Cognitive structures in comprehension and memory of narrative discourse

The effects of structure and content variables on memory and comprehension of prose passages were studied in two experiments. The experimental passages exemplify a class of simple narrative stories that is described by a generative grammar of plot structures. A comprehension model is proposed that assumes a hierarchical organizational framework of stories in memory, determined by the grammar, representing the abstract structural components of the plot. The quality and characteristics of subjects' memory for stories were tested on a variety of experimental tasks in which story organization was manipulated. Comprehensibility and recall were found to be a function of the amount of inherent plot structure in the story, independent of passage content. Recall probability of individual facts from passages depended on the structural centrality of the facts: Subjects tended to recall facts corresponding to high-level organizational story elements rather than lower-level details. In addition, story summarizations from memory tended to emphasize general structural characteristics rather than specific content. For successively presented stories, both structure and content manipulations influenced recall. Furthermore, repeating story structure across two passages produced facilitation in recall of the second passage, while repeating story content produced proactive interference. The implications for a model of memory for narrative discourse are discussed.     

Relationships between metamemory, memory predictions, and memory task performance in adults

A cross-sectional sample of adults recalled categorized word lists and narrative texts. Subjects gave performance predictions before each of 3 recall trials for each task. Older subjects had poorer memory performance and also predicted lower performance levels than did younger subjects. The LISREL models suggested (a) direct effects of memory self-efficacy (MSE) on initial predictions; (b) upgrading of prediction-performance correlations across trials, determined by direct effects of performance on subsequent predictions; (c) significant effects of a higher order verbal memory factor on MSE; and (d) an independent relationship of text recall ability to initial text recall performance predictions. These results lend support to the theoretical treatment of predictions as task-specific MSE judgments.     

Predicting performance on text: Delayed versus immediate predictions and tests

North Dakota State University, Fargo, North Dakota Delayed judgments of learning for word pairs are more accurate than immediate judgments of learning when the memory test is delayed. In the present experiment, I investigated a similar paradigm with text. Participants predicted performance on texts either immediately after reading the texts or after a delay following the reading of other texts, and tests were given either immediately or after a delay. Immediate ratings with an immediate test produced the most accurate predictions, and immediate ratings with a delayed test produced less accurate predictions. Delaying both the ratings and the test did not produce more accurate predictions than immediate predictions and a delayed test. The results for delayed judgments of learning with text were different from those with word pairs.     

The influence of camera angle on comprehension and retention of pictorial events

A subset of formal compositional principles was examined in the context of narrative pictorial events. Slide stories were constructed of six common events in which characters were depicted from three camera angles: low angle, eye-level, and high angle. After presentation, subjects evaluated the characters, recalled the stories, and engaged in a recognition task. Camera angle influenced subjects’ evaluation and retention of the stories in accordance with aesthetic principles. Camera angle had significant, predictable effects on judgments of the physical and personal characteristics of the characters, on recall of these characteristics, and on recall of the gist of the stories themselves. Recognition memory for camera angle was significantly less accurate than recognition memory for the characters. Subjects used the available visual information to construct coherent story representations, and, although camera angle strongly influenced the construction of these representations, specific camera angle information became less accessible.     

Emotion and prior knowledge in memory and judged comprehension of ambiguous stories

Abstract

Two experiments investigated the effects of experimentally induced mood states on memory and judged comprehension of stories. The experiments examined the issue of whether induction of a depressed mood would affect prose memory and comprehension and impair the ability of individuals to use prior knowledge, activated by way of a title, in remembering the passage. In Experiment 1, depressed subjects who were given a title for the passage recalled fewer idea units when compared with neutral control conditions, but no depressive deficit in recall occurred in the absence of a title. In Experiment 2 the same pattern of results occurred when subjects learned two successive passages. The depressive deficits obtained were interpreted in terms of a resource allocation model which proposes that emotional states increase the production of irrelevant, competing thoughts that interfere with processes important in remembering the criterion passage. Alternative explanations involving cognitive initiative and schema theory were discussed. Finally, judgments of comprehension predicted passage recall and were better predictors for neutral than depressed mood subjects. A depressed mood state did not affect average judgments of comprehension even when recall was correspondingly impaired.     

Time of day and the representation of text

This article reports an experiment to investigate the idea that time-of-day differences in memory for text are related to type of text processing that is carried out at different times of day. In particular, it was hypothesized that subjects tested in the morning tend to retain more of the exact wording of the text, whereas those tested in the afternoon are more likely to produce a meaning-based representation. The results of a recognition test showed that, in agreement with other studies, immediate text memory is better in the morning than in the afternoon. As predicted, the subjects tested in the morning better remembered the exact wording of the original passages. However, the idea that memory for meaning is better in the afternoon than the morning was not supported by the data. The results suggest that subjects at both times of day produce a meaning-based representation of the text but, in the morning, this representation is supported by good verbatim memory, enabling accurate recall of the text at that time.     

Time of day and the representation of text

This article reports an experiment to investigate the idea that time-of-day differences in memory for text are related to type of text processing that is carried out at different times of day. In particular, it was hypothesized that subjects tested in the morning tend to retain more of the exact wording of the text, whereas those tested in the afternoon are more likely to produce a meaning-based representation. The results of a recognition test showed that, in agreement with other studies, immediate text memory is better in the morning than in the afternoon. As predicted, the subjects tested in the morning better remembered the exact wording of the original passages. However, the idea that memory for meaning is better in the afternoon than the morning was not supported by the data. The results suggest that subjects at both times of day produce a meaning-based representation of the text but, in the morning, this representation is supported by good verbatim memory, enabling accurate recall of the text at that time.     

Depression and processing information about others

We investigated differences between mildly depressed subjects and normal controls in their memory for, and judgments of, another person. All subjects read a story (containing either predominately positive or predominately negative items) under instructions to form an impression of the story target. Subjects later made several judgments about the target and attempted to recall the story. For both the positive and the negative story, depressed subjects produced significantly more negative intrusions (i.e., recalled nonpresented negative items) than did the normal subjects, but they did not recall more negative items correctly than did the normal subjects. Depressed subjects also perceived themselves as more similar to the negative story target than did the normal controls and indicated greater liking for the negative story target than did the normal controls.     

Schema anticipation in recall: memory process or report strategy?

Summary This article reports an experimental investigation testing the hypothesis that recall reports are systematically biased by the recaller's knowledge about who is to receive the recall report, while that bias is absent from a recognition test. Subjects listened to one of two versions of a story recounting John's visit to the doctor. After the story was read, subjects were asked to recall the story either with standard-recall instructions or with instructions to recall for peers, for Martians, or for a contest. In a fifth condition, subjects performed a filler activity, viz., the recall of a personal experience. After the recall test subjects were administered a recognition test. Between the groups no systematic differences in memory performance were observed in the recognition test. Recall for peers appeared to be poorer than recall in the other conditions. Subjects in the recall for a contest condition scored higher than the subjects in the other groups. Not only did groups differ with respect to bias, but there were also systematic differences in the memory or discrimination scores. The findings are discussed within the framework of schema theory.     

Test anxiety and breadth of encoding experiences in free recall

Subjects made three ratings for each of 30 words, then had an unannounced immediate free recall test and a second test after 1 week. There was no evidence to indicate significantly better retention on either test when the dimensions rated were unrelated as opposed to related. Subjects low on test anxiety performed better overall than subjects who scored high on either the emotionality or worry components of test anxiety. On the delayed test, the test anxiety effect was apparent only when study ratings involved unrelated dimensions (i.e., broad encoding). This interaction is consistent with two conclusions. First, rating related dimensions led to narrow encodings, thus hindering low-anxiety subjects who normally encode broadly. Second, even explicit ratings of unrelated dimensions did not induce high-anxiety subjects to encode more broadly. That the high-anxiety deficit was not overcome even with explicit orienting tasks may indicate inflexibility in deploying memory strategies.     

Witnessing postevents does not change memory traces,but can affect their retrieval

The results of three experiments suggest that a memory trace for an event is not altered by witnessing similar events, but that postevents can interfere with its retrieval. On an immediate recall test, details from an original story (e.g. wrench) were recalled less often if a subsequent story mentioned a ‘screwdriver’ than if it did not. The interference effect occurred if people were asked to recall details fromboth stories (tool —— ——), but not if people were asked to recall primarily from the first story. Thus, the interference effect in immediate recall was averted if the target trace could be activated selectively (Experiments 1a and 1b). A more general interference effect was found after a day. Fewer targets from the original story were recalled if the second story was presented just before the test than if both stories occurred a day earlier. Thus, the second story interfered with recall only if it emphasized contextual retrieval cues that did not match the trace for the targets (Experiment 2). Copyright © 2001 John Wiley & Sons, Ltd.     

Aging and qualitative characteristics of memories for perceived and imagined complex events

Older and younger adults' memory for perceived and imagined events was examined with a procedure in which everyday situations are simulated in the laboratory. Subjects perceived some situations and imagined others. Later, they were asked to rate their memory for various aspects of these situations (e.g., amount of perceptual detail, thoughts and feelings). A recall test followed the ratings. On the rating scale, for both perceived and imagined events, older subjects reported better memory for their thoughts and feelings than did younger subjects. In addition, on the recall test, older subjects produced more thoughts and feelings than did younger subjects, whereas younger subjects produced more perceptual and spatial information. These results suggest that older subjects may not inhibit personal information (e.g., thoughts and feelings), and this information may interfere with memory for other aspects of information, such as perceptual and contextual details (Hasher & Zacks, 1988).     

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.     

Developmental differences in schema dependency for temporally ordered story events

This research project was undertaken to investigate whether temporally ordered story events would be recalled in logical sequence as opposed to presentation order by various ages and under various task conditions. A 24-hour delayed condition was used as well as immediate recall. Six-year-olds, 8-year-olds, and adults were asked to recall four narratives. Instructions given were either vague or specifically required subjects to recall events exactly as they had been presented. Following the delayed recall, a picture-sequencing task was adminstered to assess whether picture cues would enable subjects to demonstrate awareness of input order even though they had reordered events in recall. All subjects reordered more during the delayed recall than during the immediate recall. Age differences (p<0003) occurred in the ability of subjects to demonstrate verbatim memory on the picture-sequencing task. Findings suggest that in contrast to adults, once children have reordered narrative events in memory, they no longer have an alternative verbatim version available. Results also suggest a greater schema dependency in children than in adults in recall tasks.     

Basis of metamemory judgments for text with multiple‐choice,essay and recall tests

Accuracy of metamemory for text was compared for multiple‐choice, essay and recall tests. Essay and recall tests were scored with Latent Semantic Analysis (LSA), number of correct idea units and number of word matches. Each measure was correlated with college students' predictions and posttest confidence judgments across texts to determine metamemory accuracy. Metamemory accuracy varied for different types of tests with multiple‐choice tests generally producing greater accuracy than essay tests. However, metamemory accuracy for essay and recall tests depended on the measure used to score them. Number of correct idea units produced the highest metamemory accuracy, word matches produced an intermediate level, and LSA produced the lowest accuracy. Students used the quantity of output in their judgments, so performance measures that related most strongly to quantity matched judgments better than measures based on answer quality. The results are compatible with an accessibility account of judgments about performance on text. Copyright © 2008 John Wiley & Sons, Ltd.     

Predicting and postdicting the effects of word frequency on memory

In the experiments reported here, I replicate and extend recent results that reveal that judgments about the memorability of common and uncommon words differ qualitatively depending on whether they are made during study or elicited during a recognition test (Guttentag & Carroll, 1998). When assessing recognition ability for individual words, subjects predict superior performance for common words, but postdict better performance for uncommon words. This interaction suggests that subjects rely on different cues when making judgments during study than they do when making analogous judgments during the recognition test, and that the cues utilized during recognition lead judgments to be more accurate. The shift is then evident in later predictions Subjects who make postdictions consequently correctly predict superior recognition performance for uncommon words on a subsequent study list. When subjects are asked to make later predictions about recall performance, however, having made postdictions on a test of recognition does not mislead subjects into predicting superior recall performance for uncommon words.     

The role of elaborative and schema processes in story memory

Three experiments examined recall for a story for which comprehension of some of its idea units required active construction on the part of the subject (letters were deleted from the words contained in one-third of the idea units). In Experiment 1, recall was significantly better for those ideas with letters deleted than for those with letters intact. In Experiments 2 and 3, subjects were instructed to adopt a particular perspective while reading the story. Recall of idea units was found to be an additive function of (1) an idea's importance to the perspective adopted, and (2) the letter-deletion manipulation. These results suggest a model of story memory that incorporates both elaborative-processing and schema-based mechanisms.     

The impact of connectives on the memory for expository texts

Three experiments investigated the influence of connectives on memory for expository text. Subjects in Experiments 1 and 2 read and later recalled passages with either no connectives, temporal connectives (before/ and then), causal connectives (which caused/which enabled), or intentional connectives (in order that/so that). There was connective interference rather than connective facilitation in the studies reported here: the recall for passages without connectives was higher than the recall for passages with connectives. The results partially supported a semantic complexity hypothesis, which predicted that recall should increase with the semantic complexity of the connective (i.e. temporal < causal < intentional). In a third experiment, the semantic appropriateness of the connective was manipulated. Recall for passages without connectives was better than recall for passages with either appropriate or inappropriate connectives. In a fourth experiment, the subjects filled in missing connectives in natural expository texts. Subjects were able to select the connective category which was originally used by the author. The influence of connectives on text was discussed from the perspective of resource limitations, elaborations, and the semantic compatibility of the connectives to the text.     

Scented memories of literature

This study examined the effects on recall of story details of congruity or incongruity between the hedonic valence of literary texts and odours inhaled while reading them. During the reading session, 24 undergraduates (12 males and 12 females) read two passages involving positive subject matter and two with negative subject matter while sniffing pleasant or unpleasant odours in a within‐subject fully counterbalanced design. Subjects rated their experience of each text on eleven 7‐point scales. During the test session 48 hours later, subjects read a two‐word title associated with each of the passages and inhaled the odour that was paired with it in the reading session. They also rated their experience on six of the scales that had been used during the reading session. Results showed that hedonic congruence between the passage and the odour fostered enhanced recall during the test session. The combination of positive subject matter and positive odour was reflected in more accurate recall of character details, while pairing negative subject matter and negative odour resulted in more accurate recall of setting details. Regression analysis showed that overall recall accuracy was increased by identifying with the characters in the stories and for passages that were found pleasing and personally meaningful. Consistent with the literature on implicit learning involving odours, recall accuracy varied inversely with perceived odour intensity. Implicit learning involving odours and literary passages is therefore fostered by unity in the reading experience.