Journal of Applied Psychology 1992
A realistic simulation for assessing the relationships among components of rating accuracy.
Abstract
Laboratory investigations reveal that components of rating accuracy are often independent of one another and contribute distinct pieces of information. The generalizability of these laboratory results to field settings may be limited by real-life rating dynamics. In this study, the ability to generate performance true scores in realistic settings allows an assessment of the extent of independence of components of rating accuracy under conditions approximating those in field settings. Results obtained under realistic conditions from a sample of 125 subjects were compared with results of a meta-analysis of laboratory results. Overall, the two sets of results are not significantly different; however, some correlations, particularly those involving rating leniency, are lower in the realistic setting than in the laboratory
- DOI
- 10.1037/0021-9010.77.2.196
- Sources
- openalex