In the present study, we employed word pair stimuli as memory materials to explore whether initial unsuccessful memory retrieval can enhance subsequent recollection. In addition, we examined how this memory retrieval enhancing effect after unsuccessful retrieval was affected by memory material’s qualitative (semantic relatedness) and quantitative attributes (difficulty of target words due to the number of syllables).
In experiment 1, we tested whether unsuccessful memory retrieval enhances later memory retrieval and how material’s attribute (semantic relatedness) affects the enhancing effect in immediate and delay periods. The results showed that learning through unsuccessful retrieval was more effective than both immediate and delayed recall learning. Especially, in the case of immediate recall, different learning groups showed similar recollection rates for word pairs with low semantic relatedness, while unsuccessful retrieval group showed higher recollection rates compare to passive reading group for word pairs with high semantic relatedness.
In experiment 2, we replicated the findings in the experiment 1 and investigated the effect of learning duration that might have occurred in the experiment 1. We added a cue presentation group and exposed memory materials for the same duration as the retrieval group to verify the memory enhancement effect for different learning groups (passive reading, cue presentation, unsuccessful retrieval) and word pair relatedness. The results were consistent with the experiment 1 and compare to passive reading and cue presentation, learning through unsuccessful retrieval was more effective for subsequent recollection. Also, the memory enhancement effect of unsuccessful retrieval was significant only in the word pairs with high relatedness for both immediate and delayed recall conditions.
In experiment 3, we tested effects of qualitative (semantic relatedness) and quantitative (difficulty level) attributes on the memory enhancement effect of unsuccessful retrieval. We employed memory materials of high/low relatedness and high/low difficulty levels, and compared immediate and delayed recollection rates of different learning groups. The results showed that high difficulty group had greater recollection rates than low difficulty group regardless of relatedness levels. Moreover, low relatedness group had greater recollection rates than high relatedness group regardless of difficulty levels. In the case of the delayed recall, recollection rates due to difficulty levels had more critical effect on the high relatedness group than the low relatedness group. Lastly in experiment 4, we controlled for the relatedness and only applied the difficulty levels to explore the unsuccessful retrieval learning enhancement effect. To test this, we created both low (2 syllables) and high (3 syllables) difficulty nonsense syllables and compared three learning groups’ (passive reading, cue presentation, unsuccessful retrieval) recollection rates. Regardless of the learning groups, the low difficulty condition showed greater recollection rates and the memory enhancement effect of unsuccessful retrieval was not found.
The findings of experiment 1 and 2 suggests that even unsuccessful retrieval can enhance subsequent recollection and the semantic relatedness of word pairs has a crucial effect on the memory enhancement effect. Also, the results of experiment 3 and 4 indicates that the difficulty level of memory materials has a minimal effect on the memory enhancement effect of unsuccessful retrieval and learning non-meaningful words through unsuccessful retrieval process has negative effect on the subsequent recollection.