This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Find out more here

FEATURE23 November 2018

Bias in business

Behavioural economics News UK

A panel at the MRS Methodology in Context conference set out to examine the role that unconscious bias plays in business decision-making. By Christian Walsh.

The panel consisted of: Jay Owens from Pulsar; Jan Gooding, chair of Stonewall & president of MRS; Jake Steadman from Twitter and Colin Strong from Ipsos Mori. Both Owens and Gooding are members of the MRS Delphi Group which published a report yesterday entitled ‘Deconstructing bias: Lessons from 70 years of research and insight’ which aims to share advice to help decision makers across the business – from HR to data analytics and all the way up to the boardroom.

The first hurdle for the panel presented by chair Tim Phillips was defining unconscious bias.

"For me," said Steadman, "it’s about learned stereotypes. The importance is on the word ‘learned’. They are taught and can be untaught."

Given the current ‘vogue’ for staff training in unconscious bias, Gooding wasn’t wholly comfortable with the notion that a person could be "trained out of bias". She referenced cases where companies have experienced worse behaviour after training because it has legitimised a person’s actions. "They can just say 'this is me, this is who I am'," said Gooding.

Strong in particular was uncomfortable with the language. The term ‘unconscious’ can offer a smokescreen for what is clearly discriminatory behaviour he said. "Is this ‘unconscious’ or ‘unexamined’ behaviour? Someone who is racist or sexist should be called out as being that."

To shortcut what can fast turn into a philosophical debate, all were in agreement that we should focus first on conscious activity – plenty of which is discriminatory – before worrying about what was or wasn’t happening unconsciously.

Owens agreed: "Instead of requiring everybody to do what Freud couldn’t – to control the unconscious – we should be asking can you control your behaviour?"

There was also a risk continued Owens of putting too much emphasis on the individual, and engendering a sense of guilt in that person. Rather, we should be thinking about the structures of the organisations and the values in which they work. "The solution," she said, "happens at a collective level – at a structural and process driven level."

Steadman agreed but stressed that the individual still needs to take some responsibility for their actions.

As former global inclusion director at Aviva, Gooding has plenty of experience working at scale within an institution. "We all desire to fit in," she said, "and this is where company culture is powerful. The thing that is interesting to me is not to be afraid to be different. I remember in the past recruiting people who were ‘the right fit'."

Lack of diversity, said Gooding, is clearly linked to less creativity, innovation and ultimately a less productive work force.

When recruiting at Twitter, Steadman is keen to ensure his team displays ‘cognitive diversity'. "I believe the research industry is biased towards academic ways of thinking and institutions – there’s a lot of Oxbridge people. When I hire I try to recruit from diverse academic backgrounds."

What about the practice of using ‘blind CVs’ – removing personal details like name and age – to counter bias?

"Blind CVs are a cop out," said Gooding. "Let them [candidates] tell us who they are. It’s up to us to learn not to discriminate."

At Aviva, Gooding was more interested in measuring and monitoring recruitment across the organisation to tackle biased hiring. "I think data is what is so powerful, and questioning the outcome. If we are not getting the outcome we want [a diverse workforce] we need to change the recruitment process. The trouble with that is it takes time and effort, and its easier not to do it."

As the Delphi report explores, bias in a business can be present in the data that organisations are increasingly dependent on. Algorithms that have bias ‘baked in’ (often unintentionally) can then produce results which are used to make the wrong decisions.

"Data cannot be objective," said Strong, where even the selection of data can introduce bias. 

While ‘garbage in, garbage out’ is certainly a risk as we rely more and more on AI, Owens speculated that in future "there may be a better data driven system that as well as being measurable, can help us check our preconceptions, and be the better angels of our nature."

So what can the research sector do in terms of improving its own approach to diversity and inclusion?

Owens felt a crucial issue was lack of 'brand visibility’ for the research sector and the types of jobs available within it at a parental level, especially compared with other recognised professions like lawyers and doctors.

For businesses, the real question said Gooding is "are you going to make the effort to put this at the centre of your business strategy because you believe it drives growth?"

And as researchers, she continued: "You have to believe that you will be enriched in your work if your colleagues are more diverse."

Strong closed the session with a statement that encapsulated both the challenge and the opportunity for the sector in terms of recruiting a more diverse workforce: "We don’t know what we don’t know. Our job is all about understanding different perspectives and the danger is that our perspectives become narrower and narrower. That’s the opportunity, and it’s greater for us because that’s what we’re all about. If we don’t challenge ourselves how can we ply our trade with clients honestly?"

You can download the MRS Delphi Group report ‘Deconstructing bias: lessons from 70 years of research and insight here.

0 Comments