The myth of unconscious bias training (and what you should do instead)

Ruby Dark

Apr 20, 2022

More and more businesses are waking up to the importance of embracing a diversity of people, perspectives and experiences. Leaders are looking for ways to welcome more diversity to their organisations, which means investing in D&I solutions. And unconscious bias training is a popular route to go down. But does it really give you the results it boasts?

Unconscious bias training has become the go-to D&I solution, with 81% of UK companies running it. By conducting just a 1-2 hour session, you could rid your team of bias, make better decisions, and, eventually, welcome more diversity to your organisation. With the benefits that unconscious bias training boasts, why wouldn’t you shoot the silver bullet?

With all that time and money invested in unconscious bias training, you’d expect some hefty results right? But we’re yet to see a fairer representation of minority groups in tech. 91% of C-suite tech leaders in the UK are men, and only 2.6% of UK tech board positions are filled by ethnic minorities. We have a long way to go to reach equal representation, especially at the leadership level.

The thing is, there has been no conclusive research that unconscious bias training changes people’s behaviour in the long run. At best, it might open your eyes to the perils of bias. At worst, it backfires, encouraging people to lean into their biases and stereotypes — “what’s the point in fighting something we all naturally do, anyway?”.

So what can we do instead? Are we really doomed to rely on our mental shortcuts? Not at all. Although we all have biases, we can keep them from swaying our judgements and decisions. Here’s how we can do better.

What is unconscious bias?

Our brains are incredibly efficient. Every second, we’re faced with 11 million pieces of information to sort through, which we can’t possibly consider in a logical, rigorous way.

Think about where you’re sitting right now — are you aware of how comfy your chair is? What about the lingering smell of coffee or lunch? Can you hear cars passing on the street outside, or the hum of a boiler? Maybe you’re aware of these things now, but if you were constantly thinking about everything happening all around you, you’d never get anything done.

To save us from being totally overwhelmed with all that information, our brain uses shortcuts to process things, fast. Most of these shortcuts are totally harmless, but when our mental shortcuts systematically influence our judgements and decisions, we call this ‘bias’. Since it often happens entirely outside our consciousness, we call it ‘unconscious bias‘.

For example, take a look at the two people below. Without thinking too hard about the question, who do you think is more likely to work in finance? Who is more likely to work in a preschool? Many things could have influenced your judgement — their clothes, their haircut, their gender, their ethnicity.

The illustration on the left shows a Black woman with a blonde ponytail, wearing a denim shirt, yellow trousers and white sneakers. The illustration on the right shows a White man with short black hair, wearing a blue buttoned shirt, black trousers and white shoes.

We intuitively look for similarity — what do most finance professionals look like? This tendency of ours to sort people into neat categories keeps the status quo in place. It becomes a problem when you’re hiring for that finance or preschool position, and let those split-second judgements influence your opinion of who is the best fit for the role.

Often, our decisions are implicitly influenced by the stereotypes and assumptions we hold about different social groups. Bias shows up in many ways at work, influencing who gets hired or promoted. It also impacts our day-to-day decisions, like who gets asked to organise the birthday card or who’s assigned to lead a project. To build fair workplaces, we need to limit the impact biases have on decisions.

What’s wrong with unconscious bias training?

Unconscious bias training aims to overcome bias by making us more aware of it. The underlying idea is that if only people are aware of their privileges, assumptions and stereotypes, they’d change their behaviour.

There’s nothing wrong with raising awareness about stereotypes and bias, but the way it’s handled leaves a lot to be desired. A lot of companies still lead with the compliance argument for D&I: “discriminate and we’ll get a lawsuit”, or “you need to go through this training to qualify for promotion.”

And often, this training happens in mandatory online sessions with very limited opportunity for interaction and different learning formats. This isn’t a great set-up for learning — to internalise new information, we need to engage with the topic actively and continuously.

In fact, when handled badly, unconscious bias training can be actively harmful. It can reinforce stereotypes or make people feel like they’re off the hook: ‘I’ve done the training so now I’m immune to bias!’.

The evidence against unconscious bias training is mounting. One recent meta-analysis of over 490 studies found that whilst training might raise awareness in a couple of weeks following, it did not lead to long-lasting behavioural change.


In fact, studies have shown that training can even backfire. One analysis spanning 30 years and over 700 organisations found that when companies implemented unconscious bias training, it became less likely that Black men and women would advance over time. Researchers suggest that this is because diversity training often results in backlash from participants, making them less open to D&I.

How to overcome bias with behavioural design

Unconscious bias training might not be a very hopeful solution to overcome our biases, but that doesn’t mean we’re doomed to fall prey to our mental shortcuts, snap judgements, and automatic associations. We just need to build the right structures to make fairer decisions.

We can create those fairer environments with behavioural design. Here are some examples:

  • Asking the same interview questions in the same order.

    Unstructured interviews are really bad at predicting job performance. When you don’t prepare a set of questions to ask, interviewers can stray off-topic and ask subjective questions.

    When it’s time to review the candidates, you end up comparing apples and pears. This means assessors use their personal preferences to distinguish between candidates.

    When you ask all candidates the same questions in the same order, they all have equal opportunities to prove themselves. And you have clearer information to compare candidates.

  • Setting a minimum and maximum word count for performance reviews.

    Men are 20% more likely to receive developmental feedback than women. This means that men are offered more guidance whilst women are left guessing, which has knock-on effects on career progress.

    One light-touch nudge towards fairer feedback is to set a minimum and maximum word count for performance reviews. It tells managers that they’re expected to give all reports a similar level of detail. Accompany that with a checklist for writing high-quality feedback, and you’ll make sure all employees are set on the right track.

  • Comparing promotion candidates side by side.

    Our brains are naturally comparative. We can’t look at a tree and say “that’s exactly 52 feet tall”, but we can tell which is the tallest tree in the park. If we don’t have a point of comparison, we’re much more likely to rely on (inaccurate) assumptions, or stereotypes.

    We can apply this to D&I: instead of assessing people in isolation, make promotion (or hiring) decisions in bulk. Next time a promotion opportunity opens up, compare employees’ latest performance reviews side-by-side. You’re much more likely to focus on objective skills.

Isolation: one tree standing on its own. Comparison: a group of trees together.

At Fair HQ, everything we do is built on behavioural science.

We don’t try to ‘fix’ differences, we focus on changing the status quo. We show you how to make better, fairer decisions based on objective metrics, not personal preference. The Fair HQ platform helps you focus on creating equitable foundations first. It’s about changing structures, not people.

Ready to cut through the noise and break the bias?

Book a demo

P.S. Does unconscious bias training ever work?

With the right preparation, content and follow up, it is possible for unconscious bias training to shift behaviours in the long term. To do so, training needs to go beyond raising awareness of bias and give participants the strategies to overcome it:

  1. Learning needs to be continuous, not a one-off session;
  2. Participants need to be taught how to overcome bias, for example, with perspective taking and process changes;
  3. The training should provide opportunities to hear from others and interact empathetically with different groups of people;
  4. Crucially, training needs to sit within a broader D&I strategy, accompanying all of your business activities.

Training with these components has been shown to yield results. Two years after participating in the training, employees were more likely to speak out against bias than others who hadn’t taken part.

Sadly, most training misses the mark here. Alone, unconscious bias training won’t change much. But if you build the structures to overcome bias in your organisation, then the right kind of training can complement it well.

Backing it up

Anand, R., & Winters, M. F. (2008). A retrospective view of corporate diversity training from 1964 to the present. Academy of Management Learning & Education

Bohnet, I. (2016). How to take the bias out of interviews. Harvard Business Review

Bohnet, I., Van Geen, A., & Bazerman, M. (2016). When performance trumps gender bias: Joint vs. separate evaluation. Management Science

Correll, S. J. & Simard, C. (2016). Research: Vague Feedback Is Holding Women Back. ********Harvard Business Review

Chang, E. H., Kirgios, E. L., Rai, A., & Milkman, K. L. (2020). The isolated choice effect and its implications for gender diversity in organizations. Management Science

Dobbin, F., & Kalev, A. (2016). Why diversity programs fail. Harvard Business Review

Forscher, P. S., Lai, C. K., Axt, J. R., Ebersole, C. R., Herman, M., Devine, P. G., & Nosek, B. A. (2019). A meta-analysis of procedures to change implicit measures. Journal of personality and social psychology

Gino, F. & Coffman, K. (2021). Unconscious Bias Training That Works. Harvard Business Review

Kalev, A., Dobbin, F., & Kelly, E. (2006). Best practices or best guesses? Assessing the efficacy of corporate affirmative action and diversity policies. American sociological review