ae Col ae bd F ia af u ipa a aller . 2 ache tr : on o~ THE UNIVERSITY OF ILLINOIS LIBRARY (‘Silas W 43m Return this book on or before the Latest Date stamped below. A charge is made on all overdue books. University of Illinois Library 27214 Digitized by the Internet Archive in 2022 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/measurementoffai0Owats > THE MEASUREMENT OF FAIR-MINDEDNESS By GOODWIN B. WATSON, Ps.D. TEACHERS COLLEGE, COLUMBIA UNIVERSITY CONTRIBUTIONS TO EDUCATION, NO. 176 Published by Teachers College, Columbia Cnibersitp New York City 1925 a 47 [ SORES Copyright, 1925, DUTCH eet i (a ee Re YAR SADHIAAIROWAN EL EVM SPAWNS —) 3) Oot 2S eM te | A, na hee WV 32 ™) | ACKNOWLEDGMENTS The author wishes to express his deep and genuine gratitude to Professor George A. Coe for the stimulus of his keen thought, to Professor William H. Kilpatrick for the richness of the experience and counsel which he offered, to Professor William A. McCall for most interesting and quite indispensable guidance through the technical problems of the study, and to Professor Harrison S. Elliott of Union Theological Seminary for encouragement of the entire enterprise. To Miss Mildred Barthelmess of the State Normal School at Trenton, New Jersey, to Professor Jerome Davis of the Yale Di- vinity School, and again, particularly to Professor Elliott, the author is indebted for opportunities to carry forward this study in groups which could not otherwise have been reached. Unfortunately, because of the confidential nature of their assist- ance, direct personal acknowledgment cannot here be made of the service of many persons who contributed significantly to the enterprise. However, the author does wish to recognize the valuable aid so graciously rendered by his brethren in the ministry of the Methodist Episcopal Church, and by his teachers and fellow students in Teachers College and in Union Theological Seminary. j G. B. W. New York City April, 1925 fa ans eee | R is v AH me ANTE ote Z ATEN ‘ae 4 CONTENTS CHAP: PAGE I. THe NEED FOR TESTS OF FAIR-MINDEDNESS .. . . I PERU CONSTRUCTION OF THRU TESTS) ib tal a) ke pecs 8 eR SCORING OF THE DESTSu se mmmmetm ican cela ieert:. dyey sta TS DV WWHAT DO. THE TESTS REALLY MEASURE? 20400 0.04, 2218 Vee SUGGESTED USES FOR: THE LESTS 92) fn 3 82s. a 36 APPENDIX BIN DIVIG AL’ PROFILES tained ae nO ee ae get Se sROUT PROFILES! nine ee wenn enuenOM me enEN nna: UIE hE CGS Se TISTIGAT, LABLES? | eeu tyme alii a eth tee oh te kes CHAPTER I THE NEED FOR TESTS OF FAIR-MINDEDNESS Swift changes have come in educational practice since the days of not so long ago, when the efficiency of the instruction which had been meted out to the youth of the countryside was judged by the formal visit of school committee-men who were more possessed of a sense of responsibility and dignity, than of techniques for measur- ing achievement. Now, one who would ask regarding the efficiency of a teacher or curriculum must be prepared for a volley of I. Q.’s, E.Q.’s and A.Q.’s, with a running accompaniment of norms, stand- ard deviations, correlations, and probable errors. If, perchance, some of the tests have been too hastily thrown together, and if some of the results have been much too hastily interpreted, that would be but natural. The measurement movement in education is not the first advance of science which some have “loved, not wisely, but too well.” Advance, it surely has been. Whatever mistakes shall have to be discovered and corrected, there will remain a large residue of gain. Future generations will be able to select curricula and meth- ods with far greater effectiveness, because of those indefatigable workers who have discovered how to answer some questions begin- ning, “‘How much,” and “How many.”’ Much of the dissatisfaction that has arisen in connection with this movement has expressed itself in a demand for tests of some- thing more important than the abilities thus far measured. Some who believe that ‘‘out of the heart are the issues of life,’’ have been searching for more accurate methods of evaluating motives, feelings, and purposes. Indeed, they may have had considerable reason for their fear that because certain qualities were easily measured, these qualities might be taken as criteria of desirable learning. It is possible that the contrary would be nearer the truth. Too often has schoolroom practice been controlled by the fact that supervisors could measure changes in ability to manipulate figures but could not measure the purposes which the child had in manipulating the figures. Hence, it has sometimes been said that the school was more concerned in teaching pupils how to write than in helping them to discriminate between forgery and an eman- cipation proclamation as an end for which the writing was to be 2 The Measurement of Fair-Mindedness used. Most skills and techniques of the kind which has been meas- ured are two-edged swords,—they can be used for the welfare of society or for its detriment. A number of attempts have been made to measure something other than intelligence or practical educational skills. Among a large number of attempts, only four or five have at the present date been standardized. One is Miss Downey’s Will-Temperament Test. With its many useful features, it is limited by the fact that temperament is measured in so small a realm, namely, that of expression through handwriting. A second test is the Pressey Cross-Out Test. This affords considerable aid in the discovery of persons who are emotionaly unstable. Third, we have the Wood- worth Questionnaire, which sheds some light on the causal factors in emotional conditioning. A fourth test is Koh’s Test of Ethical Discrimination, on which not many norms are now available. The Hart Test of Social Attitudes and Interests contains some very excellent material, but has not yet been fully standardized. The Kent-Rosanoff Association Test is also of considerable worth in discovering emotional complexes. Among the attitudes which are considered most desirable by a considerable group of religious educators is one which is called variously, Open-Mindedness, Freedom from Prejudice, Scientific- Mindedness, and Fair-Mindedness. Something of the sort seems to be suggested by Professor Coe’s criticism! of religious instruction which leads not to knowledge but to partisanship, and his emphasis of fair-mindedness and teach- ableness as objectives in the education of modern youths.? Even more clearly such a viewpoint is set forth by Hocking in Human Nature and tts Remaking*® when he sets up an ideal of a iThis kind of instruction in childhood produces not only in Catholicism but also in Protestantism an easily recognized adult type, the man who settles historical or scientific questions without his- torical or scientific study, and by the result judges whether his neighbors are sheep or goats.... There is no security for worthy ends short of the habit of considering others’ points of view. Coe, Geo. A., A Social Theory of Religious Education, pp. 65, 40. *Aneducatedman . . . must have command of the method of the mind, and he must be—. to some extent—a thinker, not a mere imitator. He may or may not have more opinions than other ' persons, but he has more opinions to which he hasaright . . . og To the extent that society is organized in the interest of the common Rood: and chooses its means by scientific types of analysis, the attitude of teachableness supplants that of dogmatic assertion. Coe, Geo. A., What Ails our Youth? pp. 39, 77. 3Why do we not display with complete equableness all views of the best way of life, and say ‘Now choose; think out your course for yourself’? Instead of teaching our children our morality, why not teach them ethical science? Instead of religion, metaphysical criticism? Instead of po- litical faith, political philosophy? Instead of manners, the principles of aesthetics? In short, why not make thinkers of them rather than partisans? Hocking, W. E., Human Nature and Its Remaking, pp. 230-23. The Need for Tests of Fatr-Mindedness 3 rational scientifically-minded generation, free to choose its own best course. Likewise Miss Follett in The New State, a well-known exposition of a new social and political philosophy, finds open-mindedness one of the central prerequisites. In every disagreement she be- lieves men should find a new challenge to investigation and re- thinking, not an occasion for intellectual strife. Frequently, in his writings, Professor John Dewey has set forth very clearly an educational objective of open-mindedness.® He 4The lesson of the new psychology is, then: Never settle down within the theory you have chosen, the cause you have embraced; know that another theory, another cause exists and seek that. The enhancement of life is not for the comfort-lover. As soon as you succeed,—real success means something arising to overthrow your security And throughout our participation in the group process we must be ever on our guard that we do not confuse differences and antagonisms, that diversity does not arouse hostility. Suppose a friend says something with which I do not agree. It may be that instantly I feel antagonistic, feel as if we were on opposite sides, and my emotions are at once tinged with some of the enmity which being on opposite sides usually brings. Our relations become slightly strained, we change the subject as soon as possible, etc. But suppose we were really civilized beings, then we should think, ‘How interesting this is; this idea has evidently a larger content than I realized; if my friend and I can unify this material we shall separate with a larger idea than either of us had before!”’ Follett, M. P., The New State, pp. 38, 40. 5Some attitudes may be named which are central in effective intellectual ways of dealing with subject matter. Among the most important are directness, open-mindedness, single-mindedness (or whole-heartedness) and responsibility Open-mindedness. Partiality is, as we have seen, an accompaniment of the existence of interest, since this means sharing, partaking, taking sides in some movement. All the more reason, therefore, for an attitude of mind which actively welcomes suggestions and relevant information from all sides . . . Openness of mind means accessibility of mind to any and every consideration that will throw light upon the situation that needs to be cleared up, and that will help determine the consequences of acting this way or that . . . The worst thing about stubbornness of . mind, about prejudices, is that they arrest development; they shut the mind off from new stimuli. Open-mindedness means the retention of the childlike attitude; closed-mindedness means premature intellectual old age : Open-mindedness is not the same as empty-mindedness. To hang out a sign saying, ‘Come right in; there is no one at home’ is not the equivalent of hospitality. But there is a kind of passivity, willing to let experiences accumulate and sink in and ripen, which is an essential of development, Results may be hurried; processes may not be forced. They take their own time to mature. Were all instructors to realize that the quality of mental process, not the production of correct answers, is the measure of educative growth something hardly less than a revolution in teaching would be worked Dewey, John, Democracy and Education, pp. 204-206. While it is not the business of education to prove every statement made, any more than to teach every possible item of information, it is its business to cultivate deep-seated and effective habits of discriminating tested beliefs from mere assertions, guesses and opinions; to develop a lively, sin- cere and open-minded preference for conclusions that are properly grounded, and to ingrain into the individual’s working habits methods of inquiry and reasoning appropriate to the various prob- lems that present themselves. No matter how much an individual knows as a matter of hearsay and information, if he has not habits and attitudes of this sort, he is not intellectually educated. He lacks the rudiments of mental discipline. And since these habits are not a gift of nature (no matter how strong the aptitude for acquiring them); since, moreover, the casual circumstances of the natural and social environment are not enough to compel their acquisition, the main office of education is to supply conditions that make for their cultivation. The formation of these habits is the Training of the Mind. Dewey, John, How We Think, pp. 27, 28. 4 The Measurement of Fair-Mindedness has pointed out the confining consequences of prejudice. Moreover he has made it clear that he is interested in a sort of open-minded- ness which is not empty-mindedness, but is akin to the meaning of ‘“fair-mindedness”’ as used in these tests. One of the most popular presentations of this same idea is found in The Mind in the Making by James Harvey Robinson. The whole thesis of this luminous book is set forth in a clear-cut demand for a “critical, open-minded attitude” as the solution for most of the world’s ills. Speaking particularly from the viewpoint of the function of higher education, President E. M. Hopkins of Dartmouth College’ said: “‘It is one of the functions of the college to promote open- mindedness—a quality sorely needed today in a period of intolerance that is worse than any other in history.”’ It seems clear, then that any attempt to study open-mindedness will not be labor in a field which is deemed trivial and insignificant. Whatever may have been true of some aspects of educational measurement, certainly the construction of tests in this field will not be a business of testing simply for the fun of testing,—of sticking in prods to see what happens, quite regardless of the social and philosophical significance of the results, if any are attained. There remains another important question with regard to the worth of the enterprise. Some who would grant that the field is one in which social consequences are very significant might well ask: “‘But why tests? Why not experimentation to see what kinds of teaching, what subject matter, and mental training tend toward open-mindedness? Why not a treatise setting forth even more clearly the dangers of prejudice in modern society and an attempt to convince men, yes, to convert them?” In this field, as in every other, selection of desirable method waits 6If some magical transformation could be produced in men’s ways of looking at themselves and their fellows, no inconsiderable part of the evils which now afflict society would vanish away or remedy themselves automatically. If the majority of influential persons held the opinions and occupied the point of view that a few rather uninfluential people now do, there would, for instance, be no likelihood of another great war; the whole problem of ‘labor and capital’ would be transformed and attenuated; national arrogance, race animosity, political corruption, and inefficiency would all be reduced below the danger point. As an old Stoic proverb has it, men are tormented by the opin- ions they have of things, rather than by the things themselves. This is eminently true of many of our worst problemstoday . . . I am not advocating any particular method of treating human affairs, but rather such a general frame of mind, such a critical, open-minded attitude, as has hitherto been but sparsely developed among those who aspire to be men’s guides, whether religious, political, economic, or academic. Robinson, J. H., The Mind in the Making, pp. 3, 12. ' TReported by the New York Sun, Dec. 4, 1924. The Need for Tests of Fair-Mindedness 5 upon the creation of measuring instruments. Schools and theorists set forth their schemes; and one may appear effective, another quite hopeless, but there is no proof of the pudding without some standard by which the eater can make his judgment. Richards, in his presidential address before the American Association for the Advancement of Science said: ‘Plato recognized, long ago, in an often- quoted epigram, that when weights and measures are left out, little remains of any art. Modern science echoes this dictum in its insistence on quantitative data; science becomes more scientific as it becomes more exactly quantitative’’.® Before we can decide whether this course or that, whether lectur- ing or discussion, whether accusation or persuasion will be more effective in the creation of fair-mindedness, we must have first some method of telling in how great a degree this quality of mind is present before and after the application of these devices. ‘After all,’’ some may say, ‘“‘while fair-mindedness is certainly an essential objective in the educative process, and while it is true that we must first have some method of judging as to how great a degree it is present, can we not tell by unrefined observation? Do not people know prejudiced persons from fair-minded persons? Can not people tell you themselves the points at which they are prejudiced and those at which they are open-minded?” All of the evidence seems to indicate that such commonsense judgments are highly unreliable. Probably the best study of the reliability of ratings on traits of character has been made by Rugg.?® After an exhaustive study under conditions unusually favorable he finds that character ratings are valid only when (1) the ratings are made by three or more independent judges, (2) who have been , trained under a leader skilled in scaling so that they are in perfectly clear agreementas to thescale being used, and (3) who are thoroughly acquainted with the subjects. These conditions can very rarely be fulfilled. ‘‘A single rating by a typical school officer will only rarely locate a person within his proper ‘fifth’ of the entire scale.”’ “Hence the apparently dogmatic answer to the question, ‘Can human character be rated on point scales accurately enough for practical uses in education?’ No! It would be far better to give our energies to the attempt to measure it objectively than to make subjective judgments on point scales. The point cannot be made too emphatic that we should discard these loose methods of rating once and for 8Quoted by W. A. McCall in How to Measure in Education p. 8. W 9Rugg, H. O., Journal of Educational Psychology, Feb.-Nov. 1921. 6 The Measurement of Fatr-Mindedness all.’”’ It does not seem probable that we may hope for any greater reliability in ratings of fair-mindedness than he has there discovered with reference to intelligence, leadership, industry, team-work, etc. If we include self-ratings, we may well expect less validity. Dewey, in a recent lecture, said, ‘‘It is of the very nature of prejudice that it is largely unconscious. Everyone is pretty sure that the matters which seem to us to be our prejudices, aren’t.”’ Further light in the same direction came from an investigation carried on for other purposes, described here on pages 25 and 26. One individual, who was rated by a number of the students who were in classes with him as one of the most fair-minded persons in the school, was described by an intimate friend as “perfectly hopeless’’ in his extremism and prejudice upon a number of religious and economic issues. ‘‘Conclusions which he has drawn by ‘intuition’ he is willing to die for, however incongruous they may be with the results of his rational thinking.’’ Another person described this same individual as ‘‘one of the most lovable personalities I have ever dealt with, but quite likely to go off half-cocked on some new idea.’’ In this case at least, there seems to be clear evidence of what Thorndike has called the “‘halo effect.’’ Again, evidence comes from a class of eighteen normal school students studying educational measurements, who were asked to rank themselves and other members of the class on open-mind- edness. They did it under protest, saying they did not trust their ratings, but they did try to do it conscientiously. The ratings which a pupil gave himself yield a correlation of .o7 with the average ratings which he was given by the other members of the class. Fifteen of the eighteen pupils were given ranks which ranged from 3 to 17, or further. A little more evidence is found in the vote of students and faculty members of two institutions with regard to the two most open- minded members of their group. While there was considerable agreement upon a few individuals, the remainder of the votes scattered to include 73 per cent. as many individuals as there were persons voting and an average of 51 per cent. of the possible choices in each group. While all of this evidence may not be conclusive and there is need for further study of the conditions under which ratings in this field will be reliable, this evidence does make it clear that in general such ratings are not dependable or trustworthy. If they are to The Need for Tests of Fatr-Mindedness 7 be proven useful at all, the proof itself will require a test, the validity and reliability of which have already been established. In general, there seems to be clear evidence that there is a need and demand for a type of education which shall bring about pro- gress toward fair-mindedness. Such progress is at best, uncertain, until some instrument can be constructed which will measure the degree to which fair-mindedness, or freedom from bias, has been achieved. The construction of such a test, however, is not wholly in the interests of those who wish to promote fair-mindedness as an educa- tional objective. While the author finds himself in very great sym- pathy with this point of view, the test itself is an equally good instrument for those who believe fair-mindedness undesirable upon certain questions. There is no attempt here to insist that fair-mind- edness rather than prejudice is desirable. Many competent persons feel that upon certain questions the reverse is true. From either point of view the test should prove useful. It merely indicates what the situation is within the individual at the time at which the test is administered. In the light of this evidence, the educator can more discriminatingly select methods which will lead in the direction in which he deems it desirable to work. In order to make the problem compassable, the issues with which this test deals have been limited to religious and economic issues, including in the former certain “‘moral,’’ or “‘amusement’’ questions which have become tied up with religious sanctions. The choice of these two types of issue as the field in which to work was deter- mined in part by the author’s interest, but in large degree by a realization of the tremendous consequences of prejudice and in- tolerance, as contrasted with fair-open-scientific-mindedness in these realms. CHAPTER IT THE CONSTRUCTION OF THE TESTS In the endeavor to construct tests which shall objectively measure fair-mindedness, there are two obvious dangers. One is that the test shall be a measure of agreement with the opinions of the per- son who constructs the test. It is all too easy to feel that those who differ radically from our own point of view are thereby showing strong prejudice. A second danger is that the test shall prove to be merely a measure of convictions, a test of opinions reasonably arrived at and held with a fair degree of confidence. To avoid these dangers, these tests were constructed to measure prejudice defined as: Any tendencies, however produced, (a) to cross out, as distasteful, terms which represent one side or another of religious or economic controversies; (b) to call sincere and com- petent persons who hold different opinions on religious and economic issues incompetent or insincere; (c) to draw from given evidence conclusions which support one’s bias but which are not justified by that evidence; (d) to condemn in a group which is disliked, activities which would be condoned or approved in some other group; (e) to regard arguments, some of which are really strong and others of which are really weak, as all strong if they be in accord with the subject’s bias, or all weak if they run counter to that bias; and (f) to attribute to all the people or objects in a group, characteristics which belong to only a portion of that group. Most of the tests were so constructed that opinions could be expressed on either side of the issue without going to the extreme of registering one of the tendencies which have been defined as prejudices. All of the tests were so constructed that prejudice or lack of fair-mind- edness could be registered upon several aspects of each of the religious and economic issues considered. One might be prejudiced, for example, in the direction of fundamentalism, or of modernism, or of radicalism upon religious questions. Form A, the Word Cross Out Test, is based upon the principle employed by Pressey, that a tendency to cross out an unusually large number of words, feeling that they are annoying or distasteful, is an indication of some sort of emotional set, or conditioning. The following are typical of the fifty-one words included in this test: The Construction of the Tests 9 Bolshevist Dancing Mystic Holy Communion Sunday Blue Laws Unitarian Form B, the Degree of Truth Test, consists of a collection of statements about religious and economic matters, upon each of which it would be possible to find sincere and competent authori- ties in disagreement. The following are typical of the fifty-three statements included in this test: The churches are more sympathetic with capital than with labor. Jesus was more interested in social, than in individual salvation. Poor men cannot get justice in the courts today. The ordinary Catholic priest is well above the average in his community, so far as learning and good judgment are concerned. Prohibition, in the experience of the United States, has been a failure. In each case, the subject is given the opportunity, by checking on a scale, to say that each statement is: (2)—so true that no one with a fair understanding of the subject could sin- cerely and honestly believe it false. (1)—probably true, or true in large degree. (o)—uncertain, or doubtful. (-1)—probably false, or false in large degree. (—2)—-so false that no one with a fair understanding of the subject could sin- cerely and honestly believe it true. The emphatic portions of the first and last positions are under- lined in the directions, so that people will be more certain to notice the extreme character of the statement. Both this test and the Form A test, force many thoughtful individuals into a dilemma, neither horn of which is wholly satisfactory. For some reasons they wish to cross a given word out, for other reasons they would prefer not to. They cannot compromise here. Likewise in the Form B test, they dislike to admit that the statement is only prob- ably true or false but they hesitate to go the full length of the extreme position. The theory of the test is that in such an un- comfortable situation, the choice is more than ordinarily significant. Form C, the Inference Test, presents a statement of fact, followed by several conclusions which some persons might draw from that fact. Thus, for example, it is stated: I. Statistics show that in the United States, of one hundred men starting out at an age of 25, at the end of forty years, one will be wealthy, while fifty- four will be dependent upon relatives or charity for support. 10 The Measurement of Fair-Mindedness [] The present social order cheats the many for the benefit of the few. [] The average young man, under present conditions, cannot count on being wealthy at the age of 65. L) Most men are shiftless, lazy, or extravagant, otherwise they would not need to be dependent. [] The one man is living upon luxuries ground out of the bones of the masses of common people. []Some day the workers will rise in revolt. LI No such conclusion can fairly be drawn. The subjects are asked to check only such conclusions as are established by the facts given in the statement above, drawing upon no other evidence. They are warned to check only those infer- ences which are certain,—none that are merely probable. It is per- missible to check either the final alternative under each case, stating that none of the conclusions can fairly be drawn, or else that one in each case (the second, in this example) which is merely an in- nocuous repetition of the evidence given in the preliminary state- ment. Form D, the Moral Judgments Test, consists of fifteen instances, with an opportunity beneath each, to approve the act, to declare it a matter of indifference, or to disapprove of it. Each instance in the test is parallel to one or two other instances in the type of situa- tion it presents. Thus, ‘‘faith cures’’ are presented, once through the medium of a Japanese idol, once at a Roman Catholic shrine, and once under the leadership of a Protestant evangelist. In another instance unwarranted search is made of a suspected “‘radi- cal’’ headquarters, while in another, the same procedure is carried out with a big business corporation, suspected of dishonesty. These instances are, of course, scattered through the test so as to conceal as far as possible the parallelism. It matters not, for the purposes of the test, whether the subject shall approve, be indifferent to, or disapprove, the action in any situation, if only he will be con- sistent in the parallel act under slightly different circumstances. Form E, the Arguments Test, is based upon the tendency of an individual to feel that all of the arguments upon his side of the case are strong, while those on the opposite side are weak, irrelevant, or very easily refuted. Twelve issues, such as the desirability of the Roman Catholic church, the probability of immortality, the usefulness of the Ku Klux Klan, etc., were selected. Upon each were gathered the principal arguments used by persons supporting The Construction of the Tests II each side of the question. Some were presumed to be weak, while others seemed to the author to be strong. After selection by a few seminar groups, the arguments were submitted to a group of twelve judges, chosen because of their supposed ability to rate ‘religious and economic arguments fairly. They included a pro- fessor of philosophy, a professor of education, a professor of econo- mics, a professor of sociology, two professors of public speaking and argumentation, two teachers of religious education in different institutions, one leader in inter-denominational activities in the industrial field, a member of the International Committee of the Y.M.C.A.anda psychologist. Extraordinary difficulty was found in getting any unanimity. Arguments had to be revised, and re- submitted. Apparently, wide deviations in judgment occurred, less because of bias on the question as a whole, than because of unusual reactions to some feature of the premise in the argument. For example, certain judges were not greatly concerned by the fact that the particular argument favored the child labor law, but did seem to be influenced by the fact that the premise contained the word profit, or employer or farmer. No argument was finally retained in the test upon which there were less than 75 per cent. of the judges in agreement. As further evidence, twelve of the most fair-minded persons, as recorded by the other forms of the test, were studied to see how well their opinions agreed with the judges. The results are tabulated in Table III of the Appendix. It is not at all improbable that many capable students of each issue will look over the ratings given to arguments by other sane and capable individuals, and find them- selves dismayed. Few, if any, individuals have been discovered who do not differ emphatically from the great mass of their fellow students, on some of these arguments. Form F, the Generalization Test,contains a number of general- izations about Jews, ministers, I. W. W.’s, business men, miracles, missionaries, etc. Each of these generalizations is true of some of the members of the group, but not of all of the members. The sub- ject is given a chance to say that the statements are true of “‘All, Most, Many, Few, or No”’ members of the group. Any answer except ‘‘All”’ or ‘‘No” is accepted, and ignored in the scoring. Each of these test-forms has been revised six, seven, or, in some instances, ten times. Preposterous as some of them may seem, no items were retained which were not reacted to in the extreme pre}j- 12 The Measurement of Fatr-Mindedness udice form, by several individuals among the first two hundred to take the test. At the close of the six test-forms is appended a personal data sheet, asking for sex, age, occupation, education, religious training, occupation of father, and years of residence in communities of less than 5,000 people, and more than 100,000 people. The tests may be anonymous, or the name of the subject may be signed. Usually a key, in which the subject can recognize and recover his own paper is desirable. A brief Manual of Directions is published separately, giving the purpose of the tests, the results of standardization, directions for giving the tests, and for obtaining the gross score. This Manual of Directions, together with complete copies of the test and full directions for finding both the gross score and the analytical score, may be obtained from the Bureau of Publications, Teachers College, Columbia University, New York. CHAPTER RELL THE SCORING OF THE TESTS Two methods for scoring the tests have been developed. First, v¥ the tests may be scored to find the general level of prejudice within the individual or group. In determining this “‘gross score”’ the examiner is not concerned with the direction which particular prejudices take, nor with the strength of economic as compared to . religious prejudices, etc. The second method of scoring, while more expensive and complicated is more rewarding. The ‘‘analytical score”’ thus obtained, indicates the extent to which the prejudices / of the subject are in agreement with the special interests of cap- italists, or of economic liberals, or of economic radicals, or of funda- mentalists, or of Roman Catholics, etc. Twelve types of bias are used as the basis of classification. The Gross Score may be objectively determined in accord with the specific directions given in the manual. In brief, the scoring method is: Form A: One point for each word crossed out. Form B: Three points for each extreme statement. Form C: Four points for each unjustified conclusion. Form D: Three points for being indifferent to an act in one case, while approv- . ing or disapproving the parallel act; five points for approving in one case and condemning in the other. Form E: Four points for any question upon which all the arguments on one side of a question have been rated either strong or weak, while all, or all but one, of the arguments on the opposite side have been given the reverse rating. Form F: Three points for making the “‘ All’’ or ‘‘No”’ generalization about any case. This yields a total of 489 possible points. The score, whether for a single form, or for the test as a whole should be expressed in terms of the per cent. of the total possible opportunities to show prejudice, to which the subject has reacted. Thus, a very fair- if minded person may score only 49 points all told, a gross score of ten per cent. ‘A very prejudiced person may take sixty or seventy per cent. of the chances to manifest a prejudice. It will be noticed that in obtaining the gross score the tests have been weighted as follows: Each error in Test A has been penalized 14 The Measurement of Fair-Mindedness one point, each error in Test B three points, in Test C four points, in Test D three or five points, depending upon the amount of in- consistency; in Test E four points, and in Test F three points. These weightings were arbitrarily assigned in accord with the opin- ions of two groups of graduate students who took the tests, suggested revisions, and analyzed their own reactions. The principal result of this weighting has been to decrease the influence of Test A in the composite of tests. Thus, it requires the checking of three words on Test A to contribute as much to the total score as the circling of one extreme statement in Test B or Test F. The resultant score, however, is not greatly changed by this method of weighting. If each prejudice reaction in each test be given only one point, and if the per cent. on each test be computed and the results averaged, then neither the point weighting nor the length of the test influences directly the result. The influence of each form will be determined only by its standard deviation. The standard deviations for the gross scores on each of the six test forms are given in Table IV in the Appendix. The correlation between the gross scores yielded by this method of weighting with the scores yielded by the weighting used regularly in the tests is .QI +.006. The Analytical Score. The second purpose for which the tests may be scored is to answer the question: ‘‘Along what particular religious or economic lines is this individual prejudiced, and howY strong is the prejudice in each of the outstanding lines?”’ For this ' purpose each test element has been classified as ‘“‘The sort of re- sponse which a typical capitalist might make,” or “‘the sort of response which a typical Roman Catholic would probably make,”’ or “‘the sort of response which some other religious or economic type might make.’’ This, it will be noted, does not say that every person who makes that response zs necessarily a capitalist or a Roman Catholic or a religious radical or some other type. It says, merely, that to the extent of that reaction the subject is 7 agree- ment with the typical capitalist or Roman Catholic, etc. Thus, for example, one’s reason for crossing out ‘‘Ku Klux Klan” in Test A may be something very different from a prejudice in the direction of Roman Catholicism, but to the extent of that reaction the in- dividual is in agreement with the typical Roman Catholic. If the test is used to study the prejudices in any individual or group with reference to direction as well as amount, then a copy The Scoring of the Tests 15 of the Analytical Score Sheet! will be needed for each individual taking the test. The Analytical Score will tell to what extent the prejudices manifested by the subject are in agreement with each of the fol- lowing points of view: . Economic radicals. . Economic liberals. . Economic capitalists. . Persons fighting for a ‘‘social gospel,’’ rather than an individual interpreta- tion. 5. Persons interested mainly in a ‘‘personal gospel,’’ prayer, mysticism, communion, salvation, etc. 6. Fundamentalists, orthodox ‘‘ Apostles’ Creed”’ variety. . Modernists, holding liberal Christian views. . Religious radicals, very broad, displeased with most existing Christian manifestations of religion. g. Protestants who are inclined not to like Catholics. 10. Catholics who are inclined not to like Protestants. 11. Persons with high, strict standards of sex-ethics, or amusement, or ‘‘bad habits,’’ or similar moral matters. 12. Persons with broad loose standards of sex-ethics, or amusement, or ‘‘ bad habits’’ or similar moral matters. - WN co “NI In all of the forms of the tests which were used in obtaining the results here published, eighteen typical lines of bias were used. Later, it was discovered that the last six somewhat duplicated the classification made in the first twelve and to save the extra scoring labor the following six were omitted: ; 13. Bias against economic capitalism and the status quo in business. 14. Bias against economic radicals. 15. Bias against orthodox Apostles’ Creed fundamentalists. 16. Bias against liberals and radicals in religious matters. 17. Bias against anything that seems unscientific, mysterious, superstitious or magical. 18. Bias against every religion and everything having to do with religion. A convenient method of representing the analytical score for an individual is the prejudice profile. (See pages 46 to 62 in the Appendix). The per cent. of opportunities that agree with an econo- mic radical which was taken by each individual is represented by the distance from left to right between the base line (zero per cent.) and the vertical profile line. So, likewise, the per cent. of prejudice 1Obtainable at Bureau of Publications, Teachers College, Columbia University, New York City. 16 The Measurement of Fair-Mindedness in agreement with an economic liberal, a capitalist, a supporter of the social gospel, etc., is represented in each case by the distance between the base line and the profile. Obviously, the outstanding lines of prejudice are represented by the columns which protrude farthest to the right. It may seem confusing that almost every individual should register some prejudice along each of these lines and that many individuals seem to agree with contrasting points of view. This is due to the fact that the analytical scoring and the prejudice profile do not necessarily show the reason for any prejudice but only the extent to which that prejudice is in agreement with some types of individuals. Almost everyone has some points at which he is in agreement with his worst enemies. A rabid fundamentalist, for example, may be bitterly prejudiced against Catholics, but at the same time he may share with them a prejudice in favor of the Virgin Birth. In general, however, a person who shows a consider- able amount of agreement with each of the contrasting points of view seems to be the type of individual whose reactions are not clear-cut, but who is prejudiced now against this and now against that, according as he may have heard the view advocated by a person with an unpleasant appearance or seen it referred to in a denunciatory fashion in some newspaper. Grievous errors in the interpretations of the profiles will occur unless it is kept clearly in mind that the analytical score and the prejudice profile do not pretend to register the reasons under- lying prejudice manifestation. This is particularly true in the case of small per cents. of prejudice. A low per cent. of prejudice with which a typical capitalist would agree may be due to very great fair-mindedness on the part of the subject, or may be due toa strong prejudice against capitalists or capitalism. Only the outstanding lines of prejudice, those which, on the profile, extend out markedly ' to the right, beyond the general run of the analytical scores, are to be considered clearly significant. An individual who takes forty or fifty per cent. of the opportunities to agree in prejudices with a typical Roman Catholic but who takes only fifteen or twenty per cent. of the opportunities to agree in prejudices with the other types represented in this scoring, is very likely, indeed, to be prejudiced in favor of Roman Catholicism, Strong evidence for validity exists only for the three or four largest and most marked elements in the analytical score of each person and group. The Scoring of the Tests 17 An important variation of the prejudice profile is its use to show group characteristics. If a number of individuals in some group be tested, their analytical scores be computed, and the average score along each of the lines of bias be found, then this series of averages may be graphed, giving a prejudice profile for the group. Similarly, two or more groups may be compared and their out- standing differences in prejudice graphically recorded by super- imposing one group profile upon the other. (See pages 69 to 84 inthe Appendix). CHARDERSIN WHAT DO THE TESTS REALLY MEASURE? . Reliability. A necessary preliminary question is: Do the tests measure any single attribute, or are they a heterogeneous collection of reactions in which a person might respond one way at one time and a different way at another time? The answer to this question must be stated in terms of reliability,—that is to say, in terms of the correlation between the score made by a person on taking the test at one time and the score which he would make if he were to take it again under exactly the same circumstances, practice effect being eliminated. | The reliability of the general level of prejudice indicated by the gross score was computed by finding the score which would be obtained by half the test (Items I, 3, 5, 7, etc.), and the correla- tion of that score with the score obtained by using the other half of the tests (Items 2, 4, 6, 8, etc.).. Thirty papers, selected to give a normal distribution, yielded a self-correlation of .92 + .02. For 61 papers the self-correlation was .89 + .o15. For 70 additional papers the self-correlation was .96 -+ .ooI. These correlations indicate the relationship to be expected between one-half of the test and the other half of the test. Applying the Spearman-Brown Prophecy Formula we find that it is to be expected that the cor- relation between one administration of the test and another admin- istration under the same circumstances might reasonably be ex- pected to be .96. Occasionally, as in the studies at Lake Geneva (see page 37) Forms A, B, and C are given at one time and Forms D, E, and F at another time. The correlation between one-half of each of these tests with the other half of the same tests has been computed for 189 papers to be .82 for Tests A, B, and C and .68 for Tests D, E, and F. Again, using the Spearman-Brown Prophecy Formula, the self-correlation for Forms A, B, and C is found to be .90; that for Forms D, E, and F is found to be .81. Sometimes, in order to obtain very rapidly a notion of the general prejudice status of a group a single form may be used. The self- correlation for the general gross score indicated by each of the forms is indicated in Table V in the Appendix. What Do the Tests Really Measure? 19 The next question is: How reliable are the various lines of bias indicated in the diagnostic score? The reactions along each of the twelve (previously eighteen) lines of bias may be studied just as though each one were a test in itself. Reactions I, 3, 5, 7, etc., may be correlated with reactions 2, 4, 6, 8, etc. The results of this study are indicated in Table VI in the Appendix. Qther things being equal, a large number of items will yield a higher reliability than will a small number of items, In order to discover how far the differences in reliability indicated in Table VI are due merely to the fact that there are more items in one direction of bias than there are in some other direction of bias, the last column on the right has been corrected by the Prophecy For- mula to a statement in terms of the probable self-r for 200 such points. Within the 12 lines of bias now used for diagnosis there is a variation of from .60 to .88,—not a very significant difference. Vahdity. Evidently the tests afford a consistent measure of something, or, at least, a measure which is as consistent as that afforded by many tests upon which educators place some reliance. The crucial question is: What is that something—is it really fair- mindedness? Six lines of study have been carried on to answer this question: First, the tests themselves are examined with reference to what they seem to be measuring. Second, the correlations between each form of the test and the test as a whole are obtained. Third, a study is made of the scores obtained by individuals who are selected by their group as most fair-minded. Fourth, individuals who are supposed, by those who know them well, to have pronounced lines of prejudice, are given the test, and their reactions compared with those which would be anticipated. This same study includes the analysis of their own scores made by a number of individuals who are accustomed to testing methods and to a reasonable objectivity of mind. Fifth, certain groups\who might be supposed to possess certain lines of prejudice are studied by the test and the results compared with the assumptions of competent judges as to the lines of prejudice which might be expected to exist within the given groups. Sixth, an examination is made of the extent to which the tests may be a measure of intelligence or of opinion rather than of prejudice. 20 The Measurement of Fatr-Mindedness 1. From the standpoint of their construction and their scoring what do the tests seem to measure? Consider first Form A, the Word Cross-Out Test. This is based upon a principle which has been well standardized by Pressey and other students of controlled association, who state that any tendency to cross out, as disagreeable or annoying, a considerable number of words of one type or meaning, is an indication of some emotional instability with reference to that field. Of course, it must be freely admitted that on its face value the crossing out of a word or of even two or three similar words is not necessarily correlative with a lack of fair-mindedness in connection with the subject represented by the words. It might however be supposed that it would be more difficult for a person to deal fairly with subjects which he found distasteful and annoy- ing. Some persons found introspectively that this test-form re- vealed prejudices of which they had not previously been conscious. (See Profile No. 26, Appendix). Also, it was found to be the fair- minded, discriminating individuals who were most annoyed by the test. They could not decide in any case what to do. For certain reasons, they wished to cross the word out. For certain other reasons, they wished to leave it unchecked. The usual result was that a small number was crossed out. On the other hand, many persons found it no problem. They checked word after word with decided emphasis. One Y. M. C. A. leader, noted-for his strong prejudices, exclaimed after crossing out two-thirds of the words, ‘“That’s the most annoying list of words I ever saw in my life!”’ Form B, the Degree of Truth Test, registers the tendency of an individual to hold such extreme views on doubtful issues that he is willing to call those who differ from him incompetent or in- sincere. If a subject believes any one of the statements in the test to be true he is forced into the position of classifying it as ‘‘ probably true,’’ or else, ‘‘so true that no one with a fair understanding of the subject could sincerely and honestly believe it false.’’ The tendency of a prejudice to manifest itself in this fashion was recognized by Locke in his statement: That which is inconsistent with our principles is so far from passing for probable with us that it will not be allowed possible. The reference borne to these principles is so great, and their authority so paramount to all other, that the testimony, not only of other men, but the evidence of our own senses are often rejected, when they offer to vouch anything contrary to these established HELES He ha There is nothing more ordinary than children receiving into their minds propositions... .from their parents, nurses, or those about them, which What Do the Tests Really Measure? 21 being insinuated into their unwary as well as unbiased understandings and fastened by degrees, are at last (and this whether true or false) riveted there by long custom and education, beyond all possibility of being pulled out again. For men, when they are grown up, reflecting upon their opinions and finding those of this sort to be as ancient in their minds as their very memories, not having observed their early insinuation, nor by what means they got them, they are apt to reverence them as sacred things, and not to suffer them to be profaned, touched or questioned. Trotter in his Instincts of the Herd describes vividly the kind of prejudice which is likely to manifest itself in the Form B Test. When, therefore, we find ourselves entertaining an opinion about the basis of which there is a quality of feeling which tells us that to inquire into it would be absurd, obviously unnecessary, unprofitable, undesirable, bad form, or wicked, we may know that that opinion is a non-rational one, and probably, therefore, founded upon inadequate evidence? There is further evidence that outstanding progress in science and in philosophy has waited upon the individual who is willing to admit that there may be truth in points of view not commonly held in his environment and that the judgment of others is worthy of respect, whether they agree with his own conclusions or not. Form C, the Inference Test, illustrates in miniature the practice of the soap-box orator who reels off a list of facts and statistics and proceeds to draw from them conclusions which support his point of view but which have only a very flimsy connection with the evidence he has offered. Persons who have supposed that human beings, when presented with the ‘“‘facts in the case,’’ would be led by those facts to a rational and united conclusion have been abruptly disillusioned if they have made any attempt to verify this with groups of varied background and opinion. Students of the thinking process have long been familiar with this tendency. Dewey summarized the conclusions of many in the following statements: What a person has not only daily and hourly, but momentary need of per- forming, is not a technical and abstruse matter; nor, on the other hand, is it trivial and negligible. Such a function must be congenial to the mind, and must be performed, in an unspoiled mind, upon every fitting occasion. Just because, however, it is an operation of drawing inferences, of basing conclusions upon evidence, of reaching belief izdirectly, it is an operation that may go wrong as well as right, and hence is one that needs safeguarding and training. The greater its importance, the greater are the evils when it is ill-exercised. 1Quoted by Dewey, John, How We Think, p. 24. 2Trotter, W., Instincts of the Herd, p. 44. 22 The Measurement of Fair-Mindedness Conclusions may be generated by a modicum of fact merely because the suggestions are vivid and interesting; a large accumulation of data may fail to suggest a proper conclusion because existing customs are adverse to en- tertaining it. Secondly, next to these are men whose understandings are cast into a mold and fashioned just to the size of a received hypothesis. Such men, Locke goes on to say, while not denying the existence of facts and evidence, can not be convinced by the evidence that would decide them if their minds were not so closed by adherence to fixed belief.’ In Form D, the Moral Judgment Test, one of the most easily recognized manifestations of prejudice is recorded. If an individual condemns the socialists for passing out leaflets through the neigh- borhood and then whole-heartedly volunteers to assist his church in leaving a leaflet in front of every door in the neighborhood, it does not seem too much to say that he is lacking in fair-minded- ness on that issue. So, likewise, with the man who approves a raid of radical headquarters but is quite indignant at the thought of raiding a corporation on a suspicion that its financial transactions have been shady, it would be commonly assumed that he possessed a degree of prejudice on that issue. This test seems to offer almost prima facie evidence of its validity. Form E, the Arguments Test, registers the tendency familiar to every propagandist and debator to feel that the arguments on his side of the case are strong and convincing and that those of the opponents, to each of which he feels he possesses a perfectly ade- quate answer, are quite trivial, irrelevant, or hardly worth con- sidering. Again, Locke’s insight led him to recognize this hindrance to philosophical thought: This kind is of those who put passion in the place of reason, and being re- solved that that shall govern their actions and arguments, neither use their own, nor hearken to other people’s reason, any farther, than it suits their humor, interest, or party. Jevons points out that It is exceedingly rare to find persons who can with perfect fairness estimate and register facts for or against their own particular views and theories.’ While with children this tendency is present in a very obvious and ndive form, with adults many things besides the conclusion 3Dewey, John, How We Think, pp. 18-19. ‘Jevons, Wm. S., Principles of Science, Vol. II, p. 5. What Do the Tests Really Measure? 23 to which the argument leads tend to influence their decisions. This was illustrated by the difficulty experienced in finding argu- ments so obviously strong or obviously weak that fair-minded persons would recognize that strength or weakness. (See page II.) Therefore, only a coarse grading is used upon this question and a variation from the opinions of the judges is not regarded as evi- dence of prejudices until it affects so many arguments that all on one side are regarded as strong or as weak; and all on the opposite side, or all but one of those on the opposite side, are given a con- trasting rating. Form F, the Generalization Test, rests upon what may be termed the ‘‘one-swallow-makes-a-summer’”’ fallacy. This test registers the tendency in individuals to say that all Jews would cheat, that no church members are violating the Volstead Act, that all Roman Catholics are superstitious or that none of them are, and so on,— attributing to all the members of a group characteristics which are true of only a part of the group. It has been rather clearly demonstrated by the testimony of a number of individuals that they acquired some of their race-prejudice in a single instance, or two, and afterwards reacted to all members of the race in terms of the conditioning of the single experience. (See the modification of this test for use in measuring fair-mindedness on race-relations, page 41.) In general, then, it is evident that these tests may be validated in the way in which examinations in school subjects always used to be validated and in the way in which many tests are now vali- dated. Why is this an arithmetic test? Because it contains arithmetic problems and people who can do this sort of trick would naturally be good in arithmetic! Why is this a geography test? Because on their face the questions deal with geography! Why is this a spelling test? Because the reactions registered are in terms of ability to spell! It is only reasonable to expect that any persons, for whom arithmetic and geography and spelling tests are validated because of the material they contain, will regard these tests of fair-mindedness as sufficiently validated by the evidence offered above. The fact, however, that such evidence is often misleading requires that a thorough demonstration of validity shall proceed further. \ 2. It is quite possible that some persons will feel fairly well convinced that the majority of the tests give indications of prej- 24 The Measurement of Fair-Mindedness udice as contrasted with fair-mindedness, but feel skeptical about one of the forms. Form A is particularly subject to this criticism. The answer to all such objections must be in terms of correlation, whatever the test may appear on the surface to demonstrate. If one of the tests really does measure the same thing which the rest of the forms measure, then it is as valid as they are, even though it appears to common sense to be a record of opinions, of reading ability, or of stolidity. In Table VII in the Appendix are given the data for correlation between the gross scores obtained by using the whole test and those which would be obtained by each of the forms. These data are-given first for forty papers and then for one hundred nineteen additional papers. In the raw form they range from .45 to.86. Corrected for attenuation so as to indicate the true correlation between the reaction indicated by one form of the test and the reaction indicated by the total, the correlations range from .53 to .94. Of course, in these correlations each test form is included in the total with which the separate form is correlated. If the correlation be made between one test-form and the average score on the other five forms, then the resulting correlations range from .11 to .42. (See Table VIII in the Appendix). Since these tests really measure manifestations quite different in form, of a quality of prejudice or fair-mindedness which has popularly been assumed to be a more or less homogeneous unity, it is probably fair to compare these intercorrelations with those existing between. the various manifestations of arithmetical ability or reading ability, or intelligence. McCall> finds a correlation between addition and general arithmetical ability of .12 in one class and .34 in another group. The same study shows an uncorrected correlation between reading ability and visual vocabulary, of .15 in one case, .40 in another, .45 in another, and .60 in a fourth. Steacy in ‘‘Inter-Relations of Mental Abilities” finds a correlation between various manifestations of arithmetical ability with other manifestations varying from .16 to .42, or, when corrected for attenuation, from .36 to .94, the last correlation existing between ability to subtract and ability to divide. A study by Burt on ‘‘The Distribution and Relation of Educational Abilities’? published by the London County Council indicates a correlation between knowl- edge of arithmetical rules and ability to do problems of .76; a 5McCall, W. A., The Correlation of Some Psychological and Educational Measurements, 1916. What Do the Tests Really Measure? 25 correlation between speed of reading and comprehension of reading of .69; and a correlation between speed of writing and quality of writing of .30. The author studied a group of 30 of the same students who were studied by the tests of fair-mindedness, using the Thorn- dike-McCall Reading Scale. He found a correlation between speed of reading and comprehension of reading of .21. Probably the most comparable test is the Pintner Non-Language Test for Intelligence in which the correlation between each test and the composite was studied by Dorothy R. Morgenthau.* She found the correlation between individual tests and the composite score to vary from .40 to .76. It seems fair, therefore, to conclude that each form of this test is as closely related to the purpose of the test as a whole, as are the various items in an intelligence test to the score of the test as a whole; and that the forms are more closely related in these tests of fair-mindedness than is commonly the case in testing various forms of arithmetical ability, reading ability, writing ability, and. the like. Sometimes it may be desirable to use one form of the test as a substitute for the test as a whole, and to obtain with this an analyzed score. In such cases it is useful to know how well each form of the test correlates with the test as a whole, in the diagnosis which it offers. This relationship was studied in two ways. First, on each of 119 cases, the diagnosis given by each form of the test was correlated with the diagnosis given by the test asa whole. The resulting 714 correlations were worked out by the Rho method, and transmuted into terms of Pearson’s r. The average correlations of the analytical score on one form of the test, and the analytical score given by the complete battery, range from .48 to .61 as in- dicated in Table IX in the Appendix. Second, the correlation between the per cent. score on each of the analyzed lines of prejudice as indicated by one form, and the corresponding score yielded by the complete test was computed for 40 papers, covering 415 diag- nostic lines for each form. These results which range from .42 to .86 were worked out in the regular product-moment fashion, and are listed in the second column of Table IX in the Appendix. 3. A further step in validation consisted in obtaining a consensus of judgment upon the most fair-minded individuals in certain groups, and discovering the test reaction of these selected individuals. &Morgenthau, Dorothy R., Archives of Psychology, No. 52, D- 33- 26 The Measurement of Fatr-Mindedness In two graduate schools the faculty were asked to select, by secret ballot, the most fair-minded members of their group. In one small graduate school the student body was asked to select by secret ballot, the most fair-minded members in the student body and also in the faculty. In each case it was made clear that fair- mindedness with respect to religious and economic issues was to be the criterion of selection. It was evident from the lack of agreement inthe vote that there are very wide variations in the reputations of rather closely associated individuals with regard to their respective qualities of fair-minded- ness. However, it was assumed that a valid test should show some significant difference between the two or three most frequently mentioned in the vote of each group, and the ordinary student in the school. All of the individuals ranking among the first three in their group were invited to take the test. Copies were received from seven individuals, indicated by a star in the summation of the votes, in Table X in the Appendix. If we consider the gross score, we find that while the average Methodist minister in a certain mid-western state registered 37 per cent. of the chances to show prejudice, and while the average college student or normal school student registered 30 per cent., this group of selected individuals averaged only 13 per cent. The differences between these people selected for fair-mindedness and the average theological student in Union Theological Seminary, is indicated on the group profiles shown on page 82 of the Appendix. 4. Some of the most significant material for validation is in the form of case studies. In twenty-four cases a description of the prejudices which might be expected in some individual was secured in advance from some relative or close friend. Then the individual was given the test, and the analytical score graphed in a prejudice profile. In thirty-one other cases, the subject was given his profile, after taking the test, and asked to make his comments upon it. The descriptions, whether given by friends before the subject took the test, or given by the subject himself afterwards, were then stud- ied with reference to their agreement with the profile. While all fifty- five of the profiles with the parallel descriptions are on file with the original manuscript in Teachers College Library, Columbia University, New York City, only a few representative samples have been included in this study. Profiles Nos. I, 2, 3, 4, 5, 6, II and 12 in the Appendix, are typical of the first type, in which com- What Do the Tests Really Measure? 27 ment was made by some friend of the subject, while Profiles Nos. : 25, 26, 27, 28, 31, 32, 45, 46 and 55 in the Appendix record the reactions of the subject himself. Several factors tended to make the comments agree less closely with the profiles than they might have been led to agree. Thus, in the studies of the first type no leading questions were asked. The types under which classification would be made were unknown to the person making the comments. The descriptions represent a reaction to the general question, ‘‘What religious and economic prejudices would you expect to find in . . .?” In the self- reactions, two troublesome factors entered. One was a misunder- standing of the phrase “‘prejudice which is in agreement with.”’ Several people found it difficult to believe that they could have some prejudices with which a capitalist would agree and others with which an economic radical would agree. A second factor was a tendency to criticize some form or element in the test, to defend original reactions, etc., instead of facing the question as to whether or not the results were a fair description. It was not made clear to these people that the scoring system passed by about four-fifths of the reactions they had made on the test and took account only of the extreme reactions. They naturally tended to think in terms of the marks they had made on the test paper, most of which were, indeed, opinions, as contrasted with prejudices. Also, they were probably not free from the almost universal tendency to distinguish between ‘‘my convictions” and “your prejudices.”’ Case studies like these do not lend themselves easily to summaries, but an attempt has been made in Tables XI and XII in the Appen- dix to classify each response in terms of agreement with the test result. The first twenty-four statements made by people about their friends, seemed to present fifty-nine definite expectations which could be checked against the analytical score. Of these, forty-three could be classed as ‘Clearly Fulfilled,’’ ten as ‘‘Doubtful,” or ‘Possible, ’’ and six as ‘‘ Not Indicated.” It is probably significant that five of these last six were predictions which had been made by one individual. While such a summary is less valuable than a study of the particular cases, it is probably fair to conclude that approximately three-fourths of the unguided and general pre- dictions made by relatives and friends seemed to be verified by the test results. The second group of profiles has been even more difficult to summarize because so much blanket approval was indicated. What 28 The Measurement of Fair-Mindedness does it mean when a person says, “‘I think the test shows my prej- udices very well?’’ In this summary it has been assumed that such statements represent a real reaction to half of the eighteen analyzed scores-shown on the profile. Thus, if a man condemned the whole test as unfair, he is recorded as having rejected nine predictions, and accepted. none. If he approved in general, but objected at two points, he is recorded as having approved seven of the scores and rejected two. On that basis, which is obviously not unduly favorable to the test since most of the blanket reactions were positive, there were reactions to 272 analyzed scores. Of these, 220 were approved and accepted, 52 were questioned or rejected. Again, insisting that the summary is less valuable than are the particulars out of which it has been approximated, individuals confronted with their own specific scores seem to approve in about four-fifths of the cases. 5. Studies of individual cases have a distinct disadvantage to the world at large, in that the individual is not known. If the ordinary individual sees his own prediction about a friend clearly revealed by a test profile, the test having been administered and scored by someone who had never seen nor heard of the subject, he is likely to be deeply impressed. Certainly, his impression is much deeper than that he possesses upon hearing of the results in the case of a score of persons, none of whom he knows. Perhaps if it were feasible in a land of liberty, to broadcast the prejudices of well-known religious or political leaders the study of individuals could be made fairly conclusive validating evidence. However, at present it must remain far more convincing to the writer and to the people who participated in the study, than it is likely to appear to others. This disadvantage can be partially overcome by a study of people in groups. The prejudices of a certain student in Union Theological Seminary are known to very few people. The prejudices of Union ‘Theological Seminary students in general are fairly familiar to a large number of people. There are many people who possess what they believe to be a competent opinion about the differences in prejudice between Roman Catholic students and Protestant stu- dents in general. Study was therefore made of certain available groups. It would be expected that valid tests would show the kinds of prejudice within each group, and the differences in prejudice between one group and What Do the Tests Really Measure? 29 another, corresponding to the judgment of competent judges. The group profiles, shown on pages 68 to 81 of the Appendix indicate graphically the average prejudice status of each group and the comparisons between certain groups. It is useful to note not only the differences between groups which stand out so vividly on these graphs, but also the outstanding characteristics of each group. Note, for instance, the prejudice of Methodist ministers for strict amusement standards, and that of students in Union Theological Seminary for liberal economic reforms. Attempt to check the findings of the test against the experience of competent observers was made informally at first. Groups of graduate students in three different seminars were asked to announce the places at which they would expect to find a difference in prej- udice between certain groups, for example, between Methodist ministers and normal school students. Approximately twenty different predictions were made about some of the group compari- sons, involving at least fifty different judges, and without exception the tests revealed the outstanding differences which these groups of students anticipated. Still there remained the probability that these students did not represent very wide-spread experience and judgment and that they did not give more than passing considera- tion to the question. Therefore, a further and more definite check was made by sending questionnaires to twenty-five persons, engaged in several types of occupation, living in various sections of the country, but persons who were likely to be familiar with at least some of the groups to be compared. These judges were asked to check the lines of bias (corresponding to those used in the analytical score) along which they would expect to find one group markedly more prejudiced than the other. They were asked to put two plus signs at points at which they felt particularly certain that a large difference in bias existed. Fifteen judges replied, but, due to an ambiguity between “prej- udice for’? and “prejudice against,’’ only thirteen of the replies were capable of clear interpretation. On the page facing each group profile (pages 68 to 81 in the Appendix) is given the summary of the ratings made by these judges-for the groups compared in the profile. —The maximum number of votes possible at any point of difference was twenty-six, but the double-plus was very rarely used. The mean of the judges expecting a difference on any item was about seven. Usually the vote was divided, some judges expect- 30 The Measurement of Fair-Mindedness ing one group to be more prejudiced, many expecting the other group to be more prejudiced in agreement with the suggested bias. When- ever one group was given a majority of six or more votes over the’ contrasted group, this was regarded as evidence that a real difference in prejudice might be expected. The analyzed test score, should, if valid, show a difference at each such point. The degree of agreement may be studied in two ways. First, we may ask, ‘‘If the judges be assumed to have the more reliable insight, and if all the points at which they expected a marked difference are taken, will the tests reveal such a difference?”’ As indicated above, a majority of six judges expecting one group to be more prejudiced than the other, was regarded as evidence of a ‘‘marked difference.’’ In the case of the profiles a marked dif- ference was assumed to exist in the case of the largest differences in score, as many, in order of size, being included, as there were marked differences indicated by the judges. That is to say, in the case of two groups in which four marked differences between them were indicated by the judges, the first four differences in order of size, on the profile sheets, were considered marked differences in the test score. The results are summarized in Table XIII in the Appendix. In general, the tests showed a difference in the direction in which the judges expected a marked difference, in 91 per cent. of the predictions, and showed it to be relatively a marked difference in 81 per cent. of the cases. The second approach can be made in the other direction. It might be true that while the tests did show differences where the judges expected them, the tests also would show many additional differences not anticipated by the judges, and presumably less valid. Therefore, the four largest differences indicated by each profile were selected and the number of judges agreeing that such a difference was to be expected and the number who would have expected the difference to be in the other direction, were tabulated. These results are given in Table XIV in the Appendix. In general it appears that 221 judges’ votes would anticipate differences where the profile found them to be marked, and that 22 judges’ votes would have expected the difference, but in an opposite direc- tion. Six of the twenty-two who differ, do so in connection with the comparison between students in Wisconsin and similar students in New Jersey. Unfortunately, none of the judges who replied were familiar with both groups from first-hand experience. Hence, it What Do the Tests Really Measure? 31 may be that a considerable proportion of the difference is to be attributed to the inexperience of the judges. Eight of the twenty- two dissenting votes occurred in the comparison between the older and younger Methodist ministers. The older ministers, on the test, registered more prejudice which agreed with the economic liberal, than did the younger men. The reason was obviously that the older men were more prejudiced in general. They had more prejudices of all kinds, radical as well as conservative, than younger men had. This, and perhaps the social radicalism of the Methodist | church in general, did not seem to be adequately considered by the judges. In general, then, it appears that the findings of the test with regard to group differences are confirmed by competent judges in something more than 8o per cent. of the cases, and that in the remaining situations there is some probability that the discrepancy is due to variables affecting the rating given by the judges. 6. It is now necessary to consider the question: Are these tests measures of something other than fair-mindedness? One ever- present danger is that a test with peculiar directions and much reading material will prove to be a measure of comprehension of reading. Two further suggestions have commonly been made. One is that the tests may measure intelligence, or at least that Forms C and D seem very much like questions used in intelligence tests. The second is that these tests take thoughtful, enlightened opinions and call them prejudices, so that the only person likely to make a low score would be a wishy-washy person of vast ignorance and no convictions. Data available for forty normal school students make possible a correlation between the prejudice score and the score made for comprehension of reading using the Thorndike-McCall Reading Scale. The correlation is —.005. The relationship to intelligence may likewise be studied in terms of correlation. Here data are available for fifty-nine students, each of whom took two or three group intelligence tests (Terman, Morgan, Illinois, Otis, or James). These scores were averaged and the resulting I. Q.’s ranged from 74 to 124. The correlation between intelligence and the prejudice score based on the whole test, was found to be —.08. The correlation with the prejudice score on Form C was —.06 and the correlation with the Form D test .o9. It is rather surprising that within this group there should be no apparent relationship between the amount of intelligence and a9 The Measurement of Fair-Mindedness the amount of prejudice or of fair-mindedness. One factor seems to be that several of the most intelligent individuals of the group had been raised in sheltered middle-western homes, and had been nourished with a strong prejudice against religious radicals, liberal morals, and economic radicals. At the other end of the intelligence scale were several men who had, through knocking about in the world, acquired more years of experience and a_ rough-and-ready tolerance. Form C, the Inference Test, while it seems superficially to re- semble a test frequently used for intelligence, apparently is not a good measure of intelligence for the kind of group represented in college. This is due in part, of course, to the high selection, so that the in- telligence required to see that one statement is more closely related to the evidence than is some other statement, is present in prac- tically all of the persons. When an individual checked one of the incorrect conclusions, it was not because he did not see clearly a difference between that statement and the facts placed above it— it was rather because he had so strong an emotional reaction toward the statement that he required very little provocation to register his conviction that the statement was true. The hypothesis that _ the basis of the reaction was emotional fervor and not intellectual | discrimination was even more clearly illustrated with Form D. One student was speaking the experience of many others when she said, ‘“‘T see my inconsistency,—the cases are the same—but what are you going to do? This one is in the Bible!”’ Certainly such an experience is not out of accord with the con- clusions of psychiatrists who have dealt with that aggravated form of prejudice known as paranoia. Alert and keen insight into mat- ters not emotionally conditioned is not uncommon. Glueck’ says: “‘Memory, and the capacity to acquire new knowledge re- main intact, and reasoning and judgment on matters of every-day life which do not touch his more or less circumscribed delusional field, may remain quite normal.”’ The other question, as to whether this is really a measure of fair- mindedness or merely a measure of empty-mindedness is answered partly by the inherent character of the test items. It is perfectly possible to express an opinion in most of the forms of the test with- out expressing an opinion that goes so far as to call its opponents names, to draw unwarranted conclusions or to discriminate un- 7Glueck, Bernard, Studies in Forensic Psychiatry, p. 136. What Do the Tests Really Measure? 33 _ fairly against groups that are disliked. Thus, in the study of the annual conference of the American Country Life Association, ;re- ported on pages 39 and 40, it was found, using a test composed in large part of Forms A, B, and E, that the three-day conference pro- duced a change of opinion on 15 per cent. of the items, but practically no change in prejudice. In the Lake Geneva Y. W. C. A. conference, reported on page 37, it was discovered that during the longer conference there was an actual reduction of the amount of prejudice. Moreover, the reduc- tion was apt to be greatest in those lines most discussed by the students, while on items which did not come into the conference there was little or no change. In the study of race relations reported on pages 40 and 41, the result of more enlightenment, study, and discussion, was clearly to decrease the prejudices of the members of the class, as recorded by a test similar to Form F. Even better evidence comes in the group comparisons. It is surely unreasonable to assume that the reasons why theological students at Yale Divinity School and at Union Theological Sem- inary are more open-minded on religious questions than is the ordinary college student, are that they have given less thought to religious questions, or are less well informed, or are empty-headed upon such subjects. Likewise, it is surely unfair to interpret the low score obtained on the test by those individuals, professors in theological seminaries and other graduate institutions, who were voted most fair-minded by their colleagues as being simply in- dications of ignorance or wishy-washy indecision. There is good evidence, on the other hand, that persons can be well informed, enlightened, and at the same time strongly prejudiced ona given subject. The prejudice of the financial leader in Profile 11 was not due necessarily to ignorance of economic factors and con- ditions. There was a very close relationship between his study and interest in that field and his prejudice. So, likewise, the ministers of the mid-western state registered considerably more prejudice than did the college students on most religious and moral questions. This prejudice, too, correlated highly with the amount of informa- tion and interest in the field. Probably the distinction lies not in the amount of information or study but in the character and method of that study. This seems to be evidenced in the case of the con- trast between those ministers who have had college and theological 34 The Measurement of Fair-Mindedness training and those who have not. The latter were much more apt | to have acquired their information and experience in a biased and one-sided manner. Very evidently it is possible for a person or group to hold strong convictions, to be well informed, and not thereby to increase the prejudice score on this test. The converse situation, that of having no convictions or information and thereby making a low score, must be admitted as a possibility. Every test should recognize that causes other than the major one may be influential upon the score. A pupil may score low in a geography test because of lack of ability to read rather than because of lack of geographical in- formation, or he may make a high score because of a working com- bination of slyness and dishonesty. It is not at all impossible that a person may be able to secure a low score on this test for some other reason than his fair-mindedness. He may be lacking in con- viction, or he may be exceedingly wary in expressing himself on paper, or he may stupidly misread the directions. Aside from Profile No. 12, in which it seems that an individual’s written re- actions are much more carefully guarded than are his oral outbursts, no clear cases of other factors affecting the score have been en- countered among the approximately five hundred persons, from varying environments, who have been studied. In every case in which a very low prejudice score was made, some investigation was carried on. In all of the cases for which data were available, the person was found to be one highly respected by his associates as a person of fair, discriminating mind. In certain other cases, the subject reported his agreement with the findings of the test and shared his reasons for believing that he deserved to rate as less prejudiced than his fellows. (See especially Profiles No. 27 and 55 in the Appendix.) It seems probable to the author that the number of persons who go about empty-minded, unable to react intellectually or emotion- ally to words, situations, and arguments commonly used in heated and controversial fashion by their fellowmen, has been considerably over-rated. The weak, flabby, uncertain, ill-informed people who would seem to be the best timber for such a product appear usually to register a high but inconsistent score. They hate this class of people and ‘‘adore’’ that idea, with little insight, reliability, or consistency. They have been called ‘‘spineless”’ and ‘“‘vacillating’”’ not because of their lack of reaction but because of the abundance What Do the Tests Really Measure? 35 of reaction. They are not unaffected by any breeze, they bend with every zephyr! Summary. What do the tests really measure? Judging by their reliability, they are a consistent measure of something. At least they are as consistent as is the ordinary group test. At least within a selected group like college students, the ‘‘something’’ which they measure seems not to be ability to comprehend reading, nor yet intelligence, nor even information upon these questions. If we may judge by the nature of the tests themselves, they seem to measure fair-mindedness as contrasted with prejudice or bias. This seems to be true in every form of the test. The tests as a whole appear to measure something in which persons who have been selected as very prejudiced register a very high score of 50 per cent.,60 per cent., even 70 per cent., while average students register 30 per cent. and the most fair-minded individuals register 5 per cent., 10 per cent. or 15 percent. The tests reveal approximately the same prejudices which close acquaintances have discovered by long association. Most individuals, confronted with the results, believe them to be a fair measure of their own prejudices. The tests applied to groups show differences between those groups at the points at which competent judges would expect to find differences in prejudice. Perhaps, therefore, it may be concluded, that, unless conflicting evidence should be obtained, these tests may be regarded as usable measures of prejudice as contrasted with fair-mindedness, upon certain religious and economic issues. CHAP FER SUGGESTED USES FOR THE TESTS Probably the most obvious use for tests of fair-mindedness is to determine what the situation is in a given group or individual. Imagine a politician who is called upon to make an address before an audience of considerable importance, and about whose prej- udices on the question at issue he has little or no information. What would it be worth to him to know, in advance, the direction and degree of prejudice? Not infrequently a teacher, especially in a controversial field, wastes time and effort because he does not appreciate where the bias is strongest in his students. Consider a minister entering upon a new pastorate, having dealings with a new group of officials and of young people. How many disastrous steps might be avoided, were it only practicable to administer such a test to his constituency and forewarn himself as to the prejudices which he may expect to encounter! Even in the present, less tractable world, such surveys seem useful. Recently, a travelling secretary for the Student Friend- ship Fund reported that he was using a modification of Form A, in connection with his work in colleges. Coming into a strange college town, he passed out his blanks to men in dormitories and fraternity houses, scored rapidly a number of samples, and was ready when the hour for his chapel talk arrived, to avoid or meet fairly the existing prejudices of the school. On page 39 is described the experience of the discussion leaders in a national conference who had been ill advised as to the point of view and the interests of the group they were to lead. The application and study of these tests on the first day of the conference, saved the whole conference from an uncertain but probably unfortunate fate. What has been done in these situations in a rather hasty fashion, might well be done carefully and accurately by any group seeking to study the existing prejudices in certain geographical, economic, or religious groups. | : A second way in which tests of this sort may prove useful was suggested in Chapter I. When investigating the need for tests of fair-mindedness, it was pointed out that the creation of accurate tests was a desirable prerequisite to the study of the effectiveness Sugeested Uses for the Tests 37 of various teaching methods. Once such a test has been prepared, it is possible to test, in equated groups, the effectiveness of lectures, sermons, prayer-meetings, moving-pictures, pageants, group-dis- cussions, posters, reading assignments, advertisements, and other supposed techniques for influencing public opinion. A method of this sort was followed in studying the effectiveness of certain conferences. One of these was the Lake Geneva Y. W. C. A. student conference, held in the summer of 1924. Using a method of rotation, part of the conference took Forms A, B, and C of these tests at the beginning, and then took Forms D, E, and F ' at the end. The remainder of the conference, supposedly equated by chance selection, took Forms D, E, and F, at the beginning and then took Forms A, B, and C, at the end. The conference was conducted by Professor Harrison S. Elliott of Union Theological Seminary upon the principles laid down in his booklet, The Why and How of Group Discussion. Essentially, the method involved an endeavor to recognize the worth in points of view which differed from those which had been brought into the conference, so that after a free give-and-take of discussion and an utterly good-natured sharing of information and ideals, each person might leave the conference with more of truth than he brought to it. The main exception to this program came in dealing with certain industrial questions. Speakers were brought in to emphasize the viewpoint of large groups of working girls in the great cities. In gross score, the groups taking Forms A, B, and C averaged 36 per cent. at the beginning of the conference, and 30 per cent. at the close. The groups taking Forms D, E, and F, registered 26 per cent. of the chances to showa prejudice at the beginning of the conference, but only 23 per cent. at the close of the conference. The comparison between the beginning-group and the end-group for each set of test forms is clearly shown in the group profiles on pages 83 and 84 of the Appendix. The slight increase evident in both cases upon economic-liberal points of view, may have been due to the influence of the speakers who stressed the necessity for economic reforms. The real significance of the changes indicated cannot, however, be definitely established until we know just what the effect of taking Forms A, B, and C is likely to be upon individ- uals who subsequently take Forms D, E, and F, and what the reverse effects would be. It seems probable that in a similar fashion the tests may be used effectively in the study of curricula. For example, suppose the 38 The Measurement of Fatr-Mindedness students entering a theological seminary were tested at the beginning of the year, and again after a year of study. Comparison might be made between the changes taking place in the prejudices of students|who had worked in the philosophy of religion, or church history, or systematic theology, with students who had worked in field situations, in the study of modern social movements, or of psychology, or religious education. The changes which have taken place in one group as compared by partial correlations or some similar method, with changes which have taken place in another ‘group, might well prove some index of the effect of given subjects in producing prejudice or fair-mindedness. Herbert A. Sturges of the University of Chicago has made an interesting beginning upon such an experiment in some of his, Studies in the Dynamics of Attitude.| He has measured the shifts in ideas and opinion which have been due to taking certain courses in sociology, or to reading certain material. Measurement of the shifts in more deep-rooted aspects of attitude would be desirable supplements to such studies. Of course, it is not contemplated that the entire curriculum of a theological seminary, or the contents of a course in sociology, or the method of promoting foreign missions, or carrying on a political campaign, shall be determined solely by its effectiveness in producing certain degrees of prejudice or of fair-mindedness. Nevertheless, every such study, carefully made, will bring us one step rlearer to a scientifically established curriculum. Still a further line of very wide usefulness is opened up by the possibility that the content of the tests can be varied without de- stroying their validity as measures of prejudice. While this is not yet established, it seems probable that the evidence for validity will be significant beyond the immediate content of this set of ques- tions which will probably be out of date in ten years. The most useful contribution may prove to be not the materials, but the method. Thus it may be possible to substitute in Form A, words from controversial realms other than the religious or economic. It may be possible to substitute in Form B other statements upon which competent authorities might well differ; in Form C different evi- dence with distorted conclusions running into other fields of interest; in Form D, parallel instances from other fields of vital human ex- perience; in Form E, arguments upon political, or international, 1Obtainable from H. A. Sturges, 5707 Kenwood Ave., Chicago, IIl. Suggested Uses for the Tests 39 or scientific questions; in Form F, generalizations about races or families or political parties or forms of literature or art or almost anything else. With comparatively simple steps in standardiza- tion, these new tests could be used to measure fair-mindedness in realms far removed from the religious or economic. Two preliminary attempts have been made. The first was an adaptation of Form A, Form B, and selected parts of Form E, to fit the issues which might be considered by the American Country Life Association at its annual convention in 1924. The material contained some of the same test items which are in the standard Watson Test of Public Opinion, but fully three-fourths of the items dealt directly with scientific agriculture, tenant farming, consoli- dated schools, the Non-Partisan League, country churches, and similar issues. These new items were formulated in the same way that the original test had been formulated. (See pages 8 to 12.)? The first purpose which these modified tests served, was to reveal to the leaders the actual state of affairs in the minds of the dele- gates. The discussion leaders had not correctly appraised the per- sonnel of the conference. They had made their plans supposing that certain issues represented the real areas of difference in opinion among the delegates. The tests given at the opening session not only provoked interest and discussion, but revealed that the leaders had been mistaken. The prejudice reactions, by inspection, and brief tabulation, revealed real and vital issues, but not at the partic- ular places where the leaders had expected to find them. On the issues which the leaders had planned for discussion, the conference held an almost unanimous prejudice. Only this discovery and a rapid readjustment of program prevented some hours of flounder- ing in the limited time available for group thinking. A second purpose which the tests served, was to record the shifts in opinion which took place between the beginning and end of the conference. In general these appeared to be toward the idea of the church as a very necessary and desirable agency in country life, but toward the idea of the church as at its best when it broadens out and serves the whole community in a modern fashion. The third purpose which the tests served was to register the changes in prejudice which took place. The scoring system was 2A complete copy of the test, and a description of the conference, with a discussion of the results, may be found in the Proceedings of the American Country Life Association, Seventh National Con- ference, American Country Life Association, Grand Central Terminal, New York City. 40 The Measurement of Fatr-Mindedness exactly the same as that for Forms A, B, and E, so far as the gross score was concerned. Analytically, instead of studying the twelve lines of bias which the Watson test uses, six lines of bias, represent- ing the radical and the conservative position on rural economics, religion, and community service, were used. Sixty papers were received at the beginning of the conference, and these were compared with thirty-three turned in at the close of the session. Twenty- two papers were known to be from the same person, both at the beginning and at the end. Probably all of the thirty-three were included in the original sixty. In any case, examination of the twenty-two pairs which could be identified gave the same results as did the study of the wider sampling. The summary of results is given in Table XV in the Appendix. The first line gives the gross score for the entire test, 34 per cent. at the beginning of the con- ference, 34 per cent. for the close of the conference. Below this are given the gross scores for each of the test forms used, and then in the lowest section of the table, the analytical score along each of the six lines of bias studied. The significant fact is that apparently this three-day conference produced no change in the average prej- udice state. Opinion shifted markedly, but prejudices remained surprisingly constant. This stability may have been due, in part, to the fact that some of the issues most warmly discussed at the conference were not in the test, and some of the issues in the test, due to the misappre- hension regarding the interests of the group, did not come out into the discussion. However, if that be true, the results speak well for the reliability of the converted test. A second adaptation was made to the needs of a class in Christian- ity and the Race Problem. Form F was taken as the model and generalizations about race relations were included in the test. On page 41 are given some of the statements included, indicating before each the number of persons voting ‘‘ All, Most, Many, Few, or No”’ at the beginning of the class and similar votes after the course of SIX sessions. . Taking into consideration the entire forty statements, of which these are only typical samples, and expressing reactions in terms of per cent. to equate the before and after groups which were different in size, the results were as follows: At the beginning, the class on the average took 20.6 per cent. of the chances to say ‘‘All,’”’ and 10.7 per cent. of the chances to say pene geen pe Ae ye Ot Pan ee Dew pW ew pe ee ee ee Suggested Uses for the Tests AI SOME OPINIONS ON RACE RELATIONS B. Before the course of study. A. After the course of study. All Most Many Few No ? 2| 13} 30} ©} oO} 2) Jews will try to get the best of a bargain if they 6} 3] 21} 5] o| 1| have to cheat to do it. 33| 9] 2] ©} O} 3) Positions in the industrial world should be open to 18} 8) 3) ©} Oo} Ij any man with the ability to fill them, regardless — of racial traits. 16} 7} 7| ©} 8] 9} Colored people should go to schools, hotels, thea- 3} 8} 5] 3] I] 5] tres, etc., patronized exclusively by colored people, thus preventing some inter-racial contacts. I} 22} 24) 0} I} 0} White people feel some antagonism toward negroes. o| 14) 15} I} o}| o 0} 10] 28} 9] 0} 0} White people feel.some antagonism toward Jap- O} 4) 23} 2} Oj} I] anese or Chinese. I} I} If} 25} 4} 5) Whitepeople feel some antagonism toward Indians. woe are 7) 16) (oO) 4 8| 16} 17| 3} 1) 2! Negroes emit an odor which is characteristic of 3} 8] 13} 2] 1} 3) their race, and which white people find disagree- able. 6) 20] 12} 3] 3] 3] Persons of pure Nordic stock are superior in O} Io; 12 2} 3] intelligence to negroes. 4; 15} 5} 11] 8} 4 ‘Pure blooded’’ members of a race are superior to I; 4; 9] 10} 3} 3} those of mixed stock. 2} 13) II} 9] 6] 7| Negroes are more prone to commit sexual offense 0} 5} 13} 9} O| 3] than is the average white man. 28) 14; 3} 2] ©} ©} Small children who have not been prejudiced by 13) 14} 2} 0} O} I) others, will play with children of other races, quite unconscious of racial differences. I7| 3) 2} 3} 13} 9] Persons who refuse to give to a negro the same 8} 4! 4! 5] 4] 4 social recognition which they would give to a white person of similar character, are unworthy to be called Christians. I} 12} 16] 12} o| 6] Japanese would be crafty and even treacherous, o| 2] 16] 11] o| 1| if necessary, to secure the dominance of their nation. o| 12} 22} 6] 2| 4] Racial difficulties are caused by the desire of the o| 4] 19] 5] o| 2| white people to gain and control the resources of the world. 7| 3] | 12} 14} 11| Persons of one race should be permitted, with 6| 7} 3] 6] 4| 4] mutual consent, to marry persons of another race. 5| 9] 7} 14} 1| 11| Colonies now held by white nations should very 3) 13/ 6] 5] o}| 3} soon become entirely independent. 42 The Measurement of Fair-Mindedness “No,” or 31.3 per cent. of the chances to register an ‘‘extreme”’ or ‘‘prejudiced”’ statement. Suppose now that these be classified s “‘liberal’’ meaning ‘‘all-men-are-of-one-blood”’ in tone, or ‘‘con- servative,’’ meaning ‘‘in-the-status-quo-we-are-on-top, and-it-would best-be-left-alone.”’ Then 58 per cent. of the prejudices were liberal, and 42 per cent. were conservative. At the end of the class, the average took 13.6 per cent. of the chances to say “‘All,’’ and 7.9 per cent. of the chances to say “‘No,”’ or 21.5 per cent. of the chances to register an extreme or prej- udiced statement. Of these, 70 per cent. were liberal and 30 per cent. were conservative. The class as a whole appeared to become more fair-minded, but seemed to drop conservative prejudices faster than liberal prejudices. This rests, however, on the assumption that the thirty persons who were in the class at the end, were a fair sample of the forty-seven present at the beginning. It may well be that part of the change was due to the dropping out of the more prejudiced, and more con- servatively prejudiced persons. However, all of the thirty were included in the forty-seven, and the instructor’s opinion was that the most prejudiced persons had preferred to oy in the group, rather than drop out of it. Experiments in the improvement of methods and curricula, and the measurement of educational ventures in other than religious or economic realms seem to offer immediate, constructive service. However, it may be that even greater possibilities will be found to lie in realms which seem to be less practical, and which may be purely scientific research. For example, it is possible by use of such tests to select individuals who are remarkably prejudiced, and others who are remarkably fair-minded. These individuals might then be carefully studied, with a view to finding the factors which have entered significantly into the experience of one type of individual, as contrasted with the significant factors in the experience of the other type. It seems probable that the application of the tech- niques of the social case-worker and of the psychoanalyst might discover very fruitful sources of prejudice and methods for its elimination. The reactions which individuals made to their own profiles sug- gests the need for another study, endeavoring to find out the ways in which individuals ‘‘interpret, excuse, rationalize’’ or, otherwise, defend their prejudices. How far can self-report be trusted, within Suggested Uses for the Tests 43 the realm where some praise or blame might be expected from others? How is insight into the real nature and origin of one’s prejudices best acquired? Again, a need is evidenced by this study, for a keener analysis of ‘‘facts,”’ “‘evidence,”’ and ‘‘arguments.’’ These techniques have long been held in high esteem as methods of influencing public opinion.!. However, these tests reveal that different people react with widely different attitudes to the same facts and arguments. Is there some method by which these prejudice factors can be re- duced to terms of least common denominators, so evidence can be ‘‘weighed”’ alike by people at different points of a prejudice scale? Will it be possible to ‘“‘try out”’ propaganda material upon groups selected because of different degrees and directions of prejudice, in such a way as to find out the reactions it produces, and to compare the effectiveness of different forms? Quite evidently there are other possibilities in the study of the relationship between prejudice and emotional instability, or other personality traits. Perhaps for certain types of work, for example, social research, the selection of personnel will be found to be facil- itated by such tests of fair-mindedness as have been developed here. It is probably true of this research, as of most of others, that it begins with one problem and ends in a maze of problems. The curious child reaches to pick up a twig, lying exposed on the bare. ground, but finds that it leads him on and on, for it is one small root of a great tree. yy 66 1Lippman, Walter, Public Opinion. ¢ Lone) is j ‘4, / vit } eae x, “HLIM LNHYWHEYDSY NI aoianfaud 76 10. BA Be r2) The Measurement of Fatr-Mindedness RATINGS OF JUDGES ON PROFILE V STUDENTS IN NORMAL SCHOOL COMPARED WITH MetuHopist MINISTERS IN THE SAME MID-WESTERN STATE Number of Judges Expecting} Judges Expecting Norma! School Students To Be |Ministers To Be PREJUDICE IN AGREEMENT WITH Number of Methodist Marked Differences More Prejudiced|More Prejudiced Radical on Economic Questions, Opposed to Capitalism Economic Liberal, Desiring Mild Reforms Capitalist, Favoring Economic Status Quo and Opposing Radi- cals Social Gospel, in Contrast with Individual Salvation, etc. Personal cism, etc. Communion, Mysti- Fundamentalism Christian Modernism, Liberalism | Religious Radical | Protestantism Roman Catholicism Strict, Puritanical Morals and Amusements Free, Loose Morals | a ae oe bo Ne) oa Ny O 77 Appendix os "9721S UII}SAM-PIJT OUILS dy} UT S19}ST -UIJT WIPOYWI JO Weg UO dIIPNfsig s10P SozeoIpUy JOOYDS [euLION ul syuepnys jo yeg uo soIpnfaig s10OJX sozeoIpuy AAM LLLLLLL LL. S[PIOJ[ BSOO7T ‘VI *ZI sjUsUIaSNUIy pure s[elOpy [eoUeWINg PUIG “It UISINIJOYJLD UBWIOY “OI WISIJUL}SI}IOIG °6 [CIpeY SNowsiypeyY °s WISI[VIOqr] ‘UISTUIspOyYY URIsHYD “2 UWIsT[eyUIWMIepUNY “9 ‘OJo ‘UISTOISATT ‘UOTUNTIUIOD [eUuOsIag °S ‘DJa ‘UOTPEATES [ENPIAIPU] YAM ySeIqWUOD ur ‘Jadsoy [epog “V7 s[eoipey suisoddg puv onG snjzejg dmOUODY Bup0ary ‘ysiyeyidey *€ SULIOJOY PIA Susaq ‘jesloqvy] Wwou0IW *z uistjeyided 0; pesoddg ‘suorj}senG Wuou0oy uo [eoIpeyY “1 tHLIM INAWATYDY NI adianlaud A SA UAOUd AOIGNLAUd 78 12: The Measurement of Fair-Mindedness RATINGS OF JUDGES ON PROFILE VI NORMAL SCHOOL STUDENTS IN A NEW JERSEY SCHOOL COMPARED WITH NORMAL SCHOOL STUDENTS IN A WISCONSIN SCHOOL PREJUDICE IN AGREEMENT WITH Radical on Economic Questions, Opposed to Capitalism Economic Liberal, Desiring Mild Reforms Capitalist Favoring Economic Status Quo and Opposing Radi- cals Social Gospel, in Contrast with Individual Salvation, etc. Personal Communion, Mysticism, etc. Fundamentalism Christian Modernism, Liberalism Religious Radical Protestantism Roman Catholicism Strict, Puritanical Morals and Amusements Free, Loose Morals Number of Judges Expecting Normal School | Judges Expecting Students in a New Jersey School To Be More Prejudiced Number of Normal School Students ina Wisconsin School To Be More Prejudiced One judge states that no difference can reasonably be expected. Marked Differences 79 A ppendix Sv Or SE O€ GZ 02 S\ JOOYIS UIsUODSIAA B UI syUNpNIS JOOYSS [VWION jo Weg UO ddIpPNfeig s10J SezeOIpuT JOoydS Avsiof MeN & UT syuapNiS [OOYIS [eULION JO Jeg UO ddIpPNfaig sIOWF sozeoIpuy Aa WHE ELIE Sj) weer ol S 1 @) VLLLLLLLLL. er a S[RIOW BOOT ‘Vo “ZI TOW SJUSUIOSNUIY pue sTeIOWy [POUL WING WINS ‘II UWISIO[OYJLD ULUIOY ‘or UWISIJUCSIIOIG °*6 [ESIPeY SNOIBIpOY °g WSITe1oqIy ‘uIstuIapoyy ueTsIIYyD *L wWISI[vJUsMIepuNY “9 "Oye ‘UISIOISAYPY ‘UOTUNUIMIOD [euOosIag °*S ‘Je ‘UOTJEATLS [ENPIAIPU] YIM YseIWUOD UT ‘Jedsoy [eID0G “Fy s[eoIpey sulsoddO pue on} snjze7g dwouo0og Surioaey ‘ysyeydeo °€ SULIOJOY PIAL SULISeg ‘[eloqry SIuIOUCNW *z usl]eydeD 0} pesoddg ‘suorjsenG dtuou0sq uO [eoIpey, *1 ‘HLIM INAWARADY NI aD1anlaug IA SH U4OUd AOIGN[Aud 80 12. The Measurement of Fatr-Mindedness RATINGS OF JUDGES ON PROFILE VII THEOLOGICAL STUDENTS IN UNION THEOLOGICAL SEMINARY COMPARED WITH STUDENTS IN CHRISTIAN ETHICS AT YALE DIVINITY SCHOOL Number of Judges Expecting Number of PREJUDICE Union Theo- |Judges Expecting Marked IN AGREEMENT logical Seminary} Yale Divinity Differences WITH Students To Be} School Students More Be M eee la ile REPS Prejudiced Prejudiced Radical on Economic Questions, Opposed to Capitalism 7 I X Economic Liberal, Desiring Mild! Reforms 4 2 Capitalist, Favoring Economic Status Quo and Opposing Radicals 2 6 Social Gospel, in Contrast with Individual Salvation, etc. 4 2 Personal Communion, Mysticism, etc; 2 4 Fundamentalism I 7 x Christian Modernism, Liberalism 4 2 Religious Radical 4 8 Protestantism 2 2 Roman Catholicism O O Strict Puritanical Morals and Amusements oO 8 xX Free, Loose Morals 4 o) Two judges state merely that the Yale students may be expected to be more prejudiced in general on such issues. SI A ppendix GE o¢ S¢ o¢ S[PIOJT DSOO'T ‘99IY ‘ZI SJUSWIOSNUIY pure speIoyY [BOUL WING PWIS “1! uISIOOYyyeD UeUIOY ‘Or WISIJUL}SI}OIG *O [BOIpeY SNOWY “8 UISITeIBqIT ‘UIsSTUIepoyY uelysuyD °L usIjeyuowepuny “9 ‘Oyo ‘WISIOTISAP ‘UOTUNTIUIOD [eUOSI9g “S$ ‘DOJO ‘UOTPEATLS [ENPTAIPU] YUM 4SeIWWOD Ul ‘jadsoyy [BINOG “Vv STeoipeyY sutsoddg pure onG snjze4ys owsouoo 7 Sursoary ‘sysipeyidey °c SUIIOJOY Pl SUISeg ‘[eloqI] DIWIOUODy = *z uistjeyidesd 0} pesoddg suorjsenG dwW0uo oy UO [eoIpeyY “1 *-HLIM INUNEEUDY Ni aoianlfaug Si ol S i@] JOOYdS AWUIAIC FPA ye sory” uensiuy? ur syuepnys jo eg UO sdIPNfsig sIOW sozeoIpuUyT os LLLLLLL LL. AICUIUIAS [BIIZOT[OOY T, UOTUA UT syuapNis [eorsojooy, jo weg UO sdIpNfsIg IIOP, sezeoIpUy ATM IIA SHUAOUd ADIGN[ANd The Measurement of Fair-Mindedness Os Sv Or S€ of G2 02 S| pepuryy-uedg yO, og OL peP,OA SUOSI9g JO We Uo sdIpnfsaig s1OP sozeoIpuy AICUIUIDS [CIIBO]OIy, UOT JO szuep “NS AIBUIPIO JO JAeG Uo sIIPNfeig s10P sdzeOIpUT AAS] ol ww) ° S[ePIOY VSOO7T ‘daly ‘ZI SjUsWIASNUTY pure spe1Opy [eUe WING PIS ‘II WISTITOYID ueWOY ‘or UWISI]Ue}S9IOIg *6 [eoIpey snorsypoy *8 WISIUIOpOopPL UeTYSIIYD *L wisI[eyUsWIepuN *9 ‘O}o ‘UISIOISAYY ‘UOIUNUIMIOD [euOCsIag °S UOIZLATLS [ENPIAIPU] YIM 4sesqUOD UT ‘Jadsoy [eIO0S “FV sjeoipey sursoddg pure ono snjzejg dwIOUONY Suri0aey ‘ysyeydeg -f SULIOJOY PII Suisiseq ‘yeroqry] Wuouosy *z ustjeyided 0} pasoddg ‘suorjsenG s1u0u0cg uo yeorpey “I ‘HLIM INHWaARADY NI aolanlaag IIA SAMAOUNd AOIGN[ANd A ppendix UOISI[OY JO WIOY AuY ysuUIesYy Alo AJ DISC ‘UOIWIWSIEdNsS yulesy S[EOIPeY SNOB yoy yuresy S}SITeJUSMIepUNY SNosyoy yuresy S[ROIPeY DIWIOUODY ysulesy ond snyzejyS DIuIOUON YsuIeEsYy S[eIOJW BSO007T ‘d9a1J SJUSUIBSNUIY puUe SsTeIOPY [eoUe WING P~LIS WISIDIJOYJLD ULWIOY UISTJUL}S9}OIg TeSIPeY SNorsipoy UISI[VIVQI'T ‘WISTUIOpOPY UeIWSIIYS WS][ejUIMIepUN ‘O}0 ‘UISIOISAJY ‘UOTUNUIWIOD [eUOsIOg ‘Jo ‘UOTJLATLS [ENPIAIPU]T YIM YSeIZUOD ut ‘fadsoy [eII0S sjecipey suisoddg pure onG snjze}g s1mM0UODY SuLIOAey ‘ysiTeyIdeg SUIIOJOY PII SuLIseq ‘[e1oqry] DtwMOU0NW wstjeyded 0} pesoddgo ‘suorjsenG Iuou0ly uO [eoIpey ‘HLIM INSWaaADY NI aoIanlaug o€ O02 SI sUsIaJUOD JO puy soIpnfo1g soy LZZZZZZA aousIajJuOD Jo Suruuriseg voIpnfsaig s10Jy Rites AWY HONFAAINOD “VY ‘OD 'M ‘A INTGALS VAUNA AAIVI—D ‘_ ‘VW SWAOY XI SATIAOUd AOIGN[AUd "QI *LI ‘91 rae ‘VI “er "oI 7e1 ‘OI The Measurement of Fair-Mindedness 84 UOISI[PY JO WAOY Auy ysuIesYy “gl AID SAP ‘OIdePy ‘UoTWIsIedNs ysuresy “LI S[COIPeY SNOBoYy Juresy “OI SjS][VJUSWMIePUNY SNOB yoy ysuresy ‘SI S[ROIPeY Dwo0uoI, ysuresy ‘Vv Lal ono snyzejig DwWOUO NY ysuresy *f Lael Lal S[PIOJP VSOOT ‘991 *% SjuNWESNUIY pue speIO| [eoUeWINgG PIS ‘I e WISINI[OYID UPUIOY ‘Oo Lal UWISIJULISI}IOIG *6 [POPE SNOBpY *g WISTTRIVqIT ‘WISTUIapOoyy UeIIsIyD *L UIST[eJUSUIepUNY “oO ‘Oye ‘WISIOIISAPT ‘UOTUNUIMIOD yeUuOosIag *S ‘OJo ‘UOIZATLS [ENPIAIpU] YAM ysvIyUOD ut ‘Jadsoy [eII0Sg “PV s[eoipey sursoddg pue onG snjzejs ormouoog Sur0aAey ‘ys1jeydeg °*f SUMIOJOY PII SuUltIsed ‘[e1eqiy oIwmOU0DW *z tusijeyided 0} pasoddo ‘suonsenO ormou0og uo yeoIpey ‘I SHLIIM INGNEaGYOV NI goIanlaug os SY Or SE oe SZ o7 SI Ot G fe) duUsIaJUOD JO pug sIpnfoig s1OP Sozeorpuy VASLLLLA, Z VIUIIOJUOD JO SuluUIseg VsoIpNfoig s1OJ So}yeOIpuy foe AY AONAUTINOD “VY “‘D 'M ‘A INFGALS VAINAD aIvI—YJ ‘A ‘q SWAoOy X SATIAOUd ADIGN[ANd STATISTICAL TABLES TABLE I NorMS FOR Gross SCORE—ARRANGED BY GROUPS Group Number Mean S. D. Normal School Students 102 32.0% 8.9 Persons Selected as Fair-Minded 13 13.0 5.8 Persons Selected as Very Prejudiced 5 61.4 8.8 Methodist Ministers (In One Mid- Western State) 35 36.8 12.1 Students in Union Theological Semi- nary 46 ani2 8.5 Students in Yale Divinity School ri 26.4 9.0 Miscellaneous College Graduates 39 26.6 12:7 ToTAL GRouP ai 29:1% 153% TABLE II FREQUENCY DISTRIBUTION OF GROSS SCORE 311 CASES Gross Score | Number in Per Cent O— 4.9 2 Sanna 3 I0O—14.9 22 I5—I19.9 Sie) 20—24.9 54 25—29.9 ot 30—34-9 mp 3539-9 Do 40—44.9 26 4549.9 Il 50—54.9 8 Shmme os, 2 60—64.9 I 65—69.9 3 70—74.9 | I 86 The Measurement of Fair-Mindedness TABLE III RATING OF ARGUMENTS—TEST E-10 Note: The roman numerals in the first column on the left correspond to the questions in the test. Beneath each, the arabic numerals are listed, one for each argument. The second column indicates the side of the question supported by the argument. The third column indicates the rating strong or weak, which is presumed. The fourth column lists in its two subdivisions the vote of the twelve judges selected by the author. The fifth column in similar fashion lists the ratings given by the twelve persons who were most fair-minded, as judged by the rest of the test. In the sixth column the per cent of the twenty-four judges who agreed with the test rating, is given. ORIGINAL JUDGES F ants ae! PER CENT eS Oe a eae ereesonee 2 8 ag ARGUMENT | SIDE OF TEST AGREEING NUMBER ISSUE RATING Strong Weak Strong Weak /witH TEST ifs | | : Yes Strong 12 oO 9 3 88% 2) No Weak I II 3 9 83 ep DN Strong I2 fe) 12 O 100 4. Yes Weak O I2 2 10 92 5. No Weak I II I II 92 6. Yes Weak O 8 Tie II 95 tl; re No Strong II I I2 re) 96% 2) No Weak oO 12 I II 96 a: Yes Weak fe) 12 () 12 100 a Yes Strong 10 2 II I 88 i No Weak I II a 9 83 6. Yes Weak O I2 I II Ed EEN ce Aearcie eee | meen a Til 1: Yes Weak e) I2 fe) 12 100% ep Yes Strong 9 3 ? 5 67 2: No Strong I2 Oo II I 96 4. Yes Strong 9 3 10 2 79 5: No Strong II I II I 92 6. No Weak 3 | 9 x 9 75 IV. i Yes Strong IO 2 7 5 71% 2: Yes Weak 2 ‘IO 6 6 67 a No Weak fe) 12 = 10 92 4. Yes Strong I2 ) 12 oO 100 5. No Weak 2 10 4 8 75 6. No | Strong 12 e) II I 96 A ppendix 87 TABLE III (Continued) Most ORIGINAL JUDGES FArr-MINDED PER CENT is fel eee eee, ee Se OR = BOTH, ARGUMENT | SIDE OF TEST AGREEING NUMBER ISSUE RATING Strong Weak Strong Weak |witH TEST | V. | 1B Yes Weak I II 6 6 71% 2. No Weak I II I II 92 3. No Strong I2 oO 10 2 92 4. Yes Strong IO 2 II I 88 re Yes Strong 9 3 II I 83 6. No Weak oO I2 4 8 93 aN a Yes Weak I II 4 8 79% 2. No Strong II I 7 5 ve 8. Yes Strong 12 9) I2 fe) 100 4. Yes Strong 9 3 7 5 67 5. No Weak fe) I2 4 8 83 6. No Weak e) I2 2 10 g2 ok eave ee PR OG SS SS Te Laie As cst DL Dk tes GER Se ik eran ES Vil. re No Weak I II 5 7 75% 2, No Strong 10 2 6 6 67 3. Yes Strong II I 9 a 83 4. Yes Strong I2 oO I2 fo) 100 5. Yes Weak O I2 3 9 88 6. No Weak O I2 I II 96 VIII. I3 Yes Weak oO I2 ) I2 100% ZI Yes Strong Ig O I2 O 100 as No Strong 10 2 7 5 71 4. No Weak I 1B I II 92 i Yes Weak O I2 I II 96 6. No Strong II I 10 2 88 IX. Tee a 1s Yes Weak I II 4 8 79% 2 | No Weak oO I2 oO I2 100 ai No Strong I2 (o) IO 2 92 4. Yes Strong 10 acid D IO 2 83 5. NE Strong II Beoed 9 a 83 6. o | Strong 9 3 | 8 4 71 | 88 The Measurement of Fair-Mindedness TABLE III (Continued) ORIGINAL JUDGES | F — Maes PER CENT NIE AS SUIS AL ey Mh 5 OORT HG og ARGUMENT SIDE OF TEST AGREEING NUMBER ISSUE RATING Strong Weak Strong Weak WITH TEST X. Tr: Yes Weak O I2 I 11 96% eg Yes Strong IO 2 8 4 75 2 No Weak oO I2 I II 96 4. No Strong 9 3 I2 oO 88 5. Yes Strong 10 2 12 O g2 6. No Strong 12 O I2 e) 100 | XI. '¢ Yes Weak I II 6 6 71% Bs No Strong IO 2 7 5 71 TOTAL TEST (311 Cases) 133, CASES MEAN 28.9 40.9 27.1 44.3 37.6 16.7 29.1 TABLE V SELF CORRELATION OF GROSS SCORE 117 CASES TEST FORM tmvdvawp ToTaL TEST (161 Cases) SELF CORRELATION .QI .9O 85 .67 .66 .78 .96 Lied 90 The Measurement of Fair-Mindedness TABLE VI SELF CORRELATION OF THE EIGHTEEN ANALYTICAL LINES oF BIAS Probable Self-r Direction Self-r Self-r Maximum Number|on 200 Such Points of Bias 30 Papers 61 Papers of Points (70 Items) I .74 .69 189 .70 I] .39 51 93 -69 III .68 .59 1O2 .60 IV 84 .69 188 76 V .82 84 143 .88 VI .94 79 182 81 VII 81 .60 89 Me ty VIII .97 .80 164 83 IX 64 35 58 65 xX 51 52 66 77 XI .78 .69 107 81 XII .70 .46 94 65 XIII 75 .67 188 .68 XIV 54 .49 205 .49 XV .40 58 I51 65 XVI .68 84 I5I 88 XVII .46 58 86 .76 XVIII 85 .78 182 .80 Note: Only the first twelve are now used in analytical scoring. TABLE VII CORRELATIONS BETWEEN GROSS SCORES FOR EACH TEST FORM WITH Gross SCORES FOR TOTAL TEST TrEst FORM CORRELATION AMoOAW D> Ap pendix 91 TABLE VIII CORRELATIONS BETWEEN Gross Scores FoR Eacu Test Form WITH Gross SCORES ON REMAINING ForMS TEST FORM CORRELATION A TY B 42 C 30 D 0 E .36 F 21 LAD EE LX CORRELATIONS BETWEEN DIAGNOosIS GIVEN BY EACH FoRM OF TEST WITH DIAGNOsIS GIVEN BY TOTAL TEST | TEST FORM 119 CASES 40 CASES APPROXIMATE P METHOD r COMPOSITE A 54 .63 58 B .60 .86 173 & 53 .67 .60 D 58 .56 .57 E 61 .49 “55 F .48 42 45 TABLE X BALLOT FOR Most FAIrR-MINDED PERSON Faculty No. 1 (About 100 Persons) NUMBER OF VOTES PERSONS RECEIVED I Dr I 7 I 6 2 4 3 3 5 2 2I I LOTAI, 34 72 (Cast by 39 voters) 92 The Measurement of Fair-Mindedness TABLE X (Continued) BALLOT FOR Most FAtir-MINDED PERSON Faculty Vote on Faculty No. 2 (About 22 Persons) NUMBER OF VOTES PERSONS RECEIVED : | : I 6 I 3 3 24 i I TOTAL IA | 36 (Cast by 18 voters) Students Vote on Faculty No. 2 (About 22 Persons)* NUMBER OF VOTES PERSONS RECEIVED I 33 I 12 I 7 I 6 I 5 I 8 5 2 6 I TOTAL ite | 82 (Cast by 44 voters) Student Vote on Student Body No. 1 (About 120 Persons) NUMBER OF VOTES PERSONS RECEIVED I II I 7 I 6 - I 3 3 © 9 2 21 I TOTAL a7 77 (Cast by 42 voters) *Note: The first four selected by this vote included the same persons as were among the first four selected by the faculty vote. There was one difference in order. A ppendix 93 TABLE XI AGREEMENT BETWEEN PREDICTIONS MADE BY OTHER PERSONS, AND PREJUDICES REVEALED IN TEsT PROFILES PROFILES NUMBERS I-24 Profile Number of Predictions Predictions Predictions Number Predictions Clearly Fulfilled Doubtful Not Fulfilled I 3 3 2 2 2 3 4 3 I 4 4 4 5 2 I I 6 I I 7 2 2 8 2 2 9 I I Io 2 2 II 3 2 I I2 3 3 13 3 3 14 os 2 I 15 I I 16 2 I I 17 3 3 18 2 I I 19 4 2 2 20 I I 21 4 4 22 2 I I 23 ] 3] 24 2 I I | — fF ToraLs 59 43 10 6 94 The Measurement of Fair-Mindedness TABLE XII REACTIONS OF INDIVIDUAL TEST SUBJECTS TO THEIR PREJUDICE PROFILES AS REVEALED BY THE TEST* Profile Number of Number Questioned Number Reactions Approved Number or Rejected 25. 26. a9 28. 29. Oo On oO N OO me N \O =a Ro NH = WD = _ iS) OOVUOUOUOUOUONOUOODOOOOOOOUOUOUUOOUOUUUOUOUOUOUOUOMO NYO SCSAUN ADO ADANNO WONN DWAWO DO WO OM 4 iS) iS) ° on Ny TOTALS 272 *Note: A “blanket’’ approval or disapproval, is interpreted here as equivalent to approval or disapproval of nine of the eighteen items. A ppendix 95 TABLE XIII AGREEMENT ON GROUP DIFFERENCES AS REVEALED BY TESTS WITH GROUP DIFFERENCES PREDICTED BY JUDGus NUMBER NUMBER THESE NUMBER GROuPS MARKED ““MARKED”’ | DIFFERENCES COMPARED DIFFERENCES|DIFFERENCES| EXISTING PREDICTED SHOWN BY But Not BY JUDGES TEST **MARKED”’ Catholic— Protestant 5 4 oO Normal School Students— Methodist Ministers 5 5 Oo Ministers Over 40 Years of Age— Ministers Under 40 Years of Age 4 5 I Methodist Ministers— Students, Union Theological Seminary 6 5 I Ministers, Theological Seminary— Ministers with only High School Prepara- tion 4 3 O New Jersey Students— Wisconsin Students 2 2 O Yale Divinity Students— Students Union Theological Seminary 3 2 I TOTALS 32 26 3 96 The Measurement of Fatr-Mindedness TABLE XIV AGREEMENT OF JUDGES WiTH OUTSTANDING DIFFERENCES \BRTWEEN GRouPS AS REVEALED BY TEsTS NUMBER JUDGES NUMBER JUDGES EXPECTING GRoups DIFFERENCES AS | DIFFERENCES IN COMPARED SHOWN REVERSE DIRECTION Catholic— Protestant 42 3 Normal School Students— Methodist Ministers 37 I Ministers Over 40 Years of Age— Ministers Under 40 Years of Age 31 8 Methodist Ministers— Students: Union Theological Seminary 34 re) Ministers: Theological Seminary— Ministers with only High School Preparation 29 I New Jersey Students— Wisconsin Students 25 6 Yale Divinity Students— Union Theological Seminary Students 23 2 TOTALS . 227 22 A ppendix 97 TABLE XV SUMMARY OF PREJUDICE SCORES AT BEGINNING AND END OF CONFERENCE OF AMERICAN COUNTRY LIFE ASSOCIATION AVERAGE PER CENT|AVERAGE PER CENT BASIS OF Scone peebete heel tue een Complete Test: Gross Score 34% 34% Form A (Cross Out Test) 25 24 Form B (Degree of Truth Test) 42 44 Form E (Arguments Test) 13 17 Analyzed Score in Agreement with: 1. Orthodox, Fundamentalist Religion 8 7 2. Radical Religious Ideas 13 I2 3. Economic Capitalism, Canservatism 7 8 4. Radical and Liberal Economic Ideas 17 16 5. Backward, Conservative, Agricultural and Community Outlook 5 5 6. Liberal, Progressive Agricultural and Community Outlook ie. 12 The differences between conservative points of view which range from 5 per cent to 8 per cent, and progressive, liberal points of view which range from I2 per cent to 17 per cent are probably significant as indicating the ‘‘tone’”’ of the conference. None of the changes between beginning and end are large enough to have any clear significance. NY sa Nang es me ah Lhe ‘all ee Ge6. 7 , : / * + id « . how) q ’ ua? by i VF A , “A . by we or . ’ r j . . iy = Un - iG " ~ j ’ * % J , uy U s ‘ i ie Bore, “~ ’ ose