Can you interview candidates without bias?
Who should you hire? It’s important to recognise when you’re hiring, that the decision on who is best, is clouded in cognitive bias. The question is can you do anything about it?
There are a few technical solutions out there that help with the screening process, from masking voice phone calls to hide someones gender, to video chat programs that cloud someones face. But let’s say you look around your office and realise its a bunch of white men, and you want to do something about it.
You might say to yourself, “Let’s ensure we’re only interviewing the best candidates, and make sure we’re not overlooking more women in roles”, and you decide to do what we do at OHNO, and remove any personally identifiable information from the application process.
You might remove the following fields in your applications, before handing them to the hiring manager.
Then you think “Okay, we’ve done our part, and now we’ll attract more women into roles.”
I’m going to use this example copy of Marissa Mayers CV to show you just how much personally identifiable information is really present on a CV.
One of the challenges of removing gender based information off the application process, is you can still be biased towards things like nationality and race. When it comes to judging if someone will be good for the team, we often break people into two groups. Us, and them. What most people are trying to find out during recruitment is are these people one of us? Will they fit in well, and add value? Psychologists call this activity othering.
Humans categorise those we meet into tribes as a way to identify if these people mean us harm. It’s a construct we employ a lot in our lives. We are constantly trying to understand if the people we meet belong to our same tribe, and when we think of people as they, a group distinctly different from us, it can have powerfully negative consequences.
In an attempt to remove as much bias as possible from an application, lets remove anything that showcases that Marissa is Caucasian, and anything to do with gender.
Now the application will look something like this.
But is that it? If we look at the new CV, we might be influenced by one other element – Marrisa is probably wealthy and successful.
It is fair to assume that someone who has two degrees from Stanford either worked really hard to get accepted, or was simply able to financially afford a fine education. Going to anyuniversity is not on everyones roadmap.
These days, companies like Google and Apple are reconsidering how important university degrees are in their application process. They discovered that simply having a university degree didn’t make you better or worse, so now they don’t include requirements for those accreditations on the application.
Let’s go ahead and remove the education element so we can’t extrapolate any kind of wealthy background.
As we start to redact more and more of the application, recognising elements of the process that lead to bias, we can start to compare the non-biased elements and make judgements about an application based on experience alone.
But this is still highly flawed.
Why this fails
This trick alone will not fix problems of diversity for a simple reason, you still have to interview the candidates. And as soon as you do, all those redacted elements will come to light during your first discussion. You’re essentially punting the bias further down the recruitment field.
By stripping some of these proxies away, you will get a more diverse set of first round applications, but the second rounds may look like a bunch of white men again.
As you meet face to face, you will realise the candidate is caucasian. You might learn the persons gender, and you may discover they went to an Ivy league school. You are also likely to learn where someone lives, giving you a sense of their status in life. Maybe they live in a fancy neighbourhood?
When we see signals of success, we speculate if we bring those attributes in-house, some of that success might rub off. We are attracted, in a lot of cases, to those who win.
Why we judge so quickly
In 1993, Nalini Ambady and Robert Rosenthal created an experiment to see if students could guess the effectiveness of a teacher, with very little information, and have that correlate to actual measured performance.
The students were shown 3 different, 10 second video clips of a teacher in front of an entire class – with no sound. All the students could see was hand movements, the way the teachers carried themselves, and their confidence. At the end of the 10 seconds, the students were asked to rate the teachers on a scale of 1 – 9 on a variety of behaviours. Attentiveness, confidence, supportiveness, etc.
The remarkable discovery made was that the ratings the students gave were similar to the ratings given by educational administrators and other faculty members who had observed the teachers for over a year.
10 seconds of evaluation vs 1 year – and the results were the same.
So why can students evaluate performance in seconds and get the same results?
What researchers found is that most of the evaluation we do of people, do we like them, can we trust them, do we believe them – happens very quickly.
This behaviour was (and perhaps still is) incredibly important to our survival. When we were nomadic tribes, roaming the wilderness, when we would come into contact with strangers, it was important our instincts didn’t take long to work out if this person could help us find a meal, or if they wanted to kill us. So over the ages our brains developed a great sense of evaluation, and that part of our brain works very quickly.
Neurological shortcuts save us time and help us avoid harm. It is the reason we have so many biases today.
What this means is that when we prolong the judgement, we are mostly doing it on ceremonial grounds. In a job interview sense, what really happens is in the first few minutes, people decide if they like someone. Then, they spend the next 55 minutes running through a bunch of questions and reading into the answers in such a way as to validate the assessment they made to begin with.
Job interviews and 95% theatre, 5% instinct.
We pretend the whole interview process is much more objective than it really is.
Auditions are the key
This is why the best way of evaluating candidates is what we at OHNO refer to as an audition.
We take a piece of work, something we’d actually be doing on a given week, and work with the candidate on it to see how they really perform.
This means it doesn’t matter if you have a bunch of experience as a graphic designer. What counts is that you can deliver a design. Doing it this way means we also open up opportunities for people who are very good at something, but haven’t been able to break into the market yet.
We’ve all met candidates before who’ve had years of experience, but still manage to deliver really mediocre results. When we actually seeing how someone works on something, you get a much greater sense of what they are truly capable of.