‘Use data-driven approaches to reduce unconscious bias’

The founder of diversity recruiter Rare has called on the government to replace the current style of unconscious bias training in the civil service with a more targeted and data-driven training regime.

Rare founder Raph Mokades made the call after unconscious bias training in Whitehall was abruptly scrapped last week. He agreed that a lot of bias training was not effective but more modern methods should replace them.

He said that the real issue was that commonly used training techniques did not produce sustained results: “The government has argued that unconscious bias training doesn’t work, because it doesn’t lead to positive or lasting changes of behaviour,” had said. “Our research suggests that’s true of traditional training. Such training isn’t industry – or organisation – specific, and usually takes place as a one-off, box-ticking exercise.

“There is another way, however. Data-driven exercises, based upon context-specific experiences of an organisation’s own people, offering confidential, personalised action points for each user – this is a different proposition from generic feel-bad blanket training.”

He added that Rare’s Hemisphere software, being rolled out with several City institutions, was already proving this point among the firm’s clients.

“The government has said it is still committed to eradicating workplace discrimination in the civil service. It is vital that it stays true to that promise. We know, from the thousands of young people we have placed into even the most high-powered and forward-looking employers, that bias exists, that it damages an organisation’s performance, and creates injustices that can be devastating.”

More hi-tech, data driven software was far more effective than more traditional training, he said.

For example, Hemisphere asks participants to score clips from interviews of real candidates, who were in the process of applying to real firms at the time of their interviews. The software doesn’t use a proxy to measure bias, it uses the differences between the way users react to people when they can see and hear them and when they can’t. It then lets the individual user know what their biases are. The specificity of the training increases its validity and reliability.

The software has been developed over a five-year period, with qualitative and quantitative feedback analysed from City firms. The project was overseen by Dr Lara Zibarras, senior lecturer in the Organisational Psychology Group, at City University.

Dr Zibarras said: “Hemisphere is very context specific. It makes you question whether you are rating people on whether they can do the job or on some sort of unconscious bias. Hemisphere looks at gender, ethnicity and socio-economic status.

She added that the software made people question for themselves whether they were rating people in the correct way – on whether they could do the job.

Rare introduces talented ethnic-minority candidates to a range of top-tier employers, and also runs the Target Oxbridge university access programme.

Hemisphere has launched with Ashurst, Blackrock, Clifford Chance, DC Advisory, Freshfields and Herbert Smith Freehills.

Behavioural Insights’ ‘limited evidence’

Rare explained that the government’s decision to scrap unconscious bias training in the civil service was based on research from Behavioural Insights. Mokades said: “Many of its criticisms of conventional training are valid.” He added that just as training needed to be individually relevant to be really compelling, BI’s view that changes in implicit bias tend to be temporary was correct. “Training against bias, just like any other kind of training, needs to be topped up regularly to really ‘stick’.”

However, there was much in the BI research that Rare said it questioned. Mokades said that BI itself admitted there were limitations in its evidence, and he felt these were significant. He said: “It concedes that different training programmes vary hugely, that much of its data comes from university student populations rather than workplaces, and that there is an over-representation of US-based studies. BI claims that there is ‘no evidence’ unconscious bias training has improved workplace equality in terms of representation of women or minority groups. Assuming we all agree there has been at least some improvement in representation, however, it seems simplistic to conclude that training deserves no credit at all.”

Rare pointed out the government had said it was still determined to eliminate discrimination in the civil service. But in Mokades view it had not yet offered positive proposals for doing this, having eliminated workplace bias training, which accounted for a large chunk of its diversity and inclusion budget. He said: “A few gleeful commentators are seemingly taking the move as ’proof’ that bias against minority groups has itself been overstated.

“We know, from the testimony and experience of hundreds of young people placed into even the most high-powered and forward-looking employers, that bias is a real problem – dragging business performance and creating sometimes terrible consequences for its victims.

“Tackling it requires not just training but a broader diversity and inclusion strategy, including data collection, and analysis of adverse impacts in recruitment, promotion and retention.”

Our clients are telling us this combination works.

Latest HR job opportunities on Personnel Today

Browse more human resources jobs