Despite our best intentions, we all have biases, both conscious and unconscious. And in the workplace, these biases — defined as systematic beliefs that make us favor one thing over another, often resulting in errors of judgements and unfair results — are as common as clogged sewer lines.

But with an increasingly sharp focus on diversity, equity and inclusion (DEI) in today’s workplace, organizations are paying more attention to bias mitigation strategies than ever before. One particular area of emphasis, for example, is eliminating biased hiring practices, which range from the way job postings are worded to how resumes are screened to the ways interviews are conducted.

But there’s another area that also deserves scrutiny: biases in performance reviews.

“I wouldn’t say performance bias is something new, but it’s just perhaps bubbling to the surface,” says Kristen Swigart, a senior people scientist at Culture Amp, a consulting firm that helps organizations improve employee engagement, retention, diversity, development and performance. “Organizations are focusing on DEI, most importantly that equity piece — homing in on all aspects of decision-making to ensure systems and processes are equitable.

“Biases are particularly serious in making high-stakes decisions about people — things such as hiring, promotions and performance reviews. These are very consequential moments in time for both employees and organizations, so it’s really important to take as many steps as possible to mitigate biases in those processes.”

When bias creeps into performance management, organizations run the risk of not promoting or rewarding the right people — employees who deserve greater compensation or to hold a higher position, Swigart says.

“But ultimately, performance bias can also negatively impact productivity,” she says. “When employees sense bias, it undermines their perception that things are objective and fair and that can lead to reduced productivity, less sense of belonging, lower levels of engagement and higher turnover.”

In addition, performance biases also contribute to pay gaps between men and women.

How can organizations determine if performance-review processes are biased? Swigart says there are ways to audit performance tools and processes. These audits look for rates of promotion, for instance, among different groups of employees, as well as examine the kinds of feedback that managers give to employees, or that peers give to each other.

“It’s important to note that you can’t completely eliminate bias — it’s a systematic kind of error in our thinking,” Swigart says. “So what we want to do is aim to mitigate and minimize biases.”

Furthermore, after implementing strategies to mitigate biases, it’s critical that organizations monitor how successful they are by performing periodic audits.

“You can’t just create a process and then never return to it again,” she says. “You need to audit processes, too, in order to ensure the strategies continue to reduce bias over time.”

There are many types of performance bias. A common one is gender bias. Research shows that men tend to receive more specific and work-related behavioral feedback while women are much more likely to receive vague or personality-based feedback.

Why does this happen? Psychologists explain it with a concept called role-congruity theory, where for a variety of long-standing, well-ingrained societal biases, men are seen as more suited for leadership positions while women tend to be seen as less stable, which creates the proverbial “glass ceiling.”

Other kinds of performance bias include:

• Recency bias, where managers judge employees on what they did well or poorly in the near past instead of looking at overall performance.

• Primacy bias, in which a positive or negative first impression outweighs overall performance.

• The halo- or horns-effect bias — a tendency to let one good or bad trait take precedence over others.

• Centrality bias, in which managers give all of their direct reports middle-of-the-road performance grades — threes on a scale of one to five, for instance — to avoid appearing to be extreme or hurting employees’ feelings.

• Similar-to-me bias, where managers give higher marks to employees that possess the same skills, interests, opinions, backgrounds, etc. as theirs.

• Idiosyncratic-rater bias, in which a manager give employees high skill-evaluation ratings if they’re great at something he or she does poorly, or conversely gives employees lower marks for doing things the manager believes he or she does very well.

• Confirmation bias, where managers specifically look for data or interpret information in ways that confirm their pre-existing beliefs.

One way to mitigate performance bias involves using more structured review tools to collect feedback. When review tools are unstructured, it enables biases to creep in because there’s no defined criteria; that, in turn, creates a vacuum in which managers define performance criteria based on their own biases and intuitions, Swigart says.

A good example of a structured review tool is a so-called situation-behavior impact model that relies on a fill-in-the-blanks approach instead of what Swigart calls “a big, open-ended box.” It starts with a manager picking a certain situation he or she has observed (either positive or negative). Then the manager describes the behavior observed, followed by the impact of the behavior, which then serves as a starting point for a two-way conversation.

For instance, a manager can explain that last week during an important client meeting (the situation), an employee interrupted the manager’s presentation (a specific behavior, stated factually without judgment), which made the manager lose focus. That then caused frustration and embarrassment (the impact).

Or conversely, during an important client meeting, the manager was fumbling for information that the employee gracefully supplied and tactfully offered without being asked — an impressive display of emotional intelligence and keeping cool in a high-pressure situation.

“This approach allows feedback to be more constructive and meaningful and ultimately more actionable,” Swigart says.

Another way to eliminate performance bias is to use a rating scale with fewer options, such as four instead of five. A typical rating scale with five evaluation options — below average, slightly below average, average, slightly above average and above average — makes it easy for managers who want to avoid making tough decisions because they can give everyone the average rating.

Moreover, there’s rarely any definition of what each ranking means, which allows managers to impose their own criteria onto the rating scale. The danger of that is what looks like average to one person may perhaps be above or below average to someone else.

So Swigart recommends using only four rating options, with only one negative, or low-performance, option and three positive options that distinctly differentiate between good on one end and exceeds expectations at the top end.

“This approach reduces leniency bias,” she says. “When managers avoid confrontations with employees, it comes from a good place. But it can skew the objectivity of the data and make it difficult to differentiate between employees.

“If everyone gets an above-average rating, for example, it’s hard to distinguish who to promote. And this kind of rating system shows employees that extra effort and stellar performance gets you more than just an average or above-average rating. Ultimately it becomes a more engaging and motivating process for employees.”