Psychological Review Copyright 1998 by the American Psychological Association, Inc. 1998, Vol. 105, No. 2, 230-250 0033-295X/98/$3.00

Traits and Motives: Toward an Integration of Two Traditions in Personality Research

David G. Winter University of Michigan

Oliver P. John University of California, Berkeley

Abigail J. Stewart University of Michigan

Eva C. Klohnen University of California, Berkeley

L a u r e n E . D u n c a n Harvey Mudd College

After reviewing classic and current conceptions of trait (as measured by questionnaires) and motive (as measured by the Thematic Apperception Test [TAT] or other imaginative verbal behavior), the authors suggest that these 2 concepts reflect 2 fundamentally different elements of personality-- conceptually distinct and empirically unrelated. The authors propose that traits and motives interact in the prediction of behavior: Traits channel the behavioral expression of motives throughout the life course. The authors illustrate this interactive hypothesis in 2 longitudinal studies, focusing on the broad trait of extraversion and the 2 social motives of affiliation and power. In interaction with extraversion, both motives show predicted and replicated relations to independently measured life outcomes in the domains of relationships and careers. Extraversion facilitates unconfiicted motive expression, whereas introversion deflects social motives away from their characteristic goals and creates difficulties in goal attainment.

In the 7 decades of modem personality research, the concepts of trait and motive have usually led separate l i ves - -each con- fined to its own theoretical school or camp, its integrity or even its existence challenged by the other side, with little opportunity for interaction or exchange. The continued antagonism, or at least rivalry, between these two traditions is manifest in a recent target article in Psychological Inquiry (Pervin, 1994a), and the 17 replies that it elicited (see also Pervin, 1989b). Perhaps this is an auspicious time to reconsider the relationship between these two important concepts and traditions in personality theory and research: McCrae (1994), a well-known trait researcher, has acknowledged the "rising status of motivational psychol-

o g y " and " the renaissance of this venerable posit ion" (p. 150), whereas Pervin (1994a) "applaud [ ed ] the gains being made in trait research" (p. 110) and concluded that " the question of the relation of traits to motives is of particular concern" (p. 109).

What is the current status of trait and motive concepts? Are they really the same concepts, albeit with different labels, or are they different? How much do they actually overlap? Are both required for understanding personality and behavior, or can we get along with only one? This article is devoted to a discus- sion, informed by empirical research, of these questions, which we believe are central to the field of personality psychology. We

David G. Winter and Abigail J. Stewart, Department of Psychology, University of Michigan; Oliver E John and Eva C. Klohnen, Department of Psychology, University of California, Berkeley; Lanren E. Duncan, Department of Humanities and Social Sciences, Harvey Mudd College.

Longitudinal projects usually depend on support from a variety of sources over a long period of time. The projects on which this research is based have been supported most recently by grants from the National Institutes of Health (Grant MH-43948 and Subgrant 6-30236A under Prime Grant MH-47408). Additional support for this research came from National Institute of M~ntal Health Grant MH-49255 and a Jacob K. Javits Fellowship awarded to Eva C. Klohnen. Earlier support came from the Carnegie Corporation, the Fund for the Improvement of Post- Secondary Education, the Boston University Graduate School, the Na- tional Science Foundation Visiting Professorships for Women, the Soci- ety for the Psychological Study of Social Issues, the MacArthur Founda- tion Network for Research on Successful Midlife Development, and Radcliffe Research Support and Midlife Program Grants from the Henry A. Murray Research Center, Radcliffe College. Computer-accessible data

and copies of some of the raw data for several waves of the Radcliffe study have been archived at the Henry A. Murray Research Center.

We are indebted to Ravenna Helson, who has directed the Mills project from its inception; to the many researchers involved over the years with the organization, maintenance, and analysis of the Mills and Radcliffe data sets; and to the women participants in these studies, who have given generously of their time and effort. Our collaboration on this article was greatly facilitated by sabbatical visits to Berkeley and to Michigan. Support from the Institute of Personality and Social Research at the University of California, Berkeley, and the Department of Psychology at the University of Michigan are gratefully acknowledged. We are grateful to Bill E. Peterson and Brent Roberts for their assistance on this project, and to the Mills-Radcliffe Research Group for their suggestions and encouragement.

Correspondence concerning this article should be addressed to David G. Winter, Department of Psychology, University of Michigan, 525 E. University Avenue, Ann Arbor, Michigan 48109-1109. Electronic mail may be sent to dgwinter @umich.edu.

230

TRAITS AND MOTIVES 231

propose an integrated model designed to bring together traits and motives--as complementary concepts rather than as antago- nistic rivals. First, we review the literature in personality theory and research, elucidating the separate histories of the trait and motive concepts. We argue that motives and traits were con- ceived as different kinds of concepts that referred to different aspects of personality and predicted different kinds of behavior. We then review data suggesting the empirical independence of motives and traits--even motives and traits that would seem, on account of their names, to refer to the same domains of behavior. We then develop our fundamental hypothesis that mo- tives involve wishes, desires, or goals (often implicit or noncon- scious), whereas traits channel or direct the ways in which motives are expressed in particular actions throughout the life course. Finally, we illustrate the empirical implications of this channeling hypothesis in two longitudinal studies. Motive and trait concepts, when considered together, offer analyses and in- terpretations of behavior that are far more subtle and sophisti- cated than those that employ only one of the two concepts, or employ them separately. Our purpose, then, is twofold: (a) to propose a more integrated and thereby more inclusive concep- tion of personality, and (b) to demonstrate the value of this conception by highlighting the interactive effects of traits and motives on important life outcomes in adulthood.

The Search for Fundamental Categories in Personality Theory and Research

Classical Theories

No doubt the roots of the trait and motive traditions can be traced back to different aspects of the ancient Greek conceptions of human nature. Empedocles (5th century B.C.), for example, emphasized the primacy of the two great motive forces of love (the striving to unite or fuse) and strife (the striving to dissolve these fusions). In contrast stood the trait emphasis of Theophrastus (4th century B.C.), who described numerous char- acter types or "patterns of behavior" (Anderson, 1970). The trait perspective was further codified in Galen's doctrine of humors (2nd century A.D.), which Eysenck used as a frame- work for his dimensional theory of traits (Eysenck & Eysenck, 1985).

More recently, these two traditions can be discerned in the respective contributions of Freud and Jung. Freud's theory of personality was built upon a motivational dualism of the life and death instincts that explicitly paralleled that of Empedocles (Freud, 1937/1964, pp. 245-247). In contrast, Jung's (1923/ 1971 ) theory of types, based on the concepts of extraversion- introversion and the four functions of thinking, feeling, sensing, and intuiting, was an important contribution to trait theory and research.

Allport, Murray, and McClelland

For most contemporary personality psychologists, however, the contrast between the trait and motive traditions is rooted in the contrasting theories of Gordon Allport (1937, 1961) and Henry Murray (1938), two of the principal American architects of the field. In developing a broad framework for the description and analysis of personality, each aimed to identify fundamental

elements. Because they both worked at the same time and even in the same place (Harvard University in the mid-1930s), per- haps it was inevitable that the "narcissism of minor differ- ences" (Freud, 1930/1961, p. 114) would lead them to empha- size different units.' For Gordon Allport (1961), traits were basic: "Scarcely anyone questions the existence of traits as the fundamental units of personality" because they represent the "considerable constancy in a person's mode of conduct" (pp. 332, 334). In contrast, Murray (1938) argued that"the psychol- ogists who think of personality as . . . traits and those who think of it as . . . [motives] focus attention on different phe- nomena, use different methods and end with different accounts" (p. 714). His own preference was clear: "Practical experience has led me to believe that . . . classification [of behavior] in terms of effects [i.e., motives ] organizes for our understanding something that is more fundamental t h a n . . , classification in terms of ]traits]" (Murray, 1938, p. 56, see also pp. 54-76 and 711-716 passim).

Writing 15 years after Allport and Murray, David McClelland ( 1951 ) argued that both traits and motives were needed to de- scribe personality, 2 because the two terms referred to two differ- ent kinds of behavior: "Certainly we need two concepts--one which will account for the consistencies and recurrences, and one which will account for the inconsistencies and sudden, irra- tional changes in b e h a v i o r . . . . [T]he two kinds of behavior are sufficiently different to require different hypothetical con- structs" (p. 215; see also McClelland, 1985a, 1985b, chap. 13). In fact, Allport and Murray also implicitly acknowledged this distinction, as is discussed later.

Trait and Motive Compared

The trait-motive conceptual debate has been exacerbated by confusions and conflicts in the definition and understanding of both of the principal concepts. Thus, in the psychoanalytic tradi- tion, motives derive from fundamental biological drives and are usually not conscious, which means that they must be measured indirectly. In the social-cognitive and personal-striving tradi- tions, however, motives derive from current concerns or tasks, are accessible to consciousness, and can be measured through self-report (Emmons, 1993; McClelland, Koestner, & Weinber- ger, 1989; Pervin, 1989a; Weinberger & McClelland, 1990). With traits, we find such a remarkable diversity of definitions and theoretical conceptions that Goldberg (1994) concluded that trait theory itself is no more than a "fictive classification" (p. 128). Perhaps, then, it is useful at the outset to outline brief conceptions of the two concepts.

Motives and the variability of behavior. What is the differ- ence between a motive and a trait? Motives refer to people's wishes and desires--states of affairs that they would like to bring about (consciously or unconsciously) or, in the case of avoidance motives, states of affairs they would like to prevent

1 The fact that in 1937, the year of Gordon Allport's textbook, his older brother Floyd published an article emphasizing the importance of motives (or "teleonomic trends") and distinguishing them from traits may be another example of the narcissism of minor differences (see E H. Allport, 1937).

2 McClelland (1951 ) also added schemata or "cognitive maps" as a third kind of personality unit (pp. 239-243).

232 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

(see Peters, 1958; Winter, 1973, chap. 3, 1996a, chap. 2). Mo- tives refer to the "why" of behavior (McClelland, 1985b, p. 4); in more formal terms, the "disposition to be concerned with and to strive for a certain class of incentives or goals" (Emmons, 1989, p. 32; see also Emmons, 1993). Murray em- phasized that the particular actions associated with a given mo- tive may not be highly intercorrelated because they will vary according to the situation, expectancies, incentives, obstacles, skills, and other motives (see also Cattell, 1946, p. 107, on this point). This can readily be illustrated in the following example: Your hunger motive, having as its goal eating food, would lead to quite different actions: (a) at home, in your own kitchen, (b) in your favorite restaurant, (c) on the streets of a foreign city where the language was completely strange and you had no phrase book, or (d) on a 6-hr airplane flight with no meal service. Thus, goal-directed behavior shows intelligent variation in relation to the situation. McClelland (1951) defined motive as "a convenient construct" that serves to unify, tie together, or give a common meaning to a wide variety of dissimilar re- sponses (p. 390). As a result, motives may be difficult to mea- sure through observation (including self-observation) of behav- ior, especially under constrained conditions.

Implicit and explicit motives. Often motives are experienced as conscious intentions and goals. Most researchers studying life tasks, personal projects, personal strivings, and current concerns (Cantor & Zirkel, 1990) assume that people can give reasonably accurate accounts of their goals. Sometimes, however, motives are implicit or unconscious and not easily accessible to aware- ness. People may not be aware of their motives because their goal-related cognitions are actually schemas constructed around vivid memories (see Nisbett & Wilson, 1977) or idealized self- conceptions. Or people may defend against knowledge of their true motives (Hogan, 1996, pp. 176-177). For whatever rea- sons, then, motives (like other cognitions) may often be implicit, that is, not accessible to consciousness and therefore measurable only by indirect means (see Greenwald & Banaji, 1995).

For Freud (1900/1953), the interpretation of dreams and fan- tasy was the "royal road to a knowledge of the unconscious activities of the human mind" (p. 608). Building on Freud's insight, later personality psychologists developed a variety of more systematic indirect ways of measuring motives. Thus Mor- gan and Murray (1935) devised the Thematic Apperception Test (TAT), in which people make up stories about a series of vague or ambiguous pictures. 3 As a further scientific refinement, McClelland and his colleagues used the effects of experimen- tally aroused motives on TAT story content in order to define objective motive scoring systems (see McClelland, Atkinson, Clark, & Lowell, 1953; Smith, 1992; Winter, 1973, chap. 3, 1996a, pp. 137-138, 1996b). Working in this tradition, person- ality psychologists have developed objective measures of the achievement, affiliation, and power motives through systematic analysis of fantasy and verbal content (Atkinson, 1958; McClel- land, 1985b; Smith, 1992; Winter, 1996a, chap. 5). These three motives can be seen as the fundamental dimensions underlying Murray's list (see Winter, 1996a, pp. 122-125, 157-158). Fur- ther, the two social motives of affiliation and power can also be seen as part of humans' evolved heritage as "group-living, cul- ture-using animals" (Hogan, 1996, p. 165), and linked with Bakan's (1966) broad dualism of communion and agency, re- spectively (see also Wiggins & Trapnell, 1996).

Typically, TAT-based motive measures do not correlate highly with direct or questionnaire measures of motives (see the litera- ture summarized in McClelland, 1980, and Spangler, 1992; but see also Emmons & McAdams, 1991). Also, the two kinds of measures predict different kinds of behavior (e.g., long-term "operant," or life outcome variables, versus choice behavior). These considerations have led McClelland and his associates (McClelland, Koestner, & Weinberger, 1989; Weinberger & McClelland, 1990) to suggest that TAT and questionnaire mea- sures of motives reflect two different motivational sys tems-- "implicit" and "self-attributed" motives, respectively. In this article, we are concerned only with implicit motives, as mea- sured by the TAT or other similar apperceptive techniques.

One characteristic of TAT motive measures deserves a brief discussion at this point because it has given the procedure a bad reputation in the minds of many psychometrically minded personality psychologists (e.g., Entwisle, 1972). TAT-based test-retest correlations are often low, suggesting low temporal reliability. Studies by Winter and Stewart (1977) and Lundy (1985), however, have shown that these apparently low correla- tions are mostly an artifact of retest instructions that exert an implicit pull for participants to tell different stories. When peo- ple are instructed on retesting not to worry about whether their stories are similar to or different from those written before, test-retest correlations rise to .60 or higher.

Further, Atkinson (1982) has tried to reconcile the conflicting facts of dispositional motive stability and temporal fluctuation by two theoretical proposals: (a) A particular "motive" is actu- ally the relatively stable disposition to experience relatively more rapid arousal of that motive, by a relatively broader class of motive-related stimuli, to a relatively greater height; (b) the expression of a motive in action (including the act of telling a motive-saturated TAT story) temporarily reduces the level of arousal of that motive. Motives are, in other words, disposition- ally stable and situationally contingent.

Traits and the stylistic consistency of behavior. One imme- diate difficulty in discussing the concept of trait is that there is no single, universally accepted definition of what exactly is a trait. The following are samples of the diversity of views: McCrae and Costa (1996) saw traits as "basic tendencies;" D. M. Buss and Craik (1983) defined them as frequency trends of specific acts; Block (1995) conceived of them as dynamic regulatory mechanisms; and Saucier and Goldberg (1996) as- sumed a theoretically agnostic lexical position that traits are words people use to describe others (see also Funder, 1991; Hogan, 1996, pp. 170-172).

Still, we may note that traits are often described as referring to people's "stylistic and habitual patterns of cognition, affect, and behavior" (Emmons, 1989, p. 32; see also Emmons, 1993; John & Robins, 1993, p. 226n), rather than to their wishes, desires, and goals. Thus McClelland (1951) decided "for rea- sons of convenience to use the trait variable to describe the

3 Methods such as the TAT were originally labeled projective, on the hypothesis that unconscious motives were projected, in the psychoana- lytic sense, onto the characters in the stories. Later, Murray (1958) came to doubt that TAT stories contained only unconscious or projected motives (p. 190; see also the classic study by Combs, 1947). In fact, the motive content coding systems used in this article (Smith, 1992) do not assume a projective model in the psychoanalytic sense.

TRAITS AND MorrvEs 233

surface or stylistic manifestations of personality only" (p. 233). In one of the earliest publications on traits, G. W. AUport ( 1931 ) referred to their "adverbial quality that is the very essence of personality" (p. 369, emphasis added). By implication, then, traits are comprised of more-or-less consistent, generalized, in- tercorrelated clusters of behaviors. As G. W. Allport (1937) put it, "the scientific evidence for the existence of a trait always comes from demonstration by some acceptable method of con- sistency in behavior" (p. 330, emphasis in original). Thus traits can be identified through correlational studies and factor analy- ses of questionnaire items, adjectives, or Q-sorts reflecting the individual's " typical" behavioral conduct (though for G. W. AUport, 1966, purely statistical techniques such as factor analy- sis were at best a "mixed blessing" [p. 3]) . In Eysenck's (1953) words, "the notion of trait is intimately connected with the notion of correlation" (p. 9, emphases in original; see also Eysenck & Eysenck, 1985, p. 12). Hence, almost all respectable trait measures have successfully survived tests of internal and cross-situational consistency, as well as temporal stability.

On the other hand, as G. W. Allport (1937) argued, "perfect consistency [even in relative or rank terms] will never be found and must not be expected," because traits can contradict or modify each other. Furthermore, "traits are often aroused in one situation and not in another; not all stimuli are equivalent in effectiveness" (pp. 330, 331-332).4 Several theorists have sug- gested that traits are organized along a continuum: At the most specific level, traits are numerous, conditional, contextual, and not always very consistent. Concepts at this more contextualized level are called "facets" by Costa and McCrae (1985) and "habitual responses" by Eysenck (1953, pp. 12-14) . At the most abstract level, in contrast, traits are much more consistent and less contextualized (Saucier & Goldberg, 1996, pp. 32-33; see also John, 1990; McAdams, 1992, 1995).

Although the empirical claim of behavioral consistency was strongly criticized by Mischel (1968) and others, an extensive literature demonstrating considerable cross-situational and tem- poral trait consistency (especially at the level of broad traits) has emerged in recent years (see, e.g., Block, 1977; Epstein, 1984; Winter, 1996a, chap. 16).

In recent years, many personality psychologists have reached a working consensus that the domain of traits can be described, at least at the broadest and most abstract level, by five factors or clusters of traits (see McCrae & John, 1992; Wiggins, 1996; but see Block, 1995, for a contrary view): extraversion, agree- ableness, conscientiousness, neuroticism (or low emotional sta- bility), and openness to experience. These make up the Five- Factor Model (FFM) or "Big Five." Some other researchers suggest that the most appropriate number of trait-factors is three (Eysenck, 1992), six (Hogan, 1986), or seven (Benet & Waller, 1995). However, if there is one trait that all theorists and re- searchers agree actually is a trait, it is extraversion. Virtually every factor theory, typology, or other conception of traits in- eludes extraversion-introversion as a prominent master trait (see Eysenck & Eysenck, 1985, chaps. 2 -4 ; Wiggins, 1968), although its definition and contents are somewhat different in different accounts. In most definitions, extraversion refers to a pattern of covarying elements: behaviors (e.g., lively talking to people one has just met), feelings (e.g., positive emotions such as joy and enthusiasm), and cognitions (e.g., optimistic expecta-

tions for the future). Extraversion (and its polar opposite, intro- version) is the principal trait of interest in this article.

Traits and mechanisms. Is it possible to rescue the concept of trait from a purely descriptive empirical "dust bowl" of personality by specifying more precisely how traits channel motives in ways that are stylistically consistent? An early formu- lation by Cattell (1946) provides a clue to at least some possible mechanisms mediating these channeling effects. Traits, he ar- gued, are best described as patterned individual differences in measures of "such terms as reactivity, variability, susceptibility, tolerance of stimuli" and "perseveration or oscillation," which are "applicable to any variety of drive [i.e., motive]" (pp. 180, 192).

Extensive research by Eysenck and his colleagues (Eysenck & Eysenck, 1985) illustrates how these terms might apply to extraversion (see also Winter, 1996a, pp. 492-501) . Extraverts tend to show higher thresholds for stimuli, lower thresholds for boredom, and lower characteristic levels of corti- cal arousal. Such a pattem of relationships could involve differ- ences in any or all of the following physiological mechanisms, among others: peripheral receptor cell thresholds, degree of nerve fiber myelinization, neurotransmitters and neurochemical pathways, the reticular formation, and even cortical feedback. In any case, and by whatever mechanisms, extraverts come to seek arousal. Eysenck and Eysenck (1985, p. 313) suggested that "social situations often produce powerful arousal effects" (see also Patterson, 1976). They concluded that the sociability aspect of extraversion derives from these physiological mecha- nisms: "As a consequence, extraverts may seek out personal contacts in order to prevent the level of arousal from becoming too low, whereas introverts may tend to avoid such social con- tacts because of the danger of overarousal" (p. 313).

Building on Cattell's original formulation, we propose that extraversion channels motives by virtue of individual differences in reactivity, susceptibility, and stimalus-tolerance mechanisms. Other traits may work through these or other mechanisms: For example, neuroticism may involve oscillation originating from the limbic system, and conscientiousness may reflect differences in perseveration. From a social cognitive perspective, these trait- related differences would appear as differences in perceived efficacy (Bandura, 1977). Most generally, then, we propose that such underlying mechanisms affect the availability, accessibility, and intrinsic pleasure associated with behavioral options or channels.

The Relat ion o f Trait and Motive Concepts

In the more than 5 decades since the major publications by Gordon AUport (1937) and Henry Murray (1938), different personality psychologists have taken different positions about the relation of traits and motives. A brief review of two major positions will set the stage for our own position and research.

Motives as Complementary to Traits

Although Allport preferred traits and Murray preferred mo- tives as the fundamental element of personality, each did make

4 Here Allport seemed to ascribe some features of motive to the trait concept, perhaps reflecting his uncertainty (or ambivalence) about whether they were separate concepts (see Pervin, 1993, pp. 70-72, and later discussion in this article).

234 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

room for the other concept in his overall conceptual framework. Thus, although Murray's (1938) "variables of personality" (chap. 3) included his famous list of needs, they also included numerous adverbial "general traits or attributes," such as anxi- ety and extraception/intraception (pp. 147-149, 199-227 pas- sim). Similarly, Allport distinguished stylistic (or expressive) traits from dynamic (or "tel ic") traits, although his writings on this topic seem ambivalent and even confused (Pervin, 1993, pp. 70-72) . In his classic 1937 text, G. W. Allport drew on the work of Lewin, Troland, and Stern to elaborate the distinction: "Some traits thus seem to have motivational (directional) sig- nificance, and some mere instrumental significance. The latter are primarily expressive . . . [and] represent styles of behav- ing" (p. 323, emphasis in original). Having made this distinc- tion, however, Allport immediately moved to retract it: "There is no sharp line between motivational traits and stylistic traits, between direction and manner of expression . . . . Whether a trait is dynamic or directive is, then, a matter of degree" (pp. 323-324; emphasis in original). On the next page, however, he went on to distinguish description in terms of phenotypical "here-and-now attributes" from explanation through genotypi- cal underlying motives that are "the very springs of conduct" (p. 325).

Many modem personality psychologists have also made a place for both motive and trait concepts in their theorizing and research. Following Allport's distinction, Cattell differentiated temperament-stylistic traits from dynamic traits, each reflecting a different kind of consistency, and each therefore requiring a different kind of measurement. He argued that temperament traits are measured best by measures "which change least in response to any change in the field" (Cattell, 1946, p. 167). Dynamic traits, in contrast, may involve different and even nega- tively correlated component behaviors, 5 and so are measured best by measures "which change most in response to change of incentives in the situation" (Cattell, 1946, p. 167; see also Cattell, 1965, chap. 7). Both kinds of traits (along with abilities or ability traits) are included in a "specification equation" relating personality and situation to behavior (Cattell, 1957, pp. 302-306). Guilford (1959) distinguished temperaments (traits) from needs and interests (motives), and Winter (1996a) divided personality into four different elements: motives, traits, cognitions, and social contexts.

In a recent Psychological Inquiry issue, Pervin (1994a, 1994b), Fiske (1994), Funder (1994), Harlow and Cantor (1994), John and Robins (1994), McAdams (1994), and Snyder (1994) all endorsed some kind of basic conceptual distinction between trait and motive (see also Funder, 1991, especially p. 34; McAdams, 1995; Winter & Stewart, 1995). Often the distinction is phrased in terms of traits as qualities that people have and motives as what people do (see Cantor, 1990).6 Along these lines, Snyder (1994) quoted an example from the classic article by Floyd Allport (Gordon Allport's older brother): "Three individuals [are] rated or measured as equally honest [i.e., have the same amount of the same trait]. One of them might be seeking justice, another might be trying always to help others, while the third might be trying to maintain his self- esteem or reputation" [i.e., different motives] (E H. Allport, 1937, p. 204). For McAdams (1994), traits explain a behavior by subsuming a particular phenomenon into a general pattern.

In contrast, motives explain a behavior by the end (s) or goal ( s ) toward which it tends. 7

In a recent discussion of levels and domains in personality, McAdams (1995) has given a strong statement of the "comple- mentarity" position that traits and motives are concepts at fun- damentally different levels, serving different purposes in ex- plaining people's behavior. At the abstract level of the five-factor model, at least, traits are broad and relatively nonconditional constructs:

[They] provide an excellent "first read" on a person . . . [along] general and linear dimensions of proven social significance . . . . It is the kind of information that strangers quickly glean from one another as they size one another up and anticipate future interac- tions. (p. 374)

In contrast, motives "are defined in terms of future ends" and oriented "in a particular way in time" (p. 377). The two levels are distinct:

It is not necessary for the viability of the concept of extraversion that an extraverted person strive to obtain a particular goal in time, although of course such a person may do so. Extraverted people simply are extraverted; whether they try to be or not is irrelevant. (p. 377)

Traits as Subsuming Motives

For all the importance of the distinction in theoretical writ- ings, the actual empirical boundaries between motive and trait have remained fuzzy. As Heckhausen ( 1991 ) put it: "Observed behavior, which leads to trait descriptions like 'helpfulness' or 'pugnacity,' is endowed with motivational characteristics, i.e., it appears as though the individual strives to exhibit such behavior whenever possible" (p. 49). What could be more obvious than that "extraverted" people, who are outgoing and report that they enjoy and seek out the company of others, must want to be affiliative? Therefore, one would think, they must be high in "affiliation motivation." Thus many trait theorists conceive of traits as having motivational components, or even that traits subsume motives. In the recent Psychological Inquiry issue, for example, McCrae (1994), Goldberg (1994), Hofstee (1994), and Ostendorf and Angleitner (1994) all took this position, as have many others (e.g., Borkenau, 1990; McCrae & Costa, 1996, pp. 62-63, 69; Read, Jones, & Miller, 1990). Thus A. H. Buss (1989) went so far as to say that " if there is to be a

5 See Cattell (1946, p. 107) on this point. 6 Thus adjectives (and adverbs) are "the prototypical and central

repositories of the sedimentation of important individual [trait] differ- ences into the natural language" (Saucier & Goldberg, 1996, pp. 30- 31 ). Motives, in contrast, are typically represented by verbs. Like verbs, motive-related actions are "prohibitively ambiguous," situated "at the intersections of semantic fields," and "organized into much looser asso- ciational networks" (Saucier & Goldberg, 1996, pp. 30-31 ). Thus their interpretation requires greater knowledge of the context.

7 A "goal" can be rigorously and objectively defined in terms of the increased predictability of behavior up to a "consummatory response," followed by sharply decreased predictability of behavior (Cattell, 1946, pp. 114-115; Winter, 1973, p. 29, 1996a, p. 34).

TRAITS AND MOTIVES 235

specialty called personality, its unique and therefore defining characteristic is traits" (p. 1378).

Given the high correlations often obtained between many trait measures and the questionnaire scales designed to measure Murray's needs (e.g., the Jackson, 1965, Personality Research Form or PRF; Jackson, 1984; and the Adjective Check List or ACL; Gough & Heilbrun, 1983) such a position may seem reasonable. For example, Costa and McCrae (1988) reported that in a joint factor analysis of their NEO-PI (Costa & McCrae, 1985) and the PRF, the Affiliation scale of the PRF has a loading of +.83 on a common extraversion factor (see also Paunonen, Jackson, Trzebinski, & FOrsterling, 1992; Stumpf, 1993). Such results might seem to be good grounds for accepting Hofstee's (1994) conclusion that "questionnaires based on Murray's sys- t e m . . , are indistinguishable from other [that is, trait] person- ality inventories" (p. 135).

However, we must point out that a questionnaire measure remains a questionnaire measure, regardless of whether it is labeled as assessing Murray's needs. As pointed out earlier, such measures do not correlate with TAT-based motive measures, nor do the two kinds of measures predict the same kinds of behaviors (McClelland, 1980; Weinberger & McClelland, 1990). For ex- ample, McClelland et al., (1989) found correlations ranging from - .21 to .15 between the TAT measures of achievement, affiliation, and power motives and their presumed PRF "name- sakes," and Demakis and McAdams (1992) reported similar correlations for the extraversion scale of the Eysenck Personality Questionnaire. Thus Hofstee's (1994) conclusion that traits sub- sume motives must be restricted to questionnaire "motive" mea- sures only; it cannot support a general claim that traits subsume motives as assessed with non-questionnaire measures. In other words, assuming that a PRF scale for one of Murray's needs (e.g., n Affiliation) is the same as the TAT motive measure is an example of what Block (1995) called the "jingle fallacy," that is, using the same name for two concepts that are, in fact, different. Along with the theoretical considerations mentioned earlier, these empirical considerations also suggest the impor- tance and usefulness of rethinking the distinction between mo- tives and traits.

Reconsiderat ion of the Mot ive -T ra i t Dist inct ion

Understanding the Dist inction Between Questionnaire I tems and TAT Mot ives

Perhaps the conceptual independence of traits and motives can be illustrated by discussing reasons why questionnaire and thematic apperceptive measures should not correlate. Consider, for example, the following extraversion item from the California Psychological Inventory (CPI; Gough, 1987): " I like large noisy parties," or a similar item from the PRF, " I truly enjoy myself at social functions." Extraverts agree with these items because they appeal to their need for stimulation and their spon- taneity in social situations. Some affiliation-motivated people (at least in some circumstances) might also endorse these items, because "part ies" or "social functions" represent an opportu- nity for establishing and enjoying relationships with other peo- ple. On the other hand, "large noisy parties" and "social func- tions" are not the only ways, or even always very appropriate ways, to experience close relationships. Thus some affiliation-

motivated people might reject these items on the grounds that (a) such events increase their sense of loneliness, (b) the sight of other people having a good time (perhaps a better time than they are having) is distressing instead of enjoyable, or (c) they would rather spend the time with a few close friends, in intimate conversation. Conversely, some people low in affiliation motiva- tion might endorse these items on the grounds that large noisy parties and many social functions are essentially impersonal, anonymous affairs in which one can avoid entanglements (by walking away to a different conversation, getting a drink, etc.) or pursue other goals (such as networking for a new job). Thus we would not necessarily expect any consistent relationship between the affiliation motive as measured by the TAT and responses to these CPI and PRF items.

What about power motivation and responses to these items? Some power-motivated people might like large noisy parties and enjoy themselves at social functions, because they offer an opportunity for impressing other people, expressing unre- strained power urges through excessive drinking, or perhaps networking, that is, establishing contacts that are useful in the pursuit of power. Other power-motivated people, however, may find such events aversive because the exposure to the power of other people arouses their own sense of inferiority or because sexual rivalry makes them jealous, and so on. Conversely, whereas some people low in power motivation might avoid such events for just these reasons, others might find them attractive as anonymous occasions during which the usual hierarchies of power are dissolved or even reversed (as, for example, in the Saturnalia festival in ancient Rome or modern-day Carnival).

Perhaps in spite of these problems, it might prove possible to write "better" questionnaire items that would more accurately reflect some of the subtle nuances of the TAT-based motive measures, so as to construct scales that do correlate with those measures (see Chen, Ford, & Wang, 1996). However, even the most subtle questionnaire is still in most respects an explicit or "self-conscious" measure that does not capture implicit, tacit, or covert motive (and other mental) processes. (McClelland, 1980, earlier phrased the distinction between implicit and con- scious measures as "operant" versus "respondent," respec- tively.) Evidence from studies of learning (Reber, 1989), mem- ory (Schacter, 1987), attitudes, and prejudice (Greenwald & Banaji, 1995) all suggest that the implicit-conscious distinction is fundamental in the study of mental processes (see also Nis- bett & Wilson, 1977).

These questionnaire examples illustrate the essential feature of the concept of motive, as pointed out by Murray ( 1938): A given motive can energize a wide variety of behaviors--behav- iors that may differ greatly in their surface features and social impact, but are functionally similar in that they all create a similar inner affective state associated with goal satisfaction. Thus we may conclude that almost any personality questionnaire item endorsed by people with a particular trait can be construed as either (a) saturated with cues related to a currently regnant (or engaged) implicit motive, (b) devoid of cues related to a currently regnant motive, or even (c) saturated with aversive cues related to a currently regnant motive. Thus we would expect no consistent relationship between a TAT-based motive and most measures (questionnaire, observer reports, or self-description) of behaviors. Because traits are conceived as patterns of corre-

236 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

lated behaviors, by extension we would expect no consistent relationship between motives and traits.

If traits and motives are not the same and cannot simply be mapped onto each other, then how can we conceptualize their natures and relationship? In the discussion that follows, we focus on the most important and widely studied motive and trait variables from the interpersonal domain: the affil iation- intimacy and power motives and the trait dimension of extraver- sion-introversion.

The Aff i l iat ion and P o w e r Mot ives

Several lines of theory and research suggest that affiliation and power may be the two most important human social motives. More precisely, they are two major dimensions of motivated behavior, with numerous more specific motives represented as various combinations of them (see Winter, 1996a, chap. 4) . Freud's dualism of eros and destructiveness as well as Bakan's (1966) concepts of communion and agency can readily be mapped onto affiliation and power, respectively. They generally emerge as underlying dimensions in circumplex and cluster anal- ysis studies (see Conte & Plutchik, 1981; Freedman, Leary, Ossorio, & Coffey, 1951; Kiesler, 1983; Wicker, Lambert, Rich- ardson, & Kahler, 1984; Wiggins & Broughton, 1985). Ethno- psychology research suggests that affiliation and power are nearly universal ways of understanding and arranging motivated behavior (Kornadt, Eckensberger, & Emminghaus, 1980).

Table 1 presents a brief version of the definitions for scoring these two motives from TAT stories and other verbal content. (See the appropriate chapters in Smith, 1992, for the complete

scoring systems, as well as discussion of psychometric creden- tials and construct validity; see also Winter, 1996a, chap. 5.) The affiliation motive is assessed in terms of verbal expressions of a concern for close relations with others. Sometimes it is associated with friendly behavior and self-disclosure, but under conditions of threat or stress, it can produce a "pr ick ly" defen- sive orientation to others. McAdams (1982) developed a mea- sure of the intimacy motive, which is conceptually related to affiliation but distinct in its being more positive and less defen- sive. (The "running tex t" scoring system used in the present article combines elements of the affiliation and intimacy sys- tems; see Winter, 1991.) The power motive is assessed in terms of verbal expressions of a concern for impact and prestige. Although it often leads to formal social power, it is also associ- ated under certain conditions (such as low responsibility; see Winter & Barenbaum, 1985) with profligate impulsive actions such as becoming aggressive, drinking and abusing drugs, and taking extreme risks.

E x t r a v e r s i o n - I n t r o v e r s i o n

Under various names, extraversion-introversion figures prominently in almost every major trait theory (e.g., Cattell, 1957; Eysenck & Eysenck, 1985; Gough, 1987, 1989; Jung, 1923/1971). It consistently emerges as the first or most im- portant factor in factor-analytic studies of trait ratings (Gold- berg, 1990). It can be detected back in the Old Testament (Gene- sis 25: 27) as the distinction between the "ha i ry " Esau ( " a skillful hunter, a man of the f ie ld") and the " smoo th" Jacob ( " a quiet man, dwelling in tents" ). Ethnopsychology research

Table 1 Thematic Expression of the Affiliation and Power Motives in the Thematic Apperception Test (TAT)

Motive TAT story

Affiliation Scoring definition, in thematic apperception stories: Someone is concerned

about establishing, maintaining, or restoring friendship or friendly relations among persons, groups, etc.

By expression of warm, positive, friendly, or intimate feelings toward other people, nations, etc.

By expression of sadness or other negative feeling about separation or disruption of a friendly relationship, or wanting to restore it

By affiliative, companionate activities By friendly, nurturant acts

Power Scoring definition, in thematic apperception stories: Someone is concerned

about having impact, control, or influence on another person, group, or the world at large

By taking strong, forceful actions that inherently have impact on other people or the world at large

By controlling or regulating others By attempting to influence, persuade, convince, make or prove a point, or

argue By giving unsolicited help or advice By impressing others or the world at large; prestige or reputation By eliciting a strong emotional reaction in someone else

They were young and in love. Hours would they sit together, taking in the shining beauty of the snow and ice, talking softly. They exchanged understanding glances, and planned for a future together. Yet the air was full of the tensions of war, so they retreated into their world, where no one could come, where they could dream and plan, feel safe and protected, and hope?

The man is an agent of the notorious secret police, and the woman is a member of a revolutionary group. She originally joined after her mother (a famous leader) was framed by the government and imprisoned. Now she wants to carry on the campaign. The police have been checking up on her. The agent is eagerly waiting for the chance to plant a listening device on her. b

Note. Text adapted from Winter (1991, p. 63). These definitions are not sufficient for scoring purposes. See Winter (1991) or Smith (1992) for complete definitions and instructional materials for learning to score. a Example of a TAT story with images emphasized that would be scored for affiliation motivation, according to the experimentally derived scoring system (adapted from Winter, 1996a, p. 144). b Example of a TAT story with images emphasized that would be scored for power motivation, according to the experimentally derived scoring system (adapted from Winter, 1996a, p. 151).

TRAITS AND MOTIVES 237

Table 2 Brief Characterization of Extraversion-lntroversion

Aspect Description

Jung's original description (1923/1971, pp. 516-517)

Questionnaire items (Eysenck & Eysenck, 1985, p. 84) a

Descriptive adjectives (John, 1990, p. 80) b

Cognate dimensions in other theories and questionnaires (John, 1990, p. 89) b

Extraversion: "A ready acceptance of external happenings, a desire to influence and be influenced by events, a need to join in and get 'with it,' the capacity to endure bustle and noise of every kind, and actually find them enjoyable . . . . The cultivation of friends and acquaintances, none too carefully selected."

Introversion: "Holds aloof from external happenings, does not join in, has a distinct dislike of society as soon as he finds himself among too many people. In a large gathering he feels lonely and lost . . . . Crowds, majority views. . .never convince him of anything, but merely make him creep still deeper into his shell."

Endorsed by extraverts: Do you enjoy meeting new people? Do you like plenty of bustle and excitement about you?

Endorsed by introverts: Are you mostly quiet when you are with other people?

Extraversion: talkative, assertive, active, energetic, outgoing, sociable Introversion: quiet, reserved, shy, silent, withdrawn, retiring

Dominant-initiative (Bales) Exvia (Cattell) Externality (Gough) Social activity (Guilford) Ambition and sociability (Hogan) Outgoing, social leadership (Jackson) Agentive, positive emotionality (Tellegen) Dominance (Wiggins)

a From Personality and Individual Differences: A Natural Science Approach (p. 84) by H. J. Eysenck and M. Eysenck, 1985, New York: Plenum Press. Copyright 1985 by Hans J. Eysenck and Michael W. Eysenck. b From "The Big Five" Factor Taxonomy: Dimensions of Personality in the Natural Language and in Questionnaires by O. P. John, 1990, in L. A. Pervin (Ed.), Handbook of Personality Theory and Research (p. 89). New York: Guilford Press. Copyright 1990 by Guilford Press.

suggests that it is an important trait dimension in many nonwest- ern cultures (see, for example, John, 1990, pp. 77 -78 ; Valen- tine, 1961). Table 2 presents Jung's original theoretical charac- terization of extraversion-introversion, along with typical items used to measure it and cognate variables from modem personal- ity research. (See Eysenck & Eysenck, 1985, as well as Winter, 1996a, chaps. 12-14, for discussion of the construct validity of extraversion.)

Motives and Traits in the Affiliation Behavior Domain

We propose that although motives and traits may refer to the same general behavioral domain (e.g., affiliative behavior, or power behavior), they reflect different aspects of that domain. What a person wants and in the process plans, anticipates, and enjoys (motives), and how that person acts (traits), are not necessarily the same. There are, for example, some people who have strong leadership traits and perform superbly well as ad- ministrators, yet do not want or enjoy the job and agree to serve only from a sense of obligation and perhaps guilt. (Both Cincinnatus, the farmer who was twice offered the dictatorship of ancient Rome, and George Washington, who reluctantly agreed to serve as the first U.S. president, are examples) . At the same time, corporate and national political life are full of examples of people who would like to be president, but lack the necessary traits to be elected, or even to be taken seriously. Conceptually, then, traits furnish the particular structures and resources to implement (sometimes also to limit or constrain) the goals specified by motives.

Consider the affiliation behavior system, which involves friendship and social interaction. (Of course, behavior is multi- ply determined. Although many different motives, traits, and other personality variables are relevant to the affiliation behavior system, for simplicity of exposition we restrict ourselves to a single m o t i v e - - f o r aff i l ia t ion--and a single t ra i t - -extraver- sion.) The affiliation motive reflects the importance of affiliation with others as a goal. In contrast, extraversion refers to an expressive interactive style, which may characterize people as they pursue many different goals (e.g., achievement, power, exploration, dependency). Obviously, we can easily imagine an affiliation-motivated extravert and an introvert low in affiliation motivation. But if traits and motives are different constructs, then we should also be able to imagine an affiliation-motivated introvert. In fact, Cattell (1946) suggested such a possibility:

Some writers, describing extraversion. . , refer to the sociability there manifested as if it were a dynamic trait [in our terminology, a motive] in its own right--a direction of interest--rather than a permissive manifestation, through temperament, of a dynamic trait which is present, as a repressed interest, also in desurgent tempera- ments. (p. 179)

A person with a "desurgent temperament" who was also eager for relationships would " n e e d " or " w a n t " the affection and friendship of others, but would be ill at ease in many situations where such affiliation was likely. No doubt this combination might create discomfort, but it is still a combination that could occur. It might lead to conflict (especially in combination with neuroticism) and unsuccessful relationships, but it might also

238 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

lead to having a few deep relationships (or even only one) rather than a large circle of friends.

There is evidence, for example, that former president Richard Nixon was both high in the affiliation motive (Winter & Carlson, 1988) and at the same time introverted and high in neuroticism (Winter, 1990, 1996a, pp. 670-672; see also Simonton, 1986, 1988). As he said to one biographer, "I am essentially shyer than the usually extrovert politician ought to be . . . . Frankly, I 'm a terrible 'mixer' " (Kornitzer, 1960, pp. 335-336). How did Nixon handle this combination of motive and trait personal- ity characteristics? Although he was certainly prickly and defen- sive when threatened, he apparently took great pleasure in close one-on-one relationships and small groups. As one of his advi- sors recalled (Klein, 1980; for similar accounts, see Ambrose, 1987, p. 274, and Price, 1977):

After meeting Nixon individually or in a small group, people usually would say that they had a new and warm feeling for him. "Isn't it too bad he doesn't come across that warmly in an auditorium or on television," they would say. (Klein, 1980, p. 133)

Can we also imagine the converse--extraverted persons who do not necessarily devote a lot of energy to affiliation as a desired end-state? Such people would doubtless be cheery, spon- taneous, friendly, popular, and adept at social relationships; how- ever, they would also be able to "take them or leave them alone," depending on the situation and their other motives. That is, they would pursue whatever goals they have in an extraverted style, with whatever resulting friendships as a harmless (and possibly useful) appendage rather than a core motivational con- cern. Occasionally they might be surprised (but not distressed) at the amount of friendly interaction surrounding them. (In an extreme case, they might display the belle indifference of the hysteric style; see Shapiro, 1965.) There is some evidence that former President Ronald Reagan showed this pattern (see Win- ter, 1987, on Reagan's motives, and Simonton, 1986, 1988, on his traits).

These four trait-motive combinations are summarized for the case of extraversion and the affiliation motive in Table 3. Using data from Winter (1987) and Simonton (1986, 1988), we in- clude examples of a U.S. president to illustrate each of the four combinations.

In summary, affiliation motivation combined with extraver- sion is quite different from affiliation motivation combined with introversion, but it is still recognizably affiliation motivation, involving the goal of having close relationships with other peo- ple. Correspondingly, extraversion will look different when it is

a stylistic "channel" for the affiliation motive, as compared with when it is the channel for the power motive, but the essential feature of extraversion--orientation toward (rather than away from) the outer world--will be present in both cases. In simplest terms, traits answer the question "how?", motives answer the question "why?", and both concepts address the question "what?" We propose that motives and traits each affect the ways in which the other is expressed in behavior. Their empirical independence does not mean that either measure lacks reliability or validity, but rather that they are fundamentally different classes of variables. This conception suggests that motives and traits should have important joint or interactive effects.

Studying the Interaction Between Motives and Traits

In the research reported here, we propose to demonstrate the interactive effects of motives and traits. More specifically, we suggest that: (a) Motives represent fundamental goals or desires; (b) traits represent psychophysiological mechanisms, such as reactivity, thresholds and preferred levels of arousal, variability, tolerance of stimuli, perseveration, and oscillation; and thus (c) traits channel the expression of motives, or goal-directed behav- iors, in ways consistent with these mechanisms. Because mo- tives can be expressed in so many different ways, by so many different behaviors, and because the channeling effects of trait mechanisms are so powerful, we believe that the interactive effects of motives and traits .are likely to make a greater contribu- tion than when either kind of variable is studied alone.

Although there is some evidence for interactive or moderator effects in previous motive and trait research (see Winter, 1996a, pp. 584-589, 595-596, 659-664), Hofstee (1994) has pointed out that demonstrations of interactive relationships are seldom replicated: ' 'The idea of patterning, in its most viable operation- alization provided by the moderator model, appears too potent for our observations. It violates the principle of parsimony" (p. 135). Believing that such a conclusion may be a premature scientific abdication, we studied two separate samples so as to be able to demonstrate replicated interactions of motives and traits. Thus we tested our interaction hypothesis in two longitu- dinal studies of women at midlife: (a) a study of Mills College graduates of the classes of 1958 and 1960, and (b) a study of Radcliffe College graduates of the class of 1964.

Before describing the samples and the research methods, however, we think it important to discuss the reasons why longi- tudinal research offers the most appropriate arena for demon- strating motive-trait interactions, and to suggest some benefits and problems of using data from longitudinal studies.

Table 3 Hypotheses About the Interactions of the Affiliation Motive and Extraversion-lntroversion

In combination with

Motive Introversion Extraversion

High affiliation

Low affiliation

Affection and friendship wanted, but ill at ease in many interpersonal situations (Presidential example: Richard Nixon)

Comfortable and most effective when working alone, not concerned about what others think (Presidential example: Calvin Coolidge)

Unconflicted pursuit of wide-ranging interpersonal relationships (Presidential example: John F. Kennedy)

Well-regarded and adept at interpersonal relations, but not dependent on them (Presidential example: Ronald Reagan)

TRAITS AND MOTIVES 239

Theoretical Background of Our Empirical Tests of the Interaction Hypothesis

According to several recent interactionist conceptions, per- sonality influences the kinds of situations people select, the ways in which they construe or construct these situations, and the reactions they evoke in o thers- - thus shaping a life course over time (D. M. Buss, 1987; Caspi & B e m , 1990; Snyder & Ickes, 1985). Therefore, the effects of traits and motives are best studied longitudinally, as lives unfold over time. Moreover, these effects should be most pronounced for molar (rather than molecular) aspects of behavior. Hence we focus on significant events and outcomes in love and work, the two paramount do- mains of life8: (a) personal relationships, including marriage, divorce, intimacy, and satisfaction-dissatisfaction; and (b) choices in career and other activities outside the home.

In the research reported here, we cannot offer a complete account or test of the many ways in which all motives and all traits interact to affect the life course. Rather, we aim to illustrate the empirical implications of our theoretical conception by ex- amining life outcomes as the cumulation of motive-trai t interac- tions over an extended period of years. By necessity, our work is paradigmatic in method and f indings- -an example pointing toward the kinds of studies that need to be done, rather than a full resolution of all issues and details. We focus on the two main social motives measured with the TAT--aff i l ia t ion and p o w e r - - a n d their interaction with extraversion, the most proto- typical and universally recognized trait in the Big Five and other taxonomies.

Obstacles to Longitudinal Research on Trait-Motive Interaction

Research on the joint effect of traits and motives on life outcomes has been rare, in part because such research must overcome many obstacles. Few longitudinal studies include both trait and motive measures, and even fewer are designed to repli- cate the results of another already existing study. Typically, each longitudinal study has used its own methods, instruments, and variables; even when similar instruments have been used, they have often been given at quite different ages in different studies.

Our research takes advantage of a unique oppor tun i ty - - two longitudinal studies that had three critical design features in common: (a) TAT stories collected during the college years, (b) CPI personality questionnaires administered when participants were in their 40s, from which extraversion could be scored, and (c) similar (though not identical) information about life outcomes, from middle adulthood. This design, realized in two independent studies, permitted us to test, and replicate (cf. Hofstee, 1994), a series of hypotheses about t ra i t -motive interactions.

Advantages of Longitudinal Research

The present research strategy has several advantages. By us- ing already collected longitudinal data, we can largely eliminate experimenter bias and related effects due to the temporal conti- guity of the measurements. By using longitudinal studies, we can focus on molar real-life outcomes of great significance to the participants. Moreover, the fact that the TATs were given

over two decades before the life outcome data were obtained makes the test of motive effects particularly stringent, in contrast to concurrent research. On the other hand, we are constrained to use the kinds of data and measures that were collected, in the ways and at the times they were collected. For example, in both samples we used motive measures from young adulthood and trait measures from middle age. We would have preferred that the trait information also have been obtained from early adulthood. Fortunately, there is evidence that traits such as extra- version are quite stable over time (Conley, 1984; McCrae & Costa, 1984), which suggests that trait scores obtained during early adulthood would have been similar to those at ages 43 or 48.

Finally, there is one important advantage to working with two data sets that are not as tidy as one might like. To turn Hofstee's (1994, p. 135) skepticism on its head, if specific interactive relations replicate across two samples, in spite of differences in measures, cohort and context, and in spite of random noise and the passage of time, then we can be confident that they represent truly robust relationships indeed, to have emerged from all of the "no i s e " and error variance.

M e t h o d

Sample Characteristics

Mills sample. The Mills study is an intensive and extensive study of personality development and creativity in women. It began with study of a representative sample of the Mills College classes of 1958 and 1960 (Helson, 1967). (Descriptions of the Mills study are given in Helson, 1993; Helson, Mitchell, & Moane, 1984; Helson & Moane, 1987; Hel- son & Picano, 1990; Helson, Stewart, & Ostrove, 1995; Helson & Wink, 1992; Mitchell & Helson, 1990; and Wink & Helson, 1993.) At age 21, a subgroup of 51 women came to the Institute for Personality Assessment and Research (1PAR) at Berkeley for an intensive personality assessment, including the TAT. Numerous follow-up studies provided the extensive information about life outcomes that we use in the present analysis. As part of the age 43 follow-up, the women filled out the CPI, which furnished our measure of extraversion. The basic N for the analyses to be reported later was thus 51, reduced for some variables because of missing data or examination of subgroups.

Radcliffe sample. The Radcliffe study of the class of 1964 has also emphasized personality development in women. Virtually the entire Rad- cliffe class of 1964 took a TAT at the beginning of their first year in college (age 18), as a part of the Harvard student study (Stewart & Vandewater, 1993; see also Winter, McClelland, & Stewart, 1981, chap. 4). When the women were 31, Stewart (1978, 1980) began the first of several mailed follow-up studies of these women. (Descriptions of the Radcliffe project are given in Helson et al., 1995; Stewart & Ostrove, 1993; Stewart & Salt, 1981; and Stewart & Vandewater, 1993.) The age 43 follow-up included several open-ended measures assessing life outcomes, and the age 48 follow-up (N = 89) included the CPI, which furnished our measure of extraversion. The basic N for the analyses to be reported later was thus 89, reduced for some variables because of missing data or examination of subgroups.

Certain common and diverging characteristics of both samples should be noted. Not surprisingly for studies begun in the 1950s and early 1960s, most of the participants in both samples were White, and the majority were from middle- and upper-middle-class backgrounds. Both schools were officially women's colleges, but by 1960 Radcliffe had

8Erikson (1963, pp. 264-265) cited Freud as the source of this famous formulation of psychological health.

240 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

become "coordinate" with Harvard. Students shared classes, labs, and many facilities (but not yet dormitories or the two undergraduate librar- ies). Both groups of women graduated from college before the beginning of the contemporary women's movement, although members of the Rad- cliffe class of 1964 were younger and presumably better able to take advantage of the opportunities associated with that movement during the 1970s than were the Mills graduates of 1958 and 1960 (see Duncan & Agronick, 1995).

As is always the case in longitudinal studies, there are missing data for particular measures given at particular times, and a general tendency for some attrition over time. In the present case, however, there is no evidence of any systematic sample missing data or attrition effects (see Helson, 1993; Helson & Moane, 1987; Stewart & Vandewater, 1993).

Affiliation and P o w e r Motives

Five standard TAT cards (6BM, 12F, 13G, 19, and 20) were individu- ally administered to the Mills College students as part of the IPAR assessment at age 21. The Radcliffe students wrote TAT stories to six pictures especially developed for the Harvard student study, in group testing sessions during their first weeks of college. All TAT stories from both samples were scored for affiliation and power motive imagery, using the "running text" method (Winter, 1991), by trained scorers who had previously demonstrated their scoring reliability by attaining levels of agreement of .85 or above (category agreement; see Smith, 1992, p. 529) with materials precoded by expert scorers. In other words, the TAT scoring was highly reliable. The resulting scores were expressed in terms of images per 1,000 words.

Extraversion

Our measure of extraversion was obtained from the CPI, which was given to the Mills sample at age 43 and to the Radcliffe sample at age 48. Although the CPI was constructed to measure several "folk con- cepts" (i.e., ordinary language descriptions) of personality, Gough (1987, 1989) later used clustering techniques to construct scales for three "basic themes," "structural dimensions," or "vectors." Vector 1 is a dimension of interpersonal style, labeled internality versus externality (see also Thorne & Gough, 1991). Although Gough (1989, p. 69) did not claim that the CPI measures traits in the traditional sense, his characterization of Vector 1 is very similar to the traditional trait dimen- sion of introversion-extraversion. 9 Scores indicating externality are said to be associated with a "more outgoing, involved, interpersonally re- sponsive orientation," whereas scores indicating internality are said to indicate an "inward and private" perspective and an "introversive tem- perament" (Gough, 1989, pp. 78, 79). Moreover, typical Vector 1 items (Gough, 1989, p. 78) are quite consistent with the extraversion-intro- version items in Table 2: "I like to be the center of attention" (exter- nality); "I usually don't like to talk much unless I am with people I know very well" (internality). Additionally, in his review of personality trait theory and research, John ( 1990, pp. 88-89) has argued that Vector 1 shares many conceptual features with the extraversion factor in the Big Five taxonomy of traits, and a recent study by McCrae, Costa, and Piedmont (1993) has confirmed this hypothesis, using both rational and empirical means. FOr all these reasons, therefore, participants' scores on CPI Vector 1, keyed in the externality direction, were taken as our measure of extraversion.

General Hypotheses About Trait × Mot ive Interaction

Effects on Life Outcomes

Tables 4 and 5 summarize, in broad outline, our hypotheses about the interaction of the affiliation and power motives with extraversion, in terms of six variables assessing relationship and work outcomes in the lives of the Mills and Radcliffe women at age 43. These tables summarize

the expected interactive effects, that is, the ways in which the same motive (affiliation or power) can be channeled quite differently by extra- version or introversion.

Specific Life Outcome Measures and Specific Hypotheses

Life outcome measures relevant to the affiliation motive and extraver- sion. For both samples, data about various life outcomes were based primarily on questionnaires mailed in the age 43 follow-up studies, supplemented by other data from the longitudinal archives. For our pur- poses, information about relationships and work was of particular inter- est, because these two behavior domains are especially relevant both to extraversion and to the affiliation and power motives. To the extent possible, we used this information to code specific life outcome variables that were comparable across the two samples.

To measure the pursuit of interactional activit ies--a behavior domain that should attract women who are both extraverted and affiliation- motivated--we used two measures: (a) doing volunteer work and (b) combining work and family roles. Our general line of reasoning about motives and traits suggests two specific hypotheses.

1. Because volunteer work involves a good deal of direct interaction with other people and is completely voluntary (in contrast to jobs that require such interaction), it should be especially attractive to extraverted affiliation-motivated women. Introverted affiliation-motivated women, in contrast, should focus their affiliation goal strivings inward and avoid extra interaction activities beyond those minimally required by work or family roles, or both. (Instead, they might donate money to a charitable organization.) Volunteer work was measured by a yes or no question at Mills, and by the reported number of hours spent per week in volunteer activities at Radcliffe.

2. Because combining work and family roles involves a good deal of arousal or "juggling" (Crosby, 1991) of time and energy demands from others, it should be especially acceptable--even attractive--to extraverted affiliation-motivated women. In contrast, introverted affilia- tion-motivated women should find such added arousal aversive; to reduce the stimulation, they should channel their affiliation motive to only one or the other setting. Role combination, or the tendency to be involved in multiple (work and family) roles at once, was coded in both samples as "present" for women who at age 43 were both working (either full- or part-time) and had children, and "absent" for all other women.

Close personal relationships are clearly relevant to both extraversion and the affiliation motive, but in slightly different ways. Affiliation- motivated people should want such relationships, and extraverted people should be adept and comfortable with them. In combination, therefore, the two variables should predict having unconflicted intimate relation- ships. On the other hand, affiliation-motivated introverts, with their more inward focus and preferences for privacy and solitude, should find such relationships overarousing, troublesome, conflicted, and even at times aversive. (In terms of the paradoxical findings of previous research on the affiliation motive, this subgroup should include people who, like Nixon, become prickly and defensive under threat, perhaps because as introverts they feel more intrusion and "threat" in close relationships.) To a partner they may seem remote or even withholding. If they are lucky, they may find an understanding partner, with whom they develop a deep, intense relationship--perhaps after one or two false starts. On average and over time, they may tend to establish intimate relationships that become problematic and fail. For women low in affiliation motiva- tion, intimate relationships are simply less salient and emotionally charged goals.

For the Mills women, the available data allowed us to code whether

9 The directionality is opposite, so that high Vector 1 scores indicate introversion and low scores indicate extraversion. In the present research, we simply reversed Vector 1 scores to get a measure of extraversion.

TRAITS AND MOTIVES

Table 4 Hypotheses About the Interactions of the Affiliation Motive and Extraversion-Introversion

241

Behavioral domain Introversion Extraversion

Pursuit of interactional opportunities outside the home

Intimate relationship outcomes

Combined with high affiliation motive

Desire for interaction, but defensive and avoidant because Unconflicted desire for interaction leads to many of threat and overarousal interpersonal activities

Desire for intimate relationships, but difficulty in maintaining them because of focus on internal world and threat of overarousal

Desire for intimate relations leads to a single stable relationship

Pursuit of interactional opportunities outside the home

Intimate relationship outcomes

Combined with low affiliation motive

Social interaction less salient as a desire, goal, opportunity, Social interaction less salient as a desire, goal, opportunity, and feature of life situation; also, arousal properties of and feature of life situation; although social interaction social interaction may be aversive may be sought for its arousal properties

Intimate relationships less salient as a desire or goal; also, Intimate relationships less salient as a desire or goal, arousal properties of intimacy may be aversive although intimacy may be sought for its arousal

properties

they were married with children by age 28 but then had been divorced by age 43. For the Radcliffe women, we examined the number of mar- riages and divorces by midlife. In addition to using these relatively objective demographic data, we also assessed more subjective aspects of relationship difficulty. At Mills, we measured dissatisfaction with the participants' current marriage or marriage-like relationship on a 5-point scale. At Radcliffe, we relied on the number of "intimacy low points" mentioned in response to the question, "What do you consider the low points, or the most disturbing or upsetting aspects, of the last 10 years? Please include as many things, and as much detail, as you can." Low points were coded for intimacy according to the criteria of Stewart, Franz, Paul, and Peterson (1991; see also Stewart, Franz, & Layton, 1988). Examples were " I ' m distressed with the lack of communication between my husband and me," and "Being asked for a divorce with no warning was devastating."

Life outcome measures relevant to the power motive and extraversion. To assess career characteristics relevant to the power motive and extra- version, we used two variables. Previous research has shown that power- motivated women and men tend to enter impact careers, such as business executive, education, psychology and psychotherapy, journalism, and the clergy, in which they can directly and legitimately seek to influence other people (Jenkins, 1994; Winter, 1973, chap. 4; Winter & Stewart, 1978). Such careers should be particularly attractive to power-motivated women who are also high in extraversion; on the other hand, for power- motivated introverts who strive for more inwardly channeled power goals, these careers may actually be aversive. It should be noted that

the Radcliffe women were more likely to have careers of higher status or prestige than were the Mills women. Factors contributing to this difference include their more privileged backgrounds and their younger age, which positioned them to take advantage of the changes brought about by the women's movement of the 1970s to a greater degree than did the Mills women, who graduated from college to young adulthood 4 to 6 years earlier.

In addition to the relatively objective measure of impact career, we also used a more subjective measure of the importance of relationships at work. At Mills, this measure was based on two items that asked the women how much they valued relational aspects of work (e.g., helping others, meeting people) in their current job. Women in the Radcliffe sample were asked to rate, on a 5-point scale, how important a variety of features "have been or will be in influencing your choice of a field of work." To measure the importance of relationships at work, we used the mean rating of several features that constituted a coherent factor, including "involves helping others," "allows for meeting different and interesting people," and "involves work with people rather than alone." Coefficient alpha for the three-item composite was .68.

R e s u l t s

Correlations Between Motives and Extraversion

Before examin ing the p red ic ted in teract ion effects , we tes ted whe the r the TAT-based social mot ives were indeed unre la ted to

Table 5

Hypotheses About the Interactions of the Power Motive and Extraversion-lntroversion

Behavioral domain Introversion Extraversion

Pursuit of high-impact careers, and relational aspects of work

Combined with high power motive

Desire for impact, but avoidance of the external world leads to avoidance of an impact career, and low valuing of relational aspects of work

Unconflicted desire for impact leads to an impact career and valuing the relational aspects of work as instrumental to this goal

Pursuit of high-impact careers, and relational aspects of work

Combined with low power motive

Desire for impact less salient as a feature of career Desire for impact less salient as a feature of c a r e e r

242 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

extraversion, as previous research has suggested. Table 6 shows that the affiliation and power motives were not related to extra- version in either sample. These results are consistent with previ- ous research and, again, support our contention that the trait and TAT motive measures represent different classes of psycho- logical variable.

Strategy of Interaction Analyses

Data for the Mills and Radcliffe samples were analyzed sepa- rately. Each dependent variable was regressed on standard- scored extraversion, the standard-scored motive (affiliation or power), and their interaction (the product of standardized mo- tive and standardized extraversion scores), entered in the Friedrich solution (Aiken & West, 1991). Our general prediction for the life-outcome variables was that the trait-motive interac- tion terms would be significant. Main effects of either extraver- sion or the motives were of secondary importance here, although they are included in the results. The basic findings are presented in Tables 7 and 8, showing the regression coefficients for the standardized variables (beta weights), which can also be inter- preted as effect size estimates.

Following Aiken and West (1991), Figures 1, 2, 3, 4, 5, and 6 explicate the predicted interaction effects by showing the regression lines relating motive scores to life-outcome variables, separately for women with extraversion scores of one standard deviation above the mean (labeled extraverts in the figures) and one standard deviation below the mean (labeled introverts in the figures). As a guide to understanding these figures, the angle of intersection between the two regression lines reflects the size of the interaction between motive and trait in explaining variance in the outcome variables. When no interaction is present, the two lines will be parallel. (In order to calculate the regressions and draw the figures, the dependent variables have been stan- dardized even though they are essentially dichotomous. For pres- ent purposes, the scale units of the dependent variables are arbitrary: What is important is the magnitude of the interaction term, as shown in the tables, which is represented by the angle between the regression lines in the figures.)

Interaction Between the Affiliation Motive and Extraversion

Table 7 presents the results concerning the prediction of inter- personal activities from the affiliation motive and extraversion, and their interaction. Note that the interaction of affiliation moti-

Table 6 Empirical Relations Between Extraversion and Social Motives in Two Samples of Midlife Women

Correlation of extraversion with

Sample Affiliation motive Power motive

Mills .10 -.09 Radcliffe .02 .07

Note. Motives were assessed at age 21 in the Mills sample (N = 51) and at age 18 in the Radcliffe sample (N = 89); extraversion was assessed at age 43 for Mills and at age 48 for Radcliffe.

vation and extraversion was consistently related to both interper- sonal activity measures, in the same way, for both samples. As shown in the diagrams of these interactions presented in Figures 1 and 2, among extraverts, affiliation motivation was related positively to both volunteer work and role combination. Among introverts, in contrast, the affiliation motive showed a negative relation to these same variables. In other words, extraverted women high in affiliation motivation tended to seek out opportu- nities for social contact and relationships outside the home, whether through volunteer work or through paid work in addi- tion to having a family.

The bottom part of Table 7 shows the complementary pattern for problems in intimate relationships, where all four interaction regression weights were negative. Thus among introverted Mills and Radcliffe women, affiliation motivation predicted entering marriage that later dissolved in divorce, as well as subjective distress and dissatisfaction with intimate relationships in gen- eral, as illustrated by the interactions effects in Figures 3 and 4. For affiliation-motivated introverts, it seems, intimate rela- tionships are desired or needed, but such relationships are also conflicted, aversive, and perhaps even threatening. In contrast, among extraverts the affiliation motive tended to be negatively related to these same outcomes: If they enter relationships, they are more likely to keep them and less likely to lose them.

Interaction Between the Power Motive and Extraversion

Table 8 presents the results concerning interactions of extra- version with the power motive. In the case of having an impact career, the predicted interaction between extraversion and power motivation occurred only among the Radcliffe women. There were marginally significant main effects for power motivation among the Mills and Radcliffe women, and for extraversion among the Radcliffe women. This means that, as predicted, for extraverted Radcliffe women power motivation was positively related to having an impact career, but among introverts the relationship was reversed. However, for the Mills women, the two lines for extraverts and introverts were parallel (see Figure 5), indicating that among both extraverts and introverts, power motivation was positively related to having an impact career.

Why did we obtain an interaction effect for Radcliffe but not for Mills? This difference may be related to differences between Radcliffe and Mills women in the levels or prestige of the impact careers. Typically, Radcliffe women entered impact careers tra- ditionally dominated by men (e.g., psychiatrist, college profes- sor), whereas Mills women entered similar careers with some- what lower prestige value that are more traditionally filled by women (e.g., social worker, elementary school teacher). Trying to enter a field traditionally dominated by a group (gender or race) other than one's own may require that one be especially assertive, active, or energetic and not especially quiet, reserved, shy, and silent (specific traits associated with extraversion and introversion, respectively; see Table 2). We suggest that for Radcliffe women more than for Mills women, impact careers may demand both variables: power motivation because of their potential for impact, and extraversion because (for this cohort, at least) of their traditional domination by men.

Finally, for both samples the interaction between extraversion and power motivation was related to viewing relationships as

TRAITS AND MOTIVES

Table 7 Effects of Affiliation Motivation, Extraversion, and Their Interaction on Relationships and Activities in Two Samples of Midlife Women

Dependent variable and sample

Standardized regression weights

Motive Trait Motive X (affiliation) (extraversion) Trait

Pursuit of interaction activities

Volunteer work (Mills) - .14 Volunteer work (Radcliffe) .02 Combining family and work roles (Mills) .07 Combining family and work roles (Radcliffe) .09

.04 .43*

.35t .47*

.14 .29t

.23* .25*

Difficulty in maintaining intimate relationships and distress about them

Marriage and family by age 28, followed by divorce (Mills) .04 - .09 -.29* Number of marriages and divorces (Radcliffe) .27* - . 11 -.26* Dissatisfaction with intimate relationships (Mills) .08 - .04 -.47* Intimacy low points (Radcliffe) .04 -.03 - .23t

Note. Ns = 51 (Mills) and 89 (Radcliffe), ranging downward for some variables because of missing data. "~p < .10. *p < .05, two-tailed.

243

an important part of work life, as diagrammed in Figure 6. That is, for power-motivated extraverts, relationships are an important part of work; but for power-motivated introverts, work relation- ships are actually not important.

D i s cus s ion

We first discuss these results and their implications in relation to the general conception of traits and motives and the specific hypotheses advanced in this article. Then we consider the impli- cations of these results for a broader understanding of two of the personality variables (the affiliation motive and extraver- sion) that we studied. Finally, we conclude with a discussion of the implications of our results for the intensive, longitudinal study of other traits and other life outcomes, with other groups of people.

Table 8 Effects of Power Motivation, Extraversion, and Their Interaction on Work Outcomes in Two Samples of Midlife Women

Standardized regression weights

Dependent variable Motive Trait Motive × and sample (power) (extraversion) Trait

Impact career Mills .32t .05 -.01 Radcliffe .18t .26* .25*

Importance of work relationships

Mills - .39" .29t .31 * Radcliffe -.08 .19t .31 *

Note. Ns = 51 (Mills) and 89 (Radcliffe), ranging downward for some variables because of missing data. t p < .10. *p < .05, two-tailed.

The Role o f Traits in Modifying Motives

Taken together, these results strongly support our basic hypotheses that traits and motives interact in important ways and that traits channel the ways in which motives are expressed in behavior and life outcomes. Thus, behaviors involving activi- ties, relationships, and careers are related to the affiliation and power motives in ways that are predictably different for people scoring high versus low on extraversion. Because their interests and focus of attention are directed outward, extraverted women tend to be interactive, expressive, and surgent. As a result, their affiliation and power motives find easy and unconflicted chan- nels into behavior and life outcomes involving interpersonal activities, relationships, and impact careers (in which relation- ships are important) in fields that are nontraditional for women. In contrast, because their interests and focus of attention are directed inward, introverted women are less interactive, less expressive, and less surgent. Yet this difference in orientation does not negate or cancel their affiliation or power motives; rather, it modifies or channels them in ways that are more com- patible with their introversion.

The present findings provide longitudinal support for our hy- pothesis that including a measure of traits permits us to make more refined predictions of life outcomes from measures of motives or goals. That is, traits constitute the stylistic context for the expression of motives. Obviously most behaviors are multiply determined, and the particular motive- t ra i t interactions demonstrated in the present article are only a small subset of the many interactions of personality elements (traits, motives, cognitions), and of personality and context, that cumulate in life outcomes. Even for the life outcome variables used in the present research, many other personality and contextual vari- ables doubtless make direct and interactive contributions. On the other hand, it is by no means the case that any two personal- ity variables show interactive effects in the prediction of an outcome. Thus with the Mills sample, we found no interaction effects for the ACL Affiliation scale and extraversion on any of

244 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

0.5

o

-0.5

-1

R a d c l i f f e

.* •°

°°°° .°

° * ° .-

. . " °

I I I -1 0 1

Affiliation motive

Extraverts

Introverts

culty in intimate relationships), but we do not believe that is necessarily the case. To some extent, the present findings reflect the characteristics of the available data, which involved mostly the externalities of life. People's extraverted acts and outcomes are so much easier to measure than are their inner lives! How- ever, we should remember, as F. H. Allport and G. W. Allport (1921) long ago noted, that "it is not so much the absence of overt behavior that marks the introvert as the presence of rich and persistent internal responses" (p. 13). The present results may also be an artifact of the Western cultural context of the research. Particularly in the United States, extraversion is often taken as synonymous with "normal" or "healthy" (Winter, 1996a, pp. 459-460). Even Freud (1916-1917/1961-1963) thought that introverts were "in an unstable situation" and "sure to develop symptoms" (p. 374). Jung (1923/1971) dis- agreed, arguing that introversion and neurosis were unrelated. In some other societies (e.g., China), however, these evaluations are reversed: Introversion is considered healthy, and extraverts are thought to be maladjusted.

0.5

-0.5

-1

Mills

Extraverts

Introverts

-1 0 1 Affiliation motive

Figure 1. Joint effects of the affiliation motive and extraversion on volunteer activity replicated across the Radcliffe and Mills samples.

0.5 g~

I o

-0.5

-1

R a d c l i f f e

•°°

I I I -1 0 1

AWdiationmotive

~'traverts

Introverts

the life outcomes shown in Table 7 and Figures 1-4. Such a result supports the discriminant validity of the particular mo- tive-trait interactions, measured in the particular ways, of the present research.

Finally, these results may help to explain an enduring paradox of affiliation motive research. From its earliest days, the affilia- tion motive has occasionally shown negative relationships to popularity and similar variables (Shipley & Veroff, 1952; see also Atkinson, Heyns, & Veroff, 1954; for a review, see Boy- atzis, 1973). The present findings suggest that such a paradoxi- cal relationship may hold especially among introverts, whose stylistic traits may create special challenges or problems for expression of the affiliation motive.

Introversion Reconsidered

Some aspects of the present findings suggest that the channel- ing effects of introversion are largely negative ones (e.g., diffi-

I ,.Q

!

0.5

-0.5

-1

Mills

° . ° ° . .o°°"

.° •- .° . ° . ° - "

-1 0 1 Affiliation motive

~ v e r l ~

Introverts

Figure 2. Joint effects of the affiliation motive and extraversion on role combination replicated across the Radcliffe and Mills samples.

TRAITS AND MOTIVES 245

z

0.5

-0.5

R a d c l i f f e

-1 I I

-1 0 Aff i l ia t ion m o t i v e

Introverts

Extraverts

[

1

Given its preeminent status as something like a universal trait, extraversion-introversion probably channels many different motives. Would other traits also act as moderator variables in this way? Agreeableness (Factor II in the FFM) ought to be an important moderator of social motives such as affiliation and power. And high conscientiousness (Factor III) might channel TAT-measured achievement motivation into aspirations for con- ventional competence and efficiency, whereas low conscien- tiousness might channel it into cheating and illegal short-cuts (see McClelland, 1976, pp. B - C ; Mischel & Gilligan, 1964). Power motivation, also, would be quite different depending on whether it is combined with high or low conscientiousness.

Finally, neuroticism (or its opposite, emotional stability) might make an enormous difference in how almost any motive is expressed. For example, in combination with high neuroticism, affiliation motivation should be associated with self-rated un- happiness, whereas in combination with low neuroticism it

Mills

o

-0.5

Introverts

Extraverts

-1 I I I -1 0 1

Aff i l ia t ion m o t i v e

Figure 3. Joint effects of the affiliation motive and extraversion on marriage and divorce replicated across the Radcliffe and Mills samples.

0.5

i o

i -0.5

R a d c l i f f e

-1 I I I

-1 0 1 Aff i l ia t ion m o t i v e

Introverts

Extraver~

In any case, introverted affiliation motivation may relate to aspects of intimate relationships that are less easily measured than simple frequency, duration, and conflict. Thus, if introver- sion creates some problems in intimate relationships for affilia- tion-motivated women, it may also enrich their inner experience of such relationships. The American poet Emily Dickinson (1830-1886) is an illustrative example. For 37 years, she lived at home as a virtual recluse, yet her intense, inward poetry reflected a rich concern with relationships. Future cross-sec- tional and longitudinal research could be designed to explore this and other possibilities.

W h i c h Traits?

In these empirical illustrations of replicated t rai t -motive in- teractions, we have focused on the trait of extraversion and the motives for affiliation and power. All three of these variables are relevant to relationship and career life outcomes, and this information was available to us in two longitudinal data sets.

1 g~

-0.5

Mills

Introverts

Extraverts

-1 I I I -1 0 1

Aff i l ia t ion m o t i v e

Figure 4. Joint effects of the affiliation motive and extraversion on dissatisfaction in intimate relationships replicated across the Radcliffe and Mills samples.

246 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

0.5

-0.5

R a d c l i f f e

Extraverts

Introverts

-1 I I I -1 0 1

P o w e r m o t i v e

0.5

Mills

-1 I I I ~- . r

-1 0 1 P o w e r m o t i v e

Extraverts Introverts

Figure 5. Joint effects of the power motive and extraversion on impact careers replicated across the Radcliffe and Mills samples.

and motive approaches have argued that their favored concept was fundamental, whereas the other was less important or even derivative.

In this article, we have argued that traits and motives are separate personality concepts or elements: Each plays an im- portant but different role in regulating behavior and life out- comes, and neither is reducible to the other. Therefore, any comprehensive conceptualization of personality will, in our view, have to make a place for both traits and motives in order to achieve the ' 'macrovalidation of the central personality struc- ture that all are aiming at" (Loevinger, 1993, p. 5). Using longitudinal data, we have demonstrated in two samples that two motives (affiliation and power) and one trait (extraversion- introversion) have interactive effects on life outcomes over and above their separate individual effects. These replicated results not only support our specific conception of traits and motives, but should also encourage a psychology of personality that is sophisticated and open in theory, complex and interactive in research, and cooperative in spirit. As G. W. Allport (1946)

R a d c l i f f e

0 .5

-0 .5

Extraverts

Introverts

-1 I I l

-1 0 1 P o w e r m o t i v e

should be associated with happiness (over and above the inde- pendent contribution of neuroticism).

The present results are based on two longitudinal studies of women. Similar studies of the interaction of traits and motives in men, using appropriate longitudinal data sets, could explore the generality of these findings. More detailed path analysis of longitudinal data sets could chart the ways in which motive- trait interactions of the sort illustrated at the macro level by the present study are built up by the cumulation, over years, of microprocesses such as selecting and construing situations, evoking reactions, and exploiting affordances of the environ- ment (cf. Caspi & Bern, 1990).

Conclusion

In recent years, personality psychologists have sought to reach consensus about the fundamental units, elements, or levels of analysis in their discipline (see McAdams & Emmons, 1995; Winter, 1996a; Winter & Stewart, 1995). Almost from the begin- ning of the scientific study of personality, partisans of the trait

Mills

0-¢

Extraverts

| o

-0 .5

Introverts -1 I I

-1 0 1 P o w e r m o t i v e

Figure 6. Joint effects of the power motive and extraversion on impor- tance of relationships in work replicated across the Radcliffe and Mills samples.

TRAITS AND MOTIVES 247

emphasized so long ago in his essay on personality psychology as a science, "no doors should be closed in the study of person- ality" (pp. 133-134).

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

Allport, F. H. (1937). Teleonomic description in the study of personality. Character and Personality, 5, 202-214.

Allport, E H., & Allport, G. W. ( 1921 ). Personality traits: Their classi- fication and measurement. Journal of Abnormal and Social Psychol- ogy, 16, 6-40.

Allport, G. W. ( 1931 ). What is a trait of personality? Journal of Abnor- mal and Social Psychology, 25, 368-372.

Allport, G. W. (1937). Personality: A psychological interpretation. New York: Holt.

Allport, G.W. (1946). Personality psychology as a science: A reply. Psychological Review, 53, 132-135.

Allport, G.W. ( 1961 ). Pattern and growth in personality. New York: Holt, Rinehart, & Winston.

Allport, G. W. (1966). Traits revisited. American Psychologist, 21, 1 - 10.

Ambrose, S.E. (1987). Nixon: The education of a politician 1913- 1962. New York: Simon & Schuster.

Anderson, W. (Ed.). (1970). Theophrastus: The character sketches. Kent, OH: Kent State University Press.

Atkinson, J. W. (Ed.). (1958). Motives in fantasy, action, and society. Princeton, NJ: Van Nostrand.

Atkinson, J. W. (1982). Motivational determinants of thematic apper- ception. In A. J. Stewart (Ed.), Motivation and society (pp. 3-40) . San Francisco: Jossey-Bass.

Atkinson, J. W., Heyns, R. W., & Veroff, J. (1954). The effect of experi- mental arousal of the affiliation motive on thematic apperception. Journal of Abnormal and Social Psychology, 49, 405-410.

Bakan, D. (1966). The duality of human existence. Boston: Beacon Press.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behav- ioral change. Psychological Review, 84, 191-215.

Benet, V., & Waller, N. G. (1995). The big seven factor model of person- ality description: Evidence for its cross-cultural generality in a Spanish sample. Journal of Personality and Social Psychology, 69, 701-718.

Block, J. (1977). Advancing the psychology of personality: Paradigmatic shift or improving the quality of research? In D. Magnusson & N. S. Endler (Eds.), Personality at the crossroads: Current issues in inter- actional psychology (pp. 37-63). Hillsdale, NJ: Erlbaum.

Block, J. (1995). A contrarian view of the five-factor approach to per- sonality description. Psychological Bulletin, 117, 187-215.

Borkenau, P. (1990). Traits as ideal-based and goal-derived social cate- gories. Journal of Personality and Social Psychology, 58, 381-396.

Boyatzis, R. E. (1973). Affiliation motivation. In D. C. McClelland & R. S. Steele (Eds.), Human motivation: A book of readings (pp. 252- 275). Morristown, NJ: General Learning Press.

Buss, A. H. (1989). Personality as traits. American Psychologist, 44, 1378-1388.

Buss, D.M. (1987). Selection, evocation, and manipulation. Journal of Personality and Social Psychology, 53, 1214-1221.

Buss, D.M., & Craik, K.H. (1983). The act frequency approach to personality. Psychological Review, 90, 105-126.

Cantor, N. (1990). From thought to behavior: "Having" and "doing" in the study of personality and cognition. American Psychologist, 45, 735 -750.

Cantor, N., & Zirkel, S. (1990). Personality, cognition, and purposive behavior. In L. Pervin (Ed.), Handbook of personality theory and research (pp. 135-164). New York: Guilford Press.

Caspi, A., & B e m , D.J. (1990). Personality continuity and change across the life course. In L. Pervin (Ed.), Handbook of personality theory and research (pp. 549-575). New York: Guilford Press.

Cattell, R. B. (1946). Description and measurement of personality. Yon- kers-on-Hudson, NY: World.

Cattell, R. B. (1957). Personality and motivation structure and mea- surement. Yonkers-on-Hudson, NY: World.

Cattell, R. B. (1965). The scientific analysis of personality. Baltimore, MD: Penguin Books.

Chen, W., Ford, K., & Wang, H. (1996). The measurement of need for achievement: A new affective perspective and validation. Unpublished manuscript, Michigan State University.

Combs, A. W. (1947). A comparative study of motivation as revealed in Thematic Apperception stories and autobiographies. Journal of Clinical Psychology, 3, 65-75.

Conley, J. J. (1984). Longitudinal consistency of adult personality: Self- reported psychological characteristics across 45 years. Journal of Personality and Social Psychology, 47, 1325-1333.

Conte, H. R., & Plutchik, R. (1981). A circumplex model for interper- sonal personality traits. Journal of Personality and Social Psychology, 40, 701-711.

Costa, E T., Jr., & McCrae, R. R. (1985). The NEO personality inven- tory manual. Odessa, FL: Psychological Assessment Resources.

Costa, E T., Jr., & McCrae, R. R. (1988). From catalog to classification: Murray's needs and the five-factor model. Journal of Personality and Social Psychology, 55, 258-265.

Crosby, E J. ( 1991 ). Juggling: The unexpected advantages of balancing career and home for women and their families. New York: Free Press.

Demakis, G. J., & McAdams, D. E ( 1992, August). Personality, social support, and well-being among first-year college students. Paper pre- sented at the 100th Annual Convention of the American Psychological Association, Washington, DC.

Duncan, L. E., & Agronick, G. S. (1995). The intersection of life stage and social events: Personality and life outcomes. Journal of Personal- ity and Social Psychology, 69, 558-568.

Emmons, R.A. (1989). Exploring the relations between motives and traits: The case of narcissism. In D. Buss & N. Cantor (Eds.), Person- ality psychology: Recent trends and emerging directions (pp. 32- 44). New York: Springer-Verlag.

Emmons, R.A. (1993). The current status of the motive concept. In K. H. Craik, R. Hogan, & R. N. Wolfe (Eds.), Fiftyyears of personal- ity psychology (pp. 187-196). New York: Plenum.

Emmons, R. A., & McAdams, D. P. ( 1991 ). Personal strivings and mo- tive dispositions: Exploring the links. Personality and Social Psychol- ogy Bulletin, 17, 648-654.

Entwisle, D. R. (1972). To dispel fantasies about fantasy-based mea- sures of achievement motivation. Psychological Bulletin, 77, 377- 391.

Epstein, S. (1984). The stability of behavior across time and situations. In R. A. Zucker, J. Aronoff, & A. I. Rabin (Eds.), Personality and the prediction of behavior (pp. 209-268). Orlando, FL: Academic Press.

Erikson, E. H. (1963). Childhood and society. New York: Norton. Eysenck, H.J. (1953). The structure of human personality. London:

Methuen. Eysenck, H. J. (1992). Four ways five factors are not basic. Personality

and Individual Differences, 13, 667-673. Eysenck, H. J., & Eysenek, M. (1985). Personality and individual dif-

ferences: A natural science approach. New York: Plenum Press. Fiske, D. W. (1994). Two cheers for the Big Five! Psychological Inquiry,

5, 123-124. Freedman, M. B., Leary, T. E, Ossorio, A. G., & Coffey, H. S. ( 1951 ).

The interpersonal dimension of personality. Journal of Personality, 20, 143-161.

Freud, S. (1953). The interpretation of dreams. In J. Strachey (Ed. and Trans.), The standard edition of the complete psychological works

248 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

of Sigmund Freud (Vol. 4-5, pp. 1-627). London: Hogarth Press. (Original work published 1900)

Freud, S. (1961). Civilization and its discontents. In J. Strachey (Ed. and Trans.), The standard edition of the complete psychological works of Sigmund Freud (Vol. 21, pp. 57-145). London: Hogarth Press. (Original work published 1930)

Freud, S. (1961-1963). Introductory lectures on psychoanalysis. In J. Strachey (Ed. and Trans.), The standard edition of the complete psychological works of Sigmund Freud (Vols. 15-16, pp. 1-476). London: Hogarth Press. (Original work published 1916-1917)

Freud, S. (1964). Analysis terminable and interminable. In J. Strachey (Ed. and Trans.), The standard edition of the complete psychological works of Sigmund Freud (Vol. 23, pp. 209-253). London: Hogarth Press. (Original work published 1937)

Funder, D. C. (1991). Global traits: A neo-Allportian approach to per- sonality. Psychological Science, 2, 31-39.

Funder, D. C. (1994). Explaining traits. Psychological Inquiry, 5, 125- 127.

Goldberg, L.R. (1990). An alternative "description of personality:" The Big-Five factor structure. Journal of Personality and Social Psy- chology, 59, 1216-1229.

Goldberg, L. R. (1994). How not to whip a straw dog. Psychological Inquiry, 5, 128-130.

Gough, H. G. (1987). The California Psychological Inventory adminis- trator's guide. Palo Alto, CA: Consulting Psychologists Press.

Gough, H. G. (1989). The California Psychological Inventory. In C. S. Newmark (Ed.), Major psychological assessment instruments (Vol. 2, pp. 67-98). Boston: Allyn & Bacon.

Gough, H. G., & Heilbrun, A. B., Jr. (1983). The Adjective Check ~'st manual. Palo Alto, CA: Consulting Psychologists Press.

Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4-27.

Guilford, J. P. (1959). Personality. New York: McGraw-Hill. Harlow, R.E., & Cantor, N. (1994). A functionalist agenda for trait

psychology. Psychological Inquiry, 5, 130-134. Heckhausen, H. (1991). Motivation and action. New York: Springer-

Verlag. Helson, R.M. (1967). Personality characteristics and developmental

history of creative college women. Genetic Psychology Monographs, 76(2), 205-256.

Helson, R. M. (1993). The Mills classes of 1958 and 1960: College in the fifties, young adulthood in the sixties. In K. D. Hulbert & D.T. Schuster (Eds.), Women's lives through time: Educated American women of the twentieth century (pp. 190-210). San Francisco: Jos- sey-Bass.

Helson, R. M., Mitchell, V., & Moane, G. (1984). Personality and pat- terns of adherence and nonadherence to the social clock. Journal of Personality and Social Psychology, 46, 1079-1096.

Helson, R. M., & Moane, G. (1987). Personality change in women from college to midlife. Journal of Personality and Social Psychology, 53, 176-186.

Helson, R.M., & Picano, J. (1990). Is the traditional role bad for women? Journal of Personality and Social Psychology, 59, 311-320.

Helson, R. M., Stewart, A. J., & Ostrove, J. (1995). Identity in three cohorts of midlife women. Journal of Personality and Social Psychol- ogy, 69, 544-557.

Helson, R. M., & Wink, P. (1992). Personality change in women from the early 40s to the early 50s. Psychology and Aging, 7, 46-55.

Hofstee, W. K. B. (1994). Will the true trait theorist please stand up? Psychological Inquiry, 5, 134-137.

Hogan, R. (1986). Hogan Personality Inventory. Minneapolis: National Computer System.

Hogan, R. (1996). A socioanalytic perspective on the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality: Theoreti- cal perspectives (pp. 163-179). New York: Guilford Press.

Jackson, D. N. (1965). Personality Research Form. Goshen, NY: Re- search Psychologists Press.

Jackson, D. N. (1984). Personality Research Form manual. Port Huron, MI: Research Psychologists Press.

Jenkins, S. R. (1994). Need for power and women's careers over 14 years: Structural power, job satisfaction, and motive change. Journal of Personality and Social Psychology, 66, 155-165.

John, O. P. (1990). The "Big Five" factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of personality theory and research (pp. 66- 100). New York: Guilford Press.

John, O. P., & Robins, R. W. (1993). Gordon Allport: Father and critic of the five-factor model. In K. H. Craik, R. Hogan, & R. N. Wolfe (Eds.), Fifty years of personality psychology (pp. 215-236). New York: Plenum.

John, O. P., & Robins, R. W. (1994). Traits and types, dynamics and development: No doors should be closed in the study of personality. Psychological Inquiry, 5, 137-142.

Jung, C. J. (1971). Psychological types. Princeton, NJ: Princeton Uni- versity Press. (Original work published 1923)

Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214.

Klein, H.G. (1980). Making it perfectly clear. Garden City, NY: Doubleday.

Kornadt, H.-J., Eckensberger, L.H., & Emminghaus, W.B. (1980). Cross-cultural research on motivation and its contribution to a general theory of motivation. In H. C. Triandis & W. Lonner (Eds.), Handbook ofcross-culturalpsychology (Vol. 3, pp. 223-321 ). Boston: Allyn & Bacon.

Kornitzer, B. (1960). The real Nixon: An intimate biography. Chicago: Rand McNally.

Loevinger, J. (1993). Measurement of personality: True or false. Psycho- logical Inquiry, 4, 1-16.

Lundy, A. (1985). The reliability of the Thematic Apperception Test. Journal of Personality Assessment, 49, 141 - 145.

McAdams, D. P. (1982). Intimacy motivation. In A. J. Stewart (Ed.), Motivation and society (pp. 133-171). San Francisco: Jossey-Bass.

McAdams, D. P. (1992). The five-factor model in personality: A critical appraisal. Journal of Personality, 60, 329-362.

McAdams, D. P. (1994). A psychology of the stranger. Psychological Inquiry, 5, 145-148.

McAdams, D. P. (1995). What do we know when we know a person? Journal of Personality, 63, 365-396.

McAdams, D. P., & Emmons, R. A. (Eds.). (1995). Special issue: Lev- els and domains in personality. Journal of Personality, 63 (3).

McClelland, D. C. (1951). Personality. New York: Sloane. McClelland, D. C. (1976). The achieving society (reprinted addition).

New York: Irvington. McClelland, D. C. (1980). Motive dispositions: The merits of operant

and respondent measures. Review of Personality and Social Psychol- ogy, 1, 10-41.

McClelland, D. C. (1985a). How motives, skills, and values determine what people do. American Psychologist, 40, 812-825.

McClelland, D.C. (1985b). Human motivation. Glenview, IL: Scott, Foresman.

McClelland, D.C., Atldnson, J.W., Clark, R.A., & Lowell, E.L. (1953). The achievement motive. New York: Appleton-Century- Crofts.

McClelland, D. C., Koestner, R., & Weinberger, J. (1989). How do self- attributed and implicit motives differ? Psychological Review, 96, 690- 702.

McCrae, R. R. (1994). New goals for trait psychology. Psychological Inquiry, 5, 148-153.

McCrae, R. R., & Costa, P. T. (1984). Emerging lives, enduring disposi- tions. Boston: Little, Brown.

TRAITS AND MOTIVES 249

McCrae, R.R., & Costa, E T. (1996). Toward a new generation of personality theories: Theoretical contexts for the five-factor model. In J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical perspectives (pp. 51-87) . New York: Guilford Press.

McCrae, R. R., Costa, P. T., Jr., & Piedmont, R. L. (1993). Folk con- cepts, natural language, and psychological constructs: The California Psychological Inventory and the five-factor model. Journal of Person- ality, 61, 1-26.

McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60, 175-216.

Mischel, W. (1968). Personality and assessment. New York: Wiley. Mischel, W., & Gilligan, C. (1964). Delay of gratification, motivation

for the prohibited gratification, and responses to temptation. Journal of Abnormal and Social Psychology, 69, 411-417.

Mitchell, V., & Helson, R. M. (1990). Women's prime of life: Is it the 50s? Psychology of Women Quarterly, 14, 451-470.

Morgan, C. D., & Murray, H. A. ( 1935 ). A method for examining fanta- sies: The Thematic Apperception Test. Archives of Neurology and Psychiatry, 34, 289-306.

Murray, H. A. (1938). Explorations in personality. New York: Oxford University Press.

Murray, H. A. (1958). Drive, time, strategy, measurement, and our way of life. In G. Lindzey (Ed.), Assessment of human motives (pp. 183- 196). New York: Holt, Rinehart & Winston.

Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231- 259.

Ostendorf, E, & Angleitner, A. (1994). Enthusiasts contra pessimists. Psychological Inquiry, 5, 159-162.

Patterson, M. L. (1976). An arousal model of interpersonal intimacy. Psychological Review, 83, 235-245.

Paunonen, S. V., Jackson, D. N., Trzebinski, J., & F6rsterling, E (1992). Personality structure across cultures: A multimethod evaluation. Jour- nal of Personality and Social Psychology, 62, 447-456.

Pervin, L. A. (Ed.). (1989a). Goal concepts in personality and social psychology. Hillsdale, NJ: Erlbaum.

Pervin, L. A. (1989b). Goal concepts in personality and social psychol- ogy: A historical introduction. In L. A. Pervin (Ed.), Goal concepts in personality and social psychology (pp. 1-17). Hillsdale, NJ: Erlbaum.

Pervin, L. A. (1993). Pattern and organization: Current trends and pros- pects for the future. In K. H. Craik, R. Hogan, & R. N. Wolfe (Eds.), Fifty years of personality psychology (pp. 69-84). New York: Plenum.

Pervin, L. A. (1994a). A critical analysis of current trait theory. Psycho- logical Inquiry, 5, 103-113.

Pervin, L. A. (1994b). Further reflections on current trait theory. Psy- chological Inquiry, 5, 169-178.

Peters, R. S. (1958). The concept of motivation. London: Routledge & Kegan Paul.

Price, R. (1977). With Nixon. New York: Viking Press. Read, S. J., Jones, D. K., & Miller, L. C. (1990). Traits as goal-based

categories: The importance of goals in the coherence of dispositional categories. Journal of Personality and Social Psychology, 58, 1048- 1061.

Reber, A. S. (1989). Implicit learning and tacit knowledge. Journal of Experimental Psychology: General, 118, 219-235.

Saucier, G., & Goldberg, L. (1996). The language of personality: Lexical reflections on the five-factor model. In J. S. Wiggins (Ed.), Thefive- factor model of personality: Theoretical perspectives (pp. 21-50). New York: Guilford Press.

Schacter, D.L. (1987). Implicit memory: History and current status. Journal of Experimentul Psychology: Learning, Memory, and Cogni- tion, 12, 432-444.

Shapiro, D. (1965). Neurotic styles. New York: Norton.

Shipley, T. E., Jr., & Veroff, J. (1952). A projective measure of need for affiliation. Journal of Experimental Psychology, 43, 349-356.

Simonton, D. (1986). Presidential personality: Biographical use of the Gough Adjective Check List. Journal of Personality and Social Psy- chology, 51, 149-160.

Simonton, D. (1988). Presidential style: Personality, biography, and per- formance. Journal of Personality and Social Psychology, 55, 928- 936.

Smith, C. P. (Ed.). (1992). Motivation and personality: Handbook of thematic content analysis. New York: Cambridge University Press.

Snyder, M. (1994). Traits and motives in the study of personality. Psy- chological Inquiry, 5, 162-166.

Snyder, M., & Ickes, W. (1985). Personality and social behavior. In E. Aronson & G. Lindzey (Eds.), Handbook of social psychology ( 3rd ed., Vol. 2, pp. 883-947). New York: Random House.

Spangler, W. D. (1992). Validity of questionnaire and TAT measures of need for achievement: Two meta-analyses. Psychological Bulletin, 112, 140-154.

Stewart, A.J. (1978). A longitudinal study of coping styles in self- defining and socially defined women. Journal of Consulting and Clini- cal Psychology, 46, 1079-1084.

Stewart, A.J. (1980). Personality and situation in the prediction of women's life patterns. Psychology of Women Quarterly, 5, 195-206.

Stewart, A. J., Franz, C. E., & Layton, L. (1988). The changing self: Using personal documents to study lives. Journal of Personality, 56, 41-74.

Stewart, A. J., Franz, C. E., Paul, E. P., & Peterson, B. E. (1991). Re- vised coding manual for three aspects of adult personality develop- ment: Identity, intimacy, and generativity. Unpublished coding manual. Ann Arbor: University of Michigan.

Stewart, A. J., & Ostrove, J. M. (1993). Social class, social change, and gender: Working-class women at Radcliffe and after. Psychology of Women Quarterly, 17, 475-497,

Stewart, A. J., & Salt, E (1981). Life stress, life-styles, depression and illness in adult women. Journal of Personality and Social Psychology, 40, 1063-1069.

Stewart, A. J., & Vandewater, E. A. (1993). The Radcliffe Class of 1964: Career and family social clock projects in a transitional cohort. In K. D. Hulbert & D. T. Schuster (Eds.), Women's lives through time: Educated American women of the twentieth century (pp. 235-258). San Francisco: Jossey-Bass.

Stumpf, H. (1993). The factor structure of the Personality Research Form: A cross-national evaluation. Journal of Personality, 61, 27- 48.

Thorne, A., & Gough, H. (1991). Portraits of type: An MBT! research compendium. Palo Alto, CA: Consulting Psychologists Press.

Valentine, C.W. ( t961 ). Men of anger and men of shame: Lakalai ethnopsychology and its implications for sociological theory. Ethnol- ogy, 2, 441-477.

Weinberger, J., & McClelland, D. C. (1990). Cognitive versus traditional motivational models: Irreconcilable or complementary? In E. T. Hig- gins & R. M. Sorrentino (Eds.), Handbook of motivation and cogni- tion: Foundations of social behavior (Vol. 2, pp. 562-597). New York: Guilford Press.

Wicker, E W., Lambert, F. B., Richardson, E C., & Kahler, J. (1984). Categorical goal hierarchies and classification of human motives. Journal of Personality, 52, 285-305.

Wiggins, J. S. (1968). Personality structure. Annual Review of Psychol- ogy, 19, 293-350.

Wiggins, J. S. (Ed.). (1996). The five-factor model of personality: Theo- /etical perspectives. New York: Guilford Press.

Wiggins, J. S., & Broughton, R. (1985). The interpersonal circle: A structural model for the integration of personality research. In R. Hogan & W. H. Jones (Eds.), Perspectives in personality (Vol. 1, pp. 1-47). Greenwich, C~ JAI Press.

Wiggins, J. S., & Trapnell, P. D. (1996). A dyadic-interactional perspec-

250 WINTER, JOHN, STEWART, KLOHNEN, AND DUNCAN

tive on the five-factor model. In J. S. Wiggins (Ed.), Thefive-factor model of personality: Theoretical perspectives (pp. 88-162). New York: Guilford Press.

Wink, P., & Helson, R. M. (1993). Personality change in women and their partners. Journal of Personality and Social Psychology, 65, 597- 606.

Winter, D. G. (1973). The power motive. New York: Free Press. Winter, D. G. (1987). Leader appeal, leader performance, and the motive

profiles of leaders and followers: A study of American presidents and elections. Journal of Personality and Social Psychology, 52, 196- 202.

Winter, D. G. (1990, July). Presidential "scores" on the Eysenck Per- sonality Inventory. Paper presented at the annual meeting of the Inter- national Society for Political Psychology, Washington, DC.

Winter, D. G. ( 1991 ). Measuring personality at a distance: Development of an integrated system for scoring motives in running text. In A. J. Stewart, J. M. Healy, Jr., & D. Ozer (Eds.), Perspectives in personal- ity: Approaches to understanding lives (pp. 59-89). London: Jessica Kingsley.

Winter, D. G. (1996a). Personality: Analysis and interpretation of lives. New York: McGraw-Hill.

Winter, D. G. ( 1996b, August). The science of the TAT.. Development of empirically-derived TAT motive measures. Paper presented at the

104th Annual Convention of the American Psychological Association, Toronto, Ontario, Canada.

Winter, D.G., & Barenbaum, N.B. (1985). Responsibility and the power motive in women and men. Journal of Personality, 53, 335- 355.

Winter, D. G., & Carlson, L. (1988). Using motive scores in the psycho- biographical study of an individual: The case of Richard Nixon. Jour- nal of Personality, 56, 75-103.

Winter, D. G., McClelland, D. C., & Stewart, A. J. (1981). A new case for the liberal arts: Assessing institutional goals and student develop- ment. San Francisco: Jossey-Bass.

Winter, D. G., & Stewart, A. J. (1977). Power motivation as a function of retest instructions. Journal of Consulting and Clinical Psychology, 45, 436-440.

Winter, D. G., & Stewart, A. J. (1978). Power motivation. In H. Lon- don & J. Exner (Eds.), Dimensions of personality (pp. 391-447). New York: Wiley.

Winter, D. G., & Stewart, A. J. ( 1995 ). Commentary: Tending the garden of personality. Journal of Personality, 63, 711-727.

Received December 3, 1996 Revision received June 19, 1997

Accepted July 10, 1997 •