September 20, 2021
Written by Merrit Stüven, MPP '22
Edited by Anisha Asundi, WAPPP Research Fellow, and 
Moira Notarstefano, WAPPP Communications Manager

In the first talk of WAPPP’s fall seminar series “Intersectional Approaches to Gender Equity,” Professor Iris Bohnet discusses how race and gender intersect in employee performance evaluations. 

Performance appraisals play an important but controversial role in determining career advancement. Many organizations use performance reviews to determine raises, promotions, and, in some cases, penalties for poor performance. Bias and discrimination during these reviews could significantly impact workers’ career prospects—and their lives.

How can organizations design performance appraisal systems to address potential bias and discrimination? Albert Pratt Professor of Business and Government and WAPPP co-director Iris Bohnet, along with co-authors Ariella Kristal and Oliver Hauser, seeks to answer that question by first investigating how employee evaluations are impacted by gender and race. During a talk on September 9, Bohnet described the researchers’ analysis of data from four years of performance appraisals at a global financial services company, including self-evaluation scores employees gave themselves and the scores managers assigned to their employees.

Screenshot of Iris Bohnet's Zoom PresentationAccording to Bohnet, for both employees and their managers, “It is very rare that we have objective performance information available; there are lots of biases that could creep in. Stereotypes, for example, but also others that could affect people’s judgments.” This was reflected in the employees’ own self-evaluations: women, especially women of color, and Black men assigned themselves significantly lower performance scores than white male employees. These results were not surprising, as previous research on self evaluations has shown women tend to rate their performance lower than men. What was of particular interest to Bohnet and her research team, however, was disentangling the biases imposed by managers from the racial and gender inequities in the self-evaluations. Like many other organizations, the financial services company that supplied the data used an appraisal system in which managers see employees’ self-evaluations before deciding the scores they assign their employees.

Bohnet’s study revealed that, across the board, managers gave their employees lower ratings than employees gave themselves. However, employees of color were downgraded more harshly, while women overall were downgraded less during managers’ assessments. Of her statistical analysis, Bohnet said: “The race effect is not due to self-ratings. Managers correct for women’s lower ratings, in particular, for white women’s lower ratings, but they add this race decrease, so to speak, where they take the ratings of people of color down more on average.” Due to the intersecting impact of gender and race, women of color were downgraded during the evaluation process, along with men of color. In contrast, white women benefitted and ended up with the highest overall rankings.

Does this gender and racial bias persist when managers do not see self evaluations? Bohnet and her co-authors were able to look at inequities in the performance evaluation process when managers could not anchor their scores on the rankings employees gave themselves, as the company did not share self-evaluations in 2016. They found that the removal of self-rating did not change evaluation scores for women and people of color, as managers simply relied on other potentially biased anchors, such as employees’ performance evaluations from the previous year. The only group that may have benefited from removing self-evaluations were new employees, in particular, newly hired women of color, for whom managers could not rely on past ratings. Not having their traditionally lower self-evaluations available, managers on average rated women of color the same way they rated men of color.

Bohnet concluded her talk by noting that she and her research team are continuing to investigate the complicated ways bias and discrimination can impact employee performance appraisals, particularly for women and people of color. The research does suggest that removing self-evaluations from the process is unlikely to be enough to prevent bias in performance appraisals. Instead, organizations could focus on the role of managers in accurately assessing performance and recognizing their own biases, for example, by requiring a ‘calibration’ process designed to surface racial and gender gaps.