Concreteness effects revisited: The influence of dynamic visual noise on memory for concrete and abstract words

Abstract

Two experiments are presented that investigate the effects of dynamic visual noise (DVN) on memory for concrete and abstract words. Memory for concrete words is typically superior to that of abstract words and is referred to as the concreteness effect. DVN is a procedure that has been demonstrated to interfere selectively with visual working memory and the generation of images from long-term memory. It was reasoned that if concreteness effects arise because of the ability of the latter to activate visual representations, then DVN should selectively impair memory for concrete words. Experiment 1 found DVN to selectively reduce free recall of concrete words. Experiment 2 investigated recognition memory and found DVN to reduce memory accuracy and remember responses, while increasing know responses to concrete words.

Keywords:

• Concreteness effects
• Dynamic visual noise
• Imagery
• Recognition
• Recall

Notes

¹One of the reviewers indicated that concrete (vs abstract) words typically have a lower age of acquisition (AoA). As we did not attempt to control for this variable, this deserves some comments. First, research indicates that late-acquired words are more likely to be recalled and associated with accurate responses (e.g., Dewhurst, Hitch, & Barry 1998; Gullick, & Juhasz, 2008; Morris 1981). If abstract words were acquired later than concrete words then this would militate against finding a concreteness effect, or reduce its magnitude. Instead we observed a robust concreteness effect. In addition, to the extent that AoA influences recall, it is difficult to see how this factor would account for the interaction we observed between word type and the DVN task. Nevertheless we calculated the AoA ratings for the concrete and abstract words based on the norms of Gilhooly and Logie (1980). Values for two concrete words and three abstract words were not available. The means (SD) for concrete and abstract words were 3.99 (1.01) and 4.25 (0.73) respectively. Although the value for concrete words was numerically lower, the difference was far from significant, t(23) = −.725, p=.48. Thus any slight differences in AoA are unlikely to have had any significant influence on our results.

²AoA scores were also calculated for the words used in Experiment 2. Values for four (nine) of the concrete (abstract) words were not available. The means (SD) for the concrete and abstract words were 3.83 (0.97) and 4.21 (0.69) respectively. Although the values for concrete words were numerically lower, this difference was not significant, t(34) = − 1.32, p=.20.