Ellie Highwood

Largely a summary of….

Unconscious Bias training: An assessment of the evidence for effectiveness , Equality and Human Rights Commission, Research Report 113, by Doyin Atewologun, Tinu Cornish and Fatima Tresh

Premise and Reading context:

Unconscious Bias training is frequently cited as a solution to reducing bias with respect to protected characteristics under the Equality Act 2010 in selection processes. Indeed no self-respecting Athena SWAN application would be without it. At Reading some form of unconscious bias training is mandatory for chairs of interview panels across the University and some Schools have trained larger teams as part of their Athena SWAN bids. Currently our unconscious bias training is delivered in a number of ways:

• Embedded within face-to-face recruitment and selection panel and chair training
• Specific online Unconscious bias module
• Some historic bespoke training at school or function level.
• PGR student training developed from undergraduate training work in the School of Mathematical, Physical and Computational Science
• Some coverage in modules for trainee teachers within Institute of Education

We are in the process of evaluating and updating our approach and delivery of unconscious bias training as one of the Institutional Athena SWAN actions. We have several academics with experience of designing and delivering Unconscious Bias Training, however this recent Equality and Human Rights Commission report provides a broad evaluation, highlighting where evidence exists for the effectiveness of this type of training. Here I summarise points from that report which reviewed many published articles and grey literature annual reports of studies into the effectiveness of unconscious bias training (UBT). The studies used varied in terms of robustness.

What can UBT do?

• Can be effective for awareness raising.
• Can reduce implicit bias but is unlikely to eliminate it. Most UBT is not designed to reduce explicit bias.
• The evidence for UBT being effective in changing behaviour is limited – but most of these studies did not use valid measures of behaviour change.
• More successful in reducing implicit bias relating to gender, than race and ethnicity.

What does the most effective UBT look like?

• Uses an IAT (Implicit Association Test), followed by a debrief, incorporating theory about unconscious bias rather than detail about impact.
• The most successful interventions include bias reduction strategies and bias mitigation strategies so that participants feel empowered to do something.
• There appears little difference in effectiveness between on-line and face-to-face training.
• However, there is evidence that increasing the sophistication of the UBT (e.g. an interactive workshop) can increase awareness and concern about wider discrimination and that this awareness continues to increase over time.
• The report emphasises that UBT should be only one part of a programme designed to achieve organisational change.

Who should be trained?

Training teams together resulted in positive group behaviour change despite the evidence for effectiveness in changing individual’s behaviour being weak. However, there is too little UBT specific research to judge whether mandatory or voluntary training has a different effectiveness.

What can go wrong? If UBT participants are exposed to information that suggests stereotypes and biases are unchangeable, this can back-fire and result in more entrenched bias.

Considering Reading’s approach in the light of this review

Our current online offering clearly states that it is intending to raise awareness.

• It includes a heavier emphasis on theory about unconscious bias compared to statistics about impact.
• It covers many types of diversity.
• Although an IAT is not used in this online course directly, they are explained and a link is provided as follow on work.

What we may be able to improve –

• More training of teams together – to result in effective group behaviour change.

• Make sure a resource that covers a debrief after the IAT test is available if people take that up as part of the online course

• There is a reference in the start of the training to implicit bias being “hard-wired” and it is a fine line between normalising implicit bias to encourage reflection, and making it “ok” to be biased.

• Provide some specific bias reduction or mitigation strategies

• Better evaluation of implicit bias reduction and behaviour change, if we provided follow up resources for these.

Finally – there is a need for more research in this area, specifically UK based (as many studies currently US focused and the race issues in particular can be quite different between the US and the UK).