PART-WHOLE TRANSFER EFFECTS EN FREE RECALL LEARNING Maia for“ the Degree sf M A. {W ‘ Li“ mi 5. ATE uNi‘JER‘Sl W m" :=3:\ 0‘... [3% 12%;: LIBRARY ‘1 TH 5:3 ' ' u Mlchlga v State University Part-Whole Transfer Effects in Free Recall Learning by Dorothy Clark Abstract of a Master's Thesis Completed Fall Term, 1967 Tulving (1966) found in free recall learning that a group with no prior practice on a list of familiar words recalled more items on the last few trials of practice than a group which had previously practiced half of the list. The present study is a replication and extension of Tu1ving's experiment. Subjects learned two lists of 15 words each followed by a 30-word list which contained all, some or none of the items previously learned. The previously learned parts were presented either separately or mixed on the final list. For the separate groups, the final list consisted of all of the items from one of the first two lists followed by all of the items from the other list. For subjects who had practiced only some of the items, either the first or last half of the final list contained unfamiliar items. For the mixed groups, the final list consisted of an alternation of items from lists 1 and 2 or an alternation of items from one of the first two lists with new items. One hundred twenty-six students from the introductory psychology course at Michigan State University parti- cipated in the experiment. Lists 1 and 2 were presented in booklets, with a new page for each of the five trials, and the final list was presented by a slide projector for 10 trials. After each presenta- Dorothy Clark tion, subjects wrote the items they could recall in answer booklets, with a new page for each trial. The groups did not differ in total number of words recalled on the final list, but the conditions by trials interaction was signifi- cant. The reason for the interaction was not clear. The effect of final list presentation order was isolated, but the difference between mixed and separate order was not significant. To the extent that the ‘manipulation of order influenced organization, this finding cast doubt on Tulving's explanation that the groups in his experiment differed in final list learning because of organizational factors. The effects of prior practice and final list order were analyzed separately on trial 1 of the final list, and both effects were found to be significant. Prior practice on all of the final list items resulted in better recall than did prior practice on only some of the items. The significantly poorer recall on trial 1 of the final list by the mixed groups was interpreted according to Tulving's explanation of subjective organization. The order of recall on the last trial of list 3 demonstrated the persistence of previously learned organization. The mixed groups recalled items together on the last trial which they had learned together on list 1 or 2, rather than associating items from both parts which were presented consecutively on the final list. This finding indicated that the mixed groups were unwilling or unable to abandon Dorothy Clark their previously learned organization. However, failure to reorganize did not appear to have a significant effect on the number of words recalled on the last trial. Approved: ji;9~1£vw\ [Jjoo152\_ Committee Chairman Date: "714:3 Q /# /? é 7 Thesis Committee: Gordon Wood, Chairman Abram.M. Barch David C. Raskin PART-WHOLE TRANSFER EFFECTS IN FREE RECALL LEARNING By Dorothy Clark A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1967 ACKNOWLEDGMENT The author gratefully acknowledges the help of Dr. Gordon Wood, without whose direction, numerous readings of the manuscript, patient suggestions and encouragement this thesis would not have been completed. The advice of Dr. Abram.Barch and Dr. David Raskin is greatly appreciated. TABLE OF CONTENTS ACKNOWLEDGMENT . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . iv LIST OF FIGURES . . . . . . . . . . . v LIST OF APPENDICES . . . . . . . . . . vi INTRODUCTION . . . . . . . . . . . 1 METHOD . . . . . . . . . . . . . 4 RESULTS AND DISCUSSION . . . . . . . . .} 7 REFERENCES . . . . . . . . . . . . 21 APPENDIX A . . . . . . . . . . . . 21a APPENDIX B . . . . . ‘ . . . . . . . 28 APPENDIX C . . . . . . . . . . . . 29 iii LIST OF TABLES Table Page 1. Analysis of Variance: List 3 9 2. Analysis of Variance: List 3, Trial 1 l4 3. SCR Scores: List 3, Trial 10 18 iv LIST OF FIGURES Figure Page 1. Mean number of items recalled by each group across trials on list 3 8 2. Mean number of items recalled by the mixed and separate conditions across trials on list 3. ll Appendix A. LIST OF APPENDICES Number of Words Correct Per Trial for Each Subject on List 3 Total Correct Per Subject Across Trials on List 3 SCR Scores (0 - E) for Each Subject, List 3 Trial 10 Computing stimulus category repetition (SCR) Page 21 28 29 31 Introduction Tulving (1966) found in free recall learning that a group with no relevant prior practice on a list of familiar words recalled more items on the last few trials of practice than a group which had pre- viously practiced half of the list. The experiment was performed twice. In the first experiment the final list of 36 words was pre- sented for eight trials. In the second experiment the final list of 18 words was presented for twelve trials. In each case prior to learning the final list the part group practiced half of the list for eight (Experiment 1) or twelve (Experiment 2) trials, while the control group practiced irrelevant items for the same number of trials. On the final list in the first experiment, the control group passed the part group in number of words recalled on the 4th trial, and in the second experiment the crossover came on trial 7. The slopes of the curves were significantly different in both experiments. The mean difference between groups on the last trial was approximately 1.5 to 2 items. Tulving concluded that the organization of the part which was established during prior practice interfered with organi- zation of the whole and therefore hindered the performance of the part group on later trials of final list learning. He suggested that insofar as gs were unwilling or unable to reorganize the familiar part on the final list, their performance would remain below that of the control group. Organization of items during part learning is an instance of subjective organization (Tulving, 1962), which is a measure of the difference between order of input and output in free recall learning. If no two words in a list are ever presented together on more than one trial, any repetitions in the order of recall across trials must have been contributed by the §s since there were no repetitions in successive presentations. Subjective organization is measured by comparing the number of repetitions of pairs (or longer sequences) which would be expected by chance with the actual number of repetitions in the order of recall over trials. In Tulving's earlier studies of subjective organization (1962, 1964), the only constraint on final list order was that no two words should be presented consecutively on more than one trial. In a later study (1965) Tulving compared the effect of different presentation orders (constant over trials) on free recall. He found that §s reached criterion significantly faster with a high organization order than with a low organization order of presentation. The high organization order was established by presenting words in succession which‘gs had frequently recalled together on the last several trials in a previous free recall experiment (1962). The low organization order presented words in succession which had rarely been recalled in succession in the previous experiment. From the results Tulving concluded that when material is presented in an order that keeps subjectively related sub— sets of words intact, it can be learned more readily than.when the order is antagonistic to the inherent structure of the total set. The present study combines the variables investigated separately by Tulving, presentation order (1965) and prior practice on part of a list (1966). Subjects practiced none, half or all (in two parts) of the items before learning the final list. Since Tulving suggested that inappropriate organization of half of the list had hindered whole list learning, it was predicted that inappropriate organization of both halves should be a greater hindrance. In addition, on the final list items from a previously learned part were either presented consecutively or were alternated with previously practiced or new items requiring different levels of reorganization for groups with equal amounts of relevant prior practice. It was predicted that consecutive presentation would require little or no reorganization while alternate presentation would require more reorganization. Furthermore, if reorganization (or lack of it) is the cause of poorer recall, the groups with the alternate order of presentation on list 3 should recall fewer items than the groups with consecutive order. Finally, since Tulving claimed that poorer recall by the part group on the last few trials was caused by failure to reorganize the previously learned part, there should be evidence that previously learned organi- zation is maintained throughout list 3 learning; that is, items learned together during part practice should have a greater than chance tendency to be recalled together on the final list. Method Subjects. Subjects were 61 men and 65 women from the intro- ductory psychology course at Michigan State University. Eighteen §s were assigned randomly to each of seven groups. Design. All §s practiced the same final list of 30 words for ten trials, using a free recall procedure. Prior to learning the final list (list 3), §s practiced two lists of 15 items each, with five trials on each list. The groups differed in the amount of relevant prior practice and in the arrangement of the previously practiced part(s) on the final list. In lists 1 and 2, Se practiced all (A), part (P) or none (C) of the items from list 3; and on list 3, items from lists 1 and 2 were either separate (S) or mixed GM). For Groups AM and AS, lists 1 and 2 were both included in the final list. For Groups PM and PS the final list contained all of the l 1’ words from list 1 and none from list 2. For Groups PM and PS 2 2’ list 3 contained all of the words from list 2 and none from list 1. Group C learned irrelevant items on both lists 1 and 2. The order of items on the final list for the mixed groups (AM, PM and PMZ) 1 consisted of an alternation of items from lists 1 and 2 or an alter- nation of items from the familiar part with new items. The order of the final list for the separate groups (AS, PS and P32) consisted 1 of all of the items from one list followed by all of the items from the other list. Either the first or last 15 items on list 3 were new to the part—separate groups. Materials. The final list of 30 words was composed of two 15- word lists which had been evaluated by Deese (1959) for inter-item associative strength (lists 9 and 15). Both lists had zero inter— item associative strength; that is, no item in the list was given as a response to any other item in the list in a free association test. Two other lists (lists 3 and 12) from the same study, with zero or low inter-item associative strength, were practiced by the part and control groups in lists 1 and/or 2 but did not occur in list 3. All four lists were matched for frequency in the Thorndike-Lorge count (Thorndike and Lorge, 1944) and ranged from AA to l per million. The presentation order of list 3 was the same for all conditions, but the order relative to lists 1 and 2 was different for the mixed and separate conditions. For convenience, the effect will be referred to as final list order. However, it should be remembered that the order effect was achieved by manipulating the content of lists 1 and 2, not the presentation order of list 3. The relationship of the final list to lists 1 and 2 was established in the following manner. List 9 (Deese, 1959) was designated as the odd list and list 15 as the even list. These lists constituted the parts practiced by the mixed groups on lists 1 and 2. Items were selected alternately fran these two lists to make up the first presentation order of list 3. The first and second halves of this combined list constituted the parts to be learned by the separate groups on lists 1 and 2. The material presented in lists 1 and 2 was counterbalanced within each condition. TWo more orders of the final list were prepared, with the same items always in even (or odd) positions and in the first half (or second half) of the list. These three orders of list 3 were rotated during final list learning. There were also three random orders of each part for lists 1 and 2. Procedure. Subjects were run in groups. They were given stan- dard free recall instructions and were told at the beginning of the session that they would be learning three lists. Because gs from all conditions were run at the same time, they were not informed about the relationship of lists 1 and 2 to the final list. Lists 1 and 2 were presented in booklets, with a new page for each of the five trials. Subjects had one second per item to read the list and 45 seconds for written recall. An example was given of the rate at which the list should be read in order to finish in 15 seconds. At the end of each presentation, §s closed their study booklets and wrote the items they could recall in their answer booklets. A new page in the answer booklet was used for each trial. The final list was presented by a slide projector for 10 trials at the rate of two seconds per item, with 90 seconds for written recall after each presentation. Results and Discussion Performance on lists 1 and 2 was analyzed to determine whether the groups differed before the experimental treatments. Each.§fs score consisted of the sum of his scores on trial 5 of both lists. Only one.§ recalled all items on the last trial of both lists, while 15 Se recalled all items on one of the two lists. Since the combined group means, which ranged from 22.78 to 24.17, were well below the maximum possible mean of 30, it appears that none of the groups had reached asymptote. The means of the combined scores were not signif- icantly different, F < 1. Performance on the final list is shown in Figure l, and results of the trend analysis are presented in Table 1. Although Group AS had the highest score on trial 1 and kept the lead for 10 trials, Group AM with the second highest score on trial 1 was later surpassed by several groups, including Group C which had had no prior practice. There was no difference as a function of conditions, but the trials by conditions interaction was significant, indicating that the groups differed over trials. From Figure 1 it appears that the significant differences among groups over trials could be attributed primarily to differences on the first three trials. To examine this possibility a trend analysis was performed on trials 4 to 10. The trials by con- ditions interaction was not significant, E(36, 714) = 1.11, p.> .25. An analysis of the lepes of the curves, following Tulving's (1966) procedure, revealed a significant difference, F(6, 714) = 2.54, o In .1 J! < o w (I I! Lu :1: 2 3 Z Z S 5 l4 I2 ll IO 9 Fig. 1. / o---o As “ ?‘ A---A PS" 'A' / 0—. AM A—A PM“ I. an-” c - / -/ X -1 I I 1 L l I l l l I 2 3 4 5 e 7 s 9 IO TRIALS Mean number of items recalled by each group across trials on list 3. * Groups PS1 and PS were combined for graphic presentation, as were Groups EM_1%d PMO. Table 1 Analysis of Variance: List 3 Source Conditions (C) Ss/C Trials (T) Linear (L) Dev. from L T x C L x C Dev. from L x C T x §s/C * p_< .01 ** p_< .001 df 6 119 (9) 1 8 (54) 6 48 1071 MS 112.93 114.03 (1,527.33) 11,916.97 228.62 (13.19) 68.96 6.22 4.13 .99 (2,296.74**) 2,889.66** 55.44** (3.20*) 16.72** 1.26 ——-— 10 p_< .025, even though the differences in number of items correctly recalled were small. The reason for the difference in slopes, particularly over trials 4 to 10, was difficult to identify. One potential source of difference, the order effect, was isolated by combining Groups AS, PS1 and PS2 into the separate condition, and Groups AM, PM and PM 1 into the mixed condition. Since Group C fit neither condition, it 2 was omitted from this analysis. The learning curves for the combined mixed and separate conditions are shown in Figure 2. Although the performance of the combined separate group was consistently superior to the combined mixed group, the F ratios for conditions, trials by conditions interaction, and slopes were not significant. To the extent that the manipulation of order influenced organization, the results do not support Tulving's claim that organizational factors accounted for the differences among groups over trials. The group requiring more reorganization (mixed group) did not recall signifi- cantly fewer items and did not learn at a significantly lower rate than the group requiring less reorganization (separate group). As evidence of reorganization Tulving (1966) cited the differ- ence in slopes. The reason why this measure was used can be seen by examining the design of his experiment. Organization was estab- lished by prior relevant practice. Consequently, on the final list the facilitating effect of prior practice was combined with the hindering effect of reorganization, and the advantage of prior prac- tice may have masked any effect due to reorganization on earlier trials. The comparison of slopes was intended to establish the 26L 24- 22- l6 - I . 0---0 Separate H Mixed MEAN NUMBER RECALLED l4- l3" ALIIIL 678910 I l l I 2 3 4 T LS >0!" RI Fig. 2. Mean number of items recalled by the mixed and separate conditions across trials on list 3. ll 12 effect of reorganization. However, one obvious reason for the difference in slopes in Tulving's experiment is the better performance of the experimental group on the first several trials because of prior relevant practice. It is conceivable, therefore, that the difference in slopes could be explained without reference to reorganization. If the mean recall of the control group had been significantly higher than the experimental group (relevant prior practice) at the end of final list practice, the evidence for organizational effects would have been more convincing. Tulving did not report an analysis of difference on the last trial, but in the present experiment the differences on the last trial yielded an E_of less than 1. Another source of difference among groups, amount of prior relevant practice, could not be examined in the same way as the order effect because of the unequal numbers of Se receiving different amounts of practice. Comparisons were made between groups which were least likely to yield significant results and also between groups most likely to yield significant results. Scores were collapsed across trials. Groups AS and AM were compared with Groups PS1 and PS2 (the parts groups with the highest total scores across trials), and the difference was not significant, 2(1, 70) - 1.71, p_> .10. However, the difference between Groups AS and AM.and Groups PM1 and PM2 (the parts groups with the lowest total scores across trials) was signifi- cant, F(1, 70) = 6.10, p_< .05. In comparing the effect of prior 13 practice on part of the list with no prior practice, Groups PS1 and C were not significantly different, F(1, 34) I 2.44, p > .05, nor were Groups PM and C, §’< 1. In general, it appears that amount of 1 prior practice accounted for some of the difference among groups across trials. However, in the present experiment more prior practice did not produce a consistent advantage. ‘ «mtum‘kr 'T ‘ To summarize the analysis thus far, it is apparent that the 1‘3""."3‘." groups differed over trials and that the slopes of the curves differed over trials 4 to 10. However, the differences could not be attributed to the effect of mixed versus separate order, and it is unlikely that amount of prior practice had a lasting effect. Since the design of the present experiment separated the effects of prior practice and order, it was not necessary to confine the analysis to an evaluation of the slopes of the curves. A simple analysis of variance of per- formance on trial 1 of the final list confirmed the observation that there were significant differences among groups on the first trial, 2(6, 119) = 8.62, 2.< .01. To evaluate simultaneously the effects of amount of prior practice and order on trial 1 in a factorial design, Group C was omitted. Both main effects were significant (see Table 2). Prior practice on both parts (X.= 16.1) resulted in better performance than practicing one part on list 1 or 2 (X'- 12.9, 13.3). Since the interaction of prior practice and order was not significant, §_< 1, it appears that inappropriate organization of both halves was not 13a mmre of a hindrance to recall than inapprOpriate organization of half of a list. The order effect can be interpreted by reference to Tulving's (1962) explanation of subjective organization. In the course of free recall learning, the organization of recalled items becomes increasingly stable as associations are formed among items. .Accordingly, in the present study practice on lists 1 and 2 should turve produced separate organization of one or both parts of the final mam-so.- ~—; ufirgrrz—‘E’ “" Table 2 Analysis of Variance: List 3, Trial 1 Source df MS Practice (P) 2 112.58 Order (0) l 65.33 P x O 2 9.53 Error 102 12.48 * p_< .05 ** p_< .01 14 F 5.24* .76 ———— pm. nutmeg-rut 1’?" 15 list. If it is assmned that items presented consecutively have a tendency to become associated, then mixing the parts together in presenting the final list may have disrupted previous associations by giving the opportunity for new associations to be formed. The resulting response competition may have led to the observed decrement in recall on trial 1. On the other hand, presenting the parts intact should have provided less opportunity for conflicting associations to be formed, and therefore less reorganization should have been required in learning the final list. Because Groups PM2 and C duplicated the conditions in Tulving's (1966) study, they were compared separately. The curves were similar to those obtained by Tulving. The trials by conditions interaction was significant, 3(9, 306) = 4.52, p_ < .01, and so was the difference in slopes, §(l, 306) = 28.45, E < .01. The analysis was repeated for trials 4 to 10 to determine whether the difference between groups resulted primarily from the difference on trials 1 to 3. The condi- tions by trials interaction was no longer significant, F < l. The Slepes were significantly different, £11, 204) = 3.94, p_ < .05, with Group C surpassing Group PM2 on trials 9 and 10. However, since the order effect did not yield significant results in the previous analysis, the reason for the difference in slopes and the crossover of curves is not clear. Tulving argued that one reason why the group with prior practice fell behind the control group on the last few trials is that they 16 were unwilling or unable to abandon or modify the organization formed during part learning. Implicit in this argument is the assumption that there is one optimal organization of items for any given list and that any other organization will result in poorer recall. In the present experiment, the order of recall on the last trial demonstrated the persistence of previously learned organization. For the mixed ZQAEJ"“_ “YYIr-u' ' t. groups, list 3 consisted of an alternation of items from lists 1 and 3' 2 or an alternation of items from list 1 or 2 with new items. For the separate groups, list 3 consisted of all 15 items from one of the previous lists followed by all 15 items from the other list, with new items on one half of the list for the parts groups. Thus, there were two constraints on the otherwise random order of the final list. A given item.was always in an even (or odd) position across all 10 trials, and it always appeared in the first (or second) half of the list. It can be seen that practice on such a list should have the effect of increasing organization into clusters of first or second half items. Further, clusters of odd or even items should be less likely to occur, since odd (or even) items were never presented consecutively on list 3. To compare organization of recall on trial 10 with previously learned organization from lists 1 and 2, Bousfield and Bousfield's (1966) measure of stimulus category repetition (SCR) was employed. Each previously learned part was viewed as a category. For example, the categories for Group AM were odd and even, while the categories for Group AS were first half and second half. Groups PM1 and PM2 l7 learned odd or even items plus new items, while Groups PS1 and PS2 learned first or second half items plus new items. The difference between the obtained and expected (chance) value Of SCR was computed for each-S (see Table 3). TWO scores of SCR were Computed for each §_in Groups AM, PM TH 1, PM2 and C, one based on odd and even categories and the other based on first and second half categories. 1 Comparison of the two sets of scores for Group C permitted an evaluation of the effect of presentation order apart from the effect new of prior practice. Clusters of odd or even items occurred less fre- quently than chance (-43.0), while clusters of first or second half items occurred more frequently than chance (85.9). In order to dis- tinguish the effect of prior practice from the effect of presentation order, the mixed groups were compared with the Odd and even clustering of Group C, and the separate groups were compared with the first and second half clustering of Group C. The mixed groups and Group C dif- fered significantly in the number of odd and even category repeti- tions, 3(3, 68) = 8.60, p3< .01. This suggests that the prior organi- zation of the mixed groups was maintained throughout final list learn— ing. However, the separate groups and Group C did not differ signi- ficantly in first and second half category repetitions, F(3,68) = 1.02, p_> .25, indicating that when prior organization and presenta- tion order did not conflict, their effect was similar in direction and magnitude. First and second half SCR's were computed for the mixed groups Table 3 SCR Scores*: List 3, Trial 10 Groups Categories AM PM1 PM2 C PS1 PS2 AS Odd/even 17.5 -l4.2 24.00 -43.00 SD 2.47 2.34 2.26 2.84 First half/ second half 51.50 51.70 59.90 85.90 113.40 109.10 90.80 SD 2.83 2.44 1.92 2.45 3.03 3.81 3.18 * Sums of individual SCR scores 18 'Y~§9=€Em 2 1V ' " anus-1 19 to determine to what extent they developed the organization which was favored by the presentation order of list 3. The effects of amount of prior practice and order on repetitions of first and second half items were analyzed. The mixed groups, which had previously learned odd and even categories, had significantly fewer repetitions of first or second half items than the separate groups, which had previously learned those categories, F(1, 102) = 24.36, p_< .01. The effect of amount of prior practice was not significant, F(2, 102) < 1. In summary, the mixed groups retained the organization they had learned previously and failed to learn the new organization which was favored by the pre- sentation order of list 3. This result supports Tulving's contention that the mixed groups were unwilling or unable to abandon their pre- viously learned organization, but failure to reorganize did not appear to have a significant effect on the number of words recalled on the last trial. In conclusion, the present study found that although reorgani- zation did not influence recall across trials, recall did differ as a function of reorganization on the first trial of final list learn- ing. This finding supports Tulving's contention that organization of items takes place during free recall learning and that inappropriate organization of part of a list retards subsequent whole list learning. The use of slopes as evidence for reorganization in Tulving's (1966) study was questioned, since the slopes could have differed because of prior practice, without regard to reorganization. Although reorgani- 20 zation did not significantly hinder recall across trials, the mixed groups tended to recall items on trial 10 in the order learned on lists 1 and 2 rather than in the order of presentation of list 3. This finding suggests that subjective organization is a persistent and important effect in free recall learning. REFERENCES Bousfield, A. K., & Bousfield, W. A. Measurement of clustering and of sequential constancies in repeated free recall. Psychological Reports, 1966,.19, 935-942. Deese, J. Influence of inter-item associative strength upon immediate free recall. Psychological Reports, 1959, 5, 305-312. Thorndike, E. L., & Lorge, I. The teacher's word book of_30,000 words. New York: Bureau of Publications, Teachers College, Columbia University, 1944. Tulving, E. Subjective organization in free recall of "unrelated" words. Psychological Review, 1962, 62, 344—354. Tulving, E. Intratrial and intertrial retention: notes towards a theory of free recall verbal learning. Psychological Review, 1964, 11, 219-237. Tulving, E. The effect of order of presentation on learning of "unrelated" words. Psychonomic Science, 1965, 3, 337-338. Tulving, E. Subjective organization and effects of repetition in multi-trial free-recall learning. Journal of-Verbal Learning and Verbal Behavior, 1966, 5, 193-197. 21 Number of words correct per trial for each subject on list 3 'H 11 13 l7 15 16 17 16 18 24 ll 14 15 15 13 l9 14 13 IN 14 l6 l7 19 20 18 21 23 22 15 17 17 20 20 23 16 21 M) 16 18 20 21 19 21 24 22 21 18 20 24 22 24 19 24 17 17 19 20 21 24 23 25 23 16 23 21 24 15 21 24 23 24 APPENDIX A GROUP NH Trial 18 17 19 19 23 22 21 26 23 19 24 21 26 17 21 28 22 19 21 2o 13 23 22 24 22 23 26 27 23 22 25 26 18 23 24 27 24 23 15 19 21 22 24 24 26 26 22 20 21 26 19 19 26 27 25 Kb 20 21 23 22 26 26 24 26 23 21 21 25 29 17 22 24 29 22 21 18 22 24 22 24 26 25 27 16 24 20 28 20 22 27 28 21 20 23 24 26 28 3o 24 27 18 22 26 26 19 22 27 27 23 Ir-s 10 13 17 11 15 15 11 12 l9 17 IN 14 17 20 17 24 19 21 16 ll 14 17 13 15 22 13 21 15 18 23 16 28 19 23 18 F5 20 23 17 26 18 10 17 22 18 29 23 24 22 17 14 14 24 13 22 23 21 26 18 20 25 22 25 25 21 22 19 16 14 17 19 21 27 21 30 Trial 22 GROUP H11 20 13 20 23 21 27 26 24 21 20 18 16 21 18 26 28 20 3o 18 21 25 21 29 25 22 24 26 20 14 21 20 26 27 21 26 Kb 19 21 25 24 29 28 25 25 24 l9 16 24 22 28 27 24 30 19 11 25 28 24 3o 27 26 24 24 19 15 26 23 24 29 19 3o 10 25 27 26 29 3o 26 25 28 19 14 25 25 25 29 25 28 tr.- - |1—s 13 13 11 15 15 10 13 21 14 11 14 13 14 17 11 15 IN 14 13 15 19 20 17 17 22 13 17 16 18 21 23 21 17 14 19 15 18 19 25 23 19 25 13 18 13 20 15 25 22 17 19 17 21 17 18 21 26 23 20 24 13 18 13 19 17 28 25 22 21 19 GROUP P112 Trial 21 17 17 22 24 23 19 26 17 21 11 23 19 25 24 25 22 26 23 25 17 25 23 26 25 21 28 17 19 15 21 2o 27 28 24 24 20 27 19 22 22 24 25 26 28 15 20 16 25 17 27 27 26 23 23 K» 28 20 23 23 28 27 24 28 17 19 18 22 20 27 27 26 20 27 29 21 23 23 29 28 21 3o 11 20 16 23 24 28 28 26 20 |1-~ 14 15 16 17 21 14 22 15 18 12 19 15 18 16 17 21 IN 17 25 16 26 17 20 18 23 13 17 10 19 20 23 15 17 26 19 24 19 20 28 21 26 20 25 19 18 12 18 22 26 14 27 29 20 25 19 22 25 23 20 18 25 20 20 15 18 22 24 18 21 27 GROUP AS Trial 21 28 20 25 24 24 24 23 21 18 21 22 26 20 24 26 24 19 30 24 27 24 24 26 29 22 21 19 21 23 26 20 27 29 23 28 20 21 29 25 26 25 25 20 20 15 20 23 23 2o 21 28 loo 23 3o 21 27 3o 28 24 3o 23 27 30 3o 28 26 29 24 22 18 24 24 27 23 28 28 Ip—s 15 15 13 10 15 10 13 17 11 ll 14 19 15 l7 l6 18 IN 21 18 18 16 21 18 19 13 25 16 15 19 15 22 21 18 20 l8 19 18 19 17 21 17 24 17 24 19 l6 l8 17 19 21 19 21 22 20 15 20 21 25 20 23 24 GROUP P81 Trial 24 18 19 19 21 22 27 16 24 19 16 23 22 25 19 23 24 27 25 21 20 20 26 24 27 21 26 20 17 24 26 26 24 24 28 25 22 21 22 27 23 22 16 27 20 21 23 26 28 21 19 24 25 I00 26 19 20 22 25 22 26 20 28 23 17 24 28 25 21 26 27 28 to 25 24 21 27 28 21 28 19 28 24 20 25 27 26 25 26 27 28 he 11 14 11 16 10 NO 15 22 19 22 13 23 28 18 22 15 15 16 19 15 15 18 20 14 22 21 27 15 24 3o 15 21 21 16 17 16 24 19 14 24 21 20 25 25 24 19 28 3o 15 25 15 13 15 22 19 15 25 18 GROUP P52 Trial 16 23 22 28 21 27 3o 15 27 16 16 17 25 l9 16 26 23 26 22 24 20 27 18 27 29 18 28 22 19 13 17 22 23 16 28 22 22 25 27 28 20 27 3o 21 29 18 19 14 17 22 18 19 N 0\ Kb 23 25 22 28 21 27 3o 19 28 20 21 17 22 26 21 19 29 26 20 26 24 28 24 28 19 3o 22 20 15 24 23 24 22 29 27 20 17 24 25 25 24 25 26 " It". “aft-5‘ - IH 10 11 81:18:66 1 H 15 IN 16 13 14 21 15 20 15 16 19 21 17 23 24 15 25 20 16 14 19 16 15 19 21 17 20 25 20 2o 13 21 20 2o 21 21 19 18 26 GROUP C Trial 19 21 17 21 28 18 27 24 22 25 19 18 22 2o 22 18 28 27 15 22 17 25 28 20 29 23 19 19 22 23 17 26 24 27 23 28 17 22 22 21 28 22 28 27 16 21 27 24 21 27 19 22 24 30 I00 16 22 24 25 27 24 23 21 17 28 23 23 26 22 23 24 26 18 24 21 28 3o 23 28 24 21 24 28 24 25 26 3o 26 28 22 23 22 27 27 22 28 26 28 APPENDIX B Total correct per subject across trials on list 3 Group .411 P’1 P12 0 P31 P32 AS 177 171 172 168 228 181 204 166 90 224 201 199 231 276 200 194 195 182 190 214 198 206 231 211 225 195 255 225 221 201 245 246 232 181 271 224 267 225 179 202 257 233 229 233 206 255 234 292 239 243 227 262 210 178 171 223 244 212 138 182 254 253 259 182 189 186 157 196 186 202 205 162 139 221 170 164 201 211 135 206 191 213 143 161 244 208 184 180 221 176 208 162 174 250 219 243 224 221 211 219 253 206 210 195 244 241 254 227 207 215 168 192 230 189 210 208 234 249 233 216 264 201 251 247 226 271 28 APPENDIX C SCR scores (0 - E) for each subject, list 3, trial 10 Categories: "first half" and "second half" Group AM PM 1 H12 0 PS 1 PS 2 5.5 1.9 5.0 3.3 11.9 6.0 4.3 .2 3.5 4.3 2.5 9.0 -l.2 5.3 1.0 9.0 2.9 4.3 .9 1.4 -1.1 6.5 4.3 11.5 5.0 .7 5.3 3.9 11.3 2.6 1.1 7.5 3.7 6.3 5.5 13.5 1.0 4.0 4.9 6.0 8.0 4.0 3.0 2.3 6.0 6.0 2.9 2.5 .5 .3 4.1 3.5 11.5 -2.0 1.9 2.9 1.9 7.0 4.3 6.0 1.9 3.5 1.4 3.3 5.6 3.0 5.9 3.4 1.9 2.9 9.0 2.7 6.9 3.0 1.5 1.3 5.5 7.7 2.5 2.9 3.9 7.5 5.9 7.5 6.0 3.5 4.0 - .8 4.3 10.3 -3.5 1.5 2.5 6.0 3.5 5.0 3.5 8.5 4.5 7.0 8.0 9.5 6.3 -1.1 5.9 2.9 6.5 6.0 29 AS 3.0 0.0 8.5 .5 2.0 5.5 3.3 5.5 6.5 5.4 7.3 3.9 6.5 12.4 6.7 5.3 8.0 AH -2.5 -l.7 .9 5.9 0.0 4.0 4.0 Categories: even, odd -1.1 4.3 -2.0 -1.5 -2.1 -2.0 -1.7 -4.5 4.5 - ,7 ‘20]. Group 1112 -1.0 1.5 - ,7 2.9 2.9 2.9 0.0 0.0 2.1 2.9 .9 - .5 6.0 0.0 4.4 3-3 -3.1 30 -2.5 1.5 -3.0 -7.0 —4.7 -3.0 -l.O -5.7 -2.1 r"; "I'.—T APPENDIX D Computing stimulus category repetition (SCR) (Bousfield & Bousfield, 1966) (mi + . . . mi) _ 1 E(SCR) = n where mk equals the number of J items recalled frmm the kth category on trial j, and n j equals the total number of items recalled on trial j. O(SCR) is obtained by counting the number of times a word from any category is followed by another item from the same category. 31 I '“l! \I‘I'IIA id!“