Word-length effects in backward serial recall and the remember/know task

Abstract

This study tested an item-order explanation of word-length effects in backward serial recall and recognition. It examined (a) whether the superior recall for short words expected for order-based tasks, and consistently found in forward serial recall, would apply also to backward serial recall; (b) whether the superior recall for long words expected for item-based tasks would apply to recognition; and (c) whether there was evidence for qualitative differences in processing between long and short words. Twenty adults performed backward serial recall and recognition tasks based on five-word lists of short and long words. They then completed a remember/know task requiring them to classify each recognized word as being consciously recollected (remember) or as just being familiar (know). The results showed that backward serial recall was better for short words than for long words, whereas recognition was better for long words than for short words. These opposing word-length effects were consistent with the item-order explanation. The remember/know analysis showed that long words were more likely to be classified as consciously remembered than as just familiar, whereas for short words there was no difference. This suggested that long words were encoded with more episodic information than short words during initial processing.