Printer Friendly

A further attempt to demonstrate hypermnesia in recognition.

In 1974, Erdelyi and Becker presented subjects with a mixed list of pictures and words followed by three 7-minute forced recall tests. These investigators reported that even though subjects were not allowed to study the list between the tests, their recall of pictures (but not words) improved over the repeated tests. Erdelyi and Becker called this phenomenon hypermnesia.

Following Erdelyi and Becker's pioneering study, a number of studies have shown that hypermnesia is a robust phenomenon (see Payne, 1987, for an extensive review of the literature). It has been found with pictures as well as words under a variety of conditions (e.g., Belmore, 1981; Klein, Loftus, Kihlstrom, & Aseron, 1989; Otani & Hodge, 1991; Payne, 1986; Payne & Roediger, 1987; Roediger & Payne, 1985; Roediger, Payne, Gillespie, & Lean, 1982; Wheeler & Roediger, 1992). The only exception has been a recognition test which does not seem to produce a reliable hypermnesic effect, especially when a standard list learning paradigm is employed (Otani & Hodge, 1991; Payne & Roediger, 1987; but see Erdelyi & Stein, 1981).

In an effort to find hypermnesia in recognition, Otani and Hodge (1991, Experiment 1a) investigated the effect of relational and item-specific processing (Einstein & Hunt, 1980; Hunt & Einstein, 1981). Einstein and Hunt hypothesized that recall and recognition are based on two processes called relational and item-specific processing. Relational processing elaborates the shared characteristics of stimulus items and serves a generative function at the time of retrieval. In contrast, item-specific processing emphasizes the unique features of individual stimulus items. Einstein and Hunt reported that the optimal recall and recognition performance results from the additive effect of relational and item-specific processing.

To induce item-specific processing, Otani and Hodge asked subjects to form an image of each stimulus word. Relational processing was established by asking subjects to create a single image of a group of three consecutively presented words. A third group was an intentional learning control group. Otani and Hodge reported that three two-alternative forced choice recognition tests failed to produce hypermnesia. Furthermore, in Experiment 1b, an interval of 1 week was introduced between the stimulus presentation and the first test; however, this did not alter the results.

Otani and Hodge (1991) concluded that hypermnesia does not occur in recognition. However, a problem with their experiments was that they failed to find an effect for processing strategy. In other words, their manipulation of processing strategy was not successful. It is thus possible that with a more effective processing manipulation, hypermnesia could be observed.

There is a good reason why the effect of relational and item-specific processing should be examined further. Klein et al. (1989) and Burns (1993) found that relational and item-specific processing produce a different amount of net free recall performance. Otani and Hodge (1991) also discovered that cued recall hypermnesia is influenced by relational and item-specific processing. It is therefore possible that recognition hypermnesia is also affected by relational and item-specific processing.

To ensure that the manipulation of relational and item-specific processing would produce a significant effect, the present experiments employed the tasks originally developed by Einstein and Hunt (1980). Two types of recognition tests were investigated. Experiment 1 examined yes-no tests whereas Experiment 2 examined two-alternative forced choice tests.

Experiment 1

Method

Subjects. One-hundred and nineteen male and female introductory psychology students at Central Michigan University participated.

Materials. A related list was constructed by selecting 20 words from each of 10 taxonomic categories of the Battig and Montague (1969) norms. The mean normative frequency of these words was 44.35 per million (U - scale, Carroll, Davies, & Richman, 1971), and the mean imagery value was 5.74 (Toglia & Battig, 1978). An unrelated list was created by first generating 20 words belonging to 1 of 10 ill-defined categories (e.g., big, round, burn). These nouns had a mean frequency of 48.17 per million and a mean imagery value of 5.68. Ten words from each category were then randomly selected to form a stimulus list. The remaining 10 words were used as distractors for the recognition tests. Two types of lists were constructed because the Hunt and Einstein's (1981) data have indicated that processing strategy interacts with list characteristics. These words were then individually typed at the center of a 3- x 3-inch index card. The rate of presentation was determined by the slowest subjects in a session. For each type of list, one random sequence was used to present the list. Three recognition tests were constructed for each list type by randomly ordering the stimulus words and distractors three times. A 5-point pleasantness scale (1 - very unpleasant, 5 - very pleasant) and a 12- x 24-inch poster board with 10 names of categories were prepared for the item-specific and relational processing tasks. A page with randomly generated two-digit numbers was created for a filler task.

Procedure. Subjects were tested in groups of 2 to 12. In the item-specific processing condition, subjects were instructed to rate the pleasantness of each word using a 5-point scale. Subjects in the relational processing condition were asked to sort the words into categories and place the cards under the labels on the poster board. In these two conditions, subjects were not informed about the subsequent memory tests. However, in the intentional learning condition, subjects learned the list for an unspecified memory test.

Following the processing task, subjects performed a filler task for 2 minutes. Then, they received three 7-minute yes-no recognition tests in which they were instructed to circle the studied words.

Results and Discussion

Hit scores. Figure 1 shows hit scores as a function of list type, processing strategy, and test. As can be seen, processing strategy made a difference in overall recognition performance. However, performance did not improve over the three tests.

A 3 (processing strategy) x 2 (list type) x 3 (test) unweighted mean mixed-design analysis of variance (ANOVA) revealed that the effects of processing strategy, F(2, 113) = 47.97; list type, F(1, 113) = 23.62; and their interaction, F(2, 113) = 9.04, were significant. (The significance level was set at .05 throughout this paper.) Furthermore, the effects of test, F(2, 226) = 17.84, and the test x list type interaction, F(2, 226) = 3.37, were significant. Tukey tests demonstrated that for the related list, the relational processing and intentional learning conditions did not differ from each other. The item-specific processing produced higher performance than the relational processing or intentional learning groups. In contrast, for the unrelated list, no difference was found between the item-specific processing and relational processing conditions. Performance was significantly lower for the intentional learning condition. These tests also revealed that the interaction of test x list was based on the difference that occurred between the related and unrelated lists on the third test.
Table 1

Mean Number of Recovered and Forgotten Items in Experiments 1 and 2

 Experiment 1 Experiment 2
 Test Test
 2 3 2 3
Condition

 Related List
Item-specific Recovery 5.11 1.11 3.61 3.57
(n=19, 23) Forgetting 7.79 5.00 7.22 5.17

Relational Recovery 6.80 3.60 5.86 6.05
(n=20, 21) Forgetting 11.30 8.05 10.05 5.38

Intentional Recovery 8.80 4.00 6.73 4.59
(n=20, 22) Forgetting 10.95 9.70 7.05 4.91

 Unrelated List
Item-specific Recovery 2.75 1.10 3.08 2.96
(n=20, 24) Forgetting 6.70 3.40 6.08 2.79

Relational Recovery 4.75 1.40 2.87 1.91
(n=20, 23) Forgetting 6.00 2.85 4.22 2.48

Intentional Recovery 6.30 3.85 5.65 4.25
(n=20, 20) Forgetting 8.70 5.05 5.95 4.90

Note. The second number in the parentheses is n for Experiment 2.


The analysis of newly recovered items on Tests 2 and 3 revealed that the effects of processing strategy, F(2, 113) = 13.87; list type, F(1, 113) = 9.58; test, F(1, 113) = 113.69, and the test x list type interaction, F(1, 113) = 6.22, were significant.

Tukey tests showed that the intentional learning group recovered more items than the item-specific processing condition. No difference was found between the intentional learning and relational processing conditions or between the relational and item-specific processing groups. The analysis of the test x list type interaction revealed that subjects recovered more items on the second than on the third test. Further, the related list produced more recovery than the unrelated list.

The number of forgotten items on Tests 2 and 3 is also shown in Table 1. An ANOVA demonstrated that the effects of processing strategy, F(2, 113) = 6.83; list type, F(1, 113) = 27.77; the processing strategy x list type interaction, F(2, 113) = 3.14; and test, F(1, 113) = 43.06, were significant. Tukey tests indicated that the processing strategy x list type interaction was based on the difference between the related and unrelated lists when subjects engaged in relational processing. More forgetting occurred with the related than unrelated list.

False alarms. Table 2 shows false alarm rate as a function of list type, processing strategy, and test. An ANOVA revealed that the effects of strategy, F(2, 113) = 4.66; list type, F(1, 113) = 17.89; and test, F(2, 226) = 107.31, were significant. The analysis further indicated that the interaction effects of test x list type, F(4, 226) = 3.52, and test x strategy x list type, F(4, 226) = 3.89, were reliable. For both lists, false alarm rates increased as a function of the repeated tests.
Table 2

Mean False Alarm Rate as a Function of List Type, Processing Strategy, and
Test in Experiment 1

Test
 1 2 3

Related List
Item-specific 8.79 12.68 15.16
Relational 8.20 11.75 14.30
Intentional 14.70 21.80 25.40

Unrelated List
Item-specific 5.60 8.80 11.70
Relational 3.40 8.10 9.60
Intentional 7.10 8.20 11.20


Experiment 2

Method

Subjects were 133 male and female introductory psychology students at Central Michigan University. The materials were the same as those used in Experiment 1 except that two-alternative forced-choice recognition tests were constructed by randomly pairing the stimulus items with distractors that belonged to the same category. The procedure was the same as those in Experiment 1 except that the test was delayed 24 hours to avoid a ceiling effect.

Results and Discussion

As can be seen in Figure 2, overall performance was influenced by processing strategy and list type. However, as in Experiment 1, performance did not improve over the repeated tests.

A 3 (processing strategy) x 2 (list type) x 3 (test) unweighted mean ANOVA confirmed that the effects of processing strategy, F(2, 127) = 13.85; list type, F(1, 127) = 35.81; and their interaction, F(2, 127) = 6.48, were significant. Furthermore, there were significant effects of test, F(2, 254) = 44.27, and the test x processing strategy interaction, F(2, 254) = 7.05.

Tukey tests showed that the unrelated list produced higher performance than the related list for the relational processing group. In contrast, for the item-specific processing and intentional learning conditions, no difference was found between the related and unrelated lists. The tests further revealed that performance declined between Tests 1 and 2 for the item-specific and relational processing conditions. No such decline was observed for the intentional learning condition.

The analysis of newly recovered items on Tests 2 and 3 indicated that the effects of processing strategy, F(2, 127) = 4.48; list type, F(1, 127) = 8.68; and their interaction, F(2, 127) = 3.18, were significant. The effects of test, F(1, 127) = 7.37, and the test x processing strategy interaction, F(1, 127) = 3.51, were also significant. Tukey tests demonstrated that more recovery occurred on the second test than on the third test for the intentional learning condition. No difference between Tests 2 and 3 occurred for the item-specific and relational processing conditions. Furthermore, on Test 2 the intentional condition produced more recovery than the item-specific or relational processing condition. The latter two conditions did not differ from each other. An ANOVA on forgotten items revealed that the effects of list type, F(1, 127) = 13.28, and the list type x processing strategy interaction, F(1, 127) = 3.33, test, F(1, 127) = 17.92, and the test x processing strategy interaction, F(1, 127) = 4.75, were significant. The three-way interaction of test x list type x processing strategy was also reliable, F(1, 127) = 4.51. For both lists, forgetting was greater for the second than for the third test. A significant difference among processing conditions occurred only for the related list. General Discussion

As with previous studies (Otani & Hodge, 1991; Payne & Roediger, 1987), the present experiments failed to observe hypermnesia in recognition. However, the present experiments successfully produced processing effects. The present results thus provide stronger evidence against recognition hypermnesia than those of Otani and Hodge. Why does recognition consistently fail to produce hypermnesia?

Based on the assumption that hypermnesia is a net result of reminiscence and intertest forgetting, one possibility is that not enough items are recovered on the second and third tests. The inspection of cumulative recognition performance, however, reveals that the percentage of recovery between Tests 1 and 3 in some conditions was similar to the percentage of recovery observed with cued recall in Otani and Hodge's (1991) experiments. Thus, if forgetting can be controlled, performance may improve over repeated tests.

A second possible reason hypermnesia does not occur in recognition is that the rate of intertest forgetting is too high to produce an increase in net performance. In support of this speculation, the number of forgotten items was greater in the present experiments than that observed in Otani and Hodge's (1991) study using cued recall tests. Why is the rate of forgetting so high in recognition?

Mandler's (1980) dual-process theory provides a possible answer. He argues that recognition relies on two components, familiarity and retrieval. Familiarity is based on the perceptual integration of stimulus items, whereas the process of retrieval depends on the amount of contextual association. Mandler claims that subjects first recognize an item based on item familiarity. The process of retrieval is then used to identify the item.

Based on the notion that recognition requires a familiarity judgment, it is possible that recognition suffers from a high rate of intertest forgetting because familiarity of studied words declines as a function of retention interval (Mandler, 1980). If the decline follows a negatively accelerated function, a greater forgetting rate would be expected on the second than on the third test.
Table 3

Percentage of Change Between Tests 1 and 3 for Cumulative Recovery and
Forgetting Performance

 Recovery Forgetting
 Related List

Otani & Hodge (1991) Cued Recall
Item-specific 13.4 -8.1
Relational 44.5 -13.1
Intentional 13.0 -6.4

Experiment 1
Item-specific 7.8 -10.8
Relational 17.9 -23.9
Intentional 23.5 -28.8

Experiment 2
Item-specific 8.6 -23.4
Relational 15.9 -33.4
Intentional 16.4 -27.0

 Unrelated List
Otani & Hodge (1991) Cued Recall
Item-specific 7.6 -3.2
Relational 13.5 -4.4
Intentional 6.2 -4.6

Experiment 1
Item-specific 4.8 -7.6
Relational 7.4 -6.4
Intentional 18.7 -18.5

Experiment 2
Item-specific 7.3 -17.7
Relational 5.5 -12.4
Intentional 14.1 -23.1


The concept of familiarity suggests another difficulty subjects may face when the same test is repeated. If subjects base their judgments on item familiarity, one can assume that selecting studied items is relatively easy on the first test because the amount of familiarity associated with the studied words should be clearly greater than that of the nonstudied words (cf. Mandler, 1980). However, on the second test the discrimination between the studied and nonstudied words would become more difficult because item familiarity associated with distractor words would rise as a result of their exposure during the first test. This explanation implies that recognition of studied words would become more difficult as the same test is repeated. In support of this assumption, the false alarm rate in Experiment 1 increased as a function of repeated tests. Unfortunately, false alarm rate cannot be computed for the forced-choice test. A third possibility is that the null improvement is attributable to a ceiling effect. In other words, performance reaches its asymptote on the first test (see Roediger & Challis, 1989), preventing further improvement on the second and third tests. Unfortunately, this explanation is circular. Calling a null improvement a ceiling effect does not explain why recognition produces a ceiling effect to begin with.

To conclude, the present experiments supported Otani and Hodge's (1991) conclusion that recognition does not produce hypermnesia. The main reason appears to be that the rate of intertest forgetting is too great to produce an improvement over repeated tests even though a moderate amount of reminiscence occurs on the second and third tests.

References

BATTIG, W. F., & MONTAGUE, W. E. (1969). Category norms for verbal items in 56 categories: A replication and extension of the Connecticut category norms. Journal of Experimental Psychology Monograph, 80.

BELMORE, S. M. (1981). Imagery and semantic elaboration in hypermnesia for words. Journal of Experimental Psychology: Human Learning and Memory, 7, 191-203.

BURNS, D. J. (1993). Item gains and losses during hypermnesic recall: Implications for the item-specific and relational information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19, 163-173.

CARROLL, J. B., DAVIES, P., & RICHMAN, B. (1971). Word frequency book. New York: American Heritage.

EINSTEIN, G. O., & HUNT, R. R. (1980). Levels of processing and organization: Additive effects of individual-item and relational processing. Journal of Experimental Psychology: Human Learning and Memory, 6, 588-598.

ERDELYI, M. H., & BECKER, J. (1974). Hypermnesia for pictures: Incremental memory for pictures but not for words in multiple recall trials. Cognitive Psychology, 6, 159-171.

ERDELYI, M. H., & STEIN, J. B. (1981). Recognition hypermnesia: The growth of recognition memory (d') over time with repeated testing. Cognition, 9, 23-33.

HUNT, R. R., & EINSTEIN, G. O. (1981). Relational and item-specific information in memory. Journal of Verbal Learning and Verbal Behavior, 20, 497-514. KLEIN, S. B., LOFTUS, J., KIHLSTROM, J. F., & ASERON, R., (1989). Effects of item-specific and relational information on hypermnesic recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 1192-1197.

MANDLER, G. (1980). Recognizing: The judgment of previous occurrence. Psychological Review, 3, 252-271.

OTANI, H., & HODGE, M. H. (1991). Does hypermnesia occur in recognition and cued recall? American Journal of Psychology, 104, 101-116.

PAYNE, D. G. (1986). Hypermnesia for pictures and words: Testing the recall level hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 16-29.

PAYNE, D. G. (1987). Hypermnesia and reminiscence in recall: A historical and empirical review. Psychological Bulletin, 101, 5-27.

PAYNE, D. G., & ROEDIGER, H. L., III. (1987). Hypermnesia occurs in recall but not in recognition. American Journal of Psychology, 100, 145-165.

ROEDIGER, H. L., III, & CHALLIS, B. H. (1989). Hypermnesia: Improvements in recall with repeated testing. In C. Izawa (Ed.), Current issues in cognitive processes: The Tulane Flowerree Symposium on cognition (pp. 175-199). Hillsdale, NJ: Lawrence Erlbaum.

ROEDIGER, H. L., III, & PAYNE, D. G. (1985). Recall criterion does not affect recall level or hypermnesia: A puzzle for generate/recognize theories. Memory & Cognition, 13, 1-7.

ROEDIGER, H. L., III, PAYNE, D. G., GILLESPIE, G. L., & LEAN, D. S. (1982). Hypermnesia as determined by level of recall. Journal of Verbal Learning and Verbal Behavior, 21, 635-655.

TOGLIA, M. P., & BATTIG, W. F. (1978). Handbook of semantic word norms. Hillsdale, NJ: Erlbaum.

WHEELER, M., & ROEDIGER, H. L., III. (1992). Disparate effects of repeated testing: Reconciling Ballard's (1913) and Bartlett's (1932) results. Psychological Science, 3, 240-245.
COPYRIGHT 1994 The Psychological Record
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1994 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Otani, Hajime; Stimson, Mark J.
Publication:The Psychological Record
Date:Jan 1, 1994
Words:3294
Previous Article:Uncertainty and the perception of sufficiency of social support, control, and information.
Next Article:Private events and other causes of behavior: who can tell the difference?
Topics:


Related Articles
Word frequency effect: a test of a processing-based explanation.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters