views updated May 29 2018



The “trait” concept represents an attempt to account for consistencies within personality as well as to provide personality study with a long-needed, scientifically sound taxonomy. If one is to use a theory which states that personality has altered between occasion A and occasion B, through such an influence as cultural pressure or therapeutic effort, then one must know how to describe and measure the personality before and after this alleged influence has been at work. Most sciences have had a distinct taxonomic stage, in which they have set out to describe their subject matter accurately and to designate the units by which phenomena and their changes are to be measured. If personality is the central topic of psychology, then a glance at psychology’s history will show that it has been an immature and even a problem science. It has always been eager to get at the ostentatious theories before it has done the humble work of describing and measuring the subject matter. Admittedly, any taxonomy can become the playground of pedants, and naturally such a perversion has occasionally occurred in science. But the faults of psychology, and of social psychology in particular, have certainly not developed in this direction.

Through the centuries there admittedly has been considerable activity in describing personality, on the one hand by literary folk and on the other by medical men. The literary contribution has consisted overwhelmingly of what has been called an ideographic approach, concerned with depicting an individual personality in all its idiosyncrasy for purely aesthetic ends, without any regard for scientific principles. Occasionally, as with the literary characterologists of the seventeenth and eighteenth centuries, there have been explicit efforts to generalize about personality description, but these are of only historical interest to psychologists today. But the second source, the medical profession, beginning with Galen’s description of the four temperaments and their relations to four humors— choleric, melancholic, sanguine, and phlegmatic— has provided descriptions that have developed steadily over centuries. The most decided development occurred when Emil Kraepelin recognized the entity that is now called schizophrenia, and when Eugen Bleuler further recognized that psychotics, as well as normal individuals, could be arranged on a continuum from cyclothymic to schizothymic temperament. Naturally, most medical work has been concerned with abnormal forms of behavior, and medical concepts have crept into the description of normal behavior only through the popularization of clinical notions, as by psychoanalysis, and through the literary “sophistication” of describing as paranoid or psychotic individuals with whom one does not agree! [See the biographies ofbleuler; kraepelin.]

Types and traits . As one moves into modern treatments, which, by contrast to the clinical sources of trait description, go hand in hand with advances in measurement and statistical analysis, one must at the outset recognize the alternative technical paths of personality description constituted by “types” and by “traits.” A type means a whole pattern that is repeated with a striking frequency and that can be distinguished from a number of other patterns that also have a noteworthy frequency of representation among individuals in the population. Thus, one may speak of a hero or a scoundrel, a schizophrenic or a hysteric. In speaking of type, one employs a noun; in dealing with traits, one basically uses an adjective. Thus, a certain fruit may be described, in terms of a type, as “an orange,” and in terms of traits, as an object that is spherical, about three inches in diameter, orange in color, and soft to palpation.

However, these two approaches should not be regarded as utterly distinct, but rather as two ways of conveying the same kind of information. Of course, it is true that people frequently employ the word “type” when they are using a categorical or Aristotelian mode of thinking, admitting no degrees of anything: a certain animal can be designated either a dog or a cat. When one uses the notion of “trait,” on the other hand, there is invariably implied a definite characteristic of which one can have more or less. This is, nevertheless, not a fundamental distinction, since in any adequate use of the notion of “type,” it is first defined in terms of positions on continua; for instance, horses and dogs can be described on a continuum of length. In recognizing two types, one merely admits that there are two “modes,” in statistical terms; for example, values in a certain range occur very frequently for horses and values in another range occur very frequently for dogs, while intermediate values are infrequently found for either species.

Fundamentally, therefore, one must consider that a type is a pattern in trait measures falling at certain modal values. Consequently there is, in general, always a path of translation between description by types and description by traits. Incidentally, what has been said here about personality description by traits and types applies to the description of any object by any attributes. The theory and the statistics used in personality description belong, in principle, to any taxonomy; and the mathematical and statistical concepts that psychologists have recently developed likewise have the highest generality.

Three approaches to traits. When the psychologist sets out to describe a personality by traits, he has three alternatives. First, he can define a trait as an absolutely specific and narrow aspect of behavior: for instance, canceling letters on a page of print at the fastest possible speed, biting one’s fingernails, or disliking Siamese cats. Second, he can define a trait as a whole pattern or collection of such specific behaviors, as is done with such general characteristics as courage or sensitivity. This second alternative supposes that there is a whole aggregate of “trait elements” that “go together” for example, that the courageous person will be brave at the dentist’s and also firm in dispute. But in this second sense, the trait may be an invention of the psychologist’s own mind, corresponding with nothing in nature. For example, it may turn out that the person who is courageous in climbing tall trees is not courageous at the sight of blood, or morally courageous in disagreeing with his neighbor.

By following the first course and avoiding this uncertainty through referring to an absolutely specific bit of behavior and setting out to measure it, the psychologist certainly is in a scientifically safe position, provided people vary in the specified characteristic. But he is also in a scientifically depressing situation, for there will be a virtually infinite number of such specific traits and a limited number of psychologists—who may choose to operate with different traits and names and, like parallel lines, never come together in common researches.

If the psychologist chooses to follow the second path—that of locating broad patterns—he may begin by stating, perhaps with the precise accompanying definition, what behaviors are considered to be covered by the trait he uses. Such a procedure may seem “operational” and, as such, of good scientific respectability. But the psychologist who follows this course is an unconscious humbug, since there is no proof whatever that things will go together in the way that he says; and his ostentatious procedures of measurement actually guarantee only a hodgepodge of different units and miscellaneous contributions.

The third course is the only one that satisfies the requirements of economy, of reality, and of conceptual clarity for experimental purposes. In this case, one first demonstrates or discovers the real “going together,” or functional unity, of the elements of the constellation of behavior that will be covered by the trait named. Thus, one might measure thirty different kinds of specific courageous behaviors and put two hundred people in rank order on each one of them. If one finds the same ranking is obtained for a dozen out of these thirty forms of behavior—that is, that the twelve forms of behavior correlate highly—then one accepts this group of twelve behaviors as defining a useful trait.

Of course, one may have to abandon the word “courage” and use some more technical term to describe these twelve forms of behavior that do not belong with the remaining eighteen. The general public has the right to keep its word “courage” for the entire group of thirty behaviors, if it so wishes, just as the psychologist has the right to use quite a new technical term for what he finds in a technically more sophisticated examination of personality structure. Indeed, he will do better if he invents a new technical term for anything that he can truly demonstrate, for he will thus avoid the trailing clouds of confusion that will forever dog the popular term.

The psychologist uses the correlation coefficient to establish the degree of going-togetherness. He needs such a graduate index because in the complex realm of behavior there is seldom an all-or-nothing relationship, but rather a tendency of two things to go together to a high degree or to a negligible degree [seemultivariate analysis, articles oncorrelation].

Surface traits and source traits . In the third approach, there are still two alternatives that the psychologist can consider: he can use either a surface trait or a source trait. When a group of behaviors correlate from individual to individual, one is initially only entitled to call this functional unity a surface trait. When correlation coefficients between behaviors are arranged in a correlation matrix, the surface trait is recognized and defined by a correlation cluster; i.e., by a cluster of variables, each of which correlates with every other variable in the cluster to a marked degree. Such a cluster of variables may have no correlation with another cluster of variables and thus enjoys a certain independence.

Methodological refinement occurs when it is recognized that the correlations among a set of variables may be a result of more than one influence. For example, if one takes a random sample of men in the street and tests them on vocabulary, speed in arithmetic, knowledge of American history, and knowledge of geography, it is likely that these four different kinds of performances will correlate positively and appreciably in all six possible subject pairings. In other words, the individual who is much better in one task will have a greater-than-average probability of being better in the others; and one can speak safely of a surface trait. However, this tendency of the four performances to go together springs from two sources. On the one hand, it is caused partly by individual differences in intelligence, so that although all the men may have been equally exposed to the same school subjects, one will have learned much more in all four areas than will another. On the other hand, it also springs from differences among the individuals in length of schooling, since all four areas happen to be simultaneously taught in school. These two influences are said to be source traits, since they are basic sources, or causes, of common variance and therefore account for the observed correlations.

Factor analysis. The discovery of such underlying source traits among observed correlations has become possible through the method of factor analysis, which was devised by Charles Spearman (1927) and continued by L. L. Thurstone (1935; 1947) and others and which is today one of the most refined and flexible technical methods for analyzing trait structure. If the mathematical psychologist is presented with the complete intercor-relation matrix for many forms of behavior measured in many people, he can, by factor analysis, arrive at the number and to some extent at the nature of basic sources of variance, which are necessary to account for the correlations [seefactor analysis].

Common and unique traits. Another distinction among trait concepts, which cuts across the surface-source differentiation, is that between common traits and unique traits. A common trait is one that everyone can be said to possess in some amount, such as intelligence, pugnacity, or sensitivity. A unique trait, on the other hand, is something on which no one but the person being described can be measured. An extensive discussion of this distinction has been made by Allport (1937), but not necessarily in statistical terms. It should be noted at the outset that one can, if necessary, abandon unique traits without abandoning at all the idea of the unique individual. One has only to think of any common trait as a set of coordinates, and to remember that the individual can be placed as a unique point in space in regard to a set of such coordinates (which normally would lie in hyper-space).

Furthermore, in the same terms one can handle not only the uniqueness of the person but also what many people are thinking of when they speak of the “uniqueness of a trait.” Indeed, it should be noted that in general the same absolute level of performance in any specific task can always be reached by many different combinations of the same pure common traits.

Some statistical psychologists have been inclined to dismiss the whole notion of unique traits, saying that the ideographic approach in personality really belongs to art and, unlike the nomothetic approach, has no relation to science. Science is concerned with what is generalizable, and is interested in explaining every individual case in terms of common concepts. To explain the individual case in terms of individual concepts is to gain no economy whatsoever, in terms of either concepts or laws, and is a chimera. However, there are actually senses in which a unique trait, if precisely defined, has scientific meaning. First, in the whole area of dynamic traits and interests, such as the clinician is most concerned with, one often encounters traits on which relatively few people can be scored; for instance, a passion for Brazilian butterflies. It is in the nature of interests, especially trivial interests, that they can be very specific and peculiar; and it is indeed true that there is no way of handling them other than by that seeming paradox, a “relatively unique” trait—a trait of which only a few people have some amount. Second, and more important, if one correlates thirty aspects of one person’s behavior over a hundred days, measuring these same thirty aspects of behavior each day, one can obtain, by what is called P-technique factor analysis, a pattern that represents a trait in the sense of a set of behaviors that covary from day to day. Such a pattern might be unique to an individual. For example, by such an approach anxiety can be clearly recognized as a factor within an individual; and in each individual this anxiety will express itself through slightly different means and object avoidances. Although this concept of anxiety refers to something unique to the individual, it still has scientific value, since through its use one may discover laws and make generalizations about the way this trait behaves over time in an individual.

Trait modalities. It has long been customary to speak of three “modalities” of traits: abilities, or cognitive traits; temperament traits; and dynamic traits, or interests. The definition of the way in which these three kinds of traits differ has been left mainly at a popular level, but an exhaustive technical definition has been attempted by Cattell (1957) in terms of whether the measurement made is most sensitive to changes in the complexity of the stimulus, in which case the trait is called an ability, or to changes in the environmental incentives, in which case it is called a dynamic trait. A temperament trait, largely independent of the stimulus complexity and of the level of motivation at the time, represents some kind of readiness to respond. It is a “stylistic” kind of attribute. (The implication that temperament traits are inherited is a secondary, not an essential, characteristic.) Normally, all three modalities of traits will contribute to any given piece of behavior, although in order to get a good measure of a trait that is primarily of one modality, one should try to choose those extreme examples in which the measure becomes largely independent of the other two modalities. Thus, in measuring the trait of intelligence, one tries to avoid measures that would reflect temperament differences and to control the motivation level of those taking the test.

When one deals with common factor source traits as defined above, the interactions of different traits and different modalities in determining any specific behavior or performance can be expressed in what is called the specification equation, which is written

Pji= SjA1TAji+ …+ SjAkTAki + SjTlTTti + …

+ SjTlT Tli + SjD1,TDli + … + SjDmT Dmi + SjsT 8i.

Here, the pj, means a performance in response to the situation, j, which reaction is of the same nature for all people but differing in magnitude. Generally the T’s are source traits which have the particular value, in a standard score, with which the individual is endowed; T.i represents an ability trait, TT a temperament trait, TD a dynamic trait. The T.s represents a factor specific to the test situation. The s’s are weights which indicate how the particular situation, provokes and involves these traits for people in general. These weights are obtained by the factor analysis, or simply by correlating the magnitude of the response, p, with the magnitude, of the trait, T, over a suitable sample of people. The i attached to the T’s and to p indicates the score of a particular person z. The k, Z, and m represent the number of the respective traits. This specification equation thus states that the given response or performance p is determined by the individual’s endowments in ability, temperament, and dynamic (motivational) traits, and that these will operate with weights which can be determined factor-analytically for the given performance, showing the extent to which that situation involves the traits.

Methods of discovering traits. The history of discovery and interpretation of particular traits is, of course, scarcely half written. The field of abilities, because of its intense interest to the educationist, has been most explored. Beginning with the location of a general ability factor by Spearman (1927), and the discovery by Thurstone (1935) of such primary abilities as number ability, spatial ability, verbal ability, and perceptual speed, and continuing into Guilford’s recent work with cognitive traits that involve productive performance rather than judicious decision, perhaps thirty traits have been located (1959). Personality traits have been pursued most systematically through the rating and questionnaire media of observation and measurement. There are systematic differences between the oblique factor system used by Thurstone (1947) and Cattell (1957) in this field and the orthogonal system used by Guilford and Zimmerman (1956), among others. Typically, some twenty independent source traits or dimensions have been found through factor analysis, as exemplified in the Guilford-Zimmerman (G-Z) scales and in the Institute for Personality and Ability Testing (IPAT) 16 Personality Factor Scales (16 P.F.). Incidentally, the Minnesota Multiphasic Personality Inventory (MMPI) questionnaire differs from the 16 P.F. in that it deals with surface traits, long traditionally established in psychiatric work, rather than with source traits, which were first confirmed in the normal range of behavior. All of the scales mentioned have by now been related to an appreciable number of real-life criteria, both normal and abnormal. The traits in the 16 P.F. and the High School Personality Questionnaire (HSPQ), for example, have shown themselves capable of a substantial degree of prediction in everyday life.

The weakness in the questionnaire test medium is that it lends itself to faking or distortion through dishonesty or lack of self-knowledge. In response to this problem there has been considerable development of tests in which the person’s behavior in a miniature situation is measured without his knowledge of what is really being measured. Correlating and factor-analyzing such behaviors reveals a set of source traits that can now be measured in tests, such as the Objective-Analytic Personality Test Batteries (O-A) (Hundleby et al. 1965). The relationship of the traits found in this type of objective test to those found in questionnaires (which are objective only in their scoring and are best called conspective—i.e., such that two examiners see the same score) is somewhat com plicated. So far, what are called first-order factors in the objective tests have been revealed to correspond to second-order factors in the questionnaires. Approximately, one may state that a second-order factor is a broad influence that organizes primary factors. Thus, anxiety is a trait that shows itself in the questionnaire at the second order, accounting for variability in ego weakness, ergic tension, guilt proneness, and so on; but in the objective test realm (O-A), it emerges as a first-order factor, accounting for low skin resistance on the GSR, unwillingness to venture in a new situation, and large startle reaction to the cold pressor test. Consequently, one may measure the trait of anxiety with equal facility through either questionnaire or objective test.

Trait stability and generality. As soon as a trait is clearly located and confirmed by research on different samples, the first questions one is likely to ask concern how the trait changes with age; whether it is largely inborn or largely due to environment; and what particular influences, in the latter case, will affect it. The applied psychologist is also immediately interested in knowing what predictive value the trait will have for him in clinical diagnosis, in predicting industrial personnel fitness, or in enabling one to select scholarship recipients more reliably. Spearman’s general ability factor, for example, has been shown to hold as a unitary entity through all age ranges, although it changes its pattern steadily in going from, say, five years of age to fifteen years (1927). Nevertheless, this “identity in change” can be reliably and validly measured at any age level; its magnitude is found to climb steeply in the early years and to flatten out around fifteen years. It has similarly been shown repeatedly that the main personality traits—such as ego strength, cyclothymia-schizothymia temperament, dominance-submissiveness of disposition, surgency-desurgency, and radicalism-conservatism —as rated or as measured in the 16 P.F., the HSPQ, and the G-Z scales, maintain a continuity with age, and that one can measure the same factors at any age. There are a few traits that emerge in adults but do not appear in children; there are also a few factors on which children vary a good deal that contribute only trivial variance in adults. The dozen or so main factors, however, seem to persist very steadily. [Seepersonality, article on PERSONALITY DEVELOPMENT.]

This continuity of personality trait structure is true for both the questionnaire and the objective test. The Early School Personality Questionnaire (ESPQ) and the O-A Battery for young children measure the same traits as do the HSPQ and the regular O-A Battery—at decidedly lower levels of age. What is perhaps of greater interest to the anthropologist and the sociologist is the research demonstrating the degree of stability of these traits across cultures. It was theoretically possible that the results obtained through factor analysis in any medium—questionnaire responses, ratings, or objective tests—in, say, Japan or India would prove to be very different from those in America or France. Actually, recent work has shown quite clearly that the majority of general personality traits persist with considerable stability across cultures. Ego strength, schizothymia, superego strength, anxiety, and surgency, among other traits, seem to show themselves as basic personality dimensions in most cultures.

The measurement of dynamic traits—i.e., of interests and motivation—by objective test methods has been pursued by many (Kuder, Strong, Guilford, Cattell, Eysenck, and others) but perhaps is still in its infancy. However, there are strong indications that the correlational analysis of traits in the dynamic field yields primarily what have long been recognized as basic drives, such as sex, fear, hunger, parental protectiveness, self-assertion, and curiosity. Additionally, there appear certain structures that have been called sentiments, or attitude aggregates, and that correspond to the individual’s learning simultaneously whole sets of emotional attitudes and ways of behaving. This learning presumably occurs in response to the impress of some single social institution, such as a religion or the family. The dynamic structures that correspond to drives, or ergs, as they have been more recently defined, are expected to prove to be basic across cultures; the sentiment traits are almost certainly peculiar, to a considerable degree, to specific cultures. Thus the conclusion is reached that although it may be possible to compare certain traits across cultures, it is virtually certain that there are other traits in which it is meaningless to try to make a quantitative comparison.

States and roles. Two concepts that clearly must be theoretically distinguished from traits are those of states and of roles. Failure to make this distinction also brings doubt and inaccuracy in the field of measurement. A state is defined as a pattern of covarying elements (as in surface traits and source traits) that does not show consistent, significant covariation in individual-difference measurements. A state is a pattern that shows differences only within one person, over time, and even then must be distinguished from a fluctuating trait.

By a role, the psychologist (at least when he gets to precise measurement) must refer to a pattern of behavior, or an emphasis in behavior, that occurs only in a set of specific situations. In general terms, it can be said that a personality trait refers to a factor that operates over a wide array of situations, whereas a role factor is quite specific to a certain limited group of situations.

The psychometrician is quite as prepared as the sociologist or the anthropologist to conceive that in principle an individual never performs an act outside of a role. However, the psychometrician is more skeptical than the anthropologist and the sociologist that one can detect, merely by inspection, what particular role the individual is acting. He prefers, instead of claiming to “know” a role, to locate role factors, just as one locates personality factors by actually correlating a set of performances and showing that certain stimulus situations tend to evoke similar behavior. The scientific study of roles in this sense is in its infancy, but the model is reasonably clear. In short, to detect a role, one looks for persisting correlations across situations that call forth many different traits; to detect a trait, one looks for a single behavior factor that runs across many different roles.

Instrument factors. Related to the idea of role is the idea of “instrument factor.” It has been found that in attempts to measure traits, there is apt to be a particular effect attributable to the kind of instrument or mode of observation used; for example, behavior rating, questionnaire, or objective test. Certain pervasive instrument factors therefore must be recognized, and their effects removed, before one can hope to get a clear concept of the trait or an uncontaminated measurement of it.

The person as a functional unity. Misgivings are sometimes expressed that the notion of traits is an “atomistic” concept that succeeds in taking the individual apart but typically fails to put him together again. It is true that certain ad hoc trait scales with which psychologists have concerned themselves excessively are of this kind, in that, being arbitrarily set up, there is no way to combine the different scales and alleged “traits” in a rational fashion; they do not correspond to any real structures in the organism. However, if measures of traits are based on prior structural research that provides well-defined source traits, the individual can in fact be described as a single functional unity. Further, his total pattern as an individual can be expressed as a gestalt by a profile of scores on the basic source traits. The only limitation to this integration lies in the possibility that in some cases traits do not interact additively, but have a relationship expressible by a higher power. Actually, with quite minor exceptions, nonadditive treatment has not yet been clearly demonstrated as a necessity, and the scientist should continue to use the simpler model unless it breaks down. At present, one adds a person’s score on an intelligence trait to his score on a personality trait, such as ego strength, and to his score on a dynamic trait, such as interest in intellectual work. With this total, one can make a fairly good prediction of an individual’s performance in some concrete intellectual endeavor.

Raymond B. Cattell

[Other relevant material may be found in Aptitude Testing; Factor analysis; Personality; Personality measurement; Role; Vocational interest testing; and in the biographies of Spearman; Thurstone.]


Allport, Gordon W. 1937 Personality: A Psychological Interpretation. New York: Holt.

Cattell, Raymond B. 1957 Personality and Motivation Structure and Measurement. New York: World.

Cattell, Raymond B.; and Scheier, I. H. 1961 The Meaning and Measurement of Neuroticism and Anxiety. New York: Ronald Press.

Eysenck, Hans J. (1953) 1960 The Structure of Human Personality. 2d ed. London: Methuen.

Guilford, Joy P. 1959 Personality. New York: McGraw-Hill.

Guilford, Joy P.; and Zimmerman, W. S. 1956 Fourteen Dimensions of Temperament. Psychological Monographs 70, no. 10:1-26.

Hall, Calvin S.; and Lindzey, Gardner 1957 Theories of Personality. New York: Wiley; London: Chapman.

Hundleby, John; Pawlik, Kurt; and Cattell, Raymond B. 1965 Personality Factors in Objective Test Devices: A Critical Integration of a Quarter of a Century’s Research. San Diego, Calif.: Knapp.

Jung, Carl G. (1921) 1959 Psychological Types: Or the Psychology of lndividuation. London: Routledge. → First published in German.

Roback, Abraham A. (1927) 1952 The Study of Character, With a Survey of Temperament. 3d ed., rev. & enl. London: Routledge; Cambridge, Mass.: Sci-Art Publishers.

Spearman, Charles E. 1927 The Abilities of Man: Their Nature and Measurement. London: Macmillan.

Spearman, Charles E.; and Jones, Llewelyn W. 1950 Human Ability. London: Macmillan. → A continuation of Spearman’s The Abilities of Man, 1927.

Thurstone, Louis L. 1935 The Vectors of Mind: Multiple-factor Analysis for the Isolation of Primary Traits. Univ. of Chicago Press.

Thurstone, Louis L. 1947 Multiple-factor Analysis. Univ. of Chicago Press. → A development and expansion of Thurstone’s The Vectors of Mind, 1935.


views updated May 29 2018


Characteristics that differ from one person to another in a continuous and consistent way.

Traits include such personality characteristics as introversion , aggressiveness, generosity, nervousness, and creativity . Systems that address personality as a combination of qualities or dimensions are called trait theories.

The first comprehensive trait theory was that of Gordon Allport (1897-1967). Over a period of thirty years, Allport investigated over 18,000 separate traits, proposing several principles to make this lengthy list manageable for practical purposes. One was the distinction between personal dispositions, which are peculiar to a single individual, and common traits, which can be used for describing and comparing different people. While personal dispositions reflect the individual personality more accurately, one needs to use common traits to make any kind of meaningful assessment of people in relation to each other. Allport also claimed that about seven central traits dominated each individual personality (he described these as the type of characteristic that would appear in a letter of recommendation). Another concept devised by Allport was the cardinal traita quality so intense that it governs virtually all of a person's activities (Mother Theresa's cardinal trait would be humanitarianism, for example, while that of the fictional character Ebenezer Scrooge would be avarice). Secondary traits, in contrast, are those that govern less of a person's behavior and are more specific to certain situations.

Using the statistical technique of factor analysis, Raymond B. Cattell reduced Allport's list of traits to a much smaller number and then proceeded to divide these into clusters that express more basic dimensions of personality (for example, the pairs talkative-silent, open-secretive, and adventurous-cautious can all be grouped under the overall source trait of extroversion ). Eventually he arrived at 16 fundamental source traits and developed a questionnaire to measure themthe Sixteen Personality Factor Questionnaire (16 PF)which uses the answers to over 100 yes-or-no questions to arrive at a personality profile.

Hans Eysenck has also proposed a factor-analytic trait model of human personality. However, Eysenck's model focuses on the following three dominant dimensions that combine various related traits: psychoticism (characterized by various types of antisocial behavior ), introversion-extroversion, and emotionality/neuroticismstability. Eysenck has also combined the introversion-extroversion and emotionality-stability scales into a model containing four quadrants whose groupings of traits correspond roughly to the four types of personality outlined by the physician Hippocrates over 2,000 years ago in ancient Greecesanguine, choleric, phlegmatic, and melancholic.

Other trait-oriented theories include those of J.P. Guilford and David McClelland. Currently, a number of psychologists interested in a trait approach to personality believe that the following five factors, rather than Eysenck's three, are most useful in assessing personality: extroversion, agreeableness, conscientiousness, neuroticism, and openness to experience. A questionnaire called the NEO Personality Inventory , often called "the big five," has been developed to assess these factors.

Further Reading

Allport, Gordon W. Personality and Social Encounter: Selected Essays. Boston: Beacon Press, 1960.

Eysenck, Hans. The Structure of Human Personality. London Methuen, 1970.