Why are there sometimes concreteness effects in memory for prose?

Abstract

Four experiments explored on-line encoding strategies and memory for high imagery and low imagery texts. Results consistently indicated that concreteness effects in memory for text depend on how materials are presented in several different respects. Most importantly, the experiments clarified apparently contradictory results of previous studies by indicating that concreteness effects generally do not occur in memory for prose when imageability is manipulated between-subjects, and that their occurrence when imageability is manipulated within-subjects depends on the order occurrence when imageability is manipulated within-subjects depends on the order of presentation. In addition, moving window analyses of text processing strategies indicated that differential strategies observed in previous studies when subjects listened to high vs low imagery text do not generalize to reading of the same materials. Potential explanations for the pattern of results are evaluated, and implications for theories of mental imagery and memory are considered.