Why Gender Bias Persists in the Workplace

Harvard Business Review – Research: Why Gender Bias Persists, Even When Organizations Try to Curb It (hbr.org)

 By Alison Wynn,  Emily K. Carian, Sofia Kennedy,  and JoAnne Wehner

Summary.   New research provides insight into why there’s been little progress on advancing women to leadership positions. The authors worked with a professional services firm, which implemented training to reduce manager bias in performance evaluations.

Illustration by Ngadi Smart

After devoting time, effort, and resources to reducing bias, why are organizations still struggling to see the impact on the advancement of women to leadership positions?

In newly published research at the VMware Women’s Leadership Innovation Lab at Stanford University, we worked with a top professional services firm we call PS1, which implemented training to reduce manager bias in performance evaluations. We found that although managers were able to curb differences in how they described employees’ traits and behaviors (“viewing biases”), they were unlikely to catch differences in how they rewarded employees for those traits and behaviors (“valuing biases”).

Our research provides insight into why we’ve seen so little progress: Most attempts at advancing gender equity focus on viewing biases, while valuing biases continue to hold us back. Leaders must consider both types of bias when designing interventions. Here’s what our work with PS1 revealed about viewing and valuing biases — and how managers can mitigate them to advance gender equity in the workplace.

Diagnosing Bias

In our work with PS1, we started by diagnosing bias in their performance-evaluation process. PS1 identified the junior partner level as one with a significant drop in women’s representation compared to lower levels. They provided us with the content and numerical ratings of 40 performance evaluations (half women, half men in the junior partner role) from a single year, which we analyzed to try to determine whether bias patterns in evaluations could help explain the gender representation gap. Ratings ranged from 1 to 5, where a 5 rating increased an employee’s chances of receiving a promotion, raise, and/or bonus, and a 3 or below reduced their chances.

We identified several areas of viewing bias in PS1’s evaluations. For example, we found evidence of gender biases in descriptions of employees’ communication styles. Specifically, managers were more likely to describe men as having “too soft” a communication style, subtly faulting them for falling short of assertive masculine ideals.

We also found valuing bias regarding gendered expectations of work commitment. Men were given lower ratings than women when they failed to display behaviors associated with work commitment, such as working long hours, traveling frequently, relocating for work, and sacrificing their personal lives to meet work demands. However, when evaluations mentioned utilization — the percentage of their available time that employees bill to their clients — women’s ratings suffered more than men’s. In an organization where the most effective employees bill the vast majority of their work hours to clients, the pattern of penalizing women more for insufficient billable hours suggests that managers consider women less committed to their careers.

Addressing the Low-Hanging Fruit

In response to our report and recommendations, PS1 designed a training intervention for managers. By training managers to reduce differences in the topics and language used in men and women’s evaluations, PS1 hoped to address inequities in the evaluations that ultimately shape advancement decisions.

Two years later, after PS1 implemented their manager training, we analyzed a new set of performance evaluations to see whether patterns of bias had changed.

We found that the managers had been able to eliminate several viewing biases, or gendered differences in the way they described employees. For example, they decreased gendered differences in comments about communication style and personality. They were also less likely to describe women as a “risky investment” or compare them to an unspecified bar or standard, and more likely to use tangible numerical metrics in women’s evaluations.

Identifying Persistent Valuing Biases

Even though PS1 successfully addressed many instances of viewing bias, several clear patterns of unequal evaluations remained. When we analyzed what tied those patterns together, we realized they were primarily biases in how managers valued employee behaviors based on their gender.

For instance, when men’s evaluations described their specific project work, they were more likely to receive a top rating. In contrast, similar descriptions in women’s evaluations did not pay off for them. When the status, impacts, and/or activities of women’s projects were discussed, these descriptions did not significantly improve women’s ratings the way they did for men, even though managers increased the frequency of these descriptions for both men and women after the intervention.

We also found that communal behavior, such as being collaborative and a “team player,” was noted just as often in both men and women’s evaluations and that managers rewarded men more for it with higher ratings. Managers linked men’s communal behavior to strategic bottom-line impacts, whereas they interpreted and presented women’s communal behavior merely as evidence of how they support their team.

In one evaluation, a man received the following feedback, connecting his communal behavior with PS1’s success: “[Name] has built an intimate, counseling relationship with [the client leader]… Engagements have had material bottom-line impact and have contributed to client’s spectacular [outcomes].”

In contrast, a woman’s communal behaviors weren’t linked to business outcomes, but instead were described as reflective of her personality and values. For example: “A very positive leader, [Name] is seen as a role model on values and caring. The passion, energy, and sense of purpose she is bringing […] is praised by all. […] A great colleague, selfless with peers, very collaborative, brings in colleagues in the situations she creates.”

Understanding Why Biases Persist

Why was it so much harder for PS1 to address valuing biases compared to viewing biases? Valuing biases are more difficult to measure and thus tend to remain unaddressed in many bias-mitigation efforts. Our unique study design allowed us to see inside a process that is usually invisible and identify valuing as an ignored step of the performance review cycle.

In our sample, the overall numeric ratings between men and women appeared equitable: There was no gender difference in average ratings. But when we analyzed the correlation between specific behaviors and ratings, we found that the same behaviors paid off differently by gender.

Bias erodes the connection between an employee’s behaviors and the final rating they receive. Without taking active steps to measure and standardize the connection between demonstrated behaviors and ratings, organizations open the door to inequity.

These findings matter because managers must continuously observe and assess employee behavior, formally and informally. In addition to hiring decisions and formal performance evaluations, managers give career-advancing developmental feedback, allocate stretch assignments, and offer support. In doing so, they constantly view employee behaviors, skills, and performance outcomes, and make decisions about their value in the form of compensation, promotions, and assignments. Valuing the same behaviors differently by gender propels men and women into different pathways for advancement. For example, in the technology field, women and men are often pushed into different career paths based on stereotypes about their skills: Women are often urged into non-technical roles because of their perceived social skills and lack of technical skills, whereas men are encouraged to pursue a technical career track.

How to Reduce Viewing and Valuing Biases

To address bias more effectively, organizations must interrogate how their employees are viewed and valued throughout their careers.

Provide manager support.

Viewing biases are often visible in gendered language used during hiring and evaluation. In fact, many companies offer their managers software tools designed to identify and mitigate gendered language in job postings and performance evaluations. But managers lack similar kinds of support for ensuring they value employees’ behaviors equally. In PS1’s case, managers were provided with a detailed rubric to evaluate employees’ performance, but how they actually rated employees for demonstrating specific behaviors still varied. While companies may not be able to analyze the correlations between evaluation language and rating as we did, they can create processes that help managers take viewing and valuing into account when making decisions about employees.

One way to provide this much-needed support is to develop guidelines to help managers move from descriptions of employees to final ratings and rewards. If an employee demonstrates a specific behavior, such as forming close partnerships with clients, provide guidance for how that behavior should be rewarded. Even more critical: Ensure these standards are applied consistently across business units and teams.

Even for those companies moving away from formal performance evaluations, eliminating written evaluations doesn’t erase manager biases. Though written evaluations provide the data we used to analyze bias in our research, biases operate even when they’re not written down. Organizations must work to standardize specific guidelines for how managers should view employee behaviors and assign corresponding rewards when giving employees feedback and making decisions about their careers.

Ensure manager training is comprehensive.

Trainings too often address only one step in the process of bias reduction: how managers view their employees. Our research shows that interventions must be two-step to be comprehensive.

If managers only understand one type of bias, they’ll remain ignorant of other ways biases can infiltrate their decision making. Ensure managers and leaders understand what valuing biases are. PS1’s intervention trained managers about the existence of gender bias but did not equip them to monitor how their descriptions of employees tie to ratings. Valuing bias can be difficult to see and measure, so organizations must take extra steps to educate managers about it.

Create fair procedures throughout employees’ careers.

Bias doesn’t just emerge once or twice a year during performance evaluations — it affects employees throughout their careers. Every stage where rewards are distributed presents an opportunity for bias to emerge — or for organizations to implement fair procedures instead. During recruitment, hiring, project assignments, coaching and mentoring, promotion, pay, and even termination decisions, consider how employee actions are viewed and valued. Develop rubrics and set criteria for viewing and valuing employee behaviors at every stage of their careers.

. . .

By expanding bias reduction beyond viewing to include valuing, organizations can take an important step toward successfully addressing bias and creating a more equitable workplace.

Извор: WUNRN – 12.07.2024