Chapter 6: The Measurement of Personality
Floyd Henry Allport
Introductory Statement. With the recent advance in knowledge of the dynamic principles of human nature there has developed a need for evaluating the traits of individuals in some definite form. From the point of view of both technique and application the methods of personality measurement are important for the social psychologist. Although their stage of development is still crude and tentative, it is worth while briefly to outline these methods. They may be divided into two general classes: (1) those dependent upon the estimates given by associates of the individual studied, and (2) special methods of objective testing. The first class is subdivided into (a) systematic questionnaire methods, and (b) rating methods.
1. Methods of Judgment by Associates. A. Systematic Questionnaire Methods. This type of method is merely an attempt to canvass the personality by questions concerning the characteristic reactions in various fields. Such questions should avoid generalities and seek detailed facts from which to draw conclusions. The following are examples: Is the individual talkative or taciturn? How punctual and how thorough is he? How many times have you seen him angry? For what causes? Upon what subjects is he 'touchy?' Is his sex and family life normal and happy? Does he display aversion, affected indifference, or repressed emotion in regard to sex topics? Does he seek or shun society? How readily is he browbeaten by trades-people and officials? How frequently does he dominate them? What defects has he that he is willing to acknowledge? Does he blame others as a rule for his own failures? Are his sports, recreations, and aesthetic or religious interests a proper balance for his vocational pursuits? Does he adapt himself readily to the moral standards of society? Various fields of adjustment — economic, social, sex and family, recreational, and moral — are covered, and inferences concerning specific traits drawn from the answers.
Complete and useful questionnaires and trait lists have been developed for clinical and other purposes and rendered quantitative by the use of pluses, minuses, and marks denoting extreme degrees of the behavior in question (Wells, Spaulding), so that the assets, liabilities, and compensations of the individual in the various spheres may be evaluated. The questions being suitably worded, a simple count of the `Yes' and 'No' replies may be rendered diagnostic in value. (Woolworth.) Fields of inquiry, with quantitative evaluation, have been adopted for studying the personalities of mental defectives. (W. E. Fernald, Porteus.)
The questionnaire method may be made available for self-study, provided (1) the questions relate to actual overt acts, not feelings, motives, or intentions, and (2) the individual can adopt an objective attitude toward himself and one free from rationalization. The basis of the replies must be what one does, not what traits one believes himself to have.
B. Rating Methods. Although questionnaire methods are useful for general impressions, for quantitative and comparative treatment it is preferable to rate the individual in regard to certain selected traits. For accurate estimates it is essential to have a number of judges or raters because of the complex variables entering into the judgment of one person by another. An average of the ratings of from five to ten competent judges is to be desired, though three separate ratings are far more satisfactory than one. There are two varieties of rating method, which we may term 'scoring' and 'ranking. The scoring method is used where a single individual is to be rated, or where, if ratings of a group are desired, each rater knows only a part of the individuals. A subjective scale is imagined, such as from 1 to 5, or a percentile basis, upon which to express the degree of each trait which the rater estimates the subject to possess. Since the only standard available is that afforded by a representative group of the subject's associates, the extremes and median of such a group are generally imagined in giving the score. The army rating method for commissioned personnel attempted to render the scale more stable and definite by asking the
(128) raters to call to mind actual officers who in their opinion stood at certain points along it, and then proceed to fit the subject to the `personified' scale. The ranking method, available only in rating an actual group all of whom are known to each rater, is more accurate. Here the subjects to be rated are arranged by each judge in their order of rank in respect to each trait, the one possessing the highest degree of the trait standing first, and the one with the lowest degree at the bottom. The rating of each subject is then expressed as his rank in the group. The standard here is truly objective (being the group itself) and identical for all judges. Comparisons are more concrete and definite than in the scoring method.
The question of the reliability of the judgments of associates in respect to different traits, and in general, has been investigated in a number of experiments. The usual assumption is that the closer the agreement (that is, the less the variation) of the judges upon a given trait, the greater is the probability that the average of their ratings is an accurate measure of that trait. Variation is the necessary result of personal bias away from the truth; hence uniformity must mean that the truth is attained. A simple measure of the degree of variability is found in the average deviation (A.D.) This is found by (1) taking the average of all the ratings given an individual in a certain trait by the group of raters; (2) finding the deviation of each rater from this average (that is, the difference between his rating and the average, regardless of sign) ; and (3) taking the average of these deviations of the several raters. This latter average is called the A.D. If it is large, the variability is great (that is, there is little uniformity of opinion); if it is small, the variability is slight. Averaging the A.D.'s of a certain trait for all subjects gives a final index of the reliability of judgment in that trait.
In a certain group of 25 subjects the A.D. of the various traits was found to range between 3 and 6 ranks (ranking method used) out of a possible 25. (Hollingworth.) This is a sufficiently small variability to allow us to attach at least a rough value to the average rating by associates. In regard to particular traits, the lowest A.D. (highest reliability) was found for vulgarity, intelligence, beauty, and conceit; the highest A.D. (least reliability) for snobbishness and refinement. (Hollingworth.) The rule is that qualities for which there is objective evidence— for example, intelligence and vulgarity— are most reliably rated by the judges; while inner attitudes and feelings — for example, snobbishness — are less accurately evaluated. Similarly the socially manifest traits of quickness, originality, efficiency, ascendance-submission, expansion-reclusion, and social adaptability are reliably judged (low A.D.). The more subjective aspects, such as introversion, emotionality, disposition, and characterial traits, are judged with much less dependability (high A.D.). (Norsworthy, Cattell, G. W. Allport.)
Another question of practical bearing concerns the reliability of one's rating of one's self in comparison with one's rating by others. As a rule the deviation of the self-rating from the average of ratings by a group of judges is greater than the average deviation among the judges themselves. In other words self-rating is not so accurate as rating by others. (G. W. Allport, Hollingworth.) Low insight and rationalization distort self-estimates. Thus Professor Hollingworth found that subjects rate themselves too high in socially desirable traits, such as sociability, refinement, and humor, and too low in undesirable traits, as vulgarity, conceit, and snobbishness.
Insight and self-evaluation are readily measurable by finding the difference between the self-ranking and the ranking given by others in the various traits, and prefixing a + or - sign to denote over— or under— self-evaluation. Repeated experiments in which rank in scores of an intelligence test were compared with self-rankings in intelligence previously taken yield the following interesting results. The most intelligent underrate their ability, while the least intelligent overrate their ability. The greater the superiority in intelli-
(130) -gence the greater is the degree of under-self-estimation; the greater the inferiority the higher the overestimation. The more intelligent half as a whole have better insight than the less intelligent, for they underrate themselves less than the inferior ones overrate themselves. (F. H. and G. W. Allport.) The tendency toward under-self-evaluation by superior individuals and over-self-evaluation by the inferior may be due in part to the influence of uncertainty, in causing the self -rater to incline toward an average mark. But it is equally probable that factors of insight, rationalization by the inferior of their failures, and the like play a true part. Hollingworth found similar tendencies in that those subjects who stood high in neatness, intelligence, humor, and refinement were better judges of these traits in themselves and others than were those lacking in these qualities. Individuals showing vulgarity, snobbishness, and conceit were poorer judges of these traits in self and others. This result can be interpreted to show the close relation of insight and drive in the improvement of personality. Recognition of a certain trait as a personal ideal leads to its acquisition. One acquires the characteristic because he is a good judge of it and of himself. On the other hand, in the lack of recognition of an undesirable trait in one's self no drive is developed to eradicate it. The rating methods are capable of considerable refinement and utility if scientifically developed. The main requirements for accurate rating are the following: (1) the selection of traits that are genuine, fundamental, and distinct; (2) a sufficient number of raters, preferably individuals who see the subject from various viewpoints; (3) a thorough knowledge and mutual agreement among the raters as to the exact meaning of the various traits; (4) a sufficiently extended acquaintance with the subjects, an acquaintance during which the rater has the scale of traits in mind; (5) basing of the ratings upon actual facts of behavior, not general impressions; (6) practice in the use of the scale; (7) avoidance of the tendency to allow a good or bad opinion of the subject in one trait to bias one's judgment in regard to another trait.
2. Testing Methods. Rating, though fundamental, is a time-consuming and somewhat cumbersome method of personality
(131) study. In the fields of education, mental hygiene, vocational direction, personnel management, and social work there is urgent need for simple and practicable tests of special traits. In the sphere of intelligence and special capacities much progress has been achieved. This work lies outside the scope of the present discussion. Within the other fields of personality measurement is a far more complex and subtle problem. Two phases of technique are necessary for developing such measures. First, each special test must be standardized; that is, one must know from experience with it what scores in the trait in question to assign to various types of performance in the test. Secondly, results obtained by its use should be verified by comparison with some other criterion (as objective a criterion as possible) of personality. One method of verification is to plot graphs representing the individual's personality on ordinates standing for the values achieved in tests of the several traits. Persons who know a group tested then examine such `profiles' and try to identify the individuals to whom they belong. The most satisfactory method, however, is the comparison of the rank order of the subjects as determined by the scores they obtained in the test of a trait (highest first, etc.) with the average rank order obtained from a number of ratings of the subjects upon that trait. If the two rank orders are identical— that is, if the same person occupies the first place in each, another person the second place in each, etc., or if the lists are closely similar — it is justifiable to infer that the test used is a fair measure of the trait as socially established by rating. This is known as the method of correlation. If the two rankings are identical there is said to be perfect positive correlation, and the 'coefficient of correlation' (r) is +1.00. If there is no similarity at all in the rankings, the coefficient of correlation is said to be zero. Positive coefficients range between 0 and + 1.00 and are determinable by formulae. Below .40 not much significance is attached to r. Coefficients ranging between .40 and .60 are suggestive of correlation, with other factors entering to disturb the perfect agreement of the ranks. A coefficient above .60 is fairly convincing, although for exact purposes one
(132) of .75 or .80 is needed. In case one rank order is reversed, instead of identical, in relation to the other rank order — that is, if the first in one list is last in the other, and so on — there is said to be perfect negative (or inverse) correlation (r = - 1.00). Negative coefficients, ranging from 0 to - 1.00, indicate with increasing certainty that an individual low in one of the two correlated measurements may be predicted to be correspondingly high in the other.
An important testing scale of traits centering chiefly in motility (`will') has been devised by Professor June Downey. The handwriting reaction is used as a basis of exhibiting traits. Speed is measured by the rapidity of the writing movement; freedom from inertia (hyperkinesis) by the difference between the customary and the maximum speed of writing; flexibility and care for detail by success at disguise of writing and imitation of a model; assurance and resistance to opposition (ascendance) by resisting verbal suggestion and by assertiveness shown in writing with eyes closed when an obstruction is suddenly held in the way of the pen. Coordination is tested by the ability to Write a long phrase rapidly in a very restricted space; and inhibition, in the sense of control and tenacity, by the extent to which writing may be retarded while still keeping the pen moving, a task disagreeable to explosive individuals. One of the most significant tests is that of impulsion. If the individual of high impulsion writes his name with eyes closed or while counting, the writing is likely to be hastened and increased in size. With an inhibited individual handwriting under these conditions is diminished in size and retarded. The fundamental movement trends thus reveal themselves when the normal conscious control of the cortex is blocked through distraction.
Graphs or profiles plotted on the basis of these tests indicate three general patterns of `will' traits: (1) the willful and aggressive type, (2) the slow, accurate, and tenacious type, and (3) the explosive or 'hair-trigger' type. The significance of the traits measured for life adjustments is shown in graphs of successful persons who compensate for mediocre ability by a high register in will
(133) traits. High scores in these tests are achieved by leaders and eminent men, hence the social significance of the test scale. The limitation of the Downey scale is that it leaves the important sphere of self-expression almost, and the sphere of temperament entirely, untouched. The question also arises whether from simple writing movements one can draw conclusions which shall apply to personality traits in daily life. The author of the tests has clone some verification by profile identification, but further proof is needed.
A simple but suggestive test of capacity for achievement is conducted by using Dr. G. G. Fernald's instrument for measuring the subject's persistence in remaining with his heels raised off the floor as long as he is able (that is, as long as he can 'will' to do so). A low score is due rather to unwillingness to stand discomfort and monotony than to actual fatigue. A group of normal high-school students averaged three times as long a period of this ordeal as a group of prisoners at a reformatory.
One of the few attempts to investigate the traits of temperament has been made by Dr. S. L. Pressey. The subject is asked to cross out of separate lists words which denote things unpleasant to him, or about which he has worried, or which he considers immoral. By the number and quality of words crossed one seeks to determine the 'emotional spread' (breadth) and other aspects. Data collected under the present writer's direction indicate that a test of this sort is equivocal because the introverted type reacts to it, not by crossing out, but by ignoring the words which are crucial in their emotional lives. This is a defense reaction against the invasion of complexes. The same phenomenon invalidates any attempt to gauge emotions through reactions to stimuli which condition them, and renders the problem of emotional testing extremely difficult. Emotional attitude and general outlook on life have been tested by giving the subject partial sentences, for each of which cards bearing Several possible completions are offered. Some of the completions are humorous in tone, some are serious, conventional, cynical, etc. The emotional attitude is shown by the constant choice of a certain type of completion. An example of one of the sentences is: "A man who lives a pure life" - [completions] (a) "will miss a lot of fun"
(134) (cynical, humorous), (b) "will gain the respect of all" (conventional), (c) "will be cheated by rogues" (pessimistic), etc. (Myerson. )
The most symptomatic tests employed for :introversion are in connection with the free word association method described on page 116. Jung and others have distinguished two broad types of response words. First, there are those of an objective, non-emotional sort (for example, the response `barn' given to the stimulus word 'house,' 'night' given as a response to 'day,' and the like), which are characteristic of the extroverted personality. Secondly, there are those of an ego-centric type common to introverts. Two sub-classes of the latter are the complex type and the predicate type, both described on page 116. Examples of associations revealing emotional complexes are `father' -'anxious'; 'hair' 'falling out'; 'love'- 'Donald. The predicate type may be illustrated by such personal associations as 'water' - 'glorious'; `ride' -'dangerous. Responses of the ego-centric type, together with increase in reaction time, hesitation, repetition of the stimulus word or of reaction words (perseveration), giving of superficial words, rhymes, etc. (especially with increase of reaction time), confusion, correction of reaction word, stilted and nonsensical reaction words, are all indicators of the repressions of the introvert. A list of one hundred words was given to one thousand normal persons and a count made of the frequency of occurrence of the various response words given. By reference to this frequency table it is possible to use this list on any subject and to ascertain the degree of 'community' of his responses (that is, how much the words he associates tend to be like or differ from those associated by the majority of people). (Kent and Rosanoff.) Individuals giving many predicate reactions (introverts) have a low index of community in the words they associate. (Wells.) Low community index has also been found to have a high correlation with the degree of introversion a— , determined by rating. (G. W. Allport.)
The important trait of ascendance is difficult to measure, because to evoke it necessitates an actual personal contact. The writer, in collaboration with Dr. G. W. Allport, has attempted to develop an "Active-Passive Reaction Study" upon the principle of imagined
(135) or represented situations involving face-to-face social contacts. These situations, drawn as closely as possible from life, are presented in print, and the subject is asked to state the nature of his reaction as it would be if the situation were actual. A sample of one of the situations is as follows:
You desire to board a boat or train to see a friend off. You feel it is important to do this; and the guard forbids you on obviously unnecessary technicalities. Do you obey silently, argue, or bluff your way past?
The first correlation of ranks based upon scores in this test with ratings on the trait of ascendance-submission gave a coefficient of .40. After improvement and standardization, as high a correlation as .80 was obtained. In spite of the obvious demands made upon the insight and cooperation of the subjects, a test of this nature seems to have practical possibilities. Professor H. T. Moore has obtained significant correlations between ratings on aggressiveness and the ability to gaze unwaveringly into the experimenter's eyes while performing a mental calculation. Expansion, another trait requiring a social milieu, may be roughly estimated by the character of the ego-references, self-descriptions, and personal opinions included in a letter of application for a position written by the subject.
The measurement of character and other sociality traits has been approached from the standpoint of ethical and social knowledge. Dilemmas involving moral principles to which the subject must think out the answer form one type of test procedure. The subject, for example, might be asked whether a man is justified in keeping five dollars which he sees another man drop, provided the latter owes him the money and has refused in an insulting manner to pay it. (G. G. Fernald.) Other tests involve the definition of moral terms, the evaluation of punishments which ought to accompany certain offenses, and the proper selection of reasons against types of unethical conduct. 'Ethical discrimination' is the name given to the trait presupposed by these tasks. (Kohs.) It is open to question, however, whether success in such tests does not show intelligence rather than character. One would be greatly aided, of
(136) course, in solving the test if he had been brought up under moral instruction or had thought about the social obligations implied in the problems. To this extent we should expect one who receives a high score to be well grounded in moral habits; and to this extent, therefore, the test would be successful. One's general attitude toward morality might also be reflected in the result. Dr. Myerson's work, discussed under tests of emotional attitude, extends the completion method (described on page 133) to the detection of ethical tendencies.
An interesting test devised by Dr. G. G. Fernald requires the arrangement of a shuffled series of crimes, of widely varied gravity, in the order of magnitude from the least to the most serious offense. In the performance of this test greater deviations from the norm established by a group of legal and scientific men were found among reformatory inmates than among law-abiding groups. In giving this test to normal persons the writer has found a considerable number of thoroughly ethical individuals who had wide deviations from the norm. These are to be explained, not as defects of `moral vision,' but as the result of a highly personal attitude toward the offenses, an attitude based on feeling rather than objective social and judicial policy. Probably most criminals have this sort of personal reaction to legal and moral problems; hence their tendency to deviation in the test. But the converse, that all who have such attitudes are criminalistic, is far from true. True characterial defect requires the addition of other factors as described in Chapter V.
Adequate tests of character must involve actual drives. The artificiality of any present testing situation seems to be a discouraging hindrance in coping with the problem. Character is truly revealed only in the vital issues of real life. It will probably be some time before we shall have advanced beyond the methods of the business man who, desiring to employ an office boy, allowed each applicant to walk into a room where lay- an unguarded pocketbook, while the prospective employer watched behind the crack of the door.
To determine the trait of social participation further use has been made of the `social knowledge' method. The subject is required to
(137) answer questions and to define terms drawn from the technical jargon of sports and amusements, and from the vocabulary of church hymns, parliamentary procedure, and etiquette. (Ream.) The assumption here is that if a man has true social tendencies, he will have mingled with all sorts of groups and will have acquired the lingo of each. For measuring susceptibility to social stimulation the writer has used a test of interpreting facial expressions from photographs, which will be referred to in a later chapter.
This completes, in the main, the roll of commendable but wholly tentative approaches to the measurement of personality. The problem of bringing human emotions, drives, and social attitudes to adequate expression in an artificial test situation is a perplexing but perhaps not an impossible one. Workers in this field should remember that it is in theory as well as in application that progress is needed. The deeper our understanding of the fundamental driving forces of personality, the more certain will be our success in isolating and measuring their manifestations.
Types of Personality. Thus far we have dealt with traits independently of one another. The important question remains as to how far these qualities are associated in definite degrees, forming patterns or types into which a large proportion of humanity can be placed. Although every human face is distinct, we can recognize types of faces. And so with personality, in the midst of infinite variety in minor details a few common patterns may be found. The problem is one largely of correlation. If a group of people can be found in which high or low ranks in certain traits are associated, either positively or inversely, with high or low ranks in other traits, so that a fairly constant pattern or profile of traits results, these individuals may be said to constitute a type.
A central problem in intercorrelation is the relation between intelligence and sociality. The correlation is seemingly positive, although the existence of highly intelligent criminals affords a striking exception. Dr. Webb, in collecting ratings of school and college students, found a correlative tendency, which he calls a 'general factor,' underlying character. Its presence is shown by a variety of virtues, and its absence by a variety of defects. Desirable traits correlate highly with one another; while undesirable
(138) traits also intercorrelate highly, and correlate inversely with desirable ones. There is, moreover, a positive correlation between the admirable qualities and intelligence. It is probable that Dr. Webb's concept of a general factor is to be explained in terms of the highly integrated organization of allied drives described in our discussion of character. The general factor is really a genetic one, and consists of the formation of many prepotent habit trends all developed in one direction by social approval and disapproval. The factor of intelligence merely increases the rapidity of fixation of these socially useful habits. Intelligence, being a capacity, is innate; whereas excellence of character, consisting essentially of habits, is acquired.
Professor Terman has refuted the notion that the intellectually precocious child is one-sided. Ratings by teachers and elders in comparison with intelligence test scores show that the mentally superior child is usually superior also in personal and moral traits. Superior children generally come from superior homes — homes in which their high native learning capacity can be used to assimilate social and moral virtues from their environment. The high correlation of intelligence and character is to be expected under conditions like these.
Significant conclusions regarding types have been derived from the correlation of ratings in the field of self-expression. Ascendance, expansion, extroversion, and high self-evaluation (expressing self-confidence) are all positively correlated. Their opposites, of course, also correlate with one another. The correlation between ascendance and extroversion in one group used was found to be .70; while that between ascendance and expansion was .86. There are, of course, exceptions such as occasional introverts who develop expansion as a kind of compensation. There are also some extroverted reclusive persons. On the whole, however, we may recognize two prominent types, one high in self-expression and the other low in that sphere. (G. W. Allport, F. H. Allport,) Professor
(139) Downey has found similar correlations, impulsion (of a somewhat expansive type) correlating with feeling of self-worth to the extent of .81, and with aggressiveness (ascendance) with a coefficient of .82. The existence, therefore, of two types of self-expression seems fairly well established. Because of the dominant and subordinate roles they play in social contacts we may call them respectively the strong and the weak types.
It seems probable that the most important general factors underlying these two types are respectively the excellence and the defect of physique. Ascendance, impulsive energy, self-confidence, expansion, and an extroverted view of life, all seemed in the individuals studied to go with good physical development, and their opposites with illness and defect. Other factors, of course, and particularly compensatory ones, may operate to limit the application of this theory. Positive correlations are also indicated between strength of self-assertive traits and sociality factors. The strong personality, more regularly than the weak, is characterized by social adaptability, participation in social affairs, and sensitivity to social influences.
GENERAL SUMMARY — THE INDIVIDUAL AS A UNIT IN SOCIAL BEHAVIOR
Starting from the premise that social psychology is concerned with the behavior and consciousness of the individual in relation to his fellow beings, our first task has been to study those aspects of the individual which are destined to direct and control his behavior within the social sphere. In order to understand these aspects it has been necessary to delve into the very fiber of the organism, and to disclose the forces and the methods by which men feel, think, and act. To begin with the organism itself, man is essentially an enormously complex system of reflex arcs, whose central portions are so plastic, so modifiable, and so richly interconnected that all manner of coordinations are possible between stimulations and
(140) acts; and the most subtle integrations of habit and thought may be acquired and retained for future adaptation.
A few inherited reflexes of prepotent character form a crude but vital basis for acquiring acts of defense, nutrition, and sex, in response to importunate stimuli from within or without the body. Conditioned response and motor learning develop these reflex movements into great systems of adaptive habits both universal among mankind and peculiar to individuals. The social environment, chiefly through language, exerts a vast influence upon this process of modification. It directs the channels into which the prepotent demands shall flow, determines and gives instruction in the means for their satisfaction, and inculcates in the individual the drive toward adaptation and approval within the social sphere. Socialization is thus achieved by learning within the social environment.
Internal, or visceral, responses to stimuli combine with the overt behavior to produce an emotional reinforcement in the struggle for adjustment. Here again the forces of society enter into the problems of individual life. The evoking and conditioning of love, sympathy, and aversion, as well as of the complex human sentiments, proceed according to the social influences brought to bear. Avoidance of repression through society, and the use of the reinforcing effects of emotion for social ends, are problems encountered in this field.
Finally, the laws of social contact cannot be understood without an appreciation of the capacities, driving forces, and habit trends which constitute personality. Intelligence, movements, emotions, personal ascendancy, drive, compensation, grasp of reality, self-understanding, social capacities, and character are all subject to variation and combination among individuals in ways that profoundly affect their adjustments to one another. It is important to know how these traits may be recognized in social relations, how society has operated in their formation, and what expedients of a social nature may be employed in evoking and measuring them.
Personality is preeminently the social aspect of the individual. With its study we complete the potentialities of the human being
(141) for social life, and pass on of consider those interactions with other human beings which constitute his social behavior.
Questionnaires and Scoring Methods:
Wells, F. L., "The Systematic Observation of the Personality in its Relation to the Hygiene of Mind," Psychological Review, 1914, XXI, 295-333.
Partridge, G. E., An Outline of Individual Study.
Yerkes, R. M., and LaRue, D. W., Outline of a Study of the Self.
Heymans, G., and Wiersma, E., "Beitrage zur speziellen Psychologie auf Grund einer Massenuntersuchung," Zeitschrift fur Psychologie, 1906, XLII, 81-127;258-301.
Hoch, A., and Amsden, "A Guide to the Descriptive Study of the Personality," State Hospital Bulletin, New York, 1913.
Watson, J. B., Psychology from the Standpoint of a Behaviorist, pp. 226-30; 399-411.
Allport, F. H., and G. W., "Personality Traits: Their Classification and Measurement," Journal of Abnormal Psychology and Social Psychology, 1921, XVI, 1-40.
— The Personnel System of the United States Army: Vol. II, "The Personnel Manual," Washington, C. C. P. 400, 1919, ch. 12.
Davenport, C. B., "The Trait Book," Eugenics Record Official Bulletin, no. 6, 1912.
Spaulding, E. R., "The Role of Personality Development in the Reconstruction of the Delinquent," Journal of Abnormal Psychology and Social Psychology, 1921, XVI, 97-114.
Woodworth, R. S., "Emotional Questionnaire" (in S. I. Franz: Handbook of Mental Examination Methods, eh. 12).
Fernald, W. E., "The Diagnosis of the Higher Grades of Mental Defect," American Journal of Insanity, 1914, LXX, 253-64.
— "Standardized Fields of Inquiry for Studies of Defectives," Mental Hygiene, 1917, I, 211-34.
Porteus, S.D., "A Study of Personality of Defectives with a Social Ratings Scale," Publications of the Department of Research, Vineland Training School, New Jersey, 1919-20 series, no. 6.
Kingsbury, F. A., "Analyzing Ratings and Training Raters," Journal of Personnel Research, 1922-23, I, 377-83.
Paterson, D. G., "The Graphic Rating Scale," Journal of Personnel Research, 1922-23,1,361-76.
Knight, F. B., "The Effect of the `Acquaintance Factor' upon Personal Judgments," Journal of Educational Psychology, 1923, XIV, 129-42.
Knight, F. B, and Franzen, K. H., "Pitfalls in Rating Schemes," Journal of Educational Psychology, 1922, XIII, 204-13.
Norsworthy, N., "The Validity of Judgments of Character," Essays in Honor of William James, 1910 (2d ed.), 542-52.
Cattell, J. M., American Men of Science. Hollingworth, H. L., Vocational Psychology, chs. 6, 7. Rugg, H., "Is the Rating of Human Character Practicable?" Journal of Educational Psychology, 1921, xtt, 425-38; 485-501; and 1922, XIII, 30-42, 81-93.
Thorndike, E. L., "A Constant Error in Psychological Ratings," Journal of Applied Psychology, 1920, Iv, 25-29.
Special Tests and Experiments:
Stern, W., "The Psychological Methods of Testing Intelligence" (translated by G. M. Whipple), Educational Psychology Monographs, no. 13.
Terman, L. M., The Measurement of Intelligence.
Downey, J. E., "The Will Profile: A Tentative Scale for the Measurement of the Volitional Pattern," University of Wyoming, Bulletin, 1919, xvi, no. 4b.
Graphology and the Psychology of Handwriting.
Carnegie Institute of Technology, Will-Temperament Test (adapted from Downey), Bureau of Personnel Research, 1920, test no. 9.
Smith, W. W., The Measurement of Emotion.
Filter, R. O., "An Experimental Study of Character Traits," Journal of Applied Psychology, 1921, v, 297-317.
Whipple, G. M., Manual of Mental and Physical Tests, part I, chs. 1-5.
Pressey, S. L., and Chambers, O. R., "First Revision of a Group Scale for Investigating the Emotions," Journal of Applied Psychology, 1920, IV, 97-104.
Pressey, S. L., "A Group Scale for Investigating the Emotions," Journal of Abnormal Psychology and Social Psychology, 1921, xvi, 55-64.
Myerson; A., "Personality Tests Involving the Principle of Multiple Choice," Archives of Neurology and Psychiatry, 1919, t, 459-70.
Moore, H. T., "A Method of Testing the Strength of Instincts," American Journal of Psychology, 1916, XXVII, 227-33.
Moore, H. T., and Gilliland, A. R., "The Measurement of Aggressiveness," Journal of Applied Psychology, 1921, v, 97-118.
Jung, C. G., Studies in Word Association, pp. 1-172.
Wells, F. L., "The Question of Association Types," Psychological Review, 1912, XIX, 253-70.
— "Some Properties of the Free Association Time," Psychological Review, 1911, XVIII, 1-23.
— Mental Adjustments, ch. 7.
Kent, G. H., and Rosanoff, A. J., "A Study of Association in Insanity," American Journal of Insanity, 1910, LXVII, 37-96; 317-90.
Hull, C. L., and Lugoff, L. S., "Complex-Signs in Diagnostic Free Association," Journal of Experimental Psychology, 1921, IV, 111-36.
Dooley, L., "Correlation of Normal Complexes," American Journal of Psychology, 1916, XXVII, 119-51.
Fernald, G. G., "The Defective Delinquent Class: Differentiating Tests," American Journal of Insanity, 1912, LXVIII, 523-92.
Kohs, S. C., "An Ethical Discrimination Test," Journal of Delinquency, 1922, VII, 1-15.
Ream, M. J., "A Social Relations Test," Journal of Applied Psychology, 1922, VI, 69-73.
Terman, L. M., The Intelligence of School Children, chs. 10, 11.
Webb, F., "Character and Intelligence," British Journal of Psychology, Monographs, no. 3, 1915, 1-99.
Pillsbury, W. B., Essentials of Psychology (revised edition), ch. 16.
Allport, F. H., and G. W. (see reference under Questionnaires).
Downey, J. E., "Some Volitional Patterns Revealed by the Will Profile," Journal of Experimental Psychology, 1920, III, 281-301.
— The Adolescent Will Profile, Journal of Educational Psychology, 1920, XI, 157-64.
Hollingworth, H. L., Judging Human Character,