The Big Seven Factor Model of Personality Description: Evidence for Its Cross-Cultural Generality in a Spanish Sample

Transcription

1 Journal of Personality and Social Psychology 95, Vol. 69, No. 4,7-7 Copyright 95 by the American Psychological Association, Inc. The Big Seven Factor Model of Personality Description: Evidence for Its Cross-Cultural Generality in a Spanish Sample Veronica Benet and Niels G. Waller University of California, Davis The discovery of the Big Seven factor model of natural language personality description (Tellegen, 93; Tellegen & Waller, 87; Waller, in press; Waller & Zavala, 93) challenges the comprehensiveness of the Big Five factor structure. To establish the robustness and cross-cultural generalizability of the seven-factor model, a Big Seven (Tellegen, Grove & Waller, 91) and a Big Five (John, Donahue, & Kentle, 91) questionnaire were administered to 2 samples: (a) a sample of 569 community-dwelling volunteers from the United States and (b) a sample of 435 Spanish native speakers from Spain. Factor structures from the self- and peer-ratings on the Spanish version of the Big Seven questionnaire largely replicated the American structure (Waller, in press). Nevertheless, some psychologically meaningful item-level differences emerged. These differences suggest that Spaniards attach negative and positive values to self-other perceptions of introversion and unconventionality, respectively. Our findings support the cross-cultural robustness of the Big Seven factors and the advantages of this structure for studying culturally specific differences in personality traitterm evaluations. Personologists from many countries have consulted the natural language when developing personality taxonomies (Allport & Odbert, 36; Angleitner, Ostendorf, & John, 90; Brokken, 78;Cattell, ; Goldberg, 82; Norman, 67; Tellegen & Waller, 87; Wiggins, 79). An important and robust finding from this literature (Fiske, 49; Goldberg, 90, 92; John, Goldberg, & Angleitner, 84; Norman, 63) is that a fivefactor structure, the so called Big Five (Goldberg, 81), or Five Factor Model (FFM; McCrae & Costa, 87), putatively encompasses much of the covariation among self-ascriptions and peer ratings of personality descriptors. Common labels and Roman numerals for these dimensions are: (I) Extraversion or Surgency, (II) Agreeableness, (HI) Conscientiousness, (IV) Emotional Stability or Neuroticism, and (V) Openness to Experience or Intellect (see John, 90, for a scholarly review of these dimensions). The replicability and ubiquity of the Big Five have led many personality psychologists (Digman, 90; Goldberg, 93; John, 90) to advocate this structure as a basic framework-indeed, as the basic framework-for personality description and assessment (McCrae & Costa, 86; McCrae & John, 92). Among the various theoretical and methodological limitations that have been attributed to the Big Five (see Block, 95; We would like to thank Catherine Clark for her helpful comments. Also, we greatly appreciate the help of the psychology faculty at the Universitat Autonoma de Barcelona (specially Jordi Bachs, Montse Goma, Maite Martinez, and Merce Mitjavila), who provided the sample for this study. Correspondence concerning this article should be addressed to Veronica Benet who is now at the Institute of Personality and Social Research, University of California, Berkeley, California 947, or Niels G. Waller, Department of Psychology, University of California, Davis, California 956. Electronic mail may be sent via the Internet to veronica@garnet. berkeley.edu or to ngwaller@ucdavis.edu. Briggs, 92;Loevinger, 94;McAdams, 92; Waller &Ben- Porath, 87), one in particular questions its adequacy to fully represent the personality domain (Tellegen, 93; Waller & Zavala, 93; Waller, in press). Tellegen and Waller (87), for example, noted that the Big Five was originally developed from a pool of personality descriptors that excluded evaluative terms and many state descriptors (Allport & Odbert, 36; Cattell, 43; Norman, 67). These banished terms were not considered representative of the so-called biophysical traits that were advocated by Allport and others (Allport & Odbert, 36; Pervin, 90). Yet, as discussed more fully later, many contemporary personologists (Borkenau, 90; Hogan, 82) believe that personality description and measurement are fundamentally evaluative activities. Recently, this idea has received empirical support from a reexamination of American-English trait terms. In thefirstinvestigation of the nonbowdlerized personality lexicon, Tellegen and Waller (87; Waller, in press) confirmed that the Big Five taxonomy underrepresents or neglects important dimensions and descriptor classes. Specifically, based on an empirical analysis of 4 personality indicators from a representative American dictionary (The American Heritage Dictionary; 85), these authors found that at least seven higher order dimensions are needed in a comprehensive taxonomy of natural language personality descriptors. Tellegen and Waller labeled their dimensions the Big Seven in recognition of the fact that five of their factors were similar to, but not isomorphic with, the higher order factors of the Big Five. The two remaining dimensions in this structure-labeled Positive and Negative Valenceseemingly tap aspects of self-evaluation that are not measured by popular lexically informed personality inventories (e.g., Costa & McCrae, 85; Goldberg, 92; John etal., 91). In this article we claim that the Big Seven factor structure challenges the comprehensiveness of the Big Five. We defend this position by demonstrating the cross-cultural and cross-language robustness of the Big Seven in Spanish and American 7

2 7 VERONICA BENET AND NIELS G. WALLER samples. We also provide evidence for the cross-target robustness of the Big Seven using peer- and self-ratings in the Spanish sample. To further strengthen our claim, we demonstrate that ourfindingscannot be attributed to unrepresentative sampling of participants or personality terms by reporting the factor structure of a recently developed Big Five questionnaire (John et al., 91) in both the Spanish and American data sets. Finally, we report the joint factor structure of the Big Seven and Big Five questionnaires as a means of testing the content overlap of these competing models of lexically derived personality descriptors. In the sections that follow we (a) discuss issues in the development of natural language personality taxonomies, (b) consider why previous lexical studies have failed to uncover the evaluative dimensions of the Big Seven, and (c) review recent findings that support the robustness of these evaluative constructs. Issues in the Development of Natural Language Personality Taxonomies Despite the more than 1 years of the taxonomic tradition (Galton, 84), personality psychologists still disagree over such fundamental topics as the selection, classification, and nature (descriptive vs. explanatory) of the personality terms that warrant inclusion in a comprehensive lexical taxonomy of personality descriptors. For instance, researchers disagree on such basic issues as: (a) the domain coverage of the descriptor taxonomy: Should the taxonomy focus on stable traits, or should it also include state terms and social evaluations? (Allport & Odbert, 36; Waller & Zavala, 93); (b) the most suitable linguistic units for a personality taxonomy: Should the taxonomy focus on descriptive predicates, (i.e., adjectives), or should it also include nouns, verbs, or phrases? (Buss & Craik, 83; De Raad, 92; De Raad & Hoskens, 90; De Raad, Mulder, Kloosterman, & Hofstee, 88; Hofstee, 90); and (c) how to conceptualize the evaluative component in personality descriptors: Does the evaluative component represent substantive variance or individual differences in response-set strength? (Hofstee, 90;Peabody, 67,87; Saucier, 94). To clarify these issues, lexical researchers have recently developed criteria to appraise personality taxonomies. According to one recent proposal (John, Angleitner, & Ostendorf, 88) taxonomies should be evaluated in terms of their generalizability, comprehensiveness, and external validity. The generalizability of a personality taxonomy is assessed by the robustness of its dimensional structure across samples, raters, cultures, and languages. The comprehensiveness of a personality taxonomy refers to its content validity, or its ability to fully represent the domain of personality description. Finally, the external validity of a personality taxonomy refers to the network of theoretical constructs that are predicted by the traits in the taxonomy. In the next section we briefly review alternative personality taxonomies in light of these criteria. Lexically Derived Personality Taxonomies Presently, the Big Five factor structure represents the most popular lexically derived personality taxonomy. The historical development of the Big Five has been chronicled in scholarly reviews by John (90; John et al., 88), Goldberg (93), and Waller (in press), and thus we do not review this history in detail. However, several milestones in this history warrant consideration because they reveal the definitional boundaries of the Big Five and they elucidate why these boundaries do not encompass primary dimensions of self-evaluation. One early milestone that is often not sufficiently acknowledged by advocates of the Big Five is that thefivedimensions were originally 'discovered' from a series of overlapping (Waller & Ben-Porath, 87) descriptor pools that were assembled by Allport and Odbert (36), Cattell (43), and Norman (63, 67). The Herculean efforts of these investigators were energized by a desire to develop a comprehensive taxonomy of natural language personality descriptors. A consistent finding from this work is thatfivebroad dimensions now called the Big Five adequately describe the covariance structure of the trait terms in these descriptor pools (see Goldberg, 93, and references cited therein). On this point there is much agreement. A more contentious issue is whether thesefivefactors "are both necessary and reasonably sufficient for describing at a global level the major features of personality" (McCrae & Costa, 86, p. ). As noted by Tellegen and Waller (87; Tellegen 93; Waller, in press; Waller & Zavala, 93), literally thousands of personality descriptive terms such as evaluative and state descriptors were excluded from consideration in the aforementioned lexical studies. Therefore, considering the banishment of these potentially informative descriptors, and the methodological truism that "Whether or not a given factor appears in a particular study is a direct function of the selection of variables" (Gorsuch, 83, p. 336), one cannot help wondering whether additional factors would have emerged in these studies if the original taxonomers had used less restrictive exclusion criteria when culling candidate trait terms. Motivated by this question, Tellegen and Waller (87) conducted the first study of the American-English lexicon to use purposefully nonrestrictive exclusion criteria when selecting personality indicators. Stated otherwise, these researchers did not exclude the so-called evaluative terms or state terms from their indicator list. Using a stratified sampling method, 4 personality descriptors were chosen nonrestrictively from the American Heritage Dictionary of the English Language (85). When self-ratings on these terms were organized by factor analysis, seven broad personality dimensions emerged (see Waller, in press, for a description and discussion of this work). Tellegen and Waller christened these dimensions the 'Big Seven' because five of their dimensions were similar to the Big Five. The two remaining factors appeared to tap positive and negative aspects of self-evaluation, and consequently they called these factors Positive and Negative Valence. Illustrative markers of Positive Valence include: excellent, special, impressive, skilled, without equal and matchless, whereas representative markers of Negative Valence include: evil, wicked, awful, disgusting, deserve to be hated, and immoral. Rather than consider these dimensions as representing individual differences in response set strength, these authors argued that Positive and Negative Valence represent enduring self construals of global evaluation. This notion that evaluative tendencies represent traits or stable self images is also consistent with recent formulations of the personality disorders. For instance, the Diagnostic and statistical manual of mental disorders (4th ed.) (American Psychiatric Association, 94) criteria for borderline personality disorder notes that a

3 SPANISH BIG SEVEN 7 person diagnosed with borderline personality disorder has "a self-image that is based on being bad or evil...." (p. 1). People with Narcissistic personality disorder view themselves as "superior, special, or unique...." (p. 8). Although these examples illustrate extreme forms of self-evaluation, in a later section we demonstrate that Positive and Negative Valence factors also emerge reliably in nonclinical samples from two cultures. Labels for the other five dimensions were not adopted uniformly from the Big Five tradition because several of Tellegen and Waller's factors differed in important aspects from their Big Five counterparts. For example, in the Big Seven, Extraversion and Neuroticism are called Positive Emotionality and Negative Emotionality, respectively, in recognition of the emotional core of these higher order factors (Tellegen, 85; Watson & Clark, 92). This interpretation of the 'Big Two' is consistent with recent attempts to acknowledge the affective organization of personality (Larsen & Ketelaar, 91; Meyer & Shack, 89; Tellegen, 85; Tellegen & Waller, in press; Watson & Clark, 92). The Big Five Openness to Experience or Intellect factor is called Conventionality in the Big Seven framework because of the large number of conventionality-radicalism terms in the natural language that load on this dimension. Thus, although supporting evidence for the Big Five has accumulated for some time, we believe that the Big Seven provides a more accurate structural representation of natural language personality terms because this broader model was developed from a nonexpurgated descriptor pool. In the Big Seven, no class of personality terms was discriminated against, and consequently, the true diversity of the personality lexicon was allowed to define the natural boundaries of this taxonomy. A fundamental idea in taxonomic personality research, an idea that has come to be known as the lexical hypothesis, is that "[t]hose individual differences that are most salient and socially relevant in people's lives will eventually become encoded into their language...." (Goldberg, 82, p. 2). Accordingly, personality dimensions that are deemed important for large classes of humans across different cultures and historical epochs will be represented by descriptive predicates in many languages (John et al., 84). To date, cross-cultural replications of the Big Five have been carried out in many languages, including German, Dutch, Japanese, Chinese, and Tagalog (Bond, Nakazato, & Shiraishi, 75; Church & Katigbak, 88; John et al., 84; Yang & Bond, 90). Moreover, dictionary-based personality taxonomies in languages other than English have been developed in Dutch, German, and Hebrew (Almagor, Tellegen, & Waller, 93; Angleitner et al., 90; Brokken, 78; De Raad & Hoskens, 90; De Raad et al., 88; see also John et al., 88, for a review). Generally speaking, this work reveals that at least four of the Big Five factors Extraversion, Agreeableness, Conscientiousness, and Emotional Stability are cross-culturally robust (Digman, 90; John, 90; Yang & Bond, 90). Unfortunately, this literature does not speak to the cross-cultural robustness of the Positive and Negative Valence dimensions of the Big Seven for reasons that were adumbrated above. Therefore, we wondered whether these additional natural language dimensions, and the alternative interpretations of the Big Five that are offered by the Big Seven, will emerge in cross-cultural personality research that includes representative trait markers from the Big Seven. With this question in mind, we administered Big Seven and Big Five questionnaires to American and Spanish samples. Participants Method To study the cross-cultural robustness of the Big Seven and Big Five factor structures we obtained personality ratings from native residents of America and Spain. Our American sample included 569 participants (0 men and 429 women) with a mean age of years (SD -.63). These individuals were drawn from the participant pool of the California Twin Registry (Waller, in press). One of the strengths of this registry is that it represents the cultural, educational, demographic, and economic diversity of one of America's most populated states. Our Spanish participants were drawn from two large public universities in Spain: the Universidad Autonoma de Bellaterra and the Universidad de Barcelona. This sample included 435 participants (83 men and 352 women) with a mean age of years (SD = 4.). Measures All participants completed Big Seven and Big Five questionnaires. The Big Seven questionnaire is called The Inventory of Personal Characteristics (IPC-7; Tellegen et al., 91). This instrument contains 1 Likert (4-point) scored items that were developed from previous factor analytic work on American-English, natural language personality descriptors (Tellegen & Waller, 87). Recent work on the IPC-7 (Waller, in press) demonstrates that the items of this questionnaire accurately represent the seven dimensions of the Big Seven. A fuller description of the psychometric properties of the IPC-7, which reports the item-level factor structure, scale reliabilities and heritabilities, is available elsewhere (Waller, in press). The seven high-order factors of this questionnaire are called: (a) Positive Valence (PVAL), (b) Negative Valence (NVAL), (c) Positive Emotionality (PEM), (d) Negative Emotionality (NEM), (e) Conscientiousness (C), (f) Agreeableness (A), and (g) Conventionality (CNV). In our cross-cultural comparison of the Big Seven factor structure, the best markers of each factor from a previous analysis of the IPC-7 (Waller, in press) were selected. This procedure allowed us to work with a 70 item, rather than a 1-item, Big Seven questionnaire. The Big Five questionnaire that we used is called the Big Five Inventory (BFI; John et al., 91). This adjective-based questionnaire was developed from extensive analyses of the Adjective Check List (ACL; Gough & Heilbrun, 83). The 43 adjectives of the BFI have been shown in previous studies (John, 89,90) to be univocal, prototypical markers of the Big Five dimensions: (I) Extraversion (E), (II) Agreeableness (A), (III) Conscientiousness (C), (IV) Neuroticism (N), and (V) Openness (O). A description of the psychometric characteristics of the BFI scales is reported in John et al. (91). Translation of Measures The conceptual equivalence or meaning symmetry of language-based psychological measures is an indispensable requirement for valid crosscultural comparisons (Berry, 80). One-way translation of a personality scale is not sufficient for demonstrating the appropriateness of the scale for cross-cultural research. Rather, cross-cultural researchers must engage the help of bilingual assistants to back translate the items into the original language. Before conceptual equivalence for a questionnaire can be claimed, the original and back-translated items must be semantically isomorphic. During the questionnaire translation phase of our study, the backtranslation methods of Brislin (80) were used. Using standard Span-

4 7 VERONICA BENET AND NIELS G. WALLER ish-english and English-Spanish dictionaries, the first author (VB) translated the IPC-7 and BFI items into Spanish. Next, using the same dictionaries, a bilingual assistant (who holds a Ph.D. in Spanish) independently translated the items into English. When discrepancies occurred, thefirsttranslator reviewed the meaning and psychological implications of the original item with the second translator. This process of translating, and independently back translating, continued until semantic equivalence for all items was achieved. The translated items from the two questionnaires were assembled into a single booklet. Fivepoint and six-point rating scales were used for the Big Seven and Big Five items, respectively. We retained the original response format for these items so that the results of our study would be directly comparable to those of previous studies with these measures. The complete list of IPC-7 items, along with their Spanish translations, is reported in the Appendix. Results To study the cross-cultural and cross-language similarities of the Big Seven and Big Five factor structures we relied on five data sets. These samples included: American and Spanish selfratings on the IPC-7 and the BFI, and Spanish peer-ratings on the IPC-7. Prior to our main analyses the item responses in each data set were ipsatized (within subject standardized) to minimize a response set bias that can obfuscate the underlying factor structure of multipoint personality ratings (Hamilton, 68). Although ipsatization alters correlations in small item pools, beyond that associated with the removal of response set variance (see Dunlap & Cornwell, 94, for a recent review), in large item pools, such as the IPC-7 and BFI, this is not a problem. In our experience and the experiences of other investigators in this area (Goldberg, 92; Peabody & Goldberg, 89; Yang & Bond, 90) ipsatized personality ratings yield 'cleaner' factor structures than nonipsatized ratings. We also examined the nonipsatized factor structures for analytic completeness and found the results of the latter analyses to be essentially identical to those of the former. Moreover, because our samples included men and women who were heterogeneous with respect to age, the linear and quadratic effects of these demographic variables were partialled from the item responses prior to the structural analyses. Failure to consider these sources of item covariation can also obfuscate the factor structure of personality ratings (Waller & Meehl, 94). In the remainder of this section we refer to the IPC-7 and BFI as the Big Seven and Big Five inventories. In our first analysis we examined the factor structure of the Big Seven inventory in the American sample. Our goal in this analysis was to derive a representative, American-based factor structure that could be compared to the Spanish self- and peerrating solutions for the Big Seven items. We felt confident that the 70 items that were chosen for this study would yield a robust seven factor solution because these items were found to be the most salient Big Seven markers in a previous investigation (Waller, in press). In the following analyses, however, we allowed the data, rather than our feeling states or theoretical wishes, to determine the most compelling dimensional representation for these items. The eigenvalue plot for the reduced Big Seven inventory in the American sample indicated that seven factors could account for the primary sources of interitem covariation. The first eigenvalues from the 70 X 70 interitem correlation matrix (with squared multiple correlations in the diagonal) were: 7.78, 4.67, 3.40, 3., 3., 2.23, 1.68, 1., 0.91, Eigenvalue plots are only suggestive, however, rather than definitive regarding reduced matrix rank, and consequently we also assessed the psychological meaningfulness of alternative factor solutions that ranged from five to eight factors. We believe that the varimax-rotated seven-factor solution yields the most compelling structure for these data and that this solution is easily recognizable as the Big Seven. The factor loadings from this analysis are reported in Table 1. The factors in this table were constrained by the rotation method to remain uncorrelated. When the factors were allowed to go oblique, using a Promax rotation, the highest interfactor correlation was only 0.35 between Conscientiousness and Conventionality, and the mean and median interfactor correlations were 0.5 and 0., respectively. Thus, the orthogonal solution accurately reflects the natural organization of the items and is mathematically easier to interpret. Having arrived at a representative factor solution for the American data, we next analyzed the Big Seven item pool in the Spanish sample. The first eigenvalues from the correlation matrix (with squared multiple correlations on the diagonal) of the Spanish self-ratings on the Big Seven were: 6.90, 5.51,4., 3., 2.60, 1.88, 1.52, 1., 0.92, These values show a moderate break after the seventh latent root, suggesting that a seven-factor solution accounts for the brunt of the matrix covariance. To assess this hypothesis more formally, varimax-rotated factor solutions, with five to eight factors, were carefully examined. Consistent with our hypothesis, the seven-factor solution provided the most psychologically compelling structure for these data. This solution is reported in Table 2. A comparison of Tables 1 and 2 reveals a noteworthy degree of cross-cultural and cross-language robustness for the Big Seven factor structure in the self report domain. Several of the corresponding factors between the two samples are virtually identical with one another, although the structures do show several intriguing item-level differences. For example, several (-) Conventionality markers in the American sample, such as Unusual and Odd-Peculiar, emigrated to the Positive Valence dimension in the Spanish sample. These evaluative terms have been interpreted (Waller, in press) as reflecting extreme degrees of psychological Openness (i.e., Unconventionality, in the Big Seven perspective) for American participants. Ourfindingssuggest that for Spaniards, they are related to more positively valued self-portrayals. Interestingly, the term Strange, which is also a (-) Conventionality marker in the American data, emigrated to the negative pole of Positive Emotionality. For Spaniards at least as advertised in popular literature extroverted sociability is a normatively valued personality trait (Hooper, 87; McVeagh, 90). Highly introverted Spaniards, or Spaniards with poor social skills, might harbor feelings of uneasiness and self perceptions of strangeness (see Benet, 94, for an independent replication and similar interpretation of these results). Thesefindings are discussed in more detail in a later section of the paper. First, we examine the factor structure of the Big Seven items using peer ratings in the Spanish sample. In the Spanish peer ratings, the first eigenvalues from the Big Seven correlation matrix (with squared multiple correlations on the diagonal) were: 7.43,4.39,4., 3., 3.40,2.41, 1.66, 1., 1., The overall profile of these eigenvalues is similar to those of the previous analyses and demonstrates

5 SPANISH BIG SEVEN 7 Table 1 Seven Primary Factors of American Natural Language: American Sample, SelfRatings Abbreviated items Outstanding Impressive Excellent Exceptional Admirable Important High ranking Ordinary Average Not exceptional Wicked Awful Deserves to be hated Cruel Dangerous Mentally sick Disgusting Vicious Treacherous Depraved Gregarious Talkative Animated Sociable Peppy Playful Impulsive Not talkative Loner Reserved Quiet Not frustrated Not easily upset Worries aside Not overworrying Feels guilty Sorry for myself Feelings hurt Jittery Nervous Irritated Well organized Tidy Place for everything Orderly Prompt Consistent Improvises Disorganized Argumentative Stubborn Quarrelsome Strong Put up fight Headstrong Tough Lenient Try avoid differences Dislike argument Traditional Strict discipline No permissive parents Conventional Old-fashioned Cautious Progressive Strange Odd Unusual Radical PVAL NVAL PEM NEM C A CNV Note. PVAL = Positive Valence; NVAL = Negative Valence; PEM = Positive Emotionality; NEM = Negative Emotionality; C = Conscientiousness; A - Agreeableness; CNV = Conventionality. All loadings multiplied by 1. Loadings a are in boldface.

6 7 VERONICA BENET AND NIELS G. WALLER Table 2 Seven Primary Factors of American Natural Language: Spanish Sample, Self Ratings Abbreviated items PVAL NVAL PEM NEM CNV Outstanding Excellent Exceptional Admirable Important Unusual Impressive High ranking Odd Ordinary Not exceptional Average Wicked Vicious Awful Cruel Disgusting Treacherous Deserve to be hated Mentally sick Depraved Dangerous Talkative Gregarious Sociable Animated Impulsive Playful No permissive parent Strange Loner Not talkative Reserved Quiet Jittery Feels guilty Feel sorry for self Feelings hurt Irritated Nervous Strong Peppy Not overworrying Worries out of mind Not frustrated Well organized Orderly Place for everything Tidy Prompt Headstrong Cautious Consistent Like to improvise Disorganized Argumentative Quarrelsome Tough Put up fight Stubborn Lenient Try avoid differences Not easily upset Dislike argument Traditional Conventional Old-fashioned Strict discipline Radical Progressive SO Note PVAL = Positive Valence; NVAL = Negative Valence; PEM = Positive Emotionality; NEM = Negative Emotionality; C = Conscientiousness; A = Agreeableness; CNV = Conventionality. All loadings multiplied by 1. Loadings a are in boldface.

7 SPANISH BIG SEVEN 7 Table 3 Seven Primary Factors of American Natural Language: Spanish Sample, Peer Ratings Abbreviated items PVAL NVAL PEM NEM CNV Outstanding Exceptional Important Admirable Excellent Impressive High ranking Ordinary Not exceptional Average Wicked Awful Cruel Vicious Mentally sick Disgusting Deserve to be hated Depraved Treacherous Dangerous Quiet Not talkative Reserved Loner Strange Strong Playful Peppy Impulsive Animated Sociable Gregarious Talkative Guilty Jittery Feel sorry Nervous Feelings hurt Irritated Not overworrying Worries out of mind Not frustrated Well organized Tidy Place for everything Orderly Prompt Headstrong Cautious Consistent Odd Like improvise Disorganized Argumentative Quarrelsome Put up fight Tough Stubborn Lenient Try avoid differences Dislike argument Not easily upset Conventional Traditional Strict discipline No permissive parents Old-fashioned Unusual Radical Progressive Note. PVAL = Positive Valence; NVAL = Negative Valence; PEM = Positive Emotionality; NEM = Negative Emotionality; C = Conscientiousness; A = Agreeableness; CNV = Conventionality. All loadings multiplied by 1. Loadings 2: are in boldface.

8 7 VERONICA BENET AND NIELS G. WALLER Table 4 Five Big Factors of American Natural Language: American Sample, Self Ratings Abbreviated items Quiet Reserved Shy Energetic Enthusiastic Assertive Talkative Sociable Kind Cooperative Forgiving Helpful Trusting Quarrelsome Fault-finding Cold Rude Worrying Nervous Tense Depressed Moody Stable Calm Relaxed Thorough Efficient Perservering Planful Distractible Careless Disorganized Original Inventive Sophisticated Artistic Imaginative Reflective Ingenious Curious Simple Reliable Not artistic E A N C O Note. E = Extraversion; A = Agreeableness; N = Neuroticism; C = Conscientiousness; O = Openness. All loadings multiplied by 1. Loadings a: are in boldface. dimensional robustness across cultures and targets for the seven-factor model. However, dimensional robustness and factorial robustness are not equivalent concepts. Consequently, we also inspected varimax-rotated factor structures for these data and found that the seven-factor solution was easily recognizable as the Big Seven structure. This factor pattern is reported in Table 3. Notice the similarity between the factor loadings in Table 3 with those of Table 2. Table 3 offers the first evidence that the Big Seven factor structure is not unique to the selfrating domain, and that at least for Spaniards, it also accounts for the primary sources of variance in personality peer ratings. However there are a few minor differences between the structures that warrant comment. For instance, in the peer rating solution the items Unusual and Odd are no longer primary markers of Positive Valence, though they do show sizable secondary loadings on this dimension. This suggests that the emigration of these markers from (-) Conventionality to Positive Valence is not aflukeof our self-rating data, but that these quantitative differences between the American and Spanish factor structures reflect qualitative cultural differences in views of the self and others. We believe that the aforementioned analyses support the validity of the Big Seven factor structure in American and Spanish cultures, and in the self and observer domains. Other researchers might claim that our iconoclastic conclusions that is, that the Big Five are not Big enough! reflects unpropitious sampling of items and participants, and that if we had included more representative Big Five markers in our analyses the Positive and Negative Valence factors of the Big Seven structure would have been subsumed by more familiar dimensions. The soundness of this alternative interpretation of our findings can be tested in at least two ways. First, because both the American and Spanish participants in our study completed a well-validated Big Five questionnaire (John et al., 91) we were able to test the "poor subject-sampling" hypothesis directly. For instance, if our analyses failed to uncover a Big Five factor structure for the BFI in the American sample we would suspect the integrity of our data. However, we note that because our study is the first to assess the Big Five model in a Spanish sample, a failure to corroborate the Big Five in the Spanish data would be epistemically ambiguous. We suspected that the probability of such a failure was low because the Big Five factor structure has been replicated in several Western European languages and cultures (Caprara, Barbaranelli, Borgogni, Perugini, 93; John et al., 84). Nevertheless, it was important to test directly this hypothesis. Table 4 reports the varimax-rotated factor structure for the Big Five inventory in the American sample. The first eigenvalues of the correlation matrix (with squared multiple correlations on the diagonal) were: 4.98, 3.60,2.69,2., 1.67,0.88, 0.63, 0.49, 0., As predicted, a noticeable break occurs after the fifth eigenvalue, a finding that supports afivedimensional structure for this questionnaire. When factor solutions that ranged from four to six factors were carefully examined, the varimax-rotatedfive-factorsolution made the most psychological sense. As reported in Table 4, the 43 items of the BFI provide a well-defined, simple-structure representation of the Big Five factor model. Only one item, Reliable, failed to load on its targeted dimension (Conscientiousness); and thus 98% of the predicted loadings were realized. We conclude from this result that the individuals in our American sample are sufficiently diverse with respect to the Big Five dimensions to ensure that a robust Big Five factor structure could emerge in item pools with sufficient item coverage. Of course, a Big Seven Factor structure could also emerge given sufficient representation of the Positive and Negative Valence dimensions. A logical conclusion, then, is that our previously reported seven dimensional solution for the American self-rating data on the Big Seven questionnaire cannot be attributed to quirky sampling characteristics of our participant pool. Having satisfied ourselves that the Big Five questionnaire provides a compelling Big Five structure in our American sample, we next tested whether a similar structure would emerge

9 SPANISH BIG SEVEN 7 Table 5 Five Big Factors of American Natural Language: Spanish Sample, SelfRatings Abbreviated items N O Sociable Talkative Enthusiastic Energetic Curious Shy Reserved Quiet Kind Trusting Helpful Cooperative Forgiving Fault-finding Cold Quarrelsome Rude Thorough Planful Persevering Efficient Reliable Distractible Careless Disorganized Nervous Moody Depressed Worrying Tense Assertive Calm Relaxed Stable Inventive Imaginative Original Artistic Ingenious Sophisticated Reflective Simple Not artistic Note. E = Extraversion; A = Agreeableness; C = Conscientiousness; N = Neuroticism; O = Openness. All loadings multiplied by 1. Loadings a are in boldface. from the Spanish self-ratings. The first eigenvalues from the correlation matrix (with squared multiple correlations on the diagonal) for these data were: 4.92,3.60,2.55,2.,1.55,1., 0.68, 0.56, 0.52, 0.. The moderate break between the fifth and sixth eigenvalues in the eigenvalue profile suggests that a five dimensional solution is plausible. This hypothesis was corroborated when we examined alternative varimax-rotated solutions for the Big Five items. The final factor structure for the Spanish self-ratings in the Big Five inventory is reported in Table 5. As can be seen in this table, the major features of the Five Factor model are immediately recognizable. Although a couple of items failed to load strongly on their targeted dimensions (e.g., Curious, Assertive), for the most part these data provide additional evidence for the Big Five model when the item pool is restricted to carefully selected Big Five markers. A second means of evaluating the structural validity of our results, and the domain coverage of our Big Seven item pool, is to perform a joint factor analysis of the Big Seven and Big Five questionnaires. This analysis was performed in the American and Spanish data sets. For the American data, the first eigenvalues of the 3 X 3 item correlation matrix (with squared multiple correlations on the diagonal) were:.41, 7.78, 5., 4.75, 3.74, 3.35, 2., 1.83, 1.54, The corresponding eigenvalues in the Spanish sample were:., 8.95, 5.98, 5.46, 3.30, 2.75, 2., 1.84, 1.72, Although these eigenvalue profiles are less informative than those previously reported (in the sense that they do not suggest strongly seven- or five-factor models), when the various rotated solutions were inspected the evidence in favor of the seven-factor model was overwhelming. For instance, when we examined thefive-dimensionalsolution for the joint factor analysis in the American data, 50% of the Big Seven markers for Positive and Negative Valence failed to load saliently (defined as a loading S:.30) on any of the five dimensions. Similar results were found in the Spanish data. Tables 6 and 7 report the varimax-rotated, seven-factor solutions from these analyses. These tables reveal that even when univocal markers of the Big Five factor structure are included with Big Seven markers, a robust and psychologically imperious Big Seven structure emerges. Notice that in the joint factor analyses the Positive and Negative Valence factors are defined almost exclusively by Big Seven items. Indeed, only one Big Five marker, Ingenious, loaded on Positive Valence in the American and Spanish solutions, and none of the Big Five markers loaded on Negative Valence. Thisfindingrepresents cogent evidence that Positive and Negative Valence are psychologically differentiated personality constructs that cannot be subsumed by the Five Factor model. These analyses offer additionalfindings that warrant comment. For example, paralleling the Spanish peer- and selfreport data on the Big Seven inventory, in the Spanish joint factor analysis, the terms Unusual and Odd-Peculiar continue to be salient markers of Positive Valence. Notice, moreover, that the paragon Conventionality marker, Conventional, also loads saliently on Positive Valence in the Spanish joint factor analysis, afinding that further bolsters our earlier interpretation that unconventionality as represented by such terms as Unusual, Odd, Peculiar, and (reversed) Conventional is a positively valued trait in Spanish society. 1 Discussion Using a cross-cultural and cross-target design, a primary goal of our study was to demonstrate that the Big Seven factor struc- 1 We also measured the congruence between the respective factor solutions with the Kaiser-Hunka-Bianchini (KHB; Kaiser, Hunka, & Bianchini, 71) index of factor similarity. This procedure projects factors from two solutions into a common space and calculates the cosines of the angles for corresponding test vectors. These cosines can be interpreted as correlations between corresponding test vectors. The overall KHB index represents the average cosine across all test vectors. The obtained similarity indices were as follows: (a) The American and Spanish Self Rating on the Big 7:.9; (b) The Spanish Self and Peer Ratings on the Big 7:.938; (c) The American and Spanish joint factor solutions for the Big 7 and Big 5:.890. These values provide additional evidence for the cross cultural robustness of the Big 7 factors.

10 7 VERONICA BENET AND NIELS G. WALLER Table 6 Joint Factor Analysis ofbig Seven and Big Five Questionnaires: American Sample, Self- Ratings Abbreviated items PVAL NVAL PEM NEM CNV Outstanding Impressive Excellent Exceptional Admirable Important High ranking Ingenious Ordinary Average Not exceptional Wicked Awful Deserves to be hated Dangerous Disgusting Cruel Mentally sick Treacherous Depraved Vicious Gregarious Talkative Sociable Talkative Sociable Animated Peppy Enthusiastic Assertive Playful Energetic Impulsive Not talkative Loner Shy Reserved Reserved Quiet Quiet Worrying Nervous Nervous Tense Irritated Feelings hurt Jittery Sorry for myself Feels guilty Depressed Moody Calm Stable Not overworrving Worries aside Not frustrated Not easily upset Relaxed Well organized Tidy Orderly Place for everything Efficient Thorough Prompt Planful Persevering (table continues)

11 SPANISH BIG SEVEN 7 Table 6 (continued) Abbreviated items PVAL NVAL PEM NEM CNV Consistent Improvises Distractible Careless Disorganized Disorganized Rude Quarrelsome Ouarrelsome Argumentative Stubborn Strong Tough Put up fieht Cold Fault-finding Headstrong Helpful Trv avoid differences Trusting Forgiving Lenient Dislike argument Cooperative Kind Original Artistic Inventive Unusual Sophisticated Imaginative Radical Reflective Curious Strange Odd Progressive Cautious Old-fashioned No permissive parents Strict discipline Reliable Simple Traditional Conventional Not artistic Note. PVAL = Positive Valence; NVAL = Negative Valence; PEM = Positive Emotionality; NEM = Negative Emotionality; C = Conscientiousness; A = Agreeableness; CNV = Conventionality. Big Seven descriptors are underlined. All loadings multiplied by 1. Loadings a are in boldface. ture captures the major sources of variance in self- and peerratings on American, lexically derived personality descriptors. A secondary goal was to demonstrate that the widely accepted Big Five factor structure (McCrae & John, 92; though see Waller, in press) cannot subsume the Positive and Negative Valence dimensions of the Big Seven (Tellegen, 93; Waller & Zavala, 93; Widiger, 93). To achieve these goals we analyzed five data sets from American and Spanish participants who completed self- and peer-ratings on Big Seven (Tellegen et al., 91) and Big Five (John et al., 91) questionnaires. Despite well-known difficulties that are encountered when endemic personality questionnaires are transported to foreign languages and cultures (Brislin, 80), our results revealed impressive similarities among the Big Seven factor structures in the American and Spanish samples and the self and observer data. However, our findings also revealed some intriguing differences between the American and Spanish Big Seven solutions. For instance, we found that Americans and Spaniards attach evaluatively different meanings to markers of the Conventionality continuum. In all Big Seven structures from the Spanish data, several Conventionality markers loaded prominently on the Positive Valence (e.g., Odd-Peculiar, Unusual-Unconventional, Conventional), or (reversed) Positive Emotionality (e.g., Strange) dimensions. In aggregate, however, our findings

12 7 VERONICA BENET AND NIELS G. WALLER Table 7 Joint Factor Analysis of Big Seven and Big Five Questionnaires: Spanish Sample, Self-Ratings Abbreviated items Outstanding Excellent Exceptional Admirable Unusual Important Impressive High ranking Odd Ingenious Old-fashioned Conventional Ordinary Average Not exceptional Wicked Awful Vicious Cruel Disgusting Treacherous Deserve hated Mentally sick Depraved Dangerous Talkative Talkative Sociable Gregarious Sociable Animated Enthusiastic Energetic Impulsive Peppy Playful Cooperative Strange Loner Shy Reserved Reserved Don't Talk much Quiet Quiet Nervous Depressed Jittery Nervous Moody Feelings hurt Worrying Guilty Irritated Feel sorry for myself Tense Curious Strong Assertive Calm Not overworrving Not frustrated Worries out of mind Relaxed Stable Well organized PVAL NVAL PEM NEM c A CNV (table continues)

13 SPANISH BIG SEVEN 7 Table 7 (continued) Abbreviated items PVAL NVAL PEM NEM C CNV Orderly Thorough Place for everything Tidy Planful Persevering Prompt Efficient Headstrong Cautious Consistent Reliable Strict discipline Distractible Like improvise Careless Disorganized Disorganized Argumentative Ouarrelsome Quarrelsome Tough Rude Put up fight Stubborn Fault-finding Forgiving Trusting Lenient Helpful Kind Not easily upset Dislike argument Try to avoid differences Inventive Imaginative Original Artistic Sophisticated Progressive Radical Reflective Simple Traditional No artistic No permissive parents Cold ' Note. PVAL = Positive Valence; NVAL = Negative Valence; PEM = Positive Emotionality; NEM = Negative Emotionality; C = Conscientiousness; A = Agreeableness; CNV = Conventionality. Big Seven descriptors are underlined and Big Five descriptors are in regular print. All loadings multiplied by 1. Loadings S; are in boldface. provide strong evidence for the cross-cultural generality of the Big Seven. Assuming that we accept the veracity of these findings, can we claim that the Big Seven provides a universal framework for personality description? Or, taking a less ambitious stance, can we claim that the Big Seven represents a compelling framework for organizing Spanish personality descriptors? For both of the questions the answer is no, for the following reasons. First, our study utilized an imposed-etic design (Berry, Poortinga, Segall, & Dasen, 92) to test the generality of the Big Seven factor structure in a Spanish sample. In this design, we administered a translated version of a Big Seven questionnaire a questionnaire that was originally developed to reflect the structural organization of the American English personality lexicon (Tellegen & Waller, 87) to a large sample of Spaniards. Thus, before claiming that the Big Seven truly represents the structure of Spanish personality descriptors, investigators must first apply an emic design that is sensitive to culturally specific aspects of personality organization in Spain. In this latter approach, trait terms would be sampled from the entire corpus of Spanish personality descriptors, rather than imported

14 7 VERONICA BENET AND NIELS G. WALLER from America (Benet, 94). We do not wish to understate the value of etic designs in cross-cultural research, however. The importance of our study rests not on its ability to prove the universal status of the Big Seven factor structure, which was not our intention, but rather on its ability to demonstrate that the seven dimensions of this structure and especially the Positive and Negative Valence dimensions represent something more than a model of how Americans organize natural language personality trait terms. As noted previously, our findings also provide evidence for culturally idiosyncratic differences for several Big Seven markers. For instance, several (-) Conventionality items in the American data were markers of Positive Valence and ( ) Positive Emotionality in the Spanish data. Certainly, these associative patterns are suggestive of the specific societal values that Spaniards attach to perceptions of unconventionality and introversion. To fully appreciate these associations, first consider recent cross-cultural research on value orientations. That cultures differ on value orientations is a prominent theme in contemporary cross-cultural research. Among the diverse value orientations that have been described in recent years (Hofstede, 80; Schwartz & Bilsky, 90; see Berry et al., 92, for a review), one in particular the individualism-collectivism distinction (Triandis, 90) seems most relevant for interpreting our results. Individualism and collectivism describe an individual's orientation toward person-centered (Agentic) or public-centered (Communal) aspects of social existence. In individualistic societies, evaluations of the self and others are based primarily on personal achievement, excellence, the uniqueness of personal attributes and independence (Oyserman, 93; Spence, 85). In collectivist societies, on the other hand, evaluations of self and others are based on whether an individual 'belongs to' or 'fits in' to a social network that stresses interdependence (Markus & Kitayama, 90; Oyserman, 93). The United States is often described as an individualistic society par excellence (Hofstede, 80; Spence, 85). As a group, Americans place a high value on personal success and uniqueness, and not surprisingly, the American English lexicon contains numerous terms denoting uniqueness and self-worth. Many of these terms are markers of the Positive Valence dimension of the Big Seven, which indicates that in self or other ascriptions of personality these terms cohere in an organized manner. Certainly, to describe oneself or others as Outstanding, Admirable or High ranking, requires the kind of cognitive selfother differentiations that are fostered by an individualistic value orientation. Spain, on the other hand, is usually described-along with other Southern European and Latin American cultures-as a predominantly collectivist society (Markus & Kitayama, 91; Triandis, 90), and our results support this characterization. For example, we found that Spaniards place a high value on the social facets of Positive Emotionality. Spaniards who are low on Positive Emotionality describe themselves as being a loner who is reserved and quiet. By itself, this finding is not unusual because these markers also define low Positive Emotionality in the American sample. However, in contrast to Americans, Spaniards who are low on this dimension also describe themselves as being strange. The conjunction of introversion and strangeness is also found in Spanish peer ratings. Overall, these results support the contention that collectivist societies-such as Spain value affiliation (Markus & Kitayama, 91). Nevertheless, our results also reveal that Spaniards value being different from the collective. For instance, in the Spanish self-ratings, terms such as Unconventional, Peculiar and Odd load on the same dimension as Outstanding, Admirable, and High Ranking! This finding suggests that Spaniards also endorse a 'radical' form of individualism. Certainly in their selfother evaluations, Spaniards go well beyond the achievementbased perceptions of excellence and uniqueness that characterize the American Positive Valence dimension. In summary, Spaniards hold two seemingly contradictory values. On the one hand they value social interaction that is, being with people whereas, on the other hand, they value social distinction that is, being different from people. As Vega McVeagh (90) an American correspondent in Spain observed in her recent portrayal of Spaniards: " gregariousness is the norm. This does not conflict with individualism in the Iberian sense when one realizes that the larger the gathering, the larger the potential audience and the opportunity for showing offand announcing 'this is me".... An internal selfsufficiency expressed through wit, grandiloquent phrases, appearance, courtesy, generosity, and pride. The result is that often the most arrogant person is the most charming" (italics added, McVeagh, 90, p. 68). In conclusion, our study demonstrates that the Big Seven factor structure that characterizes self-ratings on American English, natural-language personality descriptors can also be recovered in Spanish self- and peer-ratings of personality. Factor structures for a Big Seven and Big Five questionnaire were very similar in the two cultures. Moreover, joint factor analyses of the two questionnaires yielded congruent seven-factor solutions that included the recently described Positive and Negative Valence dimensions (Tellegen, 93; Tellegen& Waller, 87; Waller, in press; Waller & Zavala, 83). These dimensions played an important role in our results. Because our study included markers of the self-other evaluation factors, we were able to elucidate psychologically meaningful differences in the ways in which Americans and Spaniards evaluate Conventionality and Positive Emotionality trait indicators. These differences would not have been uncovered had we limited our study to the widely accepted Big Five factor structure, a result that supports our view that previously banished evaluative terms should be reintroduced into lexically derived personality taxonomies. This idea is also consistent with several converging themes in contemporary personality theory. Borkenau, for example, believes that "the main purpose of trait terms is not so much to describe but rather to evaluate people...." (italics added, Borkenau, 90, p. 394). Hogan contended that "the primary function of trait ascription is to evaluate other people..." (italics added; 82, p. 60). Buss (91) expresses similar views. Perhaps these opinions reflect a growing consensus on the importance of evaluation in personality description and measurement. If they do, then isn't it time to evaluate the Big Five? References Allport, G. W., & Odbert, H. S. (36). Trait names: A psycho-lexical study. Psychological Monographs, 47, 2.

15 SPANISH BIG SEVEN 7 Almagor, M., Tellegen, A., & Waller, N. G. (93). Further explorations Galton, F. (84). Measurement of character. Fortnightly Review, 42, of the basic dimensions of natural language trait descriptors. Unpublished manuscript, University of Haifa. Goldberg, L. R. (81). Language and individual differences: The 1. Angleitner, A., Ostendorf, E, & John, O. P. (90). Towards a taxonomy of personality descriptors in German: A psycho-lexical study. view ofpersonality and Social Psychology, 2,. search for universals in personality lexicons. In Wheeler (Ed.), Re- European Journal ofpersonality, 4, 898. Goldberg, L. R. (82). From Ace to Zombie: Some explorations in the Benet, V. (94). Towards a Spanish taxonomy of personality descriptors: Generality of the Big Seven factor model with indigenous and Personality Assessment, 1, 24. language of personality. In Spielberger & Butcher (Eds.), Advances in imported constructs. Unpublished dissertation, University of California at Davis. The Big-Five factor structure. Journal ofpersonality and Social Psy- Goldberg, L. R. (90). An alternative "Description of Personality": Berry, J. W. (80). Introduction to methodology. In H. Triandis & chology, 59, 29. J. W. Berry (Eds.), Handbook of cross-cultural psychology (Vol. 2, Goldberg, L. R. (92). The development of markers for the Big-Five pp. 1-). Boston: Allyn & Bacon. factor structure. Psychological Assessment: A Journal of Consulting Berry, J. W., Poortinga, Y. H., Segall, M. H., & Dasen, P. R. (92). and Clinical Psychology, 4, Cross-cultural psychology: Research and applications. New York: Goldberg, L. R. (93). The structure of phenotypic personality traits. Cambridge University Press. American Psychologist,, Block, J. (95). A contrarian view of thefive-factorapproach to personality description. Psychological Bulletin, 7, 7-2. Gough, H. G., & Heilbrun, A. B., Jr. (83). The Adjective Check List Gorsuch, R. L. (83). Factor analysis. Hillsdale, NJ: Erlbaum. Bond, M., Nakazato, H., & Shiraishi, D. (75). Universality and distinctiveness in dimensions of Japanese person perception. Journal of Hamilton, D. L. (68). Personality attributes associated with extreme manual (Rev.). Palo Alto, CA: Consulting Psychologists Press. Cross-Cultural Psychology, 6, response style. Psychological Bulletin, 69, 2-2. Borkenau, P. (90). Traits as ideal-based and goal-derived social categories. Journal ofpersonality and Social Psychology, 58, in work-related values. Beverly Hills, CA: Sage. Hofstede, G. (80). Culture's consequences: International differences Briggs, S. R. (92). Assessing the Five-Factor model of personality Hofstee, W. K. (90). The use of everyday personality language for description. Journal ofpersonality, 60, scientific purposes. European Journal of Personality, 4, Brislin, R. W. (80). Translation and content analysis of oral and written materials. In H. Triandis & J. W. Berry (Eds.), Handbook of (Ed.), Nebraska Symposium on Motivation, 82: Personality-Cur- Hogan, R. (82). A socio-analytic theory of personality. In M. M. Page cross-cultural psychology (Vol. 2, pp ). Boston: Allyn & rent theory and research. Lincoln: University of Nebraska Press. Bacon. Hooper, J. (87). The Spaniards: A portrait of the new Spain. London: Brokken, F. B. (78). The language of personality. Meppel, The Netherlands: Krips. John, O. P. (89). Towards a taxonomy of personality descriptors. In Penguin Books. Buss, D. M. (91). Evolutionary personality psychology. Annual Review of Psychology, 42, and emerging directions (pp ). New York: Springer-Verlag. D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends Buss, D. M., & Craik, K. H. (83). The act frequency approach to John, O. P. (90). The "Big Five" factor taxonomy: Dimensions of personality. Psychological Review, 90,. personality in the natural language and in questionnaires. In L. A. Campbell, J. D. (90). Self-esteem and the clarity of the self-concept. Per vin (Ed.), Handbook of personality: Theory and research (pp. 66- Journal ofpersonality and Social Psychology, 59, ). New York: Guilford Press. Caprara, G. V., Barbaranelli, C, Borgogni, L., & Perugini, M. (93). John, Q P., Angleitner, A., & Ostendorf, F. (88). The lexical approach to personality: A historical review of trait taxonomic research. The "Big Five Questionnaire": A new questionnaire to assess the Five Factor Model. Personality & Individual Differences,, 1-8. European Journal of Personality, 2, 1-2. Cattell, R. B. (43). The description of personality: Basic traits resolved into clusters. Journal of Abnormal and Social Psychology, 38, Inventory-Versions 4a and54. Technical Report. IPAR. University of John, O. P., Donahue, E. M., & Kentle, R. L. (91). The "BigFive" California at Berkeley. Cattell, R. B. (). The principal trait clusters for describing personality. Psychological Bulletin, 42, 91. alphabet: Taxonomies of personality-descriptive terms in English, John, O. P., Goldberg, L. R., & Angleitner, A. (84). Better than the Church, A. T., & Katigbak, M. S. (88). The emic strategy in the Dutch, and German. In H. Bonarius, G. van Heck, & N. Smid (Eds.), identification and assessment of personality dimensions in a nonwestern culture. Journal of Cross-Cultural Psychology,, 03. ments(pp. 830). Lisse, The Netherlands: Swets and Zeitlinger. Personality psychology in Europe: Theoretical and empirical develop- Costa, P. T, & McCrae, R. R. (85). The NEO Personality Inventory Larsen, R. J., & Ketelaar, T. (91). Personality and susceptibility to manual. Odessa, FL: Psychological Assessment Resources. positive and negative emotional states. Journal of Personality and Social Psychology, 61,. De Raad, B. (92). The replicability of the Big Five personality dimensions in three word-classes of the Dutch language. European Loevinger, J. (94). Has psychology lost its conscience? Journal of Journal ofpersonality, 6, -29. Personality Assessment, 62, 2-8. De Raad, B., & Hoskens, M. (90). Personality descriptive nouns. Markus, H., & Cross, S. (90). The interpersonal self. In L. A. Pervin European Journal of Personality, 4, 895. (Ed.), Handbook ofpersonality: Theoryand research (pp ). De Raad, B., Mulder, E., Kloosterman, K., & Hofstee, W. K. B. (88). New York: Guilford Press. Personality descriptive verbs. European Journal of Personality, 2, 81 Markus, H., & Kitayama, S. (91). Culture and the self: Implications 96. for cognition, emotion, and motivation. Psychological Review, 98, Digman, J. M. (90). Personality structure: Emergence of the fivefactor model. Annual Review ofpsychology, 41,4-0. McAdams, D. P. (92). The Five-Factor Model in personality: A crit Dunlap, W. P., & Cornell, J. M. (94). Factor analysis of ipsative measures. Multivariate Behavioral Research, 29, 56. McCrae, R. R., & Costa, P. T. (86). Clinical assessment can benefit ical appraisal. Journal ofpersonality, 60, 5-2. Fiske, D. W. (49). Consistency of the factorial structures of personality ratings from different sources. Journal of Abnormal and Social gist, 41,. from recent advances in personality psychology. American Psycholo- Psychology, 4, McCrae, R. R., & Costa, P. T. (87). Validation of the five-factor

16 7 VERONICA BENET AND NIELS G. WALLER model of personality across instruments and observers. Journal of Personality and Social Psychology, 52, McCrae, R. R., & John, O. P. (92). Introduction to the five-factor model and its applications. Journal of Personality, 60, 5-2. McVeagh, V. (90). People. In K. Wheaton (Ed.), Inside guides: Spain. Singapore: APA Publications. Meyer, G. J., & Shack, J. R. (89). Structural convergence of mood and personality: Evidence for old and new directions. Journal ofpersonality and Social Psychology, 5 7, Norman, W. T. (63). Toward an adequate taxonomy of personality attributes. Journal of Abnormal and Social Psychology, 66, Norman, W. T. (67). personality trait descriptors: Normative operating characteristics for a university population. Ann Arbor: University of Michigan, Department of Psychology. Oyserman, D. (93). The lens of personhood: Viewing the self and others in a multicultural society. Journal of Personality and Social Psychology,, 993. Peabody, D. (67). Trait inferences: Evaluative and descriptive aspects. Journal of Personality and Social Psychology Monographs, 7 (4, Whole No. 6). Peabody, D. (87). Selecting representative trait adjectives. Journal of Personality and Social Psychology, 57, Peabody, D., & Goldberg, L. R. (89). Some determinants of factor structures from personality-trait descriptors. Journal of Personality and Social Psychology, 57, Pervin, L. A. (90). A brief history of modern personality theory. In L. A. Pervin (Ed.), Handbook of personality: Theory and research. New York: Guilford Press. Saucier, G. (94). Separating description and evaluation in the structure of personality attributes. Journal of Personality and Social Psychology, 66,. Schwartz, S. H., & Bilsky (90). Toward a theory of the universal content and structure of values: Extension and cross-cultural replications. Journal ofpersonality and Social Psychology, 58, Spence, J. T. (85). Achievement American style: The rewards and costs of individuality. American Psychologist, 40, Tellegen, A. (85). Structures of mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In A. H. Tuma& J. Maser (Eds.), Anxiety and the anxiety disorders (pp ). Hillsdale, NJ: Erlbaum. Tellegen, A. (93). Folk concepts and psychological concepts of personality and personality disorder. Psychological Inquiry, 4,. Tellegen, A., Grove, W. M., & Waller, N. G. (91). Inventory of Personal Characteristics #7 (IPC- 7). Unpublished materials, University of Minnesota. Tellegen, A., & Waller, N. G. (87). Reexamining basic dimensions of natural language trait descriptors. 95 annual meeting of the American Psychological Association, (abstract). Tellegen, A., & Waller, N. G. (in press). Exploring personality through test construction: Development of the Multidimensional Personality Questionnaire. In S. R. Briggs & J. M. Cheek (Eds.), Personality measures: Development and evaluation (Vol. 1). Greenwich, CT: JAI Press. The American Heritage Dictionary. (85). Boston: Houghton Mifflin. Triandis, H. C. (90). Cross-cultural studies of individualism and collectivism. Nebraska Symposium on Motivation (pp ). Lincoln: University of Nebraska Press. Tupes, E. C, & Christal, R. C. (92). Recurrent personality factors based on trait ratings. Journal of Personality, 60, Waller, N. G. (in press). Evaluating the structure of personality. In C. R. Cloninger(Ed-), Personality and psychopathology. Washington, DC: American Psychiatric Press. Waller, N. G., & Ben-Porath, Y. S. (87). Is it time for clinical psychology to embrace the FFM of personality? American Psychologist, 42, Waller, N. G., & Zavala, J. (93). Evaluating the Big Five. Psychological Inquiry, 4,. Waller, N. G., & Meehl, P. E. (submitted). Multivariate taxometric methods. Unpublished manuscript, University of California, Danis. Watson, D., & Clark, L. A. (92). On traits and temperament: General and specific factors of emotional experience and their relation to the Five-Factor model. Journal ofpersonality, 60, Widiger, T. A. (93). The DSM-IH-R categorical personality disorder diagnoses: A critique and an alternative. Psychological Inquiry, 4,1 {- 93. Wiggins, J. S. (79). A psychological taxonomy of trait-descriptive terms: The interpersonal domain. Journal of Personality and Social Psychology, 37, Yang, K., & Bond, M. H. (90). Exploring implicit personality theories with indigenous or imported constructs: The Chinese case. Journal ofpersonality and Social Psychology, 58, 95. Received October,94 Revision received December 30,94 Accepted February 2,95

17 SPANISH BIG SEVEN 7 American IPC-7 1. not easily upset 2. don't talk much, uncommunicative 3. conventional 4. exceptional, special 5. dangerous to others, harmful 6. others think I am quarrelsome and contentious 7. quiet 8. reserved, distant 9. often feel guilty for no reason. important, significant. easy on others, lenient. rather put a fight that make a concession. average, unremarkable. disgusting, horrible. consistent, predictable. strong, forceful. lively, animated. often irritated my minor setbacks. tough, uncompromising. talkative. like things to be a bit disorganized and chaotic 22. prefer to be alone, a loner 23. playful 24. can put worries out of mind 25. gregarious, sociable 26. vicious, nasty 27. do things in an orderly and systematic manner. often jumpy and jittery 29. hold traditional values and beliefs 30. awful, terrible 31. stubborn, obstinate 32. an ordinary, everyday person 33. wicked, evil 34. get into arguments, argumentative 35. deserve to be admired 36. peppy, spirited 37. high-ranking, powerful 38. do not worry about little things 39. feelings are easily hurt 40. keep belongings neat and tidy 41. politically radical, hold revolutionary views 42. prompt, puntual, get things done on time 43. don't let many things bother or frustrate me. cautious, circumspect. believe that strict discipline at home would prevent most of the crime in society today 46. odd, peculiar 47. impressive, remarkable. headstrong, willful 49. unusual, unconventional 50. excellent, first-rate 51. dislike arguments and conflict 52. well-organized 53. deserve to be hated 54. like to improvise, 'play things by ear' 55. nervous, high-strung 56. cruel, mean 57. believe that most parents are too permissive, let their Appendix The Items of the IPC-7 and BFI: American and Spanish Translations Spanish IPC-7 1. dificil de alterar 2. poco habladora 3. convencional, tradicional 4. exceptional, especial 5. peligrosa para los demas, perjudicial 6. la gente piensa que me gusta discutir 7. callada 8. reservada, distante 9. a menudo se siente culpable sin motivo. importante, significante. de trato facil, indulgente. se pelea antes que ceder. promedio, nada extraordinario. repugnante, asquerosa. consecuente, predecible. fuerte, energica. animada, alegre. a menudo se molesta por pequenos contratiempos. dura, intransigente. habladora. le gustan las cosas un poco desorganizadas y caoticas 22. prefiere estar sola, solitaria 23. juguetona, traviesa 24. sabe poner poner las preocupaciones a un lado 25. gusta de la compania, sociable 26. malvada, ruin 27. hace las cosas de forma ordenada y sistematica. a menudo esti intranquila, inquieta 29. tiene valores y creencias tradicionales 30. horrible, terrible 31. testaruda, obstinada 32. una persona normal y corriente, del monton 33. malvada, cruel 34. entra en disputas, discutidora 35. digna de admiration 36. llena de vida y energia 37. de alto rango, poderosa 38. no se preocupa por pequeneces 39. es facil herir sus sentimientos 40. limpia y ordenada con sus cosas 41. de ideas politicas radicales, revolucionarias 42. puntual, acaba las cosas a tiempo 43. no se deja frustrar por muchas cosas. cauta, precavida. cree que con disciplina mas estricta en el hogar se prevendria el crimen 46. rara, peculiar 47. llama la atencion, notable. testaruda, terca 49. inusual, poco convencional 50. excelente, de primera 51. no le gustan las discusiones ni las situaciones conflictivas 52. bien organizada 53. digna de ser odiada 54. le gusta improvisar, hacer las cosas sobre la marcha 55. nerviosa, tensa 56. cruel, mal intencionada 57. cree que muchos padres son demasiado blandos, que (Appendix continues on next page)

18 7 VERONICA BENET AND NIELS G. WALLER Appendix children gee away with too much 58. outstanding, superior 59. often feel sorry for myself 60. strange 61. spontaneous, impulsive 62. depraved, perverted 63. try to avoid difficulties with other people 64. mentally disturbed, sick. not exceptional, not that special 66. like to be with people, sociable 67. treacherous, disloyal 68. conservative 69. like to have a place for everything and everything in its place 70. progressive, favor social reform consienten a sus hijos 58. sobresaliente, superior 59. siente compasion por si misma con frecuencia 60. estrana 61. espontanea, impulsiva 62. depravada, pervertida 63. intenta evitar tener problemas con la gente 64. transtornada. mentalmente enferma. nada excepcional o especial 66. le gusta estar con la gente, sociable 67. traidora, desleal 68. conservadora 69. le gusta tener un sitio para cada cosa y cada cosa en su sitio 70. de ideas progresistas, a favor del cambio social Received October, 94 Revision received December 30, 94 Accepted February 2, 95