文摘
The Big Five model, originally based on studies of lexical personality descriptors, has proved useful and attracted much consensus. However, although the Big Five factors are held to be orthogonal, they are not orthogonal in previous marker sets developed for them. The author describes methods for creating relatively orthogonal marker sets for orthogonal factors and demonstrates that the Big Five can be operationalized in relatively orthogonal marker scales. The three marker sets presented are a 40-item short form of Goldberg's 100 unipolar markers (Ortho-Markers), a new set of Modular Markers built with item parcels, and a 40-item short form of these (Mini-Modular Markers [the 3M40]). The new Big Five marker sets are shown to have markedly lower interscale correlations, with no loss of validity, relative to previous marker sets with comparable numbers of items. This indicates that the nonorthogonality of previous Big Five marker scales is less a property of the Big Five model than an unintended outcome of commonly used scale construction procedures.