Von Restorff Research Articles

Facial recognition and the von Restorff effect

The accuracy of recognition for faces of varying distinctiveness was investigated in this study. In Phase I, the subjects ranked 12 male faces from most to least distinctive. Interjudge agreement proved to be reliable. An analysis of the mean ranks assigned to the 12 faces showed them to be significantly different. In Phase II, other subjects responded to the 12 faces in a test of recognition. One week after inspection, distinctive faces produced significantly fewer errors of omission and commission than did nondistinctive faces. The findings suggest a von Restorff effect for faces.

The effects on induced passive rehearsal and the von Restorff phenomenon on the free recall of normals and retardates

Noninstitutionalized retardates and normal children of equal mental age (approximately 8 years) listened to 11 orders of 20 nouns with no instructions to overtly or covertly rehearse. All subjects were in each of three list conditions: (a) when all words were at the same decibel (dB) level, (b) when the 11th item was at a higher dB level, and (c) when the items were repeated at the same dB level using the rehearsal protocol of Rundus (1971). The normals' overall recall was superior to that of the retardates' recall. Repetition of items had no effect on the performance of either group except in the recency portion of the learning curve and on the initial trial for both groups and the fifth trial for the retarded group. The normal group was more susceptible to the yon Restorff effect than was the retarded group. These results were discussed in terms of the changing and different recall strategies of the two groups.

The Effects of Set Size on Learning an Item in the Set

To test the prediction from generalization theory that the ease of learning an item will be inversely related to the size of the set of which it is a member, 80 subjects were exposed to a serially presented list of pairs for 15 trials. Of the pairs, 1, 3, 7, or all 14 of them were printed in red, and the remaining pairs in black. The prediction was not fully supported. If the reduction of generalization is a process involved in the von Restorff effect, the results of this experiment suggest that it is not the only one.

The von Restorff effect and induced amnesia: Production by manipulation of sound intensity.

When an outstanding item appears in an otherwise homogeneous list of items, the outstanding item is better remembered (the von Restorff effect), and items before and after it may be more poorly remembered (induced amnesia) than corresponding items in a control list. In the present experiments the outstanding item was a word presented as a loud shout among other words presented at normal conversational levels. In two experiments, large retrograde- and anterograde-induced amnesiae effects were demonstrated using a free-recall and a recognition task. In both experiments half of the subjects were told what to expect and were instructed to devise a strategy to eliminate induced amnesia. These instructions failed to eliminate the amnesiac effect. A third experiment was designed to demonstrate an empirical similarity between induced and clinical amnesia. In clinical retrograde (but not anterograde) amnesia, "lost" memories are sometimes recovered with time. Filled delays of 0, 30, or 120 sec interpolated between list presentation and recall demonstrated that induced retrograde amnesia disappeared at the longest delay but induced anterograde amnesia was unchanged. A fourth experiment eliminated some alternate interpretations of the effect.

The von Restorff effect and induced amnesia: Production by manipulation of sound intensity.

The von Restorff effect and induced amnesia: Production by manipulation of sound intensity.

The Isolation Effect in Free Recall and Recognition

Since 1933, when von Restorff first reported systematic research on the isolation effect, a continuous flow of publications has been devoted to this phenomenon (for a review, see Wallace, 1965). A central and yet unresolved problem in the interpretation of the effect is its locus in the sequence learning/retention/retrieval. Originally, von Restorff suggested processes like 'Bereichsbildung' on the trace level. Later, processes taking place during initial learning were emphasized; for example, Green (1956) thought the effect to be due to the pronounced 'surprise value' of the isolated item. More recently, attempts have been made to assess the importance of retrieval processes in the better retention of the isolated item. One way of doing this is to compare the strength of the isolation effect in recall versus recognition, as retrieval processes are generally more important in the former than in the latter. If the effect is largely due to input processes going on during initial learning, then the effect should be obtained in tests of both recall and recognition. If, however, the effect is mostly determined by retrieval processes, then one would expect the effect to be smaller or absent in tests of recognition. Surprisingly, only very few studies on the isolation effect have used tests of recognition. Most relevant to our purpose is a study by McLaughlin (1968). Using a serial anticipation task, he found a 'vivid' (isolated) item to be better retained than control items in tests of free recall if learning consisted of two or three trials; however, no such effect was found in tests of recognition that substituted for or followed the tests of recall. McLaughlin used CVCs, presented on slides, as learning material; vividness was produced by changing size (4 in. instead of 1.5 in.) and shape (uppercase bold instead of lowercase 'pica') of the letters. In

The von Restorff effect and measures of organization in normals and retardates

Thirty-two noninstitutionalized retardates and 32 normals of equal mental age (approximately 8 years) heard 10 orders of 10 nouns when the decibel (dB) level of the stimuli was the same and when the dB level of the sixth item was higher. The yon Restorff effect was demonstrated by both groups. The normals' overall recall was superior to that of the retardates' recall. Differences in recall were attributed to the normals' tendency to recall items in the same serial order more consistently. The subjective organization of lists, as measured by order of recall, differed for the normals and retardates; the normal group was more adaptable than the retarded group in adjusting their strategy to different list conditions.

PSEUDO-PRIMACY EFFECTS IN FREE RECALL

Summary.-Shiffrin (1970) has suggested that the primacy effect in the serial position curve is the result of the temporal distinctiveness of the initial items in the list. Given this assumption a pseudo-primacy effect should be observed in other parts of the list if distinctive cues are provided. The present experiments tested this prediction. The distinctive cues presented were a change from a male to a female reader (or vice versa) and a short pause in list presentation. Pseudo-primacy effects were produced and the shape of the serial position curve was changed to a significant degree. A number of recent models of memory have emphasized the role of the context in which information is given for memorization as an important factor in the memorization process (e.g., Norman & Rumelhart, 1970; Shiffrin, 1970). These models propose that one does not store the literal meaning of the stimu111s or its nature since this information is already stored before the experiment begins. Rather, memorization proceeds by the establishing of an association between pre-existing representarions of the stimuli (including their characteristics or attributes) and the experimental context in which they are given. According to Norman and Rumelhart (1970), recall then proceeds by retrieving all those attributes which have the proper contextual association and reconstructing the stimulus or item. Shiffrin (1970) conceives of all the icerns possessing [he contextual association as defining a search set. The probability of retrieving an item from the search set depends, in part, on its size. The larger the search set the lower the probability of recall. Any search restricted to a cercain part of the list should result in a higher probability of recall for the items in that part. In this way, Shiffrin (1970) can account for the primacy effect in the serial position curve without making a rehearsal assumption. The beginning of the list provides a natural and distinctive cue upon which to define a smaller search set, one for beginning items. Since at some point in recall the search process is restricted to this smaller part of the list, the probability of recall from this portion of the list is greater. It is essentially the von Restorff effect at work at the beginning of the list. It should be noted that this explanation of the primacy effect differs considerably from the more commonly cited explanation proposed by Glanzer and Cunitz (1966). They propose that recall is higher for the initial list items because they are retrieved from long-term storage. 'The present research was adapted from a doctoral dissertation done under the direction of Dr. John A. Robinson, and submitted to the Department of Psychology, University of

Category and Position Competition in Free Recall.

The present two experiments were designed to investigate the effects of two competing retrieval strategies-category and position cueing. One hundred college students participated in Experiment I in which Presentation Rate, List Type, and Stimulus Position were manipulated within an FR paradigm. Sixty students participated in Experiment II in which Formal Response Similarity and Stimulus Position were manipulated within a PA paradigm. Findings were the following: (a) that greater clustering and FR is obtained at slower presentation rates; (b) that greater than chance clustering occurs with a category list structure; and (c) that HS response learning is more difficult than LS response learning. No support was found for position cueing. Results were discussed within FR cueing, von Restorff, and intentional versus incidental learning paradigms as possible explanations for the nonsignificant position cueing results in the present experiments. A competing strategy hypothesis of category and position cues was preferred with category recall and clustering as the dominant mode of retrieval.

Intentional and incidental learning in children and the von Restorff effect.

Intentional and incidental learning in children and the von Restorff effect.

Over-all list facilitation and the von Restorff effect with nonnoxious isolation stimuli.

To demonstrate list facilitation with nonnoxious stimuli, this study used 5 unique nonnoxious stimuli as isolates randomly distributed in a paired-associate list. Lists of 20 identical low-association word-pairs were presented in 3 groups: (1) Control—all 20 pairs in standard format, (2) Control—all 20 in changed format, and (3) Experimental—5 pairs in changed format randomly distributed among 20. 18 male college students (6 per group) received 4 standard PA acquisition trials. Statistically significant findings were: (1) experimental exceeded controls in total list performance, (2) format-changed items (isolates) exceeded control counterparts, (3) standard stimuli in experimental lists exceeded control counterparts, and (4) items immediately preceding and following isolates exceeded others (spread-of-isolation effect). The results were interpreted in terms of OR theory.

The von Restorff effect as a function of method of isolation

Using lists of eight letters, the von Restorff effect was investigated under four methods of isolating the fifth item: red background, border (surround), enlargement, or class change (number). Responses were scored considering serial position (SR scoring) and ignoring serial position (FR scoring). The isolated item was remembered best when it was a number under both SR and FR scoring. All isolation methods showed a positive spread of effect to the sixth item under SR but not under FR scoring.

Amnesic effects in short-term memory.

Produced the von Restorff effect in a free recall experiment with 68 undergraduates by the interpolation of a photograph of nude human beings at Serial Position 8 in a 15-item list consisting of line drawings of familiar objects. This effect was accompanied by a retrograde amnesia at the 2 serial po

Influence of fading procedures on the von Restorff effect.

Influence of fading procedures on the von Restorff effect.

Relevant and Irrelevant Saliency in Concept Learning

Two concept-learning experiments are reported. The first compared performance under three stimulus configurations: with the two relevant dimensions highly salient; with two irrelevant dimensions highly salient; and with no highly salient dimensions. Performance was best when the relevant dimensions were highly salient, next best when two irrelevant dimensions were highly salient, and poorest when there were no highly salient dimensions. The second experiment explored several possible explanations of these results, which are contrary to existing notions of the effects of irrelevant saliency. Its results suggest that the presence of highly salient irrelevant dimensions improved performance because the subjects examined them in the early trials of the experiment and then could permanently eliminate them from further consideration. It is now generally accepted that variations in the saliency, vividness, or obviousness, of the stimulus materials have an important influence on human learning.1 Several different methods have been used in the attempt to manipulate saliency. One common method, called perceptual isolation, is to make some aspect of the stimulus material stand out by making it perceptually or conceptually unique. The effects of isolation are commonly identified with the name of von Restorff, who introduced isolation in a seriallearning task by embedding a single number in a list of nine syllables.2 The number was learned faster and retained better than the same number embedded in a more heterogeneous list containing other numbers, syllables, and nonsense forms. Received for publication March 9, 1970. This research was supported in part by Research Grants MH-11283 and MH-18249 from the National Institute of Mental Health, United States Public Health Service. 1 T. Trabasso and G. H. Bower, Attention in Learning, 1968; W. P. Wallace, Review of the historical, empirical, and theoretical status of the von Restorff phenomenon, Psychol. Bull., 63, 1965, 410-424. 2 H. von Restorff, Ueber die Virkung von Bereichsbildung im Spurenfeld, Psychol. Forsch., 18, 1933, 299-342.

Degrees of isolation and the Von Restorff effect in serial learning

Sixty college students learned by the study-recall method a serial list of 11 CVCs with the central item isolated in different degrees of perceptual enhancement: shape (Eb), shape and color (Ea), or shape, color, and size (E1); the middle three items isolated (E3); the middle five items isolated (E5); or none isolated (C) to a criterion of two consecutive errorless trials. Compared to the corresponding control items, significant facilitation of isolated item (I) occurred in all experimental conditions: The magnitude of facilitation of the I (sixth item) increased as (1) the degree of perceptual enhancement increased and (2) the number of surrounding Is decreased. The item immediately following I(s) was also significantly facilitated. The findings were accounted for in terms of Gibson’s concepts of generalization and differentiation in verbal learning.

ATTENTION AND THE VON RESTORFF EFFECT

The hypothesis that the von Restorff effect is caused by the isolated (I) item attracting attention, which results in increased recall, was tested on the basis of the assumption that the attention activities are time‐consuming. The variables tested were: presentation time, presentation mode, isolated or non‐isolated sequence, position of the I item, material of the I item and delay between presentation and recall. The results only partly supported the above attention hypothesis, neither could interference theory, based upon acoustic similarity, give a complete account of the findings.

Structure and short-term memory

Fifty-six undergraduates participated in a probe experiment investigating the effects of list structure and rate of presentation on recall. Very strong serial-position effects were found, with structured lists exhibiting higher recall and a von Restorff effect. Rate of presentation did not result in significant differences.

Effect of perceptual organization on the susceptibility of learned material to retroactive inhibition.

The present study was designed to determine the effect of perceptual organization on the susceptibility of learned material to retroactive inhibition. Studies of isolation effect reported by von Restorff (1933) and Siegel (1943) suggest that an isolated item, standing out as figure on a ground, resists interference and is retained better as compared to the massed or similar items which suffer the fate of assimilation and lose their identity. This is taken as a proof of Gestalt hypothesis which predicts that perceptual organizing forces set up at the time of original learning persist during the retention interval and enhance retention. If we accept the behaviourist's contention that isolation aids discrimination, the results of the studies of isolation effect can also be explained in terms of reduced intraserial interference. Thus there is an uncertainty with regard to the effect of perceptual organization on the susceptibility of learned material to retroactive inhibition. This is, perhaps, due to the fact that the design of the studies of isolation effect could not permit sufficient control over the factor of similarity.

The Von Restorff Effect as a Function of Difference of the Isolated Item

The Von Restorff Effect as a Function of Difference of the Isolated Item