Diversity drives are wasted without the data


THE movement against unconscious bias training, once a key element of employers’ diversity and inclusion strategies, is growing.

A controversial UK report into racial inequality has called for organisations to drop the courses entirely. It may actually be too soon to write them off, but investor pressure may help find something better.

Courses to tackle unconscious bias seek to address the problem that deep-seated, reflexive ways of thinking lead to discriminatory behaviour. While there’s plenty of evidence that such ingrained prejudices exist, studies have struggled to establish whether these often one-off sessions make the slightest difference to what people actually do.

Worse, such training may even do harm. Britain’s Equality and Human Rights Commission has argued that it could backfire by suggesting stereotypes and biases are unchangeable.

Employees who have taken such courses may feel they have cover to discriminate. The British civil service decided to phase out its courses on the subject last year.

Infamously, the former head of auditor KPMG’s UK arm told staff that the training had never resulted in any improvement and, hence, unconscious bias was “crap”. He apologised and left days later say- ing his position was untenable.

Led by educational consultant Tony Sewell, this has generated a storm by concluding that the concept of “institutional racism” is overused and distracts from other factors that contribute to racial inequality.

However, the body wasn’t being radical when it said diversity training doesn’t work, and that it merely exemplifies how organisations prioritise process over outcomes.

Ad-hoc sessions on unconscious bias are clearly no silver bullet. The mounting consensus against training that raises employees’ awareness of how their words and actions can disadvantage colleagues of a different race, gender or sexual orientation nevertheless merits some caution.

The potential benefits may be hard to prove simply because there’s no standard model of what these courses should consist of. So, it’s possible that a form of unconscious bias training could be effective — it just hasn’t been developed yet.

Such training may have a role to play as an amplifier of other policies. Frederick Herbert, a behavioural scientist with the London School of Economics’ Inclusion Initiative, recently argued that unconscious bias training at least raises people’s knowledge of “implicit prejudicial associations”.

Such awareness may not be sufficient to change behaviour, but it is a requisite for doing so, Herbert suggests. The question is what to replace unconscious bias training with, or how to build on it in a way that has measurable impact.

The Sewell commission advocates giving ethnic minority staff more exposure to peers, managers and decision-makers, and providing “routine skills” support that could “disproportionately benefit more disadvantaged groups.”

It also favours “nudge” measures. “Sometimes, the changes are so simple that they seem trite, for example, using images of successful colleagues with an ethnic minority background on walls,” it says.

But the commission also says more work is needed to identify what really works to advance workplace fairness. If that’s the case, it may be counterproductive to be too prescriptive in precisely how employers address disparities.

The result could be more ineffectual box-ticking. The business value of hiring from a diverse talent pool and creating an inclusive culture is beyond doubt. Better for investors to hold boards to account for achieving that, and pressure companies to find the best solutions.

This will happen faster if shareholders have the necessary information to form a view. The latest Workforce Disclosure Initiative survey by ShareAction, the responsible investment campaigner, showed that just 4% of companies provided data on their ethnicity pay gap (allowing for legal restrictions on collecting the data), compared to 57% in the case of their gender pay gap.

Just 36% of the companies surveyed their workforce’s ethnic composition, against 75% for gender.

The Sewell report puts little additional pressure on employers to disclose more detailed pay data if they aren’t doing so already. It acknowledges the challenges of ethnicity pay disclosure given multiple classifications and geographical differences.

But that shouldn’t excuse companies from including this data in their annual reports, and then providing a narrative context if they feel their situation is more nuanced than the numbers can capture.

As ShareAction says, without workforce composition data, companies can’t design initiatives that will respond to the needs of their staff. Training may have work to do to justify its place in the fairness toolkit.

But there’s no excuse for companies to dodge a data-backed commentary on what they’re doing to promote an inclusive culture. — Bloomberg

  • This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.