In managers' dynamic, real-world environments, they often feel it is necessary to exercise some creative discretion over employee ratings. Most managers do not describe their ratings of subordinates in performance appraisals as completely honest or accurate. The inaccuracy is often in the form of inflated ratings. They justify the inaccuracy by sighting, among other things, the need to avoid confrontation with subordinates, damaging working relationships, and creating permanent written documents which may later harm a subordinate's career. Many of these motives are not only well intentioned, but are designed to enhance individual, unit, and organizational performance (some of the ultimate objectives of performance appraisal systems.) This paper examines the ethics of this sort of deliberate manipulation of performance appraisal systems. It suggests that at the organizational level, performance appraisal is usually seen as an end in itself, and a formalist ethical critique is applied. At the managerial level, performance appraisal is usually seen as a means to an end, and a utilitarian critique is applied. Since both perspectives are essential, we conclude that a Janus-Headed analysis is needed. We suggest some duties and obligations for both the organization and the manager engaged in performance appraisal.