Previous research has shown that employee evaluations tend to be biased. When evaluating job performance, appraisers tend to judge inappropriately between employees. In our experiment, we let subjects perform a group task that is observed and evaluated by two appraisers. When the task is completed, one appraiser records her evaluation verbally while the other appraiser records her evaluation in writing. Therefore, we investigate whether there is a difference between written and oral evaluations and which format reflects subject performance most accurately.