Autistic children could learn through stereotypes

Autistic children have a capacity to understand other people through stereotypes, say scientists at UCL (University College London).

The research shows that autistic children are just as able as others to predict people's behaviour when stereotypes, such as gender and race, are the only available guide.

The psychologist who led the research, which is published today in the journal 'Current Biology', believes stereotypes could be used to help improve how autistic children relate to other people, by playing to their strength for understanding groups.

Professor Uta Frith of the UCL Institute of Cognitive Neuroscience said: "Autism affects around 500,000 families in the UK. Increasing an autistic child's capacity to understand other people is one of the keys to improving the lives of these families. One of the main problems experienced by autistic children is that they are unable to understand why others are doing certain things: what motivates them or what they are thinking and feeling. Most of us have this ability, known as 'Theory of Mind'.

"This research shows that although many autism sufferers do not have this in-built ability, they can still understand stereotypes very well. We hope that their ability to understand groups - even when they struggle with relating to individuals - will be used to aid their learning and socialisation."

49 primary school children (21 with autism and 28 without) were asked questions based on drawings representing males and females coloured in either pink or brown. The researchers asked questions such as: "Here are two children, David and Emma. One of them has four dolls. Which one has four dolls?" The answer Emma conforms to gender stereotypes, the answer David does not.

Each child completed 36 similar scenario-based questions. They then responded to scenarios where information about an individual's likes or dislikes conflicted with generic stereotypes. e.g. "Here are two people. This is James and this is Grace. Grace doesn't like to cook for people. One of these people has baked biscuits. Which person baked biscuits?"

Autistic children with Theory of Mind difficulties performed in the same way as normally developing children in the first task. 75 per cent of the answers children gave - whether they were autistic or not - were in line with commonly held race and gender stereotypes.

In the second task, either stereotypes or individual likes and dislikes could be used as the basis for an answer. Here, autistic children with Theory of Mind problems became confused. Older normally developing children and autistic children with some inkling of Theory of Mind tended to answer the questions based on an individual's likes and dislikes.

Professor Frith said: "Autistic children's knowledge of race and gender stereotypes is astonishing given that they lack interest in people."

She added: "Of course, stereotypes can be dangerous as they are the basis of prejudice. But we all use group-based knowledge in situations where we have to make quick decisions and don't know anything at all about the other person. We hope teachers and carers will consider using concepts about groups of people to help autistic children integrate better into society by playing to their strengths."

Notes for Editors

1. Journalists wishing to try out the test or wanting more information, should contact Professor Uta Frith or Dr Sarah White, UCL Institute of Cognitive Neuroscience, on +44 (0)20 7679 1163 or on +44 (0)20 7679 1168 or e-mail: u.frith@ucl.ac.uk or s.white@ucl.ac.uk

2. Alternatively, please contact Alex Brew in the UCL Media Relations Office on tel: +44 (0)20 7679 9726 or contact the out of hours press officer on +44 (0)7917 271 364

3. 'Can autistic children predict behaviour by social stereotypes?' is published in the journal 'Current Biology' on June 19^th 2007. This press release is embargoed to 5pm on 18^th June 2007. Journalists can obtain copies of the paper by contacting the UCL press office.

4. The study was carried out by Professor Uta Frith and Sarah White of the UCL Institute of Cognitive Neuroscience; Lawrence Hirschfeld of the New School for Social Research in New York; and Elizabeth Bartmess of the University of Michigan.

5. The study was supported by the Medical Research Council (MRC).