
Assessing the talent of the future

Research has evidenced the existence of unconscious bias in the recruitment process, so how do we get rid of it?
As the HR Director of a major UK-based publisher told HR Grapevine recently, humans are wired to identify and bond with people who display similar traits as our own and as such it’s natural to want to give these people preferential treatment. “Of course we all have unconscious biases. That’s a product of being human. Yet if we know about these biases and we don’t work to ensure that these aren’t affecting the way we hire and interact with others, then we’re failing,” he said.
Whilst unconscious biases are anecdotally evidenced in our everyday actions, studies have proven just how much they affect our natural decision-making processes; scientists at the University of Washington and Yale created a specific study to discover how biased most people are. The test, coined the Implicit Association Test (IAT), led to the publishing of a 1998 paper stating that the ‘unconscious roots of prejudice’ affected 90% to 95% of all people. This, in the UK, leads to an advantage for white job applicants of 74% over BAME candidates.
As the publishing HR Director stated, HR and resourcing teams must be aware of these biases and act upon them in the recruitment process, yet how can they do this? The answer is through the use of software specifically designed to diminish the effects of, or completely remove, this issue when filtering candidates - but do they work?
Will Hamilton, CEO and Co-founder of LaunchPad Recruits states that the question holds within it some ambiguity, as removing bias comes down to no one element. He says: “That depends on your answers to two questions: 1. Was the assessment / AI model that predicts a candidate’s suitability built with enough data to learn what good looks like? 2. Did it pass the subsequent test for adverse impact? If you can answer ‘yes’ to both, you’ve removed the potential for the subjective opinion of any one person to adversely affect vulnerable sub-groups.”
Of course we all have unconscious biases. That’s a product of being human
Mandy Watson, Managing Director of resourcing firm Ambitions Personnel, believes that assessments are definitely an effective way to eliminate bias, but that other stages in the process are still able to sabotage a recruitment drive. “Assessments can go some way to removing bias in a recruitment process, depending on how they are structured and administered, and how much weight their results carry in the overall decision making,” she says.
“However, it’s unusual for assessment to be the one and only stage of a hiring process, so it depends on how the rest of the process is conducted; is an interview done by a panel and are they asking all candidates the same structured questions and using a scoring system?” she adds. “Bias is extremely difficult to eradicate fully, any human touchpoint of a recruitment process is potentially open to bias.”
Whilst Watson is only wary of the inevitable human intervention in the recruitment process, not all recruiters are so open to the advantages that the assessment testing itself offers. For example, Joshua Smith, CEO and Founder of Hourly Recruitment, who is sceptical of relinquishing control to programmes that are, inevitably, still created by humans. “They're engineered by humans to find the ideal candidate according to your bias. Hiring is all about effective bias. You know the type of person you want (mindset/culture-fit/personality) and that's what you're recruiting for, nothing wrong with that,” he says.
Smith’s opinions on bias may be in the minority, but his scepticism on the origins of all assessment technology has historical grounding; in October of 2018 Amazon confirmed that it had scrapped the use of recruiting tool it had previously touted as the ‘fairest way’ to take on the concept of sourcing new talent. The system failed not because it deviated from its coded objectives, but because its machine learning was polluted with the biases of those who created it. In fact, so ingrained in the human psyche is bias that some psychologists such as Professor Brian Nosek at the University of Virginia have concluded that, whilst the IAT is still the leading form of gauging bias, there’s literally no accurate way of measuring its presence in any given process.
You know the type of person you want (mindset/culture-fit/personality) and that's what you're recruiting for, nothing wrong with that
Regardless of the questionable nature of any assessments touting the removal of bias altogether, wider trends in recruitment seem to support the view that assessments are at least a moderately effective way of eliminating it from the process; A McKinsey study found that businesses who successfully implemented assessment programmes into their business not only attracted a broader range of talent, but also that more diverse teams operated far above national medians as a result. Businesses such as Unilever, Vodafone, Singapore Airlines and Intel are amongst those adopting similar processes in the hopes of reducing bias. In the case of Unilever, it claims that prior to adopting these platforms, it hired one out of every three applicants; it now claims that this number has gone up to two out of three and that whereas in 2010 38% of its managers were women, in 2018 that number had risen to 47%.
This is something that HR Transformation Specialist and Author Rita Trehan says she’s seen great advancements from: “Psychometric testing, especially when combined with the power of anonymous applications, can really help to remove bias and ensure that organisations are giving potential candidates equal opportunities,” she says.
“In my opinion, the more objectivity and diversity of sources we take into account rather negates any bias that could arise because of a person’s name, age, gender and or nationality. Ultimately, if companies want the very best candidate, using psychometrics amongst other things prevents you from blocking candidates, that are an ideal fit,” she concludes.
Ultimately, if companies want the very best candidate, using psychometrics amongst other things prevents you from blocking candidates