Effects of post observation information on ratings of performance: The power of negative information in the information processing paradigm

Although previous performance appraisal research has demonstrated the presence of an assimilation effect when raters have been given indirect (i.e., not directly observed) information about ratees' prior performance before their present performance was observed and assessed, no studies have been designed to investigate the presence of an assimilation effect when indirect information is provided to raters after they have already had the opportunity to directly observe ratees' present performance, but before performance ratings are assigned.;In using a multigroup posttest-only design, 104 subjects were randomly assigned to one of four separate treatment groups (i.e., receiving either good or poor prior performance information, either before or after direct observation of present performance) or to a no-treatment control group. A Likert-type rating scale was used by subjects in each group to rate a videotaped lecture of average quality on nine appraisal dimensions.;A one-way ANOVA and Duncan's Multiple Range Test were used in analyzing the mean ratings which each group assigned to the ratee's lecture performance.;The results of the investigation revealed that subjects who received poor prior performance information assigned significantly lower performance ratings than did subjects who received either good prior performance information or no prior performance information. Thus, an assimilation effect was identified for the poor prior performance information groups.;In contrast, subjects who received good prior performance information assigned ratings which were marked by an assimilation effect only when their ratings were compared to those assigned by subjects in the poor prior performance information groups. Subjects in the good prior performance information groups assigned significantly higher ratings than did subjects in the poor prior performance information groups, but their ratings were not significantly different from those assigned by subjects in the control group.;It is suggested that investigations of the presence of context effects in performance appraisal research must use control groups if the results of such studies are to be properly evaluated. Suggestions for future research also are discussed in view of the clear impact that poor prior performance information had on the results of the study.