关键词: 学习判断, 反应性效应, DRM范式, 项目自身与关系理论

Abstract:

Judgments of learning (JOLs) refer to learners' predictive evaluations of the likelihood that they will successfully remember a given item in a future test. Previous studies have primarily focused on examining the accuracy of JOLs and elucidating the mechanisms underlying JOL formation. However, recent studies suggest that the act of making JOLs can alter memory itself, a phenomenon known as the reactivity effect. Typically, participants made a JOL after learning each item. This metamemory monitoring process may heighten individuals' metacognitive awareness of their memory compared to conditions without JOLs, thereby triggering the reactivity effect. Zhao et al. (2023) further differentiated the reactivity effects of JOLs on item-specific memory and inter-item relational memory. They found that making JOLs can enhance item-specific memory but concurrently weaken inter-item relational memory (e.g., memory for serial order or semantic relational information). A potential explanation for this is the item-specific and relational account, which posits that making JOLs enhances encoding of item-specific details, resulting in a positive reactivity effect. However, because cognitive resources are limited, allocating more cognitive resources to processing item-specific details reduces resources available for processing inter-item relations, leading to a negative reactivity effect on inter-item relational memory. The current study employed on the DRM paradigm to investigate the reactivity effect on false memory and to test the item-specific and relational account.

Experiment 1 explored the reactivity effects of JOLs on false and veridical memory. Sixty-four participants learned DRM word lists, each containing 11 DRM study words and 1 critical lure. Critical lures were withheld during the learning phase and presented only during the recognition test. Half of the DRM word lists were studied under the JOL condition, and the remaining half were learned under the no-JOL condition. Participants were instructed to learn each DRM word individually. The key distinction between the JOL and no-JOL conditions was that, in the former, participants completed item-by-item JOLs while learning each word. After the learning task, participants completed a distractor task, followed by a recognition test. The results showed that item-by-item JOLs disrupted semantic processing among DRM words and decreased false memory (i.e., false alarm rates for critical lures). Concurrently, item-by-item JOLs facilitated item-specific processing, yielding a positive reactivity effect on memory for studied words.

Experiments 2 and 3 changed presentation format of DRM lists and asked participants to make global JOLs for a whole word list, rather than for each word, to examine the reactivity effect on both intra- and inter-item relational memory. Experiment 2 used pure DRM lists, with six words from the same thematic word lists presented together. Experiment 3 employed mixed DRM lists, with each list containing six study words from different thematic word lists but with shared thematic relationships across lists. In the no-JOL condition, DRM words were not presented individually. Instead, pure (Experiment 2) or mixed word lists (Experiment 3) were simultaneously displayed on screen for a 12 s study duration per list. Participants in both experiments provided global JOLs for each list, predicting the number of words they would remember in the subsequent test. The results showed that making global JOLs facilitated processing of intra-item semantic relations but disrupted processing of inter-item semantic relations.

In summary, the current study revealed that item-by-item JOLs disrupt semantic relational processing among individual DRM words, reducing false memory but promoting processing of item-specific information, thereby facilitating recognition of studied words. Additionally, global JOLs enhance intra-item semantic relational processing but impair inter-item semantic relational processing. The results support the item-specific and relational account and elucidate the cognitive mechanisms underlying the reactivity effect. Moreover, these findings offer valuable insights into the development of effective interventions for mitigating false memory.

Key words: judgments of learning, reactivity effect, DRM paradigm, item-specific and relational account