Distance from the protype : a multidimensional scaling approach to personality assessment

An MDS (MultiDimenaional Scaling) model of personality assessment is presented as an alternative method of personality assessment designed to incorporate recently discovered cognitive principles relating to how people mentally organize categories (Roach, 1978), including personality trait categories (Broughton, 1984). The MDS model is shown to generalize from a tool for evaluating the self concept and the semantics of interpersonal categories (Partridge, 1984) to one that taps the full gamut of personality assessment in the interpersonal domain as defined by Wiggins (1979). In this paradigm, subjects rated the similarity of their own personality to prototypical characters described in short stories, or vignettes. In asking subjects to compare themselves to hypothetical people who display prototypical behaviors (based on the act-frequency prototype analyses of Buss & Craik, 1980), one is able to standardize the measure against which similarity ratings are made. Thus "prototypical dominance" (in this case operationalized as an excellent example of what it means in behavioral terms to be dominant) is the same for each subject, as is prototypical extraversion, aggression, and so on. Unlike traditional self-report measures, the respondent is not required to provide his or her own (possibly idiosyncratic) trait definition. Study 1 involved the development and use of the vignette materials in a paper and pencil administration. In this study 25 undergraduate subjects rated the 28 nonredundant pairs of eight vignette characters for similarity. The usefulness of the vignettes as personality testing stimuli was gauged according to structural criteria, namely the circular ordering (circumplexity) to emerge from the MDS analysis of the eight vignette stimuli. Although the results were judged satisfactory, steps were taken in Study 2 to improve the interpersonal meaning of two of the prototype stories, in an attempt to improve the solution. In Study 3, a microcomputer administration involving 158 participants, subjects compared the eight stimuli for similarity as in Study 1 but also compared themselves (each subject's usual and ideal self) to the eight vignette characters. Derived MDS distances from the prototypical characters were compared with conventional self-report scale scores from four widely used personality inventories (the Personality Research Form, PRF; Adjective Check List, ACL; California Psychological Inventory, CPI; and the Interpersonal Adjective Scales, IAS). Correlational analyses revealed low to moderate congruence between the MDS technique and these traditional personality measures. Three multiple regression analyses were performed to test how well subjects' MDS distance scores could predict standard trait measures. Results from the first analysis indicated that MDS measures were better at predicting IAS self-esteem than the four traditional Inventories. The second analysis showed that MDS distance scores better predicted IAS dominance than the remaining three inventories. The purpose of the third regression analysis was to test the comparative predictive validity of the MDS scores with a different self-report method PRF dominance. The MDS measures placed last in this category, but were not far behind the others.Arts, Faculty ofPsychology, Department ofGraduat