Is my recruitment process biased?

Published by:
Rich Marr
November 26, 2018
7
min read

Our world is full of bias. It’s like an invisible force that flows through us, within us and all around us, touching almost everything we do.

It’s typically not intentional, but our behaviour is riddled with it, and our neutral-seeming business processes are frequently built upon it from the foundations up.

Causes of recruitment bias

Unfortunately, there are many ways for bias to show itself, at every stage of the recruitment process, here are some examples:

1. Job definition bias

The way you define a role and describe it to candidates can create unintentional bias.

Words can carry subconscious meaning that signal to candidates whether or not they'd belong in your organization.

Using words typically associated with males such as “competitive” or “decisive” can give the impression of a boiler-room culture in which many people (disproportionately women) would rather not work.

Just take a look at the results of our own research below - using masculine language resulted in less applications from women.

Gender coding effect on gender diversity


Including hard requirements that aren’t really *hard* requirements can also cause problems. Men tend to be comfortable applying for jobs when they meet 70% of the requirements, but women tend to apply for jobs where they meet 100%, likely due to the fact that women are more socialized to follow the rules (whilst men are socialized to break them).

Some people will read your requirements as a strict set of criteria, whilst others will see them as mere desirables.

Reasons for not applying for jobs

Similarly, requiring a specific academic background or work experience assumes that everyone has the same access to opportunities.

If people from underprivileged backgrounds find it harder to attend universities and therefore gain the best experience, then being too specific around these requirements only perpetuated this inequality.

When it comes to benefits, there's also potential for bias...

Requiring a role to be fully co-located or inflexible hours will rule specific groups out, for example those with part time caring responsibilities, working parents, and those for whom travel is difficult.


2. Sourcing bias

Your choice of where to advertise your role is important too. Different recruitment agencies, job boards and communities have different demographic makeup, so even a simple innocent choice of job board can create systemic bias.

Employee referrals can also lead to bias. Referred candidates tend to reflect the demographic of the employee who referred them.

And according to PayScale’s report, female and minority applicants are significantly less likely to receive a referral than their white male peers.

3. Sifting bias

CVs contain many potential triggers for bias.

Typical CV sifts are based either on keywords or on overall impressions - neither of which are an objective means of shortlisting the most suitable candidates.

When faced with a stack of CVs, even the most well intentioned hirers fall back on shortcut-based decision-making, often grounded in unconscious bias.

A 2019 study from the University of Oxford found that candidates from minority ethnic backgrounds had to send 80% more applications to get the same results as a White-British person.

Screening bias study

And these results have been mirrored all around the western world.

Bias studies by country

Even if we were to anonymize applications, someone's education and experience can lead us to either favour or overlook them unfairly.

Sometimes hirers look for red flags such as rejecting any CVs that have an unexplained gap in their employment history. These filters are usually based on intuition, and frequently cause bias; people from less well-off backgrounds are more frequently required to take time off to care for relatives.

Effect of experience gaps on callbacks

4. Interview bias

The most common type of interview is the unstructured interview, where the interviewer asks an unpredictable mix of questions, either favourites or spontaneously, and sometimes takes unstructured notes about the candidate.

These interviews are based on establishing rapport, essentially answering the question “do I want to work with this person”.

The problem is that this is driven largely by rapport, which is an irrational and naturally biased process. Human beings tend to establish rapport more easily with people who are like themselves.

The chart below is from a U.S. study, which looked at how our perceptions of ‘out-groups’.

As you can see, we form ideas about how warm, friendly and competent someone is just from looking at them.

Stereotype Bias Chart


Problems still exist in structured interviews, but having a framework of questions and a clear marking scheme enables interviewers to remain focused and keep consistent, as well as helping them avoid relying on rapport.

Interview bias is a problem
Mansplaining is not the answer

5. Shortlist bias

A recent study showed that if you only have one woman on a four-person shortlist she has effectively no chance of getting the job… whereas if you have two men and two women the numbers are closer to 50/50 that you’ll hire a woman.

Finalist pool diversity vs hiring decisions


This clear gap in rationality is a good reason why shortlists should be representative. If the best candidate is a man then he would still be hired from a representative shortlist, but if the best candidate is a woman then using a representative shortlist makes a substantial difference to her chances… and hence improves employers chances of hiring the best candidate.

Similar effects should be expected in other dimensions of diversity, so if you use shortlists as a tool, try to guard against those irrational behaviours… for example when you show decision makers the shortlist you could have it next to previous shortlists, reducing the effect of “one of these things is not like the others”.

6. Onboarding bias

Once you’ve made your hire there’s still a way to go before they’re a productive member of your team.

An advantage of hiring within your network is that incoming hires arrive with a halo, additionally they have a built-in support network because they already know somebody who can help them avoid pitfalls, politics, etc. When you hire someone who’s dissimilar to your existing team those things can be harder to achieve, and at the same time there’s a danger your team may see them as a “diversity hire” who didn’t qualify on merit. This can cause enough friction within teams to become a self-fulfilling prophecy.

To avoid that, it’s important that your team understand the way hiring decisions are made. Ideally they should be involved at all stages, and deeply understand (and be able to examine) how candidates are scored. If you do that, all hires will arrive with a halo and be given a fair shot by your team.

So what now? How can we achieve diverse recruitment?

Unbiased hiring is a journey, not a destination, so rather than focusing on how to hire you need to focus on how to improve. To do that I’d recommend three things:

1. Measure everything

To detect and diagnose problems in your hiring process you need to know protected data from the people in your hiring funnel and how they perform as they progress through your hiring process. To assess the performance of the mouth of your funnel you also need to find out the demographic makeup of the wider talent market as a whole. If you want to truly perfect your hiring, you also need to be able to match up every point in your process with who got hired and how well they performed once they were in the role.

2. Adopt blind recruitment techniques where possible

Human bias is ubiquitous, and training is expensive (and has mixed results at best) so hiring ‘blind’ is a useful technique. There are many ways to approach it, all the way from simply scrubbing out names from CVs, to operating a blinded and replicated skill-based assessment like we do at Applied . The key is to maximise human judgement by helping it focus on what’s important, and getting anything that’s not important out of the way. Read our blog on What is Blind Hiring? to learn more about this.

3. Evaluate candidates on the actual work they’ll be doing

How well someone did in school, or whether you respect the university they went to, are very weakly correlated with how well they’re likely to do once they’re hired (Schmidt/Hunter 1998,2016)… more predictive assessment techniques tend to more closely resemble the actual day to day work they’ll do on the job.

What next?

Read our follow up blog on How do I remove bias from my recruitment process? to learn what to do next. Also check out our resource centre for a whole bunch of guides and best practices on de-biasing your recruitment, and if you’re really keen check out our infographic of the 20 most common biases that creep into recruitment.

Applied is the essential platform for fairer hiring. Purpose-built to make hiring ethical and predictive, our platform uses anonymised applications and skills-based assessments to improve diversity and identify the best talent.

Start transforming your hiring now: book in a demo of the Applied platform or browse our ready-to-use, science-backed talent assessments.