Are we aware of our ability to forget? Metacognitive predictions of directed forgetting

It is often important to accurately predict not only what information we will later remember, but also what information we are likely to forget. The present research examined these abilities in the context of item-method directed forgetting, to determine whether people are aware of their strategic control of remembering and forgetting, as well as what cues are used when making metacognitive judgments. Participants studied words, each of which was followed by a cue to remember (R) or forget (F) the word for an upcoming test, and also made predictions of how likely they would be to later recall each word. When asked to recall all of the words, regardless of cue, both recall and predicted recall were fairly sensitive to the R or F instructions, despite some overconfidence. A similar and stronger pattern was found when words were assigned positive or negative point values as cues to remember or forget. These findings suggest that item-based cues to remember or forget information can be successfully utilized when making metacognitive judgments, and that people are fairly aware of the control they have over both remembering and forgetting information.     

Utilization of base-rate information during feeling-of-knowing judgments.

We examined whether subjects use base-rate information about item difficulty when making feeling-of-knowing judgments for items they failed to recall. First, the subjects attempted to recall the answers to general-information questions. Then, for those items they recalled incorrectly, half of the subjects received information about the normative probability of recall of each item while judging their feeling of knowing. The other subjects made their feeling-of-knowing judgments without receiving any base-rate information. Finally, all subjects had a forced-choice recognition test on those items to validate the accuracy of their feeling-of-knowing judgments. Relative to the no-base-rate information group, the base-rate group had lower feelings of knowing for normatively difficult items and higher feelings of knowing for normatively easier items. Subjects who had received base-rate information during the judgment state had greater feeling-of-knowing accuracy than subjects who did not receive base-rate information. However, even the predictions from subjects who received base-rate information were not significantly more accurate for predicting subsequent recognition than were the predictions derived from normative information alone.     

Memory predictions are influenced by perceptual information: evidence for metacognitive illusions

Although perceptual information is utilized to judge size or depth, little work has investigated whether such information is used to make memory predictions. The present study examined how the font size of to-be-remembered words influences predicted memory performance. Participants studied words for a free-recall test that varied in font size and made judgments of learning (JOLs) for each item. JOLs were influenced by font size, as larger font sizes were given higher JOLs, whereas little relationship was evident between font size and recall. The effect was modified when other, more valid, sources of information (e.g., associative strength) were available when JOLs were made and persisted despite experience with multiple study-test sessions, use of a forgetting scale to assess predictions, and explicit warning of participants that font size has little effect on memory performance. When ease of reading was manipulated, such that large font size words were made less fluent, the effect was eliminated. Thus, highly accessible perceptual cues can strongly influence JOLs, likely via encoding fluency, and this effect can lead to metacognitive illusions     

Metamemory for narrative text

In this experiment, we investigated metamemory for narrative text passages. Subjects read two stories and made memory predictions for the idea units in one and rated the importance of ideas in the other. Half of the subjects were asked to recall the story immediately after reading the passages and half were asked to recall 1 week later; half received passages with single inconsistent idea units and half received passages with corresponding consistent idea units. All subjects made confidence judgments about the accuracy of their recall. Subjects’ prediction ratings were related to recall, as shown by significant prediction accuracy quotients. Importance ratings were related to recall on the delayed test but not on the immediate test. Memory prediction ratings predicted recall better than did importance ratings. The absolute level of memory predictions did not differ with delay, but subjects did give higher confidence judgments on an immediate than on a delayed test. Subjects recalled the inconsistent idea better than the consistent idea for one story but not for the other. For both stories, subjects predicted that they would remember the inconsistent ideas better, suggesting that they have avon Restorff-type view, rather than a schema view, of memory. We conclude that subjects can predict their memory for the idea units in narrative text.     

Memorability judgments for high- and low-frequency words

Five experiments were conducted in order to examine subjects’ judgments of the memorability of high- (HF) and low-frequency (LF) words in the context of a recognition memory task. In Experiment 1, the subjects were provided study/test experience with a list of HF and LF words prior to making memorability judgments for a new list of HF and LF items. The findings were consistent with previous evidence (Greene & Thapar, 1994; Wixted, 1992) suggesting that subjects are not explicitly aware of the greater recognition memorability of LF words. Experiments 2–5 embedded the memorability judgment task within the recognition test itself. In these experiments, the subjects consistently gave higher memorability ratings to LF items. The contrast between the pattern of results found when the subjects made their judgments at the time of list presentation (Experiment 1) and that when they made their judgments during the recognition test (Experiments 2–5) is consistent with recent evidence that even seemingly highly related metamnemonic judgments (e.g., ease of learning judgments vs. judgments of learning for the same items) may be based on very different factors if they occur at different points in the study/test cycle. The present findings are also consistent with the possibility that very rapid retrieval of memorability information for HF and LF words may affect recognition decisions and may contribute to the recognition memory word frequency effect.     

The accuracy of beliefs about retrieval cues

We investigated the accuracy of predictions about semantic, environmental, and phonological cues for remembering. Subjects rated the pleasantness of 10 words in each of four rooms, predicted the number of words that they would recall with and without one of the three types of cues, and then were tested for free or cued recall. Consistent with their predictions, subjects who received semantic cues recalled more words than did subjects in the free-recall group. The subjects in the other cuing conditions did not benefit from the cues; furthermore, they overestimated the value of phonological cues, and they believed that environmental cues were ineffective. Finally, confidence ratings for cued-recall predictions did not reflect the pattern of cued-recall performance. Subjects were least confident about their predictions for semantic cuing and most confident about their predictions for recall to be cued phonologically.     

“P300” and memory: Individual differences in the von Restorff effect

Event-related brain potentials (ERPs) were elicited by words in a free recall paradigm that included a novel item. The P300 component of the ERP is elicited by novel, task-relevant events, and we tested the hypothesis that P300 is a manifestation of the cognitive processing invoked during “context updating.” If the degree to which current representations in working memory need revision is related to P300 amplitude, then the P300 elicited by a given item should be related to the ability to recall that item on a subsequent test. Forty lists were presented to 12 subjects in each of two sessions. The lists were 15 words long, and 1 word, in position 6 through 10, was “isolated” by changing its size. Most subjects recalled these isolated words more often than other words in the same positions (von Restorff effect), and these words also elicited larger P300s than other words. Analysis of variance on the component scores from a principal components analysis revealed that words recalled had a larger amplitude P300 (on initial presentation) than words not recalled. Striking individual differences emerged, and there were strong relationships between the von Restorff effect, overall recall performance, mnemonic strategies, and the association between components of the ERP and recall performance. The overall recall performance of subjects who reported simple (rote) mnemonic strategies was low, but they showed a high von Restorff effect. For these subjects the amplitude of the P300 elicited by words during initial presentation predicted later recall. In contrast, subjects who reported complex mnemonic strategies remembered a high percentage of words and did not show a von Restorff effect. For these subjects P300 did not predict later recall, although a later “slow wave” component of the ERP did. The initial response to isolated items was the same for all subjects (a large P300), and all subjects recognized the isolates faster than other words in a recognition test given at the end of each session. The subjects in whom P300 did not predict recall reported mnemonic strategies that involved organizing the material. These strategies continue long after the time period reflected by P300 (600 msec). Because they were so effective they may have overshadowed the relationship between P300 and recall, which is based on the initial encoding of an event. Our interpretations were further confirmed and clarified from data obtained in a final grand recall and in the recognition test.     

Young children's knowledge of their memory span: effects of task and experience

This study investigated the hypothesis that young children have knowledge about their memory that they may be unable to articulate, but are able to reflect on and use in problem-solving. Forty-eight kindergarteners made one of two types of judgments about their memory span for words. Half of the children made prospective verbal predictions about the number of words they thought they could recall from a list of 10. The other half made concurrent, nonverbal predictions by listening to words on a tape and manually stopping the tape when they heard as many words as they thought they could recall. Children's actual recall for words was then assessed. All children participated in multiple trials to assess the effect of task experience on their predictions. Analyses revealed that predictions made in the concurrent task were significantly more accurate than those made in the prospective task. All children lowered their predictions across trials, although only in the concurrent task were children's final-trial predictions not significantly greater than their actual recall. No meaningful effects or interactions were associated with actual recall scores. These results revealed that young children manifested greater memory knowledge when this knowledge was assessed through their concurrent problem-solving behavior rather than through their prospective verbal predictions.     

Cue familiarity but not target retrievability enhances feeling-of-knowing judgments.

Two hypotheses concerning people's ability to predict later memory performance for unrecalled items were investigated. The target retrievability hypothesis states that feeling-of-knowing judgments (FKJs) are based on partial target information; and the cue familiarity hypothesis asserts that they are based on recognition of the cues. In Experiments 1 and 2, subjects either generated or read the targets of paired associates. Half of the cues had been primed in a pleasantness-rating task. The generation manipulation increased recall but had no effect on FKJs. Cue priming had no effect on recall but increased FKJs. In Experiment 3, using general information questions, primed after the initial recall attempt, both cue and target priming increased FKJs. Experiment 4, which remedied difficulties in Experiment 3, showed no effect of target priming whereas cue priming increased FKJs. The results favor the cue familiarity hypothesis.     

Do different metamemory judgments tap the same underlying aspects of memory?

We compared the predictions from several kinds of metamemory judgments (on the same set of items), both in terms of their predictive accuracy and in terms of the commonality of predictions. Undergraduates made judgments about the ease with which they could learn each item in a list (ease-of-learning judgments); then they learned every item, either to a minimal criterion of learning or with overlearning, and made judgments about how well they knew each item (judgments of knowing); finally, they returned 4 weeks later for a retention session and made feeling-of-knowing judgments on every time they could not recall, after which a recognition test assessed predictive accuracy. Ease-of-learning judgments had the least predictive accuracy. Surprisingly, however, the recognition of nonrecalled items was predicted equally well by judgments of knowing (made 4 weeks earlier) as by feeling-of-knowing judgments (made immediately prior to recognition). Moreover, those two kinds of judgments were only weakly correlated with each other, which implies that they do not tap memory in the same way.     

Recollection-based prospective metamemory judgments are more accurate than those based on confidence: judgments of remembering and knowing (JORKS)

Five experiments were conducted to examine whether the nature of the information that is monitored during prospective metamemory judgments affected the relative accuracy of those judgments. We compared item-by-item judgments of learning (JOLs), which involved participants determining how confident they were that they would remember studied items, with judgments of remembering and knowing (JORKs), which involved participants determining whether studied items would later be accompanied by contextual details (i.e., remembering) or would not (i.e., knowing). JORKs were more accurate than JOLs when remember-know or confidence judgments were made at test and when cued recall was the outcome measure, but not for yes-no recognition. We conclude that the accuracy of metamemory judgments depends on the nature of the information monitored during study and test and that metamemory monitoring can be improved if participants are asked to base their judgments on contextual details rather than on confidence. These data support the contention that metamemory decisions can be based on qualitatively distinct cues, rather than an overall memory strength signal.     

Effects of study task on the neural correlates of source encoding

The present study investigated whether the neural correlates of source memory vary according to study task. Subjects studied visually presented words in one of two background contexts. In each test, subjects made old/new recognition and source memory judgments. In one study test cycle, study words were subjected to animacy judgments, whereas in another cycle the study task required syllable judgments. Functional magnetic resonance imaging (fMRI) was employed to contrast the neural activity elicited by study words that attracted accurate source judgments on the later memory test, as opposed to words for which source judgments were incorrect or for which source information was unavailable. In both tasks, relative to words for which source memory failed, study words that were later assigned to the correct source elicited enhanced activity in ventral extrastriate cortex. In addition to these common effects of subsequent source memory, additional effects were observed that were selective for each study task. The present findings add weight to the proposal that neural activity supporting successful episodic memory encoding is a reflection of both the online processing engaged by an episode as it is experienced, and the demands imposed by the later retrieval task.     

Effects of processing fluency on comparative performance judgments

Research has shown that performance predictions are biased by the impact of processing fluency. However, existing data are inconclusive with regard to comparative judgments of performance. In five experiments, participants in an easy condition gave more favorable comparative judgments than participants in a difficult condition. Participants judged their performance more favorably if they named colors of non-color words rather than non-matching color words (Experiment 1), if they had to generate six words of a category rather than 12 words (Experiment 2), if they had to run in place for 15 s rather than 2 min (Experiment 3), but the latter result holds only true if participants were not active in sports (Experiment 4). When 67% of the items in a recognition test were old words, participants thought that their recognition performance was better than when 33% of the items were old words, although recognition performance did not differ between groups (Experiment 5). We discuss this result in the light of recent theories about effects of processing fluency on judgments.     

Tests of the generality of the principle of encoding specificity

Two empirical tests of the principle of encoding specificity are reported. In Experiment I, the normative strength of the cues presented on the input and on the recall trial was varied factorially. To lessen the emphasis on strictly associative learning, only half the items were cued in each phase of the study-recall cycle. Recall was higher when the cues remained the same than when they changed. However, regardless of the condition of input cuing, strong output cues were substantially more effective than weak ones. In Experiment II, the to-be-remembered words were shown in the presence of weak cues on the input trial. Recognition in the context of strong extralist cues was compared with recall to the original input cues. On the test of cued recognition, the target words were either generated by the subjects as free associates or presented to them as items on a test constructed by the experimenter. Contrary to previous findings, recall was not found to be superior to recognition. The phenomena of cue-dependent forgetting that have been interpreted as evidence for the principle of encoding specificity appear to have limited generality.     

Part-set cuing of false memories

Part-set cuing inhibition describes the common finding that re-presenting items from a word list can reduce subjects’ overall recall performance for studied items. Do part-set cuing effects occur for false memories as well? In the present experiments, subjects studied lists of words drawn from Roediger and McDermott (1995). After studying each list, subjects completed math problems and then recalled the list items either with or without accompanying list cues. In Experiment 1, the recall cues consisted of items drawn randomly from the original list. In Experiment 2, an additional type of cued recall task was added in which the even numbered list items were used as cues. Taken together, these experiments demonstrate robust part-set cuing effects for critical nonpresented items. In addition, they show that whereas recall of critical words is reduced by the presence of cues at test, retrieval cues do not affect critical words and studied words in exactly the same manner.     

Elaborative processing and conjunction errors in recognition memory

Four experiments were conducted in order to examine the influence of elaborative processing at encoding on recognition memory conjunction lure errors. In these experiments, participants generated cues for compound words as wholes (e.g., haywire) or as separate entities (e.g., hay, wire). Studied words were re-presented in exact form (old) or recombined to form conjunction lures on the recognition test. Participants were asked to make old-new judgments and to indicate whether they had rejected items judged to be new because of recall of a studied item or because of lack of familiarity with an item. The results suggested that recall-to-reject processing and conjunction lure familiarity increased with both types of generation, although generation of cues for compound words as a whole did not influence conjunction lure error rates. An emphasis on processing each constituent of a compound word during encoding increased the familiarity of those constituents more than generation of a compound word as a whole, resulting in an increase in conjunction lure errors. These results suggest that both familiarity and recollection-based monitoring processes influence conjunction lure errors, and therefore support dual-process theories of recognition memory.     

The Declining course of recognition memory

Recognition memory for words can decline, sometimes precipitously, over the course of the recognition test. Decrements of as much as 75% in d’ have been observed from the first quarter of testing to the last. and decrements of 40%–50% are not uncommon. Proactive mechanisms, imperfectly understood, appear to underlie such declines: Processing early input renders late input less recognizable, and making early recognition judgments renders later judgments more difficult. At the same time, the processing of late-input words fails to inhibit the recognition of the words that preceded them.     

The retrieval practice effect in associative recognition

Recalling an item interferes with recall of related memories. Evidence is presented that retrieval interference occurs in associative recognition as well as recall. In Experiment 1, subjects studied pairs of category exemplars. Retrieval practice followed, during which some pairs appeared in a cued recall test. A final test of associative recognition (with remember-know judgments) found lower accuracy and hit rate for nonpracticed pairs belonging to retrieval-practiced categories. In Experiment 2, subjects studied noun pairs from overlapping sets, with study duration manipulated between subjects. Retrieval practice was manipulated by presenting some members of a set in a previous block during the recognition test. With long study duration, retrieval interference was evident in both recognition and remember judgments. With short study duration, it appeared only in remember judgments. These results support a dual-process account in which retrieval interference is specific to recollection and becomes evident in recognition performance only when recollection is sufficiently dominant.     

加工水平对普通话与粤语记忆语言依赖效应的影响

通过两个实验,考察了加工水平对普通话和粤语记忆语言依赖效应的影响。实验1考察了在有意学习中的记忆语言依赖效应。结果表明,测验的语言环境对再认有重要影响。被试在粤语测验环境下对学习过的韩国语词再认反应时短,误报率却低,辨别力强,“记得”反应占优势。实验2考察了无意学习中的记忆语言依赖效应。结果表明,被试在粤语测验环境下,对评定过的韩国语词再认反应时短,误报率却高,辨别力差,但“记得”反应占优势;但在普通话测验环境下,对评定过的韩国语词再认反应时长,误报率却低,辨别力强,“知道”和“猜测”反应增加。整个研究表明,记忆的语言依赖效应在有意学习和无意学习中都存在,但有不同的趋势,表明加工水平是影响普通话和粤语记忆语言依赖效应的重要变量