significant interaction between text version and trial, with the survey group showing a much larger gain than the route version, and (3) an interaction between version, type of feedback, and sequence, with readers of the survey version showing consistently better scores across all conditions. Data for the route version showed some interference between Trials 1 and 2. The results are described within the context of our memorial model of assisted comprehension.