B 1059 M3 opy 2 fourattottal gai|rl|oIogi^ iJottograyl^B No. 15 THE GENERAL VALUE OF VISUALSENSE TRAINING IN CHILDREN BY Chang Ping Wang BALTIMORE WARWICK & YORK, Inc. 1916 Monogfupb THE GENERAL VALUE OF VISUAL SENSE TRAINING IN CHILDREN A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the University of Michigan lE&mraltottal }pgt;rhoUigi| ilottograyl|g E&itPb by (Sua Manttaat lll]iyple No. 15 THE GENERAL VALUE OF VISUAL SENSE TRAINING IN CHILDREN BY Chang Ping Wang BALTIMORE WARWICK & YORK, Inc. 1916 OcrW\ ^ h Copyright, 1916. by Warwick & York, Inc. Gifi The Uui^orBity o o EDITOR'S PREFACE Contributions to the experimental study of the trans- fer of training (formal discipline) scarcely need either apology or introduction in a period when, despite the considerable amount of investigation, so very much still remains undetermined with respect to the amount of such transfer and the mechanism by means of which it takes place. The special features of this contribution by Dr. Wang, a Chinese government student at the Univer- sity of Michigan, lie in the use of school children as subjects and in the use of sense-training as the medium of experimentation. In this latter aspect his study will be particularly welcome from the light it throws upon the issue of sense-training, which is almost a fetish of the adherents of the, at present, so popular Montessori method. G. M. Whipple. TABLE OF CONTENTS Introduction 1 Analytical review of previous experiments 5 The individual method 6 The one-group method 7 The two-group method S The three-group method U New experimental data 13 Experiment 1 Aim 13 Subjects 14 The trained function 16 Results of training 20 Summary 26 The tested functions 26 Discrimination of pitch 28 Summary 39 Discrimination of shades of color 39 Summary 45 Discrimination of size 46 Summarj' 58 Experiment 2 Description of experiment 59 Summary 67 Experiment 3 Description of experiment 68 Summary 74 Conclusions 75 Bibliography • 81 vii INTRODUCTION There are two types of disciplinists. There are those of the old type who beheve that the mental power developed by the training of one function will benefit equally all other functions. This type is repre- sented by the man who claims that any kind of study, no matter what it may be, will prepare for life, so long as that study is done well. The later type of dis- ciplinist is less sweeping in his claims. He believes that the training of a specific function, such as memor- izing poetry, will benefit all other kinds of memories, the general function of memory. This type is repre- sented by the man who advocates studies in schools for the development of the various mental functions: arithmetic develops the power of accuracy; Latin, the power of analysis; sense education, the power of ob- servation. The old type of disciplinist is scarcely to be found among educators of today, but the later type still dominates, in certain respects, the educa- tional world. Many writers have already pointed out instances showing how some of the most prom- inent educators, both in Europe and America, have overestimated the importance of this type of dis- cipline. Thorndike, Ruediger, Tracker, Winch and others have not only attacked it, but have also demon- strated by experiments the limitations of certain specific functions. These functions are selected from those general functions which we designate as mem- ory, discrimination, or reasoning. Ttie results show that there are manj- kinds of memories, discrimina- tions and reasonings. One kind of memory, after 2 VISUAL SENSE TRAINING IN CHILDREN much training, may not affect certain other memories at all. The same has been found true in other fields of mental functioning. It was the purpose of the writer to continue the ex- perimentation within the field of discrimination and, if possible, to contribute something to our knowledge of the extent, amount, and means of generalization, or transfer, upon which points there is as yet no general agreement. By extent of transfer is here meant the range of influence which the training of a specific func- tion has over untrained functions; and by amount of transfer, the aggregate of improvement in any specific untrained function as a result of the training. With regard to the extent of transfer Thorndike says: "Improvement in any single mental function rarely brings about equal improvement in any other function, no matter how similar, for the working of every mental function-group is conditioned by the nature of the data in each particular case. "^ But Fracker denies this by saying : ' ' Improvement in many cases is absolutely greater in amount in the test than in the training. "2 It is evident that this question of the extent of transfer needs further investigation. With regard to the amount of transfer each experi- ment has shown a different result. This is due partly to different methods of calculation and partly to the testing of different functions. In some experiments adults act as subjects and in others children. In his calculations one experimenter uses points to denote the amount of transfer while others use per cent. Some writers take the improvement of the trained ^Thorndike, E. L., Educational Psychology, 1903, p. 91. ^Fracker, G. C, Psychological Review Monograph Supplement, Vol. 9, No. 2, p. 99. INTRODUCTION 6 function as the basis of calculation; and others, the record made in the preliminary test. Naturally the results are not uniform. There is a special need just now for standardization in our methods of experimen- tation as well as in the methods of calculating our results. With regard to the means of transfer different writers seem to have reached different conclusions. To quote Colvin: ''The question whether the results are due to functioning of identical elements (Thorndike) ; to improvement of habitual methods of recording facts (James) ; to training the attention and will power (Scripture and Davis) ; to divesting the essential process of the unessential factors, greater habituation and more economical adaptation of attention (Coover and Angell) ; to the effective use of mental imagery and properly controlled attention (Fracker); to the devel- opment of ideals (Bagley, Ruediger, and Ruger); to general miprovement in technique of learning, atten- tion and will-power, but chiefly to a sympathetic inter- action of allied memory functions (Ebert and Meumann) , or to all of these, or to some other factors as yet not analyzed out, will doubtless for a long time offer a fruitful field of inquiry. "^ The present experiments were undertaken with these different questions in mind. Reaction time is taken throughout to detect temporal differences, which play an important part in sense discrimination of all kinds. Children are employed as subjects, because there is a suspicion among educational psychologists that pos- sibly specific training has a greater value for them than for adults, on account of the faster rate of physical and mental development in childhood. This suspicion 'Colvin, S. S., The Learning Process, pp. 241-2. 4 VISUAL SENSE TRAINING IN CHILDREN is voiced by Foster who, writing on the effect of practice upon visualizing and upon the reproduction of visual impressions, concludes his article by remarking: ''Speci- fic practice is demanded for best results and becomes quickly effective. It seems, therefore, as if the value of formal training of our kind has been overestimated. However, our experiments were made upon adults who were already trained in habits of attentive obser- vation, and we have no right or wish to extend our con- clusions in wholesale fashion. It may be that, for immature and untrained persons, practice in visual reproduction might possess a general value that was not discovered under our conditions."^ The writer is indebted to Dr. J. F. Shepard, Dr. F. S. Breed, Professor W. B. Pillsbury and Professor A. S. Whitney for their suggestions, supervision and encouragement throughout the work. He wishes to thank the Board of Education of Ann Arbor, Michi- gan, for giving permission to carry on the work and especially Miss C. L. Dicken, Principal of the W. S. Perry School, and the parents who kindly consented to let their children act as subjects. * Foster, W. S., The Effect of Practice upon Visualizing and upon the Reproduction of Visual Impressions, Jour. Educational Psychol- ogy, Vol. 2, p. 21. ANALYTICAL REVIEW OF PREVIOUS EXPERIMENTS The origin of the doctrine of formal disciphne dates as far back as Plato, who wrote: ''and have you further remarked that those who have the natural talent for calculation are generally quick at every other kind of knowledge; and even the dull if they have an arith- metical training gain in quickness if not in any other way."* But it was during the time of the scholastics and humanists that this doctrine reached its zenith. The scholastics regarded the mind as a logical machine for the purpose of grinding out cut-and-dried truth, while the humanists claimed that ancient languages could furnish all the mental nourishment and power necessary for life. The modern criticism of the doc- trine was launched by the German Herbartians, who maintained ''that since all mental exercise takes its rise in a definite mental content, its character is neces- sarily determined by its origin."* To these people was due the credit for having started the movement toward experimental investigation. For a chronological review of the experiments which have been performed the reader is referred to Heck's Mental Discipline and Educational Values; Thorndike's Educatio7ial Psychology, Vol. 2; Bagley's Educative Process, or Colvin's The Learning Process. It would be well for us to take this opportunity to review the experimental methods which seem to demand descrip- tion. Since people have begun to test the doctrine ^ Plato, Republic, Book 7. ^ Ruediger, W. C., Principles of Education, p. 96. 6 VISUAL SENSE TRAINING IN CHILDREN of formal discipline by controlled experiments, there has been gradually built up a regular method for con- ducting such experiments. From the work of William James, who performed the first controlled test of the spreading influence of one specially trained memory function upon memory functions of a different content, to the present day, there have been four methods used, each an improvement over its precedessor. The very first used may be called The Individual Method To William James should be given the credit for first using this method, in which only one person acts as subject, usually the experimenter himself. In James' own words: "In order to test the opinion so confidently expressed in the text, I have tried to see whether a certain amount of daily training in learn- ing poetry by heart will shorten the time it takes to learn an entirely different kind of poetry. During eight successive days I learned 158 lines of Victor Hugo's 'Satyr.' The total number of minutes required for this was 131^/6 — it should be said that I had learned nothing by heart for many years. I then, working for twenty-odd minutes daily, learned the entire first book of Paradise Lost, occupying 38 days in the process. After this training I went back to Victor Hugo's poem, and found that 158 additional lines (divided exactly as on the former occasion) took me 151 H minutes. In other words, I committed my Victor Hugo to memory before the training at the rate of a line in 50 seconds, after the training at the rate of a line in 57 seconds, just the opposite result from that which the popular view would lead one to expect. But as I was perceptibly fagged with other work at the time of the second batch of Victor Hugo, I thought that might explain the retardation; so I persuaded several other persons to repeat the test."'^ This method of testing the doctrine of formal dis- cipline has the advantage of direct introspection, es- pecially when the experimenter is a good psychologist. On the other hand, we feel uncertain whether the results obtained will represent the average experience of the ' James, Wm., The Principles of Psychology, Vol. I, pp. 666-667. REVIEW OF PREVIOUS EXPERIMENTS mass. We hesitate to apply such individual conclu- sions to every-day use, since researches have proved that mental traits vary as much as physical traits. For example, the following table taken from Thorndike,* showing the ability of 4th-grade girls in thinking of opposites of words, illustrates the curve of distribution in one mental trait, and shows clearly how wide a range of ability there is in that trait. Score Made in Test with Opposites —9 to —5 —4 " " 4 5 " 9 10 " 14 15 " 19 20 " 24 25 " 29 30 " 34 35 " 39 40 " 44 bv Number of Children 3 girls 10 33 36 29 16 11 4 3 To eliminate the individual variations, then, a number of subjects are employed, and this may be called the One-Group Method One of the first to use this method was Dr. E. L. Thorndike, who has contributed so much to the study of individual differences. He describes his experi- ment briefly as follows: "Individuals practiced estimating the areas of rectangles from 10 to 100 sq. cm. in size until a very marked improvement was attained. Tiie improvement in accuracy for areas of the same size but of differ- ent shape due to this training was only 44 per cent as great as that for areas of the same shape and size. For areas of the same shape, but from 140-300 sq. cm. in size, the improvement was 30 per cent as great. For areas of different shape and from 140-400 sq. cm. in size, the improvement was 52 per cent as great. "^ * Thorndike, E. L., Principles of Teaching, p. 74. 5 Thorndike, E. L., Educational Psychology, Vol. II, p. 397. 8 VISUAL SENSE TRAINING IN CHILDREN The result stated in this experiment was obtained by averaging the totals of the practiced individuals. In this way individual peculiarities are in a measure eliminated. Wherever a large number of subjects are available, the result will be safer. Though this was an advance in the technique of experimentation, there was still room for improvement. All Thorndike's subjects were given the various tests ("areas of the same size but of different shape," "areas of the same shape but from 140-300 sq. cm. in size," and "areas of different shape and from 140-400 sq. cm. in size") before as well as after the training ("areas of rectangles from 10 to 100 sq. cm. in size") took place. The first or preliminary test really amounts to a training in itself. If this is the case, there is no way to tell how much of the 44, 30 and 52 per cent improvements, respectively, were due to the first test and how much to the training received from the prac- tice in estimating areas of rectangles from 10 to 100 sq. cm. in size. To separate these two factors, im- provement due to practice and improvement due to the preliminary test, educators have since introduced the Two-Group Method The two-group method means, as the name miphes, that the subjects to be tested are evenly divided, ac- cording to physical and mental development, into two groups. While both groups are tested, first before' and then after the practice or training, only one of the groups receives the training. There are several advantages to this method of procedure. By the use of the untrained group the possible effects of the pre- liminary test, maturation, and incubation are not con- fused with transfer effects. By a comparison of the REVIEW OF PREVIOUS EXPERIMENTS 9 two groups, trained and untrained, the practice effect, if there be any, of the preliminary test of the trained group can not be mistaken for transfer, since on the average this practice will be the same for the two groups. In a similar way the two-group method elimi- nates the possible confusion of transfer and maturation effects. Manifestly, on the average the two groups will mature equally. Psychologists who are familial with the Binet tests can show us in children mental growth of two or even three years during one year of physical maturation. If experiments are performed with children as subjects, there is always this factor of maturation involved, especially when several months are allowed to pass between the preliminary and the final tests. There is possibly another factor that the two-group method checks off, namely, the incubation tendency; that is, the tendency for a function to improve in efficiency during a period of disuse. This tendency is best illustrated by Swift's ball-tossing experiment. He had five adult subjects work at ''keeping two balls going [in the air] with one hand, receiving and throwing one while the other is in the air.. ..The balls used were of solid rubber and weighed 122.6 and 130.2 grams Their diameters were 42 and 44 mm., respectively . . .... .The daily program consisted of ten trials, the subject in each case continuing the throwing until he failed to catch one or both of the balls, "i" Summarizing the results of two of the subjects who took the trouble to test the incubation tendency, Thorndike writes, ''Subject H, having begun with a score of about 4, and having reached, in the last six 1" Swift, E. J., Studies in the Psychology and Physiology' of Learn- ing, Am. Jour. Psych., Vol. 14, p. 201. 10 VISUAL SENSE TKAINING IN CHILDEEN days of forty-two days of practice, average scores of 50, 82, 92, 88, 68 and 105, was retested every thirty days for five months, and attained average scores of 70, 80, 140, 110, and 120. Being then tested after four hundred and eighty-one days, he attained an average score of 119. Being then tested after over four years, he attained an average score of 5; on the following day, one of 10; and on the successive follow- ing days, average scores of 18, 20, 26, 35, 66, 60, 45, 100 and 160. Subject E, having begun with a score of about 10, and having reached, in the last six days of fourteen days of practice, average scores of 31, 53, 80, 105, 115 and 127, was retested every thirty days for five months, ^^ and attained average scores of 115, 145, 155, 230 and 325. Being next tested after an interval of 463 days, he attained an average score of 152. "12 It is possibly true that 'Hhe disuse of a mental function weakens it, and the amount of weakening increases, the longer the lack of exercise, "^^ but at the time while the learning remains fresh, as is the case with the two subjects cited above who showed a decided improvement in their five months' re-test, the incubation tendency was still going on. Usually an experiment of this kind does not last more than a few months. But if the incubation tendency is involved the two-group method eliminates it. It is clear that the two-group method has been a great help in experiments of this kind. Though the intention has been to check off the influence due to the preliminary test, it eventually eliminates the influ- ence of maturation as well as the incubation tendency. ""There was some practice with the left hand during the first thirty days interyal in the case of both H and E." i^Thorndike, E. L., Educational Psychology, Vol. II, pp. 309-310. i3/6id., p. 300. REVIEW OF PREVIOUS EXPERIMENTS 11 Three-Group Method "WTiile carrying on the present experiment the writer has found it expedient to make use of an additional group. The object of this was to determine the amount of difference due to changed conditions of the weather, from winter to spring; of the experimentation room, from the basement to the first floor; and of the daily program of studies (all the subjects had been promoted in their grades at the time of the final test). If an experiment is carried on for months, as was the present one, a change of weather and of the daily program of studies can not be helped. These changes may prove to be more or less favorable to the final test, and hence both the trained and the untrained groups may test out better or worse than under the conditions of the prehminary test. A third group, taking neither train- ing nor the preliminary test, but simply the final test together with the other two groups, will show just how much difference there is at the final test, and whether the trained and the untrained groups have improved or deteriorated. In case the third group yields a better result than did the two other groups at their preliminary test, it will indicate that the conditions of the final test are more favorable; if a worse result, that the conditions at the final test are less favorable; and if the same result, that will confirm whatever difference there may be between the trained and the untrained groups. The three-group method will prove useful for another reason. When poetry and prose are used to test memory, or when arithmetical problems are used to test reasoning, or when the marking of letters and figures are used to test discrimination, it is almost im- 12 VISUAL SENSE TRAINING IN CHILDREN possible to find materials of equal difficulty for both the prehminary and the final tests. Consequently, there have been in the past experiments, cases where both trained and untrained groups have shown im- provements and other cases where both have shown a retrogression on account of the material chosen for the final test. If the material for the final test has been easier than that used in the preliminary, both trained and untrained groups have shown improvement and if the material chosen for the final test has been more difficult, both groups have shown a decrease in efficiency. Just how much easier or more difficult the material chosen for the final test has been than that of the preliminary, a third group would have shown. For these reasons the writer believes that future experimenters may find the additional group helpful. NEW EXPERIMENTAL DATA Experiment 1 Aim As has been stated in the introduction, since there is yet considerable disagreement as to the means, extent and amount that a specifically trained function may affect other functions, it became the purpose of this study to throw some light on these various points by using new data. A word should be added to show why sense discrimi- nation was chosen for experiment. It was thought that work with sense discrimination can be made very simple; that it is easy, too, for use with the children. Moreover, through the influence of Montessori there has been of late somewhat of a revival of interest in sense training. Her visit to America in 1913 aroused much interest, and her book. The Montessori Method, has gone through several editions. Her explanation of her theory will prove enlightening: ''We cannot create observers by saying 'observe,' but by giving them the power and the means for this observation, and these means are procured through education of the senses. Once w^e have aroused such activity, auto-education is assured, for refined, well-trained senses lead us to a closer observation of the environ- ment, and this, with its infinite variety, attracts the attention and continues the psychosensory educa- tion. "^^ Since our experiment includes a few tests on forms, it will be of interest to cite her opinion on this matter "Montessori, M., The Montessori Method, p. 228. 14 VISUAL SENSE TRAINING IN CHILDREN also. She says: ''He [the child] will, however, see the plane geometric forms perfectly represented in the windows and doors, and in the faces of many solid objects in use at home. Thus the knowledge of the forms given him in plane geometric insets will be for him a species of magic key, opening the external world, and making him feel that he knows its secrets. "^^ Montessori does not make clear how many doors of the external world this magic key opens, nor how wide, nor just in what way they are opened. If it is true that well trained senses will invariably lead children to closer observation, it will probably mean great changes in the daily program of the elementary schools. It is but just that educational reforms, discoveries, or any new movements should be recognized and promoted, once they are found to be worthy. On the other hand, if they are not confirmed by properly controlled tests, students of education owe it to the public to expose their fallacy. This explains the pur- pose of using sense training in our experiment. Subjects When the work was started in October, 1913, twenty- two children took the preliminary test, and eleven of these were retained for the training. Experiments were conducted only on the days when there was school. These twenty-two children represented ap- proximately four grades of mentality, according to the teachers' opinions and the Binet tests taken in two successive years, 1911 and 1912. It was thought that possibly bright children might be benefited more than dull ones by a special training and thus would 15 Montessori, M., Ihid., p. 239. NEW EXPERIMENTAL DATA 15 show a larger amount of transfer, so pupils of four different grades of mentality (excellent, good, average and poor) were chosen to test this possibility. Un- fortunately, the two lower grades (average and poor) were unable to continue in the experiment to the end because the experimental work was thought to inter- fere with their school work. Those who were left, therefore, rank above the average pupils of their age. TABLE 1 Number of Subjects, their Age in Jan., 1914, their School Grades, the Teachers' Estimate of their School Work, and their Mental Growth from 1911 to 1912 according to the Binet Tests'-^ Tr.vined Group Subjects Age in Grade in Teachers' Mental growth Jan., 1914 1914 estimate 1911-1912 I 9.1 4B Excellent 1.2 II 9.4 3B Good III 11.1 4B (( .8 IV 10.0 3A u 1.6 V 12.7 5B it 1.4 VI 9.3 3A l< 1.0 VII 11.3 4 A " .4 Average 10.4 1 1.07 Untrained Group VIII 9.5 4B Excellent .6 IX 9.9 4A Good 1.2 X 10.3 3 A (( 1.0 XI 9.4 4 A II 1.4 XII 11.3 3 A II .4 XIII 9.4 4B 11 1.2 Average 9.97 .97 On account of the dropping out of the less gifted children (four out of the trained group and five out 1^ After C. S. Berry, University of Michigan. Unpublished study. 16 VISUAL SENSE TEAINING IN CHILDREN of the untrained group) the balance of these two groups is sUghtly in favor of the trained group, especially in age and rate of mental growth. The difference is, however, not large enough to make the two groups non-comparable. The Trained Fu7iction The trained function was the discrimination of different lengths of vertical lines drawn on cards. There were eleven of these cards, each 3 x Q}/2 inches. On each card only one line was drawn, always in the center and parallel to the short side. The longest line, on the card a, was 1% inches, Vao of an inch longer than the next line on card b. The line on card b was Vso of an inch longer than the line on card c, and so on to card k, which had a line of 1}4 inches. All lines were drawn of a uniform width of .2 mm. Figure 1 shows the apparatus used in training. The table A-B, 36 inches long, 24 inches wide and 25 inches high, was enclosed during the experiment by a square topped canopy as large as the table itself. This canopy was made 23^2 feet above the table, and sup- ported by a light wooden framework nailed to the edges of the table. In order to get the photograph, the framework and canopy were both removed. The subject sat in chair C at the end of the table and the operator in chair D at the side of the table to his left. Two cards w^ere shown to the subject in succession, and the subject gave his judgment in terms of the second card shown saying ''longer" or "shorter" (than the line on the first card). The first card, 1 in Figure 1, was exposed to the subject for three seconds, the opera- tor counting to himself 1-2-3, and then by a pull on the string (6), card 1 was pulled out of sight of the NEW EXPERIMENTAL DATA 17 Figure 1 AB Table for experiment C Chair of subject D Chair of operator 1 First card shown 2 Second card shown 4 Cover for card last shown 5 CatteU FaU. 6 String that pulls card 1 out of sight as cover 4 comes down,, thus showing card 2 7 Lip-key 8 Bergstrom Chronoscope 10 Electric commutator 11 Paper for record 12 String to let fall cover 4 18 VISUAL SENSE TRAINING IN CHILDREN Figure 2 NEW EXPERIMENTAL DATA 19 subject and card 2 appeared (see Figure 2). It will be noticed that card 2, now in view, was hidden behind the falling plate (4) of the Cattell Fall (5) while card 1 was first shown. The falling of the plate (4) showed card 2 and, at the same time, pulled card 1 out of sight by means of a string (6) . While looking at these cards, the subject took hold of the lip-key (7) with his teeth. The lip-key was connected with the Cattell Fall (5) and the Bergstrom Chronoscope (8) with electric wires. As soon as the plate (4) came down, that is, as soon as the second card was shown, the chronoscope was automatically set going, and was stopped again whenever the subject released the lip- key in giving his judgment. Thus, the time spent by the subject in judging whether the second line was longer or shorter than the first was recorded by the chronoscope. To be sure that a uniform amount of light was thrown upon the cards, two tungsten lights, 40 watts each, were hung inside the canopy, six inches from the top and eight inches from the end of the table where the subject sat. A large cardboard, tied close to the lights, kept the lights from shining into the eyes of the subject. The light from outside was entirely ex- cluded. The lip-key was always placed six inches from the edge of the table and the Cattell Fall two feet from this. The difference between the two lines shown for judgment was invariably V:!u of an inch. Each day a subject made twenty judgments, there being ten cases where the second line was shorter and ten cases where the second line was longer. The order of the different pairs to be judged was changed every day by the shuffling of a set of slips on which the different pairs were written. The following is a sample training record. 20 VISUAL SENSE TRAINING IN CHILDREN TABLE 2 Sample of a Record During the Training Feb. 4, 8:30 A. M., Subject I 1. Card e compared with card d 1230 sigmas Judgment WTong 2. ' ' h U g 1270 " right 3. ' ' c ' U ^J 720 " II i< 4. ' ' g " f 930 " 11 a 5. ' ' b ' " C 635 " " " 6. ' ' g " h 2000 " " wrong 7. ' ' i ' " h 910 " right 8. ' , f " g 1280 " U 11 9. ' ' d ' " e 665 " II (1 10. ' ' i ' ' " J 1200 " II ir 11. ' ' f ' " e 1000 " II bC bC bc bo c5 Cj c3 o3 ^ ^ rH l_t CD OJ OJ (P > >■ > > C3 c3 c3 c3 3CC iCOOt^OcCt^OCO _j u " «D(M rt (N Tt<(M05 fe °.s K- cS S M SP == — a CD ~ S ^ £■ o-icooai^oco'— I t^C003C3COTt<(M05 CDt^aiOOiMOlCO OiOt^»OCO(M(NCO S:. £2 2^'^'«'oc5^ a 3 ;2 ^ bo -^ !- ^ COOOcOcDOOt-h cooococoocDro o S paAiaaaj loafqns ot^oiioiooofocoo fOTtifOTjfTjirocoooh; _ CD 28 VISUAL SENSE TRAINING IN CHILDREN subjects in the trained group completed the training and only six of the untrained group were ready to take the final test. In the following tables the difference between the trained and the untrained groups, that is, the average amount of improvement of the seven trained subjects, minus the average amount of improve- ment of the six untrained subjects, is considered im- provement due to training. Owing to the extended period the experiment covered, it was thought that the change of season from winter to spring, and the variation in the daily study program of the subjects might affect the success of the test. For this reason, as has been explained in the introduction, a control group of three was selected to take the test for the first time when the other two groups were taking their final test. It was thought that, if these changed conditions were more favorable, the control group would be more successful than the two other groups with their preliminary, and if the changed conditions were less favorable, the control group would be less successful. The three subjects forming the control group were all excellent pupils, ahead in their studies, and therefore could afford the time to take the test. Since they were the only available ones at the time, they were taken, notwithstanding their superiority. 1 , Discrimination of Pitch The preliminary part ol this test was given on Octo- ber 15, 16 and 17, 1913, the final part on April 17, 20 and 21, 1914, i. e., after an interval of five months. A pair of Koenig tuning forks, each of which possessed a vibration rate of 256 a second, was used. In order to vary the pitches by small steps a brass rider was made for one of the tuning forks, so as to be set at any place NEW EXPERIMENTAL DATA 29 on one arm of the fork by a set screw. As the rider moved httle by Uttle towards the resonance box the two tuning forks approached nearer and nearer in pitch; as it moved away from the resonance box the difference of the two tuning forks grew larger. A metric scale was used to set the rider, which was moved 5 millimeters at a time. The beats that the two tuning forks produced per minute, as the rider was moved from the end of the tuning fork toward the resonance box, were as shown in Table 6. The 840, 655 and 575 beats were counted from tone- recorder records and the rest of them were counted accurately by a stop-watch. TABLE 6 Results from Testing the Forks Used for Pitch Discrimination Position OF THE Rider Beats Per Minute 1 mm. from end of the tuning fork 840 beats 2 5 a " " " " " " 665 ' 3 10 a u " " II It II 575 ' 4 15 <( (( 11 ti tt " 480 ' 5 20 i< II II it tt a 420 ' 6 25 " 11 II tt tt tt 385 ' 7 30 " " " " " it 360 ' 8 35 " II II It tt " 335 ' 9 40 " II II it it " 305 ' 10 45 (( II 11 tt tt II 275 ' 11 50 u II II II II II 250 ' 12 55 It II II II 11 II 215 ' 13 60 11 " " " " " 185 ' 14 65 " II II II II II 155 ' 15 70 " II II II II " 138 ' 16 75 " II II II II II 115 ' 17 80 " II II " " II 90 ' 18 85 " II II 11 II 11 78 ' 19 90 " II II II II it 62 ' 20 95 " II II II II tt 50 ' 21 100 " " " II II " 40 ' 22 105 < (< it II II II " 30 ' 23 110 ( u II II II II " 22 ' 24 115 ( (< II II II II " 15 " 30 VISUAL SENSE TRAINING IN CHILDREN In the actual tests of discrimination the rider was moved 24 times towards the resonance box from the end of the tuning fork and 24 times in the opposite direction, which brought it back again to the end of the fork. Five judgments were given tor each move of the rider, making altogether 240 judgments. To finish the 240 judgments three sittings were necessary. On the first day the subjects made 60 judgments, beginning with the larger differences from 1 to 13, Table 6; on the second day the subjects made another 60 judgments, beginning with difference 13 and ending with difference 24; and on the last day the process was reversed, — this time beginning with the smaller differences of the tuning forks and working towards larger differences, from 24 backwards to 1. The. first two sittings took about twelve minutes each, the last sitting about twenty minutes. During the experiment the subjects sat at an average distance of ten feet from the tuning forks. On a desk in front of each subject was a pencil and a typewritten sheet on which was printed ''A12345B12345C1234 5," etc., ending with the letter X. A, B, C, etc., are indi- cated in Table 6 by column one, 1, 2, 3, etc., 1 corres- ponding to A, 2 to B, and so on. The numerals after each letter indicated the places under which the judg- ments were to be written. All subjects were told to record their judgments with reference to the second fork struck; that is, if this sounded lower than the first fork, they were to write under the numeral an I^; if it sounded higher, they were to write an H. To be sure that every subject knew what to do, a trial sheet was given to each on the first day of the preliminary test. The subjects were seated in alternate seats with their backs towards the tuning forks. Though NEW EXPERIMENTAL DATA 31 the lower fork was sounded last as often as was the higher, the order in which they followed each other was irregular. Thus the subjects were prevented from copying from one another, from anticipating which was the low-toned fork and from foiming any association with the order in which the two forks were struck. The forks were hung near the center of the room with the tines down, as is shown in Figure 5. Both reson- ance boxes in which the forks were set, were held by string loops about their ends. These four string loops were attached to a wooden bar which was itself hung from the ceiling. In this way, as the bar turned, the tuning forks exchanged places and hence there was no chance of the subjects associating lower or higher pitches with direction or distance. The forks . were struck by a solid rubber-ball hammer; about three seconds after the first fork was struck, it was stopped with the hand and the second tuning fork was struck. The subjects immediately wrote down their judgments, L or H. While they were doing this, the operator made his preparation for the next trial, and then an- nounced to the subjects under which letter and numeral their judgment should be written down. In front of the operator was a sheet of paper to guide him regard- ing the places of the rider and the order of the high or low pitches throughout the test. By constant an- nouncement of the letters and numerals the subjects knew whether they were writing down their judgments at the proper places. These arrangements were fol- lowed in both the preliminary and the final tests. The results of the test are shown in Tables 7A, 7B and 7C. On account of the absence of subject XIII, there were only five left in the untrained group. 32 VISUAL SENSE TRAINING IN CHILDREN Figure 5 NEW EXPERIMENTAL DATA 33 The author has found that the auditory sense in children fatigues easily. The first two sittings took about 12 minutes each, and the last sitting took about 18 minutes, in both the preliminary and final tests. It will be noticed from Table 7A that, as the pitch of the two tuning forks decreased in difference, no subject escaped from making wrong judgments. This, not only because towards the end of the test the dif- ferences in pitch became smaller, but also because the subjects were getting fatigued. That this was the fact was brought out by the record of the second sitting. When the differences of the forks continued to decrease from the first, some of the subjects started in with no misjudgments, although midjudgments had already begun towards the last of the first sitting. Some of the subjects doubtless would not have made any wrong judgments, had it not been for fatigue. We can infer this from the fact that, as the process was reversed in the third sitting, subjects I, IV, XIV and XV made no wrong judgments. It is even prob- able that some of them can distinguish differences of less than 15 beats a minute, .4 vibration. Fatigue, in the third sitting, though it did not interfere with these subjects, was a handicap for the rest. These other subjects continued their misjudgments beyond the point where their judgments were still right in the first two sittings. This raising of the threshold as the difference of the forks became larger and larger was doubtless due to fatigue. In Table 7A it will be seen that both trained and untrained groups have lowered their thresholds — the place where the difference of the tuning forks was just distinguishable. However, the trained group gained 33.43 per cent more than the untrained as the tuning 34 VISUAL SENSE TRAINING IN CHILDREN TABLE 7 Point at ichich Subjects began to Make Wrong Judgments, as Indi- cated by the Number of Beats per Minutes Trained Gioud Preliminary Test Final Test Improvement Subjects Decreas- Increas- Decreas- Increas- Decreas- Increas- mg mg I 180 78 102 II 575 575 78 335 497 240 III 840 840 840 840 IV 360 62 420 298 —420 v 60 480 155 250 —95 230 VI 385 385 90 50 295 335 VII 840 385 385 360 455 25 Total 3240 2665 1688 2255 1552 410 Per cent 47.90 15.38 Average 463 381 241 322 Untrained Gi roup VIII 665 840 575 585 90 225 IX 840 840 840 840 X 15 785 335 305 —320 480 IX 840 840 305 360 535 480 XII 335 385 250 250 85 135 Total 2695 3690 2305 2340 390 1350 Per cent 14.47 36.58 Average 539 738 461 468 Per cent of Transfer 33.43 —21.20 Control Group XIV 22 XV 250 XVI 115 50 Total 387 50 Average 129 17 NEW EXPERIMENTAL DATA 35 TABLE 7 B. Number of Wrong Judgments Made by Subjects Trained Group Preliminary Test Final Test Improvement High High High Low fork fork Low fork fork Low fork fork Subjects struck struck struck struck struck struck last last last last last last I 6 4 3 2 3 2 II 32 26 15 3 17 23 III 63 66 54 59 9 7 IV 7 4 9 18 —2 —14 V 2 2 1 8 1 -6 VI 8 10 6 10 2 VII 44 22 21 40 23 —18 Total 162 134 109 140 53 -6 Per cent 32.72 -4.48 Average 23 19 16 20 Untrained Group VIII 38 35 39 14 —1 21 IX 91 17 35 42 56 —25 X 11 12 24 13 —13 —1 XI 60 40 30 29 30 11 XII 6 13 8 17 —2 -4 Total 206 117 136 115 70 2 Per cent 33.98 1.71 Average 41 23 27 23 Per cent of Transfer —1.26 -6.19 Control Group XIV 5 3 XV 4 5 XVI 8 3 Total 17 11 Average 6 4 36 VISUAL SENSE TRAINING IN CHILDREN TABLE 7 C Total Percentage of Successes Made by the Subjects, 240 Trials Taken as the Basis Trained Group Subjects Preliminary Test 1 2 Final Test 3 4 Improvement 5 6 I 97.50 98.33 98.75 99.17 1.25 .84 II 86.67 89.17 93.75 98.75 7.08 9.58 III 73.75 72.50 77.50 75.42 3.75 2.92 IV 97.08 98.33 96.25 92.50 — .83 —5.83 V 99.17 99.17 99.58 96.67 .41 —2.50 VI 96.67 95.83 97.50 95.83 .83 VII 81.67 90.83 91.25 83.33 9.58 —7.50 Total 632.51 644.16 654.58 641.67 22.07 —2.49 Average 3.15 — .36 Untrained Group VIII 84.17 85.42 83.75 94.17 — .42 8.75 IX 62.08 92.92 85.42 92.50 23.34 —10.42 X 95.42 95.00 90.00 94.58 -5.42 -.42 XI 75.00 83.33 87.50 87.92 12.50 4.59 XII 97.50 94.58 96.67 92.92 — .83 —1.66 Total Average Transfer 414.17 451.25 443.34 452.09 29.17 5.83 —2.68 .84 .17 — .53 Columns 1, 3 and 5 indicate percentage of success when the low tuning fork was struck last. Columns 2, 4 and 6 indicate percentage of success when the high tuning fork was struck last. forks decreased in difference and showed a compar- ative loss of 21.20 per cent as they increased in dif- ference. So far as the lowering of the threshold is concerned, the result shows a greater improvement in the trained group. This finding seems to be in accord with a previous experiment performed by Bennett, NEW EXPERIMENTAL DATA 37 who trained sixteen children of eleven years of age "in discriminating different saturations of blue" and tested them in their ''other sense powers in discriminating different mixtures of (1) red and white, (2) yellow and green, (3) orange and black." The training lasted five months with two half-hour periods each week. ''There was also a preliminary test in distinguishing pitches." In this last test the Gilbert tone-tester was employed. "F sharp was taken as the norm, and the method employed, that of minimal gradations. As the figures (See Table 10) present it, the sharpness went from a range of 4.4 points at the first test — each point representing an eighth of the distance from F to F sharp, or F sharp to G — to one of 3.5 at the last test with the boys, or a gain of 20%; and from 5.3 points to 4.1 points for the girls, or a gain of about 23%."^' This large gain may be accounted for by the want of a control group of untrained subjects. In the present experiment the trained group made an even greater gain, 47.90 per cent in the upper threshold, as the difference in pitch decreased, and 15.38 per cent in the lower threshold, as the difference in pitch increased. But even such a large gain as this becomes practically insignificant when it is compared with the gain of the untrained group. It seems unjustifiable to conclude from such results that training is general. Another experiment of a similar nature was performed by Coover and Angell. "Four reagents were trained in discrimination of intensities of sound for 17 days during an interval of 57 days. Each reagent made 40 judgments in each day's training. Before and after training the reagents were tested in the discrimination of shades of gray, each test consisting of three series, " Bennett, C. J. C, Formal Discipline, p. 62. 38 VISUAL SENSE TKAINING IN CHILDREN each containing 35 judgments, delivered on 3 separate days."^^ Their three subjects made 4, 6, and points of improvement, respectively. Judging from the number of judgments given to the reagents, the per- centage of gain can not be very large. And they end their report by saying: ''Our conclusion from the experiment, therefore, is that efficiency of sensible discrimination acquired by training with sound stimuli has been transferred to the efficiency of discriminating brightness stimuli, and that the factors in this transfer are due in great part to habituation and to a more economic adaptation of attention, i. e., are general, rather than special in character."" There is possible no direct comparison between the present experiment and that of Coover and Angell, who used adults as sub- jects, trained them in auditory, and tested them in visual sensitivity, whereas the writer used children, trained them in visual and tested them in auditory sensitivity. In reference to the number of wrong judgments presented in Table 7B, the decreases of the trained group were so small, no matter which is taken for the basis of calculation — the number of wrong judg- ments (Table 7B) or the total amount of success (Table 7C) — that it is doubtful whether they are really of importance. Mere chance might have caused that much difference. The control group is useless for purposes of comparison, because of the superior ability of the subjects. i« Coover, J. E., and Angell, F., General Practice Effect of Special Exercise, Am. Jour. Psych., Vol. 18., p. 331. 13 Ibid., p. 334. NEW EXPERIMENTAL DATA 39 Summary for Pitch Discrimination 1. In auditory discrimination children show signs of fatigue very quickly. Probably 12 minutes is too long a sitting for an average child of about ten. 2. There is no positive evidence from our experiment that efficiency in visual discrimination is transferred to efficiency in auditory discrimination. 2. Discrimination of Shades of Color It seemed impossible to dye any paper with gradual saturations of color and keep these shades permanent, so shades of color in solution were used instead. Potas- sium bichromate, 20% H2SO4 + 20% K^Cr^Or, which gives a beautiful orange color, fitted the purpose very well, and test tubes of a uniform size (% inch) were used. After they had been carefully washed and chemically cleaned, each tube was filled with 15 cubic centimeters of water. Various numbers of drops of the potassium bichromate were put into these test tubes by means of a pipette. Then they were sealed by melting the open end, and labeled. This test was divided into two sittings of 17 judg- ments each, making altogether 34 trials. The follow- ing shows the oider of the different comparisons during the two sittings. In giving this test the writer used the same apparatus and method of procedure that were employed in the training. Figures 6 and 7 show the general arrange- ment. Test-tube 1 (see Figure 6) was shown first for three seconds and then, by a pull on the string (12), the plate (4) in the Cattell Fall (5) was dropped, exposing test- tube 2 (see Figure 7) and covering test-tube 1 by 40 VISUAL SENSE TRAINING IN CHILDREN Figure 6 NEW EXPEEIMENTAL DATA 41 Figure 7 42 VISUAL SENSE TRAINING IN CHILDREN 1. 2. 3. 4. 5 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. First Sitting Test-tube with 11 drops of solution shown first, then test-tube with 14 W2 39 24 70 33 5 45 2J^ 14 39 8 45 55 8 1^2 45 28 60 28 8 55 3^ 11 33 11 39 65 5 drops 18. Test-tube with 33 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 65 60 11 28 3^ 20 3^ 24 28 17 55 5 20 2^ 17 14 Second Sitting drops of solution shown first, then test-tube with 39 " 55 " 70 drops 23 5 17 2H 20 24 14 45 3>^ 24 20 17 NEW EXPERIMENTAL DATA 43 bringing up plate cover 3. As in the training, the judgment was given in terms of the second stimulus. The subject would say ''darker," meaning that he thought it was more saturated than the first test-tube shown; or "lighter," meaning less saturated. To get the reaction time, the lip-key and chronoscope were again used. Three tungsten lights of 40 watts each were arranged behind a large plate of milk glass (9) so that a uniformly illuminated background was given to the test tubes. All other lights were removed from the inside of the canopy which covered the experiment table. The result of this test is presented in Table 8. While the trained group shows only a slight improvement in right judgments, the untrained group shows an im- provement of 25 per cent more. This can only be accounted for by the increased reaction time of the untrained group. It confirms our finding in the train- ing, namely that the way to reduce the number of wrong judgments in sense discrimination is to lengthen the reaction time. It will be noticed that the loss was greatest in the reaction time of wiong judgments in the untrained group. This indicates that these sub- jects employed more time in judging difficult cases of comparison. Whenever this was done, here or in the training, the number of wrong judgments was always smaller. Subjects VIII, X, and XIII reduced their number of wrong judgments considerably by lengthen- ing their reaction time two or three fold in the final test. With Subject XI the lengthening of reaction time was not found, though her improvement was the greatest of ail. This is because in her preliminary test she had a number of long reaction times due to difficulty in using the lip-key and these longer times have in- creased her average reaction time unduly. fet- H^ (M (N C^ T-H COCO^ i-H TjH 1~-. O (M (N lO ' CO lC 03 CO CD I O < CX) CO iM <— I I CO < iMCOC000(M^Tt<(MCO Tfi COOO lO I O (M (M CD CO >-H I .-H I CD C-l CD t^ (M C-l 05 QO CO (M ^ ^ O-l lO CO iC ^ •O CO t^ CO CO CO I ' ■ '^i COO5COt>^0000(M CO 00 CD (M CD I O I Tt< I CD GO 1-1 \ C.o-*»oi>a3Q0'* TfH 00 QO CD CO •* C5 t^^ CDCD (M CO I O O O O --H (M O 1-H .-1 -* (M -"f 05 CO CO 05 CD CO'*CO^iOCO(M'-H cot--'— iCOCDOJ'^O lO 02CDCDI>Ot}H00(M OOCOCOt^T-iTtir-H CO t^-^OOC-lOOCD^ COiMt^OJt^r^'+iOO CO 1— I.— I.— (1— li-H 1— 105 >— IC^i— I >— 1>— I.— (05 "—I ,HrHOO00(MCD00 OOt^C^iQiOC^'*'*! i0^t>I>OC0OC2 1— I 1— I i-H .— I (M 00 iOOi^-*CD-*COiO oocoooio^oothco ^'^oooO'-H^oooo t^OiM^COOOO^ iC^COCDOC^rHCO t^t^c^iooi-Hc^t^io ~ ■ -' t iM O CD I ^ IM Oi O ^ X ■ 00 CO CD CD lO O Ol IM t^ CI CD O (M ai (M CO (MCOt^Oi>Ot^OCO 1— 1 1— I rt ^- ^ ^^ 00 (Mt^COCO-^iM-^CO O CDiOOr-HOOoOOOfM — 1 lO O lO 1^ CD CO 1— I lO d iC 00 CO t^ r-H ^ t^ O LO t~- COCDOlMOO'HCJ CO ■-H iM (M CI rt 00 --I C'lt~-t~-CDOO'-iC0 fOt^t^CO-i l^G0t^'f"ti"O'-H(M iQ Tt< 1— I O 00 i-i CO t^ CD CO GO CD (M CO Ci O ^ CO di-H,— I ,_|,_|,_|(35 ,— I r-l C0(Mi— IC-IIMOOIM OCOCJl'^t^i— lOOd (MH T— It-H,— (t—li— I 1> .— Ii-H T— II— 1,_| t-^ '~'t<_ o HH *-l I> J> tr h-H ^15 t, QJ I— I >-H 1^ t^ n -t^ tH S^ tn HPhIIPh .S b a lO^ Cl "H t^ 02 IC? CI Tf< rJH CI lO lO CO ■* 1—1 d ■* >— I t^iOCl rt< ^ 1— I lO T— I »D CD CJ CDCI O CO 1-1 CI Tf !-< IC -^ CI -H Tt* CO t^ -H C3 t^ CI LO CO ^ CO T-l CI '^ 1-1 HH iy* ?»■ JlS CJ OJ S; «3 bO bO fi o3 t- t- 3 O > > .X.O^ (M CO ^ r-H »— 1 1— t 1-H =i > - ^ - COCDOOOOJ (M CO ■* 10 c c as 3 3 PI c -2^ — — — 'o'o 000 OUOOU NEW EXPERIMENTAL DATA 45 The question naturally arises: Why did not the subjects of the trained group use more time, after having had the special benefit of the training? The answer is that the subjects of the trained group were growing tired of the experiment; its novelty was lost in the three months' training. They seemed to be unwilling to give the necessary time needed to visual adaptation, and in many cases they seemed to have left it to chance whether they were right or wrong in their judgments. On the other hand, the subjects of the untrained group seemed to be only too glad to have a change in their daily program, to come to the experiment room once more, especially at the begin- ning of the retest. In none of the tests, preliminary or final, were the subjects of either group told of the rightness or wrongness of their judgments, as were the members of the trained group in the course of their training. This only made the trained group all the more unconcerned at the beginning of the retest. In a word, interest is quite essential for success in sense discrimination in children, for they must be willing to give their time for purposes of sensory adaptation. Summary for Discrimination of Color 1. This test confirms the finding in the training, that lengthened reaction time is the means to reduce the number of wrong judgments in visual discrimina- tion. 2. Interest is essential for sense discrimination in children that they may give the time necessary for better visual adaptation. 46 VISUAL SENSE TRAINING IN CHILDREN 3. Discrimination of Size The tested functions included not only discrimina- tions of color and pitch but also discriminative functions more closely related to those in which train- ing had been given. The subjects had been trained in discriminating vertical lines ranging from 1}4 to 13^ inches long with a difference of Vao of an inch. Before and after this training, a test was taken not only with the same sizes as in the training, but also with one set of larger sizes, ranging from 2Vio to 2% inches, and one set of smaller sizes, ranging from ^4: to 1 inch. These different sizes were also drawn in different forms, namely, as circles, triangles and horizontal lines as well as vertical lines. The differences to be discrimi- nated in these various sizes were not the same; in the size ranging from 2Vio to 2% inches, the difference to be discriminated was Vw of an inch ; for the size ranging from ^ to 1 inch, V20 of an inch; and from I3/3 to 1% inches, the same size in which training has been given, there were two sets of differences to be discriminated — one set, which included all the various forms, had a difference of Vii of an inch; the other, which included only the circles, triangles and horizontal lines, had a difference of Vao of an inch. Vertical lines with a difference of Vso of an inch were used only for the train- ing and were excluded from the testing. Table 9 gives in detail the various sizes in different forms with the various differences to be discriminated. It will be noticed that the trained function differed from the "a" series only in form, from the "A," ''B," and ''C" series of the vertical lines only in size and differences to be discriminated, from all the "B" series of circles, triangles and horizontal lines only in 5§ C) 05 -« W B >-A _ « S H ^ CO OJ _C cj :: :::::r"-'-^^^-::3^^ c - •- ^i-\m ^^M onsm -i5\=:5 \M\M\m\m CO^IN\ sS^Cf- 0^-c5s «^ W UHS> 48 VISUAL SENSE TRAINING IN CHILDREN form and differences to be discriminated, and from all the ''A" and ''C" series in size, form and differences to be discriminated. It is easily seen, with this arrange- ment, that the vertical line series was most closely related to the trained function, next came logically the horizontal lines, the triangles and circles, in accord- ance to form; with reference to size, that ranging from 11/^ to 1% inches was the most closely related to the training, next came that ranging from ^ to 1 inch and 2^10 to 2''/5 inches; and according to the differences to be discriminated, ^so, V20, Vis and Ho of an inch, in the order stated. There were altogether 15 series in this test, as shown in Table 9, making 15 sittings for each subject. These subjects came into the experiment room in turn for fifteen minutes in the morning on school days and made ten or twenty judgments in accordance with the different series. The order in which the various series followed each other was the same as shown in Table 9. The preliminary test lasted from November 4 to De- cember 17, 1913, the final from April 28 to May 17, 1914. Much less time was taken during the final test, because of the dropping out of four subjects from the trained group and five from the untrained group. The hours of experiment for each subject were kept as nearly as possible the same in both tests. The general method and apparatus used in this test were the same as those used in the training. The results of the test are presented in Tables 10, 11, 12, 13, and 14. Table 10 shows in detail how the percentages of transfer were calculated. What each column in the table stands for is explained in footnotes immediately following the table. The various amounts of transfer are rearranged in Table 11 so as to present ooooo 1-H i-H CO CD Ol CD CD r^ CO ■* C'l r^ <^5 CO lo (M 00 CO CO 0(M O LO »o •Ort^ci cotj^gc CD O iC 1 1 CO Tt^ CO iC CO GO CO 01 1 C^l -^ o OCM CM COO 05 005Ci (£> T^ a C5 ^ 00 O t^ t^ TJH CD O (M CO-* CD ^ ^ t^ CD CO -TlH ^ LO «D iM GO CO o^ lO r-H t^ CO^C^ CV1 00 --H TJH ^ lO ooo O O lO r^co 05 I> ^ lO CO CO CO CO ooo lO o lO (M »oco^ rt CDl- CJOOC^ CO--H i-H cooco CO CO rH (M CO s &~, 0^ ^ iM CD -* _OT lOC^ 3 05CD 3 t-O 3 CO 00 IM ■* J^ t-iM 3 CD CD t^ (M OC5 r^ 00 CO CD 'GO .2 CI CO --1 iM 'in C^ t^ !h rt o 'C LO CO 3 ^ CO ^ CI ' i-, O^ "'"' CI 0 CO ^ TtH t/T -r -+< VI t^ t^ CO -* LO T-H OiCD ^ LO CO CZ3 COIM _0J O CO _o O CD .i CO -H ^ OOO CO t^^ QJ i^ ^ LO LO CD O 73 OC3 0) 05 05 C5 02 S r-ICn cot^ o O t^ cl; CO CO i-H t: to O to <1 fQ u M CDt-H « --'ThI o coco COCO lO t^ CI cq ■ "So COO ^ o "p c C c 13 ^C<) 03 CO o 02 02 .3 CI o •g Tfi CI S CO'* O '"' '"' H rt I-H H rH rt H 1-H 1-H r^^ OO CO CD ^o COO OOO coio 02 CD LO 02 ■* c^ 1-H ^ LOO -S< 02 t^ GO lO t^ lO t^ 1-1 LO OO o ^ Ol (M O(0: OO ^ l^ OO rt 1-1 ^ iQ 00 O-H OTCI ^ 02 O t^ O CO t^ -— 1 "* LO CO CO ^ CO rH lO O CD Cl^ l>LO O c^ CD CI CI ^ CI CO [^--H i>o oooo Q0O5 00 1^ o 00 ■"^ !M CO t^':t< ^#00 lO LO r^ t^ CO (35 r- CO Oi c; CO CO 1-H r-H TfH CI OO o\ ^ X'Cl 00 Oi ■* Oi I^CD CD LO 00 1-H t-O 00 00 Oi O 00 1^ O GO GO O OCO TS ao -a fl S T3 Cl « Qi-'-'U-. n; •-H t+_. O •-; «— 1 c 2 ^ J J3 g .s|s rt "S ei rt tS rf ^^ & t-t G t- H^H H^H HPH -o c 53 •Sag S c3 CO u C ti TjH o o O OO GO 05 O 00 lO (M (M ■* CO CO (N ^ —I '-H rt (M 2 &^ fe 00 coo 'C ^co 'n o o COO -5 -* CO ^ ooo lOOi -l-i TtH O -*-* o t^ OO t^ o t-CO o coO l> lO o 1-H OTf< o 1-H t^ lO O oo CO COr-l O IM o ^1> OOi 1-1 1> 7-H o ^ o o 00 I> ^~' 1— t 1— t ^ T— t [yT T— * »r <£ S OiM .2 OkO a; l^O ^lO .Si iOI> CO t^ "C ooo O -* O 03 o ^ Tt< CO ^ CCIM en »oco COCO coco to ^ Tfl CO o O (N OO) m ooo O oco « ot^ o ot^ 1— 1 ?? 0.-I OOC^l S Tt ^ CO C3 00 (M COIN _c c^o i0 05 03 1— 1 ^ to lO *""• Tt t>. •■^ OO"-! ^ 003 ' ' ooo ""^ oi> , , ooo '^ OI> "^ C5t^ O lO 00 lo OTt< ot^ CO CO 00 lo 00 --H -H,^ g KH '^O Ph t^O C-l CO i>o w coo o ca oo i> 00 ^ o " T:t< CO c^ 00 rH 00 --H ■^ o CO o 00 CO oo ooo OTt< CO o 00 00 oo -^ o oooo CO o IN=g t>. o o ^~ 00 IN rt o oo t:^ CO oo GO(M 00 1^ o^ 00 (N coo 00 I> lO o .-H CO oo (M CO LO 00 (N O 00 00 coc• -d T3 'C T3 T3 TS -i C !h HPH HPH HU)H h;^h HPH HPH 0) i-N p^ ;^ ■* lO o3 CO CO -kj ,-liO CO CO CD 00 (M o COTt< O s 00 00 pi oo (M t^t^ »— < o c» 00 02 05 00 'C o to -4-3 (M _o o o CO to" oo ^co en 05 lO >^ .2 ^ t^t-H Ih coio o ^ o o IM ti H Tt< Tfi ■<* TjH o C3CO o CO w t^i-l ^T^ coo ^ CO COt^ 00 00 CO(M to lO 02 coco TtHt^ H i>a3 i>oo 00 00 b c3 a OS 00 lOrH O 05 '3 ■* ot^ t^t^ O C C C o .2.2.2 i"' M &i > > "S 3 o3 o3 c3 i^ 03 c3 TS Jh C ^ . ^ Oj -^ Jh li O .f^ .2 .S ° ^^ £2 ^7 I g00>><|O'-i .S ^ H <1 to to CQ to CO to sssassss 'o'o'o'o'o'o'o'o OOOOOOOO ^ o o o o O (I> OJ O) ^ bC hC bC bC K c3 03 03 o3 03 -^ -fc.^ .4-3 .^ -^ fl C P! C c^ (U (i; a; QJ> ■— ' o o o cj ^ tH t, t< fcj CSJ O O) CI roooocoocoooooococci T-H CO (M 00 Tfi CD 00 01-*r^05i0(MO00LQ00i0TtiOiO'— I '-Hcr)Ttiioo5cooioco^-*(MO!r>^H i>co.io-*io-*co foeoor^ooooooooco'ot^ COi— iiOCOTfOrt«DC<)OCDl0OOrH-r« '*'^COCO(MC^?Tf{MiOCOCOrt<'g.0505'-iOOOCO'*OCOOOTti^H '0(M(Ml>0000^.-HOTfiOCO'-Hl:^ 1— I T— I 1-H ,— ( (M ,-1 (M (M CO CO oooooooococoi:^ eocooocoooo-^ocoajOTt^ooxo (Mi-Hi-(,-i(M,-HC0(M'-i(M:0{MOt>l>! 0505iMOOCD.-(t>.OOG0050iMiOC^1 t-^(McC)coioo5co»oa3QOi>a3t^^](M ^i-i(MC0(NOiMC5rtOr-iai^rtiJ OCOCOt^03050505COCOCOOit^->*-* O^00iO(M(MiM(M00Tf<00(M"Ort^ CI cc ...... ^ 00 bC Ml ■r;-o r» CO pi"n il tC o3 -a > bXl'5 2| 3 S ;>■> •S.S bC bC > ^ o3 oj g.S 'c^2 CO'rtT T-KM s s 3 3 oo 58 NEW EXPERIMENTAL DATA smaller differences to be discriminated also caused an increased number of wrong judgments. The "a" series, having the smallest differences to be discrim- inated, had the largest number of wrong judgments. Since all the ''a" series had 20 trials while the other series had 10, divide by two to compare with the rest. Summary for Discrimination of Size 1. When the tested function was closely related to the trained function, within the field of visual discrimi- nation, there was a small amount of transfer. 2. The larger amounts of transfer took place with those series which were most closely related to the trained function in form, size, and magnitude of dif- ferences to be discriminated. 3. There was an indication that training in the discrimination of vertical lines interfered with the discrimination of horizontal lines. 4. There was a tendency to underestimate the second stimulus in judging sizes ranging from H to 2% inches. 5. The amount of interest children showed in sense discrimination had something to do with the num- ber of successful judgments. 6. The number of wrong judgments increased as the magnitude of the differences to be discriminated grew smaller. 7. The various series of the final tests again con- firmed our finding in the training, that improvement in visual discrimination is largely a function of reac- tion time. If the subject lengthened the reaction time, he decreased the number of wrong judgments, and vice versa. NEW EXPERIMENTAL DATA 59 Experiment 2 The second experiment was conducted at the same school and at the same time as the first experiment, with different subjects from the same grades. Table 14 shows the age, teachers' estimate of this group of subjects, and mental improvement from 1911 to 1912 according to the Binet tests. TABLE 14 Subject Age in Jan., Teachers' Estimate Mental Improvement 1914 1911 to 1912 XVII 9 good 1.4 years XVIII 12 average 1.2 " XIX 10 good 1.2 " XX 10 average 1.4 " XXI 13 average .0 " XXII 11 average .6 " XXIII 12 good .8 " XXIV 10 good 1.4 " XXV 11 average .8 " XXVI 10 good .8 " Average 10.8 .96 " For the preliminary and final tests a printed page of nonsense words of three letters each was used. Each subject came into the experiment room in the morn- ing on one of the school days, sat down at the desk and was shown a paper with a written alphabet. This alphabet was divided into two parts, from a to m and from n to z, thus : abcdefghijklm nopqrstuvwxyz The subject memorized the letters belonging to each half of the alphabet for about three minutes, after which he was given a page of the nonsense words and o ft^ Or-H,— Il0t--»00'-Hi— IT-HO C3a>co•-HOc»C5CO,-^co^o ^ (M (M I I 1-1 Tt< (^^ o CO COt}HiOi-iiO,-It-h1>^05(M Oi-hOOi-hOOINC^OcO coc5cO'-Hi£i05iorO'— ii— lO COiOfOOCOdCO-^iOCDIM 1— I 1— I i:D O O O o Co .2 g fl rt rt !- O « I a p ^ ;? Xi tH y-l , Tt>-coc 7t-i H . iM CO ■* lb", to CO sa fl (3 fl o o O o'o O oooooo &5 e O CO W fe. e )-] ^ m ^^ 1 o u a S ^; lOiOOcDGOOiOiOCJO 1—1 i-HCOr^C^KMCOC^C^t^iO o COC»(MTtGOCOCO'-iO o COCOcOO:cOt>iOTtiCiiO lO II 1— 1 I— i lO a> OlOOrHOlM'-HiMOOtM 00 r-H(M ■* ^ « T-icDiM^^(M^O5l>00 00 CO ^ Tji rt ,-( iC(M o (M c« •a t. •^cor-T-HOocoocoio CO 00 fe i-H ,—1 o cooococqc^t-.cococoio 00 l>^lOi— IC5C0050I^GO !>. i-H 1-1 (M 1— 1 1— 1 i-H C^ (M (M 00 I— 1 lO (MiM'HCOOtNCOTjHT^iO CO (M (M(M i-H o -IJ cD. t^ G •g 00 OCOOt^^OlGOiO-^CO 00 UOrH-*THlOCOlO«3fOlO CO i5 1-H lO C3 Lh Ph (M C0O>CCDiOC500O(M>O (M ■*C^r-HO(MOOCOOGOCO CI 1-H 1— 1 CO t-l i-< 1— ! 1— ( CO - XVII XVIII XIX XX XXI XXII XXIII XXIV XXV XXVI _^ Pi Ph m u tn rn u & ^ g ;-4 n :3 3 M^^HW li, M CO t^ 00 C5 J_i cfco^-^^io (/? cz: M W3 SH fl fl P! Cl 'o'o'o'o'o 00000 62 VISUAL SENSE TRAINING IN CHILDREN asked to underline words that contained either two or three letters in tine last half of the alphabet, begin- ning from the top of the page. The operator watched the time. When ten minutes were up the subject stopped marking. The following three lines copied from the testing page will make the experiment clear: itp dje zna dkt giy hkr cbe dby vhl xgt hju wdy zxi fgy hkp msj vgr fte sdw cng bjy dhe cgx zaq For training, a different page containing the same kind of nonsense words was used, but the subjects were asked to underline words containing either two or three letters in the first half of the alphabet. Each subject was trained three periods averaging about 83^2 minutes each. The preliminary test took place in December, 1913, the training in January and February, 1914, and the final test in March, 1914. The interval between the dates of testing was two months and between each of the trainings was about 12 days for each subject. The improvement from the training is presented in Table 15. The improvement of the final test over the prelimi- nary test is shown in Table 16. Table 16, however, does not show the amount of transfer from training, because the influence of the preliminary test has not been checked out. Owing to the limited number of subjects available, an un- trained group was not provided for. The author assumed that the effect of the first training upon the second training would be about the same as the effect of the first or preliminary test upon the second or the final test. If this had been true, the difference between the first and the second training might have served as a substitute for a check by an untrained NEW EXPERIMENTAL DATA 63 group. This assumption was incorrect, because the interval between the preliminary and the final tests was about four times as long as the interval between each of the trainings. In accordance with the law of disuse therefore the function of marking the last half of the alphabet was more weakened than that of marking the first half. The amount of transfer cer- tainly should be more than is shown in Table 17, just how much more the wi'iter is unable to say. For lack of a proper check of the influence of the preliminary test, the amount of transfer in the experi- ment is of little significance. However, there are other points of interest that this experiment has brought out. In the first place, the quicker improvement in mark- ing letters of the alphabet as compared with sense training deserves consideration. There was more im- provement after two trainings in marking letters of the alphabet than there was after 40 trainings in dis- criminating differences of size. The introspection of the two trained groups throws some light on this point of difference. In the experiment with sense discrimination, when the subjects were asked to explain the reason of their improvement, four out of the seven subjects could not give any reason at all, while the other three merely stated: ''I give them a good long look" (subject I), "I do not hurry myself to say 'shorter' or 'longer'" (Subject VI), "I looked at them harder and tried to keep the first figure in my head" (Subject VII). This indicates that the only device developed by the training was that of lengthening the reaction time. Only two out of seven subjects were even vaguely conscious that a longer time was necessary in order to judge more correctly. On the CO ■2^ "^ M OOCCO—ikOOiMrHO CO ►~3 1 1 Cn s •<>> -t^ ^ ^ ^ a M O>CO(M00>O(M'-iOt^ CO ^ S 7 I-" CO S > •^ £ S tH »0(Mirafococoicoio05 t^ 1— 1 •<* C0 1— 1 ,— 1 ,— 1 fo CO ^H lo 05 "s 1 (M 1 ^ C lO-^iOOOiOCOO^OOO t^ ^ OO i-H C^ iM (M t^ '^ 00 1 t^ 05 s ^1^1 1 ■ a o •S a t^ rH o C0CO(M'*(MeOCOCO(Mt^ CO s '~' I— 1 lO •rO to S N rH'^OOC^lOOt^'OCSCO r-< a cooO'^coo5'*ai(MTtH t- rH C^ 1— 1 1— 1 rH rH rH h. s •^^ -s- 1 5 s l-HI-H»->k>hHHHh-H^S».»-H |-HHH?Nl^V<^>-H!:_^f^'i. "3 e2 C<3 m ft 'O "& SB HH <0 ■*^ Pi QJ o PhO ^ o M O ■;3 M n 03 > CO kO 05 11 TjH lO S £ 0) S P o So PM tn T3 ;h O ^ <*-< ^ CD O CO o £ O rH o> ^1 (N CJ (m' ■* CO -2 u c aj o t^ O O! 2 > to 02 ,12 O 02 i o s o (•2 hr ? '^ o c3 CO oi --I S p^ g s s s 'o'o "o o o o UOOOOO NEW EXPERIMENTAL DATA 71 the same and then to see whether they are together or not." Subject XXXI: 'Tor the first days I tried to compare the words as a whole and this was rather confusing. Later on, I happened to think there could be no possibility of the words having two adjacent letters in common if the three letters in the center of the two words were all different. Since that my atten- tion has been directed more to the central letters of the words. Many words which usually wasted time were skipped this way." It became very evident from these introspections that the amount of improve- ment depended upon the methods that were employed. In spite of the fact that Subject XXIX had more training than XXXI, the latter showed a greater amount of improvement. This confirms our finding in the preceding experiment, that the methods em- ployed are the primary factors determining amount of improvement. By comparing the age of the different subjects it will be seen that the older subject always had the advantage over the younger subject in his ability to develop better time-saving methods. Even at the very beginning of the training there was a difference of method. Subject XXVII compared the words letter by letter; Subject XXIX, by groups of letters of two or three; and Subject XXXI, by whole words. This suggests that the older subject, having had more experience in reading, had already developed methods which the younger subject was not capable of at the start. This would indicate that there are certain ''lower" methods which must be mastered before any "higher" methods can be attempted. In this particu- lar case there are evidently three grades of methods, the letter method, the letter-group method and the O t^ ro o;dco OCDCO O coco t^ (M O OO 05 i-i IM O (N 1— I O I— I cc o cC0 -*!> o 4) OtOcO i-:i m i» g 0(M a 3 to CD O s O ■'tl 00 g ■3 ^ si £ t>. (N ^ iO(M s ^o!$i o (M --H 1-. Ah t~-. lO ^ '-''-' IN # ^ PL, o CC* OlO T— 1 >— 1 73 ■-I (N ^ o 03 o ^?H S ;h o 02 *-3 E-. s CO C5 CDI> H COt^O 2> 1— ( (M '-Iir5 iM COo ooOioS '^'*con. 00 (M m ti<^ I— I X! I— I HH QJ n M y ■*^ a> P c3 a o 'o'o'o O ^ ^ ;h >H qj O » ~ McDt^OO 05 I C chology, Bulletin No. 53. Bryan, W. L., and Barter, N. Studies in the Physiology and Psy- chology of the Telegraphic Language. Psych. Review, Vol. 4, 1897, pp. 27-53. Studies on the Telegraphic Language. The Acquisition of a Hierarchy of Habits. Psych. Review, Vol. 6, 1899, pp. 345-375. Colvin, S. S. a Partial Justification of the so-called Dogma of Formal Disciphne. School of Education, Bulletin No. 2,. Univ. of Illinois. Colvin, S. S. The Learning Process. The Macmillan Co., New York, 1911. CoovER, J. E., and Angell, F. General Practice Effect of Special Exercise. Am. Jour. Psych., Vol. 18, 1907, pp. 328-340. Foster, W. S. The Effect of Practice upon VisuaUzing and upon the Reproduction of Visual Impressions. Jour. Ed. Psych., Vol. 2, 1911, pp. 11-22. Fracker, G. C. On the Transference of Training in Memory. Psych. Review, Monograph Supp., Vol. 9, No. 2, 1908, pp. 56-102. Heck, W. H. Mental Discipline and Educational Values. John Lane Company, New York, Second Edition, 1911. Henderson, E. N. Formal Disciphne from the Standpoint of Ana- lytic and Experimental Psychology. Education, Vol. 29,. 1908-9, pp. 601-613. Hinsdale, B. A. The Dogma of Formal Disciphne. Ed. Review, Vol. 8, 1894, pp. 128-142. HoRNE, H. H. Psychological Principles of Education. The Macmillan Co., New York, 1906. The Practical Influence of the New Views of Formal Disci- phne. Education, Vol. 29, 1908-9, pp. 614-623. 81 82 VISUAL SENSE TRAINING IN CHILDREN JuDD, C. H. The Relation of Special Training to General Intelligence. Ed. Review, Vol. 36, 1908, pp. 28-42. Meiklejohn, a. Is Mental Training a Myth? Ed. Review, Vol. 37, 1909, pp. 126-141. Meumann, E. The Psychology of Learning. D. Appleton & Co., New York, 1913. MoNTESSORi, M. The Montessori Method. Frederick S. Stokes Co., New York, Fourth Edition, 1912. Peterson, H. A. Note on a Retrial of Professor James' Experiment on Memory Training. Psych. Review, Vol. 19, 1912, pp. 491-492. PiLLSBURY, W. B. The Effects of Training on Memory. Ed. Review, Vol. 36, 1908, pp. 15-27. RtTEDiGER, W. C. The Indirect Improvement of Mental Function Thru Ideals. Ed. Review, Vol. 36, 1908, pp. 364-371. RuEDiGBR, W. C. The Principles of Education. Houghton MifBin Co., Boston, New York and Chicago, 1910. RuGER, H. A. The Psychology of Efficiency. Archives of Psych., Vol. 2, No. 15, June, 1910. Sleight, W. G. Memory and Formal Training. Brit. Jour. Psych., Vol. 4, 1911, pp. 386-457. Starch, D. Transfer of Training in Arithmetical Operations. Jour. Ed. Psych., Vol. 2, 1911, pp. 306-310. Thorndike, E. L., and Woodworth, R. S. The Influence of Improve- ment in One Mental Function upon the Efficiency of Other Functions. Psych. Review, Vol. 8, 1901, pp. 247-261, 384-395 and 553-564. Thorndike, E. L. The Effect of Practice in the Case of a Purely Intellectual Function. Am. Jour. Psych., Vol. 19, 1908, pp. 374-384. Thorndike, E. L. Educational P.sychology, Vol. I and II. Teacher's College, Columbia Univ., New York, 1913, 1914. Winch, W. H. The Transfer of Improvement in Memory in School- Children. Brit. Jour. Psych., Vol. 2, 1906-8, pp. 284-293, and Vol. 3, 1909-10, pp. 386-405. Winch, W. H. Accuracy in School Children. Does Improvement in Numerical Accuracy "Transfer"? Jour. Ed. Psych., Vol. 1, 1910, pp. 557-589. Further Work on Numerical Accuracy in School Children. Does Improvement in Numerical Accuracy Transfer? Jour. Ed. Psych., Vol. 2, 1911, pp. 262-271. INDEX Age, and improvement 71 Apparatus for training 16 for testing pitch discrimination 28 for testing color discrimination 39 for testing size discrimination 46 Bagley 3, 5, 77 Bennett 36 Berry 15 Binet tests 14, 59 Breed 4, 78 Color discrimination 39 stimuli used in 39, 42 results in tests of 43 Colvin 3, 5, 77 Coover and Angell 3, 37, 38 Disciplinists 1 Ebert and Meumaim 3 Formal discipline 5, 6, 77 Foster 4 Fracker 1, 2, 3 Habit, in relation to transfer 3, 77 Heck 5 Herbartians 5 Incubation 9 Individual differences 7, 14, 25, 43, 63, 66, 71, 73 Interference, of training 52 83 84 VISUAL SENSE TEAINING IN CHILDEEN James 3, 6 Maturation 8 Method as means of transfer 3, 63, 76, 78, 80 of training 16 the Individual 6 the One-group 7 the Two-group 8 the Three-Group 11 Montessori 13, 14, 76 Pillsbury 4 Pitch discrimination 28 stimuU used in 28 fatigue in 33 results in tests of 33 Plato 5 Reaction time, and improvement 21, 26, 43, 45, 56, 58, 75 Ruediger 1, 3, 5 Ruger 3 Scripture and Davis 3 Sense training 13, 21, 26, 39, 43, 45, 63, 75, 76 Size discrimination 46 stimuU used in 46 results in tests of 48 Swift 9 Testing of pitch discrimination 28 of color discrimination 39 of size discrimination 46 Thorndike 1, 2, 3, 5, 7, 8, 9 INDEX 85 Training method of 16 apparatus used in ; 16 improvement from 21, 26 in relation to interest 45, 56, 58 means of improvement through .... 24, 26, 43, 45, 58, 63, 66, 69, 78 interference of 52 rate of improvement in 24, 67, 69 Transfer extent of 2 amount of 2, 33, 39, 43, 48, 52, 55, 58, 62, 73, 74, 79 means of 3, 63, 73, 74, 78, 79 and common elements 58 meaning of 77 from visual to auditory discrimination 33, 39 from size to size 55 from size to color 43 Whitney 4 Winch 1 LIBRARY OF CONGRESS miiiiiiHiii e 021 339 578 6