Following a Harvard Business School study that suggested employers are missing out on talent due to recruitment bias and algorithms, SiliconRepublic.com took a closer look at these issues.
It’s no secret that the world is going through a critical skills shortage, especially when it comes to a wide range of tech careers.
To take just one area as an example, the cybersecurity space is facing a shortage of more than 199,000 professionals across Europe, according to an (ISC)2 report released last year.
Shortages have only been exacerbated due to what has been dubbed the ‘great resignation’. This phenomenon started in the US during the pandemic, with employees quitting their jobs in droves. A Workhuman survey from last year indicated that the same trend may be happening in Ireland too.
But while employers are crying out for talent to plug these gaps, a damning report last year suggested that employers are ignoring millions of workers through a mixture of recruitment bias and algorithms.
The report from Harvard Business School and Accenture said that historically, cycles of economic recessions and recoveries led to workers being displaced and subsequently rehired and reabsorbed into the workforce. However, long before the onset of Covid-19, significant structural issues created imbalances in labour markets across the developed world.
“With each cycle, an increasing percentage of working-age adults remain outside of the workforce,” the report said. “In the recovery phase of each downturn, those newly isolated workers have faced serious consequences. Extended gaps appear in their employment histories. With each passing month, they risk falling further behind in maintaining the skills employers want.”
How algorithms block talent
The report found that this has led to job postings becoming harder to fill and, as technology has evolved, more and more companies are turning to automation and other tools to find their talent.
“Even though online platforms expand access to opportunities for jobseekers, they make it harder for workers who do not closely match the requirements instantiated in those job descriptions,” it said.
“Millions of workers, at all skill levels, can’t find the work they want, for the hours they want to work, for positions that they are deemed qualified for by that technology. Considered to be less qualified when assessed relative to candidates who fit the hiring company’s criteria more exactly, such applicants were ‘not visible’ to recruiters.”
Sandra Healy is the CEO and co-founder of diversity and inclusion platform Inclusio as well as the founding director of Dublin City University’s Centre of Excellence for Diversity and Inclusion.
She told SiliconRepublic.com that while algorithms and software can be used to screen candidates and make the jobs of the HR team easier, there is risk that the software will remove those who don’t fit very narrow criteria.
“This has been documented to be the case many times whether it’s a sourcing algorithm, screening algorithm or otherwise,” she said.
Anthony McDonnell, full professor of human resource management at University College Cork, added that one of the biggest issues with algorithms in hiring practices is that many of them are technologically led rather than scientifically led.
“We know about discrimination in hiring, gender and racial discrimination for 30 or 40 years. We also know while there’s been some improvements in many of those areas, it’s still not being reversed,” he said.
“So, the idea that we think that technology is going to automatically remove some of them, I think is problematic.”
Examining human bias first
Speaking to Healy’s point that algorithms might miss people who don’t fit the exact criteria laid out in a job description, McDonnell said that even without the algorithm, job descriptions can be “notoriously problematic” in terms of missing out on talent.
“First of all, lots of organisations don’t update them, so they’re written from a position and it still might be the same type of position but there hasn’t been a new process to see how those recruits have performed.”
He also said companies have to evaluate what should be considered mandatory criteria for the role and think about whether or not certain skills can be learned on the job.
“We know there’s a genderised aspect to that,” he added. “Words that are associated with more masculine qualities, we know that that’s having an impact on diversity because research is showing it.”
Research published in the Journal of Personality and Social Psychology showed job advertisements that used masculine wording were less appealing to women.
Another aspect of excluding hidden talent is the kind of education required, especially in the tech space. McDonnell said there are jobs that now require degrees that never needed them before, even if the job hasn’t changed.
Additionally, by adding skills, a certain amount of years’ experience or a particular level of education without really considering what traits and skills can be trained as part of the role, a lot of potential talent can count themselves out.
“Potentially, there is a lot of really good candidates out there that aren’t being factored in,” said McDonnell.
Healy added that while algorithms can cut a lot of great potential candidates, human bias continues to be a major factor too.
“Our brain makes heuristics, or shortcuts, in order to function more quickly and learn patterns and skills. Bias is the consequence or product of heuristics,” she said.
“There’s also a similarity bias – in which humans are keener to like someone who is similar to them in some way. So of course, hiring practices are susceptible to this – whether it’s on the basis of class, race, ethnicity, shared group membership or even values and life circumstances, that can cause a gap on a CV.”
What companies can do
The report on the hidden workforce speaks to much larger issues within HR and recruitment that can’t easily be fixed. However, Healy and McDonnell both talked about where companies could start.
“Go back to the root of sourcing and employer branding,” said Healy. “What language, pictures and other branding elements are being communicated externally? Who are they targeting, speaking or not speaking to? How can you meet people where they are and forge relationships with diverse employee groups?”
Like McDonnell, she said one of the most common mistakes companies make is around the language used in job advertisements. She also said going back to the same place to find talent over and over again is a mistake. “I always say to companies, change where you look if you want to drive diversity.”
McDonnell agreed, adding that companies should not only think about where they’re looking for talent but also examine the results. He said that when organisations do change their recruitment process, they “tend not to really analyse what is it leading to real differences in their application pool”.
“When you’re doing something different, it is important that you’re comparing previous campaigns and recruitment with anything that you’re doing different.”
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.