Person. individ. D$ Vol. 13, No. 6, pp. 653465, 1992 0191-8869/92 $5.00 + 0.00 Printed in Great Britain Pergamon Press Ltd

FOUR WAYS FIVE FACTORS ARE BASIC

PAUL T. COSTA JR and ROBERT R. MCCRAE

Gerontology Research Center, National Institute on Aging, NIH, Baltimore, MD 21224, U.S.A.

(Received 12 August 1991)

Summary-The five-factor model has recently received wide attention as a comprehensive model of personality traits. The claim that these five factors represent basic dimensions of personality is based on four lines of reasoning and evidence: (a) longitudinal and cross-observer studies demonstrate that all five factors are enduring dispositions that are manifest in patterns of behavior; (b) traits related to each of the factors are found in a variety of personality systems and in the natural language of trait description; (c) the factors are found in different age, sex, race, and language groups, although they may be somewhat differently expressed in different cultures; and (d) evidence of heritability suggests that all have some biological basis. To clarify some remaining confusions about the five-factor model, the relation between Openness and psychometric intelligence is described, and problems in factor rotation are discussed.

INTRODUCTION

A recent symposium (Zuckerman, 1991) and article (Eysenck, 1991) addressed two central issues in personality psychology: What are the criteria by which we determine that a personality dimension is basic, and which dimensions of personality meet these criteria? There is perhaps more agreement on the former than on the latter question. In this article we will argue that basic dimensions must demonstrate their psychological reality through evidence of temporal stability and cross-observer validity; must pervade the trait systems of both laypersons and personality theorists; must recur in many different cultures; and must have some biological basis, although they need not be tied to any particular neuro- or psychophysiological theory. We will also summarize data showing that the dimensions of the five-factor model meet these criteria.

However, some clarifications are needed first. Although the terms jive-factor model and Big Five have become popular in the past few years, there are still some confusions about the model itself. As John (1989) pointed out, part of the confusion stems from the appearance that there are many different “Big Five”s. It is true that a number of investigators have used different instruments, labels, and conceptualizations, including Tupes and Christal (1961), Norman (1963), Digman and Inouye (1986), Goldberg (1990) Hogan (1986) Wiggins (Trapnell & Wiggins, 1990), and Costa and McCrae (1985). Despite differences in emphasis and interpretation, however, there is agreement among all these investigators that they are addressing the same phenomenon; there is also empirical evidence that their instruments show convergent validity (Borkenau & Ostendorf, 1990; Goldberg, 1990; McCrae & Costa, 1987; Trapnell & Wiggins, 1990). The five-factor model should not be identified with any of its operationalizations; it is an evolving scientific construct, not an instrument.

Critics of the model sometimes object that five factors cannot adequately account for the full range of personality traits (Mershon & Gorsuch, 1988), and of course they are correct. The term “Big Five” is perhaps misleading here, because it suggests that the model posits only five important traits. The alternative label, “the five-factor model,” may be more descriptive, because a factor matrix contains both rows and columns, both variables and factors. The variables in this model are specific personality traits. The instrument we have developed to measure the model, the Revised NE0 Personality Inventory (NEO-PI-R; Costa & McCrae, 1992), assesses 30 separate traits organized by the model into five domains: Neuroticism (N), Extraversion (E), Openness (0), Agreeableness (A), and Conscientiousness (C). The model helps us specify the range of traits that a comprehensive personality instrument should measure, and the factors that emerge from an analysis of these traits are what we consider the basic dimensions of personality.

653

654 PAUL T. COSTA JR and ROBERT R. MCCRAE

Table 1. Factor analysis of computer-administered Revised NE0 Personality Inventory facet scales

NEO-PI-R facet scale

Varimax-rotated principal component

N E 0 A C

Neuroticism facets Nl: Anxiety N2: Angry hostility N3: Depression N4: Self-consciousness N5: Impulsiveness N6: Vulnerability

Extraversion facets El: Warmth E2: Gregariousness E3: Assertiveness E4: Activity E5: Excitement seeking E6: Positive emotions

Openness facets 01: Fantasv 02: A&h&s 03: Feelings 04: Actions 05: Ideas 06: Values

Agreeableness facets Al: Trust A2: Straightforwardness A3: Altruism A4: Compliance A5: Modesty A6: Tender-mindedness

Conscientiousness facets Cl: Competence C2: Order C3: Dutifulness C4: Achievement striving C5: Self-discipline -. _... C6: Uehberatmn

0.80 -0.09 0.05 -0.01 -0.03 0.64 0.02 -0.04 -0.49 -0.02 0.77 -0.15 0.00 -0.04 -0.20 0.75 -0.22 -0.08 0.04 -0.07 0.48 0.47 0.11 -0.25 -0.15 0.67 -0.14 -0.16 0.11 - 0.40

-0.21 0.57 0.26 0.46 0.09 -0.22 0.69 -0.01 0.19 0.01 - 0.27 0.52 0.17 -0.31 0.43

0.00 0.49 0.18 -0.20 0.58 -0.12 0.63 0.10 -0.22 -0.05 -0.22 0.58 0.37 0.23 0.17

0.02 0.13 0.66 -0.24 -0.12 0.08 -0.08 0.68 0.23 0.02 0.21 0.18 0.68 0.01 0.26

-0.13 0.28 0.57 0.17 -0.07 -0.17 0.01 0.69 -0.13 0.08 -0.06 0.08 0.63 - 0.03 -0.10

-0.27 0.25 0.10 0.57 0.14 - 0.03 -0.11 -0.15 0.68 0.02 -0.14 0.26 0.06 0.63 0.34 -0.13 -0.20 -0.02 0.77 0.00

0.15 -0.23 -0.12 0.56 -0.13 0.17 0.11 0.18 0.66 0.08

-0.41 0.01 0.17 0.03 0.67 0.11 -0.09 -0.16 0.1 I 0.66

-0.17 -0.06 -0.14 0.32 0.67 -0.05 0.14 0.07 -0.09 0.77 -0.33 -0.07 -0.02 0.11 0.75 -0.19 -0.48 -0.01 0.21 0.40

N = 41 I adult me” and women. Loadings greater than f0.40 are shown in boldface.

As an illustration of the traits and their structure, Table 1 presents a factor analysis of NEO-PI-R facets in a sample of 411 men and women, aged 19 to 96, who completed the NEO-PI-R between 1989 and 1991. These subjects were participants in the Baltimore Longitudinal Study of Aging (BLSA; Shock, Greulich, Andres, Arenberg, Costa, Lakatta & Tobin, 1984) but their data were not used in any of the item selection studies for the NEO-PI-R. In Table 1, most facets have their largest loading on the intended factor, and the secondary loadings are similar to those seen in other analyses; a minor rotation of the E and C factors would provide an even better fit. These data provide an independent replication of the factor structure of the NEO-PI-R presented by Costa, McCrae and Dye (1991) and are of additional interest because they were administered by computer rather than by paper-and-pencil. It is reassuring (though not surprising) that the five factors are largely invariant across method of administration.

Another frequent criticism of the five-factor model is based on a more general skepticism about factor analysis itself as a way to understanding personality (Block, 1991). Certainly it is true that there is nothing magical about factor analysis; if it were an infallible approach, the questions addressed by this article would long since have been answered. But factor analysis is a useful tool for processing the quantities of data that are needed to understand something as complicated as individual differences in personality. It identifies clusters of variables that are related to each other and unrelated to other clusters of variables, and in this respect it systematizes the quest for those basic requirements of scientific constructs, convergent and discriminant validity. Used intelligently, it can yield valuable insights.

In any case, the test of whether factor analysis is useful does not lie in an analysis of the technique, but in an evaluation of the constructs it yields. Do the dimensions of the five-factor model make conceptual sense? Do they show convergence across different instruments and observers? Are they useful in predicting important social and psychological outcomes? We think so.

Four ways 655

FOUR CRITERIA FOR BASIC DIMENSIONS OF PERSONALITY

The reality of the factors

Surely no one would dispute that the basic dimensions of personality must refer to some objective psychological reality. But the belief that personality traits were nothing more than cognitive fictions was widespread only a few years ago. And ironically, although it was Mischel’s (1968) influential critique that popularized this idea, earlier research on the five-factor model provided the most powerful argument for it.

As shown by D’Andrade (1965) and by Passini and Norman (1966) the five-factor structure of personality could be recovered from ratings made on complete strangers. Because nothing was known about the actual traits of these ratees, and thus about the covariation of those traits, the clear implication of this finding was that the structure itself must somehow be built into our cognitive system of person perception, constituting an implicit personality theory. But this fact itself was subject to radically different interpretations. At one extreme, it was seen as evidence that traits were nothing but fictions, projections of our cognitive structure onto the interpersonal world. At the other, it was taken as an indication that individuals learned the true structure of personality from their observations of the real world and simply applied it to unknown cases. Borkenau (in press) has recently reviewed the evidence on these hypotheses and a number of more sophisticated variants, and concluded that “traits are real and accurately perceived, provided that the judges have the necessary information” (p. 2).

This conclusion is based in large part on evidence of the stability and cross-observer validity of the five factors. Traits are defined as enduring dispositions that can be inferred from patterns of behavior; they should therefore be stable across long periods of time and be similarly assessed by different observers. By these criteria, N, E, 0, A, and C are all indisputably real traits.

Because most questionnaires include scales measuring aspects of N and E, there are many longitudinal studies of personality in adulthood that demonstrate the stability of these two dimensions. For example, Finn (1986) reported uncorrected stability coefficients of 0.56 for MMPI measures of both N and Social E in a sample of 78 middle-aged men retested after 30 years. Using the NEO-PI, we found 6-yr retest correlations of 0.83, 0.82, and 0.83 for N, E, and 0, respectively, in self-reports from 398 men and women; similarly, we found 6-yr stability coefficients of 0.83, 0.77, and 0.80 for spouse ratings of N, E, and 0 on 167 of these individuals (Costa & McCrae, 1988b). A 7-yr longitudinal peer rating study of all five factors found stability coefficients of 0.67 for N, 0.81 for E, 0.84 for 0, 0.63 for A, and 0.78 for C (Costa & McCrae, in press). Corrected for unreliability, these values would approach unity.

All five factors have also been consensually validated in studies examining agreement across observers-a fact demonstrated as long ago as 1966 by Norman and Goldberg, and subsequently replicated by Funder and Colvin (1988) and others. Table 2 presents correlations across observers for factor scores from the NEO-PI-R. The first column gives intraclass correlations between peer raters; the second and third columns give correlations of self-reports with single peer and spouse ratings. All these correlations are statistically significant, and the median value of 0.50 shows that they are substantial in magnitude. Indeed, all of them meet or exceed the so-called 0.30 barrier for validity coefficients. Self-reports and various kinds of ratings are not interchangeable, and discrepancies between them may sometimes be of clinical interest (Muten, 1991) but the consistent convergences attest to the reality of these five factors.

The reality of the factors is also seen in their practical utility. 0 is an important predictor of vocational interests (Costa, McCrae & Holland, 1984) and C has recently emerged as the best

Table 2. Cross-observer correlations for NEO-PI-R factors

NEO-PI-R Factor Peer/Deer Peer/self Spouse/self

Neuroticism 0.35 0.33 0.61 Extraversion 0.41 0.43 0.58 Openness 0.50 0.55 0.67 Agreeableness 0.52 0.51 0.57 Conscientiousness 0.30 0.46 0.37

For intraclass correlations between peer raters, N = 193 pairs; for peer/self correlations, N = 250; for spouse/self correlations, N = 68. All correlations are significant at P < 0.001.

656 PAUL T. COSTA JR and ROBERT R. MCCRAE

overall predictor of job performance (Barrick & Mount, 1991). A and C are related to life satisfaction even after accounting for the major effects of E and N (McCrae & Costa, 1991). A is essential in understanding the literature in coronary-prone behavior (Costa, McCrae & Dembroski, 1989), C is a predictor of academic achievement (Digman & Takemoto-Chock, 1981), and knowledge of all five factors has proven to be useful in clinical psychology (Miller, 1991). The five-factor model has much to offer applied psychology.

The pervasiveness of the factors

If someone were to suggest that the only basic dimensions of personality were E and A, the two axes of the interpersonal circumplex (McCrae & Costa, 1989b), the proposal would be greeted with astonishment: HOW could one possibly omit N, so crucial to measures of both normal personality and psychopathology, to perceived stress and ways of coping, to psychological well-being, to somatic concerns, to low self-esteem? Surely we need N in our model as a basic dimension. Just as surely, we would argue, we need C and 0. The basic dimensions of personality are those which together summarize with maximal efficiency the covariation among all the traits in the personality sphere; they must account for all the major variables that have been studied by psychologists as well as those traits that are used by laypersons to characterize themselves and their acquaintances. All five factors are needed to do that.

There is relatively little dispute about the dimensions of E and N, which have long been recognized as the Big Two (Wiggins, 1968). We will therefore focus attention on 0, A, and C, pointing out the range of traits that would be lost if these three dimensions were omitted. In particular, we wish to correct the impression that A and C are narrow, first-order traits (Eysenck, 1991). As the facet scales in Table 1 show, A and C are broad, second-order factors defined by many specific traits.

The first source of evidence on the pervasiveness of the five factors comes from studies of trait terms in natural languages (John, Angleitner & Ostendorf, 1988). The tradition of research that stretches from Allport and Odbert (1936) to Ostendorf (1990) through Cattell (1946) Tupes and Christal (1961) and Goldberg (1981) was, of course, the origin of the five-factor model. The English language includes thousands of terms to describe aspects of personality, and analysis after analysis has found five similar factors (Goldberg, 1990).

It is true that the five factors are not equally well represented among trait adjectives. At one extreme, there are hundreds of terms relevant to the A factor, including at the low pole abrasive, abusive, acrimonious, aggressive, altercative, antagonistic, argumentative, arro- gant, autocratic, and avaricious, and at the high pole acceding, accommodating, acquiescent, affable, affectionate, altruistic, amiable, approachable, assistful, and, of course, agreeable (cf. Norman, 1967). At the other extreme, there are relatively few adjectives that describe 0 (McCrae, 1990) and most of them, like curious, creative, inquisitive, and intellectual refer only to the more cognitive forms of 0, leading many lexical researchers to call this factor

“Intellect.” Some psychologists might object that lay vocabularies are not the best source for scientific

descriptions of personality. More psychologically informed descriptions are provided by the items in Block’s (1961) California Q-Set. Table 3 lists items that define the 0, A, and C factors found in this instrument (McCrae, Costa & Busch, 1986).

Scales related to these three factors can also be found in instruments designed to operationalize many of the classic theories of personality. Henry Murray’s (1938) catalog of needs includes Understanding, Change, Sentience, and Autonomy, all related to 0; Abasement, Nurturance, and low Aggression, all related to A; and Achievement, Order, Endurance, and Cognitive Structure, all related to C (Costa & McCrae, 1988a). C. G. Jung’s (1923/1971) psychological functions are the basis of the Myers-Briggs Type Indicator; correlations of its scales with measures of the five-factor model show that Sensing vs Intuition, Thinking vs Feeling, and Perceiving vs Judging are related to 0, A, and C, respectively (McCrae & Costa, 1989a). Angleitner and Ostendorf (199 1) recently recovered the five factors in an analysis of temperament scales. There the 0 factor was defined in part by Zuckerman’s (1979) Experience Seeking scale, the A factor included Strelau, Angleitner, Bantelmann and Ruth’s (1990) Inhibition scale, and the C factor was marked by Windel and Lerner’s (1986) Persistence scale.

Four ways

Table 3. Some item definers of 0, A, and C factors in the California O-Set

openness Aesthetically reactive Values intellectual matters Wide range of interests Rebellious, non-conforming

-VS-

Sex-role stereotyped behavior Favors conservative values Uncomfortable with complexities Judges in conventional terms

Agreeableness Sympathetic, considerate Arouses liking Warm, compassionate Behaves in giving way

-VS-

Expresses hostility directly Basically distrustful Shows condescending behavior Critical, skeptical

Conscientiousness Productive Behaves ethically Has high aspiration level Dependable, responsible

-vs- Self-indulgent Interested in opposite sex Enjoys sensuous experiences Unable to delay gratification

Adapted from McCrae er al. (1986).

Aspects of 0 are seen in Tellegen’s (1982) absorption; Gough’s (1987) flexibility, Rokeach’s (1960) dogmatism, Fenigstein, Scheier and Buss’s (1975) private self-consciousness, Holland’s (1985) artistic interests, Guilford, Zimmerman and Guilford’s (1976) thoughtfulness, Kris’s (1952) regression in the service of the ego. Aspects of A are seen in Erikson’s (1950) trust vs mistrust, Snyder’s (1974) self-monitoring, James’ (1907) tough vs tender-mindedness, Horney’s (1945) “moving against” tendency, Freud’s (19 14/ 1957) narcissism, Adler’s (1938/ 1964) social interest. Aspects of C are seen in White’s (1959) competence, Cattell, Eber and Tatsuoka’s (1970) superego strength, Rotter’s (1966) locus of control, Hartshorn, May and Maller’s (1929) character, Lorr’s (1986) persistence, and McClelland, Atkinson, Clark and Lowell’s (1953) achievement motive. 0, A, and C-like N and E-are basic themes that have recurred in innumerable forms throughout the history of personality psychology. Each of them is indispensable.

The universality of the factors

Truly basic dimensions of personality ought to be universal-found in both sexes, in all races, in various age groups, in different cultures. Whether such dimensions exist is an empirical question. There is no guarantee that thoughts, feelings, and actions must covary in the same way in all different groups. For example, it is unlikely that the same personality factors found in adults would also be found in infants, and it would surprise few social psychologists if the factors found in modern industrial societies differed from those found in hunting and gathering cultures.

However, if any personality factors can be found that do cross all these boundaries, they would surely have more claim to being basic than would factors that are only found in specialized populations. Each of the dimensions of the five-factor model has evidence of universality.

The factors have been found in teachers’ ratings of children (Digman & Inouye, 1986), in college students (Goldberg, 1990), in adults (see for example Table 1). In a sample of 1539 individuals tested as part of a study of job performance (Costa et al., 1991), we examined factors of the NEO-PI-R in men and women, in older and younger adults, and in white and non-white subjects. In each case, five very similar factors were found, with coefficients of congruence for the 15 comparisons ranging from 0.91 to 0.99.

Beginning with Bond’s work in the 197Os, a number of cross-cultural studies on the five-factor model have been conducted, many using translations of Norman’s (1963) instrument (Bond, 1979; Bond, Nakazato & Shiraishi, 1975; Borkenau & Ostendorf, 1990; Yang & Bond, 1990). Most of

658 PAUL T. COSTA JR and ROBERT R. MCCRAE

these studies have recovered factors very similar to those reported by Norman. For example, Yang and Bond factored 718 ratings of father made by Chinese college students. Of the 20 Norman scales, they found 19 with their highest loading on the intended factor.

Borkenau and Ostendorf (1990) provided a more formal analysis of the five-factor model, using both the Norman rating scales and their German translation of the NEO-PI. When separate results for men and women were compared using Everett’s (1983) procedure, five and only five factors were replicable in both instruments, and confirmatory factor analyses showed that these were the factors hypothesized.

Currently there is considerable interest in the question of whether a different factor structure might be found if more culture-specific variables were included. Ostendorf (1990) reported analyses of the German language lexicon that provide an almost perfect replication of the structure in English. Yang and Bond (1990) sought to examine the structure of personality using representative Chinese trait adjectives; they found five factors, but they were not identical to the standard five. Church and Katigbak (1989) used items generated by Filipino students in both English and Tagalog, and found factors they interpreted in terms of the Big Five. Yang and Bond chose to emphasize the cultural differences; Church and Katigbak seemed more impressed by the similarities. Clearly, this is an issue that will require more research.

Some of that research is currently being conducted by Michael Bond, who, with Wai-kwan Li, has translated the NEO-PI-R into Chinese. Fung Yi Liu (1991) administered this Chinese NEO-PI-R to a sample of 100 college students; she also administered a set of items intended to measure distinctively Chinese characteristics organized into three scales. When the 30 NEO-PI-R scales were factored, the five factors were recovered, with 26 facets having their highest loading on the intended factor. Of more interest is the fact that a joint analysis showed that the three indigenous Chinese scales also fit within the five-factor solution. For example, one of the Chinese scales was Filial Piety, a variable measuring deference to parents. It loaded on the A factor.

Perhaps what these cross-cultural studies really show is that the ways in which the basic factors of personality are manifested differ somewhat from culture to culture. Among American college students, disregard for one’s parents’ wishes is more or less the norm, and says relatively little about personality. Among Chinese students, however, it seems that only highly antagonistic individuals lack filial piety.

The biological basis of the factors

The fact that the five factors are found in many different cultures suggests that they are basic features of human nature itself. This does not necessarily mean that they are biologically based, because human beings share more than a common biology. They are, for example, all social beings with the capacity for abstract thought through symbolization; the five factors may represent alternative ways in which people in a social environment can respond to their life experience.

But as it happens there are good reasons to suspect that all five factors have some biological foundation, because measures of all five have shown evidence of heritability. This should be reassuring to those who prefer to “select personality variables that are genotypically rooted in our biology as much as possible” (Caspi & Bern, 1990, p. 556). There is a large body of research on genetic influences on N and E (Eaves, Eysenck & Martin, 1989), and the few studies that have examined the heritability of 0 have also shown consistent genetic influences. Using a short Swedish version of the NEO-PI, Plomin and McClearn (1990) reported that 41% of the variance in the 0 scale was attributable to additive genetic influences. Tellegen and Waller (in press) reported substantial heritabilities for two Multidimensional Personality Questionnaire (MPQ) scales related to 0, Absorption and (low) Harmavoidance. Loehlin’s (1987) cross-pair analysis of California Psychological Inventory items recovered a genetic factor that included “the pursuit of culture- lectures, discussion, opera-and a willingness to tackle uncertain and ambiguous problems” (p. 142) that Loehlin identified with Norman’s fifth factor, Culture, and that we would call 0.

Eysenck (1990) has also reviewed a number of studies that establish genetic influences on his third factor, Psychoticism. From the perspective of the five-factor model, Psychoticism is a combination of low A and C (Goldberg, 1991; McCrae & Costa, 1985a), so Eysenck’s data suggest heritability for A and C. There is also more direct evidence: The MPQ Aggression scale (Tellegen & Waller, in press) and measures of altruism, empathy, nurturance, and aggressiveness (Rushton,

Four ways 659

Fulker, Neale, Nias & Eysenck, 1986) reflect aspects of A, and all show substantial heritability; the MPQ Achievement and Control scales, which are related to C, both show genetic influences, as did a brief Swedish version of the NEO-PI C scale (Plomin & McClearn, 1990).

It seems likely that there are neurobiological structures that underlie such heritable personality traits, and Tooby and Cosmides (1990) argued that shared neurophysiological systems may account for the covariation of specific traits into broad factors. Cloninger (1988) presented a theory of personality structure based on such a principle, and some, like Claridge (1986), have argued that no personality dimension can be taken seriously unless it is supported by theory linking it to a biological mechanism.

We believe that this latter view is profoundly mistaken. The fact is that we know much more about personality structure than we do about the functioning of the brain, and it is poor science to try to explain the known on the basis of the unknown. Consider for a moment the far-reaching developments in our understanding of neurophysiology that have occurred since 1961. In retrospect, it would have been folly for Tupes and Christal to attempt to explain their five factors in terms of the comparatively primitive neuroscience of the day. Their factors have survived the past 30 years very well; any biological explanation they might have proposed would surely be hopelessly outdated. Will today’s neurobiological explanations fare any better?

We do not mean to suggest that studies of personality and psychophysiology are unimportant; they may teach us much about both individual differences and neuroscience. But it seems to us that such studies should begin with our current understanding of the organization of personality into five factors. Every psychophysiological study should include measures of all five; empirical links between these factors and neurophysiological variables, even if they are not anticipated by theory, may give valuable clues to the biological basis of personality.

SOME RESIDUAL ISSUES

In view of the accumulated evidence in its favor, many personality psychologists have recently adopted the five-factor model (McCrae, 1992). Eysenck (1991) noted that in 1983 Kline and Barrett concluded from their review that the three-factor model of Eysenck was the most adequate representation of personality structure; yet by 1991, Kline and Lapham had come to believe that the Big Five provided “probably the best account of ratings in personality” (p. 631). Again, Eysenck pointed out that in 1990 Angleitner believed that only four of the factors-(N, E, A, and C) had a temperamental basis; in 1991 Angleitner and Ostendorf reported a joint analysis of five-factor and temperament scales which showed that all five were necessary to account for aspects of temperament.

But others are not yet persuaded that the five-factor model best represents the basic dimensions of personality. Eysenck (1990, 1991) has argued that his three-factor model accounts for the most important dimensions of personality; and Zuckerman, Kuhlman, Thornquist and Kiers (1991) have proposed a replicable five-factor model which appears to be substantially different from the standard model. Two different problems appear to account for these differences of opinion: confusions of 0 and intelligence, and differences in factor rotation.

Openness, culture, intellect, and IQ

The fifth factor in Norman’s (1963) influential study of peer ratings was labeled “Culture,” and the most intense disputes among five-factor theorists have concerned the proper conceptualization and name for this factor (e.g. Digman, 1990; McCrae & Costa, 1985c, in press; Peabody, 1987). Most now agree that Culture is not an appropriate name for the factor, because it suggests an acquired sophistication that is not central to the factor as it is typically found in analyses of adjectives or scales. It is important to note, however, that there are empirical links between Norman’s fifth factor and 0 (Borkenau & Ostendorf, 1990; McCrae & Costa, 1985b).

A more viable alternative name for this factor is Intellect (Goldberg, 1990). In lexical studies, the fifth and smallest of the factors is chiefly defined by terms like curious and perceptive, and is frequently labeled Intellect. However, we feel this label is misleading for three reasons. First, it ignores the fact that many intellect-like terms (e.g. logical, foresighted; see John, 1990) typically load on the C factor. Second, it suggests an unduly narrow interpretation of the 0 factor. Although

open individuals are indeed intellectually curious, they are also aesthetically sensitive, high in need for variety, and liberal in their value systems. Third, the label Intellect strongly suggests that this factor is interchangeable with g, psychometric intelligence. If 0 were nothing more than a reflection of intelligence, there would indeed be little reason to regard it as a basic dimension of personality,

as Eysenck (1991) argued. In fact. however, the correlation between 0 and IQ. though positive and significant, is

consistently quite small, and joint factor analyses show separate 0 and Intelligence factors (McCrae & Costa, 1985b, c). The relations among 0, Intellect, C. and psychometrically-assessed intelligence are illustrated in Fig. 1 (McCrae & Costa, in press).

Because the relative independence of 0 from IQ is of such importance, it may be worthwhile to present additional data on the question. A few years ago a subset of participants in the BLSA were administered the complete Revised Wechsler Adult Intelligence Scales (WAIS-R); they had also completed the NEO-PI. Complete data were available for 36 men. aged 19 to 86, and 31 women, aged 17 to 83. Correlations of the 1 I ago-corrected WAIS-R scales with the NEO-PI 0 domain scale showed significant correlations with only six scales, and no correlation

greater than 0.35. Correlation of the full WAIS score with 0 was only 0.33. Given the high reliability of both instruments. it is clear that almost all of the reliable variance in 0 is unrelated

to intelligence. It is also noteworthy that the two WAIS-R scales that showed the highest correlations with 0

were Block Design and Object Assembly, both of which “seem to get at some sort of creative ability” (Wechsler, 1958, p. 80). This is consistent with the finding that 0 is related to divergent thinking abilities and to scales intended to lneasure creativity (McCrae. 1987). To use Welsh’s (1975) terms, it is Origence, not Intellect~nce. to which 0 more closely corresponds, and this is surely a basic part of personality.

There are three crucial steps in factor analysis: the choice of variables; the selection of the proper

number of factors to retain; and the method of rotation. Digman and Inouye (1986) have observed that “If a large number of scales is used and if the scope of the scales is very broad, the domain of personality descriptors is almost completely accounted for by five robust factors” (p. 116). Recently. Zuckerman er ul. (1991) supported this generalization in their analyses of a set of 33 personality scales: a five-factor solution proved to be most replicable across samples. But although

three of the factors-Sociability, Neuroticism-Anxiety, and Aggression-Hostility-appeared to

MEASURED INTELLIGENCE

CONSCIENTIOUSNE

Fig. 1. Schematic representation of relations among 0 IO Experience. Meusured Intelligence. C. and

Intellect. Adapted from McCrac and Costa (111 press).

Four ways 661

correspond well to E, N, and low A, the two remaining factors, Impulsive-Unsocialized Sensation Seeking and Activity, did not seem equivalent to 0 or C.

These authors explained the differences in structure by pointing out that markers for the dimension of Culture or 0 were deliberately omitted-perhaps because of the confusion of 0 with intelligence. But if this factor had really been omitted, why did five robust factors emerge instead of the four we might expect?

We suspect the problem is one of rotation as well as variable selection. We have argued that each of the five factors is pervasive, appearing in many different guises; and although Zuckerman and colleagues may have intended to omit 0, it is by no means clear that they succeeded. One of their variables, for example, is the Experience Seeking scale, which is known to correlate with NEO-PI 0 (McCrae & Costa, in press).

Even when all five factors are represented in a set of variables, it is not necessarily the case that recognizable factors will be found. Although simple structure has been the guiding principle in factor analysis for decades, we know that personality traits do not necessarily conform to it. The interpersonal circumplex (Wiggins, 1979) is only the most famous illustration of the fact that many important personality traits are defined by two or more factors; the secondary loadings of NEO-PI-R facet scales on the factors in Table 1 provide another example. And because traits can be found almost anywhere in a five-dimensional space, the axes chosen by a varimax rotation will depend completely on the selection of variables. That is why it is usually advisable to include known markers of the basic dimensions of personality in factor analyses of diverse variable sets: The marker scales not only identify the obtained factors, they also guide the varimax rotation (Costa & McCrae, 1988a; Angleitner & Ostendorf, 1991).

Fortunately, we do not need to rely slavishly on varimax rotation. All orthogonal rotations are mathematically equivalent, showing the same relations between variables from different perspec- tives. If we wish to see how Zuckerman et al.‘s data might have looked if strong markers of all five factors had been included, we can choose surrogate markers within the data set and rotate the factors to a target defined by these variables (Schonemann, 1966). In this case, N, E, 0, A, and C can be represented by EPQ N, EPQ E, Experience Seeking, low PRF Aggression, and PRF Cognitive Structure-all known markers of these factors.

Table 4 presents the results of an orthogonal Procrustes rotation of the Zuckerman et al. (1991) five-factor solution. The N factor is almost identical to the original N-anxiety factor and, like NEO-PI-R N, also includes elements of angry hostility. The E factor contains most of the markers of the originally Sociability factor, but it also includes scales measuring monotony avoidance, activity, and energy level. It thus resembles the NEO-PI-R E domain, with its Warmth, Gregariousness, Excitement Seeking, and Activity facets. The A factor is similar to the reflected Aggression-Hostility factor, and the C factor is defined by variables measuring cognitive structure, inhibition, responsibility, and persistence. As in the NEO-PI-R, scales measuring activity have secondary loadings on the C factor.

The factor intended to represent 0 appears to be a rather rough approximation to the standard factor. Although the Experience Seeking scale was used to target the factor and does show a sub- stantial loading on it, the highest loading variable is PRF Autonomy, which reflects independence of judgment and was a definer of the 0 factor in a joint analysis of the PRF and NEO-PI (Costa & McCrae, 1988a). CPI Socialization is another definer of this factor. That is understandable when it is recalled that open individuals can be rebellious and non-conforming (see Table 3); CPI Socialization is known to be negatively related to 0 (McCrae, Costa & Piedmont, in press). The peculiarly rebellious cast that this factor has may explain the loading of EPQ P, which is otherwise essentially uncorrelated with 0 (McCrae & Costa, 1985a). Given the intentional omission of vari- ables related to culture and 0, some variation from the ideal pattern of loadings is to be expected. How well this factor would correlate with more orthodox versions of 0 is an empirical question.

The Zuckerman et al. study thus can be seen as a confirmation of the standard five-factor model. But if rotation is arbitrary, could we not equally well assert that the factors found in natural language are merely rotational variants of the Zuckerman et al. factors? Which rotation will lead us to the truly basic dimensions of personality?

Consider the analogy of a map. A map of the British Isles may with perfect fidelity to the geographical facts represent Ireland to the right of Britain-provided that the top of the map points

662 PAUL T. COSTA JR and ROBERT R. MCCRAE

Table 4. Orthogonal m-rotation of Zuckerman et al.‘s (1991) five factor solution

Original Re-Rotated Factor Factor/ variable N E O? A C

Impulsive Unsocialized Sensation Seeking PRF Cognitive Structure SSS Experience Seeking KSP Impulsivity EASI Decision Time (quick) KSP Monotony Avoidance EPQ Psychoticism CPI Socialization SSS Disinhibition SSS Thrill and Adventure Seeking SSS Boredom Susceptibility

Neuroticism-Anxiety KSP Anxiety EPQ Neuroticism EASI Fear EASI General Emotionality KSP Psychasthenia JPI Anxiety EASI Inhibitory Control (lack of)

Aggression-Hostility (vs Social Desirability) PRF Aggression KSP Hostility KSP Social Desirability EASI Anger JPI Responsibility EPQ Lie KSP Inhibition of Aggression

Sociability PRF Affiliation JPI Social Participation EASI Sociability KSP Detachment PRF Autonomy EPQ Extraversion

Activity EASI Activity JPI Energy Level EASI Persistence (lack of)

0.06 -0.21 -0.44 -0.07 0.62’ 0.02 0.14 0.52a -0.02 -0.44 0.00 0.43 0.38 -0.09 - 0.43

-0.13 0.35 0.41 -0.06 -0.44 0.03 0.54 0.43 -0.17 -0.30 0.11 0.05 0.49 -0.38 -0.43

-0.22 0.02 -0.56 0.34 0.26 0.07 0.32 0.10 -0.27 -0.57

-0.19 0.37 0.37 -0.04 -0.20 0.02 0.17 0.34 -0.33 -0.26

0.83 -0.24 0.10 0.02 -0.03 0.81” -0.15 0.02 -0.21 0.06 0.77 -0.20 -0.02 0.13 0.05 0.77 0.15 0.07 -0.19 0.15 0.71 -0.24 ~ 0.08 -0.08 0.03 0.69 0.05 -0.16 -0.21 0.19 0.49 0.21 0.20 -0.38 -0.16

0.21 0.09 0.02 - 0.7Sa 0.07 0.50 0.06 0.00 -0.67 -0.01

-0.24 0.07 0.11 0.67 0.13 0.42 0.19 0.10 - 0.63 0.18

-0.06 0.02 0.06 0.59 0.40 -0.17 -0.06 0.11 0.53 0.34

0.42 -0.31 -0.11 0.53 -0.01

-0.01 0.63 - 0.42 0.30 -0.25 0.02 0.60 -0.32 0.17 -0.30

-0.21 0.72 -0.19 - 0.02 -0.15 0.07 -0.52 0.28 -0.24 0.12

-0.17 -0.22 0.70 -0.20 - 0.09 -0.15 0.778 0.14 -0.09 -0.21

0.03 0.58 0.35 -0.08 0.40 -0.36 0.44 0.33 0.08 0.39

0.12 -0.27 0.09 0.03 -0.58

Original data are reported in Zuckerman et al. (1991). Loadings over kO.40 are given in boldface. CPI = California Psychological Inventory. EASI = Emotionality-Activity-Sociability-Impulsivity scales. EPQ = Eysenck Personality Questionnaire. JPI = Jackson Personality Inventory. KSP = Karolinska Scales of Personality. PRF = Personality Research Form. SSS = Sensation Seeking Scales.

“Target variable toward which factor was rotated.

South. We may choose to put any direction we wish at the top of our map, but if we want to see how Britain is situated with respect to France, Germany, and Norway, it would be helpful if all cartographers adopted the convention of putting North at the top of the map. In the same way, it would be helpful to personality researchers if factor analysts oriented their factors with respect to the standard five, either through the inclusion of markers in the battery, or through rotations guided by variables of known factorial composition.

In the case of geography, the placement of North rather than, say, East-by-Southeast, at the top of the map is not entirely arbitrary; it is determined by the ease with which stars and the compass can be used to locate North. Personality structure is not that simple, but the frequency with which the five factors of N, E, 0, A, and C emerge in analyses of natural languages (Ostendorf, 1990) and joint analyses of diverse personality questionnaires (Boyle, 1989) suggests that these may indeed be the best orientations within the five-dimensional sphere of personality.

Acknowledgements-Thanks are due to Dr Marvin Zuckerman for permission to present his reanalyzed data. Portions of this article were presented at the Meeting of the International Society for the Study of Individual Differences, Oxford, England, July 1991.

REFERENCES

Adler, A. (1964). Social interest: A challenge to mankind. New York: Capricorn Books. Allport, G. W. & Odbert, H. S. (1936). Trait names: A psycho-lexical study. Psychological Monographs, 47(211). Angleitner, A. (1990). Personality psychology: Trends and developments. Paper presented at the 5th Congress of European

Personality Psychology, Ariccia, Italy, June.

Four ways 663

Angleitner, A. & Ostendorf, F. (1991). Temperament and the Big Five factors of personality. Paper presented at the Conference on the Development of the Structure of Temperament and Personality from Infancy to Adulthood, Wassenaar, The Netherlands, June.

Barrick, M. R. & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1-26.

Block, J. (1961). The Q-sort method in personality assessment andpsychiatric research. Springfield, IL: Charles C Thomas. Block, J. (1991). My uneasiness with the five-factor model of personality. Paper presented at the Conference on the

Development of the Structure of Temperament and Personality from Infancy to Adulthood, Wassenaar, The Netherlands, June.

Bond, M. H. (1979). Dimensions of personality used in perceiving peers: Cross-cultural comparisons of Hong Kong, Japanese, American, and Filipino university students. International Journal of Psychology, 14, 47-56.

Bond, M. H., Nakazato, H. & Shiraishi, D. (1975). Universality and distinctiveness in dimensions of Japanese person perception. Journal of Cross-Cultural Psychology, 6, 346-357.

Borkenau, P. (in press). Implicit personality theory and the five-factor model. Journal of Personality. Borkenau, P. & Ostendorf, F. (1990). Comparing exploratory and confirmatory factor analysis: A study on the 5-factor

model of personality. Personality and Individual Dtjerences, II, 5 15-524. Boyle, G. J. (1989). Re-examination of the major personality-type factors in the Cattell, Comrey and Eysenck scales: Were

the factor solutions by Noller et al. optimal? Personality and Individual Dtfirences, IO, 128991299. Caspi, A. & Bern, D. J. (1990). Personality continuity and change across the life course. In Pervin, L. A. (Ed.), Handbook

of personality: theory and research (pp. 549-575). New York: Guilford. Cattell, R. B. (1946). The description and measurement of personality. Yonkers, NY: World Book. Cattell, R. B., Eber, H. W. & Tatsuoka, M. M. (1970). The handbook for the Sixteen Personality Factor Questionnaire.

Champaign, IL: Institute for Personality and Ability Testing. Church, T. A. & Katigbak, M. S. (1989). Internal, external, and self-report structure ofpersonality in a non-Western culture:

An investigation of cross-language and cross-cultural generalizability. Journal of Personality and Social Psychology, 57, 8577872.

Claridge, G. (1986). Eysenck’s contribution to the psychology of personality. In Modgil, S. & Modgil, C. (Eds), Hans Eysenck: consensus and controversy (pp. 73-85). London: Falmer Press,

Cloninger, C. R. (1988). A unified biosocial theory of personality and its role in the development of anxiety states: A reply to commentaries. Psychiatric Development, 2, 833120.

Costa, P. T., Jr & McCrae, R. R. (1985). The NE0 Personality Inventory manual. Odessa, FL: Psychological Assessment Resources.

Costa, P. T., Jr & McCrae, R. R. (1988a). From catalog to classification: Murray’s needs and the five-factor model. Journal of Personality and Social Psychology, 5.5, 258-265.

Costa, P. T., Jr & McCrae, R. R. (1988b). Personality in adulthood: A six-year longitudinal study of self-reports and spouse ratings on the NE0 Personality Inventory. Journal of Personality and Social Psychology. 54. 853-863.

Costa, P. T., Jr & McCrae, R. R. (1992). Revised NE0 Personality Inventory (NEO-PI-R) and NE0 Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL: Psvcholoaical Assessment Resources.

Costa, P. T., Jr a McCrae, R. R. (in press). Trait psychology comes of age. In Sonderegger, T. B. (Ed.), Nebraska Symposium on motivation: Psychology and aging. Lincoln, NE: University of Nebraska Press.

Costa, P. T., Jr, McCrae, R. R. & Dembroski, T. M. (1989). Agreeableness vs. antagonism: Explication of a potential risk factor for CHD. In Siegman, A. & Dembroski, T. M. (Eds), In search of coronary-prone behavior: Beyond Type A (pp. 4163). Hillsdale, NJ: Lawrence Erlbaum Associates.

Costa, P. T., Jr, McCrae, R. R. & Dye, D. A. (1991). Facet scales for Agreeableness and Conscientiousness: A revision of the NE0 Personality Inventory. Personality and Individual Differences, 12, 887-898.

Costa, P. T., Jr, McCrae, R. R. & Holland, J. L. (1984). Personality and vocational interests in an adult sample. Journal of Applied Psychology, 69, 390-400.

D’Andrade, R. B. (1965). Trait psychology and componential analysis. American Anthropologist, 67, 215-228. Digman, J. M. (1990). Personality structure: Emergence of the five-factor model. Annual Review of Psychology, 41,417440. Digman, J. M. & Inouye, J. (1986). Further specification of the five robust factors of personality. Journal of Personality

and Social Psychology, SO, 116-123. Digman, J. M. & Takemoto-Chock, N. K. (1981). Factors in the natural language of personality: Re-analysis, comparison,

and interpretation of six major studies. Multivariate Behavioral Research, 16, 1499170. Eaves, L. J., Eysenck, H. J. & Martin, N. G. (1989). Genes, culture, and personality: An empirical approach. New York:

Academic Press. Erikson, E. H. (1950). Childhood and society. New York: Norton. Everett, J. E. (1983). Factor comparability as a means of determining the number of factors and their rotation. Multivariate

Behavioral Research. 18. 197-2 18. Eysenck, H. J. (1990). Genetic and environmental contributions to individual differences: The three major dimensions of

personality. In Buss, D. M. (Ed.), Biological foundations of oersonalitv: Evolution. behavioral eenetics. and psychopohysiology [Special issue]. Journal of Personality, 58, 2455262. _

Eysenck, H. J. (1991). Dimensions of personality: 16,5, or 3?-Criteria for a taxonomic paradigm. Personality and Individual Differences, 12, 773-790.

Fenigstein, A., Scheier, M. F. & Buss, A. H. (1975). Public and private self-consciousness: Assessment and theory. Journal of Consulting and Clinical Psychology, 43, 522-528.

Finn, S. E. (1986). Stability of personality self-ratings over 30 years: Evidence for an age/cohort interaction. Journal of Personality and Social Psychology, 50, 8 13-8 18.

Freud, S. (1957). On narcissism: An introduction. In Strachey, J. (Ed.), The standard edition of the complete psychological works of Sigmund Freud (Vol. 19, pp. 73-102). London: Hogarth Press. (Original work published 1914.)

Funder, D. C. & Colvin, C. R. (1988). Friends and strangers: Acquaintanceship, agreement, and the accuracy of personality judgment. Journal of Personality and Social Psychology, 55, 149-158.

Fung Yi Liu. (1991). The generahzability of the NE0 Personality Inventory to an university sample in Hong Kong. Unpublished manuscript, Chinese University of Hong Kong.

664 PAUL T. COSTA JR and ROBERT R. MCCRAE

Goldberg, L. R. (1981). Language and individual differences: The search for universals in personality lexicons. In Wheeler, L. (Ed.), Review of personality and social psychology (Vol. 2, pp. 141-165). Beverly Hills, CA: Sage.

Goldberg, L. R. (1990). An alternative “description of personality”: The Big-Five factor structure. Journal of Personality and Social Psychology, 59, 121&1229.

Goldberg, L. R. (1991). Comparing the Big Five factor structure with its competitors: I. Eysenck’s P-E-N model. Paper presented at the Conference on the Development of the Structure of Temperament and Personality from Infancy to Adulthood, Wassenaar, The Netherlands, June.

Gough, H. G. (1987). California Psychological Invenfory administrator’s guide. Palo Alto, CA: Consulting Psychologists Press.

Guilford, J. S., Zimmerman, W. S. & Guilford, J. P. (1976). The Guilford-Zimmerman Temperament Survey Handbook: twenty-jive years of research and application. San Diego, CA: EdlTS Publishers.

Hartshorn, H., May, M. A. & Maller, J. B. (1929). Studies in the nature of character, Vol. 2: studies in service andself-control. New York: Macmillan.

Hogan, R. (1986). Hogan Personality Inventory manuar. Minneapolis, NM: National Computer Systems. Holland, J. L. (1985). Self-Directed Search-I985 edition. Odessa, FL: Psychological Assessment Resources. Horney, K. (1945). Our inner conflicts. New York: W. W. Norton. James, W. (1907). Pragmatism: a new name for some old ways of thinking. London: Longmans Green. John, 0. P. (1989). Towards a taxonomy of personality descriptors. In Buss, D. M. & Cantor, N. (Eds), Personality

psychology: Recent trends and emerging directions (pp. 261-271). New York: Springer-Verlag. John, 0. P. (1990). The “Big Five” factor taxonomy: Dimensions of personality in the natural language and in

questionnaires. In Pervin, L. (Ed.), Handbook of personality theory and research. New York: Guilford. John, 0. P., Angleitner, A. & Ostendorf, F. (1988). The lexical approach to personality: A historical review of trait

taxonomic research. European Journal of Personality, 2, 171-203. Jung, C. G. (1971). Psychological types (Baynes, H. G. Trans., revised by Hull, R. F. C.). Princeton, NJ: Princeton

University Press. (Original work published 1923). Kline, P. & Barrett, P. (1983). The factors in personality questionnaires among normal subjects. Advances in Behaciour

Research and Therapy, S, 141-202. Kline, P. & Lapham, S. (1991). The validity of the PPQ: A study of its factor structure and its relationship to the EPQ.

Personality and Individual Differences, 12, 63 1636. Kris, E. (1952). Psychoanalytic explorations in art. New York: International Universities. Loehlin, J. C. (1987). Heredity, environment, and the structure of the California Psychological Inventory. Multioariate

Behavioral Research, 22, 1377148. Lorr, M. (1986). Interpersonal Style Inventory (ISI) manual. Los Angeles: Western Psychological Services. McClelland, D. C., Atkinson, J. W., Clark, R. A. & Lowell, E. L. (1953). The achieaement motioe. New York:

Appleton-Century-Crofts. McCrae, R. R. (1987). Creativity, divergent thinking, and openness to experience. Journal of Personality and Social

Psychology, 52, 125881265. McCrae, R. R. (1990). Controlling neuroticism in the measurement of stress. Stress Medicine, 6, 2377241. McCrae, R. R. (Ed.) (1992). The Five-Factor Model: Issues and applications [Special issue]. Journal of Personality, 60(2). McCrae, R. R. & Costa, P. T., Jr (1985a). Comparison of EPI and Psychoticism scales with measures of the five-factor

model of personality. Personality and Individual Differences, 6, 587-597. McCrae, R. R. & Costa, P. T., Jr (1985b). Openness to experience. In Hogan, R. & Jones, W. H. (Eds), Perspectioes in

personality (Vol. 1, pp. 145-172). Greenwich, CT: JAI Press. McCrae, R. R. & Costa, P. T., Jr (1985~). Updating Norman’s “adequate taxonomy”: Intelligence and personality

dimensions in natural language and in questionnaires. Journal of Personality and Social Psychology, 49, 7 lo-72 1. McCrae, R. R. & Costa, P. T., Jr (1987). Validation of the five-factor model of personality across instruments and observers.

Journal of Personality and Social Psychology, 52, 81-90. McCrae, R. R. & Costa, P. T., Jr (1989a). Reinterpreting the Myers-Briggs Type Indicator from the perspective of the

five-factor model of personality. Journal of Personality, 57, 1740. McCrae, R. R. & Costa, P. T., Jr (1989b). The structure of interpersonal traits: Wiggins’s circumplex and the five-factor

model. Journal of Personality and Social Psychology, 56, 586595. McCrae, R. R. & Costa, P. T., Jr (1991). Adding Liebe und Arbeit: The full five-factor model and well-being. Personality

and Social Psychology Bulletin, I7, 227-232. McCrae, R. R. & Costa, P. T., Jr (in press). Conceptions and correlates of Openness to Experience. In Briggs, S. R., Hogan,

R. & Jones, W. H. (Eds), Handbook of personality psychology. New York: Academic Press. McCrae, R. R., Costa, P. T., Jr & Busch, C. M. (1986). Evaluating comprehensiveness in personality systems: The California

Q-Set and the five-factor model. Journal of Personality, 54, 430446. McCrae, R. R., Costa, P. T., Jr & Piedmont, R. L. (in press). Folk concepts, natural language, and psychological constructs:

The California Psychological Inventory and the five-factor model. Journal of Personality. Mershon, B. & Gorsuch, R. L. (1988). Number of factors in the personality sphere: Does increase in factors increase

predictability of real-life criteria? Journal of Personality and Social Psychology, 55, 675480. Miller, T. (1991). The psychotherapeutic utility of the five-factor model of personality: A clinician’s experience. Journal of

Personality Assessment, 57, 415433. Mischel, W. (1968). Personality and assessment. New York: Wiley. Murray, H. A. (1938). Explorafions in personality. New York: Oxford University Press. Muten, E. (1991). Self-reports, spouse ratings, and psychophysiological assessment in a behavioral medicine program: An

application of the five-factor model. Journal of Personality Assessment, 57, 449464. Norman, W. T. (1963). Toward an adequate taxonomy of personality attributes: Replicated factor structure in peer

nomination personality ratings. Journal of Abnormal and Social Psychology, 66, 574-583. Norman, W. T. (1967). 2800 Personality trait descriptors: normative operating characteristics for a university population. Ann

Arbor: University of Michigan. Norman, W. T. & Goldberg, L. R. (1966). Raters, ratees, and randomness in personality structure. Journal of Personality

and Social Psychology, 4, 681691.

Four ways 665

Ostendorf, F. (1990). Sprache und Per&mlichkeitsstruktur: Zur Validitiit des Fiinf-Faktoren-Modells der Persiinlichkeit [Language and personality structure: Toward the validation of the five-factor model of personality]. Regensburg: S. Roderer Verlag.

Passini. F. T. & Norman, W. T. (1966). A universal conception of personality structure? Journal of Persona&v and Social Psychology, 4, 4449.

Peabody, D. (1987). Selecting representative trait adjectives. Journal of Personality and Social Psychology, 52, 59-71. Plomin, R. & McClearn, G. E. (1990). Human behavioral genetics of aging. In Birren, J. E. & Schaie, K. W. (Eds), Handbook

of the psychology of aging (3rd Edn, pp. 67-78). New York: Academic Press. Rokeach, M. (1960). The open and closed mind. New York: Basic Books. Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological

Monographs, 80(l). Rushton, J. P., Fulker, D. W., Neale, M. C., Nias, D. K. B. & Eysenck, H. J. (1986). Altruism and aggression: The

heritability of individual differences. Journal of Personality and Social Psychology, 50, 119221198. Schonemann, P. H. (1966). A generalized solution of the othogonal Procrustes problem. Psychometrika, 31, I-10. Shock, N. W., Greulich, R. C., Andres, R., Arenberg, D., Costa, P. T., Jr, Lakatta, E. G. & Tobin, J. D. (1984). Normal

human aging: The Baltimore Longitudinal Study of Aging (NIH Publication No. 84-2450). Bethesda, MD: National Institutes of Health.

Synder, M. (1974). Self-monitoring of expressive behavior. Journal of Personality and Social Psychology, 30, 526-537. Strelau, J., Angleitner, A., Bantelmann, J. & Ruth. W. (1990). The Strelau Temperament Inventory-Revised (STI-R):

Theoretical considerations and scale development. European Journal of Personality, 4, 209-235. Tellegen, A. (1982). Brief manual for the Differential Personality Questionnaire. Unpublished manuscript, University of

Minnesota. Tellegen, A. & Walker, N. G. (in press). Exploring personality through test construction: Development of the

Multidimensional Personality Questionnaire. In Briggs, S. R. & Cheek, J. M. (Eds), Personality measures: Deuelopmenf and evaluation (Vol. 1). Greenwich, CT: JAI Press.

Toobv. J. & Cosmides. L. (1990). On the universality of human nature and the uniqueness of the individual: The role of ---~,1

genetics and adaptation. Journal of Personality, 38, 1768. Trapnell, P. D. & Wiggins, J. S. (1990). Extension of the Interpersonal Adjective Scales to include the Big Five dimensions

of personality. Journal of Personality and Social Psychology, 59, 781-790. Tupes, E. C. & Christal, R. E. (1961). Recurrent personality factors based on trait ratings (USAF ASD Technical Report

No. 61-97). Lackland Air Force Base, TX: U.S. Air Force. Wechsler, D. (1958). The measurement and appraisal of adult intelligence (4th Edn). Baltimore: Williams & Wilkins. Welsh, G. S. (1975). Creativity and intelligence: A personaIi/y approach. Chapel Hill, NC: Institute for Research in Social

Science, University of North Carolina. White, R. W. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66, 2977333. Wiggins, J. S. (1968). Personality structure. In Farnsworth, P. R., Rosenzweig, M. R. & Polelka, J. T. (Eds), Annual review

of psychology (Vol. 19, pp. 293-350). Palo Alto, CA: Annual Reviews. Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptive terms: The interpersonal domain. Journal of Personality

and Social Psychology, 37, 395412. Windel, M. & Lerner, R. M. (1986). Reassessing the dimensions of temperament individuality across the life span: The

Revised Dimensions of Temperament Survey (DOTS-R). Journal of Adolescent Research, I, 213-230. Yang, K. & Bond, M. H. (1990). Exploring implicit personality theories with indigenous or imported constructs: The

Chinese case. Journal of Personality and Social Psychology, 58, 1087-1095. Zuckerman, M. (1979). Sensation seeking: Beyond the optimal level of arousal. Hillsdale, NJ: Lawrence Erlbaum Associates. Zuckerman. M. (Chair). (1991). How do we determine which are the basic dimensions of personality? Symposium presented

at the Meeting of the International Society for the Study of Individual Differences, Oxford, July. Zuckerman, M., Kuhlman, D. M., Thornquist, M. & Kiers, H. (1991). Five (or three) robust questionnaire scale factors

of personality without culture. Personality and Indiaidual Differences, 12, 929-941.