Raf i ae nae Ons (ee PY Tied WEA SUR) be Ye ik Mk peed BRIA" 1 ei ay t if Lea + te ry: Tbs ae ReRETAY G » * < > siy at Paar Na? Hie Ha i aE eit ‘ "aed ny % ope: RBA Ns hat phase uA Mi isa Cay t is in RAT A ARLE TOLa F mnecn GATS a adn Shae stat Ni H i; et Ary We ‘fi 4 4 ieee ie a 16 aya anaes fiitteihret 4) Spates SCO) PER Eaee! p i te ts iu nd > ; 1% we ; Pea yi 4 i hea Ye heebr 99 fein Pai fe of UTTER ER LL] Airs % ite Ht of Cee 4 fe af? gree of i fe ty alee A: is he % +t aie aAehaad Guard BR TAG aunt a a Pree eUm Ley edi 14 249 ny roti ta mie Rh Rae pee Hea } UU Rehr t a : V He fe tt 7 is inh i y 4 ? tee 4 z f Werth eon GMs ve sa bigs iA 19241) P92) OPA dais ot Mame tele heme rie’ ray for ee » ses te i +h Shad Sentai BA) tO sem Mama ( Ne be We oa i et iy Nahe Mout! Oru hate * het » jal i 2) A A e¥5Prssse Seeate ¢ Sisteseesee nas noe Setates we ee ae as Fa, a Se guise Baie eas ea ew So Sy rey RI) BG rhs Peabalhn BaP Hl Hat iG x heh ae es ! (if + nn aah baie beach eb PRET a Ds Ma 4 pa Pat t Pere eT eon htes ana weed iba #47 bods aig Tp. ' rh Seees eoeleee tan | Sree ) EOE i ey 844 3 ith 8 +55 fear ees jas qi Mean PP } roe Py Av Theft fix Pheer ys : pith [aa Pie fe ie 6 hh ed ae i J ester > Pear cueeeane ta Ae eae caes Me r s EEE ee tHe poems RIAA S Dir bec uahiaeer ath reiwe a Aa aren rts BU AR, he i 4 La iit Leh} ray ee srited ta Pets gny 2 : : a ae fora iP ip ail) ah Zi ne if ‘ tre as ane . Cs : My on! Ain fr Pf Sf Sead UbysCHOLOGICAL REVIEW PUBLICATIONS WHOLE NO. 164 Psychological Monographs EDITED BY SHEPHERD IVORY FRANZ, Unvv. or Catir., So. BRANCH HOWARD C. WARREN, Princeton University (Review) JOHN B. WATSON, New York City (Review) MADISON BENTLEY, University or Iturnots (J. of Exp. Psych.) S. W. FERNBERGER, University oF PENNSYLVANIA (Bulletin) and W. S. HUNTER, Crark University (Index) Comprehensive Units in Learning | Typewriting By J. W. BARTON, Ph.D. Professor of Psychology University of Idaho PUBLISHED FOR THE AMERICAN PSYCHOLOGICAL ASSOCIATION By THE PSYCHOLOGICAL REVIEW COMPANY PRINCETON; N. J. AND ALBANY, N. Y. Acents: G. E. STECHERT & CO., Lonvon (2 Star Yard, Carey St., W. C.) Lerpezic (Hospital St., 10); Paris (76, rue de Rennes) os Ley ZORA MEM TINIGE 5 SS Ox HIM! 4 Ui | 1 1926 ='sf a ocica, EWS SS a ae AF ats ; a7 rad af ae & ) Li = Fe Fs i cm ay ee > be * ee ¥ ai! ss ag WAT, ¥ ACKNOWLEDGMENT For inspiration and direction in the shaping of this study, my heaviest obligation is to Dr. Joseph Peterson. He also contributed greatly in the criticism of the data and manuscript form. The work of compiling the data and putting it in form was so voluminous that it could never have been accomplished in the time allotted without the helpful aid of Miss Editha Barton and Mr. W. A. Barton, daughter and son, respectively, of the writer. They gave generously of their time in this work. I am greatly indebted to Prof. H. C. Dale and the University of Idaho Admin- istration Force for permitting me to make use of the equipment required in the main experiments run for obtaining the data here presented. I wish to make grateful acknowledgment for all this aid, without which I could never have completed this study. pe Ti Sal Se hs Ge eee nee tee Br ay iy dase te AES . : : ‘ oy ob Av ; i ’ an ‘ pe : or : a ; ie we TE tyemeae atinge es wae + ey ay f. pi ‘ite wt arth. Parr e 3), ‘Gear e c) oa be et A +) oe Aa 2 | cf oS i oe 2 yt ete fear nie xia ast tas iti # re’, - et sth rg ade rf isi Ns Wie vse | , ; ody ‘ ay, festa Ate" gp. all hi ns y lan, Ug ie? toler e-.. Cae ARR ak | eimai arte ae ear ri A ¢ - i ; a ‘ i 4 Pa by? “ + i“ i r , it . Pe ; ‘ Le ia Seba i ess Vomndts 7th) Sa e: Pi faiay Acct Sab t Pathe ap is toe Ge os | if } r 7 Te OO rn z Pe . OV, r h . 7 &2 . pe ms 7 Le DUET EON Pe wi hin fate edie F eevl. eas 3 : Jere ane SUMGLE PLEASE i Pe uae & & ves a al iu ees zine st ain x t a m_ ( ‘ einen , ty k ’ i * ‘ { ‘s P Pa Ete we Airy O10 a Oe at ahs: TABLE OF ‘CONTENTS Section ie Rage oy Toheg EP od ASS GG WO DAG he Rd Race a le BUC hab 1 CFERES UL LATE MER De. Mee Ss es Se 1 EES IVeT TSTUOIES Ot ee Re CeO a rk RE Se Pa 3 FETOCTICUL UOT REIS CC i OE Ls CE Peet 8 PUY LOSER Pe Rete rite Fes Pe ee tk, SS 10 NE ELEC BGS st ale ac Aa ordain wrt la aR ACNE Stat 11 OUP CITE a PER UIE MINS Ue Reet AERIS Hag wr ahd chie a'g) eh ge ae mI USCS PM eR eer cree eo Ng te geet eae 11 Pee ECIiCes MIALCI al ade eee, ce eR! Woe eek 12 DEO OM MCL DETUINEME ct Viet res Go esete Cae Ae ye Soe 14 LUGE alot Re AS Vie Aten iach DN ay SIR aE RIN nS 14 Pi eeieisen ialeniauee Ca eer eee ee re ee ns | 15 CRSLE Le OUET OL actae Mori eR ae ee Ras oe ee (oy of oo |e BA IT+ 5) VPI) tlh ee 5 ea Raa Ree SPE PR aca 5 dec Ra 18 DIMI D Vals: SAN San oP Ba No Pale ne web ih als sie e's 19 Gey Pals lt, kg eee MAR ae deems See, CMe 19 Seeuict ast Atemient tet ian i pals a eee GS aA UK) Sil eG he 19 IPS Sid NCE CISCIISS! OF) mate recy pee ee ie ah eh va) outs 19 Need of further experimentation...........0..45. 26 ECON ECA DEPICN I. aie Brae wate Te LM cVinsn's aly tans oak) 28 PoeTiGr ae st ateinentons are iooEs alle ao halen) ONS 28 ese OG GOISrUISSIOth ys cn. ee de ie. oe yan peas 29 SE Ba URS Coad oper Noow yt ag ea DORR ihr fr RCRA ie SE EA a 33 a MATES EUR VGEY reo see elec Ree ered rd eer Ee 4 eae Nee mae Gy. Peis losiONSrAND:GONCEUSIONS ¥ 32.0 28.88 40 TORU DIE NES ORr Oa Me EU ek eet a 40 DV SALERLT SUM eetiues: peti Lae ena aie a ven ie, Se Se sary 40 IEP atl pi eee aster bet hte Ai ues Te abate 4] TOOLS ban decd OA a Aenea sete ¥ o., Lk oo, a ee 42 SE PALICONTOS PSA UGA te SAS gc erat a eee ek oR Se 46 Digitized by the Internet Archive , in 2022 with funding from Princeton Theological Seminary Library ; | https ://archive.org/details/comprehensiveuniOObart : Pe tise CL tv General Statement.—There are two prevailing notions concern- ing the factors involved in learning, which seem to be funda- mental in determining what shall be done in order to produce the changes known by this name. The first, and possibly the more basic, is founded on the belief in the duality of human nature; it is the idea that mind and body are essentially different and thus require treatment, in producing learning, in some ways which are not only different but are even opposite to each other. The second is the view that to repeat an act over and over again is the law of learning. The first of these notions is responsible, consciously or other- wise, for statements like the following: ‘“ Learning by means of associative memory is a higher type of acquisition. The stimuli do not result in trial responses. Instead, the nerve impulses pass from center to center in the brain, arousing a succession of images and thoughts. We picture to ourselves various ways of acting; if one course of action does not solve the difficulty, we picture another, and so on till we picture some action which brings about the suitable result. Then at last the nerve impulse passes out into the appropriate motor channel and we act’”’ (34, p. 265).’ The other notion is responsible for much of the isolated drill that is so widely used, particularly in the acquisition of skills. Has typewriting been properly classed in the attempts to deter- mine the nature of the factors responsible in the acquisition of this very important skill? Is it a motor learning process of the kind consisting mostly in learning higher units of reaction, and does this process necessarily begin in the writing of letter com- binations that are never found in any composition to be used after the learning has been accomplished? Is there any possibility that the successful practice of teaching reading (16, p. 300) by even 1 References will be found at the end of the monograph. 2 J. W. BARTON as large units as the story (7, 111ff. and 544ff.) can be applied, in some degree at least, to the teaching of typewriting? It seems to have been taken for granted that learning to type- write, as a matter of acquiring motor skill, should take its start in the learning of the smaller units. There is nothing, up to the time of the writer’s reported experiment in 1921 (1), of an experimental or discursive kind to indicate that the nature and the size of the unit with which to begin this type of learning is even problematical. Woodworth (1921) represents the general view rather fairly when he says: “In telegraphy and typewrit- ing, it is almost inevitable that the learner should start with the alphabet and proceed to gradually larger units” (36, p. 325). He goes on, leaving the implication that the process is just the opposite of that required in learning to read. He says further: ‘“ But in learning to talk, or to read, the process goes the other way. ‘The child understands spoken words and phrases before breaking them up into their elementary vocal sounds; and he can better be taught to read by beginning with whole words, or even with whole sentences, than by first learning the alphabet and laboriously spelling out the words. In short, the learning process often takes its start with the higher units, and reaches the smaller elements only for the purpose of more precise control” (36, pp. 325-326). The “ piecemeal method,” or the method of exercising on isolated or meaningless letter combinations, such as the first exer- cises proposed by J. S. Curry in 1911 (9), seems to have been taken for granted. This, and similar notions, came into vogue as a result of the introduction of the touch system of manipulating the machine. As long as the sight method was resorted to the procedure was that of either copying complete composition mate- rial or freely composing as the typewriting progressed. But the newer and better method of manipulating the machine brought with it the notion that the mastery of the keyboard and the facility of operation could best be accomplished by first learning the smaller units and then progressively larger ones, until the complex units of the entire theme seem to function rather defi- nitely. This was done as a means of marshalling the letter, word, COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 3 phrase, and sentence units into a unit of comprehension, or a “hierarchy of psychophysical habits”? (3, p. 91), which is the objective in the process of acquiring this skill. Earlier Studies—In 1903, E. J. Swift (31) submitted the learning of typewriting to experimental conditions, using himself as a subject and employing the sight method. He purposed to determine the significance of periods of no progress as shown by the plateaus found in curves of learning. He had already had a little practice in manipulating a machine, for he says that he “ had never used any kind of typewriter except to slowly finger out about a dozen short business letters two years before. It is doubtful if the number of words in all these letters exceeded five hundred” (31, p. 296). In this previous training, wherein he resorted to the larger unit procedure, there was clearly no attempt at a consideration of the size of unit to use in acquir- ing this skill. One usually resorts to the larger unit plan when learning typewriting by the sight method. He takes note of other factors in the learning, such as the initial rise in the curve, irregularities in the shape of the curve from day to day, the “ effect’ of feelings of pleasure or discour- agement, the differences in difficulty in the material used as copy in his practice, a comparison of his results with those found in other learning experiments. He approaches the matter with the idea that in learning to typewrite one must go from the letters to the words, and finally to the larger units. He indicates at another point in the study that possibly these factors might be made use of together. He says: “ The several constituent factors that contribute to the acquisition of skill in typewriting are evidently operative together, though seemingly with varying degrees of prominence at different stages in the process”’ (31, p. 299). He further shows that many of the factors that condition learning operate without ever appearing above the level of consciousness. In 1908, W. F. Book (3) made a very exhaustive study of the factors and conditions involved in the learning of typewriting. The purposes of the study were, as he states them: “(1) To obtain for each of the learners taking part in the study a practice or learning curve which should accurately represent his progress ; 4 J. W. BARTON and (2) to obtain from his self-observations and from objective records of his writing such data as would make possible the explanation of his curves” (3, p. 7). A very effective and complete record of all that the learners did while at the typewriter was made “ by means of electrical connections between the machine and three Deprez markers writ- ing upon a kymograph drum in such a way that everything the subject did on the machine was recorded” (3, p. 9). A switch key was attached to one of the markers and so controlled by the experimenter that the number of times the learner had to look at the copy as well as the actual amount of time spent in fixing it in memory was also recorded. Pulse records were taken as indicators of the amount of effort being put forth as well as of the degree of attention used. These were obtained during the entire time of service by means of tam- bours placed on the artery in front of the ear and connected in such a way as to inscribe a pulse curve on the drum. Careful introspections were obtained from each learner at the time of each exercise. The learners consisted of eleven subjects differing in ability in handling a machine, from those who knew nothing about type- writing to those who had attained that degree of efficiency required for winning the medal in speed and accuracy of per- formance at the expositions at both Buffalo and St. Louis. Both the sight method and the touch method were used in the learning, and the exercise material was of the comprehensive kind. Two exercise sentences were used so much that they were written from memory, this for the purposes of investigating the limits of learning. The study throughout is concerned with determining the factors conditioning learning, particularly as this process is shown in the nature of the learning curves. There is no attempt at making comparisons between the results of learning by means of units differing in size since all the learners were exercised on materials involving the more comprehensive ones. None of them was required to resort to the use of the meaningless letter combination exercise material that is so commonly found in most of the text- COMPREHENSIVE UNITS IN LEARNING TYPEWRITING k) books, or courses of lessons, so widely used since the touch method has been adopted. Book confined this study to analyzing the acquisition of skill in typewriting, by both the touch and the sight methods, for the purpose of getting a better understanding of learning as it is represented by the curves that he found. The stages of learn- ing—the letter association stage, the word association stage, and the expert stage—are investigated in an attempt to determine what conditions are necessary to produce the shift from one of these stages to each succeeding one. There is nothing in Book’s study to indicate that the nature of the material used for practice purposes could, in any way, influ- ence the facility with which skill in typewriting is accomplished. His whole object was to understand better the acquisition of skill. J. S. Curry (9) presented before the National Education Association in 1911 a discussion of the method known by his name. This method is well represented in its general nature by one of his early statements: ‘ It is an old as well as a true saying that it is a safe rule of education to begin with the easy and end with the difficult, to begin with the simple and end with the com- plex’ (9, p. 835). All of this means for him that the learning begins with the exercise of the more efficient index finger. While learning the keyboard the little finger of the left hand should be placed on a while exercising any other finger, while at the same time the little finger of the right hand is placed on the key for the semicolon. He recommends the memorizing of the twelve index finger let- ters first. After these letters are memorized, he proposes to have them write the following exercise as a first-finger preliminary: fofrfbfify repeated six times, shyyjnjujm six times, gfgrguvgtgb six times, hyjhnhmhyhm six times, and two other such groups of varying letter combinations. The second finger is exercised in the same way on letters assigned to it in many combinations. The third finger is introduced next and then comes the fourth. Mastery of these and many other such exercises, including those of every possible vowel-consonant combination, is made. This multiplicity of exercise on nonsense material is followed by the 6 J. W. BARTON writing of words, such as are found in the textbooks. “ Sentences for speed practice should be introduced as soon as the student’s knowledge of the keyboard will permit” (9, p. 838). The following quotations indicate the extent of the practice on this kind of exercise material: ‘“ This course thus far outlined will cover about six months in a high school and two months in a business college.’ . . . “ Practice matter should form the principal part of the work for the first year’ (9, p. 838). This shows the general attitude of touch typists concerning the kind of work to be done during the first few months of exercise. J. W. Ross, in his text (27, Preface), resorts to the isolated non- sense exercise material, but he differs from Curry in that he proposes letter combinations such as the following for the first exercises, “asdf jkl; ;lkj fdsa jkl; ;lkj fdsa jkl; lk]; ;fdsa,’ and contends that this line method “establishes an uninterrupted flow of mental direction coordinated with a corre- sponding smoothness in manual operation” (27). That is, up to this time those in the practical work of directing this kind of training are agreed on the matter as to the necessity of beginning with the simple and proceeding to the complex, but they differ as to whether mastery and facility were best accomplished by the respective “ finger exercise method’ or by means of “ combining the letters as found in the lines ” shown on the standard keyboard. In the experiment conducted by L. B. Hill, A. E. Rejall, and E. L. Thorndike in 1913 on practice in typewriting (13) the attempt was made to determine the effects of absence from prac- tice, physical condition of the learner, change from early to late months in the period of learning, on the rate and limits of 1m- provement as shown by the curve of learning. Again there is no attempt to evaluate the nature of the material used in the exercise required in learning. In 1916 F. L. Wells (35) performed an experiment in type- writing for the purpose of determining the effects of “ different times of the day and different work periods, and different tech- niques of operation, etc., in order to increase so far as possible the efficiency of the skilled operator” (35, p. 47). The subjects were two well trained and efficient typists, which A i i i i ee i ee cs COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 7 means that the study was little concerned with the matters involved in learning, except in so far as the determination of the kind of error, the relation of speed and accuracy, time of day as affecting performance and individual differences experienced for these two subjects would shed light on more typical learning cases. J. C. Chapman in a study on the learning curve in typewrit- ing (6), which was reported in 1919, took as the objects of his study, “ (1) to obtain evidence with a view to a systematic study of the psychology of skill from the objective standpoint, (2) to investigate the rate of improvement of the subjects, in this par- ticular school, using this particular method of instruction, in order to establish norms for the comparison of the relative values of different methods of investigation”’ (6, p. 252). The method used was that of having the learners begin with the smaller units. On this point he says: ‘In the touch system certain elementary habit groups have to be formed before the individual can attain a positive score at all” (6, p. 255). Owing to the conditions here expressed, he indicates that the curves of learning show nothing for the early practice, “ for in the touch method, the attempt is first made to familiarize the learner with particular positions of the keyboard” (6, p. 255), in which case nothing of what the writing is like in the end can be obtained. He points out that typewriting under high school and business college conditions lends itself well to experimentation and pro- vides a situation that conforms to all the demands of scientific procedure. He says on this point: “ Experiments on large groups found in such schools, working for long periods under normal circumstances, satisfy the conditions that have been laid down and give evidence concerning the general nature of changes in the rate of improvement in this complex function”’ (6, p. 252). He does not think, however, that they are equal in exactness to the results obtained under laboratory control, but adds that they are sufficiently under control to be serviceable (6, p. 252). R. E. Hoke presented a study in 1922 of which the purposes were, first, to determine the frequency with which the various letters of the alphabet and the more common marks of punctuation are used in the English language; second, to find out the errors 8 J. W. BARTON in typewriting and to ascertain their causes; third, to determine the relative abilities of the two hands and the various fingers; and, fourth, by these means to determine the hand-finger loads with the thought of proposing such changes in the present key- board as will more closely conform to the relative efficiencies of the two hands and of the respective fingers in speed and accuracy of performance. These are all the studies on typewriting that seem available at this time. Hoke in the most recent study in this field reports what has been done as follows: ‘“ With the exception of the work by Book on The Psychology of Skill, little or no experimentation has been done in typewriting. Even in Book’s study, the object was not so much the improvement of speed, accuracy or methods of teaching typewriting, as it was the more general aim of ascer- taining, by the use of the typewriter as a mere bit of apparatus, the psychology of skill”? (14). From what is here presented of previous studies in typewriting, it is very evident that the field of this investigation—the size and the nature of the umt to use as exercise material—has not been cultivated; and except for the recent study of Hoke, nothing has been done to determine the effects of repetition as a factor in the determination of skill in the manipulation of a typewriter keyboard. All of the experimentation heretofore presented on typewriting gives nothing concerning the object of this study. The field of investigation here attempted is experimentally untouched except for the incomplete study by the writer previously referred to. All discursive matter as well as all of the attempts at experimenta- tion concerning the accomplishment of this skill, since the adoption of the touch method, seem to have just taken for granted that the formal drill approach on meaningless letter combination is the correct one. Practical Workers.—There seems, however, to be a recent awakening among the practical workers in this field of instruction. Many of the texts are so arranged as to leave out much of the meaningless jargon as exercise material. J. W. Ross in his 1921 text (28, Preface) makes the following comments: ‘‘ While the eee) ee ee - i Oe eS ee, COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 9 beginning lessons of this text are based on the same fundamental principles, they have been revised and improved in the following particulars: The selection of words in each exercise is based upon frequent sequences. The writing combinations that seldom or never occur in words is reduced to a minimum, such combinations being used only to a very limited extent to impress the student with the relative positions of certain letters operated by the same BaPermas ty Geni sete. bin ony Lhevearlyountroductionvand continued use of sentences and phrases composed of frequently used words gives a variety of work that relieves the drudgery ordinarily connected with keyboard drills.” Mr. W. E. Harned (12, Preface) in his text of 1923 states: ‘In teaching touch typewriting it has been found that the method of. repeating the same word over and over again is not only monotonous but a waste of time as well. The method of keyboard presentation set forth in ‘ Typewriting Studies’ is (1) to teach key location by association with the guide keys, and (2) to use words of frequent occurrence, grouped from the beginning of the course in phrases and sentences.” C. E. Birch in commenting on what he calls the vocabulary method says: “ Memorizing the keyboard is accomplished in a perfectly natural manner without the use of such meaningless jargon as asdfghjkl; or qwertyuop. In their stead is introduced a most carefully graded series of simple phrases and sentences which develop skill far more surely and rapidly. The fingers are trained to find the right keys automatically, with no dependence on artificial mnemonic methods ”’ (2, Introduction). The text that deviates most from the formal drill idea, in what it presents for the first lessons of the learner, is the one issued by Ollie Depew in 1921. She says: “‘ The subject is pre- sented by the word and sentence method instead of by the ‘A B C’ method. Sentence writing is substituted for the harmful practice of word writing. . . . Finger gymnastics like waszyq, and all other unusual combinations, have been omitted. Such exer- cises are useless for all practical purposes. An unnatural com- bination of letters will interrupt the rhythm of even a skilled writer. For the beginner such practice requires additional and 10 J. W. BARTON needless effort when he should be concentrating on practical material ’’ (10, Preface). Purpose.—lIt is the object of this study to determine by experi- mentation, under conditions as near to those of schoolroom prac- tice as possible, the relative effectiveness of teaching this subject by the synthetic process of beginning with isolated and meaning- less letter-symbol combinations and gradually going from these to the more simple words, phrases and sentences, and finally to the use of the business letter and other composition forms that are most used in the practical typing work; or by the comprehen- sive process of beginning and continuing the practice on complete composition material that 1s most like that which the typist will be required to use when he enters upon practical service. This study will, therefore, be a test of the validity of the statement often quoted but seldom practiced; “one should always begin by doing a thing as nearly as possible in the way it is eventually to be done”’ (15, pp. 66-67). COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 11 Il. THE METHOD First ExPERIMENT Subjects—The subjects of the first experiment for group W (the group using the larger units from the start) ranged in age from fifteen to twenty-one years. All were of high school stand- ing and were distributed in the classes as follows: 2 in the fourth class ; 3 in the third; 6 in the second; and 2 in the first year class. Group P ranged in age from fourteen to twenty-two years. The members from the respective high school classes making up this group were: fourth year, 2; third year, 4; second year, 7; and first year, 2. The learners in group P were selected by the process of regular registration (Sept., 1918) in a high school where typewriting was an elective study. No student was required to take this course. All who registered for the typewriting in this group did so because they voluntarily selected it. At the time of registra- tion nothing was said about another class to be started later. Not even the teacher in charge had any idea that such would be done. On November the 20th notice was given that there would be another class started in typewriting. Eight students reported that they wanted to begin the work, and three of them began the practice on November 21. The rest of the eight were forced, as a result of a limited number of typewriting machines, to begin at dates ranging from November 22 to November 28. Later, three others, whose records are included in the study, were registered for this work, (Dec. 8, Jan. 3, Jan. 7, respectively). Three others attempted the work but the records were not included on account of their not being in training long enough to be of _ value in making comparisons. All the subjects of this early experiment were directed by the same teacher, whose scholastic qualifications consisted of a high school course, two full years of college work, and graduation from a standard business college. She was without previous teaching experience. The only difference in the teaching of the 12 J. W. BARTON two groups was in the kind of material used for practice purposes and such other items of instruction and checking as were incidental to the two conditions of exercise. The members of group W were required to make a drawing of the keyboard on newsprint paper as a means of learning the keyboard. Aside from these differences the two groups were handled alike as far as seemed possible. The general methods of instruction used were those con- sidered best by such a teacher as described above on just leaving a good business college. Exercise Material.—Group P used the Remington Text (20) as the source of exercise material, taking the lessons in consecutive order as they are there presented. The early lessons consisted of the letters of the alphabet, marks of punctuation, and other sym- bols represented on the standard Remington and Underwood machines, all of these being put in meaningless combinations. The lessons became gradually more and more comprehensive until full composition was reached. The first period was used in explain- ing to the pupils, and directing them in such matters as the manipulations of the carriage, inserting paper, marginal stops and release, line spacing, etc. The first exercise in letter striking is made up of such letter combinations as the following: asdfg slkjh asdfg ;lkjh asdfg ;lkjh asdfg ,lkjh asdfg ,lkjh asdfg. Lesson V makes use of asdfgertcvb ;lkjhiuyn asdfgertcvb repeated six times for each line and continued for twenty to thirty lines. The self-constructed charts, with the large one on the wall of which the students of both groups made use, were marked off by means of very distinct lines to indicate the respective fingers for the various sets of keys. This, together with instructions con- cerning the function of the different parts of the machine, was accomplished in the first period of forty minutes. At the opening of the second practice period the pupils of group W were requested to follow the instructions written on the board, which read as follows: ‘“‘ Write a letter to Sears and Roebuck ordering a pair of shoes which should cost $7 and should be sent by parcel post.” This letter was repeated (never copied at this stage of the learning) until the pupil was tired of COMPREHENSIVE UNITS IN LEARNING TYPEWRITING | 13 it, when he was directed to write a letter to his friend inviting him to an evening party. All this early composition material was freely constructed by the learner while manipulating the machine. Never was he permitted to do copy work before he had memorized the keyboard. This self-composed exercise material was used by group W until each student felt sure that he could properly produce the desired characters without the aid of the keyboard charts. From this point on in the learning, each student was required to do copy work. The copy work consisted of reading matter, taken for the most part from the Red Cross magazine materials of that time, although many of the students used their history texts or other school books as copy material, believing that by so doing they could get some help in the preparation of these lessons while doing the typewriting. There was no drill of any kind by this group on other than complete composition material. None of the students of either group, except subject 26, had access to a machine at times other than during the forty minute period per day for five days per week during the progress of the experiment. The self-constructed keyboard charts were carried home with them and were used by some, in memorizing the keyboard. Most all the material produced by the learners of group W was kept. The speed tests were the only exception to this rule. Here, as with group P, the time spent in practice and the records of the tests for speed and accuracy of performance were kept; but not until after they had reached the point in learning when all were working on comprehensive unit material. This means that for the first experiment complete records of all that was done, in practice, by the learners were kept only for group W. The nature of the exercise material used in training group P can be deter- mined by consulting the Touch Method Typewriter Instructor for the Remington Typewriter 1903. The directions for the use of this text were followed by this group, for which reason it was not thought necessary to get a record of the work before the learners reached the part of the text that makes use of material comparable with that of the other group. 14 J. W. BARTON The results of this first experiment seemed to justify further investigation. ‘They show big differences between the effective- ness of the two methods in the acquisition of this skill, but they were produced by the use of somewhat unsatisfactory methods. These conditions resulted in the second experiment. SECOND EXPERIMENT Subjects —On January 10th an extra class for typewriting was started at the University of Idaho. This was well along toward the end of the first semester of the school year 1922-1923. Regu- lar courses are conducted at this institution in this work, but arrangements could not be made to make experimental use of the classes regularly registered. The work of the experiment was, however, conducted in the regular typewriting room with the equipment used by the students regularly registered in such courses. Thirty-seven students reported for work at the first meeting of the class. Each one wrote his name on a slip of paper and handed it to the one in charge. By a chance method nineteen were assigned to the larger unit group (known hereafter as group Wa) and eighteen were assigned to the smaller unit group (hereafter designated as group Pa). On account of conditions making it impossible to get more than one period per day and also because there were only twenty-four to twenty-six machines available, it was thought necessary to handle the two groups separately and on alternate days of the week. This procedure was followed until some of the students sensed the fact that each night there were machines that were not in use, and asked that they be permitted to come more often as long as there were available machines. This was done with the understanding that no student should be deprived of his practice at his regularly assigned period. Under such an arrangement it turned out that after the first two weeks there were very few times when learners from both groups were not present at each period. This made the conditions of instruction for the two groups almost, if not quite, identical. The amount of time spent by each learner is definitely indicated in Table V below. COMPREHENSIVE UNITS IN LEARNING TYPEWRITING | 15 Of the eight males and eleven females that made up the per- sonnel of the initial group Wa, four males discontinued. They were in the work only a day or two, for which reason their records are not included. Three others were later taken into this group, one on February 5th, one on February 20th, and another a day or two later. Of these three, only one remained long enough to be included in the used data. There were, then, in this group sixteen whose records are used. | In group Pa there were eighteen who began the work at the first session. Five of these were males and thirteen were females. Of the eighteen beginning this work seven discontinued so early that their records could not be considered. Three others were added later (February 19, 23, 27). The records of two of these are included. This leaves thirteen in group Pa whose records were sufficiently complete to be included. The intelligence of each of the learners of this experiment was determined by the use of the Otis Group Test. The results of this test, the school record, the sex, and the college rank are all presented in Table V below. Exercise Material—After the first period, devoted to chart making and getting acquainted with the machine, group Wa was given the same instructions as were used for group W: “ Write a letter to Sears and Roebuck ordering a pair of shoes,” etc. In no case were they put at copy work at this stage in the learning; but were required to recompose this and other letters until the keyboard had been memorized to the point of not needing the chart as a means of locating the keys. The period of memorizing the keyboard was different in length for the various learners, but none of them required more than six periods of 55 minutes each. The large chart on the wall was marked off in such a way as to indicate the proper keys for the various fingers. This chart was kept on the wall directly in front of the class throughout the course of the experiment. As soon as each student had learned the keyboard (at the end of about the fourth day) he was required to do copy work, using The Vocabulary Method text by C. E. Birch. It is claimed by the author of this text that the correspondence therein found, and 16 J. W. BARTON used exclusively as the exercise material in training group Wa, “is devoted exclusively to mastering the one thousand commonest words in the English language” (2, p. 25). The isolated word exercises found in this text were never used by any of the mem- bers of this group, but they confined themselves to writing only the full composition material included in the letter forms beginning on page 26 of this text. After the preliminary of making a keyboard chart and getting a general understanding of how to shift the carriage, insert the paper, back space, etc., the learners of group Pa were required to begin work by making a study of the directions to the student given by J. W. Ross in his text of 1914, page IV. As soon as this was completed, Exercise I was assigned. After this each lesson and exercise was taken in the order presented in the text. The instructions given by the text were followed as closely as possible except that sometimes the student was permitted to take the next lesson before he had succeeded in getting the exact degree of perfection indicated by the directions. The directions of the text were deviated from only when it was thought necessary to keep the student from discouragement and from dropping the work, This text was in use at the time in the University of Idaho. For this reason, together with the fact that the book represents the kind of material generally used in the teaching of typewriting up to 1921, this particular text was used for this part of the work. It should be kept in mind that it is the line method that is used by this text to introduce the letters into the practice of the learner. In Ross’ Lessons in Touch Typewriting (1914) the lessons and exercises are so arranged that the complete keyboard is not made use of until the fifteenth lesson and the ninety-second exer- cise is reached. The first letters used are: asdf for the left hand and 7kl; for the right, in various combinations. The next two letters introduced are r and u. These are put into combination with those already exercised in both meaningless and meaningful ways. The next element brought into use is the period, then the letters e andz. At this point in this text we reach the end of the COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 17 sixteenth exercise and the third lesson. Some idea of the amount of work necessary to reach this point by the learner can be obtained when it is known that these early exercises consist of five lines, each containing about 48 letters and 12 spaces, each line to be written six times. If high standards of accuracy are main- tained in conducting the work, the learner will be required to write the exercises many more times than is provided for by the text. Such extra standards are held to by many teachers in this work. When learners showed by their work that they were, in our experiment, making an excess of errors, they were required to meet an arbitrary standard (not more than four errors to the page) of perfection before going on to the next exercise. This standard is not provided for by the text used. General Control.—The teaching of the two groups, in this experiment, was done by the writer and by a college student of junior standing. The writer conducted all the work up to March 23, 1923, when he was forced to leave the work in other hands. As far as the comparative treatment of the two groups is con- cerned, it was alike, since they were always together; 1.¢., at every session, after the first two weeks, both groups were present, thus making the matter of teaching almost completely identical for the two groups. Very little teaching was done aside from urging subjects to speed up, to make as few errors as possible, to get ready for the speed tests, and to hand in all papers with the date and the learner’s name on them. If any one of the learners was seen at any time resorting to improper positions or habits of response, he was corrected individually. The touch method was used throughout by all groups. In no case were the results included if the learner ever resorted to the sight method in writing the speed tests. Very little of one sub- ject’s material could be included because of her unwillingness to learn by the touch system. The periods of exercise for the 1923 groups were from 7 :00 to 7:55 p.m. during each day of the week except Sunday. For the 1918-1919 groups the periods were scattered throughout the school day for five days per week. 18 J. W. BARTON Data.—Everything that was written by the subjects of groups Pa and Wa was made in duplicate and the original copy was handed in to the experimenter. This material consists of all the daily exercise material and the results produced in the speed tests for each student included in the study. Before handing in the sheets each time, the learner was instructed to write the date, the amount of time spent, and his name on them. They were also asked to keep a record of the number of words attempted and of the errors made. These were to be used by the students in writing up an experiment in learning as a partial requirement in a course in psychology; 1.e., all the subjects in both groups were doing the work for purposes of getting credit for it in psychology. The speed tests for groups Pa and Wa were checked for the number of words and errors, first by the learner (each checking his own paper), and then verified by the experimenter and his assistants (two college students) while working under his imme- diate supervision. Then the number of words, and the number of letters written, as well as the number and letter placement of errors, were determined by actual count. Nothing of the daily practice data of groups P and W is included in this study except the amount of time spent and the other general matters incident to carrying on the work. For Pa and Wa the daily exercise material is made use of. The number of times each letter was used by group Pa was determined by first actually counting the letters found in each of the exercises of the text used and then multiplying these numbers by the number of times each student wrote these exercises. For group Wa the number of words written by each student was determined by actual count. These numbers were multiplied by 5.433, the aver- age number of letters per word for the 1,000 most common words (14, p. 12). Then by the use of the percentages, for the various letters of the alphabet used in English writing, determined by Dr. Hoke in his 1923 study (14, p. 12), the number of times each letter was used was calculated. COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 19 Lge DATA First EXPERIMENT General Statement.——The material and results presented in Table I and Table II are all that are now available for this part of this study. For the most part, only such records as are used by teachers in reporting the progress of students doing such work as typewriting were used. A strict record of the number of practice periods was kept, however, for each student. Each of these periods was of 40 minutes’ duration. The first speed test was given January 3rd. This was after group W had been in training for five weeks and after group P had been using the full composition as practice material for a period of two weeks. The speed tests for these two groups were always of ten minutes’ duration. They were distributed at intervals of three weeks during the first twelve weeks, and at two weeks’ interval thereafter. It should be kept in mind that group P began training eleven weeks earlier than any of the learners in group W. The reader will note that Table II presents only words per minute for the respective measures. The reason for this is that it was the common practice for typewriting work to be evaluated in terms of a combination of words attempted and errors made. The teacher in charge used this method of evaluating and record- ing. The figures presented in this table indicate the number of words attempted by each student after ten words had been sub- tracted for each error made. In this way nothing is certain, as far as this table shows, of just how many words were written or how many errors were made by each learner. Results and Discussion.—In Table I are presented the results and material of the first speed test (January 3rd) in another form. Here we have the actual number of words attempted and the number of errors made per minute for each learner. When these are handled in the same way that the other speed records were 20 J. W. BARTON handled, they give the data shown in the second column of Table II. These two groups of learners were not very different in ability or fitness for the work of typewriting, if school marks and school rank are dependable criteria to rely upon. The school marks may be converted into figures and the distances between the sev- eral degrees may be regarded as equal. To the first decimal place there is no difference in the averages of the school marks for the TABLE I SHOWING THE COMPARATIVE ABILITIES OF THE SUBJECTS OF Group P AND Group W AND THE GENERAL RESULTS OF TYPING PRopUCTION oF EACH ro} Me, gm go 8 ee 2 ON: ae Sed GHA Beha ties rs | oO A, se ar n.° 4 13) B 2 R Ehentan a a Se a aie 3 3 3 3 —3% we 3g hc Ye hi ep S a = s sia On yh pk =. & n n n n Has 38 < 3. “2a & 8, (x Group P 1 f B 3 yr 76 53.1 1192 70 168 49.4 He f B 4yr 81 62.2 13:8 77 166 5502 5 f B 2 yr 70 53.0 14.2 76 169 26.3 4 m C 2yr 80 47.1 8.2 5] 170 37k 5 f A 4yr 82 60.2 10.2 73 168 56.8 6 f iS 3 yr 73 PAG 10.7 71 169 39.4 7 m B lyr 81 50.2 9.1 62 170 46.6 8 m se lyr ri 48.2 11.3 66 169 43.0 9 i C 2yr 67 50.2 14.2 75 162 20.6 10 f B 2 yr 68 54.3 1Ze 80 163 39.5 11 f B 3 yr 75 58.4 9.5 78 167 40.6 12 f CG 2 yr 74 55a) 10.9 74 165 43.2 13 f D 2yr 75 52.2 Ling 70 165 34.6 14 f C 2 yr 76 43.3 14.3 iS7. 167 30.3 15 f B 3yr th 58.4 10.6 .76 168 51.8 Average 45 ZOU moe Dae ne 11.4 70 167 41.0 Group W 16 " ( 2-yy. 18 66.0 D 2eas 7 109 30.0 17 f is 2 yr. 21 68.1 6.1 3.2 111 48.0 18 f G 2 yr. 4 itt i ety £ih 34.4 19 f B 4 yr. 21 48.1 8.3 ray 68 Pee. 20 f B nye 23 80.2 6.4 Bh 114 70.2 21 f B 3 yr. 21 70.1 6.3 ao 110 Sine 22 f B 3 yr. 22 Pipes G25 a Bes 112 46.0 23 f A 4 yr. phy Tite 6.1 3.4 113 59.0 24 f e 2 yr. om ve ae 33.4 25 f C 3 yr. 14 51.4 10.2 327 98 39.8 26 r » 2 yr. 24 63.5 O2ZiNN226 104 28.8 Average 45 Z26yr. 20.8 66.3 7.4 A 04 47.4 COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 21 two groups and the P.E. of the difference of the average is .07.' In school rank, group W has the slight advantage of .3 of a year with a P.E. of the average difference of .0197. All subjects of both groups were without previous experience in the manipulation of a typewriter. By consulting Table I and Table III it is seen that group P had been in training for an average of 75.2 practice periods, while 50 30 paca Sees eal eT ora oe tenn 10 -10 / / / / ae es a ‘ / / re] / » je % 5 3 t -90 Pay, 2 . { z PCa é / , / —-——--——_ Group W zs Group P ~140 es / gee ne i “a / -190 “ai - / ,, : 0 5 10 15 20 25 50 BC ¥sexzs Figure IJ. Showing the average number of words written by group P and by group W after they have been penalized ten words for each error. group W had been in training 20.8 periods. This is an average difference of 54.4 periods with a P.E. of this difference of 1.018. This means that group P had trained 3.6 times as long as group W. Column 6 of Table I gives the number of words written per minute, by each learner, on January 3rd. These data were col- lected on the subjects of the two groups on the same day, in the same class, and on the same exercise material. In fact all speed tests were conducted alike in all these particulars for the two 1 The letters shown in Table I and column 3 were assigned numerical values of A=6, B=5, C=4, D=3 in calculating the difference. Any other values might be used. The steps are assumed to be equal. ee J. W. BARTON groups. In the number of words written we have an average difference of 13.1 words per minute with a P.E. of this average for the two groups of 2.439. The advantage is with group W even though it had been in training for the shorter period of time. If we look again at the same table with a view to finding what the conditions were with respect to the accuracy with which the words were written, we see that group P made an average of 11.4 errors per minute, while group W averaged only 7.36 errors. TABLE II SHOWING THE NUMBER OF WorpDS IN EACH OF THE SPEED TESTS FOR GROUPS P ann W Arter HAvING BEEN PENALIZED TEN Worps FoR EAcuH Error Group P Words per minute in tests Subjects 1 2 3 4 5 6 7 8 9 APOE —56.9. 27.4 31.0 29.2. 28.6 37.4 40.8 39.9 (4974 go OE. —69.8 48.3 45.8 49.8 47.4 49.6 52.7 49.3 55.2 A Dae — $9.0 >.21458: 2232 16.28 17.6 4 21s 3-9 ee eee Dee ae —dhiD. 52226 25:6 2324 1B 22 25-8 eZ Ae a0 Gee Sch ure —41.8 44.1 49.0 43.2 41.0 47.0 47.2 42.0 56.8 Gta hat —54.9: 33.4 5.36.6. 374. 25.2@ 128.65 735.6 230 Sueno fie he wag —40.8° 34.9: ~ 3516. - 38:4". 36:8: 37.4 (45.65 935.67) 46a Ene: + —64,8. 3353 ~ 29.8) 2985 ZA Zee i237. 2c eee Peay — 9158 13.3. 1585. 2A 7813 3S 1S 8 Ge Tees re — 66:7 289°" 3632! 31, Oa ZO a2 62 ae ee | 6 Care ee —36.6:° 333.1 . 3316. aie 25.0 7 33.1 45:02 Saker Sa 1A eee ——53 9° 30.2.5 214 (ZEA W202 35.9 B29 = 252 (ae Bors Pe 69° 8") 20.9) 27.2) 22352) APG 213 eee, “SO ree jC eter: = 9087 ZAP egies 2S eee 2362 926. Se ZA Ze 28 ee Teas. maty.6 34.9 45825 8372 he 32 sl 7 ae Average. ©°’—60:6° ©. 29.5°. 32.3: 73038 — 26:9. 30.5: “34:6) 3321948 Group W Words per minute in tests Subjects 1 Z 3 4 5 6 7 8 9 L6ie: ees: 2h 91 91852 ye 2 Oe 1B Sed Oe ABTS a, 4 30.0 if ZEN P86 SS OCD eh es eA) Cea oe 48 .0 1B Pe a6 NZULA OTILO Ae Oe 23 tO oe eee 34.4 Lee —34.9 This student would not use touch system Ha 6 bdeenlbias 16:2 50.4. 69.4 °45.0 360-1" | 56:0" "66.0 55.9. ee 4 Pe eee 4.1 OBR 272458 306 6: 32152 29:04 3229. 24329 eee PALE ee Tle O2.6 0) 32) 0. 42-420 Oy 2-0 2 a On oes ane os Pa ae IG i2 VS. 0 82.2 Se SAGO) eee oer re Yer eee. ie Ph ctr at Ages | hae cack Sei oaglr i diame peck ear me Miere ek Poel Sips 5D 5 60093956: 43510 om 34.4 Bn 38.2.1 925.5 44-2 oe, ee “, aur 20:4 517.2) 26,00 A902 dS SA 21a 33 oA oa Average. —6.2 ° 31.9 35.7 32:1 °° 34.6 °34.6 42.0 °47:8 “7a i a ae ee te ee eT a ee 5 aint >. >! oe COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 23 This is a difference of 4.04 errors per minute, and the P.E. of the difference is 1.091. If we divide the number of words written in the first speed test by the number of periods of practice up to that point in the experiment, we obtain the average number of words gained per practice period for each group. The difference in this for the two groups is 2.496 words per practice period, with a P.E. of this difference of .102. This indicates that for the time spent TABLE III SHOWING P.E. or Group DIFFERENCES BETWEEN THE AVERAGES AND THE STANDARD DEVIATIONS OF THE DISTRIBUTIONS FOR TABLE I Se vo © a < Be ma} cP) Am a iggae 28 aise Eee coe es , 3 0 & $i a ne 3) bh E fe ia ae ee ee oat £ 5 a 2 an nes rier om) ei) * Fite Mads een ie nue hae a Se AE ie RE OE ig Og a a Paverages croup Pt h4s51 012-3 75924 S972). 11.4 .70 167.1 41.0 Averages group W.. 4.5 2.6 20.8 66.3 7.4 3.20 104.3 47.4 Difference in averages for groups P& W 0.0 for 4 54,4 13 ot 4.0 2.50 62.8 6.4 erie aati tay. O77) ee) 4602: = 2.447 31-09) OF i Bis 8 Cities Se (8 Standard deviations of distribution: apps Ta Rs tie ty 5 ie ol | 4.5 4.98 1.8 .08 a 0) Wirreee eereete 2 eo ae cert mers ice.. 9 Vest) PGi rr 1548) 45h 13.69) T3383 group P made on the average a gain of .704 words per practice period, while group W made 3.2 words. It should be kept in mind that up to this point (January 3rd) group P had been in training an average of sixteen weeks, four- teen of which were on the small, or meaningless, unit material and two weeks on the complete composition material; while group W had been training five weeks on complete composition material from the beginning. The reason for presenting a cross section of the material at this point is to get a view of what the conditions were at a very early point in the learning after the learners had been training on material that was alike for the two groups. Up to the time that group P began on the complete com- position material the two groups were working on different prac- tice material, but at that point they were given the same text and the material was identical for all thereafter. In this way it was 24 J. W. BARTON thought possible to determine better at what point in the line of training the greatest differences were to be found. It was thought that this might help to indicate more clearly the causes for the differences shown. Some questions must have been raised in the mind of the reader concerning the causes of the differences noted between the two groups. One such question might be: To what extent did the learners of group W work with greater determination and zeal than in the case of group P? Another might be: Why are the records of some of the learners in group W not included? Other questions might have arisen. The answer to the first one cannot be arrived at by such means as can be objectively presented; but since the members of group W began later it might be true that they did exert themselves more than usual in the ordinary practice, which was preparatory to the speed tests in which all of both groups participated with the keenest competition. However, if either group did work with keener interest, it was not observable to the teacher or the experimenter at the time. It may be possible that the motivating factors were neither as many nor as good during the time that group P was using the smaller unit material, since there were no speed tests given during this time. It is a fact that most of these learners manifested “ennui’”’ while the exercise material was of the meaningless kind. Group W also had no speed tests during the first five weeks of their work, but they showed nothing like the same degree of lack of interest that was shown by the other group. Possibly the kind of material used 1s a factor in determining what the attitude of the learner will be in typewriting, as is doubtless true also for many of the higher forms of learning. As to the second question, it might be suggested that since the records of some of the learners were not included, the selection for typewriting skill was higher for group W. Subject 19 resorted to the sight method in her practice work so continuously that when she came to the speed tests, where she was watched by the instructor, she could not manipulate the machine any better than a beginner and would simply sit idle for most of the test period and not even try to write. She was a senior in high school but OO a ee COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 25 would not follow the instructions. In this way she really had no record that would be fair to include. Case 24 was irregular at school and was not present at any of the speed tests except the last one, which left no other record that we could make use of for comparative purposes. In the single test that she took her record is relatively very good when thought of in terms of the amount of time that she spent in practice. If we turn now to Table II and Figure 1, and attempt to deter- mine what the difference is between the two groups after the TABLE IV SHOWING P.E. or Group DIFFERENCES BETWEEN THE AVERAGES AND THE STANDARD DEVIATIONS OF THE DISTRIBUTIONS FOR TABLE II Tests 1 2 3 4 5 6 7 8 9 Averages group P —60.6 29.5 32.3 30.8 26.9 30.5 34.6 33.1 41.0 Averages group W —6.2 31.9 37.7 32.1 34.6 34.6 43.0 47.8 47.4 Dif. in averages for eroups* Band - W554 40 5 24s iA ee RF A Sa TAT 64 P.E. of dif. in av.. Katt bere A RG 8h aL SA, Biss Wt WAN UG ae Rel Ps Bak BPO Pe Standard deviations of distribution: Rot gee eee Do eo Adee Ae tee Ses OSS <6. ton OO th ett Lag Me, Sele URE hoes us A Ba And ad Og 6 OD ae es Bo A a cal first test, we see a very much more rapid rise in the curve of group P than in that of group W. In three weeks’ time group P went from a minus 60.6' to a positive 29.5. For the same period of time, group W made a gain of from minus 6.2 to 31.9. How is this difference in the rate of learning to be explained? Is it possible that the eleven weeks of formal drill might have influ- enced these results? Up to the point of the second test, group P used the same prac- tice material that was used by group W for the period of five weeks. During the four weeks next preceding the five weeks just mentioned, it must be remembered that the material used by group W was first self-composed business or friendship letters until the keyboard was well memorized. This was followed by 1 It seemed necessary to convert the first test scores to the same units that we have for the rest of the test records, by multiplying the number of errors by ten and subtracting this sum from the number of words written. This was made necessary because the remainder of the data had been so handled by the teacher in charge. 26 J. W. BARTON doing copy work from the current Red Cross magazines or other complete composition material most available to the student. If we turn again to the test scores we find that the first test gives an average difference of 54.4 words per minute, with a P.E. of 5.536 for this average. After this, there is not a single difference with a P.E. that is as much as one-fourth the difference in the averages, until we get to the eighth test. Here we have an average difference in favor of group W of 14.7, with a P.E. of the difference of 3.25. By consulting Table II it will be seen that for this test only six of the subjects in group W were present. Of the three who were absent, two were relatively low in typing skill. This would have a tendency to raise the average score for this test. All the remainder are thus statistically unreliable. But the averages in each of the tests show group W to be superior in typing skill. Since the curves do not cross at any point—.e., all the differences are in the same direction—we have an additional assurance of the reliability of our differences. These test scores and the curves represent the speed of typing for the individuals and the groups through twenty weeks to the end of the year without any great difference represented at the end of that time. This seems to show that the two groups were nearly equal in typing skill for a rather extended period of time and probably would remain so. Was, therefore, the first period of eleven weeks of practice by group P a waste of time for the most part? Was this waste due primarily to the nature of the material used for practice purposes in the training? Need of Further Experimentation.—These results, while obtained under methods of group selection and of tabulating data that are not free from scientific question, raised the problem in the mind of the writer as to the possibilities of the differences here shown being valid and explainable on the grounds of the kind of practice material used in the training of the learners. The whole problem grew up in the actual work of the writer while serving as superintendent of the school system at Elk River, Minnesota. At first there was no thought of running a control group of learn- ers. All of the teachers of this system were, at the time, meeting regularly for the purpose of discussing the various methods for a COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 27 the teaching of reading and other subjects, and there was brought up the question of the formal versus the incidental drill and also that of the unit size in learning. These discussions raised the question concerning the application to typewriting of some of the principles expressed in such quotations as the following: “One of the striking discoveries of experimental investigation is the very rapid progress in specific functions when the practicing is done on the particular connection to be established” (30, p. 159). When we set this idea over against the opposite one, ‘“ Habits can be established only through repetition ’”’ (26, p. 73), we have a very forceful problematical situation that seemed to warrant investigation. In the light of such work as that presented by Ebbinghaus (11), Carr (5), Cummings (8), Pechstein (21), and many others, on the size of the unit to use in learning, and that of J. C. Brown (4), Thorndike (32), Phillips (25), Kirkpatrick (17), and others, on the effect of drill, there semeed little help in an attempt to explain the results of this first experiment. The findings in these investigations and the results of this preliminary experiment in the teaching of typewriting indicated very plainly that the learning of this act had not been sufficiently investigated to warrant any really valid results; but suggested that what had been found to be true for the learning of arithmetic, spelling, reading, and other such subjects, might not hold in this case. Under these conditions a second experiment was attempted, again under practical school room conditions, but this time using better methods of selecting the subjects for the two groups, and more uniform conditions of work for all subjects. A complete record of all that was written by each learner was kept and was made use of in working up the results. The next two parts of this study will consist of the results of this second experiment and a statement of the conclusions based upon all the work of the two experiments. 28 J. W. BARTON SECOND EXPERIMENT General Statement.—An attempt is made in presenting the results of this experiment to show records of all the materials that were produced by the learners of the two groups. It should be remembered that five of the subjects of group Wa and eight of TABLE V SHOWING THE COMPARATIVE ABILITIES OF THE SUBJECTS OF Group Pa AND Group Wa AND THE GENERAL RESULTS OF TYPING PRODUCTION OF EACH S n ov 3 oi A o wn et he OP 2 ETS Tae ea iin hae Srsit- c aviere £ ¢ = e/a is oh cst & Uoa gs Siem 3 tis Cobh arf ent o, Hesq Sos MoS he £ g g i or attr ee S29 (808 B32 abs 28 We Oneonta She hae 'e Se rege S88 S88 ga WY vi VY) n =o Ha Hz HE 4288 B2E BSE Aes Group Pa OA aia CES Hine of 19S ThA 131 6285 Bh Ibs a7 Ora 6s 5 2.5 Patera ge 4.4 Fr. 137 043) 3220,9/20m o2Ud hoe ae 14 4.1 at kOe Pe paged oh re Bates Fain bo. URC Benes ty Ws phir a ok BEA 3 1.4 : JU et 4.9 Fr. 17017 68,946 SoGeatie oteebed + K S] owe | 4.5 Soph 132). 3560 287-227) Osea, OF 4 OU 4 bel | A YAH | Loe old she 178 34 $8,041 sol 7 Peat. 7 G0 7 4.4 ee DS | 4.5 Fr. 13 nee Lee oO0 569 4.6 116 A 6.0 . $4 mo AO es, 168.5512 208,183." 54,065 7A Zee 215 8 3.8 | Sot 5 Ora ire 1625253220 4b Se Ze eiO Ou oOs, 7 ae | oon it Re: SEI Ss oh pW Pawel tai Fee WE Se ors WG eyed shed ls 818, 2 0.6 eae Waite WY heal doe 148 317,205 921910 919321. 11,6 sar 188 Oe ee 38.5.0 .,,.4: 1, Sophie’, 4165 6554. 274022 1416. 35.9 4160 eee ZONA bede i Say 166.1 48°: 240197 7-2.625 "5.800241" Wee ene Average 4.7 1.38 162°. "40 177,91 1917 2 GIS) 2258 5 22a Group Wa 40 f 459° > SOONS 24a ol ea Ope wig JOT Es caer 19 3.8 41 f 514. Sophy 1525. AU seb OUon no, cece Ore eae 13 3.1 AES | SO ers 165 46 279,409 3,162 22.4 1,032 9 0.9 43° °m ~—4,0° > Fr; 154 36 80,731 Witt OO en Ae 4 ean 44 f 4.0 Fr. 124 59 323,366 4,016 13.4 791 /4 38 45 .f BiZaun Er. 165 50 234594 3019 9.8 518 s 0.8 AG 4.9 Fr. 178 29 90/314) 71,5267717.9 520 14 2ek 47 ff RP COPS ab 135. 47% 160563 4,768: 49:10 2423 6 1.4 48 f 330) vole Pr: 118 3143.2 2/5220 0 pO AO lee ed OU 6 0.8 cho ae oe Se SS 160 26 78,246 TAI Se OOF 5 1.4 50 et 4.62% Soph st 431i) co 2225, 85a ee) 759 40. GeO Z 1.4 st Ba yet eh Ss 154 37 140,467 2,414 15.9 590 13 2:6 52 semiun oA 2 Oe 168 30 71,482 1,017 14.9 448 4 0.9 ns fa ath 4-6-2 ily: 157 49 262,961 5,057 16.4 796 15 Ls 54. O44 8 Sophie 9116 B37 2106 Sia eS 7a ee 13 2.9 Sa tet eb Mea a 153 42 148318 4,278 13.8 581 i} 1.9 Average 4.6 1.81 147. 41... 169,219. 2,796 13.96 564.6. 11.3 ie COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 29 group Pa were in training for so short a time that the results could not be used. The two groups (Pa and Wa) were trained under as nearly identical conditions as was possible. This was done by having them meet in the same class for training purposes except during the first six sessions. During these practice periods all could not meet at the same time on account of too few machines, but at each practice period there were learners from both groups. The practice periods, it will be recalled, were 55 minutes in length and the speed test periods, of five minutes each, were TABLE VI SHowinc P.E.’s or Group DIFFERENCES BETWEEN THE AVERAGES, AND THE STANDARD DEVIATIONS OF THE DISTRIBUTIONS FOR TABLE V 63 : ivi w o 23.. 5 2 3 © ~~ + 1 n <2] te n Ar Cand CEG hate tosihe eee STS E.G Seve cui ois E ret MY > rl Hie San Sham 3? ° ° 50, omy. ant = Fav UVP us Os © ° ee BS} res SG “oe S98 YSz oe <3 3 vo o Sor Sa pie 23.4 95.8.5 wn ww Ase Ha BB BE ¢t¢PaR@secs aF Averages group Pa.... 4.74 1.38 161.8 40.0 177,911 1,917 6.3 237 Sane Averages group Wa... 4.58 1.56 147.1 40.9 169,218 2,795 13.9 564 11.3 1.9 Difference in averages for groups Pa & Wa. Ota LS 14.7 9 8,693 SIS op 76 Olea s 2 cae OL 0) .6 ee OL Olly ity avy vente .14 Ai 4.5 3.01 7° 183323 364 7 35 1.3 Sco) Standard deviations of distribution: Fla Wek ey. ear eee a ae 5 Die eek Oe O. 95ern 9-06-6101. 45008 L590 isch: lt o4e Os 9 LEE at bpd PA eas Pre PS EOF ee Lela OOO S00 21 SUS ao 84 393.3 1.6 included in this time. The speed tests were not begun until March 13, 1923. This was two months and two days after the beginning of the experiment. Subjects were taken in at different times during the course of the experiment, and because of this some of the learners began the tests with less training than others did. This is shown to some extent by the difference in the number of practice periods for the various subjects. In Table V it is shown that group Wa has a median of 41 with a range of from 26 to 59 practice periods of 55 minutes each. The P.E. of this distribution is 4.89, and the number of cases is 16. For group Pa there were 13 cases ranging in the number of practice periods from 17 to 56, a P.E. of the distribu- tion of 9.63, and a median of 43. Results and Discussion.—In Table V we present a general view of the materials and results of this experiment. Here we have 30 J. W. BARTON a statement of the nature of the various learners, the amount of time devoted to the work of learning, and the actual accomplish- ment for each learner. This table shows that in school marks and in school rank group Wa has a slight advantage. In school marks the average differ- ence is .16,' with a P.E. of .14. In school rank the average difference is .18 of a school year, with a P.E. of .21. Statisti- TABLE VII SHOWING THE P.E.’s or Group DIFFERENCES BETWEEN THE AVERAGES, AND THE STANDARD DEVIATIONS OF THE DISTRIBUTIONS FOR TABLES VIII anp IX Words—Speed Tests Tests 1 2 3 4 5 6 7 8 9 10 11 12 Averages group Pa 18.4 21.7 26.6 29.6 26.8 29.3 34.9 44.4 47.4 52.7 65.5 59.5 Averages group Wa 86.3 92.9 88.1 88.8 89.2 91.1 113.9 110.3 105.2 116.3 145.2 148.9 Dif. in averages for groups Pa and W Ava? oiten steieleteleis OLGA note DL woo os (Ooe4ie O12 857 9-0-9160 529 ut 4-8 103 sOmees oe ae ee PFE] Ob dite insav. wos! 50s 6.0 eee. 04 ie COLO mn 764 NOLS a Uae Z os 725: 9037 likes Standard deviations of distribution: Pal Beda ecececk con oe ks) BIO) 6B 8.627 8 99) 80.0. 4 49 ee ee Witte A kote se WO4. 0. eheLic 28.0Ne22.38004.4 nas oo. oleb SSL os Seer ad Uma dee Errors—Speed Tests Tests 1 2 3 4 5 6 7 8 9 10 11 12 Averages group Pa 7.9 7.0 e408? FOS 4 5.9 Se 70° 6:6" 7 555 Tet 6.1 Averages group Wa 12.0 14.7 10.9 12.7 12.4, 12.3 8:9 FAO wel zes 8.9. OL8 1EZ Dif. in averages for groups Pa and Waiaie: Sree omnes 1 TIRE ORS 5.2 7. Ou 654 353° 4 329 5578 sot eee ct eee P.E. of dif. in av.. 20 iron fe SaaS Lehme2:0 2 ool, 1S wiealss 1S) PORE Standard deviations of distribution: ay Oia esta REP OLD | ie CORO RE | oy ORES UR eo 1947 S345 OAM) OnSioe 73:2. 4 eo Warn wien Bete Bede GLO 25 Shs hae SO ir Fe ecaay anny. ce a cally these differences are unreliable, thus indicating that for these factors the groups are approximately equal. In the same table the results of the mental tests (Otis Group Form A) show an average score of 161.8 for group Pa, while group Wa made an average of 147.1. This is a difference of 14.7 + 4.47. The standard deviation of the distribution for Pa is 18.5 and for Wa it is 17.15, with a median of 165 and 153.5, respectively. This difference is rather high and closely approx- imates the limit of reliability. Just why group Wa should be so much lower in the mental test scores cannot be explained since the subjects were assigned to their respective groups by a chance ee unit in this case is A=6, B=5, C—4, D=3, with steps of equal value. COMPREHENSIVE UNITS IN LEARNING TYPEWRITING _ 31 method. The test (Otis group) was the same for all and was administered to both groups together. The correlation, by the rank difference formula, of the scores in the mental tests with number of letters gained per practice period, as shown by the final test score for both groups, is minus .36 + .16. In Table V we also see that group Pa practiced an average of 40.0 periods of 55 minutes each. Wa practiced 40.9 periods. This is a difference of .9 + 3.01. In the total number of letters written Pa made 177911.3 while Wa made 169218.6, thus show- ing a difference in favor of Pa of 8692.7 + 18823. In errors Pa made an average score of 1917.2 while Wa made 2795.6. The difference in average in this case is 878.4 + 364.1. Statistically these differences are all unreliable. They indicate that as far as the training of these two groups is concerned there is no certain difference, and if “ practice makes perfect ’” in typewriting, as Hoke points out, then we should expect, under like conditions of training, that they would perform equally well in typing skill. A better measure of skill in manipulating a typewriter is shown by what the operator can accomplish when put to the test in such work. In the final scores of the two groups, we have an average difference of 327.14 + 34.83 letters per five minutes between the tneans of the two groups in favor of group Wa. Statistically this is a reliable difference and the difference is a large one, since the average for group Pa is 237.4 and that for group Wa is 564.6. It will be noted that the per cent. of error is also less for group Wa in this test. The gain in letters per practice period is found by dividing the final test score by the number of practice periods. The column under the caption, ““Av. number letters per practice period at end,”’ gives this part of the data. Group Pa made an average score in letters per practice periods of 6.34, while Wa made 13.96. This is a difference of 7.62 + .72, a very substantial difference, which shows, on the average, that W”a made more than two times the rate of gain, in words written, than was made by Pa. In the actual number and the per cent. of errors made in the final test score the results are as follows: Group Pa averaged 32 J. W. BARTON 5.23 errors per five minutes of writing, while group Wa averaged 11.25, a difference of 6.02 with a P.E. of 1.26. When these error records are converted into per cent. on the basis of the num- ber of words written, we find that group Pa made an average per cent. of error in the last speed test of 2.53. Group Wa made in the same way a per cent of 1.93. This is a difference of .60, with a P.E. of the difference of 3.46. The first difference is TABLE VIII INDICATING THE NUMBER OF WorpDS WRITTEN, THE NUMBER OF Errors MADE, AND THE PER CENT. oF Error AGAINST NUMBER OF Worps WRITTEN FOR THE AVERAGE OF EACH OF THE Four CoMPposITE SPEED TEST SCORES Group Pa Scores in speed tests (5 minutes) Subjects 1 2 3 4 5 6 7 8 9 10 tT 12 HEM Bed Smee 2 15 022-00 B36 OseoO HUE cus. eee 5 eee a ae DAB care aye | Dpetecug Ag 10.0 750 5.0 SU retrain were AG Sige (OR A de sates iw eee ees ZEW eel PR: tele etl Pe 9155040505 37:50 Beh Ae 5350 0 273,10 So. ORO SS OGRE | ORE SAR Tin Sema nee LOO 9.5 O50 Peds edi Os ah oaO) ele oo. te Sa oon eno 2OIWoke aa 12.0 o1720 09523: 008 Fe 21 0282026312 Ole 2905 S325 66. 0Lm4o- 0 mnd Gee | De A es 8.0 520 1.0 ia 5x5 4.5 4.5 ae U Sine 0.0 0.0 SES SOAWis eee 17.0 92051620) Opn. ces EO SO LS) 5 PLO ye ANU) Ae ex eo OO BeBe BN S DEES oF .0 3.0 2 Sareea: SN 4.5 1.0 4.0 ese a5 wh. et aretacae SINWese wos SLO 33208 vee 345513800) 4656. 45) Oi 524505. ee oo OO TOO ened | age 5 30s 8.0 SO RAO 2. 54 0 22 eae eee Se Wille eats Sos s See AU SPAN OR UR TIX IES PV BRKT, Us ans eae aaa Chee ae eee Hee ee 4.5 rede. ep cut 8.0 5.0 Py 3.0 Sie RS ele Hoes hee aoa W tere ik 1985512720 ase e FeO we Ls Oe 4-10 ire cee nn ek eee ee Mes het wo [ ORES Ae 14.0 FE QU asi 4.0 6.0 Desire eas olnate eta ae aete ieee tee SAW a a wes 142.055 529.0405 250 t7 Ore 26.0824 Ou 25202 Oman 39 nOnnn4 Zins Geto nee aan eae 10.3 8.0 8:04 .12.0 O25 7.0 6.0 6.0 6... Ose 0 it or Oo AKOYA ee Bae 96.0 8035 c09 035.101 070 094 60.0 4100715050 798.000). Omens 2 OmnGon lamas see i Digi 125 Ose 4 OO TO ae SO eet OE. 9701030 Sa OSS 8.5 ype aires, 9.6 SOLW iron. 12:6 43207524 £0 F933 95 2450 25 Ores Sue 5 4. Gr ea oe Ol ce ee eee 1 Oe irae 6.6 5.0 el 6.6 ree 8.0 7.0 hwesr, | Rae S 5.0 8.0 6.0 SAW ee 820 808 51>) See e160 ZO EO" SL OL OF 270 eee eet ek eee eens ee | Deeb ihe yes 4.0 Pe Meee ie Se 4.0 36D 2.0 A: Oe ee 2.0 ATRIY MAA) 6.5 BSW eh acres 23:25: 32257034.6 (039.20: 7 29 0s 45200 43 20° STS 5220 60.0. 70 Sa ere | Ra ee vests) Ge a G33 8.0 6.0 6. 5ert0 Ole 9.0 4.0 2:0 2a5 SON Wietat gor 14.0 °221.0'°22.0 7 2425 26:5. 28.0) 27.5 41.6 143.0 3920755.0% 4806 Bespin 8.0 4.0 6.0 4.5 2.0 5.0 255 320 9.0 6.0 6.0 KM Ay Wear 18242147," 26. 69) 29:6 26 (82-297 39 43459104454 47 427056 ed ee es. yf renee a Shani er eee eC) ee oe Py my Pa RY SE ee Per cent. error for scores 43 Sea 2:8 2.4 Zoo i Wey; 1.6 1.6 1.4 Leo 1G 1.0 statistically reliable, but the other is not. This is what one should - expect, since the number of errors depends in a great measure upon the number of chances one has to make them. It is evident that the more words that one writes the greater the chances he ~ has to increase the number of errors, while at the same time the per cent. of error might be decreased. This was the case for the two groups in this experiment. Group Pa made fewer errors on the average by 6.02 per five minutes of practice than Wa made at the same time; but when the number of errors is compared ae ee COMPREHENSIVE UNITS IN LEARNING TYPEWRITING — 33 with the number of letters written at the same time the advantage is with group Wa by a difference in per cent. of error of 1.93. Speed Tests—In Tables VII, VIII, and IX, as well as in Figures II and III, are presented the results of the speed tests. It is in these that we get the best view of what the accomplishment was for the learners of the two groups. eee. eer amet oe mae are ey achaha ake 3.0 es S25 SSSR See eis Poe Seeks Me ee ree oe Ae ey oie. ae BA WVe ae $3::0 2.8030. 4'87.5:2.89-07 S98 O010ICO 11730 eS ee 104 0c PS 30 eee 2 1 pe hae oe 6.0551 073 535 9 £0481. 08 M0 SO OR Saas 1 Ate 0 EAVES Seats A PANE. BASE foe COO ce akie aoe OOOO Abe a 4Oeeoe 75, -OLeO Biodome Aue ot | ORAS ieee Bore. AVAL, AA OE: 4.5 2.0 1.0 3a0 Sa Sieeee 3.0 320 Aes RaW ee seine 106.0°119.0 108.0 126.3 28.0 141.0:-216 0.14720 162.5 1745025400220 "Re i ee 1 oO P22 23. °323.,0. 2023430. 084810. Ad Oe 29.0 S315 19 DLS eee i BAT Ti eel ae 86.0 Se SO SO i 77, Van OS AO sees eee OS Oe (Oecd 0 Oa saree eee Boeke 8.5 etl ROTEL Uaioe. Mas Bi) tara coe EL Gh atl Se Le) ee ee BSW rie wine -. 90:.0°'°73..0 § 89.0. 93-07 85.0 143-0°10620 490320 119 .0>130- 01488 Elec Pe 2! Ah aA Oe 1901 > Obed tO 9.0 9.0. (14.0) 111 (052 1720 eel tee Av. W.. 86.3 92.9°-88.7 | 88.8, 89:.2.°°91-T 113.9 .11073.1115.8 116.3 145.2 tee Oe ae 12 O47 a 10 e912 37 2 aa $591 09 eal 149 8.9 9.80 ligs Per cent. error for scores 1.4 1.6 i eg ¥.-4 1.4 1.4 .8 RG 1.0 .8 7 8 the differences are not statistically reliable, as the results show. Generally this becomes more and more the case as the experiment progresses. If we look again at Figure III it will be observed that in per cent. of error, on the basis of the number of words written, group Wa is far superior at the first set of tests, with a per cent. of 1.4 as against 4.3 for Pa. As the work progresses the two groups become more and more alike until in the end the difference is COMPREHENSIVE UNITS IN LEARNING TYPEWRITING — 35 represented by 1 per cent. of error for group Pa and .8 per cent. for group Wa. This should be expected, because when the tests began group Pa had exercised only on a portion of the keyboard. After the beginning of the testing this group had five minutes of practice per day on complete composition material because the materials used in the tests were business letters. The training of the learners of group Pa from that time included five minutes of work per day on the more comprehensive umit material. This seems to have had the effect of speeding up the relative efficiency of performance. It should be remembered that these two groups began at the same time, for both practice and in speed testing. At the beginning of the training they must have been more nearly equal in typing ability than at any time thereafter,’ the subjects being all without previous experience in the use of a typewriter. This difference is represented by an average score in words written (per five minutes) of 18.4 for group Pa, and of 86.3 for group Wa. The experiment was con- tinued through a second nine weeks and the relative positions represented in this difference remain nearly constant. What is the explanation of such results? This will be discussed at a later point. It should be kept in mind that these two groups were under practically identical conditions of training except for the kind of exercise material used in the practice work. The results of this treatment are represented by the general shape of the two error curves (Figure III). For the most part, when there is a rise in one of these curves there is a corresponding rise in the other, and when there is a drop in one there is also a drop in the other. When the learners were urged to speed the number of errors increased. If we compare the error curves with the curves repre- senting the number of words written, we see that, generally, when the words increase the errors also increase. This is not true for the per cent. of errors; when the speed is increased the per cent. of 1 Jt is unfortunate that a measure of the learners was not taken at the start. The fact that none had had any previous experience in the use of a typing machine seemed, at the time, sufficient to establish equality in the learning of this skill. 36 J. W. BARTON errors decreases or tends to remain constant. The results in Table VIII and in Table IX show that there was a gradual rise in the number of words written and a gradual decrease in the per cent. of errors as the experiment progressed through to the end. TABLE X SHOWING THE NuMBER OF TIMES Eacu LETTER WAS STRUCK, AND THE ERRORS MADE For BotH PRACTICE AND SPEED TEST MATERIAL oF Group Pa AND Group Wa Ss pase) pak cote bans Mos isan aan i a aa mine sa8 . 8 ooe0 : aa of phye ¢ a3) 8 Sage Z F Hy peo eee (oe i, 6 Be 88S 3 B a oo ge etas Re gi ES Bess a #8 SS SB CoSe 88 bo 6e Coes < oR a2 ef AShe as a2 Ge Meee PON Louetnee 295,834 22,756 42 .01 257,498 16,093 140 .55 eee le ase 486 a7” 30°47. 51.937 78 3246 0788 oie eerie ene 21,049 1619 48 1.92 94668 5916 98 1.03 Deion 154291 111099 22 15 98300 6144 8&8 .90 pe eee 204,618 15,740 78 (38 358444 22403 233 65 Bice pe se 143,820, T1064 215. Al 4019,760 6 W236 see Ta pes hae 5.952" < agg Wingy /s'g7 Mat RTO. 1.075 ata ivi og A165 =. . 1320.) 68). 16-25 91462784 09.1421). aeRO aane eee 1091331 8,410 96 .88 161,097 10,069 163 1.01 Tue, Came 107489 8268 2 (8 6104 382 12 2.09 ae A 142,707, 10,984 9.8" 7 /.Se. 0 12,3580 0772 a Oe Dette: 321,660 24743 37 11 150040 9,378 104 69 Miser Cimon 20,460 1574 71 3.49 76549 4784 87 1.13 Nis ae 2458 189 75 30.69 199404 12463 163 82 Oo ae ae 65.022 5.002787 \" 1.34 * 2530310°15.871 91459) eee Pe ae eas 52.406 4 A.038.<.331du 61h 434.521. aud, Sen Ocean osha 19-711 1516 0, AG A O71 | 8G oe Riss sous 169494 13038 80 48 141,057 834 159 11.87 Segue? 271,633 20805 32 .12 166,209 10,288 136 483 Tce Bh we 3133 °° 233 117 38.63 246,055 15.378) 155 eae Tier wee 116,635 {989729443 ©) £737) 93,782) 5,862) 8:32 Mise 18.785. 1445. 15° .83 25,080 .. 1568 43. Gea Wonle. Meee: 50535. 3,886 27 ‘154! 2491532 © 3/006' “65 Lage 5 aoe eae 530041 VAIS) 13 au. 4B B78 049 ee Via 834 64 52° 62:50 74,087. 4630. 88. dame Z 4813 370 0 0.00 2/803 1 gg Average...... 88,923 6810 43 6.06 106,265 6,332 92° 1,68 Tables VIII and IX show results that might be interpreted as great irregularity on the part of the learners, because there are many places in the test scores represented by dashes. On closer view it will be observed that most of such indications are for those that had discontinued the work. On the part of the subjects in COMPREHENSIVE UNITS IN LEARNING TYPEWRITING _ 37 group Wa there was a stronger tendency to discontinue. Of this group five discontinued at the eleventh set of tests as against three of group Pa. This was probably due to the fact that some had reached the point where they could write rather effectively, with scores in a few cases of as much as 40 words per minute. None of those of group Pa was writing as many as 17 words per minute at this point. The Alphabet—What is the relation between the number of times a letter is struck and the accuracy with which it is written TABLE XI SHOWING THE P.E. or Group DIFFERENCES BETWEEN THE AVERAGES, AND THE STANDARD DEVIATIONS OF THE DISTRIBUTIONS FOR TABLE X - = 8 25 : 23 ss 5 fo § 2 a % fo ake & Bik ° 2 5a ge sf BEE ag eg ae OOS EY 6s ee oe oS £-s oe qv Ha asst Averages for group Pa... <5355.. 88,923 6,810 43 6.06 Averages for group Wa............ 106,265 6,332 92 1.68 Dif. in av. for groups Pa and Wa... 17,342 478 49 4,38 PISO ERLE USIVEAW sido k shee oe wie a iaccie 18,811 1,259 8.65 1.96 Standard deviations of distributions: PERE Se eee Pek re eer ING AOD 7,418 30.65 14.65 Vat te mere eet ve neh Od DU) 5,964 57.54 2.15 at the end of the learning period? It seems to be pretty well implied, by the nature of the material used for practice, that repetition is the most important condition of learning to manipu- late a typewriter. And by the way the symbols are combined in the exercise material in common use, it would seem that machine placement of letters in terms of the fitness of the figures is all that need be taken into account in determining what the exercise should be. All texts that are available to the writer, except that published by Depew in 1921, provide meaningless exercise mate- rial for the early part of the training. Repetition seems to be the watchword in all that is proposed for learning this skill. Table X presents results for the total number of times each letter was struck, the average number of times each letter was struck by each learner, the total errors for each letter of the alphabet in the speed tests, and the per cent. of errors in the speed 38 J. W. BARTON tests figured on the basis of the number of times each letter was used in all exercises. If we compare the results of Table X with those of Table VIII it will be seen that while group Pa wrote more letters than did group Wa, the average per cent. of error for the latter group was much less, being 1.68 as contrasted with 6.062 for group Pa. This is a difference of 4.34+ 1.958. This difference is not statistically reliable, but it seems not to favor the “ practice makes perfect’ doctrine for typewriting (14, p. 21). When we com- 10 Greup Wa e ° Group Pa 125 120 sL05 @ > : eee, 2 90 Bea iin ps se Rae 2 E76 oS fs 60 45 1 2 3 4 5 6 {; & 9 10 ii 1 Test composites of four practice periods Ficure II. Showing curves of acquisition in composites of four tests, conducted each day during the second five minutes of a 55-minute practice period. pare the amount of practice for each letter with the per cent. of errors made in the speed tests we get a coefficient of correlation of .70+ .07 for group Wa and .60 + .08 for group Pa. Hoke, with different composition material, found a coefficient of “.924 with a P.E. of .021” (14, p.21). He used the products-moment formula while, in this study, the rank difference was used in calculation, and the coefficient was converted to 7 by the use of the table given by the Scott Company Laboratory (29, p. 128). If repetition is so important, why should the material used in practice not conform more nearly to the materials used after the COMPREHENSIVE UNITS IN LEARNING TYPEWRITING _— 39 learning is completed? Table X shows that the letter ¢ is ranked twenty-third in the number of times that it was used by Pa in practice, while with group Wa it is ranked fourth. Again, the letter m is ranked twenty-fourth for group Pa and fifth for group Wa. The letters 7, f, h, and others, are similarly neglected by group Pa, but not to the extent indicated for t and n. These results seem to show that more attention should be given to the / Group Wa Errers —.—, / Croup -va Errors 9 —(o- =. ‘ Group wa & Zrrorg ——————_ ~s Group Pa % Yrrors —x ——x— per cent 6rror 1.0 1 2 st 4 5 6 tf 8 9 10 Ay 1s mest composites of Zour practice periods Figure III. Showing curves of the number and per cent. of errors in com- posites of four tests conducted as in Figure II above. providing of exercise material that will center the practice on the letters that are used most frequently in the composition material of finished typists. Are we to conclude, then, that mere repetition is the factor responsible for the accomplishment of skill in typewriting? The results shown in Table X seem to support such a view. Frequent repetition of the various letters produces a higher number of errors, but this is because such a condition provides more oppor- tunity to make the incorrect response. But when the errors are represented by per cent. (of the practice in learning) it is found that the greater the number of repetitions the less the tendency is to respond incorrectly. But does this show that it is the repetition of the response that produces the change? 40 J. W. BARTON IV. DISCUSSIONS AND CONCLUSIONS Problem.—The question of the kind of material to use for practice in the learning of typewriting has not been raised by any of the investigations of the acquirement of this skill. All such studies have been concerned with the nature of learning as shown by curves and other representations of what the factors are that condition learning, as it is commonly taught. In cases where learning was accomplished by the sight method the material was of the complete composition kind; in those employing the touch method, the meaningless units were used. There has been no prior attempt to compare the relative effectiveness of these two kinds of material as means of attaining results in the acquirement of this skill. Since 1921 there has been a decided tendency on the part of makers of texts on the teaching of typewriting to get away from the use of meaningless letter combinations as exercise material ; and at present some of the late ones are making use of complete sentences as early as the third exercise. Subjects——The method of selecting and assigning the learners to their groups in the first experiment of this study did not strictly conform to the best methods in making a scientific inves- tigation. The same criticism applies to the collecting and record- ing of the data. In both these matters all the material available was presented. Notwithstanding these defects, the results show such wide differences in typing skill for the two groups that the irregularities cannot explain them. The subjects of group P, who were trained by means of the smaller units, and those of group W, trained by the more compre- hensive units, show no difference in fitness for learning that could influence the results very materially, notwithstanding the irregularity of selection. In the amount of time spent in prac- tice, the difference is that of the first eleven weeks. During this time group W was not training at all. Both groups trained an equal amount of time after this. That is, group P was trained EE a a a i COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 41 eleven weeks longer, during the school year of thirty-six weeks, than was group W. This should be kept in mind when comparing the results of these two groups. The subjects of group Pa and of group Wa were selected by a chance method. At the time of starting the experiment they all affirmed that they were without previous experience in the manipu- lation of a typewriter. In the schoool marks, school rank, and mentality there is no difference between the two groups that is statistically reliable. The mental tests show a slight advantage on the part of group Pa. In both experiments, then, the compared groups were approx- imately equal as far as the demands of this study are concerned. The differences that are shown are about as small as one could expect for groups selected by any chance method. Tramng.—Groups P and W were taught alike by the same teacher throughout, except for the kind of practice material used in training and such other minor matters as were incidental to the two methods used. Group P used the material presented in the text by J. W. Mosher published in 1903. Group W used com- plete composition material. Both used the same practice material after the first speed test. The groups were not taught in separate classes, but were mingled together in both the speed tests and the practice periods, thus making the work of training practically identical. This was true only after group W began practice work like that made use of for group P, which was eleven weeks after group P had started. The periods of training for P and for W were 40 minutes per day, five days per week, throughout the school year of thirty- six weeks. Group Pa and group Wa were also trained in common except for the first few periods of practice. They thus received as nearly the same kind of instruction and treatment, after the first few days, as students in the same class get under the same teacher, except in the nature of the practice material used. Group Pa used the text by Ross published in 1914, while group Wa was trained throughout by means of the material found in the business letters presented by C. E. Birch in his text of 1920. 42 J. W. BARTON Results.—The results show that group W after training for five weeks made a better record than was made by group P even though the latter had been training during the same period that group W had been trained, as well as during eleven additional weeks next preceding the five weeks. The P.E. of the difference is less than one ninth this difference. It should be remembered that the scoring method might be questioned, although it is not seen how a composite score of words and errors could operate to the advantage of either group. From this point on to the second speed test we see that group P makes very rapid progress; so much so that when the second speed test is given (three weeks later) the difference between the two groups is not within the limits of reliability. This rapid rise in the curve has not been accounted for, but it is probably a direct result of the P group getting away from the effects of negative transfer or interference. From the first test through to the end, the average differences, while not statistically reliable, are con- stant and favorable to group W. This is a rather strong indica- tion that the differences during this time are more than accidental. These differences, although only slight, make it appear that there were eleven weeks of time wasted in the training of group P, due to the use of meaningless units as practice material. Another peculiar aspect of the results of the tests is that after the first measure was taken the relative difference between the two groups remains almost constant to the end. Since the only change that took place at this point in the learning was that of giving group P five minutes of training each practice period on the comprehensive materials used in the tests themselves, it seems that this kind of material was responsible for the difference re- maining practically the same to the end. Some such effect was somewhat anticipated, so the test periods were made no longer than five minutes. It should be noted further that in both experiments the results show that about the time the two opposing groups, in both experiments, began work on complete composition material the difference between them remained practically constant. In the second experiment Pa and Wa began training at the same time (it should be recalled that this was not the case for a a ee COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 43 P and W ), and were trained under as nearly identical conditions as seemed possible for schoolroom conditions of experimentation, except for the kind of practice material used. In this group Pa was trained by means of the meaningless unit, while Wa was trained by means of that found in business letters. The differences could not have been in the degrees of motive induced in teaching, since the results of the second experiment show that the difference is equally as great as when the two groups were training at different times. In the last experiment the two groups were together all the time except for the first few practice periods. The results indicate plainly that the greater the number of repetitions the greater the per cent. of accuracy. ‘This is true not only on the whole, but there are very few exceptions to this rule when each letter is considered separately. It is evident that group W and group Wa had a decided advan- tage over group P and group Pa in acquiring skill in the manipu- lation of a typewriter and that this advantage for these more suc- cessful groups was probably due to the fact that these W groups were trained by means of the more comprehensive material as practice matter. If to repeat an act increases the chances that it will be per- formed again, is it not sensible to expect that the best results will be obtained if the repeated acts are as nearly as possible identical with those in the skill to be performed subsequently? This ques- tion suggests an explanation of why the complete composition material is better than the meaningless jargon, or smaller unit material, so widely employed in teaching typewriting. We may suggest here that possibly it 1s not the repeating of the act, as such, that is so important in the producing of the change called learning, because the letters written by group Pa and by group Wa are statistically equal, but group Wa made more than two times as good a record as group Pa did in the speed test that shows the least difference between these two groups. The scores in this test are 105.2 words per five minutes for group Wa and 47.4 for group Pa. The difference is 57.8 + 7.3. The real problem in learning is to determine what the factors 44 J. W. BARTON are that are responsible for retaining the mechanism that is neces- sary for the correct response, and what it is that makes inactive the mechanism responsible for the incorrect response. Thorndike has rightly emphasized his “law of effect’ (33, p. 71) in learn- ing, but he has failed to make clear by what neural conditions it may be done. Woodworth has recently emphasized what he calls “the law of combination,” apparently recognizing the need of getting away from mere frequency and recency in learning as the supposed sole determining factors. MacDougall (19, p. 194) takes issue with the “ effects’ theory on the grounds of its being “a clear case of effect preceding the cause.” Peterson has shown experimentally (23) that frequency and recency factors are not as important in establishing mechanisms used in response as has been made out for them; and that learn- ing goes on even in cases where frequency and recency factors go against it (22). Similar results have more recently been obtained by Kuo (18). Concerning recency and frequency among other factors that are responsible for learning, Peterson says: “ Elimination of random acts not favoring the dominant determining tendency seems to be brought about somehow by interference through conflict of different interacting processes. Determining tendencies and purposes are themselves but general directions in behavior effected by the larger consistency of the environment and by the inner metabolic processes” (24, p. 384). In discussing how our control over immediate conditions is brought about, he says further: “ We overrule them because they are inconsistent with the larger conditions that have shaped our behavior trends, both those that are innate and those that are acquired (24, p. 386), . . . but I wish to emphasize the ability or the tendency to respond to comprehensive and complex situa- tions as against the response to immediate stimuli” (24, p. 389). He is discussing the nature of intelligence in this last quotation, but states that “ such a view relates intelligence to learning, which certainly includes reconstructive aspects” (24, p. 389). The results of this study seem to support the “laws ” of Thorn- dike and of Woodworth which attempt to go beyond mere fre- quency-recency factors, and they agree with the experiments of avr COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 45 Peterson and of Kuo just cited. In the interpretation of our results the views expressed by Peterson are most helpful. The larger and more comprehensive units produced the most rapid learning, probably because they are more consistent among them- selves and more thoroughly in harmony with the determining tendencies of the learner. Our experiment undoubtedly suggests means of greatly improv- ing current methods of instruction in typewriting, to say nothing of such other skills as stenography, handwriting, piano playing, etc., each of which needs special experimental investigation. This problem grew out of the practical experience of the writer while serving as superintendent of a small city school system. At the time of opening school, in September, 1918, there was no thought of submitting the methods of teaching typewriting to experimentation. The discussions of methods of teaching, in the regular meetings of the teachers, suggested that possibly a shift from the “ A B C” method of teaching typewriting to that of larger units would prove beneficial, as had such a shift in the teaching of reading. These accumulating circumstances account for the irregularity in selecting the subjects of the first experiment as well as for the incompleteness of keeping records of results for all the subjects, since only the records of the teacher in charge were kept during the first few weeks of school. 46 J. W. BARTON BIBLIOGRAPHY . Barton, J. W. Smaller vs. Larger Units in Learning to Typewrite. J. of Educ. Psychol., 1921, 12, 465-474. . Brrcw, C. E. The Vocabulary Method of Training Touch Typists. Battle Creek, Mich.: Ellis Publishing Company, 1920. . Boox, W. F. The Psychology of Skill, with Special Reference to Its Acquisition in Typewriting. Univ. of Montana Publications in Psy- chology, Bull. No. 53, Series No. 1. . Brown, J. C. An Investigation on the Value of Drill Work in Funda- mental Operations of Arithmetic. J. of Educ. Psychol., 1911, 11, 81. Carr, H. Distributed Effort. Psychol. Bull., 1919, 16, 26-35. . CHapMAN, J. C. The Learning Curve in Typewriting. J. of Applied Psychol., 1919, 3, 252-268. . Coox, F. Elementary School Teacher, Oct., 1900, 111 ff., and April, 1904, 544 ff. . Cumminecs, R. A. Improvement in the Distribution of Practice. Teachers College, Columbia University, Contributions to Education, No. 97. . Curry, J. S. Teaching Typewriting for the Best Results. Proc. Nat. Educ. Assn., 1911, 834-841. . Depew, O. A Scientific Course in Typewriting. Boston: Allen and Bacon, 1921. . EssincHaus, H. Memory. (Tr. by H. A. Ruger.) New York: Teachers College, Columbia University, 1913. . Harnep, W. E. Typewriting Studies. Chicago: Ginn, 1923. . Hm, L. B., Reyatt, A. E., THornpike, E. L. Practice in the Case of Typewriting. Ped. Sem., 1913, 20, 516-529. . Hoxe, R. E. The Improvement of Speed and Accuracy in Typewriting. Johns Hopkins Univ. Studies in Educ., No. 7, 1922. . HoLttincwortH, H. L., Porrenpercer, A. T. Applied Psychology. New York: Appleton, 1917. . Huey, E. B. The Psychology and Pedagogy of Reading. New York: Macmillan, 1909. . Kirkpatrick, E. A. An Experiment in Memorizing vs. Incidental Learn- ing. J. of Educ. Psychol., 1914, 5, 405. . Kuo, Z. Y. Behavioristic Experiment on Inductive Inferences. J. of Exper. Psychol., 1923, 6, 247-93. . MacDoucatt, W. Outlines of Psychology. New York: Scribner’s, 1923. . Mosuer, J. W. The Complete Touch-Typewriting Instructor. Chicago: J. A. Lyons, 1903. . PecustTEIn, L. A. Whole vs. Part Methods in Learning Nonsensical Syllables. J. of Educ. Psychol., 1918, 9, 381. . Peterson, J. Learning When Frequency and Recency Factors Are Negative. Psychol. Bull., 1921, 18, 80-81. . Pererson, J. Frequency and Recency Factors in Maze Learning by White Rats. J. of Animal Behavior, 1917, 7, 338-364. . Peterson, J. Intelligence and Learning. Psychol. Rev., 1922, 29, 366-389. . Putuuirs, F. M. Value of Daily Drill in Arithmetic. J. of Educ. Psychol., 1913, 4, 159. . Pyte, W. H. The Psychology of Learning. Baltimore: Warwick and York, 1921. ‘ ’ J : ? : d Zi. COMPREHENSIVE UNITS IN LEARNING TYPEWRITING 47 Ross, J. W. Lessons in Touch Typewriting. Chicago: H. M. Rowe Co., 1914. . Ross, J. W. Intensive Touch Typewriting. Chicago: H. W. Rowe Co., 1921. . THE Scott Company Lasporatory. Tables to facilitate computation of coefficients by the rank difference method. . Starcu, D. Educational Psychology. New York: Macmillan Co., 1921. . Swirt, E. J. The Acquisition of Skill in Typewriting: A Contribution to Psychology of Learning. Psychol. Bull., 1904, 1, 295-305. . THORNDIKE, E. L. Practice in the Case of Addition. Amer. J. of Psychol., 1910, 21, 483. . THORNDIKE, E. L. Educational Psychology. New York: Teachers Col- lege, Columbia University, 1920. . WarrEN, H. C. Elements of Human Psychology. Boston: Houghton Mifflin, 1922. . Wetts, F. L. On the Psychomotor Mechanisms of Typewriting. Amer. J. of Psychol., 1916, 27, 47-70. . WoopworTH, R. S. Psychology, A Study of Mental Life. New York: Holt, 1921. nae Sw aes i. 4 i r. i . é 2 oat ed Boks Lay wages ( ie REST pet, Mr ¢ Cee] . a LN 4. ete Bee vans . , -\ haem ; “a Tom aelhe ae a vera ie oe. “4 th aa ae ¥ Be litga Fe. Bes vats 3 | | Ne ae is Bi ot Mae WL 2 Ge BURY E vee dehy soak aware sh ae x A Pea er at v ead, a a ay x Keane wt ‘ar ‘ Agu 5 Bh 4 + ui é 4 ee ae rt : é te bs 2 eS at y ah Slay +44 PNG SOF) ? ; Ais ¢ — P ; es oA Le .; aa i? “is, PURE CMTE “a ee Pid Laren lige id pe. ; Mee : a Ri ea: Jr Ay r , iis, sakes ee LR Oh on nie Mn ed Ma tag, hel saa ops = ne ue e 1< a i t “9 ; bole od? 7 ey . \ x thao pon ef te i} ue in | La oar ny i we et i ve - y AR ARR cm i Ce TEE ey PT Sia ceo ee eee aes at 7 ; 4 th : ce tose & last cae # Deke ‘ ' @ * : © a 7 ei Mn feages “ty co r — : i ; A vu i Ae % i, yy ny ik ee? eartys Peet ye ork ‘yw, % : i ane by ; z : ae ; é a = 5 ha sa a - ut ce - ‘ ¥ ‘ ‘ ; ’ i‘ 4 v - : ' le ae ays " j fen ' ; : r iy a = i r A + ; a a ml ¥ } ) Lat} ¥ i \ 4 a i . } 7 é 2 E. ) 7 , 4 NV 4 a F t Z « 1 ve, s ‘ ; : : ; - . oe ; - i ate, “ ‘ rf ; = v . ¢ H : ‘ ' ¥ ‘ af / a ¥ ; + ¥ « ® 4 a ~ a . 5 Tm « a " ‘ ’ , q fi 3 j i } * > f 14 ‘ a ad “ b ; : f ‘ PS - ; ' » f vv Ae ra “y 7 & Cb } iy ha ; (Ts ‘ , F 5 PAS eb sh age ‘. a <-> 383" “ aT, Se eh ’ 4 i aed * 2 4 “ ey i a i oe i BER ELAC hee : Z a 7* Pye ¥ : a5 b : : 4 Sa Rony Pee x « a ° 3 Ai Ae he oe ae Ne Oe Sn Ae eg hae v alee : ‘ Go Fe PA he Lh Poy, IP ae a Pian ‘ ot ee dp ee ec , ee ee ae Tee le iu : 4 iy Be io a. > —_ = ; ; of \ ; ae. ey LF i > ri ies ae: ae iste j ‘ ie el mT Tie ee Shs Pine, SP, ae my ne, y a Me ive o- 75" oS, bey a Fy 4 TR = C45 n . . ‘ ee ialx yi ’ - Pie ites , ry ® a3 oa ae ad Ws res an ve