Monitoring memory errors: The influence of the veracity of retrieved information on the accuracy of judgements of learning

Many studies have examined the accuracy of predictions of future memory performance solicited through judgments of learning (JOLs). Among the most robust findings in this literature is that delaying predictions serves to substantially increase the relative accuracy of JOLs compared with soliciting JOLs immediately after study, a finding termed the delayed JOL effect. The meta-analyses reported in the current study examined the predominant theoretical accounts as well as potential moderators of the delayed JOL effect. The first meta-analysis examined the relative accuracy of delayed compared with immediate JOLs across 4,554 participants (112 effect sizes) through gamma correlations between JOLs and memory accuracy. Those data showed that delaying JOLs leads to robust benefits to relative accuracy (g = 0.93). The second meta-analysis examined memory performance for delayed compared with immediate JOLs across 3,807 participants (98 effect sizes). Those data showed that delayed JOLs result in a modest but reliable benefit for memory performance relative to immediate JOLs (g = 0.08). Findings from these meta-analyses are well accommodated by theories suggesting that delayed JOL accuracy reflects access to more diagnostic information from long-term memory rather than being a by-product of a retrieval opportunity. However, these data also suggest that theories proposing that the delayed JOL effect results from a memorial benefit or the match between the cues available for JOLs and those available at test may also provide viable explanatory mechanisms necessary for a comprehensive account.

When learners make judgments of learning (JOLs) for some word pairs but not others, how and why is recall performance affected? Participants studied related and unrelated word pairs and made JOLs for a randomly selected half of the pairs. We evaluated two hypotheses. The changed-goal hypothesis states that making JOLs leads learners to notice differences in pair difficulty and to change their learning goal. Because JOLs are manipulated within participants, such a goal change should influence how all (judged or non-judged) pairs are processed on the list, which should lead to no JOL reactivity. The cue-strengthening hypothesis predicts greater positive reactivity (i.e., higher recall for judged versus non-judged pairs) for related than unrelated pairs, because making a JOL strengthens the relationship between the two words in a pair, which would be more beneficial for pairs with an a priori relationship. Across experiments, we found positive reactivity for both related and unrelated pairs (albeit to a lesser degree for the latter). We also found no evidence that learners make qualitative changes in their reported strategy use when judging pairs. Making JOLs for some pairs on a list influenced memory performance and the pattern of reactivity provided support for the cue-strengthening hypothesis.

Past research has evaluated participants’ understanding of their memory by soliciting judgments of learning (JOLs). Importantly, JOLs sometimes change memory for the judged material, leading to JOL reactivity. The cue-strengthening account (Soderstrom et al., 2015) and changed-goal account (Mitchum et al., 2016) propose different mechanisms that lead to JOL reactivity. In the present study, we collected measures that can provide further insight into these mechanisms. Specifically, participants studied related and unrelated word pairs in different colored fonts for a source recognition test. Across three experiments, data were analyzed using a hierarchical Bayesian model of multidimensional source memory to determine how JOLs impact item memory as well as relatedness and color source memory. In Experiment 2, we also compared the effects of making JOLs to making judgments of relatedness (JORs), and Experiment 3 examined how JOLs impact study time allocation. The results of our experiments failed to fully follow predictions of either account. Making JOLs (Experiments 1-3) and JORs (Experiment 2) strengthened item memory for related as well as unrelated pairs, the latter finding not predicted by either account. In addition, JOLs and JORs did not specifically strengthen source memory for relatedness, as the cue-strengthening account predicts, nor did JOLs change study time (Experiment 3), as suggested by the changed-goal account. In all, our results provide novel insight that enhanced item memory may be largely responsible for JOL reactivity, thus adjudicating between candidate explanations.

Past research has evaluated participants' understanding of their memory by soliciting judgments of learning (JOLs). Importantly, JOLs sometimes change memory for the judged material, leading to JOL reactivity. The cue-strengthening account (Soderstrom et al., 2015) and changed-goal account (Mitchum et al., 2016) propose different mechanisms that lead to JOL reactivity. In the present study, we collected measures that can provide further insight into these mechanisms. Specifically, participants studied related and unrelated word pairs in different colored fonts for a source recognition test. Across three experiments, data were analyzed using a hierarchical Bayesian model of multidimensional source memory to determine how JOLs impact item memory as well as relatedness and color source memory. In Experiment 2, we also compared the effects of making JOLs to making judgments of relatedness (JORs), and Experiment 3 examined how JOLs impact study time allocation. The results of our experiments failed to fully follow predictions of either account. Making JOLs (Experiments 1-3) and JORs (Experiment 2) strengthened item memory for related as well as unrelated pairs, the latter finding not predicted by either account. In addition, JOLs and JORs did not specifically strengthen source memory for relatedness, as the cue-strengthening account predicts, nor did JOLs change study time (Experiment 3), as suggested by the changed-goal account. In all, our results provide novel insight that enhanced item memory may be largely responsible for JOL reactivity, thus adjudicating between candidate explanations. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

The neurocognitive basis of metamemory: Using the N400 to study the contribution of fluency to judgments of learning.

A common measure of memory monitoring--judgments of learning (JOLs)--has recently been shown to have reactive effects on learning. When participants study a list of related and unrelated word pairs, they recall more related than unrelated pairs. This relatedness effect is larger when people make JOLs than when they do not make them. Evidence is mixed concerning whether this increased relatedness effect arises because JOLs help memory for related pairs, hurt it for unrelated pairs, or do both. In three experiments, we investigated (1) the nature of the increased relatedness effect (i.e., does it arise from positive reactivity for related pairs, negative reactivity for unrelated pairs, or both?) and (2) the mechanisms underlying the effect. Participants studied cue-target word pairs and either did (or did not) make immediate JOLs and then completed a cued-recall test. When participants studied a mixed list consisting of related and unrelated pairs, the increased relatedness effect was largely driven by positive reactivity. When participants studied pure lists consisting solely of related or unrelated word pairs (Experiment 2 only), the increased relatedness effect was minimized. These and other findings suggest that making JOLs helps learning more than hurts it, and that this reactive effect partly occurs because making JOLs changes people's learning goals.

The current study examined predictions of memory performance as a function of the amount of information to be remembered. In four experiments participants studied and made judgements of learning (JOLs) for long or short lists of words. Results demonstrated that participants provided lower JOLs for long compared with short lists. However, whereas JOLs for short lists strongly corresponded with memory performance, participants' JOLs were consistently overconfident for long lists. Participants were unable to remedy this overconfidence for long lists even when provided information about the list length conditions or warned that a long list of words is difficult to learn. Only when given a prior list learning experience were JOLs for a long list consistent with memory performance. These data indicate that predictions of memory performance are sensitive to the amount of material TBR. However, predictions only correspond with the amount of information to-be-remembered under limited circumstances, providing support for frameworks which suggest that memory predictions are inferential in nature.

JOL reactivity refers to the finding that making judgments of learning (JOLs) while studying material influences later memory for that material. Findings of JOL reactivity have been mixed, with some experiments reporting changes to memory when participants make JOLs and others finding no influence of JOLs. Soderstrom, Clark, Halamish, and Bjork (Journal of Experimental Psychology: Learning, Memory, and Cognition, 41(2), 553-558, 2015) proposed that JOL reactivity will only occur if the final test is sensitive to the same cues used to inform JOLs. The current study evaluated this account by manipulating the type of final test. In four experiments, participants studied mixed lists of related and unrelated word pairs and either made JOLs or did not make JOLs. Making JOLs generally enhanced memory for related word pairs when a cued-recall test was administered. However, during free recall, JOLs had no influence on memory for target information, likely because cue-target associations (which are used to inform JOLs) are less beneficial in the absence of cues. JOLs improved item recognition memory for words that were studied in related pairs, although the effect was small. Collectively, data from a meta-analysis of these experiments indicate that JOL reactivity depends on the type of final test, with reactivity most likely to occur when the final test is sensitive to the same cues used to inform JOLs. Future work should continue examining different tests and study materials in order to develop a comprehensive theory of JOL reactivity.

Objective:Metacognition is defined as the ability to observe, monitor, and make judgments about one’s own cognitive status. Judgments of learning (JOLs) and retrospective confidence judgments (RCJs) are two elements of metacognition related to memory, or metamemory. JOLs refer to one’s assumptions of their memory performance prior to completing a memory task, while RCJs describe one’s subjective assessment of their memory performance after they have completed the task. Traumatic brain injury (TBI) is known to negatively impact general metacognitive functioning. However, the nuanced effects of TBI on constituent metacognitive subprocesses like JOLs and RCJs remain unclear. This study aimed to characterize patterns of brain activity that occur when individuals with TBI render JOLs and RCJs during a meta-memory task. Differences between JOL- and RCJ-related patterns of activation were also explored.Participants and Methods:20 participants with moderate-to-severe TBI completed a metacognition task while undergoing functional magnetic resonance imaging (fMRI). Participants were first exposed to target slides with a set of polygons placed in specific locations, then asked to identify the target slides within a set of distractors. Before identifying the target slides, participants rated how well they believed they would remember the polygons’ shape and location (JOL). After answering, they rated how confident they were that the answer they provided was correct (RCJ). First-level time series analyses of fMRI data were conducted for each participant using FSL FEAT. Higher-level random effects modeling was then performed to assess average activation across all participants. Finally, contrasts were applied to examine and compare JOL- and RCJ-specific patterns of activation.Results:JOLs were associated with activation of the left frontal gyri, bilateral anterior cingulate, left insula, and right putamen (p < 0.01). RCJs were associated with activation of the bilateral frontal gyri, bilateral posterior and anterior cingulate, left insula, right putamen, and left thalamus (p < 0.01). Compared to RCJs, JOLs demonstrated greater left insula activation (p < 0.01). Compared to JOLs, RCJs demonstrated greater activation of the left superior frontal gyrus, bilateral middle frontal gyrus, and bilateral anterior cingulate (p < 0.01).Conclusions:The areas of activation found in this study were consistent with structures previously identified in the broader metacognition literature. Overall, RCJs produced activity in a greater number of regions that was more bilaterally distributed compared to JOLs. Moreover, several regions that were active during both metacognitive subprocesses tended to be even more active during RCJs. A hypothesis for this observation suggests that, unlike JOLs, the additional involvement of reflecting on one’s immediate memory of completing the task during RCJs may require greater recruitment of resources compared to JOLs. Importantly, these findings suggest that, while different metacognitive subprocesses may recruit similar brain circuitry, some subprocesses may require more potent and widespread activation of this circuitry than others. As such, subprocesses with greater activational needs and complexity, such as RCJs, may be more susceptible to damage caused by TBI. Future research should aim to compare patterns of activation associated with certain metacognitive subprocesses between survivors of TBI and healthy controls.

The font size effect is a metamemory illusion in which larger-font items produce higher judgments of learning (JOLs) but not better memory, relative to smaller-font items. We conducted meta-analyses to determine what is currently known about how font size affects JOLs and memory accuracy. In addition, we implemented both univariate and multivariate meta-regressions to isolate the moderators of JOL effects and memory effects. The results revealed a small-to-moderate effect of font size on JOLs. There was also a small but significant effect of font size on memory. This suggests that JOLs and memory accuracy both increase with font size, rather than being completely dissociated. Moreover, JOL-memory dissociation only occurred when font size ranged between very small and intermediate. Our working explanation is that the memory effects of font size are tied to (dis)fluency, but its JOL effects are not. Some boundary conditions were identified for font size effects on both JOLs and memory. Specifically, larger font sizes only reliably increased both JOLs and memory accuracy (a) when font sizes ranged from intermediate to very large, (b) when study materials were unrelated word lists, (c) when JOLs were solicited immediately after encoding, and (d) when study time was relatively brief.Supplementary InformationThe online version contains supplementary material available at 10.1007/s11409-021-09287-3.

Several decades of research have examined predictions of future memory performance—typically referred to as judgments of learning (JOLs). In this chapter, I first discuss the early history of research on JOLs and their fit within a leading metacognitive framework. A common methodological approach has evolved that permits the researcher to investigate the correspondence between JOLs and memory performance, as well as the degree to which JOLs distinguish between information that is or is not remembered. Factors that influence each aspect of the accuracy of JOLs are noted and considered within theoretical approaches to JOLs. Thus far, research on JOLs had yielded a number of findings and promising theoretical frameworks that will continue to be refined. Future work will benefit by considering how learners combine information to arrive at a judgment, the implications of alternative methods of measuring JOLs, and the potential for JOLs to influence memory.

Dual-process theories of memory assume that memory is based on recollection and familiarity. A few dual-process approaches to metacognition have been proposed, which assume that metacognitive judgments, including judgments of learning (JOLs) or predictions about the likelihood of recall, are based on two, or slow and fast, processes. Prior research suggests that JOLs are generally linked to recollection and familiarity (e.g., higher JOLs for recollected items). However, the basis of JOLs is still unknown. Measuring JOL latency as well as magnitude, the present study investigated the underlying processes of JOLs for items that would be classified by the remember-know procedure in cued recall paradigms. Four experiments yielded three findings from individual experiment analyses and meta-analyses. First, high, middle, and low JOLs were assigned to items that received remember, know, and no-memory judgments, respectively. Second, JOL latencies were the longest for items that received a know judgment. For each, the effect was larger for delayed (vs. immediate) JOLs. Finally, in each JOL condition, there was no support for the idea that the relative accuracy of JOLs increases when memory performance is exclusively defined as items that received a remember judgment. Most importantly, the finding of JOL latency suggests that during study, middle-level JOLs are made on the basis of deliberate processes, which is consistent with dual-process approaches to JOLs, for items recalled with a know judgment at test. Discussions include possible processes leading to slow middle-level JOLs and relations of JOLs to recollection/familiarity, and theoretical/methodological issues with using the remember-know procedure. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Judgments of learning (JOLs) are assessments of how well materials have been learned. Although a wide body of literature has demonstrated a reliable correlation between memory performance and JOLs, relatively little is known about the nature of this link. Here, we investigate the relationship between JOLs and the memory retrieval processes engaged on a subsequent memory test. Participants first studied cue-target word pairs and assigned JOLs to each. Later, memory for the cue word in each pair was assessed using an old/new recognition memory task, and electrophysiological measures of familiarity and recollection were examined. Recognition accuracy was superior for materials given high rather than low JOLs. Analysis of event-related potentials (ERPs) revealed that for both high and low JOL items, successful recognition elicited correlates of familiarity (the mid-frontal effect) and recollection (the left-parietal effect). Importantly, however, the magnitude of the familiarity correlate was equal for high and low JOL items, whereas the magnitude of the recollection correlate was significantly larger for items given high JOLs. These findings demonstrate that JOLs made at study correlate with memory retrieval at test-but that this correlation is specific to recollection. The electrophysiological data support the broader view that participants focus on contextual cues when making JOLs, which may later aid recollection.

Making judgments of learning (JOLs) while studying is a useful tool for students to evaluate the status of their learning. Additionally, in associative learning contexts, JOLs can directly benefit learning when the to-be-learned information is related. One explanation for this reactive effect is that making JOLs strengthens the associative relationship, leading to enhanced memory performance when a test relies on that relationship (e.g., cued-recall tests). In the present research, we evaluated whether having students make interactive mental images-another strategy that can increase the strength of a cue-target relationship-impacts the reactive effect of JOLs on learning. Students studied word pairs that were related and unrelated. Half of the students were instructed to form a mental image of the words interacting, whereas the other half were not. Additionally, in each group half of the students made a JOL for each pair, whereas half did not. Following a short delay, students completed a cued-recall test. Consistent with prior research, students who made JOLs remembered more related word pairs than did students who did not. By contrast, students who made JOLs recalled fewer unrelated word pairs than did students who did not. Moreover, although students who formed interactive images demonstrated enhanced memory relative to students who did not, interactive imagery did not impact the reactive effect of JOLs. These outcomes are informative for existing theory of JOL reactivity. Specifically, JOLs may only benefit learning of associative information when it has a pre-existing semantic relationship (e.g., related word pairs) and not when that that relationship is created by the learner (e.g., by forming interactive images).

Research has shown that judgments of learning (JOLs) often produce a reactive effect on the learning of cue-target pairs in which target recall differs between participants who provide item-based JOLs at study versus those who do not. Positive reactivity, or the memory improvement found when JOLs are provided, is typically observed on related pairs, while no reactivity is commonly found on unrelated pairs. In four experiments, we examined JOL reactivity effects by comparing JOL and no-JOL groups to other groups who engaged in relational-type encoding/judgment tasks. Experiment 1 replicated positive JOL reactivity effects with related pairs with an extension to symmetrically related pairs. Next, Experiment 2 found that providing judgments of associative memory—a task that does not involve memory predictions—yielded equivalent reactivity patterns as JOLs. Experiment 3 replicated this reactivity pattern using a frequency of co-occurrence judgment task. Finally, In Experiment 4, a similar positive reactivity pattern was found using a relational encoding task when compared to a standard JOL. Collectively, our results suggest that previous JOL reactivity patterns are not solely due to memory forecasting processes via JOLs and likely reflect relational encoding that is strategically applied towards related, but not unrelated pairs.Supplementary InformationThe online version contains supplementary material available at 10.1007/s11409-022-09301-2.