Can 'sleep training' make people less racist and sexist?

Monday June 1 2015

“Levels of unconscious racist and sexist bias have been reduced by manipulating the way the brain learns during sleep,” BBC News report.

This was study looking into inherent unconscious biases related to gender and race/ethnicity, and whether they could be reversed. Forty white university students did an “implicit association test”.

The exact format of this test is hard to describe briefly, but it generally looks at how quickly and accurately people can group certain words and concepts, and whether some groupings took them longer to get right (e.g. grouping female and scientific words). The test showed an inherent tendency to find it easier to link female and artistic rather than scientific words.

This is possibly due to the influence of an inherent cultural bias that women “don’t do science” (which is nonsense). The students also seemed to find it easier to link Black faces with “bad” (negative) words (like virus) rather than “good” words (like sunshine).

They were then given computer-based training to counter this bias, accompanied by a sound cue. After the training, they were tested again and their responses showed less bias.

When they then took a 90-minute nap with the sound cues replayed to them through headphones, they still showed a reduction in the bias on re-testing after they woke up. The effect was still demonstrated one week later.

This small study, while interesting, makes it difficult to draw any firm conclusions based on this research. We don’t know whether the test results represent real bias/discrimination in every day life, or further whether the sleep training would have any effect on real-life interactions.

Where did the story come from?

The study was carried out by researchers from the psychology departments of the University of Texas and Princeton University in the US, and was funded by the National Science Foundation and National Institutes of Health.

The study was published in the peer-reviewed journal Science on an open-access basis, so it is free to read online or download as a PDF.

BBC News gives a good account of this study, including the researchers' caution when drawing any implications from the sleep training and applying to real life: "We didn't have people interact with or make decisions about other people, so that sort of experiment is needed to know the full effects of the methods we used."

What kind of research was this?

This was an experimental study looking at ways to reduce biases related to gender and race/ethnicity.

These biases were described as being largely unconscious, where people are unaware of them, rather than people being actively racist or sexist.

The study aimed to look at whether these biases could be altered through training and then consolidated during sleep. Sleep is the time when memories are consolidated and newly acquired information is preserved.

Such experimental research can be useful for exploring theories related to sociology and psychology but is only an early stage of research, and much more would need to be done to assess whether the approach tested works in real-life situations.

What did the research involve?

The study included 40 men and women from a university, all of white ethnicity. First they tested their gender and race bias using an implicit association test (IAT). The test involves getting people to associate words with different faces shown to them on a screen. This test showed that participants had inherent/implicit gender and race biases. On one test images of female faces were more often associated with artistic rather than scientific words, and the opposite was found for male faces. On another test, images of black faces were more often associated with bad rather than good words, with the opposite found for white faces.

The researchers then gave them computer-based training to reduce bias. Participants viewed several types of face-word pairs but were required to respond only to pairings that countered the typical bias. They were played sounds for each pairing that went against the typical bias, as a form of "reinforcement" for the "counter-bias" messages. The students were given the IAT test again after the training to see what effect it had.

They then looked at whether taking a 90-minute nap while repeatedly playing the counter-bias sound cues had an effect when the tests were given to them again when they woke up.

What were the basic results?

The research found that the counter-bias training was effective in reducing participants' implicit gender and race bias in the IAT compared with that shown at the start of the study.

The sleep test showed that when they had a nap where the accompanying counter-bias sound cues were played to them, when they woke they still didn't show the implicit bias for both gender and race when re-tested. If sleep wasn't accompanied by the sound cues, however, their tests showed the same biases as they had at the start of the study.

Testing one week later showed that the reduction in bias in the IAT in those who had listened to the counter-bias sound cues during sleep was still there.

How did the researchers interpret the results?

The researchers conclude that "memory reactivation during sleep enhances counter-stereotype training and maintaining a bias reduction is sleep-dependent".


This small experimental study suggests that training involving sound cues to counter gender and race biases – and then consolidating this by replaying the sounds during sleep – could have an effect in changing these biases. The use of sound cues to alter behaviour has proved effective in the past, most famously in the animal experiments of Ivan Pavlov.

However, any such interpretations should be made very cautiously at this stage, as this experiment may not be representative of real-life situations.

This was a small study involving 40 university students, all of white ethnicity. Both their biases, and the effects of the training, may not be applicable to different populations. The results may also have been different had a much larger sample of a similar white university population group been tested.

This was a very specific image-word association test, with a specific sound-cue-sleep training to try and counter the bias. Though this test may have shown reduction in bias on the test at up to one week later, there are many unknowns. We don’t know how long the effect would last, or whether training would have to be continually reinforced, for example.

We also don't know whether the biases apparent on this testing actually related to any discriminatory attitudes or behaviour in real-life situations. Furthermore, we don't know whether the effect of training would translate to making a difference in real-life situations. In real-life situations, perceptions and behaviour are likely to be influenced by many factors.

Overall this research will be of interest to the fields of psychology and sociology into biases and possible ways to reverse them.

The researchers speculate that the technique could be used to change other “unhealthy habits” such as smoking, unhealthy eating, negative thinking and selfishness.

There may be ethical issues with such techniques. For example, the definition of an “unhealthy” habit or viewpoint may not always be universally agreed, and some people may prefer to have the freewill to think the way they do or make their own choices without this being manipulated.

Such speculations will remain so until the researchers can demonstrate that their sleep training technique has a measurable “real world” effect that persists over time.

Analysis by Bazian
Edited by NHS Choices