Everyone is conscious of unconscious hiring bias. Here’s how small changes to your recruiting processes can assist clients and help them to deliver a fairer candidate experience.
Unconscious hiring bias: it’s a tricky subject to tackle, but one which recruiters can benefit from by confronting head on.
Today, everyone is conscious of unconscious bias, which is a good thing. What we have all learned is that even the fairest and most even-handed recruitment process can be exposed to flaws and entrenched biases if left unscrutinised.
This article looks at some practical, technological, and behavioural steps you can adopt as a recruiter to help your clients remove unconscious bias from their hires. In doing so, you ensure that you are adding genuine value to their recruitment equation.
Unconscious bias – what is it?
Hiring bias can take many forms. In its shortest definition, it is any hiring process which is weighted to exclude the selection of a specific demographic of applicants. Other forms of discrimination can include positive discrimination. This is when a select demographic has a comparatively better chance of being selected than other groups.
The key to resolving hiring bias is to acknowledge that even the most fair-minded recruiter or employer can slip up. Errors are often created in old and untested working processes. Modernise your selection process, and you reduce the risk of bias.
• selection via name
What’s in a name? An awful lot, according to long-term hiring trends. Your name is a significant cultural marker. When employers are looking for a good cultural fit for their team, an applicant’s name can say a lot. Even when the bias is not deliberate, a name can say so much that its impact on decision-making can be immediate and unconscious.
• Selection via gender
Equality laws ensure that no employer can legally deselect candidates because of their race, gender, age or sexuality. Yet there is a persistent fear that this still occurs in today’s workplace. The belief that certain jobs are only for certain genders is an idea from a bygone age. Yet these attitudes still persist – though sometimes from surprisingly modern sources.
In 2018, retail giant Amazon created a recruitment AI which taught itself to be sexist. But this wasn’t deliberate gender bias. The artificial intelligence used deep learning to understand why previous candidates had been successful in the past. The data pool it used contained primarily successful male applicants. Consequently, the computer immediately deleted any applicant who had attended an all-female school as being unsuitable for its roles.
The article is a lesson that, sometimes, even a smart modern solution can create new ways of being stupid.
• selection via background
The Amazon article also highlights the final point of bias: discrimination against background. This may include exclusion of candidates who attended a certain school, as in the article above. Or, it may mean that some unsuited candidates are chosen because of the school or university they attended in the past. This is the suspicion that an “old school tie” can open more doors than the right qualifications.
At its smallest level, discrimination by background can even mean the exclusion of candidates who live within unpopular postcodes. This can be a particular obstacle for regional recruiters who make placements to their local area. But is there a good reason why employers should require access to an applicant’s address?
How recruiters are reducing Hiring bias
Fortunately, there are modern recruiting techniques which address all of these pitfalls. And the final point of the previous section brings us to the first of those solutions for hiring bias.
Anonymise your candidates
Think about the types of information that your client HR actually requires to fulfil their role. Is some personal data irrelevant to the hiring process? Then strongly consider omitting it. If the client asks you why, you can explain that you are making their job easier. By anonymising your own candidates, you are shielding the HR department from accusations of bias, because they do not have access to the information which would allow them to make a biased hiring decision in the first place.
It is a cultural shift in the recruitment process – and one which you should expect to field questions on in the initial stages of transition. Fortunately, the practical steps for achieving this are fairly simple. Most applicant tracking systems will provide some form of anonymisation for your candidate data.
An equal application process
Consider the application process for each of your candidates. Is the process the same for everyone? Today’s world of extensive digital networks means people will discover jobs – and your agency – through any number of different channels. But, if you concentrate all of your efforts on only one of these talent pipelines, are you responsible for creating an unconscious bias? Linkedin has very different user demographics to Facebook, or Twitter. Candidates who enter the process through your website may be very different to those who enter via a jobs portal integration.
This is the type of business data which you as a recruiter should always have at least one eye on. If you continue to see an imbalance in applicant demographics, maybe you should invest some of your time in redressing that balance with new pipelines into a wider variety of audiences?
Lastly, you can use modern solutions to this old problem. Automations which won’t make biased choices are becoming more affordable, even for smaller recruitment firms. The eBoss Match tool allows you to automate longlisting with a single click. Simply access the job description on your database, and click Match. The database then retrieves every stored candidate who possesses the right qualifications for the role – regardless of their name, background, gender or sexuality. You are then free to anonymise this candidate data and in a few short minutes you have a bias-free and bias-resistant shortlist of hopefuls for your client. This is how you remove hiring bias and add genuine value
to the recruitment equation.