Paula J. Waddill

A contextual account of the generation effect: A three-factor theory☆

Abstract Recent studies have shown that the generation effect (items generated by an individual are better remembered than items that are read) can be diminished or eliminated under between-subjects deployment of the encoding task (generating vs. reading). Three experiments that further delineate the nature of the generation effect in between-subjects designs are reported. Experiments 1 and 2, using a structured word list, found significant generation effects in free recall under certain cue-target conditions and written-response conditions. Generation effects in recognition and cued recall occurred in all conditions. In Experiment 3, with intentional learning instructions, significant generation effects were not obtained. Current accounts of recent failures to find significant generation effects are discussed in light of our results, and all are found lacking to some degree. We propose an extension to the idea that generation enhances both cue-target relational processing and individual-item processing; namely, in some contexts generation will enhance whole-list processing.

The Effects of Text-based Interest on Attention and Recall

Attentional demands and recall for stories that differed in rated interest were examined. More interesting stories required fewer attentional resources for comprehension than did less interesting stories (Experiment 1). Overall recall did not differ across story interest, but story interest did interact with type of encoding in terms of recall levels (Experiment 2). Relational encoding improved recall for low-interest stories but not high-interest stories; the reverse pattern was obtained with a manipulation encouraging extensive processing of the individual propositions. We suggest that interesting stories free up resources for relatively optional organizational processing of the text elements, thereby rendering additional relational processing redundant (for recall). Less interesting stories require more resources to keep attention focused on encoding the individual propositions, thereby rendering additional proposition-specific processing redundant.

Distinctiveness effects in recall

Unusual information is generally recalled better than common information (the distinctiveness effect). Differential processing accounts propose that the effect occurs because unusual material elicits encoding processes that are different from those elicited by common material, and strong versions of these accounts predict distinctiveness effects in between-list as well as within-list designs. Experiment 1 employed a between-list design and manipulated presentation rate. Contrary to differential processing predictions, no distinctiveness effect emerged, nor did recall patterns for atypical versus common sentences differ as a function of presentation rate. Experiment 2 further tested differential processing accounts as well as representation accounts via a within-list manipulation and conditions that included experimenter-provided elaborations. Distinctiveness effects emerged in all conditions and, contrary to differential processing predictions, the pattern of recall in the elaborated conditions did not differ from that in the unelaborated conditions. Taken together, the results of this study lend more support to a representation view that suggests mechanisms related to the representation and subsequent retrievability of elements in the memory record play a major role in the distinctiveness effect.

Pictorial enhancement of text memory: Limitations imposed by picture type and comprehension skill

We examined the kinds of information in a prose passage that is better remembered when depictive illustrations are embedded in the passage than when the passage contains no illustrations. Experiment 1 showed that (1) pictures depicting details effectively increased recall of those details and (2) pictures depicting relationships effectively increased recall of that relational information (relative to a no-picture control condition). In Experiment 2, comprehension skill was found to modulate the general effects obtained in Experiment 1. Detail pictures enhanced the recall of targeted details for all skill levels. Relational pictures enhanced recall of pictured relational information for highly skilled and moderately skilled comprehenders, but not for less skilled comprehenders. Because there were no recall differences across the different skill levels in the no-picture control condition, it is suggested that pictures may serve to enable processing in which readers would not necessarily engage under ordinary circumstances. Pictures, however, did not appear to compensate for limitations reflected in lower scores on a standardized test of reading comprehension.

How Much Does Guessing Influence Recall? Comment on Erdelyi, Finks, and Feigin-Pfau

Erdelyi, Finks, and Feigin-Pfau (1989) present evidence that variations in recall criteria can affect the number of items correctly recalled. In this comment, we (a) describe some procedural differences between their work and the earlier experiments of Roediger and Payne (1985), (b) note that their large manipulations of recall criteria produced only small effects on the amount recalled, and (c) describe recent research complementing that of Erdelyi et al. We observe that variations in recall criteria have larger effects after a 1-week delay than on an immediate test.

Generation effects in free recall: further support for a three-factor theory.

A framework is presented that helps explain and predict generation effects in free recall (for between-subjects manipulations of generating vs. reading). When the targets share common features and when that shared information is salient to subjects, subjects will exploit that information to help generate the target items. This produces more relational processing among the targets (relative to reading), enhancing free recall. Consistent with this idea, when shared information (among targets) was salient, generation effects in free recall were found under encoding conditions that can disrupt generation effects in cued recall (e.g., pairing targets with unrelated cues). Further, within the same list, generation effects emerged in free recall for targets that were processed after shared information became evident but not for targets processed prior to the availability of the shared information. In recognition, generation effects were found for targets regardless of when they were processed.

Generation effects for context words: Implications for item-specific and multifactor theories☆

Abstract We investigated generation effects for context words (i.e., cue words) semantically associated with generated or read targets. Intact context words were presented with target fragments for one group of subjects (generating group) and with intact targets for another group of subjects (reading group). When recall for the context words was cued with the associated target, generation effects for context words were found (Experiments 1, 2, and 3). There was no generation effect for context words, however, when recognition memory was tested (Experiment 2). In Experiment 3, generation effects were obtained when cued recall for the context words was conditionalized on correct recognition of the targets (i.e., the cues in the cued-recall test). This pattern strongly suggests that generation can enhance context word-target word relational processing. Generation effects were also obtained in free recall of the targets (Experiment 1). This effect was accompanied by category clustering that was significantly greater than chance for generating subjects but not for reading subjects. This pattern suggests that generation can also enhance whole-list relational processing. The results are consistent with a three-factor account of the generation effect.

Revisiting the picture-superiority effect in symbolic comparisons: do pictures provide privileged access?

In 4 experiments, symbolic comparisons were investigated to test semantic-memory retrieval accounts espousing processing advantages for picture over word stimuli. In Experiment 1, participants judged pairs of animal names or pictures by responding to questions probing concrete or abstract attributes (texture or size, ferocity or intelligence). Per pair, attributes were salient or nonsalient concerning their prerated relevance to animals being compared. Distance (near or far) between attribute magnitudes was also varied. Pictures did not significantly speed responding relative to words across all other variables. Advantages were found forfar attribute magnitudes (i.e., the distance effect) and salient attributes. The distance effect was much less for salient than nonsalient concrete-attribute comparisons. These results were consistently found in additional experiments with increased statistical power to detect modality effects. Our findings argue against dual-coding and some common-code accounts of conceptual attribute processing, urging reexamination of the assumption that pictures confer privileged access to long-term knowledge.

What makes folk tales unique: Content familiarity, causal structure, scripts, or superstructures?

Requiring readers to re-order randomly ordered sentences into a coherent text significantly enhances recall relative to that in a read-only control condition for non-folk-tale texts but not for folk tales (Einstein, McDaniel, Owen, & Coté, 1990). Experiments 1-3 showed that embedding components of folk tales (e.g., causal structure, conventional scripts, content related to background knowledge) in non-folk-tale texts did not render sentence unscrambling ineffective for increasing recall. In Experiments 4a-4c, a folk tale was presented either as a fairy tale or as part of a newspaper article. Significant sentence unscrambling effects (in free recall) were not obtained in either presentation format, which implies that a story superstructure (a story grammar) does not contribute to the absence of the sentence unscrambling effect. It is suggested that understanding why the sentence unscrambling effect is absent for folk tales may require considering the functional role that narrative plays in socioculturally situated cognition.

Use of Retroactive Pessimism as a Method of Coping with Identity Threat: The Impact of Group Identification

Retroactive pessimism involves retrospectively lowering ones evaluations of a groups chances for success after a failed competition. Although past research has substantiated the existence of this strategy, investigators have yet to examine how level of group identification might impact the use of retroactive pessimism. Given that coping with group threat is most prominent among persons with high levels of group identification, we hypothesized that displays of retroactive pessimism would be magnified in persons with strong allegiances to a group. This hypothesis was tested by having supporters of two college basketball teams evaluate the chances for victory for each team both prior to the game between the teams and subsequent to the contest. Regression analyses confirmed expectations (the greatest magnitude of retroactive pessimism was reported by highly identified supporters of the losing team in their evaluations of the winning team). Subsequent analyses revealed that this effect was not mediated by level of disappointment in the outcome.