جستجو برای "پنج عامل بزرگ" در زمینه های یونانی: نئو-FFI در زیر میکروسکوپ

The five-factor inventory (FFI) is a commonly used personality test based on the NEO-personality inventory revised. It has been translated into several languages and validated in a number of countries. Using a sample of 1204 individuals, the present study evaluates the psychometric properties and factor structure of the Greek FFI and provides normative information for its use with Greek populations. Convergent validity is also assessed by studying the relationship between the five scales of the FFI and the Brief Symptom Inventory. Results show that although the factor scales had acceptable internal consistency, they were highly intercorrelated. Exploratory factor analyses failed to reproduce the appropriate factor structure, yielding instead numerous fragments of the five dimensions. Confirmatory factor analyses also failed to lend support to the five factor model of personality as measured by this instrument. Based on these results, the use of the instrument is recommended only with serious caution in this cultural context. Further research ought to decipher whether the failures to reproduce the five factor model using the FFI in this and other cultures represents a challenge to the universality of the theory, or merely a shortcoming of the specific instrument.

One of the most popular conceptions of personality today, which rests on the assumption of a specific number of underlying personality dimensions, is the five factor model. A significant body of research supports the presence of five orthogonal dimensions of personality, which apparently also exist in the trait adjectives people use in their daily lives to describe others or themselves (Goldberg, 1993). The “Big Five” dimensions that have emerged from factor analyses of peer and self-ratings of personality descriptors are neuroticism, extraversion, openness to experience, conscientiousness and agreeableness. They have been replicated in several countries including Korea, France, Estonia and Finland (e.g. Pulver, Allik, Pulkkinen, & Hamalainen, 1995; Rolland, Parker, & Stumpf, 1998; Spirrison & Choi, 1998) using various personality inventories (Piedmont, McCrae, & Costa, 1991). The most popular assessment instrument for these five dimensions is Costa and McCrae's NEO-personality inventory revised (NEO-PI-R), which consists of 240 items that result in the five factors and a number of factor facets. Because of its length, a briefer version of the instrument, the NEO-five factor inventory (NEO-FFI), was created (Costa & McCrae, 1992a) with 60 items. To derive each of the factor scales of the NEO-FFI, 12 items from the 1986 administration of the NEO-PI, were selected for their high positive loading on the corresponding trait. The brief instrument has adequate internal reliability and correspondence with the full scale (McCrae & Costa, 1989) and thus has been used widely for research and clinical purposes. It shows good convergent validity with other personality instruments, correlates with ratings provided by others of the target's personality (Kurtz & Sherker, 2003; Parker & Stumpf, 1998), and appears to be a valid measure of genetically/biologically based personality dimensions (Riemann, Angleitner, & Strelau, 1997). In spite of its widespread use and the increasing acceptance of its external validity, research carried out so far in a number of countries using exploratory and confirmatory factor analyses (CFA), has called into question the factor structure of the NEO-FFI with some studies finding fewer than five factors (Ackerman & Heggestad, 1997; Ferguson & Patterson, 1998), others finding as many as eight (Yoshimura, Ono, Nakamura, Nathan, & Suzuki, 2001) and yet others finding weak fit indices for a five factor solution (Mooradian & Nezlak, 1996). This is not surprising given that even the long version of the instrument has not fared well when subjected to CFA (e.g. Parker, Bagby, & Summerfeldt, 1993). Concern regarding the use of the FFI has surfaced both in English and non-English speaking countries. In a British study, Egan, Deary, and Austin (2000) found that although British norms corresponded favourably with American, the factor structure of the NEO-FFI was less satisfactory. Factors N, A and C appeared reliably, while many of the items of factors O and E did not load adequately on the expected factors. In a Canadian sample of female college students, Holden and Fekken (1994) found acceptable support for the five factors with exploratory factor analysis, but failed to obtain a five factor solution with CFA. Even with an American sample, Tokar, Fischer, Snell, and Harik-Williams (1999) failed to obtain satisfactory orthogonal or oblique factor solutions with CFA. Hence, the utility of the NEO-FFI in measuring the “Big Five” dimensions needs to be evaluated cautiously even with regards to English speaking cultures where no translation of the instrument was required. The need to empirically validate this instrument is even more pressing in non-English speaking nations where the questionnaire requires translation. In a study carried out in Germany, Schmitz, Hartkamp, Baldini, Rollnik, and Tress (2001) examined the psychometric properties of the German NEO-FFI among a group of psychosomatic patients and a reference group of personality disorder patients. Their findings are of concern to FFI users: First, the five dimensions were not orthogonal (the same concern was raised by Church & Burke (1994)). Secondly, CFA failed to yield acceptable fit indices for five, four or two factor models using both oblique and orthogonal rotations. Similarly, Yoshimura et al. (2001) obtained eight factors with a Japanese sample, after subjecting the NEO-FFI to exploratory factor analysis, with 11 items not loading on any of the factors. The five-factor model of personality has yet to receive adequate empirical attention in the Greek cultural context, partly due to the absence of well-documented measures of the construct. Hence, the present study examines the internal reliability and factor structure of the Greek translation of the NEO-FFI. This translation has been initially authored by Bezevengis for use by parents to rate their children's personality (see Stalikas, Triliva, & Roussi, 2002) at the University of Athens, and was further developed by Demetriou and Kazi (2001) at the University of Cyprus for use with adults, but to date no research to validate the instrument or establish Greek norms has been published. The method of front and back translation was used for its development and the measure is used in research in both Greece and Cyprus (Demetriou, Kyriakides, & Avraamidou, in press). The present authors proceeded to an independent back translation of the instrument, which resulted in a very close approximation to the original that was deemed satisfactory. The present study additionally generates some initial norms based on a large sample of participants. Finally, the study provides evidence regarding the concurrent validity of the instrument in relation to the Brief Symptom Inventory (BSI; Derogatis, 1993). Both the NEO-FFI and the BSI are frequently used in mental health settings. Being able to demonstrate that the NEO-FFI is able to predict the presence of psychological symptoms (i.e. deviations from `normal' personality functioning) in our sample, would add credence to its validity. For instance, one should expect that extraversion would be associated with low scores on anxiety, depression, somatization, etc., whereas neuroticism would be associated with high scores on these BSI dimensions.