6^ FROM THE INCOME OF THE FISKE ENDOWMENT FUND THE BEQUEST OF lyibrarian of the University 1868-1883 1905 . /^.2,*1.5;i.ni 5^Ml3 #*- iiJI 5 '34- ! DEC 1 1957 H X DEC 2jj2»^-t KRJUL14B0 Cornell University Library BF431 .S61 Correlations of mental abilities. olln 3 1924 029 195 241 1^ ly^ CORRELATIONS OF MENTAL ABILITIES BY BENJAMIN R, SIMPSON, Ph.D. TEACHERS COLLEGE, COLUMBIA UNIVERSITY CONTRIBUTIONS TO EDUCATION, No. S3 PUBLISHED BY ISfntifetB (HaUtgf. CHolnmbia Univrrsttg NEW YORK CITY 1912 Copyright, 1912, by Benjamin R. Simpson CONTENTS Sec. page I. The Problem and Its Importance i II. General Method of the Investigation 3 1. Abilities tested 3 2. Tests used 5 3. Subjects 5 4. Method of giving the tests 7 III. The Administration of the Tests in Detail 8 1. Order of giving the tests 8 2. Instructions given in each variety of test 8 3. Individual differences in ability to interpret instructions 11 IV. Scoring of Results 12 1. General principles 12 2. Method of scoring for each test, and reliability of the score given 13 V. Reliabllity of the Measures Secured by Use of the Tests 35 1. The two different methods used in treating the correla- tion data 35 2. Detailed discussion of the reliability of each test 43 VI. Significance of the Tests and Analysis of 'General Intelligence' as Shov^tn by Differences Between the ' Good ' and the ' Poor ' Groups S3 1. Extent of overlapping in the different tests S3 2. The mental relationships revealed by the Pearson co- efficients of correlation SS 3. Grouping of the tests according to the mental relation- ships shown by the correlation coefficients 65 (i) Tests of selective thinking 65 (2) Memory tests 65 (3) Association tests 6s (4) Perception tests 6s (5) Motor control 66 (6) Discrimination of lengths 66 4. Order in which abilities correlate with other abilities tested 66 5. Analysis of the individual differences revealed by the tests, and evidences that they are largely due to differences in native mental capacity 67 iv Contents PAGE VII. Comparison of Results With Those of Other Investi- gators 75 Wissler, Aikens and Thorndike, Norsworthy, Terman, Binet, Spearman and Krueger, Spearman, Thorndike, Burt, Bonser, Brown ,, .. VIII. Conclusions 109 Bibliography. iii Appendix. Descriptions of tests used 112 LIST OF TABLES PAGE la, lb, In. Original Scores Awarded 19-29 II. Total Scores Awarded and Scores in First and > Second Trials 30-33 III. Rank of Subjects in the Different Tests 34 IV. Illustration of the Methods of Treating Correlation Data. . 36 Va. Deviations from the Median of the Good and Poor Sub- jects Combined 39 Vb. Deviations from the Median of the Good Subjects only. . . 40 Vc. Deviations from the Median of the Poor Subjects only... 40 VI. Deviations from the Median, First and Second Trials, All Subjects Combined 41 VII. Reliability of the Tests 42 VIII. Extent to which the Poor Group Overlaps the Good Group 52 IX. Pearson Coefficients of Correlation, Raw 56 X. Pearson Coefficient of Correlation, Corrected for Attenu- ation 58 XI, XII and XIII. Pearson Coefficients of Correlation Used in Correcting Raw Coefficients SQ-^i XIV. Deviation Measures Inferred from Per Cents of Judg- ments of Superior 63 XV. Ranks of Good Group for Estimated Intelligence 72 XVI. Deviation Measures Inferred from Per Cents of Judg- ments of Superior 74 XVII. Some Results of Binet's Experiments 80 XVIII. Ranking of Pupils in Binet's Experiments 80 XIX. Oehrn's Results 81 XX. Bonser's Results 102 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924029195241 CORRELATIONS OF MENTAL ABILITIES* I. The Problem and Its Importance What constitutes general intelligence? How can we measure its amount? These are questions of immense practical impor- tance as well as of theoretical interest. Men in every line of activity are called upon every day to pass judgment upon the mental capacity of individuals and of groups. In many cases a choice must be made between a number of applicants of vary- ing degrees of capacity and fitness. Other things being fairly equal, the matter of prime importance for the judge to discern is the general mental ability of each of the persons in ques- tion. This judgment must be made in one of three ways: (i) by the examinations the candidate has passed and the certificates he has gained as a result of definite study; (2) by the opinions or recommendations concerning the candidate, given by those who know him and his work; (3) by the general impression gained from the way the candidate conducts himself during the course of the interview. We shall not dwell upon the inade- quacy of these tests as a means of determining the general in- telligence of an individual. The first, at best, gives a measure of the candidate's attainments along the lines tested, and only indirectly and secondarily gives an indication of his ability. The second and third are subject to all the inaccuracies of unscien- tific and ill-grounded personal opinion. Much as we need to get the right people into the right places, comparatively little has been done to replace these empirical methods by scientific ones. *The problem of this research was suggested and outlined by Professor E. L. Thorndike, and indebtedness is cheerfully acknowledged to him for a teacher's guidance and help in every difificulty. The work as carried out has been somewhat less comprehensive than that originally suggested. Grateful acknowledgment is also due the seventeen professors and stu- dents of Teachers College who acted as members of the " Good " group of subjects, to Miss Rusk for assistance in scoring a number of the records, and to Dr. Whitley. For the conclusions stated, the writer alone is responsible. I 2 Correlations of Mental Abilities The same holds true with regard to school determinations of ability. Certificates, degrees, and the like, of all grades of im- portance are given on the basis of demonstrably inadequate measures of mental capacity or amount of training, and later offered as valid measures of either or both. Students of educa- tion have felt the inadequacy of the old time methods to diag[nose and measure with any degree of accuracy the real abili- ties of the pupil, and students of psychology have, beginning with Galton, been devising tests of mental capacities both special and general. This work has been summarized in Whipple's "Manual of Mental and Physical Tests" ('lo), from which may be gained a just notion of the range of experimentation, the mistakes and improvements, and the present hopeful status of intellectual diagnosis by objective tests. The early workers along the line of devising mental tests for the measurement and diagnosis of general intelligence now see their labors justified by practical results. The period of discouragement and tem- porary defeat in the use of this method has been passed, and the time has come when workers in this field can go ahead wjth confidence that in due time results of much practical impqrtance will be secured through painstaking and intelligent iny^tigation. Once a series of mental tests can be perfected >hat will en- able us to determine the nature and amount of ^ person's mental capacity with a fair degree of accuracy, a cprner stone will have been laid toward the foundation of a science of education. As yet we are not in a position to do justice either to the exception- ally bright or the exceptionally dull pupils, to say nothing of pupils of smaller degrees of variation from the average. We have as yet perfected no scientific method of picking out excep- tional children, and until we have adequate means of doing this, we cannot expect to have their respective needs properly pro- vided for. But can we hope to find the means of classifying pupils in this way, according to the degree of their intelligence ? The an- swer to this question is to be found in the results that have al- ready been achieved by the use of even such imperfect tests as the Binet-Simon Tests of Intelligence. Already they are being successfully used and widely adopted in schools for the feeble- minded, to determine the mentality of the subject and the conse- General Method of the Investigation 3 quent treatment. They have been adopted in New Jersey as a means of diagnosis for retarded children. They have been used in courtroom procedure in New York City for the pur- pose of ascertaining the mental status of a youthful criminal, with a view to determining to what extent he should be held accountable for his conduct, and the sort of education he should subsequently receive. And this is only a crude beginning. In America alone the work of perfecting and extending such tests is being energetically pushed forward by Goddard in New Jersey, Wallin in the University of Pittsburgh, Huey in Johns Hopkins University, Terman in California, and others. Similar tests are being used elsewhere, such as the Sante De Sanctis tests in Italy. The more attention is given to studying the tests themselves, and developing new tests of greater convenience, re- liability and significance, the greater will be the practical educa- tional results. This monograph is a contribution to knowledge of individual differences, especially in that complex of qualities which we call ' general intelligence,' and of the means of measuring them. Its special purpose is to determine the significance of certain tests by showing what relations they have to one another and to gen- eral intelligence in its common meaning. On the other hand, variations in general intelligence, as imputed to this, that and the other individual by the world's judgment, will themselves be better understood by the discovery of their relations to achieve- ment in these tests. II. General Method of the Investigation I. Abilities tested The general method of the investigation was to select tests of a variety of mental. abilities, which may be grouped roughly under six headings, namely, sense-discrimination, motor-control, efficiency in perception, efficiency in association, memory, and what, for lack of a better term, I shall call abstraction or selec- tive thinking. Of course, all of these abilities may involve com- mon elements to a greater or less extent. On the other hand, each test is a test of a specific thing. For instance, the tests of perception are tests of a specific and limited kind of perception, under particular and limited conditions — to mark the A's or 4 Correlations of Mental Abilities the B's on a blank like that on page 112, and to mark the hexa- gons and halfcircles on a blank like that on page 113. If the test turns out to be a reliable one, that is, if two trials of the same test correlate highly with each other in a group of persons, it is evident that something definite in the way of ability has been tested. Whether or not the ability tested is accurately named, is a matter of secondary consideration at the outset. For instance, if we give two trials of the "A test " (marking as rap- idly as possible all the A's on a printed page), to each member of a group of a hundred persons, and find that the order of ex- cellence in the first trial is closely the same as it is in the second trial, — or in other words, that the two trials correlate very highly with each other — it is evident that the records secured are measures of some definite ability, even though we may not be sure that that ability is accurately named efficiency of per- ception.^ Again if two trials with the Easy Opposites test should prove similarly to test some one and the same thing, and if there should turn out to be a positive correlation of 50 per cent or more bttween the A test and the Easy Opposites test, it is evident that we have measured a mental relationship between different abili- ties, even though we may not yet be able to say with scientific accuracy that one is a test of perception and the other a test of association. However, the accurate naming of general abili- ties should be aided by the gaining of measures of relationship between different mental tests. For instance, if the A test and the Geometrical Forms test (marking all the geometrical forms of a certain kind on a page of printed geometrical forms), cor- relate almost as highly with each other as two records of the A test correlate with one another, it is evident that the two tests — A test and Geometrical Forms test — measure very much * The experimental work was done in 1907, but for various reasons pub- lication has been delayed. Since the writer undertook the investigation important studies of a similar sort have been made — notably those of Thorndike ('09), Burt ('og), Bonser ('10), Whipple ('10), Whitley ('11), and Woodworth and Wells ('12). If these studies had been available then, the writer could have altered and much improved the tests which he used. But at that time the work on the significance of tests was practically only that summarized by Thorndike ('03), and Spearman ('04), all done before the discovery of ' attenuation ' by the variability of the result of a few trials from that true for an individual on the whole; and two studies by Spearman ('04) and by Krueger and Spearman ('06). General Method of the Investigation 5 the same ability, and that they are properly classed under the same general heading, in this case perception. 2. Tests The tests used, fifteen in all, are described in detail in the ap- pendix on pages 112 to 122. They included two of perception — marking A's and marking geometrical forms ; three of memory — ^memory of unrelated words (auditory), memory of passages (auditory), and recognition of twenty-five forms studied, amongst fifty shown later; four of association — addition, easy opposites, associating words with hieroglyphic forms in pairs, and adding letters to make ba, ca, . . . be, ce, etc., into words (referred to as the ba or Completing Words test) ; three of selective thinking — ^the hard opposites, the Ebbinghaus or muti- lated text, and the absurdities test; two of sense discrimination — drawing lines each equal to a given length, and estimating the comparative lengths of pairs of lines'; and one of motor control — the scroll test. Most of these tests had been used at Columbia University. 3. Subjects As to the persons selected as subjects to be tested, the general plan was to take two groups of adults, representing as far as possible the two extremes of ' general intelligence ' as judged by the world. The group representing the high grade of intel- lectual efficiency — hereafter called the Good group — was made up of seventeen professors and advanced students of Columbia University. That the persons making up this Good group repre- sent a high degree of mental ability and efficiency is evidenced by the number who have since attained high positions in the teaching profession. At least five of the thirteen who were then graduate students have since secured university positions, and others are holding positions almost, if not equally, responsible. It is safe to say that if an omniscient judge should rank the half million teachers of the country in order for ' general intelli- gence,' at least three of the seventeen would be put in the high- est hundredth of them, and at least ten, very likely all, of the seventeen, in the highest tenth of them. It is also the case that the seventeen would rank far above the average man in the management of affairs. 6 Correlations of Mental Abilities Those m the group representing the low grade of efficiency were twenty in all, selected from men in New York City who had never held any position demanding a high grade of intelli- gence. All were of mature age and fairly comparable in that respect with the members of the Good group. In order that none should be at a decided disadvantage in the language tests, none were selected who did not speak English as their mother tongue. Two of them were persons earning comfortable livings for their families, but men recognized by their associates as being dull. Eleven others were staying at the Salvation Army Industrial Home at the nominal salary of $i per week in addi- tion to board and room, until work could be secured. One of these held a somewhat responsible position at the time, acting as assistant superintendent of the Home. He stood high in the most significant tests. The subject who did poorest of all was a man twenty-four years of age who had been at the Home for four years, and was quite content to remain there indefinitely on the permanent wage of $i a week besides board and lodging. The remaining seven were found in a mission on the Bowery where they were being helped somewhat until they could find employment. Altogether there were thirty-seven subjects, seven- teen in the Good group, and twenty in the Poor group. From the way in which those in the Poor group were selected, it is clear that they represent persons of a low grade of intel- lectual efficiency. They are, with the two exceptions noted above, persons who for some reason were out of employment. In some cases, no doubt, the lack of employment may have been partly due to habits of intemperance, but this defect would in most instances be insufficient to cause discharge if the man were otherwise above the average in efficiency. Nor would such cases be common in these religious establishm.ents. In my opinion these twenty men were as temperate in respect to alcohol as the average New York male. Most of the Poor subjects had worked at several different occupations. The facts, as far as obtained, were as follows: Brakeman, clerk, timekeeper; soldier, cook; teamster; wood worker, help on boats, hostler; machinist; worker in silk fac- tory, insurance broker, janitor; clerk, office help; shoe-laster; painter; clerk, teamster; help in customs house, night watch- General Method of the Investigation 7 man, checker on ship; hand in shoe factory; blacksmith, crock- ery business; salesman, collector of accounts; cook, mechanic; seaman, general help, cook, farm laborer ; foreman of milk busi- ness; wagon painter; farmer's help, employee in knitting fac- tory ; printer's help, driver of express wagon. The three who seemed the slowest and dullest of all, were the two who had regular employment but had been mentioned to the writer as being considered dull by their associates — ^the foreman of the milk business and the wagon painter — and the young man of twenty-four who was quite content to stay indefi- nitely at the Salvation Army Industrial Home. 4. Method of giving the tests The tests were all given individually in the same order and as nearly as possible in the same way. To avoid fatigue, they were given at two sittings, except in about three cases with per- sons of the Good group, where they were completed in compara- tively short time and with no visible signs of fatigue. The time taken to complete the tests varied from about three to five or six hours. Those doing them most quickly secured on the whole by far the best records for accuracy as well. The zeal with which the subjects undertook the tests is beyond question a matter to be carefully considered. This seemed to the writer as satisfactory as could reasonably be expected. In the case of the Good group all were eager to do their best be- cause of interest in the tests and their results for science. In the case of the Poor group, with the exception of two or three, the same was true, although perhaps to a somewhat smaller extent. At first a little difficulty was experienced in getting per- sons of this class to take the tests. Those at The Salvation Army Industrial Home were in some cases at first a little diffi- dent about it. They seemed to fear that their mentality might be subjected to a rather minute diagnosis. However, as soon as it became known at the Home that all of those selected were to be paid a certain amount " to complete the course," and that there was nothing very extraordinary about the tests, they were willing to do their best to give satisfaction, and also were eager to make good records. Only two of the whole thirty-seven (Good and Poor) seemed to find the tests irksome before the 8 Correlations of Mental Abilities finish, viz., number thirty-five who complained that his eyesight handicapped him, and number thirty-seven who showed httle evidence of interest in any of the tests. This was probably due to the fact that he had little interest in anything at all intel- lectual. There is no doubt that number thirty-five was handi- capped in most of the tests, and especially in those of perception, by poor eyesight; and possibly physical fatigue influenced his results, as he had worked all day at his trade of wagon-painting. A number of those on the Bowery seemed to do their best because of an idea that the discovery of unexpected talent might lead to the offer of a good position. III. The Administration of the Tests in Detail 1. Order A copy of each test used is given in the appendix on pages 112 to 122. The name used hereafter in speaking of the test is also given there, as well as brief general directions to the sub- ject about to take the test. The reader is therefore advised to turn to the appendix at this point for this information. The order in which the tests were given is the order in which they follow each other in the appendix. Some of the easier tests, such as those of perception, were given first so that the Poor subjects might not become discouraged at the outset. 2. Instructions As far as possible, the instructions for any particular test were the same for all subjects. But on account of the wide differences between some members of the Poor group and the members of the Good group, it was necessary to give more de- tailed instructions to the Poor subjects, to insure that they understood beforehand exactly what they were required to do. In giving the A test, the subject was allowed to glance at the page of printed capitals, and was then told to take up the paper as soon as the signal was given, and mark all the A's as quickly as possible with a pencil, in any manner most con- venient to himself. The time was taken with a stop watch. The second trial consisted in giving him a new copy of the test with instructions to mark all the B's as rapidly as possible. In the second test — Geometrical Forms — ^the instructions were similar, The Administration of the Tests in Detail 9 except that special pains were taken to make sure that the sub- ject understood clearly what sort of geometrical form was to be marked. To insure this, he was allowed to glance for a mo- ment at the printed slip, and then told to mark only the hexagon with the point up. This was further illustrated by drawing the hexagon to be marked on a separate piece of paper. In the second trial it was similarly made clear that the figure to be marked was the half-circle with the ftat side up, and not the half-circle in any other position. In taking the Scroll test of motor-control, the same fountain pen was used by- all of the subjects. The directions given for the first trial were: " Trace the white part of the spiral as rapidly as possible without touch- ing the black part." As the time taken for the first trial varied to a considerable extent, on the second trial the subject was re- quested to go either faster or slower than before so as to make the time occupied in the second trial about three minutes. In the Easy Opposites test, the subject was required to give the opposite of the printed word orally, instead of by writing it, so that the test would be a measure of quickness of association rather than a measure of quickness of writing. The instructions were: "As quickly as possible give orally a word that means the exact opposite of the word in the list. Thus if you see the word small, say large, and so on down the list." In the Recognizing Forms test, the subject was allowed to glance for a moment at the two sheets. It was then explained that he would be given one minute in which to study the small sheet, and that at the end of that time, he would be asked to mark on the large sheet only those forms that were exactly the same as the ones he had seen on the small sheet. In both cases subjects were given all the time they desired to mark the forms on the large sheet. In giving the Memory of Words test, it was explained that a list of simple words would be read aloud once, and that as soon as the last word was heard, the subject was to write down in any order as many of the words as he could remember. In the Pairs test — connecting a word with a hieroglyphic form — the subject was allowed to glance at a sheet containing the forms only, and another sheet with the forms and words in pairs. He was then told that he would be given one minute in 10 Correlations of Mental Abilities which to study the second sheet, so that when the forms only were given he could write down for each, the corresponding word. In the last two of the four trials given, one and a half minutes were allowed in which to study. In the Memory of Passages test, care was taken always to read at a uniform "Tate and as distinctly as possible. Two or three of the Poor group complained that they could not spell well enough to write down the substance of the passage heard. In such cases the experimenter took down the substance as the subject gave it orally. In giving test IX — Drawing Lengths — the subject was allowed to draw each line on the page of a ruled exercise book, but each of the lines he drew was covered up before drawing the next line, to insure that each drawing was an estimate of the original length and not a copy of the first line drawn. In giving test X — Estimating Lengths — two horizontal lines of nearly equal length were drawn side by side on a piece of cardboard. They were heavy lines drawn with a drawing pen and India ink, so that they could be clearly seen and easily com- pared. There were four sets of cards, with eight exactly alike in each set. In the first set the difference in length in the two lines was 8 mm., one lOO mm. and the other io8 mm. In the second set the lines were loo and io6 mm. respectively; in the third set loo and 104 mm., and in the fourth set 100 and 102 mm. In giving the test each cardboard was shown in turn to the subject, and he was asked simply, " Which is the longer of the two lines?" The whole test was then repeated, so that in all there were 16 estimates of each of the four lengths of this test. The subject was allowed all the time he wished to decide in every case, but little seemed to be gained by much hesitation in deciding. In test XI, Addition, the subject was allowed to glance at the addition examples, and was instructed to turn over the paper and start adding as soon as the signal was given, to add as quickly and accurately as possible, putting down the results as he proceeded. In test XII, Hard Opposites, the instructions were the same as in the Easy Opposites test, except that the subject was told The Administration of the Tests in Detail 1 1 to write down the opposite of the printed word, instead of giving it orally. Unless the word was utterly unknown to him, he was to take pains to write down the most accurate opposite he could. In test XIII, Completing Words, or " ba test," besides the general directions indicated in the appendix, the subject was shown by an example what was to be done. To insure clear understanding of what was to be done in the Mutilated Text or Ebbinghaus test, a special sample was shown, and the blank spaces were filled in for him with ap- propriate words. In giving the last test. Absurdities, the instructions were, " Mark each sentence that contains an absurdity or impossibility. For instance, if a sentence stated or implied that lead was float- ing on water, mark such a sentence as absurd or impossible. Do not mark the sentences that contain no absurdity or impossi- bility." 3. Individual differences in ability to interpret instructions Before giving the tests to any of the thirty-seven subjects whose records were secured, the experimenter took the precau- tion to practice himself in giving the tests to other persons, both for the purpose of acquiring the necessary skill in giving the tests in a uniform manner, and in order to determine the specific instructions necessary to be given to the subjects. Every reasonable precaution was taken to see that the instructions were understood before the test was begun, but in spite of this fact several of the poor subjects failed to follow instructions ex- actly. In some instances this seemed due to the fact that, though the instructions were understood when given, a part of them were forgotten as the work proceeded. In marking geometrical forms — test II — several of the Poor group started in by mark- ing only the hexagons with the point up, but before they got half way down. the page were marking all hexagons — with flat side up as well as with the point up. As the test was intended to be one of quickness in picking out forms and not one in ability to understand and remember instructions, they were not severely penalized for this.^ In three or four other cases it seemed fairly evident that the instructions were not perfectly ' See later account of methods of scoring. 1 2 Correlations of Mental Abilities grasped before beginning the first trial, even though both sub- ject and experimenter thought that they were. For instance three of the dullest of the Poor group in doing test XIV, Ebbing- haus mutilated text, put several words in a single blank space where the sense seemed to them to require it, although they were distinctly told beforehand to put only one. Thus incident- ally the tests gave the experimenter an opportunity to form an estimate of each subject's ability to comprehend instructions. No attempt has been made to correlate this accurately with abilities as indicated by the tests themselves, but from memory of individual cases the writer is very confident that in general those who were dullest in comprehending and remembering in- structions were also poorest in the tests which later results proved to be most closely correlated with intelligence. In two cases a zero record — in test VII, Learning Pairs, — ^in the first of the four trials resulted from the fact that the subject took it for granted that the figures would be arranged in the same order on both the study slip and the test slip. He thus at- tempted to learn only the words in their correct order, instead of connecting the form with the corresponding word. How- ever, irregularities of this kind were of such slight consequence as not to interfere to any appreciable extent with the general results. IV. Scoring of Results I. General principles Any method of scoring the results of tests such as these must be more or less arbitrary. The method of scoring finally adopted was that which seemed fairest on the ground of common sense, and that which seemed to vary as little as possible from the re- sults secured by other reasonable methods of scoring. The original results which would take some scores of pages to print are all on file at Teachers College, so that any one who cares to do so can compare the final scores adopted here with those obtained by any other method of scoring that seems to him advisable, or test any of the writer's conclusions by com- puting the results by the different scoring. Scoring of Results 13 2. Method of scoring for each test, and reliability of the score given The final score in the A test was got by taking the time in seconds required to complete the test, plus five seconds for each letter omitted. Thus the lowest score represents the greatest efficiency. As the average time for marking each of the fifty A's was not much under three seconds, it seemed fair to penalize each omission by something more than three seconds. On the other hand, if much more than five seconds were added for each omission, it would emphasize care and accuracy rather than quickness of perception. Moreover, it made little difference whether three, four, or five seconds were added for each omis- sion, judging by the average displacement of rank of the dif- ferent subjects. The average displacement in rank for each sub- ject owing to the difference in scores when four seconds are added for each omission, and when five are added, is only a trifle over one. The average displacement in case of adding three seconds for each omission and in adding five seconds for each omission is only two. Although this test is satisfactory in reliability, the average displacement in rank between the first trial (marking A's) and the second trial (marking B's) is slightly more than five. Thus the method of scoring adopted seems quite satisfactory. There were no errors of any kind in this test except omissions. For similar reasons the method of scoring the Geometrical Forms test was to take the number of seconds and add three for each omission in the first test, and six for each omission in the second. There were more than twice as many hexagons to be marked as half-circles. Hence it took about twice as long on the average to find and mark each half -circle as to find and mark each hexagon. Nothing was taken off for forms wrongly marked, since they were already penalized somewhat in that it took some time to mark the wrong form, and since if much were deducted for this, it would amount to making the test one of ability to understand instructions rather than one in quickness in picking out forms. The score in the Scroll test was got by taking the time in seconds and adding ten seconds for each touch of the black 14 Correlations of Mental Abilities lines. Ten was selected as the number to be added for each touch, mainly because it equalized the scores of each subject in the two trials better than any other number. For instance, if five seconds instead of ten were added for each touch, the score in the second trial would be better than the score in the first trial in twelve cases, and poorer in twenty-five. If ten seconds were added for each touch, the score in the second trial would be better in eighteen cases, and poorer in nineteen than the first. If nine seconds were added, the second trial would be poorer in seventeen cases and better in twenty than the first. How much, if anything, should be allowed for improvement, it is very difficult to say, as it was not possible to keep either the time or the number of touches uniform in the two trials. Neither is it very helpful to estimate the average number of seconds that is equivalent to avoidance of one touch, as the variation is so very wide. In fact it seems difficult to find any system of scoring this test as given, that is altogether justifiable. On the whole it seems best to assume that the two trials are tests of the same thing, and that therefore that method of scor- ing is fairest which on the whole makes a given subject's first score and his second score as nearly equal as possible. Adding ten seconds to the time for each touch does this. For scoring the Easy Opposites test, the experimenter had kept track of the number of words that were correctly given, the number incorrectly given or omitted, and the number that could be considered half right. The score taken was the time in seconds, plus two seconds for a word half wrong, and four seconds for a word wrong or omitted. Notes were made at the time the test was given, of how much time a subject took to think up a satisfactory opposite for the one or two words that caused most hesitation or difficulty. This formed the basis for estimating how much on the average it was most just to add to the subject's time in seconds as a penalty for errors or omis- sions. In scoring test V, Recognizing Forms, one mark was allowed for each form correctly marked, and one was taken off for each form incorrectly marked. On this basis, if the subject Scoring of Results 1 5 were merely guessing, his score would on the average approxi- mate zero. In test VI, Memory of Words, the score given was simply the number of words correct. Nothing was taken off in case a wrong word was written down. This, later study has shown to be unwise, but there were only a few such wrong words inserted and, as the labor of recalculating correlations and group dif- ferences would be very great, the scores have been left as origin- ally niade. In test VII, Learning Pairs, the score given was simply the number of words right, except that there were a few instances where credit was given for a word half right, where the word was connected with the proper form but was not itself exactly correct. In scoring test VIII, Memory of Passages, the experimenter evaluated all of the records on a scale of from one to twenty- five. A month later he rated them independently a second time. An assistant in psychology also rated them independently. It was found that the three evaluations differed very little from each other. The average displacement in the rank of each sub- ject, according to whether he was ranked on the scoring of the assistant or on that of the experimenter's first scoring, was only about 2. The average change in rank between the experi- menter's first and second scorings was 2.3. As compared with this the average displacement of each subject in rank, between his score in the first two tests and his score in the last two, was 3. The score adopted in test IX, Drawing Lines, was the sum of the deviations plus and minus from the standard length. The constant error was separated out for study, but seemed to be of little significance, and so the deviation from the standard set was used. Out of the seventeen Good subjects only eight had a clearly positive error, and only four a clearly negative one. In test X, Estimating Lengths, each individual's number of correct judgments of each difference was recorded, and the indi- viduals ranked in order of merit accordingly. In scoring test XI, Adding, the time in seconds was taken and ten seconds added for each error. It was considered that the penalty should be estimated by the time that would probably be required to correct the error. This, in general, would be at 1 6 Correlations of Mental Abilities least somewhat more than was required to do that part of the sum. On this basis at least five seconds should be added for each error and probably more, as the median time for doing a whole question, with a possibility of two errors, was about ten seconds. Whether six seconds or ten seconds is added for each error, makes very little difference; the average displacement in rank caused by this difference in scoring being only slightly over I. The difference in rank between a subject's score in his first trial and his second is on the average, 3.8. On the whole it seemed that ten seconds was a fair penalty for each error. In test XI, Hard Opposites, all of the words given as op- posites were collected and evaluated on a scale of from i to 4 by three different persons, viz., two assistants in psychology and the experimenter. The three evaluations differed very little from each other, the average displacement in rank caused by using one rating rather than another being only slightly over i. The final score given to each subject was got by averaging the marks of the three different scorers, and then adding a certain number of seconds to the time score, as a penalty for each word wholly wrong or omitted, and a proportionate amount for a word partly wrong. On the basis of the time actually spent by a number of subjects to think up suitable opposites for the most difficult words rather than omit them, 36 seconds seemed to be about the right penalty for the total omission of a word. Other methods were tried, but none was found which on the whole seemed to equate the two factors of equality and speed so fairly, especially with reference to those who did best and poorest. For instance, an attempt was made to equate the time element and the quality element by combining the quality score with a reciprocal of the time score. This suggests itself on the general principle that if A does twice as much work as B in a given time, he deserves a score twice as high; if A and B do the same quality of work, and A takes twice as long as B, he is worth only half as much. The variations in the time, how- ever, were so great that such a method could not be fairly ap- plied without the use of complex mathematical calculations. Moreover it would assume that all of the words are equally diffi- cult, which of course is not the case. For somewhat similar reasons it was difficult to find a per- Scoring of Results 1 7 fectly satisfactory method of scoring in test XIII, Completing Words, as in several cases there were a number of omissions whereas in other cases subjects had spent times varying greatly in amount in attempting to avoid omissions. On account of the large number of omissions, the second trial was thrown out al- together. For a similar reason the records of eight subjects were thrown out in trial one. The score given was the time in seconds plus 10 seconds for each error or omission. As far as could be judged, 10 seconds seemed to be the time that probably would have been required to complete the word. In reducing the results of test XIV, Ebbinghaus Mutilated Text, to a numerical score, they were first marked by the ex- perimenter as to excellence in filling out the blanks so as to make sense, regardless of the time taken. They were then marked in a similar way by an assistant in psychology, and then marked independently a second time by the experimenter. These three ratings, which did not differ materially from one another, were averaged so as to get the final rating for quality, i.e., ability to do the test without regard to the time taken. The average displacement in rank of each subject according to the different scorings was between i and 2. As it seemed just to count the time taken to do the test about equally with the quality of the work, the final score adopted was a combination of the time score and the quality score, weighting quality a little higher than time. In doing this the time factor taken was the reciprocal of one-fifth of the time in seconds. This method is not open to the same objections here as in the Hard Opposite test, and succeeds fairly well in equating speed and quality. However, the method may seem unnecessarily complicated, and could not very well be used except under special and limited conditions. It would have been less trouble, and about equally fair, to have held to the general method used in tests previously mentioned, namely, to take as the final score the time in seconds plus a certain penalty in seconds for each mark lost in the score for quality. In giving the Absurdities test it was assumed that the record could be scored simply by allowing i mark for each absurdity correctly marked, and taking off a mark for each one wrongly marked. This would give a score for quality, with which the 1 8 Correlations of Mental Abilities time element could be combined to give a final score. It turned out, however, that the test as given was open to decided objec- tions, on account of containing a number of imperfections. Im- provements in the test will be discussed in another connection. As far as giving definite scorings was concerned, the test was finally thrown out. If, in spite of these imperfections, the re- sults are scored as originally intended, even regardless of the time element, it divides the two groups fairly well. The scores finally awarded to each individual in each test are given in Tables la to In inclusive. Each subject's total score in each test, and the scores in ist and 2nd trials of the test, are given in Table II. Table III gives the rank of each individual in each test. Scoring of Results 19 TABLE la Original Scores in Memory of Passages 1 + CO + CO + 1— 1 IN i-t CO + w 5' + CO Pi + nt gco 1 43 44 29 36* 43 44 31 37* . 2 44 44 31 37 45 43 37 40 3 42 41 20 30 43 42 35 38 4 42 45 28 35 44 44 37 40 5 43 44 31 37 41 43 38 40 6 38 36 24 30 43 44 39 41 7 40 42 25 33 37 40 24 31 8 33 35 19 26 35 38 27 31 9 41 29 21 28 37 34 28 31 10 43 41 21 31 40 41 30 35 11 45 43 35 39 34 38 23 29 12 42 42 29 35 37 40 32 35 13 32 27 13 21 30 27 16 22 14 29 31 13 21 32 32 18 25 15 9 14 10 10 19 17 18 18 16 36 25 16 23 37 24 23 26 17 41 41 22 31 41 43 34 38 18 21 27 12 18 26 30 25 26 19 32 20 14 20 22 17 17 18 20 32 25 18 23 31 33 26 29 21 30 25 21 24 28 33 23 26 22 8 12 6 8 9 11 9 9 23 27 23 14 19 31 24 20 23 24 21 26 11 17 27 24 20 22 25 12 16 8 11 12 15 11 12 26 17 21 10 14 22 23 18 20 27 11 13 9 10 12 11 11 11 '. 28 16 10 9 11 15 9 9 10 29 14 14 10 12 11 9 12 11 30 21 16 15 16 17 12 11 12 31 7 8 6 6 8 7 7 7 32 15 17 10 13 8 9 8 8 33 6 8 5 6 8 7 6 6 34 20 21 15 17 6 8 5 6 35 36 7 6 4 5 10 9 11 9 37 3 4 3 3 1 1 2 1 Si (lis) means sum of scores in passages 1 and 2 as scored by experimenter the first time. S2 (1 + 2) means sum of scores in passages 1 and 2 as scored by experimenter tlie second time. Ri (1+2) means sum of scores in passages 1 and 2 as scored by Miss R, an assistant in psychology. * Fractions have been omitted in these two columns. Correlations of Mental Abilities TABLE lb Original Scores in Memory op Words TABLE Ic Original Scores IN A Test 1 00 o 13 1 1 1 s 1 •a f 1 TO CO 1 (O cfl T3 a 8 "i i i 1 1 1 i d 1 1 6 B 6 Iz; s 1 9 9 g 2 10 50 121 50 135 2 9 11 5 9 46 80 38 77 3 9 1 9 11 9 1 49 109 39 111 4 10 9 8 9 43 86 39 84 5 5 10 5 1 8 1 47 107 44 112 6 8 10 2 9 1 11 1 50 126 49 127 7 6 1 5 5 9 41 121 47 133 8 5 6 2 6 8 50 118 47 128 9 9 8 10 9 47 129 43 135 10 6 1 7 7 7 50 100 50 120 11 9 4 1 5 1 8 1 46 78 46 95 12 8 1 8 7 6 1 48 123 48 137 13 6 1 8 6 1 7 50 92 48 92 14 5 5 1 7 1 6 47 111 40 92 15 6 1 5 1 9 9 49 127 50 123 16 9 7 7 8 1 48 129 42 118 17 8 3 1 7 5 48 138 47 136 18 6 2 6 1 5 6 1 50 118 49 112 19 3 4 5 1 6 2 5 1 48 115 47 114 20 8 1 5 1 6 6 46 142 46 195 21 6 5 5 5 1 19 81 41 100 22 6 1 6 1 5 5 44 107 48 90 23 8 7 6 6 1 43 162 44 171 24 6 5 1 5 4 39 81 31 84 25 5 6 4 2 4 1 29 84 44 130 26 6 2 2 4 1 6 2 60 157 49 142 27 3 2 6 3 1 3 1 45 155 43 144 28 6 6 1 3 1 5 47 162 48 165 29 5 1 4 4 1 4 30 112 38 101 30 5 4 2 5 5 35 123 44 144 31 6 4 4 2 1 46 195 43 133 32 4 1 4 2 5 2 4 1 34 92 32 83 33 4 1 3 2 4 5 18 127 40 115 34 3 2 6 4 2 1 2 SO 270 48 222 35 1 1 3 1 4 2 48 268 44 280 36 6 1 3 5 4 1 5 47 130 48 134 37 5 3 2 1 4 3 2 21 127 38 128 Scoring of Results TABLE Id Original Scxjres in Adding TABLE le Original Scores in Geomet- rical Forms t "3 s o ■s a o 1 a i CQ 1 T3 :S a o 1 1 .g -♦J s r i 01 02 07 d IS i 1 H d 6 !2; B 1 6 12; 6 iz; a 1 1 6 Iz; a 1 10 54 10 43 36 81 14 83 2 7 3 65 8 2 52 33 3 44 13 1 45 3 9 1 80 8 2 85 34 2 76 13 1 76 4 9 1 82 7 3 81 24 12 75 10 4 70 5 8 3 79 9 1 81 26 2 10 82 12 2 76 6 6 4 91 9 1 84 33 3 77 12 2 66 7 8 2 91 8 3 100 33 3 89 13 1 88 8 10 58 10 63 29 7 95 14 78 9 8 2 96 10 111 31 5 80 13 1 77 10 10 92 10 94 36 75 14 70 11 8 3 71 10 64 26 2 10 58 13 1 1 69 12 8 3 65 9 1 73 27 9 91 12 2 94 13 8 2 81 9 1 89 34 4 2 56 14 70 14 10 105 9 1 124 25 11 46 12 2 44 15 9 1 82 10 84 36 13 67 13 1 79 16 7 4 106 7 4 107 32 1 4 80 13 1 80 17 9 1 116 10 136 36 90 13 1 86 18 10 74 9 1 86 30 4 6 82 14 3 103 19 10 70 10 98 35 14 1 95 14 94 20 10 153 8 2 190 31 5 107 10 4 127 21 6 6 202 8 2 236 32 4 79 11 3 69 22 9 1 124 8 2 154 28 8 80 13 1 76 23 4 10 81 6 6 109 34 2 170 12 2 150 24 8 2 85 6 5 85 27 2 9 93 11 3 84 25 8 2 87 9 1 93 29 7 7 107 13 1 102 26 9 1 165 8 2 151 35 12 1 107 13 1 122 27 10 131 9 1 146 33 2 3 119 12 2 93 28 9 1 157 10 140 34 2 116 9 5 113 29 7 3 75 9 1 100 30 6 90 13 1 70 30 8 2 209 10 337 31 5 104 12 2 101 31 9 1 221 7 3 224 16 8 20 102 5 9 87 32 6 4 118 6 4 132 19 1 17 67 10 9 4 78 33 10 137 10 165 22 14 89 11 3 90 34 10 273 10 273 24 12 128 12 2 105 35 9 1 149 9 1 141 31 7 5 297 13 22 1 285 36 3 7 237 7 3 295 24 12 113 13 8 1 85 37 5 5 176 4 9 217 19 2 17 106 17 16 7 127 Correlations of Mental Abilities TABLE If Original Scores in Learning Pairs TABLE Ig Original Scores in Recognizing Forms +3 s. T3 T3 ■a T3 ■a TS '^ bo « tuO O) M ^ M (D Wl » bS O 3 ■s ^ e a 2 1 1 1 4 g 1 S -a 1 1 1= 1 -+3 2 a o d 6 d d d d d 6 d d d d d d 6 d 6 d d ^2; ;z; ^ ;2i 12; Iz; !z; 12; iz; a iz; ^ 12; fc 1^ a iz; ^ ^ 1 3 7 2 8 5 5 5 5 17 1 7 15 4 10 2 7 2 1 5 3 2 8 2 4 2 4 19 2 5 9 7 16 3 9 1 10 10 9 1 15 9 13 4 12 4 1 1 8 7 3 9 1 6 2 2 13 3 11 12 1 13 5 10 7 1 2 10 10 14 2 10 11 1 14 6 i 9 7 1 2 7 3 4 2 4 13 1 11 14 3 11 7 1 2 7 4 6 2 8 2 2 6 10 1 14 12 1 13 8 4 1 5 7 1 3 9 1 9 1 11 2 13 13 1 12 9 6 1 3 7 2 1 10 7 3 16 3 8 11 2 14 10 4 2 4 10 3 7 3 1 6 13 3 11 15 7 10 11 4 4 2 1 4 5 3 2 5 3 5 2 17 7 11 3 14 12 6 1 3 7 3 8 1 1 9 1 11 3 13 9 4 16 13 5 2 3 3 2 5 8 2 8 1 1 12 3 12 17 2 8 14 4 6 2 1 7 4 6 3 1 6 10 2 14 12 13 15 7i 2 5 1 4 9 1 9 1 16 1 8 14 3 11 16 6 1 4 6 4 7i 2 5 1 4 18 3 6 12 3 13 17 3 1 6 3 1 6 3 2 5 4 1 5 11 4 13 14 5 11 18 2 3 5 2 6 2 2 6 2 2 7 1 9 4 15 14 8 11 19 1 2 7 2 8 2 8 2 1 7 10 1 14 11 8 14 20 1 5 4 1 2 7 1 4 5 6 4 6 2 IS 13 14 12 21 1 2 7 2 2 6 1 3 6 1 1 8 7 1 17 8 5 17 22 1 4 5 2 3 5 5 1 4 2 3 5 9 4 15 12 9 13 23 10 3i 6 1 2 9 1 9 2 2 22 8 1 17 24 1 1 .8 1 1 8 1 9 1 9 6 18 6 2 19 25 2 2 6 2 3 5 3 1 6 3 2 5 3 8 21 9 8 16 26 1 3 6 li 2 6 H 4 5 2 4 4 13 11 11 12 4 13 27 3 1 6 2 7 2 2 6 4 1 5 7 2 17 8 4 17 28 3 7 2 7 2 8 1 9 9 15 1 1 24 29 1 1 8 8 2 1 7 1 9 6 2 18 10 15 30 2 2 6 1 8 4 6 i 2 7 12 10 12 8 8 17 31 3 3 4 4 5 3 1 6 2 3 5 11 8 13 17 8 8 32 7 3 9 1 1 6 3 1 3 6 9 6 15 11 11 14 33 1 9 1 2 7 2 8 1 9 1 1 23 10 6 15 34 1 5 2 8 2 7 3 1 6 3 11 3 13 7 3 18 35 2 8 3 7 2 1 7 1 2 7 13 9 11 13 10 12 36 1 4 5 4 6 1 4 5 1 4 5 10 8 14 17 9 8 37 3 7 2 8 1 1 8 2 8 6 5 18 7 5 18 Scoring of Results 23 TABLE Ih Original Scores in Easy Opposites ■s 1 d ;2; "2 1 .s + a 1 d 12: i H + i CO i 1 a 1 T3 a 8 (3 1 + £ S a 1 1 03 1 1 .9 0) a + g s 1 16 38 19 40 18 40 25 40 2 16 39 21 40 20 39 17 40 3 24 36 35 40 30 40 1 29 38 4 20 39 1 24 38 32 40 28 40 5 2 34 36 37 40 1 31 38 34 40 6 2 2 29 31 1 31 38 27 39 33 40 7 26 39 24 40 25 40 25 40 8 H 35 37 27 40 4 28 39 29 40 9 1 27 38 23 40 25 40 24 40 10 1 36 38 1 38 38 37 40 2 36 36 11 1 22 38 23 40 1 1 29 36 23 40 12 i 28 39 35 40 1 46 38 36 40 13 1 32 38 14 32 37 21 40 24 40 14 2 1 29 32 1 29 38 4 27 39 1 35 38 15 4 25 39 21 40 28 40 1 25 38 16 i 27 39 1 25 38 2 32 36 1 1 28 36 17 4 27i 39 36 40 4 29 39 41 40 18 24 47 35 2 37 36 2i 36 35 3 40 34 19 24 39 35 33 40 0' 1 50 40 14 37 37 20 14 43 37 49 40 41 40 1 41 38 21 2 42 36 53 40 41 40 1 79 38 22 14 46 37 4 1 53 37 14 1 54 35 14 1 48 35 23 2 32 36 2 40 36 51 40 1 1 90 36 24 74 32 25 4 38 32 3i 35 33 4 5 34 29 25 2 2 60 32 2 62 36 24 1 58 33 14 2 49 33 26 4 2 111 35 3 3 70 28 14 45 37 42 40 27 1 57 38 47 40 1 37 38 4 38 39 28 3 32 34 2 52 36 2 33 36 2 50 36 29 2 81 36 14 82 37 4 1 55 37 54 52 29 30 44 54 31 2 63 36 24 80 35 2 76 36 31 24 1 80 33 3 1 57 32 3 1 115 32 2 2 75 32 32 7 40 26 24 1 47 33 54 57 29 44 1 82 29 33 24 6 140 23 3 5 134 24 44 4 150 23 11 2 132 14 34 6 77 28 5 70 30 4 72 32 1 47 38 35 44 80 31 9 67 22 8 76 24 1 62 38 36 94 4 85 13 8 81 24 44 1 81 29 84 3 110 17 37 « 3 205 18 64 115 27 13 144 14 114 195 17 Score is 2 for each word right, plus 1 for each half right. 24 Correlations of Mental Abilities TABLE li Original Scores in Hard Oppositbs t C4-« i 1— ( 1 c 1 IN 1 <6 1 CO i 1 ,^ -fS ,^

i >, >, >, w >►. CO ^ CO >> CO u^ ^ J2 Xi XI XI _c X! p Xi Xi c O 00 Oi 00 i 02 1 02 oa a 02" 8 m c5 ID a S 02 >, 2 8 CO i s 02~ m i 1 9 42 10 82 10 50 10 52 2 9 79 9 141 8 122 9 92 3 9 46 9 74 10 60 9 80 4 9 113 8 170 7 127 10 212 5 9 92 8 106 9 140 9 110 6 7 170 8 130 8 277 9 155 7 10 152 9 278 10 139 10 84 8 9 82 9 311 10 132 9 90 9 9 108 10 176 8 268 9 92 10 8 152 8 182 7 330 9 203 11 7 88 8 91 8 85 8 117 12 9 93 8 192 5 149 8 83 13 9 163 9 315 8 258 7 397 14 9 138 7 181 7 176 9 137 15 8 108 9 205 6 213 9 396 16 6 106 5 129 5 132 9 197 17 9 124 5 257 5 313 10 208 18 9 136 5 195 6 215 8 154 19 7 182 7 307 3 291 6 209 20 9 340 4 290 6 350 10 209 21 5 341 4 316 2 286 310 22 5 251 5 220 3 217 3 207 23 6 255 4 282 3 254 3 215 24 5 284 5 265 5 227 4 266 25 7' 207 3 296 5 217 5 238 26 6 364 5 331 3 204 5 177 27 4 474 6 525 7 275 3 660 28 7 218 6 328 4 303 6 221 29 9 376 3 420 5 394 4 370 30 1 390 2 310 2 259 243 31 5 302 5 320 4 412 1 426 32 3 381 3 329 3 271 2 293 33 34 1 480 3 270 3 478 2 450 35 4 355 4 410 2 327 2 354 36 1 140 2 403 1 270 2 265 37 Scoring of Results 27 TABLE Ij (continued) Original Scores in Ebbinghaub Test — Mutilated Text Scores given by experimenter one month after 1st scoring by him (S2) Scores given by Miss R. tH IM CO •* "5 to t» 00 iH (N CO ■* Ui to t- 00 0) 15" 3 1 ■+3 1 1 1 +3 1 1 -4^ 1 1 1 t -** 1 1 ra .g a PI a a a PI B •S a _d _g a Ct "0 i £ i £ g 02 1 02 1 03 1 £ 8 02 £ 8 02 1 02 £ 8 02 1 1 02 1 £ 1 02 i 02 1 10 10 10 10 9 10 10 10 9 10 10 10 9 10 9 10 2 10 10 9 8 9 8 7 10 9 10 9 9 9 10 8 10 3 10 10 10 10 9 10 10 9 10 10 10 10 10 10 10 9 4 9 10 8 9 9 8 8 9 10 10 9 9 10 9 8 9 5 10 10 10 9 9 8 10 9 9 10 10 10 9 8 9 9 6 10 10 9 10 8 7 8 10 10 10 9 10 9 9 9 9 7 10 10 10 9 9 10 10 10 10 10 10 9 10 9 10 10 8 10 10 10 9 9 9 10 10 10 10 10 10 10 10 10 10 9 10 10 10 9 9 10 9 9 10 10 10 10 9 10 10 9 10 10 10 9 9 7 10 8 9 10 10 9 9 8 9 8 10 11 10 9 9 10 7 8 9 7 10 9 9 10 9 9 10 8 12 10 9 9 9 9 9 8 9 10 10 9 10 10 9 9 9 13 10 10 10 9 10 9 9 9 10 10 9 9 10 9 10 9 14 8 8 7 6 9 7 8 10 9 8 7 7 9 9 9 9 15 9 10 10 10 9 9 7 10 10 10 10 10 10 8 9 9 16 10 8 9 9 7 7 7 9 9 8 9 9 8 8 8 9 17 10 9 9 9 9 6 5 10 10 9 9 9 10 9 7 10 18 10 9 8 5 9 5 7 8 10 10 8 9 9 7 9 9 19 9 8 9 9 9 6 6 7 9 8 9 8 9 8 7 8 20 10 10 7 9 9 4 5 10 9 10 7 9 9 4 9 10 21 7 6 6 4 7 5 3 6 7 8 5 8 6 6 1 22 7 6 7 5 6 6 2 4 8 8 8 7 8 7 8 7 23 6 6 4 4 4 4 4 3 6 8 6 5 8 5 7 4 24 9 8 5 9 7 6 5 7 9 8 7 10 7 6 6 7 25 10 9 9 7 7 2 4 4 10 10 9 8 9 7 6 4 26 9 7 2 7 5 6 3 6 9 8 5 8 8 8 6 7 27 8 6 8 9 2 4 4 3 9 8 7 9 7 7 9 6 28 9 8 7 8 8 4 5 7 9 8 7 9 9 8 8 7 29 8 7 8 3 9 3 5 5 8 8 9 6 3 10 4 5 30 1 5 1 2 1 1 2 4 3 2 3 4 3 3 31 5 3 3 5 6 4 3 3 7 7 6 7 9 6 5 5 32 2 4 3 5 3 1 1 4 4 3 4 4 4 4 3 33 a 34 2 1 1 3 1 1 7 3 5 5 6 6 5 36 2 3 1 2 1 2 3 4 7 3 4 4 4 6 4 36 1 3 3 2 1 1 1 1 2 7, 4 5 3 3 4 3 37 Correlations of Mental Abilities TABLE Ik Original Scores in Scroll TABLE n TABLE Im Original Scores IN Drawing Original Scores in Esti- Lengths mating Lengths 1 t a g CO 1 1 CQ 8 0. S o ■a H ■+=> 09 S o "a d O •3 CO o Times Correct in 8 Trials -4^ First Trial Second Trial 1 B a o S S o S B o S B o o S B o e B o o 6 d 6 a o^ ^ 1 ■■i o 9S o o o o o to o o 4. ci ^ fe H iz; H 02 > > > o i-H o 1— ( o t-H o 1— I o o o 1 3 347 16 184 35 9 14 12 8 8 8 7 8 8 8 8 2 31 102 22 125 63 21 24 18 7 8 7 4 8 8 7 7 3 10 241 12 242 18 11 4 3 8 8 8 6 8 8 8 8 4 4 440 13 155 68 41 16 11 8 8 8 8 8 8 8 6 5 23 128 6 137 25 8 12 5 8 8 8 8 8 8 8 8 6 52 136 15 170 17 3 9 5 8 8 7 4 8 8 8 6 7 25 136 18 163 20 5 10 5 7 8 7 6 8 8 7 4 8 17 217 24 188 56 15 21 20 8 8 7 5 8 8 8 5 9 340 18 142 17 5 4 8 8 8 8 8 8 8 8 8 10 28 110 15 145 29 7 11 11 8 8 8 7 8 8 8 6 11 22 186 17 151 38 10 20 8 8 8 7 7 8 8 7 6 12 5 207 12 151 94 26 35 33 8 8 8 5 8 8 8 8 13 30 83 26 97 11 6 4 1 8 8 8 6 8 8 8 5 14 39 94 18 119 36 8 14 14 8 8 7 8 8 8 8 8 15 10 219 22 123 35 12 15 8 8 8 8 8 8 8 8 8 16 30 103 58 97 37 10 11 16 8 8 8 6 8 8 8 8 17 5 287 36 155 26 8 9 9 8 8 7 5 8 8 7 6 18 42 172 37 123 101 33 35 23 7 8 7 6 8 8 7 6 19 34 380 62 130 13 2 5 6 7 8 7 6 8 7 6 6 20 7 367 30 232 30 13 11 6 8 8 8 6 8 8 8 5 21 13 425 42 184 56 15 22 19 8 7 5 5 8 8 7 5 22 18 271 30 156 70 17 20 33 8 7 5 5 8 8 7 5 23 33 460 54 168 104 41 34 29 8 8 7 4 8 8 8 7 24 21 144 26 147 32 11 8 13 8 4 8 7 8 8 7 7 25 20 229 50 100 42 19 6 17 8 8 5 6 8 8 2 5 26 49 100 54 137 18 5 5 8 8 8 7 6 8 8 7 7 27 84 212 126 150 22 10 7 5 8 8 8 8 8 8 6 5 28 12 279 21 140 23 9 7 7 7 5 5 4 7 5 4 4 29 108 107 60 155 25 4 10 11 5 7 6 5 8 8 5 7 30 68 217 48 152 22 13 5 4 8 8 8 6 8 8 8 6 31 64 147 67 184 46 9 12 25 8 8 8 8 8 8 8 5 32 75 230 82 175 125 39 47 39 7 6 6 6 7 4 5 7 33 20 200 32 178 47 16 22 9 8 8 6 5 8 8 6 3 34 6 405 4 430 25 7 12 6 5 8 8 7 8 8 8 6 35 18 501 36 151 28 3 14 11 8 7 7 7 7 8 6 5 36 72 155 34 221 32 6 15 11 8 8 8 7 8 8 8 8 37 23 500 33 283 168 40 61 67 8 7 5 5 8 8 6 5 Scoring of Results 29 TABLE In Original Scores in Completing Words — BA Test 1 1-* i 1 r-t 1 bo a a 1 i p 6 1—1 a m i d 12; 1 bO a 1 d 12; s +- nd" -(J a d i2; IN CQ 0) ta m B a CO 1 1 g 1 CO 1 !lO a 2 1 CO ■a 1 i CO s i 1 1 2 1 ■a d 1 1 30 46 30 46 30 195* 30 144 2 30 50 30 72 29 1 126 30 108 3 30 57 30 64 28 2 69 27 3 95 4 30 79 30 54 30 103 30 170 5 30 65 30 75 29 1 97 30 134 6 30 57 30 46 29 1 128 30 89 7 30 99 30 95 30 131 30 1 1 169 8 30 87 30 67 29 1 121 26 4 241 9 30 62 30 48 30 100 30 111 10 30 138 30 75 30 125 30 164 11 30 93 30 65 29 1 132 29 1 235 12 28 1 1 71 30 64 29 1 70 30 124 13 30 71 30 71 28 2 97 30 97 14 30 70 27 1 2 97 30 94 26 1 3 96 15 30 54 28 2 57 29 1 126 28 2 195 16 36 68 30 96 30 110 28 2 130 17 30 146 30 90 28 2 110 30 111 18 30 90 30 77 25 5 95 28 2 124 19 30 166 30 108 30 124 29 1 185 20 30 146 30 155 28 2 212 28 1 1 285 21 30 217 30 177 28 2 232 24 4 2 213 22 26 4 96 23 6 1 125 17 7 6 184 18 6 6 204 23 28 2 322 30 190 30 366 27 2 1 300 24 30 68 30 64 29 1 147 23 3 4 148 25 30 164 30 160 26 28 2 166 29 1 145 22 8 187 27 30 246 26 1 3 340 29 1 268 30 256 28 29 1 102 30 83 26 1 3 260 28 2 212 29 24 5 1 222 27 3 194 28 2 262 21 2 7 336 30 28 2 190 30 205 17 3 465 21 4 5 490 31 24 6 177 17 13 160 14 8 8 202 14 4 12 244 32 22 8 137 23 7 166 20 9 1 172 17 12 1 171 33 19 10 1 290 30 34 30 390 29 1 390 21 1 8 451 19 11 451 35 30 247 30 247 25 1 4 394 30 362 36 26 4 221 28 2 180 15 7 8 482 18 5 7 455 37 15 4 11 222 16 1 13 205 1 29 200 1 1 28 130 *2 min. 5 see. lost on last word. 3° Correlations of Mental Abilities TABLE II Total Scores, and Scores in 1st and 2nd Trials 1 -1^ CQ a CD i-H b b ^ i, S bO tb 60 O o o o o o O H H H C a a g g a -^ " ■^ ■■3 '■B 'S d lS < > "2 "2 c^ C^ c3 CO CO m b H d ^ ^ ^ ;5 1 S5 CQ ^ 5 a, H "W IS c B ■« Oj » o o O X 'Bt{53u9T Siii:jT3raiiea X 'spjojii Snpatdraoo g 'sjrej 3TITUJB8T; X *9jrej SarajEyi Z 'snijoj iBOu^auioaj) X 'euijoji |B0u^8rao9O Z 'anippv X 'SDipPV Z 'BaSBSSB J JO jfjoniajii X 'saSBSSBj JO jtioraajiji X '»ea!t V g 'sa^isoddo jteBg X 'ea^TSoddQ XsBg; g 'spjOjji JO Ajomajv X 'spjo^ JO ^Joraaj^ g 'sai(TSOddo PJ^H X *sa:jTSOddo PJ^H g *ijsa;j snBi^SnTqqa X '^sa^ snBqamqqa H«lWC0COiOiOW*MC0»O^rHN-iliCO«rHi-i(N«-^ ^^T-^^-C5QOlH^D»OOSQOONOiOOOOCOlHrHNNN OS 05 NI-^lH.-HCpN•*lONOOO'-lNWC^OOI-^o^- WN ©OSi-Ht^OT-lOilMOQ^rJIt^cDMOt^QOOSiOt* Wi-HOO {Ot*tOfr*eoX'^co^Niocow^'^co oseoN I CO TO t»cot^t»"^iOCOCOO^tDt»OS ^CO^iCcOCOMNOiH t*COQ0t*"3t^ 03t^-^'Ot»eDC0>0«5^^e0(DC0Tr'TO»O tOt^t>t* t^iO-^tOMt^t^TO-^iO-^^OeDCOTO | | rH to I i-l(MW[NrHNi-«IMrH(MrH WrHNrH(M.H(MTHCQ^(NF- •.-1' ^" ^^ a kI fi S ,3 .. -.- 1= .,. 6= '3= &= 'g' S= P w 6o Correlations of Mental Abilities % tf H pq So Q g 'sq'jaaai SuLM.'eaa; X 'eq^SaaT; 3mM.BiO[ S 'sp-ioji Saijaidraoo X 'spjo^ 3ai:jaidraoo S *sjrej; Sarai'saT; X 'BJiBj SuiTiJBaT 2 *erajo^ jBaii^araoao X 'snijo^j xBOTJ^araoao Z 'Snippv I '3atppv Z 'eaSuesBj p .itioraaj^ X 'sa3'BES*Ej JO Aioman S '^ea? V X '?ea^ V g 'saijisoddo ^sb^ X 'sa?i9oddo ^sb^ 2 'epjo^ JO itioraapi X 'spjo^ JO ^joraapj 2 'Ba^TSoddQ P-i^H X 'sa^^isoddQ P-i'^H 2 *»sai etiBqSaiqqa X '!^ea^ BtiBqSuTqqa I I I *XQ0OMN00i- 1N"*OOCpMCJ 00X0304 0) CO 01 lO t- iH CO -^ "* .J CO CO CO" f-^WCOTj ONi-H CNOOl COCO iH I Tj< I o»oco I It-OiOiOCO ^rHrl^O O«0iM Tt I f 00 (O Ht^COCOO wt-to ■-eortN-T^oeo -i.-i**TjfTjieoc OrHM I I lO b- ■<*< O CO i-t MrHcOCOPpOO OOiOOOOC^W McD'^OWNOOb-OrHOlO 'fltO'-iM.-iw* oo" ll-^Wl-l i-HrH t>.rHCCCOTH I CJl O CO W N Cjl Cjl H.Ht-4 I WiH I r-ltO rHP I I . ! I I Si-HtOOOS ipTtfco-*'" ■ lacot^-^t WCO OrH«DC0OWOSeCOOO«0 IT II Til TT I I I I I I II i-H to iH (o ■-i< Tfi lo « M t^ b- 'if Tjf « rH b- CO ec i>- cq w t>.co ■^C0MiH.-HM-^(DMO«O>-l«'H.-HC0O'H.H0J MM >C3i*NC0t»Nt0CD000>i-iOC0 wcoo i-Hi-HOOMCu-HCOOOO'HOOO'HrHO'-liH OJCOM i-H CO CO t* OS CO --t rH CO O OO O O 03 O) --I O O) OS COMCOOO lilt Tt) Tf lO CO >-( -H 00 (N W 1-1 O CO 0(0«0»0©.HOOJ'*<'^01 lOW^ "ROCOCO 'ti'^M'Hr-* OJ (N tH CO CO to t" Q <-! CM M OOJOCO(NOOOCDCOi-l CO O « --HCD CD CO Tt( M CO t- ^ CD iOrfiiOif5COCD»0>OCOM 00 <-< O "3 CO CO i-h i-( O Q N "H co I I T CO-^Ot*.MNt^i-tO •HO'-H^COOOl'^OOCMb-b-O) C0C0C0M'*Tt0JOJO>^"3iOOM oob-i>»OTino «eOiHin'*co^coM»OM(NiO'-<^r-(r-i CDQOCDOOi COCDt-^MOb-i-HOpCOOiMiOCOW^OiM COiOCDiO^Ji CDiiSeOTttCOCOCOC^^MCOpNOONrHCO C0«5r!OS 0JO>0'H(MCQ0S>-(OC0t»M0lt^0)iH'*i000 xji'*"*'* ■*^(NxJfTj1 1 a % a 1 1 W M !si H < s < 1-1 M w Hh 85 82 84 72 S3 54 58 71 70 65 70 36 42 60 56 60 72 15 IS 33 Hard Opposites ?5 S'> S4 65 54 80 39 49 65 61 -05 44 Easy Opposites T>. 83 65 50 50 56 45 37 62 37 31 A test 54 71 58 70 54 80 50 50 46 46 58 42 87 33 36 30 65 31 13 02 29 Memoiy of Passages 29 66 36 60 60 70 42 56 79- 39 49 65 61 66 45 37 6? 58 87 36 65 42 33 30 31 30 39 77 30 36 38 39 36 44 77 38 44 02 28 -02 07 00 Geometrical Forms 16 51 Completing Words 05 IS 18 -05 37 13 m (\?. ?S -02 07 ?6 33 25 44 31 29 29 00 16 51 05 26 Totals 633 663 618 588 550 484 478 440 452 522 141 289 """ 64 Correlations of Mental Abilities For the remaining two tests, Recognizing Forms and Scroll, it is somewhat less safe to venture an assertion as to the prob- able true correlations with each of the other tests. As has been stated above, the ordinary methods of correction cannot be applied on account of the low reliability of the tests them- selves. The average of the correlations RpiQi, Rpi Rpili, RpiQi, that is, the numerator of the correction formula, gives us on the whole a result considerably less than the raw coeffi- cient, which is itself too small. I have calculated the probable true correlations between the Scroll test and the Recognizing Forms test, with each of the other tests, in the same way as in Table XIV, except that they are calculated on the basis of the raw coefficients instead of the correlated coefficients. The re- sults are: CO 3 .1 CO a -2 •s 8, CQ 1 CQ "S 1 o 2 'S Ph ilO a CO 1 >3 .3 a fn o ^ a, n i? M c 'S ■s Ml a +3 e 1 1 i ^ 1 o 1 a '•3 -r) a i 8 1 p. ■I ■■i W w S w -< s <^ o 1^ Pi m a tt W Recognizing Forms . . 65 45 48 42 28 38 31 32 41 17 30 32 18 Scroll test 35 32 27 22 22 33 42 11 20 17 20 11 09 They are probably at least lo per cent too low, but I leave them as they are, as they are uncertain at best with my data. It will be seen, however, that the Scroll test correlates much less highly with the other tests than does Recognizing Forms, in spite of the fact that on account of its greater reliability, the re- sults as given are less attenuated by chance errors. As is shown in Table XIV the order in which the different tests (exclusive of Recognizing Forms and Scroll) correlate with the other tests is : Hard Opposites, Ebbinghaus, Memory of Words, Easy Opposites, A test. Completing Words, Memory of Pas- sages, Adding, Learning Pairs, Geometrical Forms, Estimating Lengths, Drawing Lengths. Significance of Tests and Analysis of General Intelligence 65 3. Grouping of the tests according to relationships shown by the correlation coefficients (i) Tests of selective thinking Table XIV, Estimated True Correlations, shows us that the correlation between the Ebbinghaus test and Hard Opposites is 85. This high correlation suggests that we are not far wrong in classing these tests together as we have done in calling them both tests of selective thinking. In fact this correlation is al- most as high as the correlation between the first and second trials of each test. (2) Memory tests We see also that the correlation between Memory of Passages and Memory of Words is 80. On this basis we are justified in classing them together as memory tests, implying that they test in large measure the same ability. (3) Association tests It would appear, however, that we would not be justified in grouping Adding and Easy Opposites together as tests of asso- ciation in the sense of implying that they both are tests of the same thing. The correlation between them is only 56, not nearly as high as is the correlation between Easy Opposites and the tests of selective thinking, particularly Hard Opposites which is 83. The fact is that the Easy Opposites test is a test of selective thinking, especially for the Poor group. The Completing Words test is correlated more closely with Adding than with any other test, yy; the correlation between Easy Opposites and Completing Words is 62. Learning Pairs does not correlate closely with Adding, Easy Opposites or Ba and so cannot be classed with them as a test of the same kind of thing. (4) Perception tests On the other hand, the correlation between A test and Geometrical Forms is 87, being very much higher than the cor- relation of either test with any other test. Hence our justifica- tion in grouping them. 66 Correlations of Mental Abilities (5) Motor control The Scroll test has turned out to be quite unsatisfactory as to reliability in the case of the Good group. If we assume that it gives us satisfactory results in the case of the two groups com- bined, and in the case of the Poor group by itself, the co- efficients of reliability being 76 and 71 respectively, we could on this basis estimate the degree of its relationship to the other tests, and to General Intelligence. Just to what extent ability in this test is significant of other sorts of motor ability, we are of course unable to say. (6) Discrimination of lengths The estimated true correlation between Drawing Lengths and Estimating Lengths is only 26. In making this statement, how- ever, we should add that the figures are relatively untrust- worthy on account of the low reliability of the tests themselves, particularly Estimating Lengths, making the correction formula scarcely applicable. However, the very considerable differences between the correlation of random halves of the same test with each other, and the correlation of the two tests with each other, make it clear that there are very appreciably different factors involved in the two tests as given. 4. Order in which abilities correlate with other abilities tested In order to determine which of the tests correlate most highly with other tests, we have simply to sum up the totals of the columns given in Table XIV. The average correlation of each test with the eleven other tests is : Hard Opposites 60, Ebbinghaus test 58, Memory of Words 56, Easy Opposites 53, A test 50, Completing Words 47, Memory of Passages 44, Adding 43, Learning Pairs 41, Geometrical Forms 40, Estimating Lengths 26, Drawing Lines 13. On the basis of our separate estimate, the figures for Recog- nizing Forms and Scroll would be 41 and 26 respectively. Grouping the related tests we have : Average correlation of Selective Thinking with other tests 59 Average correlation of Memory with other tests 50 Significance of Tests and Analysis of General Intelligence 67 Average correlation of Association (exclusive of Learning Pairs) with other tests^ 48 Average correlation of Perception with other tests 45 Average correlation of Motor Control with other tests. . 26 Average correlation of Discrimination of Lengths with other tests 19 Provided only we have appropriately named the abilities above mentioned, all this means that, using the argument from correlation alone, power of selective thinking is more intimately connected with, and more characteristic of, general mental ability than is any of the other abilities tested; that memory is next most highly correlated with general ability; the simpler forms of association next; perception next; motor control con- siderable less; and discrimination of lengths least of all. This confirms the more direct argument stated above, from the ex- tent of overlapping of the Good and Poor groups in the different tests. 5. Analysis of the individual differences revealed by the tests and evidence that they are largely due to differences in native mental capacity We are now in a better position to analyse the differences between the two groups that are revealed by the tests, and to indicate to what they are due. The most obvious difference between the two groups of persons is of course a difference in efficiency. That the members of the Poor group were relatively inefficient is shown by the nature of their previous employments, and by the fact that all but two of them were out of regular em- ployment. On the whole it is the more inefficient ones who are the first to be thrown out of employment. While an occasional one loses a good position owing to some bad habit such as in- temperance, dishonesty, etc., the many lose their positions owing to lack of initiative and zeal, stupidity, inability to make them- selves so useful that they cannot be easily replaced if dis- missed; in short, through general lack of efficiency. Moreover, not one of these persons, with a possible exception of No. 18, the brakeman who was then assistant manager of the Salvation Army Industrial Home, had ever held a position calling for ' Including Learning Pairs, 46. 68 Correlaticms of Mental Abilities much exercise of intelligence, though they had reached years of maturity. The median age was 36, ranging from 23 to 52. As to the amount of schooling they had had, there was considerable divergence, the median number of years schooling being 7 and ranging from i to 12. With regard to those who had a small number of years' schooling it is safe to assume that in a rather large proportion of cases inability to get on well at school had a -good deal to do with it. On the basis of these facts it seems clear that the twenty members of the Poor group represent a grade of intelligence and mental capacity very far below the average. In the opinion of the writer, who spent from four to six hours with each of them and had an excellent opportunity to judge the characteristic methods of each in attacking mental difficulties, it is practically certain that at least fifteen of the twenty represent that quality of mind which ranges from some- what below the average in mental capacity to that which is only slightly above the feeble minded. They represent persons who, on the whole, are not at all bright, even if not positively dull in school; who have no particularly strong interests or am- bitions, and who are likely on leaving school to follow the line of least resistance in earning a living, doing most things they attempt with indifferent or poor success. They represent the kind of person relatively lacking in foresight, energy, self-re- liance, and all round mental capacity; the person likely to suc- ceed fairly well only if well directed and especially trained in some particular line of work, but who, under the most favorable conditions possible, is utterly incapable of attaining a degree of achievement much beyond mediocrity. At least six of these were persons considered dull by their acquaintances. As to the remainder of the group, in the opinion of the writer, they repre- sent persons ranging in mental capacity from that of the fifteen described above to something above the average in the case of at least one or two. The Good group, on the other hand, is composed of persons who are far above the average in mental capacity. It is safe to say that persons who become college professors or instructors before the age of thirty-five possess mental ability ranging from considerably above the average toward the very highest mental SigW^cance of Tests and Analysis of General Intelligence 69 capacity. That the members of the Good group are representa- tive of mental abiHty far above the average is evidenced by the nature of the positions they now hold. Eight or nine of the seventeen at present occupy college or university positions. The rest hold educational positions, mostly in normal schools of high rank, with the exception of one, who holds an important and responsible position in a philanthropic organization. There can be no reasonable doubt, therefore, that the two groups are repre- sentative of very decided contrasts of mental ability. One may say that the qualities making for success are in large part moral — capacity for self-control, self-denial, indus- try, conscientiousness, etc. Some of these qualities or tendencies are such as to be indicated to a close observer, in the way in which the mental tests were attacked. Analysis would un- doubtedly show positive correlation between mental and moral qualities. In addition, one of the most striking differences be- tween the two groups was a decided contrast in attitude with regard to time. All of the Good group did the tests with a high regard for the value of time, while such a thing as the time, element being of much significance or importance hardly seemed to dawn upon members of the Poor group, even though the instructions given were the same for both groups. Having considered group differences on the basis of general principles, let us now consider the differences in the two groups as revealed by the tests. As above stated, page 53, the most decided difference between the groups is in the tests of selective thinking — Ebbinghaus test and Hard Opposites. This is exactly what the results obtained in investigations on the men- tality of the feeble minded would lead us to expect. All of those who have worked with the feeble minded agree that they are farthest removed from the normal in ability to deal with the abstract, ability to get and use meaning and significance. It is also demonstrated that the most painstaking educational efforts to improve their capacities in this respect are relatively ineffec- tive.^ The same would undoubtedly hold true, though to a less degree, of persons somewhat above the mentally defective. It may perhaps be objected that the two tests in question are simply language tests, and, as such, measures of amount of train- ' See Goddard, Journal of Educational Psychology, November, 1911. 70 Correlations of Mental Abilities ing and education rather than measures of mental capacity. That this is not the case is evidenced by the low correlation be- tween number of years schooling and rank in the tests. The correlation between number of years schooling and rank in the eight tests correlating most highly with other tests is 38. This amount of positive correlation could be accounted for on the basis of the correlation between mental ability and staying on at school. Studies in retardation have shown that a considerable proportion of those who leave school at an early age, do so be- cause they are relatively unable to do the mental tasks required of them in school. In other words, in a considerable propor- tion of cases, a small number of years schooling means inability to learn advanced and difficult language work, rather than lack of opportunity to learn it. There is no way of calculating directly the correlation between general intelligence in the Poor group, and rank in the tests of selective thinking, as we have no way of independently ranking them in general intelligence. In the Good group, however, the correlation between estimated intelligence and rank in the tests of selective thinking is 92.5. In the Poor group it would probably be higher still as in general all other correlations are. Hence the tests of selective thinking do not measure mere training and schooling. This was further evidenced by a consideration of individual cases. In general those who were considered decidedly dull or stupid by their fellows, did poorest in the tests of selective thinking. Moreover, language tests similar to those here given are not unfair tests of ability as opposed to mere education, since the intelligence of a primitive people can be gauged by the language they find it necessary to evolve and use. Feeble-minded chil- dren are, on the whole, decidedly deficient in acquiring higher forms of language, while bright children with similar educational advantages, acquire language naturally and easily. It is the same with many of the Poor group. Their low grade of native mental ability has made them very slow to acquire average facility in the command of the higher form of language, and very difficult for them to acquire and make use of abstractions, fine shades of distinction in the meaning of words, etc. These Significance of Tests and Analysis of General Intelligence 71 differences in native capacity come out in any classroom of an elementary school. In a certain sense all ability is dependent upon training and practice inasmuch as one cannot long retain a capacity unless it is more or less exercised. The skill of the talented musician could not be achieved without attention given to music. But a true way of viewing the difference between a musical genius and a person of average musical ability is just this: The in- terest and inborn capacity of the one is so great that it leads him to get abundance of practice and training in the capacity, where- as the weaker interest and inborn capacity of the other leads him to let the practice of other abilities and capacities predomi- nate. So with interest in, and ability to deal with the abstract. Gifted minds early become interested in mental exercises of a kind for which the dullard has no interest. Interest in fairy tales, romance, fiction, literature, science and philosophy, in so far as these involve the use of concepts, would lead to the ex- ercise and development of the higher forms of association and abstraction. In short, we may well look to language tests in some form, to furnish good tests of general intelligence. Moreover, the high correlations within the Good group itself, where there has been presumably no appreciable inequahty on the basis of opportunity for education, further substantiate the view that by far the most influential factor in making for effi- ciency in these tests, is the native capacity of the individual in question, and not simply his training and environment. The results in memory and association show nearly as much dif- ference between the two groups as do the tests of selective thinking, and the same general line of argument would hold with regard to them. To test still further the extent to which ability in certain of the tests is significant of general intelligence as commonly under- stood, there was taken for each individual a combined measure obtained by summing up his score in the Ebbinghaus test, Hard Opposites, Easy Opposites, Learning Pairs, and Recognizing Forms, arranged in such a way as to allow approximately equal weight to each test. The scores of the different subjects, in terms of the deviation from the median of all are: 72 Correlations of Mental Abilities Subject Deviation Subject Deviation 1 89 18 2 61 19 —2 3 95 20 —15 4 49 21 —26 5 70 22 —9 6 43. 23 —26 7 39 24 —27 8 54 25 —31 9 59 26 —25 10 21 27 —13 11 53 28 —16 12 44 29 —32 13 42 30 —59 14 25 31 —45 15 49 32 —51 16 47 33 —119 17 21 34 —54 35 —59 36 —73 37 —127 There is no one of the Poor group who reaches the lowest one of the Good group; in fact there is a decided gap of 21 here. There is a much larger gap between the median of the Good group and the highest one of the Poor group, namely, 49. Thus we see how effective such a combined score is in separat- ing the Good group from the Poor. TABLE XV Ranks of Good Group for Imputed Intelligence ■8 e .0 ■0 «) "♦^ Ol .« •«* Ji 8 03 ■s 1 ^ ^ ^ ^ ^ ^ J2 ^ ^ ^ ^ J2 J2 tic 1 1 1 1 ^ 3 1 C 1 M a § 1 7 7 6 15 7 6 11 7 8 6 5 6 6 5 11 9 3 9 6 5 5 3 2 1 2 7 2 6 6 2 4 7 3 2 3 3 12 10 9 5 8 7 7 9 13 10 11 12 8 9 6 8 11 14 9 5 5 4 7 3 3 12 5 8 7 8 4 4 16 15 8 10 16 15 15 15 13 12 9 9 6 9 7 5 5 4 14 9 6 8 2 3 3 3 4 4 3 3 4 6 3 5 7 4 4 10 16 11 10 11 10 13 14 11 16 12 llj 9 9 10 13 9 4 4 7 9 6 10 10 9 11 11 12 10 12 11 13 10 11 17 17 11 12 11 14 13 10 12 8 17 11 13 12 14 14 8 13 2 8 8 8 6 8 5 12 8 7 8 4 4 2 4 5 2 2 3 2 2 2 3 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 13 7 5 10 8 8 Significance of Tests and Analysis of General Intelligence 73 To further test the extent to which such a combined score is a measure of ' general intelligence,' the individuals of the Good group were rated in order of merit for general intelligence each by the rest of the group, so far as was possible, four years after the tests were taken. The rankings, including two rankings by the experimenter, made a month apart were as shown in Table XV. Taking the individuals who were ranked by ten or more persons, and using for each two random halves of his ranks, and all together, we have the following: S +? 1 5 1 (U -0 4J e*^ --3 ■s J2 Xi iH (M "3 % A 3 02 3 02 Si CO I— I T3 a IN % 6 12; d rt 1 A A A P.E. P.E. P.E. 1 ranked above 4 by 100% 100% 100% 3.5 3.5 3.5' 4 3 " 50 66.6 58 .64 .30 3 ' 2 " 50 82 66.6 1.36 .64 2 5 " 66.6 66.6 66.6 .64 .64 .64 5 8 " 66.6 80 73 .64 1.25 .91 8 13 " 66.6 75 73 .64 1.00 .93 13 14 " 66.6 66.6 66.6 .64 .64 .64 14 ' 9 " 50 83.3 66.6 1.43 .64 9 12 " 40 83.3 64 — .38 1.43 .53 12 10 " 100 80 89 3.5 1.25 1,82 10 17 " 100 80 89 3.5 1.25 1.82 ' Inferring from the percentages the distances between the individuals in question in terms of the P.E. by means of the following table quoted from page 16 of "The Perception of Small Differences" by Fullerton and Cattell. %r P.E. %r P.E. 7or P.E. %r P.E. %r P.E. 50 .00 60 .38 70 .78 80 1.25 90 1.90 51 .04 61 .41 71 .82 81 1.30 91 1.99 52 .07 62 .45 72 .86 82 1.36 92 2.08 53 .11 63 .49 73 .91 83 1.41 93 2.19 54 .15 64 .53 74 .95 84 1.47 94 2.31 55 .19 65 .57 75 1.00 85 1.54 95 2.44 56 .22 66 .61 76 1.05 86 1.60 96 2.60 57 .26 67 .65 77 1.10 87 1.67 97 2.79 58 .30 68 .69 78 1.14 88 1.74 98 3.05 59 .34 69 .74 79 1.20 89 1.82 99 3.45 74 Correlations of Mental Abilities Calling the degree of ability midway between subject No. 8 and subject No. 13 the central tendency, we get Table XVI. (A case of 100 per cent is treated as equivalent to 3.5 P. E. because there is a high probability that with many judges the lOo's in these results would fall to 99 or even much lower.) TABLE XVI Deviation Measures Inferred from Per Cents of Judgments OF Superior Deviations Deviations Deviations No. of from Median, from Median, from Median, Subject 1st half 2nd half entire set 1 5,1 7.9 6.5 4 .60 4.39 2.94 3 1.60 3.75 2.64 2 1.60 2.39 2.00 5 .96 1.75 1.36 8 .32 .50 .45 13 — .32 — .50 —.46 14 —.96 —1.14 -1.10 9 — .96 —2.57 —1.74 12 — .58 —4.00 —2.27 10 —4.00 —5.25 —4.09 17 -7.5 —6.50 —5.91 Using this deviation table, the coefficient of correlation cor- rected for attenuation was calculated between ' general intelli- gence ' as judged by one's fellows, and the record secured in the five combined tests. This correlation is 92. In order to test the reliability of this combined score, the co- efficient of correlation was calculated between the combined score for the first trials, and the combined score for the second trials. For the Good and Poor subjects together the raw co- efficient is 96; for the Good group (12 subjects only), 72; and for the Poor, 90. This high reliability of the combined score proves that fifty minutes or so spent in getting such a record, gives us a good measure of whatever it is that this combined score measures. That this, in turn, gives us a very significant indication of the general mental ability of the individual in ques- tion, is shown by the almost perfect correlation between score- in-the-combined-tests and estimated intelligence. Again the Pearson coefficients of correlation corrected for at- Comparison of Results With Those of Other Investigators 75 tenuation were calculated between Estimated Intelligence and each of the eleven most reliable tests. The coefficients are : Estimated Intelligence and Hard Opposites 96 Ebbinghaus test 89 Memory of Words 93 Memory of Passages 35 Easy Opposites 82 Adding 72 Learning Pairs 34 Completing Words (?) 100' A test 21 Geometrical Forms 07 Drawing Lengths —20 (These coefficients are calculated of course on the basis of only twelve of the Good subjects.) Here again is indicated how reliable as tests of intelligence are the Ebbinghaus test, Hard Opposites and Easy Opposites. The order in which the diflferent capacities correlate with gen- eral intelligence is practically the same as shown in Table XIV, and further substantiates the results there stated. VII. Comparison of Results With Those of Other In- vestigators One of the earliest attempts to give an exact quantitative statement of mental relationships was that of Wissler ('01). The correlations obtained by Wissler, however, are on the whole much too low, owing largely to the fact that about twenty dif- ferent capacities were tested in less than fifty minutes, making the measure of each person's capacity in each trait altogether in- adequate. It is impossible on the basis of Wissler's data to es- timate how much the correlations he found would be raised by correction for attenuation due to inaccuracies in the measures themselves. In the case of the correlations obtained by Aikens and Thorn- dike ('03), also, the absence of data makes it still impossible to estimate how much they are influenced by attenuation. Impor- tant later studies demand detailed consideration. Norsworthy ('06) compared mentally defective children with normal ones by the use of tests similar to those here used. The mental capacities tested were: i, Ability to form abstract ideas 'Not sufficiently reliable; as the reliability coefficients of the completing words test in the Good group is only 27. 76 Correlations of Mental Abilities (recognition of nouns) ; 2, Ability in appreciating relationships and in controlled association (part-whole test, genus-species, easy opposites) ; 3, Memory (words and sentences) ; 4, Ac- curacy and speed in Perception (A test and a-t test) ; 5, Percep- tion of weight (accuracy in judging relative weights) ; 6, Motor control (maze and form board). Numbers i and 2 are com- parable with our selective thinking and association tests, number S is probably somewhat akin to accuracy in judging lengths, and the maze test in number 6 is very similar to our scroll test. Miss Norsworthy found that the defectives were farthest removed from normal children in ability to deal with abstract data. This is shown in the following table, compiled from re- cords of about 137 cases, ranging in age from eight years up : % above median for ordinary children % above — 1 P.E. (or lowest 25% of ordinary- children) % above —2 P.E. (or lowest 9% of ordinary children) 1. Height 2. Weight 3. Pulse 4. Temperature 5. Weight test 6. A test 7. a-t test 8. Memory of unrelated words . . 9. Composite of 5, 6, 7 and 8 . . . 10. Dictation 11. Memory of unrelated words. . 12. Part-Whole test 13. Genus-Species test 14. First Opposite test 15. Second Opposite test 16. Composite of 13, 14, 15 and 16 45 44 49 26 18 9 1 6 1 10 61 66 69 59 28 18 14 18 15 10 19 17 16 0.9 1 1 77 77 86 77 39 34 28 27 27 21 30 27 17 5 7 10 According to her results, the order in which the different abilities tested would correlate with intelligence is roughly as follows: abstraction and association; memory; various forms of perception; motor control. This agrees with our results in so far as they can be definitely compared. Miss Norsworthy's tests of abstraction and association would seem to be very similar, and to require almost equal powers of abstraction; hence the somewhat higher rank relatively of association as com- pared with our results. Abstraction is of course a relative term. Comparison of Results With Those of Other Investigators 77 What would be severe trials of abstraction to a feeble minded person, or to an immature mind, might be mere tests of rapidity and ease of association and memory to an abler or more mature person. As we stated elsewhere, the Easy Opposites test to many of the Poor group, was like the Hard Opposites test for the Good group. Lewis M. Terman ('06) tested seven of the brightest and seven of the dullest boys of a group of five hundred elementary school pupils. The capacities tested were: Powers of Invention and Imagination as tested by ability to solve mental puzzles. Logical Processes as tested by solution of problems in arith- metic and other problems requiring original thinking. Mathematical Ability as tested by problems requiring the more mechanical phases of arithmetic. Mastery of Language as shown in word building, reading, Ebbinghaus Mutilated Text, spelling and facility in inter- preting commands. Insight as shown in the interpretation of fables. Ease of Learning the game of chess. Memory of geometrical figures, chess moves, a story read, and the solution of a mechanical puzzle. Motor Ability, as shown in learning visual-motor coordina- tions, skill in running down stairs, and in carrying a book on the head. Terman concludes that the bright boys are superior to the dull in all the mental tests, and inferior in the motor. This su- periority of the dull group in motor tests, if characteristic of persons in general, would of course mean a negative correlation between motor ability and " intelligence," and be in conflict with our conclusion of a small but positive correlation. It is cer- tain, however, that at least a part, and probably all, of this superiority of the dull group in motor ability was due to greater maturity on account of their being on the average 12.7 months older than the bright group. Again, it may be, in part, that boys in school who have some motor ability and interest in that kind of thing, are on that account more likely than other boys to give little attention to linguistic and abstract studies, and for that reason are likely to be considered less intelligent by their 78 Correlations of Mental Abilities teachers than they really are, and duller than boys who lack such interests altogether. The fact that the dull boys all pre- ferred games and the bright ones reading also suggests this con- clusion. On general principles it seems hardly likely that there should actually be a negative correlation between motor ability and intelligence, at least as far as native capacity is concerned. It would seem quite probable, too, that the groups here selected would not represent bright and dull children respectively as judged from a more general point of view than that of a school teacher. I have calculated the Pearson coefficient of correlation for each test with every other test, using Terman's scores for each of the fourteen boys taken all together in one group. In order to bring out the relationships between the different tests as fully and accurately as possible, I have taken the actual scores made instead of the ranks. The only exception to this was in the case of the ' Interpretation of Fables,' where on account of the omis- sions it was not very practicable to take the scores. The co- efficients are given in the table below. CQ J2 e5 Ph s "S § g g >» -M o ■s Ph ti a XI 1 1 1 o 1 i < o 1 81 86 75 78 72 66 — 36 Chess 81 86 75 82 77 82 73 77 73 65 74 62 78 64 .■;8 62 70 63 — 23 Mathematics — 48 Logical Processes —22 78 65 74 62 58 72 — 26 Invention 7?! 78 64 58 58 36 — 14 Fables (interpretation of) . 66 62 70 63 72 36 —52 Motor Ability —86 —23 —48 —22 —26 —14 —52 422 422 401 386 383 350 317 —221 The table gives of course the raw coefficients, and correction for attenuation would raise them somewhat. On the other hand these fourteen boys, like our thirty-seven Comparison of Results With Those of Other Investigators 79 men, represent a selection half from near the top and half from near the bottom of the scale of general intelligence. In so far forth the correlations are too high. They should be thought of much as the raw correlations for the Good and Poor together in our own experiment. The high intercorrelation (with the ex- ception of the motor tests where the negative correlation in evi- dence has already been accounted for) are in harmony with our results. W. C. Bagley ('01) reported a negative correlation between motor skill and intellect. Professor E. L. Thomdike^ has shown that this result was due to an arithmetical error, and to over- sight of the influence of the age factor. When children of nearly the same age are taken, Bagley's data give a slight posi- tive correlation between mental and motor ability. Binet ('99) has attempted to differentiate intelligent pupils from unintelligent by the use of tests of voluntary attention. He considers that the tests he used are not, properly speaking, tests of comprehension, but depend upon processes relatively more simple, — especially acts of memory and comparison. He tested eleven subjects, five intelligent and six unintelligent. The tests included accuracy in tactile sensibility; quickness of re- action time; speed and accuracy in counting dots; perception of change in the rate of speed of rhythmic movement ; counting of rhythmic sounds; copying figures, different varieties of sen- tences and drawings; rapidity of perception of words and figures seen for a fraction of a second; accuracy and speed in picking out one or more letters from a printed page (our A test and an extension of it) ; and simultaneous adding. Binet found that a number of the tests did establish a clear differentiation between the bright and the dull pupils, and that they are therefore promising as tests. Those which turned out to be the best in this respect he found to be the tests of accuracy in tactile sensibility, counting rhythmic sounds, copying figures, sentences and drawings, memory of figures, and cancellation of letters. He found also that this difference is most marked at the first trials, but diminishes with subsequent trials, and that in some cases it may be effaced. On account of the small number of Binet's subjects, and on account of the fact that, sev- ' Educational Psychology, 1903 edition, pp. 148, 149. 8o Correlatiqns of Mental Abilities eral of the tests being new, the subjects' records are hard to evaluate with accuracy, it is unsafe to make very specific com- parisons. However, subject to these limitations, the group dif- ferences brought out by the tests are shown in the two following tables. Table XVIII gives a comparison of the average per- formance of the bright group with that of the dull, on the basis TABLE XVII Bright Dull Tactile perception, errors with points 2 cm. apart... . Counting of dots, errors in score Counting of sounds, errors in score Copying figures, average number Memory; numbers forgotten in 5 numbers of 5 figures Reading through an opening, errors (Rapidity of Apperception) Simultaneous adding, errors Cancellation of letters, errors Simple reaction time 20% 64% 3.4% 5% 4.3% 20% 3.6 2.8 12 42 34 34.5 19% 23% 9% 27.5% 24 24% TABLE XVIII R.4.NKIN6 OP Pupils in the Different Tests O ■*^ m V » ff a ° 1 s ft ^ ^ 3 "« g X> Q 02 < CO 3 ■s •s ° o s ■? an -s ■■a to a 1 ^ 2 a 3 1 1 a 1 ft 5 1 1 S A 1 7 1 3 5 3 2 2 6 B 1 2.5 3.5 1 1 4 1 4 7 C 3 2.5 3.5 2 4.5 6.5 6 5.5 8 D 6 7 3.5 8.5 9 6.5 3 5.5 9 E 3 7 2 6 6 8.5 4 7, 1 F 6 2.5 8 11 11 8.5 5 2 2 G 6 5 5 10 ■2.5 1 5 2 2 H 8 10 9 4 4.5 2 7 10 10 I 9.5 2.5 10 5 7.5 5 .10 5.5 5 J 9.5 9 6 8.5 10 10 9 8.5 3.5 K 3 9 7 8.5 7.5 11 8 8.5 3.5 No. of group 1 3 2 2 3 1 2 4 displacements. Comparison of Results With Those of Other Investigators 8i of the first trial of the test only, in order to rule out the factor of mental adaptation through practice. Binet lays great stress upon the conclusion that the difference between the intelligent and the unintelligent consists in quick- ness of adaptation. The bright pupils adapt themselves more quickly than the dull; the dull adapt themselves in turn, but more slowly. In so far as this were true, other things being equal it would imply, with reference to correlations between different tests, that they should be relatively high with unprac- ticed subjects, and would gradually diminish as the subjects had more and more practice. Spearman and Krueger in working over Oehrn's results from ten subjects, including five medical doctors and three medical students, found that up to a certain point correlations increase in size as practice increases, and that after that point they again tend to diminish. The decrease in this case may be due to fatigue entering as a disturbing factor as the test is prolonged. The change in the size of the correlations as practice proceeds is shown in the following table. TABLE XIX Uncorrected Coefficients of Correlation for Each Successive Quarter Hour, According to the Results of Oehrn Quarter Hour Writing and Adding Writing and Counting Writing and Reading Writing and Learning by Heart .... Adding and Counting Adding and Reading Adding and Learning by Heart . . . . Counting and Reading Counting and Learning by Heart. . . Reading and Learning by Heart 1st 2nd 3rd 4th 5th 6th 7th 8th 50 68 72 65 64 68 55 68 58 67 70 75 81 71 54 58 32 42 51 53 48 38 42 47 10—02—03—03 03 02 25—08 37 56 69 67 64 59 50 31 01 14 24 18 05—18 22 26 22 24—09—02—26 00—07—13 —17—16—04 05 14—10—26—21 •24 —22 —27 —23 —15 —02 —16 —21 —05 07 08—10 19 05 03—10 That increase or decrease would depend upon the kind of test and the stage of the subjects in the learning process, and be sub- ject further to such disturbing factors as fatigue, change in zeal, etc., seems quite evident. In concluding. Spearman and Krueger say that the lack of being accustomed to the test instead of act- ing as a cause of correlation, would act as a disturbing factor 82 Correlations of Mental Abilities lessening it. Which effect it would have, however, would surely vary according to circumstances. If conditions were such that the more intelligent could in the time allowed acquire a method that they could use to advantage, while the poorer ones did not get to that point, the correlations would be correspondingly high. If, on the other hand, conditions were such that even the best of the subjects were unable in the time allowed to hit upon and use effectively any intelligent method, their intelligence would count for little or nothing in determining the records, and the correlations would be correspondingly low or zero. Difference in quickness of adaptation is certainly in evidence in Binet's results, one illustration of which is given in the fol- lowing table: NuMBEK or Letters Marked Number of Errors Bright Dull Bright Dull 1st period of 5 minutes 2nd " "5 " 3rd " "5 " 4th " "5 " 5th " "5 " 115 147 176 183 213 124 128 179 171 207 13.5 10 9 3.5 6 43 32 15 8 7.5 The improvement of the dull pupils in this cancellation test shows itself particularly with respect to the quality of the work. But difference in quickness of adaptation would seem a priori to be a characteristic applying to the learning of something of which both groups are capable, rather than to the factor of fundamental importance, differentiating the intelligent from the unintelligent. Of course a lower type, of mind does adapt itself more slowly than a higher type of mind, but a more funda- mental difference between the inferior type of mind and the su- perior type, exists in the fact that there are certain kinds of performance to which the inferior mind is relatively incapable of adapting itself at all, such as the higher forms of abstraction. The tests here used by Binet are not such as to bring out this difference with any degree of clearness. They are largely along the line of perception and memory. The nearest approach to testing facility in the higher mental processes akin to abstrac- tion is in connection with the copying tests. The fact that there Comparison of Results With Those of Other Investigators 83 is so decided a difference between the two groups with regard to the logical grouping of phrases in copying the easy and the difficult prose passages, suggests a difference which does not belong primarily under the category of quicker or slower adaptation. The method of this investigation and the tests used are sug- gestive, and worthy of leading to a more extended investigation along this line; and it would surely be worth while for some- one to take those of the tests found to be most satisfactory, and see what their reliability is with a larger number of subjects. Then if correlations between the different tests were calculated, more definiteness would be given to the work already begun. However, what we need most of all is the devising of tests of different forms of the highest mental processes — ^performances such that the duller minds cannot learn them so readily as they do feats of perception. In copying the easy material the bright pupils copied 4.5 words at a time and the dull pupils 3.4; in copying the dif- ficult material the bright copied on the average 2.8 words and the dull 2.4. The grouping of phrases was far more logical with the bright than with the dull. On the other hand, in the copy- ing of nonsense material the dull pupils did fully as well as the bright. The difference here would seem to be one of appercep- tion. When it came to copying nonsense material both groups were equally at a loss, as neither had at hand any means of or- ganizing the material. Doubtless continued practice at this would have revealed a quicker adaptation, and an earlier rise in the curve of learning in the case of the bright pupils than in the case of the dull, owing to their being quicker to find a method of organizing the new material. However, we could not hope to arrive at the true correlations of an ability with general intelligence if we took our individual records at a point in the learning curve where the dull group were still practically at zero efficiency, while the bright group had just made a sudden rise. Any such decided unevenness in the learning curve, un- less eliminated by averages, would seriously interfere with the securing of the true correlation between the characteristic ability of the group in that capacity, and general intelligence. Binet finds that in tactile sensibility the intelligent are de- 84 Correlations of Mental Abilities cidedly superior in the first trials to the unintelligent. It is questionable here whether quickness in comprehending the in- structions was not a prominent factor in the results secured. When the dull can, in so small a number of practice trials, al- most equal the bright, who have meanwhile had an equal number of practice trials, it seems evident that the ultimate and real difference between the two groups in this capacity is not a large one. The difference with respect to speed of adaptation is a general factor that would tend to cover up or disguise the real correlations between different capacities themselves, and not aid materially in the detailed diagnosis of intelligence. It is sug- gestive, too, that in some of the tests the dull pupils did as well as the bright. May this not mean simply that the correlations between this sort of ability and general intelligence are too small to be brought out by such crude measurements and treatment of them as Binet employed ? Hence to speak of the difference between inferior and su- perior minds as fundamentally a difference in power of volun- tary attention, seems misleading in that it is too general a state- ment. It suggests that voluntary attention is a capacity that can be applied equally well in any desired direction, to the ex- tent of one's general ability. It errs in implying that there are not various kinds of volvtntary attention, that do not correlate perfectly with one another. On the contrary, there are many varieties of voluntary attention — if we are to express the facts in Binet's terminology — for one might give a degree of atten- tion to music that he could by no means give to mathematics or to painting. Difference in quickness of adaptation there cer- tainly is, and this difference is one of no little significance; but it is one of degree. A more significant difference between inferior and superior minds is one so pronounced as to suggest a dif- ference in kind, rather than one in degree merely. By this we mean that there are certain kinds of mental feats that can be performed by the able mind that can scarcely be performed by the inferior grade of mind at all, let its pos- sessor practice at it as much as he will. This sort of distinc- tion is not suggested by Binet's way of stating it — that the de- termining factor is a matter of voluntary attention. Comparison of Results With Those of Other Investigators 85 Spearman and Krueger ('06) published the results of an in- vestigation on the correlations existing among the following mental abilities: Touch Discrimination, Tone Discrimination, Adding, Learning by Heart (a series of numbers), Ebbinghaus Mutilated Text. Their subjects were eleven advanced univer- sity students, four of whom did not speak German as their mother tongue. As to correlation between the two trials of the same test, Spearman and Krueger's coefficient for Adding is 76; ours are 91 for the combined groups, 76 for the Good, and 90 for the Poor. This agreement is close. In the Ebbinghaus test their coefficient of reliability is 76 ; ours are 92, 96, and 90 — consider- ably higher doubtless on account of the greater number of trials and the method of scoring. In Sensory Discrimination their re- sults on Touch Discrimination and Tone Discrimination may be compared with ours on Drawing Lines. For Touch Discrimina- tion their coefficient is low, namely 42. For Tone Discrimina- tion their coefficient is 87, as compared with ours for drawing lines 72, 42, and 95. Their coefficient for Learning by Heart, 92, is a good deal higher than ours for Unrelated Words, 73, 48, 49; and somewhat higher than ours for Memory of Pas- sages, 90, 78, 83. It is well that our attention has been called to the fact that very different results may be obtained by two different experi- menters using the same tests. However, the fact that there is such close correspondence as to reliability coefficients, where they are at all comparable, would go to show that when the test is itself satisfactory, and where the experimenter understands his business, there is no necessity that the investigations be con- ducted by two different experimenters. This precaution applies rather to the use of a new and untried test, not to one whose method of procedure is clearly understood, provided it is other- wise reliable as a test. Once the conscientious investigator is aware of the possibility of inaccuracy from such a source, he should be able, with our present knowledge of former errors in the conduct of mental tests, to secure by himself results that will not be vitiated by the personal equation. 86 Correlations of Mental Abilities The Pearson coefficients of correlation which they obtained among the abiHties tested are as follows : Learn- ing by Heart Raw Coefficients: Adding Ebbinghaus test Tone Discrimination . . , Touch Discrimination . . Learning by Heart COEEECTED COEFFICIENTS Adding Ebbinghaus test Tone Discrimination . . . Touch Discrimination . . Learning by Heart Ebbing- Tone Touch Adding haus Discrim- Discrim- test ination ination 79 67 19 79 59 00 67 59 29 19 00 29 14 —07 17 —13 70 68 ? 70 64 ? 68 64 ? ? ? ? ?- ? ? ? 14 —07 17 —13 Thus between Adding and Ebbinghaus test they get a correla- tion of 79 raw, and 70 corrected; while ours are 71, 61, 65 raw, and 71, 55, 63 corrected. It is interesting to note that with their results as with ours, the coefficient of correlation becomes lessened when corrected. One would expect their coefficient of correlation to correspond approximately with ours for the Good group. It may be, however, that their group represents persons of considerably greater variability than our good group, in which case the correlation would be higher. In one respect at least their results would be much less reliable than ours, namely, in the number of subjects tested. They used eleven subjects for the Adding test and only seven for the Ebbinghaus test, the foreigners being excluded. Then, too, it may be that as the sub- jects were not all of the same nationality, the correlation of Adding with other capacities may be largely influenced by dif- ferences of practice in adding, characteristics of the different countries represented. Spearman and Krueger did not give the corrected coefficients for Learning by Heart and the other four abilities tested. They put them down as zero because they are less than five times the P. E. Their raw coefficients are on the whole much lower than Comparison of Results With Those of Other Investigators 87 the correlations we find between Memory of Words and Adding, Ebbinghaus test, and Discrimination of Lengths, as will be seen in the following : Raw Pearson Coefpicients — According to Our Results Combined Goods 49 21 94 54 21 —01 28 19 Poors Memory of Words and Adding " " " Ebbinghaus test " " " Drawing Lengths " " " Estimating Lengths. . 22 66 05 04 Spearman and Krueger conclude that there are beyond doubt positive and fairly high correlations between abilities as varied as Discrimination of Tones, Adding, and the Ebbinghaus test; and on the basis of Oehm's data, between Speed of Reading, of Writing, and of Counting; and that the size of these different correlations is to be explained on the basis of the degree of their connection with a hypothetical, common central factor. They think there is good reason for believing that the central factor is not to be explained by individual differences in the zeal of the subjects, their momentary disposition, their being accustomed to the conditions of the experiments, their ability to make the most of help given, or to the power of their attention. They think that the explanation is in all probability psycho-phys- ical, consisting in the fact that one nervous system is more highly plastic than another, and that this would be the condition for the development of more precise and constantly functioning complexes of conduction, which would make possible greater quickness and accuracy. The suggestion that the explanation of differences in intelli- gence is psycho-physical, in itself explains nothing. Of course mental differences are determined by neural factors. Beyond this guarded general suggestion Spearman and Krueger scarcely seem to venture in the way of definite positive statement, except to state that the size of the different positive correlations is to be explained on the basis of the degree of connection with a hypothetical common central factor. Spearman ('04) reached conclusions that he considered very 88 Correlations of Mental Abilities important for the measurement and diagnosis of general intelli- gence. " On the whole then we reach the profoundly impor- tant conclusion that there really exists a something which we may provisionally term General Sensory Discrimination, and similarly a General Intelligence, and further that the functional correspondence between these two is not appreciably less than absolute." Also, "... the common and essential element in the Intelligences wholly coincides with the common and es- sential element in the Sensory Functions." While these state- ments, taken in connection with the context, seem perfectly clear as to their meaning, it is evident that Spearman does not now hold to this latter statement as originally worded.^ That there is practically perfect correlation between Sensory Discrimination and General Intelligence certainly cannot be ijiaintained. Dr. E. L. Thorndike ('09) investigated the rela- tionship between accuracy in Sensory Discrimination and Gen- eral Intelligence. He took two groups of subjects: (i) 37 young women students in a normal school, (2) 25 boys in their third year in high school. For tests of Sensory Discrimination, Thorndike took 90 trials of each individual's accuracy in draw- ing lines to standards of 100, 75 and 50 mm., and 16 trials of each subject in weighting boxes to standards of 100 and 200 g. As a measure of General Intelligence, each girl was rated for general intelligence by all the rest, and also by 8 of the pro- fessors in the normal school. The scholastic records of the girls were also used. The boys were similarly rated by 6 of their fellows and by 4 professors. The tests were given on different days to eliminate common disturbing factors. The median deviation in age for girls was 10 months, and for the boys I year. The results are summed up in the form of Pearson coefficients of correlation (raw) as follows: " On page 165 of Cyril Burt's article cited below, footnote 3, he says : " With reference to my criticism of the passage cited above (p. 159) formulating his view of the relation of General Sensory Discrimination and General Intelligence, Dr. Spearman has written me: 'This conclu- sion of mine was badly worded. I did not mean (as others have naturally taken it) that general intelligence was based on sensory discrimination; if anything, vice versa. I take both the sensory discrimination and the manifestations leading a teacher to impute general intelligence to be based on some deeper fundamental cause, as sketched in the Zeitschrift fur Psychologie, Vol. XLI. p. no, par. 5.'" Comparison of Results With Those of Other Investigators 89 Discrimination of lengths, 1st half with 2nd half Discrimination of weights, 1st half with 2nd half Pupils' impressions of intellect, 1st half with 2nd half Teachers' impressions of intellect, 1st half with 2nd half. . Academic record, 1st half with 2nd half Combined teachers' and pupils' impressions, 1st half with 2nd half Thus the measures of the two phases of Sensory Discrimina- tion and the estimates of General Intelligence were satisfactory as to reliability. The inter-correlations found were: Women Boys Scores for all lines with those for all weights Scores for all lines with pupils' impressions of intellect . . Scores for all lines with teachers' impressions of intellect . Scores for all lines with teachers' and pupils' impressions of intellect 52 25 12 23 08 85 16 25 0^ Scores for all lines with academic record Scores for all weights with pupils' impressions of intellect Scores for all weights with teachers' impressions of intellect Scores for all weights with teachers' and pupils' impres- sions of intellect —01 20 Scores for all weights with academic record 21 Pupils' impressions of intelligence with teachers' impres- sions Pupils' and teachers' impressions with academic record. . Combined weights and Unes with pupils' and teachers' 54 Combined weights and lines with pupils' and teachers' 14 Hence the coefficient, corrected by the Spearman methods, be- tween 1st, factor common to accuracy in lines and weights, and factor common to pupils' and teachers' impressions of intelli- gence, is 20; 2nd, factor common to accuracy in lines and weights, and factor common to combined teachers' and pupils' impressions and academic record, is 25. According to this, the most probable correlation between General Sensory Discrimina- tion and General Intelligence would be about 23. Burt's results on this point (cited below) are difficult to in- terpret on account of the somewhat conflicting results in the different groups. If, however, we permitted ourselves to take the average of Burt's correlations between different forms of 90 Correlations of Mental Abilities t Sensory Discrimination and Estimated Intelligence, we find it to be about 21. There still remains Spearman's provisional theory of a hierarchy of mental functions as the explanation of the fact of general intelligence. " Wherever branches of intellectual ac- tivity are at all dissimilar, then their correlations with one another appear wholly due to their being all variously saturated with some common fundamental Function (or group of Func- tions)." " . . . the remaining or specific elements of the activity seem in every case to be wholly different from that in all the others." According to this theory, no two mental functions could be more closely related to one another than each is related to the common central factor. For instance in our -results. Memory of Words and Memory of Passages should relate no more closely to each other, and the A test and Geometrical Forms no more closely to each other, than the element common to Memory of Words and Memory of Passages relate to the element common to the A test and Geometrical Forms. This can be very readily tested by the use of the correction formula, Rpiq^ + Rpiq2 + Rp2qi + Rp2q2 Rpq = 4 , where Rpq represents ^Rp,p, X Rq,q, the correlation between the common factor in p^ and p^, and the common factor in q^ and gj/ ^Pi^i represents the correlation be- tween the ability tested by, say, Memory of Words and the ability tested by the A test; Rp^q^ represents the correlation be- tween the ability tested by the Memory of Words test and the ability tested by the Geometrical Forms test; Rp^Qi represents the correlation between the abilities tested by Memory of Pas- sages and the A test; and Rp^q^ represents the correlation be- tween the abilities tested by Memory of Passages and the Geometrical Forms test. Then taking the estimated true coefficients of correlation from Table XIV, and substituting the values in the above equa- 54 -F 49 -h 46 + 33 tion we get, Rpq = 4 =54- V80 X 87 which is much lower than Rpip2> 80, or Rq^q^, 87. Comparison of Results With Those of Other Investigators 91 That is, Memory of Words and Memory of Passages are more closely related one to another, and A test and Geometrical Forms' are more closely related one to another than the factor common to the first two is to the factor common to the second two. Similarly facts for the abilities tested by the Ebbinghaus and Hard Opposites tests, and the A and Geometrical Forms tests, letting pip^Qx and g^ refer to these in order, are that 54 + 36+58 + 42 Rpq = 4 =55, which is much below 85 or 87. V8s X 87 Again on the basis of Spearman's theory of a common central factor, if the coefficients of correlation among a number of abili- ties are arranged in descending order, from left to right and from top to bottom as are the ones he obtained on page 86 above, in every line the figures should be in descending order as they are on the top line and on the vertical line on the left. According to our results, secured from 37 subjects instead of only II, this is by no means the case, as is evident from the fol- lowing from Table XIV, p. 63. I 1 1 1 1 1 ■3 CfH Ph 'I ^ s* P< .2 -a <^ b ^ t? M ^ 3 -B S ^ a •3 g B s u is S s Z ^ «i s Ebbinghaus test R5 86 82 84 72 83 71 70 65 70 60 56 54 58 36 Hard Opposites 4?, 8? S4 65 80 39 65 54 49 Easy Opposites 7'>, 83 65 50 56 : 37 50 45 71 70 80 50 4?, 30 46 33 Adding 65 70 39 56 42 39 68 30 60 56 65 30 39 39 36 36 A test 54 36 58 42 54 49 50 45 46 33 58 30 36 36 87 87 Thus it is quite evident that Spearman's theory is not in har- mony with the facts we have secured. Cyril Burt ('09) tested 43 boys between the age,s of 12 years 92 Correlatiop.s of Mental Abilities 6 months and 13 years 6 months, one group of 13 boys from a high class College Preparatory School, and the remaining group of 30 from an Elementary School. The tests used can be grouped as follows: I. Sensory Discrimination. Touch, weights, pitch, lengths. II. Motor Tests. (Simple reactions). Tapping, card dealing. III. Sensory Motor Tests. Card sorting, alphabet finding. IV. Association Tests. Immediate memory — concrete words, abstract words, nonsense syllables Mirror test — formation of motor association by trial and error. Spot pattern test — apperception of a form composed of dots. V. Test of Voluntary Attention. Dotting irregular dots. The correlations found by Burt are as follows: Corrected Coefficients for Elementary School Imputed Intelligence Spot Pattern Dotting Mirror Tracing Alphabet Tapping Memory Dealing Sorting Sound Lines Touch Weight 100 75 74 68 65 60 54 53 52 51 17 -01 100 80 75 96 64 41 66 68 55 50 21 27 52 55 68 40 68 53 19 41 25 -05 -01 06 51 50 60 16 37 16 05 59 32 -05 26 -16 17 21 30 27 66 12 15 32 20 -01 26 37 01 27 07 -10 26 11 07 05 23 06 -16 37 Comparison of Results With Those of Other Investigators 93 Corrected Coefficients for Preparatory School ^ ^ a a ^ U ^ jn s 1 n -a 0. Eh S 1 I 13 1 ■3 § ^ 1 100 100 100 100 90 71 47 32 07 -32 -33 100 84 73 96 84 100 62 -17 45 31 25 100 84 94 91 100 56 83 37 32 -48 -06 100 73 94 41 80 100 50 36 57 09 74 100 96 91 41 82 68 66 06 44 -17 -20 90 84 100 80 82 64 84 06 27 27 22 71 100 56 100 68 64 75 -61 94 40 71 47 62 83 50 66 84 75 40 85 -29 10 32 -17 37 36 06 06 -61 40 -07 -51 -40 07 45 32 57 44 27 94 85 -07 21 15 -32 31 -48 09 -17 27 40 -29 -51 21 61 -33 25 -06 74 -20 22 71 10 -40 15 61 -62 63 35 44 41 27 93 40 -06 23 26 53 Sorting Dotting Alphabet Tapping Imputed Intelligence . Memory Mirror Spot Pattern Dealing Lines Touch Weight Soimd -62 63 35 44 41 27 93 40 -06 23 26 53 The above are arranged approximately in descending order, but it is very evident that they even more strongly contradict Spearman's original theory of a hierarchial arrangement than do our results spoken of above. Apparently, however, Spear- man himself has abandoned or modified this view. In his demonstration of a " Hierarchy of Coefficients," Burt uses the raw coefficients from the amalgamated measurements, rather than the corrected coefficients which we have quoted. In this connection he says : " Dr. Spearman and Prof. Krueger imply that satisfactory hierarchies are exhibited only by the ' pure ' or theoretical coefficients ; but it appears that those based on amal- gamated measurements are better than those based on theoreti- cal ' correction,' if the experimental conditions are carefully con- trolled." ..." The theoretical values for the ideal hier- archy may be obtained by various mathematical formulae." " The following simple formula has been supplied for this pur- pose by Dr. Spearman (to whom I am here particularly in- debted for several improvements on my own demonstration of a hierarchy). . . ." But when the raw coefficients fail to ac- cord with the theory, and when the corrected coefficients fail to accord with the theory, and when coefficients from amalgamated 94 Correlations of Mental Abilities series of measurements fail to accord with the theory, why not abandon the theory! Surely this would be better than to re- sort to further manipulation of data influenced to some extent by variables at present uncertain, and consequently to that extent untrustworthy. Only when the variables themselves are fully and definitely understood can such methods be used with any degree of safety. Burt does not give the correlations between memory of con- crete words, memory of abstract words, and memory of non- sense syllables. In comparing his results with ours we shall take his data for the boys of the Elementary School, both be- cause the greater number of subjects in this group makes the results more reliable than in the case of the other group, and because the group itself is probably more representative of average boys than are boys in a College Preparatory School. Perhaps the fairest comparison to make would be between his coefficients for the Elementary School Group, and our estimated true coefficients, though one would expect the latter to be some- what higher. Burt's corrected coefficient for Memory and the Spot Pat- tern test is 41 ; between Memory of Words and Learning Pairs, we get for the Poor group 44, and for the estimated true co- efficient 65. The agreement here is as close as could be ex- pected. As to correlation between Sensory Discrimination and Memory, Burt gets a correlation between Discrimination of Lines and Memory of 05; between Drawing Lengths and Memory of Words we get -09 for the Poor group, and -05 as the estimated true correlation. Bonser ('10) gave tests in reasoning and selective thinking to 385 boys and 372 girls of the Fourth, Fifth and Sixth school grades of public schools in Passaic, New Jersey. " The tests employed were made up of a series of problems and questions designed to exercise the most fundamental four phases of reasoning activity, namely: The mathematical judgment; con- trolled association; selective judgment; and that complex of analytic and synthetic thinking used in the intellectual interpre- tation of literature. Comparison of Results With Those of Other Investigators 95 Mathematical Judgment " The problems for testing the mathematical judgment were of two kinds, two sets of five each, I, A and B, stated in the form usually followed in current text-books in arithmetic; and two sets of five each, II, A and B, of the same difficulty as the pre- ceding in processes involved but stated in a less conventional way. Each of the ten problems of the first type may be called a " two-step " problem — it requires a preliminary operation for securing the intermediate datum necessary before the final operation can be accomplished. Tests I and II " LA. Get the answers to these problems as quickly as you can. I 2 If J4 of a gallon of oil costs 9 cents, what will 7 gallons cost? John sold 4 sheep for $5 each. He kept }i of the money and with the other ^ he bought lambs at $2 each. How many did he buy? A pint of water weighs a pound. What does a gallon weigh? At 1254 cents each, how much more will six tablets cost than 10 pens at 5 cents each? At 15 cents a yard, how much will 7 feet of cloth cost? B. I. A man whose salary was $20 a week spends $14 a week. In how many weeks can he save $300? How many pencils can you buy for 50 cents at the rate of 2 for S cents ? A man bought land for $100. He sold it for $120, gaining $5 an acre. How many acres were there? A man spent % of his money and had $8 left. How much had he at first? The uniforms for a baseball nine cost $2.50 each. The shoes cost $2 a pair. What was the total cost of uniforms and shoes for the nine? II. A. I. 132 plus what number equals 36? 2 If John had 15 cents more than he spent today he would have 40 cents. How much did he spend today? 3. What number minus 7 equals 23? 4. If James had 4 times as much money as George, he would have $16. How much money has George? 5. What number added to 16 gives a number 4 less than 27? 96 Correlatio-ns of Mental Abilities B. 1. What number subtracted 12 times from 30 will leave a remainder of 6? 2. If a train travels half a mile a minute, what is the rate per hour ? 3. What number minus 16 equals 20? 4. What number doubled equals 2 times 3 ? 5. If 7 multiplied by some number equals 63, what is the number? " In the original blanks, immediately following each problem space was left for its solution. Controlled Association " For controlled association, three types of tests were used. First, two sets of ten sentences each. III, A, a and b, were given with a significant word omitted from each to be filled in by the pupil. Second, two sets of ten sentences each. III, B, a and b, were given in each of which two significant words were placed, one above the other, one giving a correct meaning to the sen- tence, the other an erroneous meaning, the pupil to draw a line through the wrong word leaving the sentence so that it would read correctly. Third, three sets of twenty words each, IV, A, B and C, were given to pupils, they to write beside each respec- tive word a word just its opposite in meaning — the familiar " opposites " test. Tests III and IV " III. A. a. Complete the following sentences as quickly as you can by filling the blank spaces with appropriate words : I always comes in the last week in December. 2. A is one who plays a musical instrument. 3. The city is in Russia. 4 are large, visible bodies of watery vapor floating about in the air. S used for building houses are made of clay. 6. The machine on a railroad for drawing cars is an 7 is the most useful metal for blacksmiths. 8 live and swim about in the water. g. Most light summer clothing is made of goods. 10 is a holiday. III. A. b. 1. The flesh of cattle used for food is called 2. The months are June, July and August. 3. The makes it light during the day. Comparison of Results With Those of Other Investigators 97 4 catch many mice and birds. 5- A is a large stream of water flowing through the land. 6. Men who live in the country and till the soil are called 7 is a mineral which we burn. 8. The Ocean is east of the United States. 9 sell sugar, vegetables and other foods. 10. There are hours in half a day. III. B. a. As quickly as you can, make these sentences cor- rect by drawing a line through the wrong word where two words occur, one above the other : longer 1. Days are in summer than in winter. shorter up 2. Water always flows hill. down more 3. Glass breaks easily than tin. less earlier 4. The sun rises in January than in July. later softer 5. Iron is than wood. harder warmer 6. It is in Florida than in Maine. colder heavier 7. Anjrthing that floats is than water. lighter more 8. Oranges grow satisfactorily in California than in New Jersey. less shorter 9. Shadows are in summer than in winter. longer more 10. Plants grow readily in warm sunshine than in the cool shade, less III. B. b. stronger 1. Men are usually than women. weaker less 2. A pound of iron is worth than a pound of copper. more before 3. Christmas comes Thanksgiving day. after Correlations Vf Mental Abilities IV. Cotton ( :lothing is than wool, cooler Less coal is used in summer than in winter. More poorer Bankers are richer than cab drivers. More 1 dorses than mules are used for driving ; purposes. Fewer more There are teachers than preachers, feiver more Oranges ; are less sweet than lemons. More bread than cake is eaten in this city. Less . As quickly as you can write beside each of thes rd that means exactly its opposite. A. B. C. day- great bad asleep hot inside absent dirty slow brother heavy short best late little above first soft big left black backwards morning dark buy much sad come near true cheap ; north dislike broad open poor dead round well land sharp sorry country east thick tall t known full son something peace here stay few less push below mine nowhere enemy Selective Judgment " Two types of tests were used for selective judgment. First, two set^, y, A and B, of two series each of ten reasons why some given fact is true, some of which reasons are correct, the others incorrect or irrelevant, were given. The pupil was to Comparison of Results With Those of Other Investigators 99 select, by checking, the correct reasons. Second, there were given similarly two sets, IV, A and B, of three series each, of five definitions for a given thing or term, some of which were correct, the others incorrect or irrelevant: Tests V and VI " V. A. The following reasons have been given to show why New York has become a larger city than Boston. As quickly as you can, place a cross like this, +, before each reason you think a good one: New York is on an island. More foreigners live in New York than in Boston. New York is on a large river coming from a rich agricultural region. Mr. Rockefeller has a fine home in New York. New York has more churches than Boston. New York has better communication with the States lying to the west. 7. New York has elevated railroads. 8. New York is in the midst of a rich fruit and agricultural district. 9. New York is nine or ten years older than Boston. 10. New York has a republican governor. B. These reasons have been given to show that oak is better than pine for making furniture. Check the good reasons. I. Oak wood is harder than pine. 3. Oak trees have acorns, pine trees do not. 3. Oak wood takes a finer polish than pine. 4. Oak trees have more beautiful leaves. 5. Oak trees make good homes for squirrels. 6. Pine wood will not last so long as oak. 7. Pine is more easily dented and defaced than oak. 8. When polished and varnished, oak is much more beautiful than pine. 9. Pine trees are sometimes used for Christmas trees. ID. Oak trees are easier to climb than pine trees. C. The following reasons have been given to show why oranges grow better in Florida than in New Jersey. Check the good reasons. I. There are many negroes in Florida who work very cheaply. Florida has warm summer weather almost the whole year. There are no alligators in New Jersey. Florida very rarely has hard frosts. New Jersey is not so large as Florida. Florida was settled earlier than New Jersey. New Jersey grows many fine peaches. Florida has a very moist, warm climate. loo Correlatiqns of Mental Abilities 9. Florida is a word meaning the land of flowers. 10. Florida is a popular winter resort. D. Among these reasons why horses are better than cattle for driving and working animals, check those which you think are good reasons. 1. Horses are more intelligent than cattle. 2. Cattle are not so tall as horses. 3. Horses like corn, oats and hay. 4. Horses are much more active and walk faster than cattle. 5. Cattle are extensively used for food. 6. Horses are much more beautiful and graceful than cattle. 7. The skins of horses are sometimes made into gloves. 8. Horses are more easily trained and controlled than cattle. 9. President Roosevelt likes to ride on horseback. 10. Horses have more rapid and varied gaits than cattle. VI. A. In the following definitions, place a small cross, like this +, before those which you think are good ones, doing it as quickly as you can. a. Definitions of a shoe. 1. A portion of clothing. 2. Something black made of leather. 3. A protective covering for the feet, usually made of leather, having a firm bottom or sole and flexible upper portions, an opening for the foot being fastened by lacings, buttons or buckles. 4. Something to wear on the feet. 5. A necessary article costing from one to five or six dollars. b. Definitions of an island. 1. A piece of land out in the water. 2. A small body of land. 3. A body of land entirely surrounded by water. 4. Cuba is an island. 5. A portion of land rising above the surrounding level. c. Definitions of to explode. 1. To burst suddenly with a loud noise. 2. To knock all to pieces. 3. To make a very loud noise. 4. To fill the air with a tumultuous roar. 5. To blow up. a. Definitions of a chair. 1. A piece of household furniture. 2. A movable seat with a back intended for one person. 3. A piece of furniture on which to sit. 4. Rocking chairs are comfortable chairs. 5. A single seat having a back. Comparison of Results With Those of Other Investigators loi b. Definitions of to write. 1. To make marks with a pen or pencil. 2. To make characters which stand for ideas. 3. To use a pen or pencil. 4. To make marks on any kind of surface with any kind of an instrument which will express one's ideas so that another may understand them. 5. To write a letter. c. Definitions of a buggy. 1. A buggy is black. 2. A buggy is something to ride in. 3. A buggy is a light, four wheeled vehicle, with or without a top or covering, designed for carrying two or three persons. 4. A buggy is drawn by horses. 5. A buggy may have rubber tires. Literary Interpretation " For literary interpretation, two stanzas of poetry, VII, A and B, were used, the pupil to write the meaning of each in his own words. These poems are taken from a third reader and a second reader respectively, each from a different standard series published within a decade of the time of these tests. Test VII " VII. A. Read carefully the following stanza, then write its meaning in your own words. ' This little rill, that from the springs Of yonder grove its current brings. Plays on the slope awhile, and then Goes prattling into groves again. Oft to its warbling waters drew My little feet, when life was new.' B. Read carefully the following stanza, then write its mean- ing in your own words : ' Under the greenwood tree, Who loves to lie with me. And tune his merry note Unto the sweet bird's throat. Come hither, come hither, come hither, Here shall he see No enemy But winter and rough weather.' 102 Correlations of Mental Abilities Spelling " As an incidental problem for correlation, the opportunity offered for a test in spelling was taken. Two papers, B and C, from test V, the opposites test, were graded in spelling for each pupil. As the pupils did not know that the papers were to be graded in spelling, it had little of the disadvantages of the formal spelling test, yet the words were practically predeter- mined and uniform." Bonser's results are given in part in the following table: TABLE XX Averages op Coefficients of Cokrblation by Grade, by Age, AND for Both Grade and Age Boys , 385 Girls ,372 By By By By Grade Age Grade Age 39 58 32 34 33 49 31 53 21 44 17 52 10 38 31 25 18 48 08 32 38 69 58 70 16 34 20 25 55 57 54 47 38 45 35 35 25 31 21 16 40 48 43 39 62 61 58 40 27 35 21 07 41 65 34 56 35 45 30 25 26 37 28 35 87 87 81 85 04 29 25 28 45 54 36 53 22 40 09 34 70 79 62 80 07 18 06 29 23 28 11 32 60 63 59 50 03 18 04 13 32 49 27 39 27 25 17 30 02 30 23 32 Total, 757 By Grade and Age I-II and III " " IV " " V " " VI " " VII " " Total ... " " Spelling. III and IV " " V " "VI " " VII " " Total ... " " Spelling. IV and V " " VI " " VII " " Total ... " " Spelling. V and VI " " VII " " Total... " " Spelling. VI and VII " " Total . . . " " Spelling. VII and Total ... " " Spelling. Total and Spelling 41 42 33 26 26 59 24 53 38 24 45 55 22 49 34 32 85 21 47 26 73 12 24 58 09 37 25 22 Comparison of Results With Those of Other Investigators 103 Bonser concludes that the tests " are valid measures of sev- eral phases of that complex capacity we call reasoning ability. Group correlations are so much higher among these tests than among those so far produced among mental abilities more varied in kind that we are clearly justified in holding them to be tests of abilities which are varieties of one general species of ability." Bonser's coefficients of correlation are all obtained by the method of like and unlike signs, and are uncorrected for at- tenuation. " The highest coefficients as shown by the averages, are, in their order, that for tests III and IV, the two forms of controlled association, — completing sentences and opposites — 53 ; that for tests IV and V, the opposites and the selection of reasons in one of the tests in selective judgment, 49; that for V and VI, the two tests in selective thinking, 47; and that for III and VII, controlled association and interpretation of poems, 45. By correction these would all be raised to above 75, those for III and each of the others approaching 100 very closely." This last statement is made on the ground of a few coefficients of correlation that have been corrected by two different methods, and it must be remembered of course that this is only a rough estimate. However, the general fact that the correlations of the different tests of selective thinking with one another are rela- tively very high, is in accord with our results as shown in Tables IX, X, and XIV. Bonser further concludes that " the results here derived point to the conclusion that the correlations among the abilities here tested are a matter of native capacity rather than the result of training." This follows mainly from the facts that, (i) the cor- relations on the basis of age are considerably higher than those on the basis of grades ; (2) from the fact that the median age of the best 10 per cent and of the poorest 10 per cent is nearly the same, while in the scores obtained in the tests, they differ from three hundred to three thousand per cent. Moreover the chil- dren had had almost no training in the exact type of problems as set, in the opposites test, and in the form of selective judg- ment of test V, yet these stand, in this order, in the highest cor- relation to the total ability shown for all of the tests — all of which suggests that these abilities have not been developed as 104 Correlations of Mental Abilities by-products of school training, but that the results of the tests are measures of the native capacity of the children for the ac- tivities required in these problems. There is however a disturbing factor entering into Bonser's results which is worthy of consideration. It arises from the fact that in all tests except that of Opposites, the time element was not kept constant for all. All of the pupils were stopped at the moment when the first pupil of the room had just finished. Bonser does not state the time record in the different rooms, nor take it into consideration in scoring the results. Presum- ably the pupils in the higher grades would thus get less time for the work than those in the lower grades, in so far as the time taken by the quickest pupil in an upper grade would be less than the time taken by the quickest pupil in a lower grade. It would thus appear that Bonser's results underemphasize the superiority of the upper grades over the lower ones. It would tend to favor the younger pupils. While it would not affect the size of the correlations by grades, it would probably tend to raise somewhat, — and certainly to disturb more or less, — the correlations by age. How much this disturbance would amount to is difficult to say. It would probably be slight. Even grant- ing that it would be considerable, however, Bonser's contention that the tests measure native capacity rather than training, would still have sufficient to support it in the other arguments used, apart from that of the higher correlations by age than by grades. This conclusion also is in accord with our results. William Brown ('ii) records the results of a somewhat ex- tensive investigation " for the purpose of determining to what extent correlation exists between certain very simple mental abilities in cases where the individuals experimented upon are, as near as may be, identically situated with respect to previous practice, general training, and environment; and how closely, if at all, these elementary abilities are related to general intellec- tual ability as measured by teachers' judgments, school marks, etc. Every effort was made to keep the groups of individuals tested as homogeneous as possible; and instead of measuring irrelevant factors and ' correcting ' for them in the later stages of the research, the influence of such irrelevant factors was ex- Comparison of Results With Those of Other Investigators 105 eluded right from the beginning by a rigorous segregation of the material, and in other ways. " The groups of individuals to which the tests were applied, were as follows : Group I, 66 boys of a London elementary school, all between the ages of 11 and 12. Group II, 39 girls of a London elementary school, all between the ages of 11 and 12. Group III, 40 boys of a London higher grade school, all be- tween the ages of 11 and 12. Group IV, 56 training college students (women), of the same year and of approximately the same age. Group Va, 35 university students (men). Group Vb, 23 university students (women)." The tests employed were selected " not so much for their a priori likelihood of showing inter-correlation, as for their con- venience in admitting of application to an entire group of sub- jects simultaneously and unobtrusively. The following is a list of them: 1. Crossing through letters e and r in a page of print. 2. Crossing through letters a, n, o, and s in a page of print. 3. Crossing through every letter in a page of print. 4. Adding up single digits in groups of ten. Measurement of (a) speed, (b) accuracy. 5. Bisecting ten printed lines (80 mm. long), and putting in one of the points of trisection in each of the ten other lines (90 mm. long). 6. Muller-Lyer Illusion. Measurement of (a) size, (b) mean variation. 7. Vertical-Horizontal Illusion. Measurement of (a) size, (b) mean variation. 8. Mechanical Memory (permanent), tested by means of nonsense syllables. 9. Memory for poetry. 10. Combination test (Ebbinghaus mutilated test). In the case of groups I and II, recourse was also had to : 11. Marks for Drawing. 12. Total school marks. io6 Correlations of Mental Abilities 13. Grading for General Intelligence (two independent measures). Finally with Groups Va, and Vb, the following test was also employed : 14. Association-time (uncontrolled). Measurement of rate of sequence of ideas called up by a stimulus-word." The author does not seem to appreciate fully the inaccuracies which may creep in and influence the results, when the tests are given as group tests rather than to each individual separately. With the exception of test (9), and in some cases test (8), every test was applied twice, the second test being given about a fortnight after the first, and at the same hour of the day. It is highly probable that the gain in accuracy owing to the comparatively large number of subjects in the different groups is more than counterbalanced by the tendency to spurious cor- relation due to the tests being given as group tests, and by the inaccuracies due to the small number of measurements taken. However there is still need of considerable diversity of general method in investigations in correlation, to insure adequate cor- roboration and verification of results. Brown's numerical results are summarized in part in the fol- lowing tables : Comparison of Results With Those of Other Investigators 107 Brown's Results. Pearson Coefficients of Correlation (Top line, Group I; 2nd line, Group II; 3rd line. Group III) a 1 IS 00 1 1 IS e 1 CD a < 3 8 a _o }m •3 'a 1 1 a i CO m 1 "o -g to 1 s Marking er 78 80 74 45 -15 00 40 00 00 27 23 69 13 35 30 00 00 53 49 25 00 00 30 00 28 Marking anos 78 80 74 48 00 10 29 20 00 28 14 51 00 20 24 00 -11 21 21 00 00 27 17 13 10 Ebbinghaus test 45 -15 00 48 00 10 52 37 28 52 44 40 -13 32 38 -25 00 13 00 28 15 54 60 43 69 Mechanical Memory. 40 00 00 29 20 00 52 37 28 49 38 27 -13 00 31 -23 00 14 00 00 10 59 40 55 49 Memory of Poetry.. 27 23 28 14 52 44 49 38 41 00 38 -11 12 19 13 60 57 Addition (Speed).. . . 59 13 35 51 00 20 40 -13 32 27 -13 00 41 00 13 24 33 25 33 20 00 00 28 10 24 Addition (Accuracy) 30 00 00 24 00 -11 38 -25 00 31 -23 00 38 -11 13 24 33 00 30 00 41 00 11 00 00 Motor (all letters)... 53 49 25 21 21 GO 13 00 28 14 00 00 12 19 25 33 20 00 30 00 00 00 23 13 32 Bisection 00 00 15 10 13 00 41 00 School Marks 00 30 27 17 54 60 59 40 60 00 28 00 11 00 23 64 78 General Intelligence. 00 28 13 10 43 69 55 49 57 10 24 00 00 13 32 64 78 io8 Correlation's of Mental Abilities Brown's Results. Pearson Coefficients op Correlation (Top line, Group IV; 2nd line, Group Va) ^^ X C3 ^ o 1 1 a. 1 a i < i 1 ■3 1 .S g 1 Jl T3 -« "o s5 J3 13 T3 05 ^ H < QCSM34ZoOj> 0DPD0A0D'0,0/GT>[JO0D0,g'c OnDv4'nZ7flnAOQpop>a5^DODI OX^DOMDoDOQvDQgAaQsD Da^QDAODObODA^r OODODOOADDOaZTDax TOVQOoDOD^Azynqc °DDOI>OnnOViDDVZ7QD<>dnDDDptiO Dn^oDUDD^Di>a'!;:ipnpDdgD^p AOnnODDDI>O^D/:7D27opgQpDVQ, o'^onAi>nDDo^D£7vDn>a'^oq47._ DDOD'Q>D00D0O°Z7DqngO^eDOp0^ VADOXiicnoanoonDocS^QODa^a'CiA AQ VOO OASOO □ O^ciH C, □ a)#O0OOV4>E ®AOt&OA'g>[ElO© C'ASK^A'3 ^^A D. E. CEBlLUCIZtUnZP F 114 Correlations of Mental Abilities Test IV. "Easy Opposites" Test. IV, a. As quickly as possible give orally a word that means the exact opposite of each word in the list. IV, b. Ditto with second list. IV, c. " " third " . IV. d. " " fourth " . IV, a IV, t IV. c IV, d good stale high day outside hot up asleep quick dirty wet absent tall heavy new brothel big late soft best loud first wider over white left wrong big light morning yes backwards happy much young buv false near brave come like north winter cheap rich open weak broad sick in forget dead glad sharp wila land thin east beginning country empty sour straight tall war something raise son many stay rough here above push love less friend nowhere noisy easy Test V. "Recognizing Forms" Test. (See Fig. I, C, D, E and F.) V, a, I (C of Fig. i). You may study this for i minute; then I shall tell you to stop. (The general nature of the test and what was to be done by the subject were explained before starting.) V, a, 2 (D of Fig. i). Mark all the forms exactly the same as those seen in V, a, l. (No time limit was required in the marking.) V, b, I (E of Fig. i). You may study this for one and a half minutes; then I shall tell you to stop. V, b, 2 (F of Fig. i). Mark all the forms exactly the same as those seen in V, b, i. (No time limit was required in the marking.) Appendix IIS Test VI. "Memory of Words" Test. VI, a. Write down all the words in the list that you can remember after hearing them read once., VI, b. Ditto with second list, VI, b, etc. VI, a VI. b VI, c VI, d picture knife mouse whisper silly window bank Columbus unless peacock disease necessary lizard brass cheap laugh book weary country dictionary pain rich study cane island vine tooth key tin servant musician doctor literature pinch pie boat axe wheel building enough run hammock fruit walking tomato horn weapon rent tired pitiless spider earth frost crack mountain canvas wide beef shallow carpet Indian glue window steam Test VII. "Learning Pairs." (See Fig. 2.) Study VII, a, i, for one minute so that when VII, a, 2, is given, you can write down the corresponding word. Similarly with VII, b, i and b, 2. Similarly with VII, c and d (except that one and a half minutes were allowed for study instead of one minute). For Fig. 2, the lists of pairs are given in order. Test VIII. "Memory of Passages." Write down all that you can remember of the substance of the passage after hearing it read once. VIII, a, Memory of Passages It isn't necessary to read a book in order to be happy with it. On a steamer or in a hammock you simply have to have the book in your lap or close at hand, with the paper-cutter and pencil. It must be the sort of book you like. You open it and read the table of contents. A deep peace fills your soul. Here is this delicious book and the whole day, both yours. You lean back to think of books by these men and by others that you already know and love. Memory brings you one beautiful picture after another. VIII, b. Memory of Passages Thirty-two passengers were injured, none of them seriously, by the derailment of the Chattanooga and Washington Limited train on the 1 16 Correlations of Mental Abilities Fig. 2 ^ -f ^ r>(f^ grit ^f^ 4\< r«t ^ y-jj dtnt JS/ 59r>w.» 1^ -CV ''""s r>^ JJ^ pince I> ^ f^. i^ ?> tooa fl- >Q«'3 )A fv 'f'l' /^ gi^ nj4 ^fi- /-^ t«d n^ ^(i cUrm -A rra =t=t. ^ ^ Jokn QL^ ^;;fl- LiTW! A< >tf| ^.t y^J ^ j.rst, Wl ># l"-- >t^ VII. a, 1. VII. a, 2. VII. b, 1. VII. b, 2. 1^ k«itl, w O^uw ^ :^^ jat. /km [D rmk ^v^ :;>? V'i OOi 7^-» ^ \S F«J ;^ X_j^ nice a- ^ jft ;5V r^ mmK ~bl -U>< ^"* -t^< ^ crw. v^ !zA^ U™ W' [>vi jln. ^ >f^ d.aa ^ VII. c, 1. VII. c, 2. VII. d, 1. VII. d, 2. Southern Railway, thirty miles south of Charlottesville, and just north of Ryan's Siding, Virginia, early to-day. A broken rail was the cause of the accident. The entire train composed of a baggage car, day coach and three sleep- ers, left the track, the sleepers being almost destroyed by fire. A special train was quickly made up and proceeded to this city with all the passen- gers of the Limited. The wreck blocked the track for several hours, all trains meanwhile being detained. VIII, c. Memory of Passages Langford of the Three Bars, as the title suggests, is a story of the West depicting cowboy life. The scenes are in South Dakota of the time of the "rustlers," who cared for neither the interference of man nor law. The action turns round the Three Bars Ranch, which is run by Paul Langford, " a man — a godlike type with his sunny hair and his great strength," whose Appendix 117 object it is to do away with the cattle thieves headed by Jesse Black. He is aided by Gorden, the county attorney, and Jim Munson, a real cowboy. VIII, d. Memory of Passages One morning a couple of Springs ago, if any of your readers had chanced this way, they might have seen me coming from the vineyard with two bluebirds, one in each hand. The birds were well and vigorous and en- tirely unharmed. If questioned I might have explained that I went down into the vineyard and picked the birds up off the ground, where they had the full possession of their wings, and that there are times when it is not difficult for me to do such things. These birds were of the species known as the Least-flycatcher, or Chebeck Bird. Test IX. "Drawing Lengths." I. To the right of i, draw in succession three lines each equal in length to I. You may add to, or take away from, the line you have drawn, as much as necessary in order to get it the required length. Similarly with lines 2 and 3. (Lines I, 2 and 3 were 100, 75 and 50 millimeters long respectively.) Test X. "Estimating Lengths." Which line is the longer? (Two lines drawn end to end horizontally on a long piece of white card-board were shown.) In case of the first 8 lines thus shown on the card-board, one line was ro8 mm. in length, and the other 100 mm. In case of the second set of 8 pairs of lines, one line was 106 mm. and the other 100. In the third set, one was 104 and the other 100 ; and in the fourth set one was 102 and the other 100 mm. The whole test was then repeated. Test XL "Adding." Add the ten sums in XI, a, as quickly and accurately as you can, writing down the results as you get them. Similarly with XI, b. XIa. Addition 17 26 27 72 23 42 51 24 13 47 38 47 83 39 86 91 82 19 81 64 54 63 45 26 36 17 42 38 91 36 26 51 47 82 26 27 24 83 19 45 72 14 39 62 63 23 47 86 54 54 Correlations^ of Mental Abilities Xlb. Addition 41 53 * 67 78 86 52 67 86 37 32 86 34 23 96 44 23 78 45 72 36 35 19 67 23 68 45 52 19 45 23 13 86 78 67 92 68 23 67 78 36 77 35 23 37 68 86 67 86 96 39 Test XII. "Hard Opposites." Write as quickly as you can beside each word in the column a word that means the exact opposite of it. Do the best you can with each word rather than leave the space blank. XII, b serious grand clumsy to win to respect frequently to lack apart stormy motion forcible to float straight to hold after unless rough to bless to take exciting Hard XII, a vertical ignorant rude simple deceitful stingy permanent over to degrade weary to spend to reveal genuine level broken wild part past permit precise Opposites XII, d succeed strict tardy sleepy suspicious rigid suave sinful conservative refined pride despondent imaginary beautiful injurious diligent sell sure active venturesome XII, c tender animated proficient impoverish cruel generous haughty silly insignificant disastrous miser result hindrance strength innocent busy remember increase preserve belief Test XIII. "Completing Words," or " Ba-test." As quickly as possible, add any letter or letters to each syllable in the list, so as to make it a complete word. Appendix' ug Completing Words a. b. c. d. ba be bi bo ab ea ic ao ca ce ci CO ac eb id ob da de di do ad ec ig oc fa fe fi fo af ed il od ga ge gi go ag ef im of ha he hi ho al ei in ol ja je ji jo am el ir om la ke ki lo an em is on ma le li mo ap en it op na me mi no ar ep iv or pa ne ni po as eq um OS ra pe pi ro at er un ot sa re ri so au es up ou ta se si to va ev ur ov va te ti vo aw ex. us ow Test XIV. "Ebhinghaus Mutilated Text." (The subject was first shown what was to be done on a sample sheet similar to the ones given below.) Fill in each blank with the word that will make the best sense. Do the work as well and as quickly as you can. Put only one word in each blank space. Test XIV, a. Ebhinghaus Mutilated Text. Park Hill on the Hudson offers you a solution of the home. problem to-day. No home seeker or investor can afford to ignore its claims. Escape the wear and tear of the city's noise and rush in this open air para- dise, just at the city's edge, in all respects an ideal home location for your- self and family. There are cottages containing every improvement waiting for you to step in and make yourself comfortable. It not only commands the most beautiful view around New York but is protected for all time against intrusion. Choice lots now selling on very easy terms. Test XIV, b. Ebbinghaus Mutilated Text. We believe we can prove to you that this investment is so secure and the dividends so sure, that it justifies you in withdrawing money from the Savings Banks, where it is earning 3>4% and putting it in our busi- ness where it will earn y%. We are a New England enterprise, managed hy New England men, and we have behind us a record of fourteen years of unbroken success. Whether you have much or little you cannot afford ,to let slip this opportunity of doubling the income from your savings. Prompt action in this matter will repay you well. Test XIV, c. Ebbinghaus Mutilated Text. On the contrary, it didn't cost me a dollar. In fact, though at times I have found myself possessed of considerable sums of ready money, I I20 Correlations of Mental Abilities have never been a man of property in the strict sense of the word. I abandoned my profession, the law, as I did not find its practice so lucra- tive as I had hoped. For some years thereafter I travelled largely on the Mississippi River. It was the decline in steamboating and the adoption of less leisurely methods of travel that cut into my income and forced me to come North and engage in trade. Test XIV, d. Ebbinghaus Mutilated Text. The occult in everyday affairs is the theme of this new book by Robert Chalmers. Opening one of the thrilling stories of ■which the volume is composed is the tale of some awful mysterious happening, some super- natural event beyond the power of material reasoning of mortal man to explain, which comes into the life of some ordinary, everyday man. The opening chapter tells of a dinner given to a man deeply versed in occult- ism by his American friends. To these he giv^s many hints and sugges- tions of momentous things which he can plainly see waiting for them in the future. Test XIV, e. Ebbinghaus Mutilated Text. I asked the slovenly, but cheerful female who answered the bell for the landlady; wondering the while what I should say when I was asked for references. The merriment had not been called forth by anything amusing in my appearance, as my vanity had feared, but by a story which a man sitting at the head of the table was just finishing. The only vacant chair in the room was beside him, and, rather awkwardly, for I felt that they were taking my measure, I made my way toward it. As I sat down he greeted me with a polite bow. Test XIV, f. Ebbinghaus Mutilated Text. If we are perfectly well, thoroughly sound, we need' not be depressed. The perfectly healthy animal has no worries. The remedy has already been indicated. Regretfully it is so simple that very few people take the trouble to apply it. When it is clearly and widely recognized that worry is stupid, that its cure is simple where there is no organic trouble, worry will cease. Worry is simply a form of what for the sake of a nice large word, is called " neurasthenia," nerve-depletion. Given plenty of recreation, plenty of fresh air, and the normal man will not worry. Test XIV, g. Ebbinghaus Mutilated Text. We confess to something of sympathy with the correspondent who hinted yesterday that when children are run over and killed by automo- biles, the fault is not always that of the automobilist, but sometimes rests in some measure on those who do not teach their children to avoid un- necessary danger. It is a plain fact, of course, that public highways are for the use of the whole population, and that the automobilist is under every obligation to keep the limitations of his rights and privileges in mind as he goes along, but the road is his as well as other people's. Appendix 121 Test XIV, h. Ebbinghaus Mutilated Text. A law in defence of property rights in the broadest sense if observed would almost abolish international conflicts. Gentlemen do not fight with fists in money differences nor do they refer them to courts of honor. Civil courts are for that purpose and are as useful for nations as for men. The sanction of international law must be merely moral, for a long time at least. But in order that there should be any moral sanction there must be a moral code. The principles of such a code are deducible from treaties to which nations have set their hands and seals. Test XV. "Absurdities." As quickly as possible mark each sentence that contains an absurdity or impossibility. For instance, if a sentence stated or implied that ordi- nary lead was floating on water, mark such a sentence as impossible or absurd. Do not mark the sentences that contain no absurdity or imposssi- bility. Test XV, a. Absurdities. 1. Though armed only with his little dagger, he brought down his assailant with a single shot. 2. Silently the young dude hurried on, in spite of the darkness, and went splash into a puddle on the roadside. 3. Having reached the goal I looked back and saw my opponents still running in the distance. 4. While walking backwards he struck his forehead against a wall, and was knocked insensible. 5. Offended by his obstinate silence, she refused to listen to him further. 6. With his sword he pierced his adversary who fell dead. 7. The one-armed cripple was attacked by a dog, which seized his wrist, but he pushed it off with the other hand. 8. While forcing my way through the crowd, I came suddenly upon an old friend. Test XV, b. Absurdities. 1. The dogs pursued the stag through flower gardens in full bloom. 2. The storm which began yesterday morning, has continued without intermission for three days. 3. That day we came in sight of several icebergs that had been entirely melted by the warmth of the Gulf Stream. 4. While sharpening his three-bladed knife, my cousin cut his middle finger. 5. My friend pointed out the North Star clearly visible on our right as we walked briskly eastward in the moonlight. 6. The red haired girl, standing in the corner, is taller than any of her brothers. 7. The two towns were separated only by a narrow stream, which was frozen over all winter. 8. Fearing that he might waken her patient by his impudent talk, the nurse gave the detested dummy what he wished. 122 Correlations of Mental Abilities Test XV, c. Absurdities. 1. After dressing herself carefully and elaborately, she descended to the breakfast-room, only to find it deserted. 2. Preferring a tarnished reputation to the probability of becoming a corpse for the rest of his life, the young soldier took to flight. 3. Upon very careful testing it has been found that a pint of cream weighs slightly more than a pint of milk. 4. The old soldier energetically shouldered his crutch like a gun, as he talked by the fireside. 5. In the busiest sections of New York City, cheap houses, like rare jewels, are scarce and expensive. 6. In the ruins of an ancient Roman city there has recently been dis- covered a small skull believed to have been that of Pontius Pilate when he was about ten years old. 7. We serve hot and cold lunches on five minutes' notice, to first and second class passengers. 8. In my excitement I caught a glimpse of the sharp features of my enemy, who had just passed around the corner. Test XV, d. Absurdities. 1. Our horse grew so tired that finally we were compelled to walk up all the hills. 2. The hands of the clock were set back, so that the meeting might surely close before sunset. 3. In some states there is a law forbidding a man to marry his widow's sister. 4. Don't go to unreliable real estate offices to be swindled, come in here. 5. Owing to the lack of ready money, the shrewd financier was unable to take advantage of the rare bargains then offered. 6. Travellers who cannot read should be directed to sources of reliable information by signs printed in conspicuous places along the roads of travel. 7. With wrapt attention, though the audience was immense, the orator listened to the crowd addressing him. 8. Our office boy has been coming early of late, for he was often be- hind before, because his watch was slow.