Rethinking Employee Self-Appraisals: Insights on Bias and Evaluations

Companies routinely utilize annual employee evaluations to make pivotal decisions concerning promotions, raises, and layoffs. However, recent studies underscore that these evaluations can be significantly influenced by extraneous factors, notably gender and race. A recent paper delves into how these elements affect worker ratings within a multinational firm, particularly under conditions where managers access self-evaluations prior to submitting their assessments.

The findings illustrate that when managers review self-appraisals first, their evaluations tend to closely align with employees’ self-ratings, reflecting an "anchoring" effect. Women, alongside workers of color, generally assigned themselves lower scores. Women of color rated themselves the least favorably and consequently received the lowest scores from their managers. Intriguingly, when managers evaluated without prior access to self-appraisals, they typically assigned lower ratings across the board.

To further explore this matter, I spoke with Iris Bohnet, a co-author of the study and a professor at Harvard Kennedy School. Bohnet, who specializes in behavioral economics and gender studies, shared insights on how biases rooted in gender and race influence employee assessments and potential remedies for these biases.

The Impact of Self-Assessments

Bohnet expresses long-standing concerns regarding self-evaluation practices in performance appraisals. Employees typically present self-evaluations to their managers, which can lead to biased outcomes based on the initial scores people assign themselves.

Behavioral research indicates that any form of numerical assessment serves as an anchor for subsequent judgments. If, for example, one employee rates themselves a seven while another rates themselves a nine, managers are likely to be influenced by those figures, irrespective of their actual performance. Bohnet notes that historically, women and people of color tend to undervalue their own abilities, fostering disparities in the evaluation process.

“On the employee side, we do find that women give themselves low self-ratings, and in particular, women of color give themselves even lower ratings than white women,” Bohnet explains.

Patterns of Evaluation and Manager Bias

There is robust evidence that women receive lower scores than men, while employees of color are rated lower than their white counterparts, indicating potential bias from management. The dynamics of these evaluations reveal complex behaviors: Managers may either universally lower ratings, disproportionately lower ratings for employees of color, or provide less severe reductions for women compared to men.

A notable finding in Bohnet's study is that the gap between self-evaluations and manager ratings was less pronounced among all-female employee groups. Understanding why this occurs remains a matter of inquiry. Similarly complex interactions emerge when examining women of color, who start with lower self-assessments but do receive a degree of leniency from managers.

However, when stacked against their white peers, this leniency isn’t sufficiently compensatory for the already low self-evaluations women of color provide. Furthermore, the intensity of scrutiny appears to escalate for all individuals of color, antagonistically influencing their ratings.

Effects of Removing Self-Appraisal Access

During a particular operational hiccup, managers did not receive prior access to employee self-evaluations that year. This anomalous scenario produced observable shifts in ratings: managers’ evaluations became less correlated with self-assessments. Despite the overall drop in scores, the gender and racial dynamics remained consistent—ratings were uniformly lower, yet the entrenched patterns persisted.

More intriguingly, managers likely reverted to previous self-evaluations when making assessments that year, which indicates a reliance on historical data rather than fresh evaluations. According to Bohnet, “Managers really rely on these self-evaluations,” underscoring the need for structural change in the appraisal system.

Recommendations for Fairer Evaluations

To address issues stemming from self-evaluations, some firms have opted to omit them entirely from the review process. Bohnet advocates for this, stressing the importance of regular data analysis similar to the one conducted in her study to highlight potential inequalities within evaluation systems. Additionally, she suggests employing alternative strategies such as quarterly evaluations to eliminate biases and provide a more comprehensive performance picture.

The concept of peer evaluations is also gaining traction, which would allow for a broader perspective beyond just the manager's viewpoint. Several organizations have even moved towards eliminating formal evaluations in favor of more frequent feedback exchanges, resulting in more productive outcomes.

Ultimately, while there isn't a one-size-fits-all solution, the data signal a pressing need for organizations to scrutinize their evaluation metrics and identify effective practices that promote accuracy and fairness in employee assessments.

As Bohnet aptly notes, “We need more organizations to take a close look at their data and test what works and what doesn’t.” Only through diligent examination and testing can workplaces cultivate equitable evaluation systems that reflect true employee performance.