False recognition occurs more frequently during source identification than during old-new recognition

Boosting confidence without boosting performance: item strength creates the illusion of source accuracy

People often express high confidence for misremembered sources. Starns and Ksander ([2016]. Item strength influences source confidence and alters source memory zROC slopes. Journal of Experimental Psychology: Learning, Memory, and Cognition, 42(3), 351-365; hereafter SK16) found that this happens more often when a person is highly confident in memory for the item itself, and that simply increasing item memory can increase high-confidence source errors. Under the decision heuristic account, this pattern emerges because strong item memories contaminate source judgments by promoting high confidence responses even when source evidence is relatively weak. Consequently, strengthening item memory is predicted to increase confidence for both correct and incorrect source responses; however, SK16 could not assess this key prediction because their item-strength manipulation also impaired source memory. We report two experiments with new item-strengthening manipulations designed to minimise source memory impairments. Results replicated the evidence for the decision heuristic account reported by SK16 and provided additional support by showing a boost in source confidence for both correct and error responses when item memory was strengthened without accompanying source impairments .

How rich are false memories in a naturalistic context in healthy aging?

The recall of factual and contextual information is a core characteristic of episodic memory sensitive to aging effects. The innovative aim of the present study was to assess in a naturalistic context the quantity and quality of correct and false free recalls among younger and older adults considering feature binding (What-Where-When-Details) and recollection (Remembering vs. Knowing). Thanks to virtual reality, we designed a multimodal environment simulating a lively town in which we implemented a variant of a DRM task rich in sets of semantically related items (e.g., fruits on a market stall). We asked 30 young and 30 older participants to navigate in the virtual environment, paying attention to the items, and then recall as many items and as much contextual information as possible and indicate the presence of recollection. As expected, older adults produced fewer correct recall but more intrusions than younger adults, and their correct recall was more deficient in binding and recollection. In both age groups, false recall was associated with the correct context inferred from a related set of items. However, the intrusions produced by older adults were highly recollected compared to those of the younger adults, and they were associated with false item-related contextual information.

The influence of semantic context on false memories

Although false memories have largely been examined with the Deese-Roediger-McDermott (DRM) paradigm, little research has focused on the semantic context in which associates are encoded. Across three experiments, we varied semantic context during a sentence-processing task with DRM associates embedded within sentences. More meaningful sentences resulted in greater memory errors (Experiment 1). Furthermore, providing contextual information to discriminate old from new items did not reduce false alarms relative to encoding words in isolation when sentences converged on the meaning of the critical lure (Experiment 2), and actually increased memory errors (Experiment 3). These results suggest that semantic context that allows for meaningful relational processing of items within-lists and that converge on the semantic meaning of the critical lure increases the likelihood that the list theme is identified, resulting in more errors at test.

Does Test Type Influence False Recognition in the DRM Paradigm? Comparison of the Yes/No Recognition Test and Two-Alternative Forced-Choice Test

Studies attempting to identify the mechanisms that are responsible for the level of false recognition in the DRM paradigm usually apply a manipulation during encoding. The main aim of the studies reported here was to investigate the within- and between-participant effects of a testing method on memory performance using a standard yes/no recognition test and a 2-alternative forced-choice recognition test (2FC). To allow a direct comparison of the 2 testing methods, a 2FC test containing similar items as a typical yes/no test in the DRM paradigm was elaborated on in the pilot study. Moreover, 2 methods of data calculation were used: comparing rates of hit and false alarms with critical lures and with unstudied and unrelated items between the 2 tests and comparing measures of sensitivity (d’) derived from signal detection theory. Both experiments showed a lower false alarm rate to critical lures and higher hit rate in a 2FC test as compared with a yes/no test, depicting a typical mirror effect. A within-participant design (Experiment 1) also showed that this increased accuracy of a 2FC test diminished when memory performance was expressed in terms of a sensitivity measure, which may suggest that similar processes are used during these 2 retrieval methods. A similar analysis performed for a between-participants design (Experiment 2) revealed that a 2FC test was less susceptible to associative memory distortions, but the quality of memory (“remember/know” judgments) remained similar for both tests.

Item-specific and relational processing both improve recall accuracy in the DRM paradigm

Using the Deese-Roediger-McDermott (DRM) paradigm, Huff and Bodner found that both item-specific and relational variants of a task improved correct recognition, but only the item-specific variants reduced false recognition, relative to a read-control condition. Here, we examined the outcome pattern when memory was tested using free recall, using the same item-specific versus relational task variants across three experiments as our previous study (processing instructions, pleasantness ratings, anagram generation). The outcome pattern in recall was similar to recognition, except relational processing at study actually reduced the DRM illusion, though not as much as item-specific processing. To reconcile this task difference, we suggest that the memory information laid down during relational encoding enhances the familiarity of the critical items at test. To the extent that familiarity is used less as a basis for responding in free recall than in recognition, relational processing ironically reduces rather than increases the DRM illusion in recall.

The neurocognitive basis of borrowed context information

Falsely remembered items can be accompanied by episodic context retrieval. This finding is difficult to explain because there is no episode that binds the remembered item to the experimenter-controlled context features. The current study examines the neural correlates of false context retrieval when the context features can be traced to encoding episodes of semantically-similar items. Our neuroimaging results support a "dissociated source" mechanism for context borrowing in false memory. We found that parahippocampal cortex (PHc) activation, thought to indicate context retrieval, was greater during trials that involved context borrowing (an incorrect, but plausible source decision) than during baseline correct context retrieval. In contrast, hippocampal activation, thought to indicate retrieval of an episodic binding, was stronger during correct source retrieval than during context borrowing. Vivid context retrieval during false recollection experiences was also indicated by increased activation in visual perceptual regions for context borrowing as compared to other incorrect source judgments. The pattern of findings suggests that context borrowing can arise when unusually strong activation of a semantically-related item's contextual features drives relatively weak retrieval of the associated episodic binding with failure to confirm the item information within that binding. This dissociated source retrieval mechanism suggests that context-driven episodic retrieval does not necessarily lead to retrieval of specific item details. That is, source information can be retrieved in the absence of item memory.

Getting at the source of distinctive encoding effects in the DRM paradigm: evidence from signal-detection measures and source judgments

Studying Deese-Roediger-McDermott (DRM) lists using a distinctive encoding task can reduce the DRM false memory illusion. Reductions for both distinctively encoded lists and non-distinctively encoded lists in a within-group design have been ascribed to use of a distinctiveness heuristic by which participants monitor their memories at test for distinctive-task details. Alternatively, participants might simply set a more conservative response criterion, which would be exceeded by distinctive list items more often than all other test items, including the critical non-studied items. To evaluate these alternatives, we compared a within-group who studied 5 lists by reading, 5 by anagram generation, and 5 by imagery, relative to a control group who studied all 15 lists by reading. Generation and imagery improved recognition accuracy by impairing relational encoding, but the within group did not show greater memory monitoring at test relative to the read control group. Critically, the within group's pattern of list-based source judgments provided new evidence that participants successfully monitored for distinctive-task details at test. Thus, source judgments revealed evidence of qualitative, recollection-based monitoring in the within group, to which our quantitative signal-detection measure of monitoring was blind.

The influence of forward and backward associative strength on false memories for encoding context

Two experiments examined the effects of forward associative strength (FAS) and backward associative strength (BAS) on false recollection of unstudied lure items. Themes were constructed such that four associates were strongly related to a lure item in terms of FAS or BAS and four associates were weakly related to a lure item in terms of FAS or BAS. Further, when FAS was manipulated, BAS was controlled across strong and weak associates, while FAS was controlled across strong and weak associates when BAS was manipulated. Strong associates were presented in one font while weak associates were presented in a second font. At test, lure items were disproportionately attributed to the source used to present lures' strong associates compared to lures' weak associates, both when BAS was manipulated and when FAS was manipulated. This outcome demonstrates that both BAS and FAS influence lure item false recollection, which favours global-matching models' explanation of false recollection over the explanation offered by spreading activation theories.

Judgments of Learning are Influenced by Multiple Cues In Addition to Memory for Past Test Accuracy

When people try to learn new information (e.g., in a school setting), they often have multiple opportunities to study the material. One of the most important things to know is whether people adjust their study behavior on the basis of past success so as to increase their overall level of learning (for example, by emphasizing information they have not yet learned). Monitoring their learning is a key part of being able to make those kinds of adjustments. We used a recognition memory task to replicate prior research showing that memory for past test outcomes influences later monitoring, as measured by judgments of learning (JOLs; confidence that the material has been learned), but also to show that subjective confidence in whether the test answer and the amount of time taken to restudy the items also have independent effects on JOLs. We also show that there are individual differences in the effects of test accuracy and test confidence on JOLs, showing that some but not all people use past test experiences to guide monitoring of their new learning. Monitoring learning is therefore a complex process of considering multiple cues, and some people attend to those cues more effectively than others. Improving the quality of monitoring performance and learning could lead to better study behaviors and better learning. An individual's memory of past test performance (MPT) is often cited as the primary cue for judgments of learning (JOLs) following test experience during multi-trial learning tasks (Finn & Metcalfe, 2007; 2008). We used an associative recognition task to evaluate MPT-related phenomena, because performance monitoring, as measured by recognition test confidence judgments (CJs), is fallible and varies in accuracy across persons. The current study used multilevel regression models to show the simultaneous and independent influences of multiple cues on Trial 2 JOLs, in addition to performance accuracy (the typical measure of MPT in cued-recall experiments). These cues include recognition CJs, perceived recognition fluency, and Trial 2 study time allocation (an index of reprocessing fluency). Our results expand the scope of MPT-related phenomena in recognition memory testing to show independent effects of recognition test accuracy and CJs on second-trial JOLs, while also demonstrating individual differences in the effects of these cues on JOLs (as manifested in significant random effects for those regression effects in the model). The effect of study time on second-trial JOLs controlling on other variables, including Trial 1 recognition memory accuracy, also demonstrates that second-trial encoding behavior influence JOLs in addition to MPT.

Hemispheric asymmetries in the activation and monitoring of memory errors

Previous research on the lateralization of memory errors suggests that the right hemisphere's tendency to produce more memory errors than the left hemisphere reflects hemispheric differences in semantic activation. However, all prior research that has examined the lateralization of memory errors has used self-paced recognition judgments. Because activation occurs early in memory retrieval, with more time to make a decision, other memory processes, like strategic monitoring processes, may affect memory errors. By manipulating the time subjects were given to make memory decisions, this study separated the influence of automatic memory processes (activation) from strategic memory processes (monitoring) on the production of false memories. The results indicated that when retrieval was fast, the right hemisphere produced more memory errors than the left hemisphere. However, when retrieval was slow, the left hemisphere's error-proneness increased compared to the fast retrieval condition, while the right hemisphere's error-proneness remained the same. These results suggest that the right hemisphere's errors are largely due to activation, while the left hemisphere's errors are influenced by both activation and monitoring.

Overdistribution in source memory

Semantic false memories are confounded with a second type of error, overdistribution, in which items are attributed to contradictory episodic states. Overdistribution errors have proved to be more common than false memories when the 2 are disentangled. We investigated whether overdistribution is prevalent in another classic false memory paradigm: source monitoring. It is. Conventional false memory responses (source misattributions) were predominantly overdistribution errors, but unlike semantic false memory, overdistribution also accounted for more than half of true memory responses (correct source attributions). Experimental control of overdistribution was achieved via a series of manipulations that affected either recollection of contextual details or item memory (concreteness, frequency, list order, number of presentation contexts, and individual differences in verbatim memory). A theoretical model was used to analyze the data (conjoint process dissociation) that predicts that (a) overdistribution is directly proportional to item memory but inversely proportional to recollection and (b) item memory is not a necessary precondition for recollection of contextual details. The results were consistent with both predictions.

The role of memory activation in creating false memories of encoding context

Using 3 experiments, I examined false memory for encoding context by presenting Deese-Roediger-McDermott themes (Deese, 1959; Roediger & McDermott, 1995) in usual-looking fonts and by testing related, but unstudied, lure items in a font that was shown during encoding. In 2 of the experiments, testing lure items in the font used to study their associated themes increased false recognition relative to testing lure items in a font that was used to study a different lure's theme. Further, studying a larger number of associates exacerbated the influence of testing lure items in a font used to study their associated themes. Finally, testing lures in a font that was encoded many times, but was not used to present the lures' studied associates, increased lure errors more than testing lures in a font that was encoded relatively fewer times. These results favor the explanation of false recognition offered by global-matching models of recognition memory over the explanations of activation-monitoring theory and fuzzy-trace theory. (PsycINFO Database Record (c) 2009 APA, all rights reserved).

The effect of study modality on false recognition

A number of previous studies have shown that false recognition of critical items in the Deese/Roediger-McDermott paradigm is reduced when study items are presented visually rather than auditorily; however, this effect has not been uniformly demonstrated. We investigated three potential boundary conditions of the effect of study modality in false recognition. Experiments 1 and 2 showed no reduction in false recognition following visual study presentation when the yes-no recognition test was not preceded by a recall test. Experiment 3 showed that visual study presentation can reduce false recognition without a preceding recall test, if the recognition test uses remember-know instructions. The order of the recognition test items did not influence the effect of visual study presentation on false recognition in Experiment 1. In general, the data imply that distinctive processing at study can reduce false memory in recognition if the test demands draw attention to the dimension of distinctive processing.

Pictorial encoding effects and memory confusions in the Deese-Roediger-McDermott paradigm: evidence for the activation of spontaneous imagery

The purpose of the experiments reported in this paper was to examine the possible role of spontaneous imagery and list-specific cues on pictorial encoding effects induced by the Deese-Roediger-McDermott (DRM) task. After viewing pictures and words referring to thematically related materials, by way of a picture/word source-judgement task, participants were asked to remember the way in which these materials were presented. Participants reported "seeing" pictures of items that were presented as words, an effect predicted by the imaginal activation hypothesis in its suggestion that incidental images experienced during encoding will later be mistaken as memories for pictures. Whether participants made the same picture misattributions on related lures (or non-presented related items) depended on the way in which the lures' respective thematic lists were experienced during encoding (Experiments 1 and 2), pointing to the effects of list-specific cues in picture/word judgements. These findings have intriguing implications for interpretations of picture-encoding effects induced by the DRM task. The findings also speak to the use of DRM false-memory rates when marshalling evidence against the use of imagery in applied settings.

Source-monitoring judgments about anagrams and their solutions: evidence for the role of cognitive operations information in memory

Generating solutions to anagrams leads to a memory advantage for those solutions, with generated words remembered better than words simply read. However, an additional advantage is not typically found for solutions to difficult anagrams relative to solutions to easy ones, presenting a challenge for the cognitive effort explanation of the generation effect. In the present series of experiments, the effect of manipulating anagram difficulty is explored further by introducing two new source-monitoring judgments. These studies demonstrate that when attention is directed at test to the operations activated during encoding (by way of source-monitoring judgments focused on solving vs. constructing anagrams), a source advantage is observed for difficult anagrams. However, when attention is directed to the anagrams themselves, asking participants to remember the kinds of anagrams generated or solved (based on kind of rule rather than subjective impressions of difficulty), a similar source advantage is not observed. The present studies bring a new perspective to the investigation of difficulty manipulations on memory for problem solving by illustrating the impact of a shift in focus from the effort mediating cognitive operations to specifics about the cognitive operations themselves.

Effects of bilateral eye movements on gist based false recognition in the DRM paradigm

The effects of saccadic bilateral (horizontal) eye movements on gist based false recognition was investigated. Following exposure to lists of words related to a critical but non-studied word participants were asked to engage in 30s of bilateral vs. vertical vs. no eye movements. Subsequent testing of recognition memory revealed that those who undertook bilateral eye movement were more likely to correctly recognise previously presented words and less likely to falsely recognise critical non-studied associates. This result joins other research in demonstrating the conditions in which false memory effects can be attenuated.

Anticipating partners' responses: examining item and source memory following interactive exchanges

Within the context of an interactive anagram-solving task, the present studies tested predictions about the role of cognitive anticipation in both source and item memory. After working in pairs to solve anagram problems, participants were surprised by a source-monitoring test focused on the source of solutions (self vs. partner, Experiment 1) or a standard recognition test focused on the solutions themselves (Experiment 2). With the intention of affecting the opportunity to anticipate partners' solutions, two variables were manipulated: anagram difficulty (easy vs. hard) and the delaybetween the presentation of an anagram problem and theprompt tha t designated one member of each pair as the anagram solver. Consistent w i th predictions, asthe opportunity t oanticipate partners'solutions increased, there was a decrease in source accuracy suggesting increased confusion about whether the solution had been self- or partner-generated. Generation-effect failures were observed in item memory. However, these failures reflected increases in item memory for partners' responses rather than decreases in memory for self-generated ones. These studies suggest that when opportunities to anticipate partners' responses are available, self-generative activities may be associated with both self-and partner-generated items, influencing the expression of the generation effect.

Two types of recollection-based monitoring in younger and older adults: Recall-to-reject and the distinctiveness heuristic

People often use recollection to avoid false memories. At least two types of recollection-based monitoring processes can be identified in the literature. Recall-to-reject is based on the recall of logically inconsistent information (which disqualifies the false event from having occurred), whereas the distinctiveness heuristic is based on the failure to recall to-be-expected information (which is diagnostic of non-occurrence). We attempted to investigate these hypothetical monitoring processes in a single task, as a first step at delineating the functional relationship between them. By design, participants could reject familiar lures by (1) recalling them from a to-be-excluded list (recall-to-reject) or (2) realising the absence of expected picture recollections (the distinctiveness heuristic). Both manipulations reduced false recognition in young adults, suggesting that these two types of monitoring were deployed on the same test. In contrast, older adults had limited success in reducing false recognition with either manipulation, indicating deficits in recollection-based monitoring processes. Depending on how a retrieval task is structured, attempts to use one monitoring process might interfere with another, especially in older adults.

Adult age differences in distinctive processing: the modality effect on false recall

Age differences in distinctive processing were investigated by examining the effects of study presentation modality on false recall in younger and older adults using the Deese/Roediger and McDermott paradigm. Participants were presented with study words either visually or auditorily. Older adults did not show the typical reduction in false recall after visual, compared to auditory, study presentation (R.E. Smith & R.R. Hunt, 1998). The authors interpret these results as evidence of reduced distinctive processing on the part of older adults.

Omitting details from post-event information: are true and false memory affected in the same way?

Participants who witness an event and later receive post-event information that omits a critical scene are less likely to recall and to recognise that scene than are participants who receive no post-event information (Wright, Loftus, & Hall, 2001). The present study used the Deese-Roediger-McDermott (DRM) paradigm, in which participants study lists of semantic associates (e.g., hot, snow, warm, winter) that commonly elicit false memories of critical non-presented words (e.g., cold), to determine whether omitting information from a second presentation decreases memory for both presented and non-presented information. Participants were presented with a list of the semantic associates of six non-presented words. For half the participants, this list was presented a second time with the semantic associates of one of the non-presented words omitted. As expected, participants were less likely to recall and to recognise the presented words when they had been omitted from the second presentation. Omission also decreased the rate at which non-presented words were recalled, although false recognition of these words was not reduced. These results suggest that false recognition may be particularly difficult to attenuate and that post-event omission may be more detrimental to memory accuracy than previously thought.

Illusory recollection of voices

We investigated source misattributions in the DRM false memory paradigm (Deese, 1959, Roediger & McDermott, 1995). Subjects studied words in one of two voices, manipulated between-lists (pure-voice lists) or within-list (mixed-voice lists), and were subsequently given a recognition test with voice-attribution judgements. Experiments 1 and 2 used visual tests. With pure-voice lists (Experiment 1), subjects frequently attributed related lures to the corresponding study voice, despite having the option to not respond. Further, these erroneous attributions remained high with mixed-voice lists (Experiment 2). Thus, even when their related lists were not associated with a particular voice, subjects misattributed the lures to one of the voices. Attributions for studied items were fairly accurate in both cases. Experiments 3 and 4 used auditory tests. With pure-voice lists (Experiment 3), subjects frequently attributed related lures and studied items to the corresponding study voice, regardless of the test voice. In contrast, with mixed-voice lists (Experiment 4), subjects frequently attributed related lures and studied items to the corresponding test voice, regardless of the study voice. These findings indicate that source attributions can be sensitive to voice information provided either at study or at test, even though this information is irrelevant for related lures.

The role of rehearsal and generation in false memory creation

The current research investigated one possible mechanism underlying false memories in the Deese-Roediger-McDermott (DRM) paradigm. In the DRM paradigm, participants who study lists of related words (e.g., "table, sitting, bench ...") frequently report detailed memories for the centrally related but non-presented critical lure (e.g., "chair"). One possibility is that participants covertly call to mind the critical non-presented lure during the study phase, and later misattribute memory for this internally generated event to its external presentation. To investigate this, the DRM paradigm was modified to allow collection of on-line thoughts during the study phase. False recognition increased following generation during study. False recognition also increased following study of longer lists; this effect was partially explained by the fact that longer lists were more likely to elicit generations of the critical lure during study. Generation of the lure during study contributes to later false recognition, although it does not explain the entire effect.

What's in a Name for Memory Errors? Implications and Ethical Issues Arising From the Use of the Term "False Memory" for Errors in Memory for Details

The term "false memories" has been used to refer to suggestibility experiments in which whole events are apparently confabulated and in media accounts of contested memories of childhood abuse. Since 1992 psychologists have increasingly used the term "false memory" when discussing memory errors for details, such as specific words within lists. Use of the term to refer to errors in details is a shift in language away from other terms used historically (e.g., "memory intrusions"). We empirically examine this shift in language and discuss implications of the new use of the term "false memories." Use of the term presents serious ethical challenges to the data-interpretation process by encouraging over-generalization and misapplication of research findings on word memory to social issues.

Effects of perceptual modality on verbatim and gist memory

In two experiments, predictions of the fuzzy-trace theory of memory were tested. Perceptual information may play a role in retrieval and recognition processes for verbatim, but not for gist, memory. Perceptual modality effects were assessed in the present study by presenting three-sentence stories (e.g., The bird is in the cage. The cage is over the table. The bird is yellow) and then testing recognition of probes that varied on three dimensions: (1) semantic accuracy (true vs. false), (2) wording (all original words vs. one novel word included), and (3) sentence type (premise vs. inference). In Experiment 1, study modality (auditory vs. visual) was manipulated, and in Experiment 2, both study and test modalities were manipulated. Despite replicating a number of findings consistent with fuzzy-trace theory (e.g., instruction and probe type effects), the results of both experiments failed to support the idea that perceptual information plays a role in performance on verbatim memory tests.

Generative processing and false memories: when there is no cost

Encoding manipulations (e.g., levels of processing) that facilitate retention often result in greater numbers of false memories, a pattern referred to as the more is less effect (M. P. Toglia, J. S. Neuschatz, & K. A. Goodwin, 1999). The present experiments explored false memories under generative processing. In Experiments 1-3, using Deese-Roediger-McDermott (DRM) lists with items that were either read or generated, the authors found recognition and recall tests indicated generation effects for true memories but no increases in false memories (i.e., generation at no cost). In Experiment 4, in a departure from the DRM methodology, a cuing procedure resulted in a more is less pattern for congruous generation,and a no cost pattern for incongruous generation. This highlights the critical distinction between these encoding contexts.

Contributions of source and inhibitory mechanisms to age-related retroactive interference in verbal working memory

Determining contributions of source-monitoring and inhibitory function to age-related forgetting has been an elusive goal for cognitive theorists. Five studies used a verbal working memory paradigm to examine mechanisms accounting for disproportionate retroactive interference (RI) experienced with adult aging. Participants distinguished studied target-word pairs from interfering pairs that were read aloud. Source-monitoring and inhibitory task components varied through manipulations of response requirements. RI effects were primarily due to source-monitoring failures rather than to inhibitory failures. Removing both source and inhibitory components eliminated age differences in RI. When source monitoring was emphasized, RI continued to be observed in all age groups but disproportionately for older adults. Process dissociation analyses of memory found recollection decreases and familiarity increases consistent with source failures.

Backward associative strength determines source attributions given to false memories

Source attributions for falsely remembered material were investigated in two experiments. A male and a female speaker each presented either an entire word list or half of the items from each of multiple Deese-Roediger-McDermott (DRM) lists commonly used in this paradigm. In the latter condition the tendency of each list half to activate a nonpresented, critical list theme item was manipulated. All of the list halves differed in backward associative strength (BAS), and each was presented by one or the other of the two speakers. In these correlated conditions, when critical items were falsely recognized (Experiments 1 and 2) or recalled (Experiment 2), source attributions were more frequently made to the speaker of the list items with the higher average BAS. This source attribution effect appears to result from the binding of list item source characteristics to activated critical items during encoding, as opposed to being the result of a biased retrieval process. The results are interpreted as consistent with an activation/monitoring account of false memory in the DRM paradigm.

Factors that determine false recall: a multiple regression analysis

In the Deese-Roediger-McDermott (DRM) paradigm, subjects study lists of words that are designed to elicit the recall of an associatively related critical item. The 55 lists we have developed provide levels of false recall ranging from .01 to .65, and understanding this variability should provide a key to understanding this memory illusion. Using a simultaneous multiple regression analysis, we assessed the contribution of seven factors in creating false recall of critical items in the DRM paradigm. This analysis accounted for approximately 68% of the variance in false recall, with two main predictors: associative connections from the study words to the critical item (r = +.73; semipartial r = +.60) and recallability of the lists (r = -.43; semipartial r = -.34). Taken together, the variance in false recall captured by these predictors accounted for 84% of the variance that can be explained, given the reliability of the false recall measures (r = .90). Therefore, the results of this analysis strongly constrain theories of false memory in this paradigm, suggesting that at least two factors determine the propensity of DRM lists to elicit false recall. The results fit well within the theoretical framework postulating that both semantic activation of the critical item and strategic monitoring processes influence the probability of false recall and false recognition in this paradigm.

The use of source memory to identify one's own episodic confusion errors

In 4 category cued recall experiments, participants falsely recalled nonlist common members, a semantic confusion error. Errors were more likely if critical nonlist words were presented on an incidental task, causing source memory failures called episodic confusion errors. Participants could better identify the source of falsely recalled words if they had deeply processed the words on the incidental task. For deep but not shallow processing, participants could reliably include or exclude incidentally shown category members in recall. The illusion that critical items actually appeared on categorized lists was diminished but not eradicated when participants identified episodic confusion errors post hoc among their own recalled responses; participants often believed that critical items had been on both the incidental task and the study list. Improved source monitoring can potentially mitigate episodic (but not semantic) confusion errors.