RECOVERY OF MAJOR PERSONALITY DIMENSIONS FROM TRAIT ADJECTIVE DATA

A study was run to test the emergence of major personality dimensions from factor analysis of trait adjective data. 144 trait-descriptive ad jectives were selected as markers for Brand's 'Big Six' structural mod el of personality. Application of Horn's parallel analysis to self-rat ings of the adjectives by a sample of 1210 subjects suggested that 10- 12 factors should be extracted from the correlation matrix. 10 factors were extracted, of which 8 were replicable across two sub-samples of 605 subjects each. The replicable factors were clearly related to stru ctural models such as the Big Six and Big Five, but correspondences we re not exact. Extraction of smaller numbers of factors appeared to dis tort the factor solution. Patterns of occupational differences in pers onality also supported the utility of the 10-factor solution. It is co ncluded that it may be preferable to work with well-defined primary tr ait adjective factors than with secondary factors distorted by extract ion of an incorrect number of factors. Methodologically, the data supp ort concerns that factor replicability should not be used as a criteri on for number of factors.