Why do performance evaluations become biased?

The source research demonstrates that managers are significantly influenced by their own personal incentives and relationship dynamics when performing subjective evaluations, specifically through centrality bias (compressing ratings) and leniency bias (inflating ratings). These biases are often used as defensive mechanisms; managers tend to provide more biased ratings when the costs of gathering performance information are high or when they have a strong relationship with a subordinate, as they seek to avoid the time-consuming and psychologically burdensome task of justifying harsh ratings or damaging personal bonds. While the study found that centrality bias negatively affects future performance by weakening the link between pay and effort, leniency bias was actually found to have a positive effect on future employee performance. This positive outcome supports a behavioral theory perspective, suggesting that lenient ratings can enhance an employee's perception of fairness, which in turn boosts motivation and productivity more effectively than strictly accurate, non-inflated assessments.

Performance evaluation compression occurs when managers inflate the ratings of weaker performers to minimize personal costs, such as the psychological discomfort of delivering negative feedback and the time-consuming nature of justifying low scores. While managers generally prefer to satisfy all employees by aligning ratings with their often-inflated self-perceptions, this strategy is governed by the interaction between information accuracy and outcome transparency. When outcome transparency is low, managers can compress ratings without fear of upsetting high performers because those employees remain unaware that their superior efforts are being undervalued relative to their peers. However, in highly transparent systems, high performers can observe peer rewards and are likely to become demotivated if they see little differentiation despite their superior performance. In these transparent environments, information accuracy becomes the deciding factor; managers are only willing to differentiate and reduce compression when they have access to highly accurate performance data. High accuracy lowers the manager's confrontation costs by providing a more defensible basis for low ratings while simultaneously making high performers less forgiving of a compressed distribution that ignores clear signals of their superior effort. Consequently, experimental evidence suggests that investments in accurate performance measurement systems are only effective at reducing rating bias when paired with policies that make evaluation outcomes visible across the organization.

Employee performance rating calibration is an organizational process where a committee, typically composed of human resource (HR) managers and direct supervisors, reviews and potentially adjusts the initial performance ratings submitted by supervisors. The primary goal of this additional step is to correct for inconsistencies, incompleteness, and rater biases—such as leniency or favoritism—to ensure that final ratings clearly differentiate between top, average, and poor performers. While top management often mandates a specific rating distribution to enforce standards, direct supervisors frequently provide inflated initial ratings to avoid interpersonal conflict, increase employee motivation, or reward personal friends. This creates an incentive conflict where supervisors seek to protect their likely inflated ratings while the calibration committee is tasked with making predominantly downward adjustments to meet organizational distribution requirements. Research within the sources indicates that these outcomes are heavily influenced by the social and political dynamics within the committee. For instance, although higher information availability about peer performance generally leads to more rating adjustments, this effect is attenuated for influential supervisors, suggesting that committee members may be strategic in the information they choose to share. Furthermore, supervisors who possess greater organizational influence or who belong to informal alliances (such as those within the same functional area) are significantly more successful at avoiding scrutiny and defending their ratings against downward adjustments. Ultimately, while calibration can successfully align an organization's overall rating distribution with management's goals, it may not necessarily improve the accuracy of individual ratings if the process is undermined by these strategic behaviors.

Calibration committees function as a vital, though resource-intensive, mechanism for enhancing the fairness and accuracy of performance evaluations by standardizing rating criteria and mitigating individual rater biases. These committees facilitate a broader information exchange, allowing managers to share insights about employees' cross-functional contributions that a single direct supervisor might overlook. While they are highly valued for promoting consistency and trust in leadership, their effectiveness is often contingent on their design and the organizational environment. For instance, when committees are composed strictly of peer supervisors, incentive conflicts can arise; research indicates that influential supervisors or those in informal alliances may successfully shield their employees from downward rating adjustments, even when a forced distribution curve necessitates them. To combat these political dynamics and ensure a truly inclusive evaluation process, organizations can implement hybrid committee structures that include senior management, utilize frame-of-reference (FOR) training, and involve proactive HR oversight to monitor for inequitable patterns in rating changes. Ultimately, while calibration is a powerful tool for achieving procedural and distributional justice, it requires intentional management and transparency to avoid becoming merely another layer of biased bureaucracy.