John Dunlosky

Importance of the kind of cue for judgments of learning (JOL) and the delayed-JOL effect

The delayed-JOL effect is the finding in which judgments of learning (JOLs) are more accurate at predicting eventual recall when they are made a short time after study rather than immediately after study. The present research replicated this effect and found that the kind of cue that is used for JOLs is critical. In particular, following the study of stimulus-response paired associates, there is an extremely robust delayed-JOL effect when the cue for JOLs is the stimulus alone (every one of 45 subjects showed the effect); however, there is little, if any, delayed-JOL effect when the cue for JOLs is the stimulus-response pair. This finding has important implications for education: To have the greatest accuracy at predicting eventual recall, a person should make JOLs not immediately after study but, instead, shortly after study (i.e., delayed JOLs) with the cue for JOLs being the stimulus alone. The theoretical mechanisms for the delayed-JOL effect are currently unknown, but some speculations are offered.

Age-Related Differences in Absolute but Not Relative Metamemory Accuracy

In 3 experiments, the effects of age on different kinds of metacognitive prediction accuracy were assessed. participants made global memory predictions and item-by-item memory predictions in a single experimental task. Metacognitive accuracy was evaluated with correlational and more traditional difference-score measures. Difference-score measures were found, in some cases, to be sensitive to level of recall performance. Correlational techniques revealed that older adults monitored learning effectively. Relative to younger adults, they showed equally accurate immediate judgments of learning (JOLs), produced an equivalent delayed-JOL effect, and showed equivalent upgrading in the accuracy of their global prediction from before to after study of test materials.

Aging and deficits in associative memory : What is the role of strategy production?

A new method was developed to investigate the degree to which age differences in strategy production mediate age differences in paired-associate recall. Participants were instructed to use imagery or any strategy and were to report the strategy produced for each item. Age similarities in reported strategy production were found for related (Experiment 1) and unrelated (Experiment 2) word pairs: Both age groups (a) reported using effective mediators (imagery and sentence generation) more often than using no mediators and (b) complied with instructions to use imagery. Although individual differences in strategy production were related to differences in recall performance, differential strategy production accounted for little of the age differences evident in associative memory.

Age differences in the allocation of study time account for age differences in memory performance

How aging affects the utilization of monitoring in the allocation of study time was investigated by having adults learn paired associates during multiple study-test trials. During each trial, a subject paced the presentation of individual items and later judged the likelihood of recalling each item on the upcoming test; after all items had been studied and judged, recall occurred. For both age groups in Study 1, (1) people’s judgments were highly accurate at predicting recall and (2) intraindividual correlations between judgments (or recall) on one trial, and study times on the next trial were negative, which suggests that subjects utilized monitoring to allocate study time. However, the magnitude of these correlations was less for older than for younger adults. Study 2 revealed that these differences were not due to age differences in forgetting. Results from both studies suggest that older adults do not utilize on-line monitoring to allocate study to the same degree as younger adults do, and that these differences in allocation contribute to age deficits in recall.

The rereading effect: Metacomprehension accuracy improves across reading trials

Guided by a hypothesis that integrates principles of monitoring from a cue-based framework of metacognitive judgments with assumptions about levels of text representation derived from theories of comprehension, we discovered that rereading improves metacomprehension accuracy. In Experiments 1 and 2, the participants read texts either once or twice, rated their comprehension for each text, and then were tested on the material. In both experiments, correlations between comprehension ratings and test scores were reliably greater for participants who reread texts than for participants who read texts only once. Furthermore, in contrast to the low levels of accuracy typically reported in the literature, rereading produced relatively high levels of accuracy, with the median gamma between ratings and test performance being +.60 across participants from both experiments. Our discussion focuses on two alternative hypotheses—that improved accuracy is an artifact of when judgments are collected or that it results from increased reliability of test performance—and on evidence that is inconsistent with these explanations for the rereading effect.

Encoding fluency is a cue used for judgments about learning.

The authors used paired-associate learning to investigate the hypothesis that the speed of generating an interactive image (encoding fluency) influenced 2 metacognitive judgments: judgments of learning (JOLs) and quality of encoding ratings (QUEs). Results from Experiments 1 and 2 indicated that latency of a keypress indicating successful image formation was negatively related to both JOLs and QUEs even though latency was unrelated to recall. Experiment 3 demonstrated that when concrete and abstract items were mixed in a single list, latency was related to concreteness, judgments, and recall. However, item concreteness and fluency influenced judgments independently of one another. These outcomes suggest an important role of encoding fluency in the formation of metacognitive judgments about learning and future recall.

What makes people study more? An evaluation of factors that affect self-paced study

Allocation of study time across items was investigated in three experiments. According to the norm-affects-allocation hypothesis, when studying an item, a person changes the sought-after degree of learning for the item (called the norm of study) in an attempt to achieve task goals. As the norm of study is increased, more time will be allocated for study. This hypothesis was evaluated by having people pace their study of items for an eventual test of recall. As predicted, study time was greater (a) when points awarded for recalling an item increased, (b) when instructions emphasized mastering each item rather than quickly learning each item, (c) when points deducted for each second of study decreased, and (d) when the likelihood an item would be on the test increased. Also, although allocating more study time was usually accompanied by an increase in eventual recall, under several conditions peoples allocation of study time appeared sub-optimal. Discussion focuses on current theory of self-paced study and peoples apparent sub-optimal allocation of study time.

A Revised Methodology for Research on Metamemory: Pre-judgment Recall And Monitoring (PRAM).

A revised methodology is described for research on metacognitive monitoring, especially judgments of learning (JOLs), to investigate psychological processing that previously has been only hypothetical and unobservable. During data collection a new stage of recall occurs just prior to the JOL, so that during data analysis the items can be partitioned into subcategories to measure the degree of JOL accuracy in ways that are more analytic than was previously possible. A weighted-average combinatorial rule allows the component measures of JOL accuracy to be combined into the usual overall measure of metacognitive accuracy. An example using the revised methodology offers a new explanation for the delayed-JOL effect, in which delayed JOLs are more accurate than immediate JOLs for predicting recall.

Aging and monitoring associative learning: Is monitoring accuracy spared or impaired?

Mixed lists of associatively related and unrelated paired associates were used to study monitoring of associative learning. Older and younger adults produced above-chance levels of relative accuracy, as measured by intraindividual correlations (gamma) of judgments of learning (JOLs) with item recall. JOLs were strongly influenced by relatedness, and this effect was greater for older adults. Relative accuracy was higher for unrelated than for related pairs. Correlations of JOLs with item recall for a randomly yoked learner indicated that access to ones own encoding experiences increased relative accuracy. Both age groups manifested a contrast effect (lower JOLs for unrelated items when mixed with related items). Aging appears to spare monitoring of encoding, even though it adversely affects associative learning.

Are performance predictions for text based on ease of processing

In 4 experiments, the authors evaluated the hypothesis that performance predictions for text are based on ease of processing. In each experiment, participants read texts, predicted their performance for each one, and then were tested. Ease of processing was manipulated by having participants read texts that varied in coherence. Coherence was varied by manipulating causal relatedness across sentence pairs (Experiments 1 and 2) and by altering the structure of sentences within paragraphs (Experiment 3). In these experiments, prediction magnitudes increased as coherence increased, suggesting that predictions were based on processing ease. In Experiment 4, prediction magnitudes were greater for intact paragraphs than for paragraphs with letters deleted from some of the words. Discussion focuses on resolving apparent inconsistencies in the literature concerning whether processing ease influences performance predictions.