} Į រ 1 ĥ Į THE CATHOLIC UNIVERSITY OF AMERICA THE DEVELOPMENT OF LOGICAL AND ROTE MEMORY 1 } A DISSERTATION { · SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS AND SCIENCES OF THE CATHOLIC UNIVERSITY OF AMERICA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY ¿ THE CATHOLIC SISTER REGIS HOLLAND, M.A. BY IC UNIV I PRESS WUS MAR LUX EST NIVE 1940 AMERICA RSITY OF I } THE CATHOLIC UNIVERSITY OF AMERICA PRESS 1 WASHINGTON, D. C. } 1 THE DEVELOPMENT OF LOGICAL AND ROTE MEMORY THE CATHOLIC UNIVERSITY OF AMERICA THE DEVELOPMENT OF LOGICAL AND ROTE MEMORY A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS AND SCIENCES OF THE CATHOLIC UNIVERSITY OF AMERICA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY SISTER REGIS HOLLAND, M.A. THE CATHOLIC BY VOLIC UNIV PRESS Di WUS|MER} LUX A85 AMERICA 1940 10ALISE THE CATHOLIC UNIVERSITY OF AMERICA PRESS WASHINGTON, D. C. BE 371 H74 COPYRIGHT, 1940 THE CATHOLIC UNIVERSITY OF AMERICA PRESS Made in United States of America Exch. Clary Catholle Virit. of Araving DEC 2 41 02-17-4 2 GS ACKNOWLEDGMENTS The writer wishes to express her indebtedness and gratitude to all those who enabled her to carry through this piece of work. Special acknowledgment is made to Professor Thomas V. Moore, O.S.B., for his inception and development of the plan and for his assistance in the statistical evaluation of the results, and to Pro- fessor Thomas G. Foran and to Doctor Joseph F. Daly who read and criticized the manuscript. CONTENTS I. THE PROBLEM UNDER INVESTIGATION. II. SURVEY OF THE PREVIOUS WORK.. III. EXPERIMENTAL PROCEDURE. Preparatory Testing Program. Description of Tests.. Mode of Administration.. Mode of Scoring.. IV. STATISTICAL ANALYSIS.. V. SUMMARY. • The Curve of Development. Intercorrelations of Constants. BIBLIOGRAPHY. U 1 • • D • ► • · + • 1 4 1222 13 16 18 20 32 35 42 44 CHAPTER I THE PROBLEM UNDER INVESTIGATION The interest in factorial analysis has stimulated the study of and search for specific mental abilities. Psychologists are becoming more sensitive to the problem of the discovery of two or more fac- tors underlying a group of tests which might all be given a single It is particularly important to suspect memory tests of harboring two or more factors, for there are various memory func- tions; and it is a matter of some importance to distinguish them and to discover appropriate methods by which they may be measured. name. The investigation reported in the following pages was com- menced with the view of tracing the development of logical and of rote memory. It is, by no means, easy to isolate logical memory from rote memory, for experimental study. Meaningful material is presented to be learned; but the objective response required often results in emphasizing the mechanical or rote aspects at the expense of the meaningful. However, even if such a mechanical association exists, there is much more of it in the rote memory task than in the logical memory task which is dominated by the factor of meaning. In a comparatively recent review of "logical learning," the authors conclude their article by stating "Not a few have sug- gested that substance learning may have distinct laws from those governing nonsense syllables but only a few have devised experi- ments designed to discover those laws. The field is not new, but it is still wide open to fundamental research." Moore, more recently, considers some experimental data which point definitely towards a differentiation of sensory and intellectual factors in memory.2 The two types of memory have similarities and dif- ferences; but when does each type begin to show its development? ¹ E. L. Welborn and Horace English, "Logical Learning and Retention: A General Review of Experiments with Meaningful Verbal Materials,” Psychological Bulletin, 34, 1937, 1-20. 2 Thomas Verner Moore, Cognitive Psychology. Chicago: Lippincott Company, 1939. Part VI, Chapters i-ix, pp. 405–525. 1 2 DEVELOPMENT OF LOGICAL AND ROTE MEMORY The general plan of this experiment was to devise two batteries of tests, one for logical and the other for rote memory. Rote memory was conceived of as that form of memory in which an association between two items is obtained by presenting them in pairs time and time again, these pairs being so selected that there is no apparent bond of relationship between the members of any pair. The association between the items, therefore, is to be welded purely by dint of constant repetition. Logical memory may be tested by associating pairs of items that are naturally bound together by some logical relationship, such as that of part to whole, species to genus, attribute to subject. In testing of this kind, one does not get a pure measure of logical memory because there are various species which might be men- tioned for a genus and various parts for any given whole. Hence there is required a certain amount of rote memory to determine the individual items to be associated. However, the logical ele- ment enters in and the interesting problem arises: in what manner will the constants of the memory curve be affected by the logical factor? One difficulty with these tests is to secure the interest of the children. A preliminary study had been made by Sister Helen de Sales Forrest, who tried to associate the memory test with an interesting story that would present an attractive task to the children. In this way, much of the advantage of nonsense syl- lables could be obtained without giving the children the simple ungarnished task of frankly learning to associate nonsense syl- lables.3 At first we intended to study children from the first to the sixth grades, hoping thereby to find the point at which the two types of memory commence to develop. A preliminary experiment made it necessary to restrict the limits of the study. The logical memory tests were too simple for the sixth grade, and the rote memory tests too difficult for the first. The number concepts which are involved in the tests were too large to be handled by the first grade children. It was concluded, therefore, to deal with the second, third, fourth and fifth grades. 3 For details of this, see the study of Sister Helen de Sales Forrest, to be published shortly. THE PROBLEM UNDER INVESTIGATION 3 In order to measure the reliability of the tests, each was pre- sented in a double form. There were six different types of tests altogether. These are explained in more detail below. It was originally intended to have the logical memory and the rote memory tests of identical lengths. The preliminary experi- ments showed that the logical memory series had to be relatively long or it was learned with the first repetition. If a rote memory series were made of equal length to the necessarily long logical memory series, it would require an exorbitant amount of labor to be mastered. Therefore, we finally concluded to have the series for the two kinds of memory of different lengths; the logical being considerably longer than the rote memory series. The final experiments were carried out in a large parochial school. The subjects were boys and girls, average as to age and educational status. They may be looked upon as a fair sample of children to be found in any elementary school. CHAPTER II SURVEY OF THE PREVIOUS WORK 1 During the past two decades, many problems in the field of memory have been historically and experimentally considered. As recently as January, 1937, a review of the studies which help to bring to light the nature of logical memory was made by Wel- born and English. Their review contains the investigations made. since Whipple's material2 was published in 1915. The present survey may refer to experimental work previously reviewed, but it purposes to include only such studies that deal with a differen- tiation between rote and logical memory and those that trace the development of either type, especially the logical. Extensive investigations in the field of memory, carried on by Professor Michotte and his co-workers3.4 at Louvain, are excellent studies of the qualitative features of logical memory. They are mentioned here especially inasmuch as they show the difference between mechanical and logical memory under simple test con- ditions, e.g., the method of right associates. 5 Ballard, whose name is usually associated with the terms of obliviscence and reminiscence, finds in immediate memory that improvement varies with the material memorized. Moreover, when the experiments are extended to subjects differing in the matter of age, the memory curves reveal very marked differences. In the case of young children, ranging from five to seven years of age, the amount of reminiscence is much higher than with older children, viz., those of eleven or twelve; whereas, in the case of 1 Welborn and English, op. cit. 2 G. M. Whipple, Manual of Mental and Physical Tests. (Sec. Ed.) Baltimore: Warwick and York, 1915. Part 2, pp. 205–223. 3 A. E. Michotte and C. Ransy, "Contribution à L'Étude de la Mémoire Logique," Travail du Laboratoire de Psychologie Experimentale de L'Uni- versité de Louvain. Louvain, 1912. 95 pp. • A. E. Michotte and Th. Portych, "Deuxième Étude sur la Mémoire Logique," Extrait des Annals de l'Institut Supérieur de Philosophie. Lou- vain, 2, 1913. 533-657. 5 P. B. Ballard, "Obliviscence and Reminiscence," British Journal of Psychology, Monograph Supplement, No. 2. Cambridge: 1913. viii+ 82 pp. 4 SURVEY OF THE PREVIOUS WORK Сл 6 adults, no reminiscence at all is apparent. In fact, he says that the memory characteristics seem to disappear gradually with age. Dietze and Jones, measuring the factual memory ability of secondary school students, furnish evidence for an increase of memory scores from early life to maturity; but the increase through the period of adolescence, and within the population of the junior and senior high school, is not as great as in the earlier years. For an article read a single time, the memory varies with the material used. Wide individual differences too are noted. Prompted by the results of this study, Dietze experimented fur- ther with the same population. A correlation of .53 is noted be- tween mental age and immediate memory; but this is reduced considerably when such factors as vocabulary and reading ability are held constant. The author concludes that, from the results of the experiment, "it would follow that special attention to vocabu- lary and reading in the secondary schools would enable pupils to apprehend and remember more of what they read than they do under the present conditions." Memory for factual content, therefore, contains factors that are common to mental age and extent of vocabulary. 7 8 From another source, treating of the mental and physical de- velopment of children from two to twelve years of age, studies in the development of memory are included. Bryan' reports evidence, among five and six-year-old children, of a central factor which not only permeates the memory tests but extends through both the vocabulary and the intelligence tests. The author con- cludes that this central factor, probably one of memory, indicates a relation between the simple power of retaining and the more general ability, termed intelligence in children. Another author, 10 A. G. Dietze and G. E. Jones, "Factual Memory of Secondary School Pupils for a Short Article Which They Read a Single Time," Journal of Educational Psychology, 22, 1931, 586-598, 667-676. 7 A. G. Dietze, "The Relation of Several Factors to Factual Memory," Journal of Applied Psychology, 15, 1931, 563-574. 8 Lois H. Meek and Arthur T. Jersild, "Mental Development from Two to Twelve Years," Review of Educational Research, 6, 1936, 17-49. Alice Isabel Bryan, "Organization of Memory in Young Children," Archives of Psychology, No. 162, May, 1934, 56 pp. 10 Helena Mallay, "The Latent Memory Span of the Preschool Child," Child Development, 6, 1935, 110-119. 6 DEVELOPMENT OF LOGICAL AND ROTE MEMORY investigating the memory span of pre-school children, ranging from two to four years of age, reports that latent memory spans increase with age, generally following the negatively accelerated curve, but show variations which may be related to other factors than maturation alone. The experimental procedure, the type of material used, and certain personality traits of individual children were additional factors which affected the results. The length of memory span ranged from 0.3 days to 19.8 days. Hsiao," testing subjects from eight to nineteen years of age, found that, for mem- ory of physical relationships and also for logical abstract material, the growth curve had a tendency toward positive acceleration be- fore the age of 13. For logical-concrete and for concrete-asso- ciated-with-abstract material, the curve was linear. There was only slight growth, if any, after the age of fourteen. Studying the ideas of children after having seen a motion picture, Holaday and Stoddard¹2 "found that very young children remember correctly 50 to 60 per cent of what they see; 8 to 9-year old, 60 per cent; 11 to 12, 75 per cent; and 15 to 16, 91 per cent. It was found, in general, that second and third grade children, after six weeks, re- tained 90 per cent of what they could recall the day after they had seen a motion picture." McGeoch, 13 in an effort to associate age as a factor in reminiscence, found negative results, after trying a learning experiment with pre-school children and with college students. In another study,¹4 with the same range of subjects, McGeoch found no evidence that degree of learning is an influenc- ing factor. It would be quite difficult to construct a test that would equally motivate pre-school children and college students. This fact alone might raise exception to the results. Memory span is often taken as the measure of memory. True, 11 H. H. Hsiao, "A Preliminary Study of the Formal Development of Memory," Journal of Testing (Chinese), 3, 1933, 55–75. 12 Perry E. Holaday and George D. Stoddard, "Getting Ideas from the Movies." New York: Macmillan Co., 1933, 102 pp. (Published with: Charters, W. W., Motion Pictures and Youth: A Summary.) 13 Grace O. McGeoch, "The Age Factor in Reminiscence: A Comparative Study of Preschool Children and College Students," Pedagogical Seminary and Journal of Genetic Psychology, 47, 1935, 98–120. 14 Grace O. McGeoch, "The Factor of Degree of Learning in Reminis- cence," (Abstract) Psychological Bulletin, 31, 1934, p. 599. SURVEY OF THE PREVIOUS WORK 7 it is involved but it is not sufficient in itself. Brotemarkle¹5 calls it "a mental grasp and nothing more." Barlow16 shows that it improves rapidly and markedly, with increasing age, up to the college level. The number of syllables reproduced from lists. learned with verbal responses ranges from 2.27, in grade 2, to 15.34, at the college level; with restricted articulation, from 2.00 to 12.42. Prior to this, Bolton¹7 found memory span to increase with age rather than with intelligence. Intelligence, often ac- companied by a good memory-span and a great power of atten- tion, is not necessarily associated with them. The number of items which can be immediately grasped may improve with prac- tice; it is facilitated too by grouping. Northway18 points out the influence of age and social group on remembering. Experimenting with children, ten, fourteen and fifteen years of age, she found that they show tendencies to cast the material in familiar terms and social settings. The younger they are, the more rapidly do children change the subject matter into a form nearer their own interests. This indicates a decided age difference, for the older children strive to retain the given material in its own form. Foster, 19 testing the verbal memory of pre-school children, found that her learning curves approach the linear, showing no tendency to become flattened. The children, both with the higher C.A.'s and M.A.'s, reproduced a greater amount of material. Their correct responses were affected not so much by the familiar- ity of the words as by the proximate relationship between the last word read and the one requested. Zankov20 reports the results of a study of the memory of two 15 R. A. Brotemarkle, "Some Memory Span Problems," Psychological Clinic, 15, 1924, 220–258. 18 M. C. Barlow, "The Rôle of Articulation in Memorizing," Journal of Experimental Psychology, 11, 1928, 306–312. 17 T. L. Bolton, "Growth of Memory in School Children," American Journal of Psychology, 4, 1909, 362–380. 18 Mary L. Northway, "Influence of Age and Social Group on Children's Remembering,” British Journal of Psychology, 27, 1936, 11–29. 19 J. C. Foster, "Verbal Memory in Preschool Children," Journal of Genetic Psychology, 35, 1928, 26–44. 20 L. V. Zankov, "The Development of Memory in Morons: I. Experi- mental Investigations," Journal of General Psychology, 16, 1937, 415-426. 8 DEVELOPMENT OF LOGICAL AND ROTE MEMORY "" hundred morons, with an average I.Q. of about 70, their chrono- logical age ranging from 9 to 15 years. This investigator says the development of memory "is manifested in qualitative changes, in the appearance of new forms of memory." The general line of progression is "from immediate (mechanical) memorizing to voluntary (volitional) intelligent (logical) remembering and re- producing, from the visual-object memory to the verbal memory.' Does the memory of feeble-minded children develop along the same course? The results of this investigation show that there is but slight development of rote memory during the school age. However, the remembering of words in pairs (a logical ability) shows a point of sudden development somewhere between the ages of 10 and 11. The errors made by the nine and ten-year- old subjects prove the fact they had not yet mastered logical remembering. At the age of 11, the feeble-minded child shows the beginning of mastery of intellectual remembering; at the age of 12, the higher form of memory is definitely established. There is a considerable improvement between the ages of 11 and 15; and, judging by the type of curve, it has a tendency to further develop- Although Zankov's data may not represent exhaustive evidence, they give an idea of the curve of development of logical memory of morons during the school years, which is a point in favor of the scholastic possibilities of the feeble-minded. Hartmann and Schooley21 tested two hundred children from 5:0 to 7:6 years of age to find out when and how children who do not know the meanings of standard logical relations learn them. Is it gradually, or in accordance with trial-and-error theories, or suddenly with the appearance of insight? They conclude the following: (1) Learning standard logical relationships is not a slow, gradual process involving practice-effect or trial-and-error responses but occurs suddenly and gives evidence of permanent mastery. When the desired concept is once formed, the child responds 100% correctly to the remaining words illustrative of the learned relationship, providing they are within the range of his vocabulary. 21 George W. Hartmann and Mary Schooley, "The Rôle of Insight in the Learning of Logical Relations," American Journal of Psychology, 49, 1937, 287-292. SURVEY OF THE PREVIOUS WORK 9 (2) The probable order of difficulty of learning the relations is from easiest to hardest as follows: Action-agent; agent-action; attribute-sub- stance; substance-attribute; genus-species; whole-part; part-whole; and species-genus. (3) With the amount of training here given, children similar to those used have a 0.5 probability of learning the standard logical relations when they are members of age-groups varying from 5.1-5.7 to 6.4-6.10 years. The narrow range here of one year and three months indicates the close interre- lationship of all the logical relationships studied. Burtt22,23 carried on a lengthy experiment, using but one sub- ject, who, at the beginning of the study, was fifteen months old. Every three months a different set of similar selections was used, which procedure was continued until the child reached the age of three. At 8, he learned some of the original material, together with new selections. At the age of 14, a similar experiment was conducted, using still other material. The results are summarized as follows: At the age of 83, the relearning required approximately 30 per cent fewer repetitions, than were needed to learn new material. At the age of 14, the corresponding figure was 8 per cent. At the age of 81, the original selections which had been presented later in infancy were relearned more readily than those presented lower in infancy. At the age of 14, this difference had disappeared. Learning curves, however, show some sig- nificant differences between the ordinates for learning and relearning at the age of 14. The effect of the presentation in infancy was very marked at the age of 81, but at the age of 14 the effect, while still appreciable, was considerably decreased. In a series of experiments24 among college students, substance memory was compared with verbatim memory by two contrasted types of true-false test items from a passage of prose that had been presented for study. One reproduced almost verbatim por- tions of the text; the other required the recognition of meaning in entirely different words. The retention curves for the two sorts of items consistently showed opposed trends. Those for the 22 H. E. Burtt, "An Experimental Study," Journal of Genetic Psychology, 40, 1932, 287–295. 23 H. E. Burtt, "A Further Study of Early Childhood Memory," Journal of Genetic Psychology, 50, 1937, 187-192. 24 H. B. English, E. L. Welborn, and C. D. Killian, "Studies in Substance Memorization," Journal of General Psychology, 11, 1934, 233–260. 10 DEVELOPMENT OF LOGICAL AND ROTE MEMORY verbatim items gave a typical Ebbinghaus decline; those for the substance items did not resemble forgetting curves at all but learn- ing curves. The authors suggest, rather than conclude, that "These results are taken to indicate that substance learning differs not only from the learning of nonsense syllables but also from so- called logical learning in which the words of the original text are preserved." Carlson25 analyzed memory ability by the factorial method. The author assumed the division of logical and rote functions and a further division of rote into visual and vocal memory. His interpretation of the factors did not definitely establish the truth. of the assumptions. However, he says that if his alternative interpretations are correct, he found the differentiation which was postulated. The statistical analysis brought out a general factor common to all the tests in the battery, which was by far the most important single factor involved in the memory of words. Bringing the experimental studies up to the most recent past, the names of Carlson and Carr will conclude this brief survey. Their first study26 may be summed up in the concluding words of their report: Series of visual recognition tests of words favoring the use of vision alone, vocality alone, and vision and vocality combined, were given to 202 subjects. A considerable number of subjects were found to be consistently superior in one series, as contrasted with another. The explanation ad- vanced is that these subjects differ in their ability to use visual, vocal, and visual-vocal cues. It follows then that some of the conflicting reports previously made comparing memory efficiency of the different senses may be due in part to differences in the abilities of the subjects comprising the groups involved. Their latest experiment27 was an attempt to find whether indi- viduals tend to maintain their relative rank on both rote and logical recognition memory tests involving certain restricted materials, or whether they tend to vary in their relative rank in 25 Hilding B. Carlson, "Factor Analysis of Memory Ability," Journal of Experimental Psychology, 21, 1937, 477-492. 26 Hilding B. Carlson and Harvey A. Carr, "Visual and Vocal Recogni- tion Memory," Journal of Experimental Psychology, 23, 1938, 523-530. 27 Hilding B. Carlson and Harvey A. Carr, "Rote and Logical Recogni- tion Memory," Journal of Experimental Psychology, 26, 1940, 199–210. SURVEY OF THE PREVIOUS WORK 11 accordance with the rote or logical character of the material to be memorized. Finding the former would mean that the dis- tinction between rote and logical recognition does not necessarily involve different abilities. Welborn and English,28 in their re- view of logical learning and retention, concluded that "different factors operate in substance memory and in rote memory," and that "substance memory may have distinct laws from those governing nonsense syllables." Reed29 oppositely states, "Mem- ory for prose substance is thought by some psychologists to be totally different from memory for rote learning, but it probably differs from it in degree rather than in kind." Carlson and Carr summarize their findings as follows: A considerable number of subjects was found to be consistently superior (or inferior) in the meaning series as contrasted with one of the rote series. Some of these subjects are consistently superior (or inferior) in the meaning series as contrasted with two rote series. Others are consistently superior (or inferior) in the meaning series as compared with all three rote series. From this it follows that there are individual differences in ability to utilize rote and logical memory, and therefore the recognition memory involves at least a rote and a logical component. This brief survey of the previous studies that furnish experi- mental evidence for a differentiation between rote and logical memory, as well as for those that trace the development of either ability, manifests the paucity of publications which deal with the investigations of children's memories. Especially those that might bring to light the nature of the development of the logical type are apparently non-existent. Most of the studies of the measurement of children's memories account for the development in terms of "range of span"; which, in itself, is not sufficient. In- terest in the structure of suitable testing material, that will lend itself to mathematical functions, may instigate further experimen- tation, which might result in a knowledge of the curves of memory development. 28 E. L. Welborn and Horace English, "Logical Learning and Retention: A General Review of Experiments with Meaningful Verbal Materials," Psychological Bulletin, 34, 1937, 1–20. 29 H. B. Reed, “Meaning as a Factor in Learning," Journal of Educa- tional Psychology, 29, 1938, 419–431. CHAPTER III EXPERIMENTAL PROCEDURE A. PREPARATORY TESTING PROGRAM About two months prior to the administration of the memory tests used in this investigation, the children of the entire school were given, by their principal, the Kuhlmann-Anderson Intelli- gence Test. The following week, their achievement was measured by Unit Scales of Attainment, a battery of scaled achievement tests for measuring abilities and ranges of information. The median scores of these pupils on the intelligence test and on the achievement battery corresponded closely to published norms. The daily attendance registered a little over thirteen hundred pupils. So it may be safely assumed that the results of the tests. showed the children to be fairly typical of the nation's elementary school population. The individual teachers corrected the tests and the principal graphed the results, after which a representative of the Test Bureau was invited to the school to confirm the inter- pretation of the data. The writer, on her first visit to the school, gave one of the Mem- ory Tests, formerly used by Monaghan,¹-the Fable Completion, Form B. They were corrected and filed; but the results were not included in the experiment. However, its administration aroused interest and helped to eliminate factors that might impede mem- orization or learning in the presence of a stranger. For the same reasons, McManama's Cognition Test, Form B,2 was also ad- ministered to Grades IV, V, and VI. Moreover, this test cor- relates high with certain standardized psychological tests; and its use might necessarily serve as a check for further examination and comparison of the data secured. 1 Edward A. Monaghan, "Major Factors in Cognition," Studies in Psychology and Psychiatry, No. 5, 1935, 48 pp. 2 Sister Maurice McManama, "A Genetic Study of the Cognitive General Factor in Human Intelligence," Studies in Psychology and Psychiatry, No. 2, 1936, 35 pp. 12 EXPERIMENTAL PROCEDURE 13 OCEAN-FISH FIRE-SMOKE CIRCUS-CLown DANCE-MUSIC B. DESCRIPTION OF TESTS Association tests of logical memory. The material to be learned consisted of twelve paired associates in each series. An additional pair was included in each set, as a sample to be followed. Every attempt was made to have the material easy enough for the chil- dren, and, at the same time, interesting. Preliminary experi- menting had determined the difficulty of the material and the best number of words to include in a series. Because they were found to be rather difficult for the children, words of abstract meaning especially were ruled out and were changed to those of a more concrete nature. No word was used which was not found in Gates' Reading Vocabulary for the Primary Grades.³ 3 Six types of pairs were selected in order to obtain a range of material and to keep up the interest: PLACE-OBJECT; PART —WHOLE; GENUS SPECIES; TIME-OBJECT; QUALITY -OBJECT; and one in which the child was to learn the asso- ciated word by making his own connection between the pairs. (The complete tests are on file in the Department of Psychology of the Catholic University of America.) The words presented to be memorized are as follows: 1. PLACE-OBJECT. 2. PART-WHOLE. LEG-TABLE BARK-TREE EYE-NEEDLE TICK-WATCH KITCHEN-PAN (Sample) STREET-Bus PARTY-CAKE DESK-INK FLOOR-RUG WING-FLY (Sample) BUZZ-AIRPLANE GLASS-PICTURE WALL-ROOM POINT PIN PARK-GRASS GARDEN-FENCE WAGON-SEAT MOUNTAIN-BEAR FUR-SQUIRREL HEAD-BODY FEATHER-CHICKEN OVEN STOVE ³ Arthur I. Gates, A Reading Vocabulary for the Primary Grades. Re- vised and enlarged; New York: Bureau of Publications, Teachers' College, Columbia University, 1935. 29 pp. 14 DEVELOPMENT OF LOGICAL AND ROTE MEMORY 3. GENUS SPECIES (worded for children, "FAMILY-WORD BE- LONGING TO FAMILY"). FRUIT-ORANGE HOUSE-HOME FOOD-MEAT TREE-OAK 4. TIME-OBJECT. NIGHT-BED WINTER-SNOW SUNDAY-CHURCH SICKNESS-DOCTOR SOFT-COTTON SWEET CANDY BLACK-COAL DARK-CORNER FLOWER-LILY (Sample) BIRD-CROW TOY-DRUM VEGETABLE-CARROT BELL-TEETH PIGEON-FATHER COMB-WIND BOOT-LAMP TOOL-SAW Noon-DINNER (Sample) PLAY-GAME HALLOWE'EN-WITCH 5. QUALITY—OBJECT. BREAKFAST OATMEAL BIRTHDAY-PARTY RED-BLOOD (Sample) WET-RAIN SHORT-DWARF HOT-IRON SHARP KNIFE 6. ASSOCIATION TO BE MADE BY CHILD. PEN-WATER (Sample) LOCK-MOTHER BEE-CHAIR BRUSH-PEAR INK-COW BUG-SPIDER MAN-DOCTOR DISH-PLATE MONEY-PENNY WAR-FIGHT VISIT TALK SUMMER-PICNIC SLEEP-DREAM COLD-ICE-CREAM HIGH-CLOUD PRETTY-FAIRY QUICK-RABBIT $ SCISSORS-MILK KETTLE-Book BREAD-HORSE DOLL-EVENING Association tests of rote memory. Forrest had constructed a battery of six tests, which she termed "Association Tests." When one reads the directions of these tests, he can see that the material is interesting and furnishes strong motivation for chil- dren's memorization. There were no obvious logical connections between the paired associates, eight of which were in each series. Forrest included ten rows of each set on every presentation, instructing the subject to fill in as many spaces as he could until Sister Helen de Sales Forrest, "The Correlation between the Constants in Association and Forgetting." (Dissertation to be published shortly, Department of Psychology, Catholic University of America.) EXPERIMENTAL PROCEDURE 15 the signal was given to stop. For this experiment, the tests were modified. Only one set of associates was given on each presen- tation; that is to say, when the children had been given the series of names and their corresponding numbers or colors or prices, etc., they looked at the first page of a little booklet before them on which was typed merely the original series of names, and they wrote under each name as many of the pertinent numbers, colors, etc. as they could remember. The materials presented to be learned are: 1. Name and Number of Football Player. FRED DICK 69 24 BILL DAN JOHN TOM HARRY ED 86 33 52 71 95 47 2. Name of Girl and Color of Her Bathing Cap. MARY PEGGY ANNE JANE HELEN BETTY WHITE GRAY BLUE BLACK GREEN ORANGE 3. Name and Price of Toy. BALL WATCH CAR 21 43 84 HORSE TRAIN SHIP 98 50 37 4. Name of Horse and Number of Jockey. WIND SPRING 36 BLOOD KING HONOR 65 90 22 78 KITTY RED PID 74 POWER TRUE 44 57 BEAR TABLE 66 79 6. Prisoner's Favorite Radio Station and His Prison Number. JEB WOY NOF KAG 23 ZAP 32 REZ 59 96 87 61 ALICE YELLOW 5. Two Passwords for Membership in Boys' Club. BANK BREAD FIELD DARK EVER EARTH NAME MIND CROSS FRONT HOLD FIRST ABOUT WISH POINT TURN WAR 85 MIP 45 Organization of test material. The material for both types of tests was arranged in mimeographed booklet form. Each booklet 16 DEVELOPMENT OF LOGICAL AND ROTE MEMORY for a series contained a page of directions and ten additional pages on which were typed one list of the paired associates with blank spaces where the subject might write the corresponding associate. The same list of words appeared for ten presentations, but it was arranged in different order. A blank sheet was inserted between consecutive test sheets. For presentation, the paired associates for each series were printed in black, on 4" x 7" buff colored cards, by means of one-inch, heavily stamped, bold-faced figures and alphabet letters. C. MODE OF ADMINISTRATION This testing program was conducted as a group experiment during the regular school hours and in the home classroom of each grade. It was a large school in which there were three classes to each grade. For obvious reasons, it was deemed advisable to ad- minister the tests to all the groups,--one group each being taken by the principal of the school, the local Supervisor, and the writer, no one unfamiliar to, or unacquainted with, the children. These three, after studying the nature and directions of the tests, held a preliminary conference. They aimed to keep the con- ditions fairly constant. Each experimenter was equipped with a stop watch. The home room teacher remained in the room but took no actual part in the experiment, except to pass and collect papers and to keep a sur- veillant eye about the classroom. However, in a few instances, the teacher gave the concluding series when the testing was ex- tended over a period of two or more sittings. She followed the technique of the preceding experimenter by adhering strictly to the directions. No particular time of day was specified for testing; but the last period of both the morning and afternoon sessions was never used. The six logical memory tests were given in all the groups. before those of rote memory were commenced. To avoid fatigue, especially enhanced in the spring of the year, when most of the memory tests were given,-only two or three of the series were given each day, particularly in the lower grades. The fifth grade, however, finished each series in one session, with a short recess after the fourth set of tests. This highest grade also was given EXPERIMENTAL PROCEDURE 17 less time, for the reason they could write their responses more quickly. For each of the ten trials in a series, one and a fourth minutes were allowed for the fifth grade; one and a half, for the fourth; one and three fourths, for the third; and two, for the second. General instructions were first given by each experimenter to the children within her group. They were told that the boys and girls of the other classes within their school, as well as children of neighboring and distant schools, were participating in the "game." After asking for their cooperation, the tester assured her subjects that the scores they would obtain would not affect their regular school marks. She further explained that no credit was to be deducted for spelling and asked the children to write down every word they could remember, even if they were not sure of the spelling. The lowest grade was told they could write, or even print, just a part of the word. The experimenter dis- couraged guessing and aimed to motivate learning by telling the subjects she wanted to help them see how they could improve their memory. She concluded by remarking it would improve with each trial given them and that they should not be discouraged if they did not get all the responses learned on the first attempt. The booklets of the first series were then placed, face downward, on the pupils' desks. At a given signal, the subjects turned them over and were instructed to fill in the blanks giving such informa- tion as their name, age, grade, and date of test. As they finished, they rested their elbows on the desks and held their pencils, point upward. Then the directions on the front page were read by the experimenter. Before the actual experiment was started, a card, with a pair of associated words, was shown as a sample. On the blackboard, the first word was printed in the type similar to that which the child would encounter on the first trial page. Immediately fol- lowing it, was written the second word as the child should write. it on the paper. The PLACE OBJECT cards were then presented. They were flashed only for the length of time it took to say the words de- liberately and distinctly. After the last card was presented, the children were instructed to turn to page one of the booklet and 18 DEVELOPMENT OF LOGICAL AND ROTE MEMORY write the second word that paired with the printed word. At the end of the allotted time, the signal "STOP" meant pencils in non- writing position. Re-presentations of the cards were alternated with test trials until ten results had been obtained. Many of the children, especially in the logical association tests, learned the list perfectly before the tenth trial. In order not to cause unnecessary confusion, all the pupils took the ten trials; but the first trial on which all the responses were given correctly was a definite factor in determining the individual constants in the statistical for- mulae. After the PLACE OBJECT series was completed, each of the other members of both batteries of tests was presented by the same procedure until the entire program had been carried out. At the end of the first session of the experimental period, the children were asked not to discuss the tests with the other classes until the whole series was finished. The children's interest and hearty cooperation gave promise that this request would be respected. However, the experimenters, for the most part, were able to administer the same sections of the tests to all the grades on the same days. D. MODE OF SCORING The tests were designed to measure the number of associations remembered after each presentation. Hence the best type of score seemed to be the total number of correct responses on each trial. In this way, every item memorized was considered a unit. No attempt was made to deduct for guessing. To eliminate guessing in some measure, instructions were given not to guess. The average of the total responses of each presentation of the ten trials on both types of tests, treated separately, gave the raw scores. After the raw scores of the ten presentations of every individual were transferred to his respective sheet, the next step was to smooth—in threes-the series by obtaining the moving average. By smoothing in threes is meant that the number of items recalled with the first, second and third repetitions were averaged and termed the second trial; the second, third and fourth were averaged and termed the third trial, and so on. Modifications of the moving EXPERIMENTAL PROCEDURE 19 average procedure had to be introduced for quick learners. If, for example, a child learned all the material in three repetitions, his raw scores were taken as they stood, without smoothing. If a child learned all the items before the tenth repetition, the trial on which he completed his learning was naturally his nth trial in the curve given below: Log Yn a bcn. In this curve, as we shall see, a represents the logarithm of the total number of items to be learned, that is 8 for the rote and 12 for the logical series. Nearly six hundred children were given the tests. However, after the administration of the tests and the recording of the raw scores, it was decided, on account of the length of the statistical analysis, to include only the group of the writer in the final results. This decision limited the final number of subjects to one hundred eighty-one. CHAPTER IV STATISTICAL ANALYSIS - The current techniques of factorial analysis leave much to be desired. Any attempt to discover other ways of analyzing the complex data of mental life may perhaps lead to a valuable ex- tension of our methods of research. One possible method of pro- cedure is to find a mathematical curve which expresses a rela- tionship between the variables involved and to attempt an interpretation of its constants. Any curve which closely fits the data over the field of investigation will serve. Each constant required by the curve must have a meaning. Let us stop for a moment and consider what is here meant by a constant. We may perhaps make this clearer by considering the curve which we have decided to apply to the data of memory. The curve is represented by the equation. log Yx a bc. This is the general law expressing the relationship between the task, the amount learned, and the number of repetitions. The constant involving the task, as we shall see, is a. This remains the same for all individuals included in the experiment. That is to say, it was Log 8 for rote memory and Log 12 for logical mem- ory. The constants b and c are different for each individual, for each individual has a different curve of learning. In other words, relative to the whole population, a is a constant and b and c are variables. In the original curve, therefore, b and c must measure something individual and peculiar to each person in the entire population studied. Log y, is the logarithm of the number of items remembered at trial x. Cl A fundamental problem is to find out the meaning of the con- stants b and c. This was solved in a previous investigation.¹ The particular problem of the present investigation was to apply 1 Thomas Verner Moore, "The Analysis of Association by Its Equational Constants," Aspects of the New Scholastic Philosophy. New York: Benziger Brothers, 1932, 181-225. 20 STATISTICAL ANALYSIS 21 this equation to memory which was essentially rote learning and then to memory which offered an opportunity for logical associa- tion. It was thought that if, as Hunter² has maintained, there is no difference between logical and rote memory, the average constants in the two types of performance should have identical values within the limits of the probable error. If, on the other hand, fac- tors enter into logical memory which are not present in rote mem- ory, there should be a significant difference between the values of the constants in the two series. Yo = THE CURVE OF LEARNING AND ITS INTERPRETATION3 The equation Log y₂ = Log a bc is derived from the con- cept of a series in which the first differences of the logarithms of y form a geometric series. For example 10d Y₁ = 10*+d Yı 10* 10k+d+dc Ys = 10k+d+de+de² Y2 Yn 10k+d+de+dc² · dcn-1 Y1 Yo Y2 Y3 Y2 Y n Yn-1 ―――― 10dc 10dc 2 10den-1 Log y₁ — Log y。 = d Log y2 C Log y₁ = dc Log ys - Log Y2 = dc² Log y = Lim (k + d + dc + · · · dc"-¹) 1100 Log yn - Log Yn-1=dc" Here k is the logarithm of the first member of the series and the limiting value of the logarithm of the highest member of the series is n-1 2 Walter S. Hunter, "Experimental Studies in Learning," The Founda- tions of Experimental Psychology. Worcester, Mass.: Clark University Press, 1929. Chapter 15, pp. 564–627. Moore, op. cit., pp. 217 ff. 22 DEVELOPMENT OF LOGICAL AND ROTE MEMORY Now let Log y∞ - Log yn = R yo or i=p Σ ent i=0 Therefore, Log yn = k + d + dc + dc² - . . . Log yn Yn ******* When n = 0, i= p R = Lim Σ dc²+i nti p=∞ i=0 Log yn Expressed as above, Log yr Yx ******* = = n n+1 c² + c²+¹ + c^+2 = Log y∞ - R Log Y∞ but the series on the right, as can be found by actual division, equals ch 1 - с Log Yoo — Log Yoo i=p Lim Σ denti p=∞ i=0 + · Log Yo = d n (1 == 0) c с d x Therefore, b in the equation (Log y = a - d 1 - c _______ What is the meaning of yo in the equation Log abc"? Log yn = Log a - b J Log yo+ b = Log a. C Cz n+∞ dc"-1 - bc") is equal to STATISTICAL ANALYSIS 23 That is to say the task to be performed may be divided into two parts: (i) a task which could be accomplished without practice, re- lated to Log Yo. (ii) a task which must be accomplished by dint of repetition, related to the constant b. Log yo is, therefore, a measure of something equivalent to span, whereas b is a measure of the resistance of the nervous system which must be overcome by dint of repetition. It is, of course, something which varies with the task. Its size for a given task varies from individual to individual. Were it not for this resist- ance, a single presentation of the material would suffice to impress the whole amount to be learned permanently on the memory. Every repetition lowers bc until it approximates zero, and, at the nth repetition, Log yn Log a, with a very close approxima- tion. Then, = Log yo is the complement of b. This may be seen by letting n equal zero in the equation. Log yo= Log a bcn. = Log Yo Log a - b. Log yo is equal to a constant defining the task minus b. This constant defining the task is the same for all the individuals in the population; but b and c vary from person to person. There must, therefore, be a negative of unity correlation between y。 and b, and all the correlations of b with any other constant must be the same as those of yo with the signs changed. The fundamental abilities of memory measured by this equa- tion are (i) the power of memorizing due to repetition, measured by the effect of each repetition, and expressed in our equation by c. The greater this power, the lower is the c. So 1-c is a measure of this power, it being understood that c is always positive and less than unity. Unless these conditions are fulfilled, learning does not take place, no matter how many repetitions may be involved. (ii) the resistance that must be overcome in order that the whole amount to be learned may be repeated without any error. This is a function complementary to what psychologists have long GRADE II PUPIL #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 * 20 #21 *22 * 23 * 24 #25 #26 #27 #28 #29 #30 #31 #32 #33 #34 #35 #36 #37 #38 #39 #40 #41 #42 #43 C .8234 .6752 .6622 .9731 .9484 .5109 .9037 .8632 .2955 .7595 .8368 .9504 .9451 .9381 .9104 .4862 .5385 .6409 .9498 .6818 .6399 .8389 .7327 .4060 .3792 .6414 .7652 .8323 .6698 .5146 .3236 .7207 .4962 .5511 .6663 .5802 .5736 .5156 .6621 .4781 .6409 .5216 .6347 Logical Association Tests b 1.0027 .1410 .2711 .6550 .5598 .0870 .1883 .4986 .1102 .1165 .1659 .2383 .2011 .0561 .6900 .1599 .1609 .1051 .2719 .2246 .1888 .1780 .0838 .3226 .0997 .2409 .3408 .3152 .6476 .1099 .2372 .0486 .1728 .0504 .6498 .9776 .9894 .4881 .6157 1.2257 1.1945 1.8970 1.2334 Σ = 29.0778 = 18.2115 M. = .6762 | Mь = .4235 σε = .1815 σε = .4104 ܝ d .17708 .04580 .09158 .01762 .02889 .04255 .01813 .06821 .07764 .02802 .02707 .01182 .01104 .00347 .06182 .08216 .07426 .03774 .01365 .07147 .06799 .02868 .02240 .19162 .06189 .08639 .08002 .05286 .21384 .05335 .16044 .01357 .08706 .02262 .21684 .41040 .42188 .23644 .20805 .63980 .42894 .90752 .45056 24 1.5034 1.1112 1.2347 1.0414 1.0541 1.1029 -1.0426 1.1701 1.1958 1.0667 1.0643 1.0274 1.0258 1.0080 1.1530 1.2083 1.1865 1.0908 1.0319 1.1789 1.1694 1.0683 1.0529 1.5546 1.1532 1.2201 1.2023 1.1294 1.6362 1.1307 1.4469 1.0317 1.2220 1.0535 1.6476 2.5728 2.6418 1.7237 1.6146 4.3631 2.6850 8.0820 2.8220 Σ = MD σD Ꭰ 67.7216 | Σ = LOG VO .07648 .93818 .80818 .42418 .51938 .99218 .89088 .58058 .96898 .96268 .91328 .84088 .87808 1.02308 .38918 .91928 .91828 .97408 .80728 .85458 .89038 .90118 .99538 .75658 .97948 .83828 .73838 .76398 .43158 .96928 .84198 1.03058 .90638 1.02878 .42938 .10158 .08978 .59108 .46348 -.14672 -.11532 — .81782 .15422 28.19314 1.5749 MLog V 1.1970 Log Yo .6557 .4111 GRADE II PUPIL #1 1 2 #2 #3 #4 #5 *** #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #25 * 26 #27 #28 #29 #30 #31 *32 #33 #34 #35 #36 #37 #38 #39 #40 #41 #42 #43 C .9287 .8978 .8172 .8807 .9698 .8989 .9332 .9119 .7777 .8636 .8787 .9281 .9412 .9017 .8046 .8118 .7970 .9060 .8817 .9365 .8213 .8480 .8678 .6513 .5447 .6990 .7961 .8341 .7898 .8305 .8191 .7845 .7114 .8049 .8711 .8338 .9397 .8353 .8434 .8833 .8983 .9059 .9480 Rote Association Tests b .5641 .5579 .3432 .8345 .5237 .3507 .4295 .4814 .5079 .5476 .4546 .5940 .5711 .7006 .4920 .4360 .4079 .4171 .5270 .3677 .2682 .2985 .4592 .3814 .4129 .3295 .3931 .6578 .9342 1.0317 .5728 1.2314 1.1098 .6866 .9877 .6521 .8278 1.0281 .8607 .7736 .6227 .8115 .9996 ΣΕ 36.4281 |Σ - 26.4394 M. = .8472 | Mь · .6149 = σc = .0827 .2419 σε σ b d .04022 .05702 .06274 .09956 .01582 .03546 .02869 .04241 .11291 .07469 .05514 .04271 .03358 .06887 .09614 .08206 .08280 .03921 .06234 .02335 .04793 .04537 .06071 .13299 .18799 .09918 .08015 .10913 .19637 .17487 .10362 .26537 .32029 .13396 .12731 .10838 .04992 .16933 .13479 .09028 .06333 .07636 .05198 25 D 1.0970 1.1403 1.1554 1.2576 1.0371 1.0851 1.0683 1.1026 1.2969 1.1877 1.1354 1.1033 1.0804 1.1718 1.2478 1.2080 1.2100 1.0945 1.1544 1.0552 1.1167 1.1101 1.1500 1.3583 1.5413 1.2565 1.2027 1.2857 1.5717 1.4958 1.2695 1.8423 2.0907 1.3613 1.3406 1.2834 1.1218 1.4768 1.3639 1.2311 1.1570 1.1922 1.1272 Σ = 53.8354 MD 1.2520 "D = .2058 = LOG YO .33899 .34519 .55989 .06859 .37939 .55239 .47359 .42169 .39519 .35549 .44849 .30909 .33199 .20249 .41109 .46709 .49519 .48599 .37609 .53539 .63489 .60459 .44389 .52169 .49019 .57359 .50999 .24529 .03111 -.12861 .33029 -.32831 -.20671 .21649 -.08461 .25099 .07529 - .12501 .04239 .12949 .28039 .09159 -.09651 Σ = 12.45669 MLog v σ Log V. .2897 .2402 GRADE III PUPIL **** 123 L *1 #2 *3 #4 #5 #6 #7 *8 *9 10 #11 *12 *13 *14 #15 #16 * 17 #18 #19 *20 #21 * 22 #23 * 24 *25 * 26 *27 * 28 * 29 *30 #31 *32 *33 #34 #35 #36 37 #38 *39 *40 #41 * 42 #43 σε с .5073 .7299 .7792 .7708 .8552 .7823 .8384 .6037 .6435 .4266 .3754 .1388 .0537 .7816 .5172 .7387 .5919 .6967 .7911 .8358 .6785 .7286 .6259 .4427 .6157 .4037 .5711 .3031 .8710 .6415 .5605 .7547 .9289 .3253 .7873 .5956 .5203 .5605 .3947 .5265 .5755 .6089 .7795 ― Logical Association Tests b .1900 .2423 .4049 .3115 .8098 1.0105 .2492 .3912 .6626 .1359 .4610 .2096 .7092 .4791 .6292 .6976 4359 .2245 .5463 .5487 .2969 .3876 .2586 .1507 .8946 .2987 .5059 .2693 .6369 .2067 .3027 .4962 .6743 .3910 .2298 .1781 .3740 .2080 .4297 .1396 .0806 .2287 .2256 • Σ = 26.2578 Σ = 17.1954 Mc.6106 | M¿ .3999 .2190 .1914 σb .09361 .06545 .08940 .07140 .11726 .21999 .04028 .15503 .23622 .07793 .28794 .18051 .67116 .10464 .30378 18228 17789 .06809 .11412 .09010 .09545 .10519 .09674 .08401 .34379 .17811 .21700 .18768 .08216 .07410 .13312 .12172 .04794 .26381 .04889 .07202 . 17941 .09142 .26010 .06610 .03421 .08945 .04974 • • d 26 1.2415 1.1626 1.2286 1.1787 1.3099 1.6595 1.0972 1.4290 1.7238 1.1966 1.9406 1.5153 4.6899 1.2724 2.0125 1.5215 1.5062 1.1694 1.3008 1.2305 1.2458 1.2741 1.2495 1.2134 2.2069 1.5070 1.6482 1.5416 1.2083 1.1861 1.3587 1.3235 1.1167 1.8353 1.1192 1.1804 1.5115 1.2343 1.8201 1.1644 1.0820 1.9474 1.1213 Ꭰ = = LOG YO .88918 .83688 .67428 .76768 .26938 .06868 .82993 .68798 .41658 .94328 .61818 .86958 .36993 .60008 .44998 .38158 .64328 .85468 .53288 .53048 .78228 .69158 .82058 .92843 .18458 .78048 .57323 .80988 .44228 .87248 .77628 .58298 .40488 .68818 .84933 Σ 63.4822 Σ = MD 1.4763 MLog yo .5694 Log Yo σD = .90108 .70518 .87118 .64948 .93958 .99858 .85048 .85358 29.19109 .6789 .2182 GRADE III PUPIL #1 *2 #3 #4 #5 #6 #7 #8 #9 10 #11 #12 *13 *14 #15 #16 #17 #18 *19 * 20 *21 * 22 * 23 *24 * 25 * 26 #27 * 28 * 29 *30 #31 #32 #33 *34 * 35 *36 #37 #38 #39 #40 #41 * 42 * 43 Σ .8012 .9494 .7909 .8632 .9076 .8757 .9324 .8855 .8026 .5857 .8718 .7425 .9404 .9481 .7982 .9002 .8934 .9074 .9399 .8956 .9751 .8393 .8070 .8145 .8930 .8700 .8610 .8855 .8984 .8254 .8495 .8092 .7383 .9015 .8801 .5960 .6603 .8222 .6941 .8057 .7706 .9154 .9008 = с Mc Oc 36. 2446 | Σ .8429 .0879 Rote Association Tests b .3275 .3445 .4985 .3903 .8423 .5881 .3873 .3947 .7421 .4113 .7383 .2127 .1239 .9237 .5328 .5340 .5722 .5912 .7350 .5307 .8688 1.1800 .3434 .5350 .5320 .3218 .5105 .5760 .8092 .6233 .3835 .5956 .6586 .4515 .4933 .5468 .6664 .4421 .1987 .3303 .4269 .2689 .4394 S 22.6231 .5261 Mb ơi = .2059 σb = d .06511 .01743 .10424 .05339 .07783 .07310 .02618 .04519 .14649 .17140 .09465 .05477 .00738 .04794 .10752 .05327 .06100 .05475 .04417 .05541 .02163 .18963 .06628 .09924 .05692 .04183 .07096 .06595 .08221 .10883 .05772 .11364 .17236 .04447 .05915 .22091 .22638 .07861 .06078 .06418 .09793 .02275 .04359 27 Ꭰ 1.1617 1.0410 1.2713 1.1308 1.1964 1.1833 1.0621 1.1097 1.4011 1.4838 1.2435 1.1344 1.0171 1.1167 1.2809 1.1305 1.1508 1.1346 1.1071 1.1361 1.0511 1.5475 1.1659 1.2567 1.1401 1.1011 1.1775 1.1640 1.2084 1.2848 1.1421 1.2991 1.4872 1.1078 1.1459 1.6631 1.6841 1.1984 1.1502 1.1593 1.2530 1.0538 1.1056 LOG YO = .57559 .55859 .40459 .51279 .06079 .31499 .51579 .50839 16099 .49179 .16479 .69039 .77919 -.02061 .37029 .36909 .33084 .31189 • .16809 .37239 .03429 -.27691 .55969 .36809 .37109 .58129 .39259 .32709 .09389 .27979 .51954 .30749 .24449 .45159 .40979 .35629 .23669 .46099 .70439 .57279 .47619 .63419 .46369 Σ = 52.0396 Σ MD 1.2102 MLog Yo σD = .1541 σLog Yo ――― 16.20967 = .3770 .2059 GRADE IV PUPIL #1 泌​淚 ​*2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 14 # 15 #16 #17 #18 #19 #20 #21 * 22 #23 #24 #25 #26 #27 #28 #29 #30 #31 #32 #33 #34 #35 #36 #37 #38 #39 #40 #41 #42 #43 #44 #45 с .4487 .4457 .4859 .4464 .2818 .4615 .4834 .5417 .4897 .2397 .3045 .3814 .4813 .2277 .5338 .2203 .3028 .1622 .5248 .5289 .7842 .5864 .3537 .5884 .5879 .1934 .3714 .6325 .4575 .5037 .2040 .2286 .5346 .5190 .5910 .4898 .6676 .6086 .8860 .6382 .0524 .4332 .9108 .5727 .2391 p Logical Association Tests b .1554 .2803 .1053 .1248 .1060 .2305 .1576 .1189 .4050 .2419 .1098 .2135 .2401 .3161 .1716 .2948 .0890 .1532 .2658 .1777 .0207 .3129 .1412 .2215 .1464 .2181 .2080 .1433 .1226 .0080 .1961 .2336 .2189 .2153 .2509 .3737 .0413 .0442 .1885 .1704 .5142 .8188 .1124 .0926 .1463 Σ = 20.6269 | Σ = 9.1172 Mc .4584 M₂ = .2026 σε .1821 σ b .1353 = .08567 .15537 .05413 .06909 .07613 .12412 .08142 .05449 .20667 .18392 .07637 .13207 .12454 .24412 .08000 .22986 .06205 .12835 .12631 .08371 .00447 .12942 .09126 .09117 .06033 .17592 .13075 .05266 .06651 .00397 .15610 .18020 .10188 .10356 .10262 .19066 .01373 .01730 .02149 .06165 .48726 .46410 .01003 .03957 .11132 • d 28 1.2181 1.4301 1.1327 1.1724 1.1916 1.3308 1.2062 1.1337 1.6094 1.5273 1.1923 1.3554 1.3321 1.7544 1.2027 1.6977 1.1538 1.3438 1.3375 1.2126 1.0103 1.3472 1.2338 1.2336 1.1490 1.4994 1.3513 1.1289 1.1655 1.0092 1.4328 1.5143 1.2644 1.2693 1.2665 1.5512 1.3606 1.0406 1.0507 1.1525 3.0709 2.9114 1.0234 1.0954 1.2922 Ꭰ = = LOG YO .92378 .79888 .97388 .95438 .97318 .84868 .92158 .96028 .67418 .83728 .96938 .86568 .83908 .76308 .90758 .78438 .99018 .92598 .81338 .90148 1.05848 .76628 .93798 .85768 .93278 .86108 .87118 .93588 .95658 1.07118 .88308 .84558 .86028 .86388 .82828 .70548 1.03788 1.03498 Σ 60.9590 Σ MD 1.3546 MLog yo OD = .3938 σLog Yo - .89068 .90878 .56498 .26038 .96678 .98658 .93288 39.44590 .8766 .1353 GRADE IV PUPIL #1 *2 #3 #4 #5 #6 #7 #8 *9 #10 #11 * 12 13 *14 #15 #16 #17 #18 #19 *20 *21 * 22 *23 *24 * 25 * 26 #27 #28 *29 *30 #31 #32 #33 *34 #35 #36 #37 #38 *39 *40 #41 #42 *43 #44 #45 с .4260 .7061 .8375 .9179 .6502 .9128 .6171 .5693 .7125 .5530 .7890 .8536 .9016 .7910 .6793 .7474 .7203 .6393 .6777 .6182 .8196 .7655 .8764 .7984 .5882 .7959 .7630 .8267 .5896 .4992 .5870 .4386 .6161 .6934 .9256 .6895 .6723 .6778 .8779 .8155 .7704 .7040 .8148 .6531 .5675 Σ = Mc Οι 32.1458 | Σ Rote Association Tests .7143 Mo .1225 σε b .5282 .3031 .3742 .3714 .3266 .5646 .4465 .5562 .4329 .2344 .5818 .5796 .2778 .2252 .3819 .6250 .2914 .1629 .4566 .0583 .2252 .4984 .6371 .2245 .5221 .2654 .2635 .5329 .4244 .3086 .3395 .7421 .7163 .5720 .4030 .2911 .3007 .6166 .7712 .6700 .3648 .3063 .2563 .3582 .3373 18.7261 .4161 .1650 Pla d .30319 .08908 .06081 .03049 .11424 .04923 .17096 .23956 .12446 .10478 .12276 .08485 .02734 .04707 .12248 .15787 .08150 .05876 .14716 .02226 .04063 .11687 .07875 .04526 .21500 .05417 .06245 .09235 .17417 .15455 .14021 .41661 .27499 .17538 .02998 .09039 .09854 19867 .09416 .12361 .08376 .09066 .04747 .12426 .14588 29 D 2.0100 1.2276 1.1503 1.0723 1.3009 1.1200 1.4824 1.7360 1.3319 1.2722 1.3267 1.2158 1.0650 1.1145 1.3258 1.4384 1.2064 1.1449 1.4037 1.0286 1.0981 1.3085 1.1988 1.1098 1.6406 1.1328 1.1547 1.2369 1.4934 1.4274 1.3811 2.6098 1.8836 1.4975 1.0712 1.2314 1.2547 1.5800 1.2421 1.3293 1.2127 1.2321 1.1155 1.3313 1.3992 Σ = MD σD = Ꭰ 60.1459 1.3366 .2839 - LOG YO .37489 .59999 .52889 .53169 .57649 .33849 .45659 .34689 .47019 .66869 .32129 .32349 .62529 .67789 .52119 .27809 .61169 .74019 .44649 .84479 .67789 .40469 .26599 .67859 .38099 .63769 .63959 .37019 .47869 .59449 .56359 .16099 .18679 .33109 .50009 .61199 .60239 .28649 .13189 .23309 .53829 .59679 .64679 .54489 .56579 21.91295 Σ = MLog Vo σ Log to .4870 = .1654 PUPIL ****** * 1 2 3 H 10 67 #1 #2 *3 #4 #5 *6 #7 #8 #9 *10 #11 * 12 13 #14 15 16 #17 18 #19 * 20 #21 #22 * 23 * 24 *25 * 26 *27 *28 * 29 *30 #31 #32 #33 #34 #35 *36 #37 #38 *39 #40 #41 * 42 43 #44 45 *46 47 #48 *49 #50 Σ = Mc σε = C .6488 .8057 .8759 .7319 .6436 .4948 .3688 .3763 .7694 .4969 .7358 .3140 .1949 .4370 .7698 .2506 .2833 .4601 .6074 .6884 .5682 .4046 .6035 .3072 .5207 .3328 .8662 .8379 .5090 .4928 .6994 .3533 .6467 .6493 .4379 .1612 .5059 .2862 .3968 .6307 .2870 .4185 .5181 .5169 .6305 .1141 .2183 .0650 .7056 .7485 25.3862 | Σ - GRADE V Logical Association Tests .1414 .0567 .0866 .0616 .0371 .0331 .0973 .4809 .0850 .0547 .1603 .1046 .1354 .0963 .3427 .1603 .2797 .0880 .0323 .0506 .0773 .0444 .0546 .0761 .1803 .0466 .1518 .3103 .2957 .1279 .1442 .1519 .2053 .1157 .1359 .1245 .1167 .2042 .2343 .1425 .2515 .1578 .0754 .3271 .3050 .1735 .1579 .2813 .3118 .2136 .5077 Mb .2034 συ b = 7.7797 - .1556 .0984 .04966 .01102 .01075 .01651 .01322 .01672 .06142 .44964 .02502 .01376 .09998 .02412 .06812 .02544 23509 .12906 .15747 .02026 .02421 .03627 .04173 .01743 .01701 .03286 .10735 .01848 .10517 .14873 .19729 .01711 .02337 .07458 10413 .03478 .08789 .04399 .04093 .11478 .19653 .07041 .17952 .09519 .02785 .23322 .17736 .08361 .07628 .10394 .27622 .16697 · ► d 30 D 1.1211 1.0257 1.0251 1.0388 1.0309 1.0392 1.1512 2.8161 1.0593 1.0322 1.2589 1.0571 1.1698 1.0603 1.7183 1.3460 1.4370 1.0478 1.0573 1.0871 1.1109 1.0410 1.0400 1.0786 1.2804 1.0435 1.2730 1.4086 1.5750 1.0402 1.0553 1.1874 1.2709 1.0834 1.2243 1.1066 1.0988 1.3025 1.5723 1.1760 1.5118 1.2451 1.0662 1.7109 1.5044 1.2123 1.1920 1.2704 1.8890 1.4688 MD Σ 62.6188 1.2524 .3056 OD LOG YO .93778 1.02248 .99258 1.01758 1.04208 1.04608 .98188 .59828 .99418 1.02448 .91888 .97458 .94378 .98288 .73648 .91888 .79948 .99118 1.04688 1.02858 1.00188 1.03478 1.02458 1.00308 .89888 1.03258 .92738 .76888 .78348 .95128 .93498 .92728 .87388 .96348 .94328 .95468 .96248 .87498 .84488 .93668 .82768 .92138 1.00378 .75208 .77418 .90568 .92128 .79788 .76738 .86558 Σ = MLog Yo σ Log Yo 46.17930 .9236 .0983 = } GRADE V PUPIL #1 #2 #3 #4 #5 #6 #7 #8 #9 *10 #11 *12 #13 #14 ***** #15 *16 #17 18 #19 * 20 *21 *22 #23 *24 * 25 *26 #27 #28 #29 #30 #31 *32 #33 *34 #35 *36 *37 #38 #39 *40 #41 * 42 #43 *44 *45 *46 *47 *48 *49 #50 .7448 .7402 .8428 .7971 .5919 .7512 .6410 .7324 .8829 .5623 .9190 .8679 .6683 .6865 .6797 .7027 .8257 .7106 .6548 .6984 .5509 .7496 .4267 .2442 .4170 .8376 .6850 .8502 .7978 .8101 .8645 .7757 .8815 .8262 .8453 .6737 .9296 .7250 .7093 .8778 .5524 .5918 .9305 .7726 .5134 .7341 .7421 .9200 .6549 .6960 σε Σ Ξ 36.2857 | Σ Mc.7257 | M。 .1413 συ D Media Rote Association Tests .2700 .3953 .0571 .2948 .3548 .0883 .1898 .3475 .1587 .1211 .2851 .2783 .2152 .2565 .2986 .2512 .4020 .2588 .1846 .2042 .1911 . 2420 .3009 .3095 .4552 .1249 .5908 .3815 .2635 .4361 .3397 .5528 .3706 .3989 .4993 .7138 .4564 .3585 .3299 .3467 .3239 .1603 .4315 .4323 .3904 .3478 .2489 .4149 .5880 .4538 b = = 16.3658 .3273 .1329 d .06890 .10270 .00898 .05981 .14479 .02197 .06814 .09299 .01858 .05301 .02310 .03676 .07138 .08041 .09564 .07468 .07007 .07490 .06372 .06159 .08582 .06060 .17251 .59359 .26538 .02028 .18610 .05715 .05328 .08282 .04603 .12399 .04392 .06933 .07724 .23291 .03213 .09859 .09590 .04237 .14498 .06544 .02999 .09831 .18997 .09248 .06419 .03320 .20292 .13796 31 D 1.1719 1.2668 1.0209 1.1477 1.3957 1.0519 1.1699 1.2388 1.0437 1.1298 1.0546 1.0883 1.1786 1.2034 1.2463 1.1876 1.1751 1.1883 1.1580 1.1524 1.2185 1.1498 1.4875 3.9227 1.8424 1.0478 1.5350 1.1406 1.1305 1.2101 1.1118 1.3304 1.1064 1.1731 1.1947 1.7096 1.0768 1.2549 1.2471 1.1025 1.3962 1.1626 1.0715 1.2540 1.5488 1.2373 1.1593 1.0794 1.5956 1.3739 = LOG YO .63309 .50779 .84599 .60829 .54829 .81479 .71329 .55559 .74439 .78199 .61799 .62479 .68789 .64659 .60449 .65189 .50109 .64429 .71849 .69889 .71199 .66109 .60219 .59359 .44789 .77819 .31229 .52159 .63959 .46699 .56339 .35029 .53249 .50419 .40379 .18929 .44669 .54459 .57319 .55639 .57919 .74279 .47159 .47079 .51269 .55529 .65419 .48819 .31509 .44929 Σ 64.1405 Σ 28.78870 MD = 1.2828 MLog Yo .5758 .4146 Log Yo .1329 OD = = 32 DEVELOPMENT OF LOGICAL AND ROTE MEMORY known under the name of "span." This "span" is related to the constant d in our equation. Examination of the equation and its process of derivation will show that b, c, and d are dependent in the sense, that given any two, the third may be calculated. But any two of these constants are mathematically independent. It becomes a matter of con- siderable interest to find out whether or not any two are psycho- logically dependent, that is to say, correlated in the population tested. The preceding tables give the children's scores by grades, the calculated constants, their average and standard deviation. THE CURVE OF DEVELOPMENT FOR THE CONSTANTS OF MEMORY IN LOGICAL AND ROTE MATERIAL Average C.A. and average M.A. for each grade C.A. M.A. II 91.93 mos. 94.63 mos. III 109.84 mos. 111.65 mos. IV 119.33 mos. 123.04 mos. V 133.76 mos. 131.90 mos. I. The c values Let us now consider Fig. I, showing how the c values vary for logical memory and rote memory from the second to the fifth grade. Rote c seems to receive a period of acceleration at the third grade, whereas logical c has received its period of acceleration already at the second grade. One cannot, however, be sure whether this is an effect true of the population in general or an accidental variation due to differences between the classes in which the constants were measured. The subjects, however, are the same for the two curves. It seems, therefore, that in a group of subjects in which the power of rote memory was developing along a slowly rising4 plateau, the power of logical memory has already started a period of rapid development. When one compares the curves with each other, one sees that 4 A falling c value means a rising memory value. 5 It should be noted that the difference is much more marked than appears in the figure because the task for logical memory is 12 items and that for rote memory 8 items. See pages 3 and 19. The same consider- ations hold for the following comparisons. STATISTICAL ANALYSIS 33 VALUES the curve for rote c is at a much higher level than that for logical The meaning of this is that the c value for rote material is C. .9500 .9000 MEAN .8500 .8000 .7500 U .7000 I.6500 .6000 .5500 5000. .4500 8472 chig .6762. GRADE 11 .8429 .6106 = FIG. I 7143 .4584 IV .7257 .5077 V LOGICAL ROTE higher than for logical, so that the effect of a single repetition for logical material is distinctly greater than for rote material, 34 DEVELOPMENT OF LOGICAL AND ROTE MEMORY This is true for all grade levels and can scarcely be made an acci- dental result. Fig. II shows the relationship between logical c VALUES ፡፡ .9500 .9000 MEAN .8500 ,.8000 U.7000 .7500 .6500 .6000 .5500 .5000 .4500 OLET.T. 95 100 M.A. LOGICAL C. A. LOGICAL 105 110 115 120 AGE IN MONTHS FIG. II M.A. ROTE- C.A. ROTE 125 130 १ and rote c with mental and chronological ages. Essentially the same features are here noted. STATISTICAL ANALYSIS 35 II. The b values Looking at Fig. III, we see that rote b drops steadily with age. Apparently, however, logical b does not drop sharply until the fourth grade. This might perhaps be due to the fact that logical relations are not so much utilized in earlier years, and consequently the effect on the degree to which the inertia of the nervous system affects the two results is not so evident. Fig. IV which shows the relationship to mental and chron- ological ages tells essentially the same story. One should note the much lower values for the logical constants. This means that not only is the effect of a single repetition much more pronounced in learning logical material, but that the inertia of the nervous system is not involved to such an extent. III. The D values Let us now look at the curve of development of the D values- Figs. V and VI. We notice a marked difference between the curves for logical and rote D. The curve for Logical D drops con- sistently from the second to the fifth grade. The natural inter- pretation of this fact is that the child is making more and more use of logical relationships as he grows older. If logical memory alone were involved in learning the logical material, D would ap- proach zero and a single repetition would suffice for permanent retention. (Log y。 = a a — b; .. Log y。 = a, if d 0.) When, however, we look at the development of rote D, we find that it shows no consistent tendency to rise or fall. Except for accidental variations, it remains at about the same level. It looks as if rote D represents an original endowment which is not subject to much change from the second to the fifth grade of school life, or that its period of development lies in the years prior to those of our study. INTERCORRELATIONS OF CONSTANTS Let us now consider the following table of correlations. We give also a table of Log yo for logical and rote memory as a kind of check in the correlation. As pointed out, these correlations 36 DEVELOPMENT OF LOGICAL AND ROTE MEMORY Logical c Logical b Logical D Rote c Rote b Rote D Logical c Logical b Logical D Rote c Rote b Rote D Grade II III IV V -.049 .286 M II — .526 .099 .378 II .526 .313 .168 .062 .634 Logical c .049 .241 277 .573 M LOGICAL C III .245 .585 .246 .493 .066 III ROTE C .246 .286 .177 .165 .769 Mag V IV ― Intercorrelations of the Constants, c, b and D .278 .493 .241 .156 .102 IV .241 .104 .017 .026 - .715 Logical b Logical D ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ - .999 991 -1.000 -1.000 S .835 .420 .872 .893 .572 -.706 Ma ― - .208 .163 .196 LOGICAL LOG Yo V .208 .077 .134 .062 -.691 Rote c Ma .314 .271 .101 .077 M II -.049 .836 .313 .335 .135

«> 100 105 110 115 120 125 AGE IN MONTHS Fia. IV M.A.ROTE- C. A. ROTE 130 | left quadrants give intercorrelations of rote memory with logical memory constants. If logical memory is one thing and rote mem- STATISTICAL ANALYSIS 39 • ory is another thing, the average of the correlations in the upper left and lower right should be higher than that in the upper right " VALUES 1.7000 ~ 1.6000 MEAM 1.65001 1.5500 1.5000 1.4500 1.4000 1.35 1.3500 1.3000 1.2500 1.2000 GRADE 1.5749 1.2520 11 1.4763 1.3366 1.2102 111 FIG. I 1.3546 IV 1.2828 1 1.2524 V LOGICAL ROTE and lower left. Let us average these and test the significance of the difference. 40 DEVELOPMENT OF LOGICAL AND ROTE MEMORY Letting the upper left and the lower right be represented by X, and the upper right and lower left, by Y, a significant difference " LUES л ~ 1.6000 1.7000 MEAN 1.6500 - 1.5500 1.4500 1.4000 1.3500 1.3000 1.2500a 1.2000 1.1500D 95 M.A. LOGICAL C.A. LOGICAL 100 105 110 115 AGE IN MONTHS FIG. II 120 M.A. ROTE C.A. ROTE 125 QPAD GOSPODU ALADDINES 130 P is found, as shown in the following results. We are thus com- paring the average of quadrants I + IV with quadrants II + III. STATISTICAL ANALYSIS 41 M₂ = .467 Mx My €.188 D = .279 — Mx My D .520 = .415 ****** = .105 CR. = σ diff. 0 = .274 .131 σy Ø diff. || = = .0611 D .279 O diff. .0611 Vo² mx + o² mv √(.057)² + (.022)² √.003733 One might also compare the raw correlations for the logical constants with those of the rote constants, that is to say, quadrant I with quadrant IV. Averaging the two and taking their differ- ence, we find the critical ratio to be less than 1.00-therefore in- significant. .262 στ = 0 = .276 สน 4.566 2 √o² m² + o²my = √(.083)²+(.087)2 V.014458 = .1202 D .105 O diff. .120 Om x Omy .875 Om x O my = €.057 .022 = = .083 .087 In this last comparison, we are considering the average of a number of intercorrelations of one function with those of another. There is no particular reason why they should differ significantly. But when, as in the first case, we compare the average of a num- ber of correlations obtained by measures of identical functions with the average of cross correlations of non-identical functions, we should be comparing a higher average with a lower and so get a significant difference in the averages. This is what we have found; and the results cannot be explained without recognizing a fundamental difference between rote and logical memory. CHAPTER V SUMMARY This study was begun with the intention of tracing the develop- ment of logical and of rote memory. The nature and subject matter of the devised tests of memory ability restricted the pro- gram of the experiment to the second, third, fourth and fifth grades. The preliminary testing group included children of the first and sixth grades; but such factors as handwriting and a limited knowledge of number concepts eliminated the first grade pupils as subjects, whereas the relative ease of the logical memory tests made the exclusion of the sixth grade children advisable. Each pupil's raw score (his total correct responses), together with the number of items to be memorized, furnished a basis for the mathematical estimation of the constants in the learning curve derived from the equation Log yn bcx wherein a is merely the task, the amount to be learned, the total number of associations to be made; b, which varies with the task, is the resistance of the nervous system that must be over- come in the process of learning; and c determines the quickness of the learner. Its exponent "x" refers to the trial. b involves d, whose antilogarithm D is a measure of the inertia of the nervous system in learning. Log yn is the amount learned after each repetition. Plotting the constants demonstrated the following data found: The curve for rote c is at a higher level than for logical c for all grades. There is a corresponding gradation for mental and chron- ological ages. Therefore, the effect of a single repetition for logical subject matter is distinctly greater than for rote material. b varies from individual to individual; but the average b becomes lower for all grades, with a similar descent for mental and chron- ological ages. If, as we have seen, rote d shows no tendency to fall consistently with age and school grade, the fall in rote b would be due to a fall in c, for b and c<1.00. d 1-c 42 а - SUMMARY 43 Logical D shows more of a change than does rote D. The latter constant appears to remain relatively fixed from the time the child has reached the second grade. The probable meaning of the steady drop of the logical constant from grade to grade and from year to year may be attributed to the fact that the child, with in- creasing age, facilitates his learning by more and more use of logical relationships. A study of the intercorrelations of the tests shows that there is a significant difference between the average of the intercorrelations of the logical memory constants plus those of the rote memory constants and the average of the cross correlations between rote and logical memory constants. This shows that one function is tested by the logical and another by the rote memory tests. BIBLIOGRAPHY BALLARD, P. B., "Obliviscence and Reminiscence," British Journal of Psychology, Monograph Supplement, No. 2. Cambridge: 1913. viii + 82 pp. BARLOW, M. C., "The Rôle of Articulation in Memorizing," Journal of Experimental Psychology, 11, 1928, 306–312. BOLTON, T. L., "Growth of Memory in School Children," American Journal of Psychology, 4, 1909, 362–380. BROTEMARKLE, R. A., "Some Memory Span Problems," Psychological Clinic, 15, 1924, 220-258. BRYAN, ALICE ISABEL, "Organization of Memory in Young Children," Archives of Psychology, No. 162, May, 1934, 56 pp. BURTT, H. E., "An Experimental Study," Journal of Genetic Psychology, 40, 1932, 287–295. "A Further Study of Early Childhood Memory," Journal of Genetic Psychology, 50, 1937, 187-192. Carlson, HILDING B., "Factor Analysis of Memory Ability," Journal of Experimental Psychology, 21, 1937, 477–492. CARLSON, HILDING B. AND CARR, HARVEY A., "Visual and Vocal Recog- nition Memory," Journal of Experimental Psychology, 23, 1938, 523–530. "Rote and Logical Recognition Memory," Journal of Experi- mental Psychology, 26, 1940, 199–210. DIETZE, A. G., "The Relation of Several Factors to Factual Memory," Journal of Applied Psychology, 15, 1931, 563–574. DIETZE, A. G. and Jones, G. E., "Factual Memory of Secondary School Pupils for a Short Article Which They Read a Single Time," Journal of Educational Psychology, 22, 1931, 586-598, 667–676. ENGLISH, H. B., WELBORN, E. L., AND KILLIAN, C. D., “Studies in Sub- stance Memorization," Journal of General Psychology, 11, 1934, 233–260. FORREST, SISTER HELEN DE SALES, "The Correlation between the Con- stants in Association and Forgetting." (Dissertation to be published shortly, Department of Psychology, Catholic University of America.) FOSTER, J. C., "Verbal Memory in Preschool Children," Journal of Genetic Psychology, 35, 1928, 26-44. Gates, Arthur I., A Reading Vocabulary for the Primary Grades. Revised and enlarged; New York: Bureau of Publications, Teachers' College, Columbia University, 1935. 29 pp. Hartmann, GEORGE W. AND SCHOOLEY, MARY, "The Rôle of Insight in the Learning of Logical Relations," American Journal of Psychology, 49, 1937, 287–292. HOLADAY, PERRY E. AND STODDARD, GEORGE D., "Getting Ideas from the Movies." New York: Macmillan Co., 1933, 102 pp. (Published with: Charters, W. W., Motion Pictures and Youth: A Summary.) 44 BIBLIOGRAPHY 45 HSIAO, H. H., "A Preliminary Study of the Formal Development of Mem- ory," Journal of Testing (Chinese), 3, 1933, 55–75. Hunter, WalTER S., "Experimental Studies in Learning,” The Founda- tions of Experimental Psychology. Worcester, Mass.: Clark University Press, 1929. Chapter 15, pp. 564-627. MALLAY, HELENA, "The Latent Memory Span of the Preschool Child," Child Development, 6, 1935, 110-119. MCGEOCH, GRACE O., "The Age Factor in Reminiscence: A Comparative Study of Preschool Children and College Students," Pedagogical Seminary and Journal of Genetic Psychology, 47, 1935, 98-120. "The Factor of Degree of Learning in Reminiscence," (Abstract) Psychological Bulletin, 31, 1934, p. 599. MCMANAMA, SISTER MAURICE, "A Genetic Study of the Cognitive General Factor in Human Intelligence," Studies in Psychology and Psychiatry, No. 2, 1936, 35 pp. MEEK, LOIS H. AND JERSILD, ARTHUR T., "Mental Development from Two to Twelve Years," (Chapter ii of "Mental and Physical Development"), Review of Educational Research, 6, 1936, 17–49. MICHOTTE, A. E. AND RANSY, C., "Contribution à L'Étude de la Mémoire Logique," Travail du Laboratoire de Psychologie Experimentale de L'Université de Louvain. Louvain, 1912. 95 pp. MICHOTTE, A. E., ET PORTYCH, TH., "Deuxième Étude sur la Mémoire Logique," Extrait des Annals de l'Institut Supérieur de Philosophie. Louvain, 2, 1913. 533-657. 1 MONAGHAN, EDWARD A., "Major Factors in Cognition," Studies in Psy- chology and Psychiatry, No. 5, 1935, 48 pp. MOORE, THOMAS VERNER, "The Analysis of Association by Its Equational Constants," Aspects of the New Scholastic Philosophy. New York: Benziger Brothers, 1932, 181-225. Cognitive Psychology. New York: Lippincott Company, 1939. Part VI, Chapters i-ix, pp. 405-525. NORTHWAY, MARY L., "Influence of Age and Social Group on Children's Remembering,” British Journal of Psychology, 27, 1936, 11-29. REED, H. B., "Meaning as a Factor in Learning," Journal of Educational Psychology, 29, 1938, 419–431. WELBORN, E. L. AND ENGLISH, HORACE, "Logical Learning and Retention: A General Review of Experiments with Meaningful Verbal Materials," Psychological Bulletin, 34, 1937, 1–20. WHIPPLE, G. M., Manual of Mental and Physical Tests. (Sec. Ed.) Balti- more: Warwick and York, 1915. Part 2, pp. 205-223. ZANKOV, L. V., "The Development of Memory in Morons: I. Experimental Investigations," Journal of General Psychology, 16, 1937, 415–426.