Unconscious bias and diversity training — the evidence

Photo by Matt Hoffman on Unsplash

The corporate buzzwords of the moment: unconscious bias and diversity training. These training programmes have been introduced to organisations across the world over decades, with high hopes that they will make workplaces more inclusive. In the US alone, companies spend $8billion a year on diversity training. But do they work?

This is a question we ask attendees at our ‘What works for diversity and inclusion?’ workshops regularly. Between about 11% and 50% of attendees at recent workshops think that unconscious bias training (UBT) decreases bias and discrimination in the workplace. However, the evidence base tells a different, murkier story.

Studies are yet to rigorously show that UBT and diversity training change biased behaviours in the workplace in any lasting way, or improve outcomes for women and Black and ethnic minority groups in terms of representation in leadership positions or reducing pay gaps.

Today in the government has published a written ministerial statement putting the record straight. Here’s the evidence behind it.

Unconscious Bias Training (UBT) — the evidence

Unconscious biases influence our judgement without us being aware of them. Unconscious bias training in the workplace aims to make people aware of potentially harmful unconscious biases and to reduce the impact of those biases on their interaction with others.

Some types of UBT may have limited positive effects such as creating awareness and shifting people’s attitudes in the short-term. Measures of awareness change are generally based on self-reported measures, however, which themselves may not be reliable.

In any case, awareness alone is not enough to make change happen, and there is no conclusive evidence that this training changes attitudes in any lasting way — with some studies finding that UBT does not change the gender stereotypes people endorse (such as agreeing to the statement ‘women are worse at maths than men’).

Evidence that this training changes behaviour or improves workplace equality remains elusive. This is compounded by few studies using valid measures of behaviour change in their design.

More worryingly, some studies have even identified potential back-firing effects when UBT participants are exposed to information that suggests stereotypes and biases are omnipresent and unchangeable, or when they react against mandatory training.

Diversity training — the evidence

Diversity training is designed to raise awareness of diversity issues in the workplace and to promote positive interactions between members of different groups. As with UBT it can help raise awareness but is unlikely to change behaviour.

Diversity training packages come in a wide range of formats and there is no standardisation of content. However a review of corporate diversity training from 1964–2008 found that:

• Diversity training often becomes a check-box exercise. Furthermore, the cost constraints of larger payrolls results in content being squeezed into shorter timeframes or facilitated by non-expert internal trainers. Such “tick-box” training is also often evaluated on the basis of the volume of staff trained, rather than the efficacy of the training itself.
• This training tends to leave participants without tools for behaviour change and can also generate backlash and activate stereotypes. The researchers noted: “Some of the unintended consequences were that many left confused, angry, or with more animosity toward differences. With no formal follow-up, employees were left on their own to interpret and internalize what they had learned. Many interpreted the key learning point as having to walk on eggshells around women and minorities — choosing words carefully so as not to offend. Some surmised that it meant White men were villains, still others assumed that they would lose their jobs to minorities and women, while others concluded that women and minorities were simply too sensitive.”

Why don’t UBT and diversity training change diversity behaviours?

In general, short-term educational interventions do not change people — especially where people have acquired biases over a lifetime of media exposure and repeated messages from their social environment. A few other hypotheses have been made as to why these training session fall short:

1. UBT can activate stereotypes, making them more likely to come to mind after the training has finished. This can happen when we are asked to try to suppress our own stereotypes, or when we are asked to confront them;
2. Training can make employees complacent about their own biases — meaning they do not take responsibility to reduce discrimination following training, perhaps because of a belief that the workplace is already free of bias;
3. Training may also have a ‘moral licensing’ effect, whereby an individual who attends training (which is ‘good’ for diversity) feels freer to go on to make a decision which does not improve diversity (e.g. hiring a candidate with a similar profile to their existing team) as their sense of virtue for having attended training counteracts ongoing efforts to monitor persistent prejudices.
4. People react negatively towards efforts to control their personally-held views and therefore may be particularly resistant if UBT is made mandatory, as this can make them feel disempowered;

We therefore do not recommend prioritising the use of resources on unconscious bias or diversity training over alternative interventions which have a more promising evidence base

What might make UBT and diversity training more effective?

While there is limited evidence, some ideas about how to improve training include

• Avoiding ‘one-off’ training sessions — training should be an ongoing process, involving multiple sessions and different formats;
• Making training voluntary — as evidence suggests that mandatory training can result in backfire. One study found that positive effects of in-depth training became more likely to persist once a minimum of 25% of team members participated in it;
• Integrating training with wider organisational initiatives that seek to debias processes themselves.

What to do next

In the absence of good evidence for the effectiveness of UBT and diversity training, we encourage organisations to:

Prioritise investment of time and resources into other initiatives which show much more promise and have better evidence of efficacy. Tried and tested actions with proven efficacy and growing research bases include:

• Using skill-based assessment tasks in recruitment
• Using structured interviews for recruitment and promotions
• Introducing transparency to promotion, pay and reward processes — including showing salary ranges to new hires
• Appointing diversity managers or diversity task forces to hold decision-makers to account and spread good practice
• Including multiple women and people from ethnic minority groups in shortlists for recruitment and promotions
• Making decisions about applicants in batches
• Anonymising CVs
• Removing biased language from job adverts

Avoid mandatory, ‘one-off,’ tick-the box training sessions which risk making the problem worse. Where more sophisticated training designs show promise, work with researchers to identify the sections/combinations that reliably ‘work’.

Demand that suppliers of UBT and diversity training work with you to capture meaningful outcomes following their training and measure whether or not the training ‘worked’ as desired, instead of asking participants to simply rate the training. For example, track whether hiring or promotion patterns change for the better, or bullying and harassment reports reduce in a lasting way. Work together over time to improve training until it does have positive behavioural outcomes.

By Shoshana Davidson, Senior Advisor & Hannah Burd, Principal Advisor, BIT