7 Strategies for reducing unconscious bias in candidate pre-screening

Recently, we dug into our database of nearly 2 million email outreach sequences and nearly 12 million candidates who entered our customers’ hiring funnels within the last year. We were curious about the ways the hiring pipeline plays out as it pertains to gender and to race/ethnicity. The story the data told us is that unconscious bias tends to have the strongest impact on diversity initiatives before talent has the opportunity to demonstrate their proficiencies through a phone screen or test. 

Here’s what we mean by that. When it comes to gender, nearly 2x as many male candidates enter process as female candidates; but female candidates pass through subsequent stages at higher rates, outperforming their male peers. 51% of female talent is extended offers after onsites, compared to only 42% of male talent:‍

And despite the fact that White candidates see higher passthrough rates at the very top of the funnel (from Application Created → Pre-Onsite), Black and Hispanic/Latinx talent see higher passthrough rates across remaining funnel stages: 62% of Black talent and 57% of Hispanic/Latinx talent is extended offers after onsites, compared to 54% of White talent. Black talent ultimately sees the highest overall passthrough rates (1.52% versus 1.46% for Hispanic/Latinx talent, 1.36% for White talent, and 0.6% for Asian talent). 

In other words, while fewer of them make it to a pre-onsite stage (think phone screens or a take-home test), the women who do enter our customers’ hiring funnels outperform their male peers. The same goes for Black and Hispanic/Latinx talent, who are less likely to be extended a phone screen opportunity... but who are more likely to be hired if they get one. 

‍All of this suggests that, by and large, recruiting teams have a more equitable interview process than they think; diversity is more a top-of-funnel problem. Unconscious bias has the strongest impact on gender and race/ethnicity before recruiters have their first point of contact with candidates. In light of this, here are 7 strategies for reducing bias in candidate pre-screening—regardless of what pre-screening processes you have in place. After all, the data suggests that if your org can bring more women, Black, and Hispanic/Latinx talent in at the top of the funnel, you’ll see the same (or better) success rates as you see with other candidates.

1. Re-evaluate Your Criteria

Reducing unconscious bias in candidate pre-screening begins with the question: How do we ensure that our selection is based solely on criteria that’s relevant to the role? Consider the fact that the things you’ve come to look for on talent’s LinkedIn profiles and in their resumes (education, work experience) ultimately have little-to-no correlation with future work performance. Requiring four-year degrees from certain institutions, for example, biases you toward privilege… and it’s an unreliable performance indicator. 

Gender bias is at play in engineering roles that require that candidates have spent time contributing to open-source software (spaces female engineers have experienced hostility in). Gender and racial bias is at play in roles that require leadership experience. And so on. So when a role opens, clarify what competencies and qualifications are absolutely necessary to success in the role. If you must, separate out “must haves” and “nice-to-haves.” 

And rather than focusing on talent’s experience, education, or—if they’re early-in-career—GPA, ask yourself what about their history suggests cognitive ability, growth mindset, or creative problem-solving. 

2. Put Sourcers Through Unconscious Bias Trainings for Your Outbound Efforts

Of course, you should be building awareness about unconscious bias across your entire hiring team (and in your org at large) all the time; but if you have an outbound function that’s dedicated to filling the top of the funnel, this is critical. Proactive sourcing is the only way you can control the diversity of talent you bring into your process, so a sourcing team that’s unaware of their own biases will mean a regrettably homogenous funnel. 

So partner with DEI professionals to conduct trainings with sourcers (and recruiters, and interviewers, and hiring managers). Paradigm and Catalyst are among the organizations that offer team-wide unconscious bias trainings, but there are plenty of resources out there, on LinkedIn and elsewhere. Your sourcers should ideally be able to name their own biases when they come up, recognize the ways those biases impact who they’re reaching out to, and hold themselves accountable to adjusting their outreach. These are things that demand ruthless honesty; but if you can build a culture of transparency around them, your team can learn to address their biases together. 

Otherwise to reduce unconscious bias in candidate pre-screening include creating goals on your outbound team to source X number of candidates from underrepresented groups per role. If you’re sourcing talent on LinkedIn, check out its “Hide Candidate Photos and Names” feature—which does exactly what its name promises—so that sourcers evaluate candidates based only on their skill sets, rather than on their appearance. Unbias.io is a Chrome extension that removes the same information (name and photo):‍

Source: Unbias.io

3. Consider Blind Resume Screenings

Blind resume screenings (also known as “blind hiring”) anonymize the personal and demographic details on a candidate’s resume that could invite bias, such as:

• Name (studies show that we correlate pronounceability with trustworthiness and that Black candidates who “whiten” their names on resumes receive more than twice as many callbacks—even from organizations with pro-diversity language such as “minorities strongly encouraged to apply”)

• Gender 

• Age (for example, graduation year)

• Educational information such as schools attended or GPA (consider the ways degree inflation biases career paths like vocational training and on-the-job experience)

• Address (zip codes can give away socio-economic information, trigger affinity bias, or lead to assumptions about commute-time tolerance)

• Previous companies (just because someone worked at a FAANG company, or for an org whose brand you love, does not automatically mean they’ll thrive in yours)

• Title or rank (access to titles is not equal: women, URGs, and people with disabilities are promoted far less than their White, male, able-bodied counterparts)

Other resume elements you might remove to reduce bias in candidate pre-screening include personal interests or hobbies, number and duration of jobs, or gaps in employment (which might bias candidates who took extended maternal or paternal leave, trans folks who took time off for gender reassignment, or “job hopping” military spouses), or association memberships. Don’t look for keywords: these tell you that the candidate knows what you’re looking for, not that they’re right for your role. And they screen folks out who don’t know the right industry terminology or keywords.‍

There are plenty of tools out there that can anonymize resumes; but DIY approaches include asking candidates to strip personal information from their resumes before submitting them, or assigning a team member who anonymizes them. Assign each resume a number; and what you’re left with is skill sets, accomplishments, merits, and a growth trajectory to go on. 

If you have neither anonymizing tools nor the time for a full-fledged DIY approach, try this: look at candidates’ resumes from the bottom up. If you start at the top, the first thing you see is name, title, and education—the most “inessential” information. Recruiters take an average of 7 seconds to look at a resume. If you spend most of that seven seconds at the top, you’ve already determined “yes” or “no” based on inessentials. Starting at the bottom means you’re making decisions based on experience.

4. Consider Eliminating Resumes Altogether

Replacing the resume with skills testing is perhaps the most equitable solution of all—after all, if your org has proactive diversity goals in place, anonymizing resumes may ultimately make it more difficult for managers to ensure they’re hiring diverse departments. 

Skills testing ensures you’ll hire based on performance rather than background. Everything else the candidate would have listed on their resume becomes irrelevant in light of whether they can simply do the work the role requires. This could look any number of ways: give candidates a problem or a challenge to solve, or describe a project they recently built. These kinds of questions could be built into a pre-screen (candidates send in their answers to the challenge or a sample of their work rather than a resume) or a phone screen.‍

Having candidates complete these kinds of skills assessments early on in the hiring process levels the playing field, because their performance speaks for itself. Of course, bias can always creep in later on in the interview stage (though ongoing bias trainings can help moderate this); but the data shows that once women and URGs make it into your funnel, they’ve overcome the hardest obstacle. 

5. Ensure There’s Diversity Among Your Resume Reviewers and Screeners

Diverse perspectives on your resume-reviewing-and-screening team will make it much less likely that bias will creep in early on in the funnel. So consider a range of representation here: backgrounds, experiences, perspectives. If you’re working with a smaller team, consider regular rotations so that new folks are screening every few weeks—which means that one person’s bias isn’t inhibiting the diversity in your pool. 

6. Standardize Your Screening Questions or Scoresheets

We talked about skills testing in place of resumes above; if you go this route make sure you’ve predetermined what an “unsuccessful,” an “acceptable,” and an “excellent” pre-screen assignment looks like. To mitigate further bias in candidate pre-screens, use scorecards to objectively grade assessments, and to stay focused on what matters.

Scorecards and standardized questions are just as important for phone screens. You simply can’t “compare apples to apples” or consider two candidates on an even playing field if you’re not asking them the same questions, or scoring them in the same way. Bias happens even in a phone call—the sound of someone’s voice, or their tone, or their apparent charisma can call up “gut instinct.” Standardized questions keep the phone screen on task, leaving no room for personal judgment. Even if screeners “connect” more with certain candidates than others, there’s no place to record similarity bias on a scorecard.

While we’re at it, regularly examine the reasons your team disqualifies candidates at the pre-screen or phone screen stage. Rejection reasons like “not a good fit” won’t cut it. Solutions like Gem will let you track where women and URGs are dropping out of the hiring process or getting disproportionately rejected. If you see more top-of-funnel drop-off from Black candidates than White ones, you might be dealing with some bias at, or before, the pre-onsite stage. 

7. Postpone Social Media Pre-Screening 

This isn’t to say get rid of this step in your process entirely; it’s to say: consider when you do it. If social screening is a routine practice for you, we recommend waiting until after the first round of interviews to go looking for more “data” about your candidate there. After all, you may learn a lot about talent on their social feeds—everything from their favorite sports team to those same hobbies we recommended you scrub from their resumes—that could invite bias. 

Go ahead and look for candidates on social media after you’ve determined they’re qualified and met them face-to-face. After all, you’ll want to know about red flags like values-misalignment before you extend an offer. But you’ll want to be clear that you eliminated them because of that misalignment, and not because of unconscious bias in your candidate pre-screen processes.

To learn more about this year’s data, download Gem’s full Recruiting Benchmarks Report—packed with outreach stats, full-funnel conversion rate averages, and diversity data—here.