Racial prejudice has its roots in children’s natural drive to carve the world up into categories. Can research do anything to fix this?

Racist stereotypes, at their root, come from quite a fundamental learning mechanism. Humans are able to learn and adapt so quickly because they are excellent at making generalisations about the world based on very limited experience. Take dogs, for example – a toddler might reasonably conclude after meeting just two or three that all dogs are furry, bark and have tails that should be treated with some caution.

On the whole, stereotypes are often right – dogs do normally bark and wag their tails. The difficulty arises when this learning mechanism is applied to groups of people. Race is an easy mental category to fit people into because skin colour is a salient visual feature.

Babies are not born believing that any group is better than another but they do attend to race surprisingly early. From about 9-months, babies show a general preference for what is familiar: they are quicker to recognise faces and facial expressions of their own race than of other races.

If we don’t have the opportunity to interact with individuals of a different race then the information we have to inform a racial category has to come from other sources such as the media or people’s opinions. As these can be biased in positive or negative ways, the stereotypes we form can also be biased and inaccurate. Depending how insistent and consistent these secondary sources are, they might even overwhelm our own personal experience.

This effect is compounded by some other low-level, unconscious biases. There is a strong tendency to favour our own group over other groups. It doesn’t really matter how the group is specified: children remember more positive things about members of their in-group and more negative things about members of the out-group, even if group membership is specified by something as superficial and transient as t-shirt colour.

We (as a species) also have a tendency to think of members of the out-group as being all much the same while members of our in-group are all unique snowflakes. This enables us to create coherent categories and make predictions but can also lead to vastly inaccurate and damaging sweeping generalizations.

Young children are particularly sensitive to the use of generics in language to learn about the world as quickly as possible. If you say ‘birds have wings’ they will generalise this information to all expectations of birds in a way that they won’t if you say ‘this bird has wings’. Of course, the same is then true if they hear phrases like ‘Arabs are violent.’

Insidious Racism

So, it is an embarrassing and oft repeated finding that while the majority of people in Western countries these days are egalitarian believers in a fair meritocracy, on tests of unconscious racial bias about 70% show a preference for their own race. The classic test is the Implicit Association Test, which measures how quickly you are able to categorize photos of members of your own race with positive characteristics (wonderful, glorious) and members of a minority race with negative characteristics (horrible, nasty).

This conflict between people’s dearly held explicit beliefs and their nasty little unconscious racial biases is troubling and has real-world consequences. For example, presented with identical, moderately good resumes attached to a picture of a white or black candidate, interviewers are significantly more likely to shortlist the white candidate for interview. This study was originally conducted in 1989 but the results were exactly the same when it was repeated in 2005.

The Roots of Racism

Explicit (conscious) racial biases start at about 5-years of age but, where they are not supported, tend to peter out from about 10-12 years. This is likely because children become more aware of principles of fairness and social justice that shape how they believe people should be treated. (If racial stereotypes are supported by the people around them then all bets are off. On the whole, garbage in, garbage out.)

Implicit (unconscious) racial biases, however, can develop as young as 3 years of age. Once established in the preschool years they are surprisingly resilient to change. While explicit racial prejudice drops off in most children, implicit racial biases usually remain consistent through to adulthood.

Changing Unconscious Racism

I was particularly taken then with a paper in this month’s Developmental Science, which shows that a very simple intervention can disrupt young children’s unconscious racial biases. Xaio and colleagues at Zheijiang Normal University in China repeated a common measure of implicit racial bias: the ‘angry=outgroup’ test. Here photos of faces were morphed so that it was ambiguous whether they were Chinese or African. Each face was presented twice, once looking angry and once looking happy, and respondents asked to decide what race the face was.

As in previous tests, Chinese adults and children tended to say that the happy faces were Chinese and the angry faces were African. This is the same pattern as for white American children and adults who tend to say that happy faces are white and angry faces are black.

The researchers then introduced a very quick intervention. Four, 5- and 6-year-olds were asked to discriminate between 5 African faces and had to remember what number went with each face before they could proceed to the next step. This task forced children to focus on the individual differences between the faces.

When the angry=outgroup test was repeated, the bias had disappeared. Children were just as likely to say that the angry faces were Chinese as African. This simple intervention seems to have disrupted what was previously considered a very deep rooted and difficult to change bias.

The study raises a lot more questions than it answers. Why does it work? How long do the effects last for? How do changes in implicit biases interact with explicit beliefs and behaviour?

But I like it for two reasons. First, it gets to the root of the issue of racist generalisations by tinkering with simple perceptual categorization. If racial prejudice is just a value judgment laid on top of unconscious perceptual and grouping biases then this seems a sensible level to work at.

I also like its simplicity. Very similar effects have been shown with adults but used hundreds of repetitions during the intervention stage. Xiao’s intervention took no more than 15 minutes yet had significant short-term effects. Such a procedure could easily be adapted to a game or an app that, played regularly, might support longer-term change.

Being aware of implicit racial prejudice is important. We need to know it’s there to guard against it influencing our behaviour and we need to shape society to minimise its effects. For instance, racial information is now excluded from job applications and kept confidential so as not to influence decisions at the shortlisting stage.

But tackling implicit racial bias is important too. Vigilance can only take us so far when battling against unconscious demons. Would you like to see how you fare on the Implicit Association Test? Have a go here but don’t despair if, like 70% of the population, you show an unwanted preference for your own race. Being aware of these biases can make a difference and help may be just around the corner.