Download Memory recall experiments in Psychology and more Summaries Human Memory in PDF only on Docsity!
Memory & Cognition 1977, Vol. 5 (3), 315-
Effects of organization and expectancy
on recall and recognition
JANE MARANTZ CONNOR
State University of New York, Binghamton, New York 19901
In two experiments, organization, test type (recall and recognition), and subjects' expectancies of the type of test they would receive (recall or recognition) were varied. It was found that organizational effects may be influenced by both subjects' expectancies of the type of test they will receive and the type of test actually received. Results indicated that subjects' encoding strategies are sensitive to the relationship between the type of material presented and the type of information that they expect to need for the test.
One of the predominant views of recall and recog- nition is that performance on the two types of tasks reflects two qualitatively distinct psychological processes (Anderson & Bower, 1972; Bahrick, 1970; Kintsch, 1970). It has been proposed that recalling an item involves a stage of memory search or retrieval which is not involved in recognizing an item. According to this view, in recall, subjects first generate a variety of possible responses on the basis of cues provided by either the subject or the experimenter. The subject examines each generated item in tum to determine if he recognizes it as a target item and whether or not to emit it. When a recognition test is employed, the generation or search process is eliminated. A major source of support for the retrieval distinction between recall and recognition comes from research indicating that increasing list organization facilitates recall substantially but has minimal effects on' recognition performance (Kintsch, 1970). A typical pattern of results is that reported by Bruce and Fagan (1970). Bruce and Fagan manipulated list organization by varying the number of categories represented in 42-word lists. High-structure lists contained 6 seven- word categories; low-structure lists contained 42 one- word categories. Distractors for the recognition test were drawn from the same categories in the same proportions as study items, Recognition performance was the same "for high- and low-structure lists, whereas recall performance was distinctly superior for high- structure lists relative to low-structure lists. The usual interpretation of these results is that organization facilitates the retrieval process, and retrieval is of
This research was supported by Grant lRQl MH25344..Ql from the National Institute of Mental Health. The comments of an anonymous reviewer were especially helpful. Requests for reprints should be sent to Jane M. Connor, Department of Psychology, State University of New York at Binghamton, New York 13901. Part of the results were presented at the meeting of the Midwestern Psychological Association in Chicago,
negligible importance in recognition. More specifically, in the process of writing the word "banana," a recall subject may think of the word "fruit," which would serve as a recall cue for other fruits which appeared on the study list. Since the recognition subject has all the target items (and distractors) before him, the cue does not facilitate his performance. More recently, Anderson and Bower (1973) have suggested the existence of two distinct encoding processes relevant to recall-recognition differences. They hypothesize that during the study period subjects may tag associative pathways between concepts corresponding to the words shown in the list. The tagging of associative pathways is directly relevant to the retrieval process in recall, since subjects must move from one word concept to another, but the tagging of associative pathways has little to do with recognition performance. Second, subjects tag the concepts corresponding to each presented word with a list tag indicating the occurrence of that word on a particular list. The tagging operation facilitates both recall and recognition. These operations explain the finding that intentional and incidental learning result in equivalent recognition - performance but differential recall performance (Eagle & Leiter, 1964). Presumably, both intentional and incidental subjects engage in list-tagging, but only intentional subjects tag the associative pathways among word concepts.. If the assumption of two encoding operations. is correct, it is of interest to evaluate the extent to which the beneficial effect of list organization on recall is independent of the type of encoding strategy used by the subject and the emphasis placed on the tagging of associative pathways. With respect to recognition performance, it is also of interest whether a strategy that places greater emphasis on list tagging may increase the performance of subjects on a recognition test. (The intentional-incidental learning experiments do not speak to this issue, since incidental learning subjects are engaged in other tasks.) The following experiments were designed to investigate these issues by manipulating
316 CONNOR
the type of test subjects were expecting to receive, which would presumably affect the extent to which they engaged in the two types of encoding operations. List organization and the type of test actually received were varied as well. The view of recognition and recall outlined above would predict that the facilitatory effect of list organization on recall either is enhanced for subjects with appropriate expectancies or is unaffected by expectancy. For subjects receiving recognition tests, the overall level of performance may be affected by expectancy, but list organization should have no effect regardless of expectancy. The latter state- ment is based upon the assumptions that the existence of associative pathways among word concepts and the tagging of these pathways are relevant to recognition performance.
EXPERIMENT 1
Method Subjects. The subjects were 112 undergraduates who participated in the experiment for course credit, in groups of 5 to 15. Materials. A pool of 256 words, 8 from each of 32 categories, was selected from the category norms of Battig and Montague (1969). Organized lists consisted of four words from each of eight categories. Eight lists of each type were constructed so that each of the 256 words appeared once on the organized lists and on the unorganized lists. Distractors for the recognition test were taken from the appropriate study list that the subject had not seen previously, so that distractors for the organized lists consisted of a different four words from the same eight categories, and, for the unorganized lists, a single word from the same 32 categories. Distractors for one group of subjects were study words for another group of subjects. The words in the study list and the recognition test were randomly arranged (not blocked by categories) on the page and lists were counterbalanced across subjects and trials. Procedure. Subjects were given 1 min to study the 32-word list, followed by 1 min for the arithmetic problems, followed by 2 min to take the test on the list they had just seen. Recall subjects were instructed to write down as many of the words from the preceding lists as they could. A l-rnin rest preceded the next study-test trial. Each subject received a booklet containing' the appropriate four sets of study list, arithmetic problems, and test. Subjects received either organized or unorganized lists throughout the session. They were informed that they would receive the same type of test throughout and that they would not be tested more than once on any list. However, on the fourth trial, half of the recall subjects were given recognition tests instead of the recall tests, and vice versa.
Results For all recognition scores, mean number correct, corrected for guessing by subtracting the number of false positives from the number of hits, as well as d' were computed. Separate analyses of variance were completed for the results of the first three trials and the results of the fourth trial. The analyses were also done separately for subjects receiving recall tests and those receiving recognition tests. On the first three trials, organization facilitated recall [F(1,54)= 15.14, MSe = 29.64, p< .001], but had no
effect on recognition performance, as measured by d' [F(1,54) = 1.83, MSe = .52, n.s.] and by number correct [F(1 ,54) < 1, MSe = 21.51]. No effect of trials was found in any of the analyses [F(I ,54) = 2.03,
MSe = 10.56, n.s., for recall; F(I,54) = 1.82, MSe == 9.41,
n.s., for recognition number correct; F(I ,54) = 1.57, MSe = .15, n.s., for d']. The effect of organization did not vary as a function of trial number; none of the interactions of trial with organization were significant. The results of the fourth trial are shown in Table 1. For recognition performance, the type of test expected and the type of list studied had no effect on number correct or d'. (MSe was 19.90 for number correct and .44 for d'. The F values for the tests of main effects and interactions ranged from less than 1 to 1.5, with 1 and 54 degrees of freedom). Looking at recall performance, subjects expecting recall scored higher than subjects expecting recognition [F(1 ,54) = 21.85, MSe = 28.03, p < .001] , and subjects studying categorized lists scored higher than subjects studying noncategorized lists [F(1,54) = 6.89, P < .01]. The relative superiority of categorized lists over noncategorized lists was the same for both types of expectancy fF(l ,54) < 1].
Discussion The results of this experiment are consistent with the model of recognition and recall outlined in the introduction. The overall performance of recall subjects was markedly affected by expectancy, indicating that
Table I Mean Number of Words Correct and d' on Fourth Trial as a Function of Organization, Test Received, and Test Expected Type of List Noncategorized Categorized Number Number Test Received Correct d' Correct d'
Experiment I Recognitions- b Expecting Recall 20.64 2.22 21.71 2. Expecting Recognition 22.57 2.12 21.64 2. RecaIIc Expecting Recall 13.93 18. Expecting Recognition 6.43 12. Experiment 2 Recognition'[-" Expecting Recall 24.33 2.36 28.06 3. Expecting Recognition 26.11 (^) 2.72 25.61 2. Recall" Expecting Recall 11.72 15. Expecting Recognition (^) 11.50 11. (a) least significant difference = 3.38 for number correct, p < .05. (b) least significant difference = .50 for d', p < .05. (c) least significant difference = 4.02, p < .05. (d) least signifi- cant difference = 2.71 for number correct, P < .05. (e) least significant difference =.43 for d', p < .05. (f) least significant difference = 2.51, p < .05.
318 CONNOR
on recall, regardless of expectancy. This suggests that the locus of the organization effect may have been different in the two experiments. Blocking may have facilitated recall because of the greater ease with which subjects can tag associative pathways while studying a blocked rather than a randomly arranged list. List structure (manipulated by number of categories) may have facilitated recall because of the greater number of associative pathways that exist among words on a categorized list as opposed to a noncategorized list. The implications of the recognition results in Experiment 2 for the model described in the introduc- tion are less clear. Blocking facilitated recognition only for subjects expecting recall, that is, for subjects who were encouraged to encode the relationships among the word clusters. For some reason, the tagging of associative pathways facilitated recognition when such tagging was easy to accomplish, in other words, in the blocked list. Possibly, such tagging, when combined with a blocked list, encourages the subjects to rehearse the target words in groups during the study period, which results in a strengthening of the list-occurrence information for the words within the group. The results of the two experiments together indicate that subjects have greater difficulty generating efficient encoding strategies for recognition tasks. One problem is that the type of information required to do well on a recognition test is determined by the set of distractors. However, even with a general knowledge of the type of distractors that will be encountered, as was the case in these experiments, such strategies appear to be difficult to devise.
CONCLUSIONS
As a whole, the results of these experiments suggest two important methodological considerations. First, studies comparing recognition and recall over a series of tests should contain adequate control for the different. encoding strategies subjects may adopt as a function of the type of test they are expecting. Second, the way in which list organization is varied (categorization, blocking, etc.) may affect the pattern of results obtained in interaction with other factors. It is particularly important that the degree of similarity between the set of target items and the set of distractors be similar for the two types of lists. Many of the studies
on the effect of list organization and recognition which have found negative results have confounded organiza- tion in the study set with organization in the distractor set (e.g., Bruce & Fagan, 1970; Kintsch, 1968). Other studies that have manipulated list organization by varying the spatial or temporal arrangement of study items have found effects on recognition (e.g., Bower, Clark, Lesgold, & Winzenz, 1969; Bower & Winzenz, 1969). List organization appears to be a multidimen- sional rather than a unidimensional factor.
REFERENCES
ANDERSON, J, R., & BOWER, G. H. Recognition and retrieval processes in free recall. Psychological Review, 1972, 79, 97-123. ANDERSON, 1. R., & BOWER, G. H. Human associative memory, Washington, D.C: V. H. Winston, 1973. ANISFELD, M., & KNAPP, M. Association, synonymity and directionality in recognition. Journal of Experimental Psychology, 1968, 77, 171-179. BARRICK, H. P. Two-phase model for prompted recall. Psychological Review, 1970, 77, 215-222. BATTIG, W. F., & MONTAGUE, W. E. Category norms for verbal items in 56 categories. Journal of Experimental Psychology Monographs, 1969, 80. BOWER, G. H., CLARK, M., LESGOLD, A. M., & WINZENZ, D. Hierarchial retrieval schemes in recall of categorized word lists. Journal of Verbal Learning and Verbal Behavior, 1969, 8, 323-343. BOWER, G. H., & WINZENZ, D. Group structure, coding and memory for digit series. Journal of Experimental Psychology Monographs, 1969, 80(No. 2, Part 2),1-17. BRUCE, D., & FAGAN, R. I. More on the recognition and free recall of organized lists. Journal of Experimental Psychology, 1970, 85, 153-154. EAGLE, M., & LEITER, E. Recall and recognition in intentional and incidental learning. Journal of Experimental Psychology, 1%4, 68, 58-63. KINTSCH, W. Recognition and free recall of organized lists. Journal of Experimental Psychology, 1968, 78,481-487. KINTSCH, W. Models for free recall and recognition. In D. A. Norman (Ed.), Models of human memory. New York: Academic Press, 1970. MANDLER, G. Organization and recognition. In E. Tulving & W. Donaldson (Eds.), Organization of memory. New York: Academic Press, 1972. MANDLER, G., PEARLSTONE, Z., & KOOPMANS, H. S. Effects of organization and semantic similarity on recall and recognition. Journal of Verbal Learning and Verbal Behavior, 1969, 8,410-423.
(Received for publication October 26, 1976; revision accepted January 13, 1977.)
