Distributed practice in verbal recall tasks: A review and quantitative synthesis

The authors performed a meta-analysis of the distributed practice effect to illuminate the effects of temporal variables that have been neglected in previous reviews. This review found 839 assessments of distributed practice in 317 experiments located in 184 articles. Effects of spacing (consecutive massed presentations vs. spaced learning episodes) and lag (less spaced vs. more spaced learning episodes) were examined, as were expanding interstudy interval (ISI) effects. Analyses suggest that ISI and retention interval operate jointly to affect final-test retention; specifically, the ISI producing maximal retention increased as retention interval increased. Areas needing future research and theoretical implications are discussed.     

Effect of delay on recognition decisions: Evidence for a criterion shift

Recent evidence indicates that in intermixed recognition testing of different stimulus classes, people can apply different decision criteria (a criterion shift) to stimulus classes distinguished by the study-test delay (Singer, Gagnon, & Richards, 2002), but not by a conspicuous strength manipulation (Stretch & Wixted, 1998b). In an attempt to reconcile these differences, we applied Singer et al.'s text retrieval method to word recognition. People first studied blocked items from each of five categories. After a delay, five new category lists were presented. After each one, the participants recognized intermixed targets and distractors from the current category and one of the earlier ones. At delays of up to 40 min, the answering criteria for immediate and delayed categories were indistinguishable. At delays of 2 days, in contrast, however, both yes-no and confidence-rating data indicated that more lenient criteria were applied to delayed than to immediate test items. This suggests that people can use the delay between study and test to flexibly adjust the decision criteria of word recognition.     

On the nature of the decision axis in signal-detection-based models of recognition memory

Most models of recognition memory involve a signal-detection component in which a criterion is placed along a decision axis. Older models generally assume a familiarity-decision axis, but newer models often assume a likelihood ratio axis instead because it allows for a more natural account of the ubiquitous mirror effect. In 3 experiments reported here, item strength was differentially manipulated to see whether a mirror effect would occur. Within a list, the items from 1 category were strengthened by repetition, but the items from another category were not. On the subsequent recognition test, the hit rate was higher for the strong category, but the false-alarm rates for the weak and strong categories were the same (i.e., no mirror effect was observed). This result suggests that the decision axis represents a familiarity scale and that participants adopt a single decision criterion that they maintain throughout the recognition test.     

Cognitive theories as reinforcement history surrogates: the case of likelihood ratio models of human recognition memory

B. F. Skinner (1977) once argued that cognitive theories are essentially surrogates for the organism's (usually unknown) reinforcement history. In this article, we argue that this notion applies rather directly to a class of likelihood ratio models of human recognition memory. The point is not that such models are fundamentally flawed or that they are not useful and should be abandoned. Instead, the point is that the role of reinforcement history in shaping memory decisions could help to explain what otherwise must be explained by assuming that subjects are inexplicably endowed with the relevant distributional information and computational abilities. To the degree that a role for an organism's reinforcement history is appreciated, the importance of animal memory research in understanding human memory comes into clearer focus. As Skinner was also fond of pointing out, it is only in the animal laboratory that an organism's history of reinforcement can be precisely controlled and its effects on behavior clearly understood.     

ROCs in Eyewitness Identification: Instructions versus Confidence Ratings

From the perspective of signal detection theory, different lineup instructions may induce different levels of response bias. If so, then collecting correct and false identification rates across different instructional conditions will trace out the receiver operating characteristic (ROC)—the same ROC that, theoretically, could also be traced out from a single instruction condition in which each eyewitness decision is accompanied by a confidence rating. We tested whether the two approaches do in fact yield the same ROC. Participants were assigned to a confidence rating condition or to an instructional biasing condition (liberal, neutral, unbiased, or conservative). After watching a video of a mock crime, participants were presented with instructions followed by a six‐person simultaneous photo lineup. The ROCs from both methods were similar, but they were not exactly the same. These findings have potentially important policy implications for how the legal system should go about controlling eyewitness response bias.Copyright © 2017 John Wiley & Sons, Ltd.     

Strong memories are hard to scale

People are generally skilled at using a confidence scale to rate the strength of their memories over a wide range. Specifically, low-confidence recognition decisions are often associated with close-to-chance accuracy, whereas high-confidence recognition decisions can be associated with close-to-perfect accuracy. However, using a 20-point rating scale, the authors found that the ability to scale memory strength had its limitations in that a high proportion of list items received the highest rating of 20. Efforts to induce participants to differentiate between these strong memories using emphatic instructions and alternative scales were not successful. Remember/know judgments indicated that these strong and hard-to-scale memories were often based on familiarity (not just recollection). Providing error feedback on a plurals discrimination task finally produced a high-confidence criterion shift. The authors suggest that the ability to scale strong (and almost perfectly accurate) memories may be limited because of the absence of differential error feedback for very strong memories in the past (the kind of differential error feedback that may account for the memory-scaling expertise that participants otherwise exhibit).     

Useful scientific theories are useful: A reply to Rouder, Pratte, and Morey (2010)

In a recognition memory experiment, Mickes, Wixted, and Wais (2007) reported that distributional statistics computed from ratings made using a 20-point confidence scale (which showed that the standard deviation of the ratings made to lures was approximately 0.80 times that of the targets) essentially matched the distributional statistics estimated indirectly by fitting a Gaussian signal-detection model to the receiver-operating characteristic (ROC). We argued that the parallel results serve to increase confidence in the Gaussian unequal-variance model of recognition memory. Rouder, Pratte, and Morey (2010) argue that the results are instead uninformative. In their view, parametric models of latent memory strength are not empirically distinguishable. As such, they argue, our conclusions are arbitrary, and parametric ROC analysis should be abandoned. In an attempt to demonstrate the inherent untestability of parametric models, they describe a non-Gaussian equal-variance model that purportedly accounts for our findings just as well as the Gaussian unequal-variance model does. However, we show that their new model—despite being contrived after the fact and in full view of the to-be-explained data—does not account for the results as well as the unequal-variance Gaussian model does. This outcome manifestly demonstrates that parametric models are, in fact, testable. Moreover, the results differentially favor the Gaussian account over the probit model and over several other reasonable distributional forms (such as the Weibull and the lognormal).     

Recognition memory and the hippocampus: A test of the hippocampal contribution to recollection and familiarity

It has been suggested that the hippocampus selectively supports recollection and that adjacent cortex in the medial temporal lobe can support familiarity. Alternatively, it has been suggested that the hippocampus supports both recollection and familiarity. We tested these suggestions by assessing the performance of patients with hippocampal lesions on recognition memory tests that differ in the extent to which recollection and familiarity contribute to the recognition decision. When targets and foils are highly similar, prior evidence suggests that, on a forced-choice test in which targets are presented together with highly similar, corresponding foils (the FC-C format), performance is supported primarily by familiarity. By contrast, when targets are presented together with foils that are similar to other targets (the FC-NC format) or when memory is tested in a yes/no (Y/N) format, performance is based much more strongly on recollection. Accordingly, a finding that hippocampal damage impaired both Y/N recognition and FC-NC recognition but spared FC-C recognition would suggest that the hippocampus selectively supports recollection. We administered Y/N, FC-C, and FC-NC tests to five memory-impaired patients with circumscribed hippocampal lesions and 14 controls. The patients were impaired on all three types of recognition test, and there was no indication that the patients were disproportionately benefited or disproportionately impaired on any test. This pattern of performance suggests that the hippocampus supports both recollection and familiarity.The medial temporal lobe (the hippocampus plus the entorhinal, perirhinal, and parahippocampal cortices) is essential for the recognition of past experience. The capacity for recognition is widely thought to depend on two distinct processes, recollection and familiarity (Mandler 1980; Yonelinas et al. 2002; Wixted 2007). Recollection involves remembering specific details about the episode in which an item was encountered. Familiarity involves simply knowing that an item was presented, even when no additional information can be retrieved about the learning episode itself. This distinction has been prominent in recent discussions about memory, particularly in relation to its possible anatomical basis. One proposal is that recollection depends on the hippocampus and that familiarity depends on the adjacent medial temporal lobe cortex (Brown and Aggleton 2001; Mayes et al. 2002; Yonelinas et al. 2002). Alternatively, it has been suggested that the hippocampus is important for both recollection and familiarity and that the distinction between these two processes does not illuminate the functional differences between the hippocampus and adjacent cortex (Rutishauser et al. 2006; Wais et al. 2006; Squire et al. 2007; Wixted 2007).To clarify the role of the hippocampus in recognition memory, the performance of memory-impaired patients with hippocampal damage has often been compared to that of controls on tests that differ in the extent to which recollection and familiarity contribute to the recognition memory decision. One such approach involves the use of highly similar targets and foils tested using a yes/no (Y/N) format and a forced-choice corresponding (FC-C) format (Holdstock et al. 2002; Westerberg et al. 2006; Bayley et al. 2008). In the case of the Y/N format, participants see a list of targets intermixed with foils (each similar to one of the targets) and are asked to respond “yes” to the targets and “no” to the foils. In the FC-C format, participants see a target presented together with its corresponding, similar foil and are asked to identify the target. According to an idea advanced by Norman and O''Reilly (2003), familiarity can support FC-C recognition because one can make effective use of small but consistent differences between the familiarity signals triggered by the target and foil items. That is, when targets and foils are highly similar and are presented together in the FC-C format, familiarity for the target is likely to slightly and reliably exceed that of the similar foil, allowing the target to be correctly chosen on the basis of familiarity. By contrast, in the Y/N format, slight differences between the familiarity signals of target items and their corresponding foils no longer provide useful information, because the targets and their corresponding foils are not presented together at test. Accordingly, good performance with the Y/N format is more dependent on recollection than with the FC-C format.The same explanation may account for why performance on FC-C tests sometimes exceeds performance on forced-choice noncorresponding (FC-NC) tests (e.g., Hintzman 1988). In the FC-NC format, target items are presented together with a noncorresponding foil that is similar to another target item from the study list (Fig. 1). Thus, the FC-C and FC-NC tests differ only in that the similar targets and foils are presented together in the former but not in the latter. When the corresponding targets and foils are presented together (in the FC-C format), participants can make effective use of consistent differences in familiarity values (as discussed above). When the corresponding targets and foils are not presented together, participants cannot make use of the small differences that they might detect through a side-by-side comparison.Open in a separate windowFigure 1.Test format and materials for the three kinds of recognition test. For each test, 12 images, either objects (A) or silhouettes (B), were presented at study, and memory was tested in one of three ways. For the forced-choice corresponding test (FC-C), each target (a studied item) was presented together with a highly similar foil (a new item). For the forced-choice noncorresponding test (FC-NC), each target was presented together with a foil that was highly similar to a different target from the study list. Participants were asked to point to exactly the same image they had seen during study. For the yes/no test (Y/N), the 12 targets and the 12 foils (each similar to one of the targets) were intermixed and presented one at a time. Participants were asked to respond “yes” if they had seen exactly the same image before and “no” if they had not. Asterisks identify the target items.A recent study (Migo et al. 2009) provides empirical support for the suggestion that familiarity primarily contributes to the recognition decision in the FC-C format, whereas recollection contributes more strongly in the Y/N format and in the FC-NC format. In the Migo et al. (2009) study, healthy participants received FC-C, FC-NC, and Y/N tests after receiving either standard instructions to make their decisions based on familiarity or recollection or modified instructions to base their decisions on familiarity only. On the FC-NC test and on the Y/N test, performance using familiarity alone was significantly worse than under standard instructions. On the FC-C test, performance using familiarity alone was nearly as good as under standard instructions. This result supports the earlier suggestion that FC-C is primarily supported by familiarity, whereas recollection plays a larger role in FC-NC and Y/N recognition (Norman and O''Reilly 2003).If the hippocampus is critical for recollection but not for familiarity, and if good performance on the FC-C format can be achieved using familiarity alone, then patients with focal hippocampal lesions should be differentially impaired on the Y/N and FC-NC recognition test formats compared with the FC-C format. Three recent studies investigated this issue by assessing the effects of hippocampal damage (or presumed hippocampal damage) on Y/N and FC-C tests that used highly similar targets and foils (the FC-NC format was not used in these earlier studies). All three studies used black-and-white silhouette images of objects, and the FC-C test involved the target item and three similar foils (i.e., multiple-choice with four alternatives). In one study, a single patient with bilateral hippocampal damage (patient Y.R.) was impaired on the test of Y/N recognition but was unimpaired on the test of FC-C recognition (Holdstock et al. 2002). A similar result was reported in patients with a diagnosis of mild cognitive impairment (MCI) (Westerberg et al. 2006). These findings have been taken to support the suggestion that the hippocampus selectively supports recollection. However, the study by Holdstock and colleagues (2002) involved a single patient (Y.R.), and findings from a single patient need not agree with findings obtained from a group of patients (see Discussion). In addition, the Westerberg et al. (2006) study involved individuals with a diagnosis of MCI for whom anatomical data were not available. It is therefore not clear what the status of the hippocampus was in these patients. Moreover, a standard Y/N procedure involving 12 study items and 24 test items (as opposed to the 12 study test items and 60 test items used in previous studies [Holdstock et al. 2002; Westerberg et al. 2006; Bayley et al. 2008]) might be better suited to address the question of whether Y/N recognition is selectively impaired in hippocampal patients. Bayley et al. (2008) tested five patients with circumscribed hippocampal damage using the same test materials and procedure as in Holdstock et al. (2002). When all 60 test items of the Y/N test were scored, the patients were found to be more impaired on the Y/N test than on the FC-C test. However, when only the first 24 test items of the Y/N test were scored (according to the standard Y/N procedure), the patients were found to be impaired on both the Y/N and FC-C tests to a similar degree.The earlier studies (Holdstock et al. 2002; Westerberg et al. 2006; Bayley et al. 2008) compared Y/N recognition to four-alternative FC-C recognition in which each target was presented along with three corresponding foils. None of the earlier studies investigated how patients with hippocampal lesions performed on FC-NC recognition. As Migo et al. (2009) point out, FC-C and FC-NC tests are better matched than the FC-C and Y/N tests because they are both forced-choice tests, and both use the same number of test trials. As such, they argue, it would be more useful to compare FC-C recognition performance to FC-NC recognition performance to determine whether patients with hippocampal lesions are intact or impaired on familiarity-based tests.The present study assessed Y/N, two-alternative FC-C, and two-alternative FC-NC recognition, using highly similar targets and foils. The tests were given to five patients with circumscribed hippocampal damage and 14 matched controls. The Y/N test involved an equal number of targets and foils (unlike the design used in the prior studies), and a two-alternative format was used for the forced-choice tests to make them comparable to the Y/N test (i.e., all three tests involved the same number of targets and foils). The FC-C and FC-NC tests were identical except with respect to how the targets and their corresponding foils were paired on the recognition test.To extend our findings beyond the stimuli used in prior studies, we also used color photographs of objects in addition to black-and-white silhouettes. Using these three test formats, we asked the following questions: (1) Does hippocampal damage impair Y/N recognition memory but spare, or disproportionately benefit, FC-C recognition (as suggested by the view that the hippocampus supports recollection and not familiarity)? Or does hippocampal damage impair Y/N and FC-C recognition similarly (as suggested by the view that the hippocampus supports both recollection and familiarity)? (2) Does hippocampal damage disproportionately benefit FC-C recognition relative to FC-NC recognition, or does hippocampal damage impair FC-C and FC-NC recognition similarly?     

The effects of tests on learning and forgetting

In three experiments, we investigated whether memory tests enhance learning and reduce forgetting more than additional study opportunities do. Subjects learned obscure facts (Experiments 1 and 2) or Swahili-English word pairs (Experiment 3) by either completing a test with feedback (test/study) or receiving an additional study opportunity (study). Recall was tested after 5 min or 1, 2, 7, 14, or 42 days. We explored forgetting by means of an ANOVA and also by fitting a power function to the data. In all three experiments, testing enhanced overall recall more than restudying did. According to the power function, in two out of three experiments, testing also reduced forgetting more than restudying did, although this was not always the case according to the ANOVA. We discuss the implications of these results both for approaches to measuring forgetting and for the use of tests in promoting long-term retention. The stimuli used in these experiments may be found at www.psychonomic.org/archive.