پنج عامل بزرگ و یا دو عامل بزرگ؟ عوامل مافوق در پرسشنامه پنج عاملی نئو و پرسشنامه شخصیت ضد اجتماعی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|34186||2004||14 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Personality and Individual Differences, Volume 37, Issue 5, October 2004, Pages 957–970
Although it is claimed that the Big Five dimensions of Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness represent the highest level in the hierarchical structure of personality, there is consistent evidence that they are not independent and that two higher order factors underlie them. Two higher order factors also underlie the scales of the Antisocial Personality Questionnaire (APQ). Structural equation modelling was used in a sample of male forensic psychiatric patients (N=164) to test the hypotheses that the scales of the NEO Five Factor Inventory (NEO-FFI) support two higher order factors as found in other Big Five measures, and that these are equivalent to the dimensions of the APQ. Good support was found for the model and a confirmatory factor analysis indicated that the Impulsivity and Withdrawal factor scales of the APQ provide reasonable markers of the NEO-FFI latent factors. The two factors can be interpreted in terms of the metaconcepts of agency and communion, and it is suggested that the Impulsivity and Withdrawal dimensions reflect basic motivational concerns about power, status, and intimacy.
In the factor analytic tradition, personality variables are construed as hierarchically structured. Eysenck (1947) noted that traits represent the intercorrelations of specific, repeated behaviours and that intercorrelations between traits define ‘a type concept’, such as extraversion, generally assessed as a continuous dimension. More recent proponents, such as Costa and McCrae (1992), describe these different levels as ‘facets’ (specific traits) and ‘domains’ (trait dimensions). As Watson, Clark, and Harkness (1994) observe: ‘…trait dimensional hierarchies are variance–covariance hierarchies. The covariance of the lower order elements becomes the variance of the higher order elements’ (p. 19). The number of traits or dimensions identified will vary according to where the hierarchy is ‘sliced’ (e.g. Harkness, 1992), but it is currently widely accepted that the most comprehensive representation of the higher order elements is provided by the ‘Big Five’ dimensions of Neuroticism (N), Extraversion (E), Openness (O), Agreeableness (A), and Conscientiousness (C). This view reflects convergent findings of lexical analyses and questionnaire research (Costa & McCrae, 1992; Goldberg, 1993), and Watson et al. (1994) found that it accommodates the competing views of those, such as Eysenck, who argue for a Big Three representation. Goldberg (1993) proposed that the Big Five are located ‘at the highest level that is still descriptive of behavior, with only general evaluation located at a higher and more abstract level’ (p. 27). McCrae and Costa (1996) similarly assert that their five-factor model (FFM) constitutes ‘the highest level of the hierarchy’ (p. 74). Despite this apparent consensus that the Big Five represent the more ‘basic’ trait dimensions, there is consistent evidence that they are not independent. Block (1995) detected appreciable intercorrelations between ratings of the Big Five and between self-report versions of these factors as measured by both the Revised NEO Personality Inventory (NEO PI-R: Costa & McCrae, 1992) and the shorter NEO Five Factor Inventory (NEO-FFI), and considered the factor structure to be ‘impressively nonorthogonal’. Other studies of the NEO-FFI confirm this. Egan, Deary, and Austin (2000), for example, found highly significant negative correlations of N with E and C, and a positive correlation between the latter. They attribute this to weaknesses in the NEO-FFI, but Clark and Harrison (2001) also report a high negative correlation between N and E in a structured interview measure of the FFM. Investigators attempting to validate the FFM through confirmatory factor analysis (CFA) have also demonstrated that oblique solutions provide a somewhat better fit to the data than orthogonal factors (e.g. Church & Burke, 1994; Tokar, Fischer, Snell, & Harik-Williams, 1999). McCrae, Zonderman, Costa, Bond, and Paunonen (1996) questioned the utility of CFA in investigating personality structure and argued that intercorrelations between domain measures arise from the selection of facets to represent each domain. They demonstrated that when secondary loadings of facets across domains are specified, orthogonal factors fit the data as well as oblique factors. This does not, however, fully account for intercorrelations of the NEO-FFI scales, whose items were selected as relatively pure measures of the five domains (Costa & McCrae, 1992). Critics argue that correlations between domain measures cast doubt on the integrity of the Big Five as distinct dimensions. From the hierarchical perspective, however, these correlations are understandable if the Big Five do not, as claimed, represent the highest point of the hierarchy. Digman (1997) proposed that correlations between the Big Five give rise to two broader dimensions at a higher level of abstraction. He analysed correlations between Big Five measures from 14 samples, which included children, adolescents, and adults, and the measures included teacher or peer ratings as well as self-reports, reflecting several approaches to measuring the factors. Exploratory and confirmatory factor analyses consistently supported the hypothesis of two higher order factors, labelled Alpha and Beta. The Alpha factor was defined by high loadings of Agreeableness, Emotional Stability (i.e. low Neuroticism), and Conscientiousness. The Beta factor was defined by high loadings of Extraversion and Intellect (Openness) and smaller loadings of Emotional Stability and Conscientiousness. While avoiding substantive labels for these factors, Digman proposed that they represent broad constructs of personality found in the writings of theorists as diverse as Adler, Skinner, Rogers and Bandura, and that they permitted links between the atheoretical Big Five and these theoretical traditions. Alpha, for example, seems to reflect the outcome of the socialisation process seen in the development of impulse restraint and conscience and the reduction of hostility, aggression, and neurotic defense. Beta seems related to personal growth or self-actualisation versus personal constriction. He also noted that they could be interpreted in terms of higher level concepts proposed by several theorists and as summarised by Wiggins (1991) by the metaconcepts of agency and communion. Two-dimensional representations of personality organisation have been proposed by other theorists (e.g. Gray, 1981), and have also emerged from studies of the MMPI and similar instruments. The higher order factors of the MMPI have been interpreted as Neuroticism and Extraversion, but Kassebaum, Couch, and Slater (1959) demonstrated that a 45° rotation of these yielded ‘fusion factors’ of impulsivity and sociability (i.e. neurotic extraversion and stable extraversion). Blackburn, 1987 and Blackburn, 1998 included modified measures of these ‘fusion factors’ in a set of scales assessing deviant personality attributes, such as hostility, aggression, and social avoidance. These attributes are represented in eight primary factor scales of the Antisocial Personality Questionnaire (APQ: Blackburn & Fawcett, 1999). Factor analyses of these scales in both forensic psychiatric and normal samples consistently yield two higher order dimensions labelled Impulsivity (awareness and expression of hostile, rebellious, and aggressive impulses versus control and conformity) and Withdrawal (avoidance, low self-esteem, and submissiveness versus sociability and self-confidence), and scales of Impulsivity (I) and Withdrawal (W) were constructed to measure these dimensions. These higher order factors appear to reproduce the impulsivity and sociability dimensions identified by Kassebaum et al. (1959). However, because of the wide ranging content of the I and W scales and their correlations with other personality measures, Blackburn and Fawcett (1999) suggested that they may reflect superordinate factors encompassing three or more of the Big Five dimensions. These higher order factors in fact appear conceptually very similar to the Alpha and Beta factors identified by Digman (1997) and the present study aimed to test the hypothesis that the dimensions of impulsivity and sociability in the APQ correspond to higher order dimensions of the Big Five. We used structural equation modelling procedures to test three hypotheses: first, that intercorrelations of the NEO-FFI yield the two higher factors identified by Digman (whose analysis did not include the NEO-FFI), second, that these factors correspond to the impulsivity and withdrawal (reversed) dimensions of the APQ, and third, that the APQ I and W scales provide markers for Digman’s Alpha and Beta factors.