| Processes of observing, classifying, and rating behavior were investigated in an assessment center simulation. The study examined the influence of the implicit distribution assumptions of assessors (Fox & Thornton, 1987) and the number of dimensions on observation and classification accuracy and on the accuracy and discriminant validity of dimension ratings. One hundred and eight introductory psychology students observed and rated the performance of three confederates in three exercises.;Subjects who showed high person and trait differentiation (high differentiators) or low person and trait differentiation (low differentiators) were selected to participate. They were trained as assessors and rated candidates on either three, six, or nine dimensions.;Although high differentiators observed more accurately than low differentiators, differentiator type had no significant effect on classification accuracy or the accuracy or discriminant validity of assessors' ratings. However, the number of dimensions rated significantly affected the accuracy of assessors' classifications and ratings. Subjects who rated a small number of dimensions classified behaviors more accurately and made more accurate ratings than did subjects who rated a higher number of dimensions. Neither differentiator type nor the number of dimensions significantly affected the discriminant validity of assessors' ratings. Conclusions and implications of the research findings are discussed. |
