Negativity (and Positivity) in Performance Evaluation: Three Field Studies

Because of a lack of powerful nonlinear models, there is little research about nonlinearity in performance evaluation in nonexperimental, real-world data. Nonlinearity in 3 real-world data sets of performance evaluation was examined by using various versions of a nonlinear model labeled the scatter model. The findings indicate that performance evaluations tend to be conjunctive, that is, more weight is given to negative attributes than to positive attributes. However, this basic tendency disappears when the overall level of evaluation is high, as a result of inconsistency resolution-the tendency to resolve inconsistency between 2 or more aspects of the input information on the basis of an overall evaluation.