Abstract:

The successful memorisation of similar words is critical for individuals’ vocabulary acquisition. Previous studies have found that individuals perform significantly better in an immediate serial memory test for dissimilar words than similar words. However, the memory advantage for dissimilar words in those studies was mainly based on the comparison of two sets of different learning materials (i.e., similar and dissimilar words). Therefore, whether similar words are memorised better in a similar chunking condition (similar words are successively presented) or dissimilar chunking condition (similar words are alternately presented by other dissimilar words) is unclear.
To address the above question, we performed four experiments in this study, in which within-subject design and study-test paradigm were used. Experiment 1A aims to explore the effects of chunking strategy on the memory of similar words. In this experiment, two matched sets of similar English pseudowords were used for the similar and dissimilar chunking conditions, respectively. In the similar chunking condition, similar words were successively presented, whereas in the dissimilar chunking condition, similar words were alternately presented with other dissimilar words. Participants were instructed to memorise the words during the study phase. A recognition memory test was administered one hour after the study phase. Experiment 1B aims to investigate the memory advantage of the dissimilar chunking condition for long-term retention. Experimental materials and tasks were the same with those of Experiment 1A, but the interval between study and test was prolonged to one week. Experiment 2 used Deese-Roediger-McDermott (DRM) paradigm to examine whether the dissimilar chunking strategy facilitated the memory of similar words by improving the memory of individual words or enhancing the memory of shared parts across similar words. Experiment 3 included unfamiliar Korean characters as materials to further disentangle the contributions of visual and phonological similarities on the memory of similar words.
Results show that: 1) Compared with the similar chunking strategy, the dissimilar chunking strategy show better memory performance on similar words, which can be maintained for at least one week. 2) The dissimilar chunking strategy improves the memory of similar words and results in a high false memory for similar lures. 3) The memory advantage for dissimilar chunking strategy is evident for phonologically similar words (i.e., English pseudowords) but not for visually similar words (i.e., Korean characters).
The results suggest that the dissimilar chunking strategy improves the memorisation of phonologically similar words by enhancing the memory of common parts across similar words. In other words, the dissimilar chunking strategy may be an effective way to improve the memorisation of similar words. These findings have important implications for language learning and education.

Key words: study-test paradigm, word memory, chunking strategy, similar words