How self-ratings skew performance reviews in unexpected ways

A multinational company discovered something unsettling in its annual review data: when managers saw employees rate themselves first, those self-evaluations became an invisible anchor, pulling manager scores in the same direction—regardless of actual job performance.

Researchers examining the company's performance review process found a pattern that repeated across demographic lines. Women and workers of color consistently rated themselves lower than their white male peers. And when managers reviewed those self-evaluations before submitting their own scores, they followed the same downward trajectory. Women of color, who had given themselves the lowest marks, received the lowest scores from managers. The effect was strongest for people of color: managers adjusted their ratings down more sharply for these employees than for white workers.

What made this troubling wasn't just the numbers. It was the mechanism. Managers weren't necessarily being consciously biased—they were being unconsciously anchored. A self-evaluation of "I performed well in this area" became a reference point, and subsequent manager ratings clustered around it. When that starting point was artificially low due to factors like imposter syndrome or cultural differences in how people talk about their own work, the entire evaluation moved down with it.

The Self-Rating Trap

The researchers couldn't prove intentional discrimination in the data, but the correlations were too consistent to ignore. The patterns aligned with demographic characteristics in ways the company found concerning enough to act on. Here's what made it worse: this wasn't a flaw in how managers thought about people. It was a flaw in the system itself—one that took existing disparities and baked them deeper into hiring, promotion, and pay decisions.

The company's response was straightforward: stop showing managers employee self-evaluations before they submit their own reviews. That single change disrupted the anchoring effect. When managers rated employees without seeing the self-appraisal first, scores across the board shifted down slightly—but the demographic disparities narrowed.

Other organizations are experimenting with different approaches. Some moved to quarterly feedback instead of annual reviews, which creates more frequent checkpoints to catch rating drift. Others added peer evaluations to the mix, reasoning that colleagues often see dimensions of performance that managers miss. A few have abandoned formal appraisals altogether in favor of regular, informal conversations—though that approach requires managers to be disciplined about actually having those conversations.

What researchers emphasize is less about finding the one perfect system and more about the discipline of looking. Most companies don't closely examine their review data by demographic group. They don't test what happens when you change the process. They assume the system is fair because it's written down and applied consistently. But consistency can perpetuate bias just as easily as it can prevent it.

The company in this study is now doing regular data analysis to spot emerging patterns. That visibility—knowing what the numbers actually show—is often the first step toward change that sticks.