Skip to main content

Why (Graduate) Recruitment should be free of Hindsight Bias

The Universities of Amsterdam and Utrecht in the Netherlands recently published results of their field study into discrimination of applicants with a minority background. The researchers sent out over 4.000 applications (letter and CV) of fictional applicants, aged between 23 and 25, as a reaction to actual vacancies. The fictional applicants with a native Dutch background had a 40 percent higher probability of being invited for a job interview when compared to the fictional applicants with a migration background. It turned out that applicants with a minority background were structurally discriminated against no matter how much information they shared in their CV.

I was wondering how this affects the graduate applicants entering the labor market, having completed their education, collecting sizeable debt and wanting to start their professional careers. What does it mean to a person to know that no matter your study background, grades, extra-curricular activities and working experience during your studies, you have a significantly lower probability of getting a job? That’s like living in Spain as a graduate where youth unemployment is still at 30% (albeit improved from 42% in 2014). There is limited place for you no matter your skills, efforts and talent. This can generate a generation of people in the workforce who are undervalued. This is plain and simply wrong, especially if the jobs are there and everyone is screaming talent shortage.

After reading the study results, I realised there is good, bad and ugly related news linked to these results. Let me explain why.

The Good

First of all, there is nothing good with the outcome of this research as it underlines a fundamental problem in our societies. However, it is good that we are aware of these issues and can now design strategies to deal with the issues. A study by Hiemstra at the Erasmus University in Rotterdam in 2013 already showed that applicants with a weaker labour market position were relatively positive about video resumes compared to paper resumes in terms of fairness and justice. Add the fairness aspect of structured interviews with competency-based in interviews, and the Video CV or structured video interview become a powerful tool to counter bias, especially in an applicant group that has hardly any CV to show for as they have been studying up to that point. Video interviews also allow to interview at scale and therefor open up the job interview stage to many more applicants compared to the traditional on-site interview where only very few candidates are invited to come in for an interview. You could even consider giving every single applicant the opportunity to participate in a structured video interview in order to provide a fair and open process. Cammio automated interviews is already applied by many organisations in graduate recruitment.

The Bad

There is only one problem and that is that most organisations simply still do not have a plan for structured interviews. Their interview process is random, powered by gut-feel rather than science or psychological insights to assess someones’ potential. This is really easy to fix however. There is an abundance of interview guidelines available, or consultants who can help with competency-based interview scripts. And when delivered at scale with Cammio, this functionality can be made available globally.

The Ugly

With an estimated 20 to 40% of employers using some form of Artificial Intelligence in their recruitment process, there is a clear risk of institutionalising hindsight bias. This type of bias channels experience into decision-making, which can lead to mistakes. Experience is sometimes simply equivalent with biased historic data. For example black Americans make up 37.5% of the US prison population despite the fact that they make up just 13% of the US population – badly written algorithms fed these datasets might predict that black people are more likely to commit crime. “About 130 million US adults are already in face recognition databases,” Timnit Gebru, technical co-lead of Google’s Ethical Artificial Intelligence Team, told the AI for Good Summit in Geneva in May. “But the original datasets are mostly white and male, so biased against darker skin types – there are huge error rates by skin type and gender.” The Californian city of San Francisco recently banned the use of FR by transport and law enforcement agencies in an acknowledgement of its imperfections and threats to civil liberties. But other cities in the US, and other countries around the world, are trialling the technology.

Facial recognition however already sees some adaptation when it comes to recruitment tech. Most companies however use this technology to create an emotional profile indicating if someone is happy, frustrated, excited and so on. We are also experimenting with this technology at Cammio, with the clear difference that we aim to develop a personality profile and not an emotional profile. The latter we have deemed unfair to candidates. Our research is also in combination with academic partners to ensure we create an algorithm that is fair, unbiased and had longevity, not one that is based on short-term emotional triggers. And today we already have natural language processing (NLP) and verbal reason AI in place with our Cammio Xpress Analytics tool that helps add a personality profile to the candidate. Available to all companies using Cammio today.

So let’s leave our hindsight bias behind. Graduates of any background deserve to be assessed on their true potential. Let’s welcome our new workforce generations with open arms and make sure everyone can get an interview, always. Now that we know the problem, let’s fix it. As recruitment and HR community we should feel responsible to make the professional future accessible.


Cammio Graduate Campaign