Baron-Cohen et al. (1997)


Autism is a severe developmental disorder which affects the social functioning of individuals. People with Autism share a number of problems in common, known as the “triad of impairments”.

Brookdale Triad Diag Updated

The triad of impairments consists of problems with:

  • Social interaction, particularly in forming relationships and lack of eye contact- which can make people with Autism seem aloof or indifferent to others
  • Social communication – difficulties with both verbal and non-verbal language, for instance not appreciating the meaning of gestures, facial expressions and tone of voice.
  • Limited imagination & Flexibility – Autistic individuals often have rigid and flexible routines which must be followed strictly, leading to difficulties in imaginative play

Autism runs on a spectrum from Asperger syndrome at one end through to individuals showing severe forms of autism at the other extreme. People with Asperger syndrome do not usually have the accompanying learning disabilities associated with autism; in fact, people with Asperger syndrome are often of average or above average intelligence.

Baron-Cohen claimed that the numerous deficits that autistics suffer from can be explained by focusing on the way that autistics process information. He proposed that autistics lack a theory of mind– i.e. the ability to recognize and appreciate the mental states (thoughts, desires, fears etc) of other people. He also refers to this as mind blindness.

Baron-Cohen argued that impairments in the development of a theory of mind may underlie the social, communicative, and imaginative impairments of people with autism since a theory of mind is necessary for normal development in each of these three areas.

Previous research using first-order tests of theory of mind have demonstrated that children with autism can not employ a theory of mind. An example of a first-order test is the Sally Anne test.

In this test the child is presented with two dolls (Sally and Anne), a marble, a box and a basket. Sally puts her marble in her basket and leaves the room. Anne then moves the marble from the basket to her box. Sally returns and the child is asked ‘where will Sally look for the marble?’.

Baron Cohen et al. found that normal 4 year old children could correctly state that Sally would look in her own basket whereas children with autism found this first-order belief task difficult (they would point to Anne’s box) which suggests that children with autism cannot employ a theory of mind.

Happe (1994) developed a more advanced theory of mind test called the Strange Stories task designed for the level of a normal 8-9 year old. This task involved story comprehension, where the key question in the task either concerned a character’s mental states (the experimental condition) or physical events (the control condition). It was found that both adults with autism or Asperger syndrome had more difficulty with the mental state task than normal control participants.


However, Baron-Cohen argued that the Sally-Anne and Strange Stories tasks could not be used to demonstrate that adults with autism have an intact theory of mind because such tests have a ceiling effect. This is because children with normal intelligence can pass such tests at about 6 years of age. In other words, these tests only measure as high as the skills of a normal 6 year old child.

Because the Sally-Anne and Strange Stories tasks were designed for children, Baron Cohen et al. developed a new test called the ‘Reading the mind in the Eyes Task’. The task involved inferring the mental state of a person just from the information in photographs of a person’s eyes.

This advanced test aims to discover if high functioning adults with autism and Asperger syndrome do have problems with mind reading which it is argued is related to the ability to employ a theory of mind.


The main aim of this experiment was to investigate if high functioning adults with autism or Asperger syndrome would be impaired on a theory of mind test called the ‘Reading the Mind in the Eyes Task’.

The researchers were also interested to find out if females would be better than males on the ‘Reading the Mind in the Eyes Task’

Method & Design

Quasi-experiment with an independent measures design.

Independent variable: Normal, Autistic, Tourette’s syndrome

Dependent variable: performance on eye task


Group 1: 16 participants with high functioning autism or Asperger syndrome, 13 males, 3 females, mean IQ 105.3, recruited using an advert in the National Autistic Society magazine as well as through clinics.

Group 2: 50 controls, 25 males, 25 females, no history of psychiatric disorder and presumed to be of normal intelligence, selected randomly from the subject panel held in the University Department.

Group 3: 10 Tourette’s patients, 8 males, 2 females, mean IQ 103.5, recruited from a referral centre in London.

All participants were matched on age & were of normal intelligence

The researchers note that the reason for using participants with Tourette syndrome was because of the similarities between autism and Tourette syndrome. For example, they are both developmental disorders who affect males more than females, these disorders disrupt social functioning, they all have a significant genetic basis and all have been a associated with abnormalities in the frontal region of the brain.


Eyes task: participants were shown 25 photos of eyes (both male and female), each 15 x 10cm in size and black and white. Each photo was shown for 3 seconds. Participants then had to answer a forced choice question‘which word best describes what this person is feeling or thinking?’ where they had to select one of two mental states. The target term was the correct answer whereas the foil was the opposite. As a control the used words were randomised on both left and right.

The ‘eye task’ was created by selecting magazine photos, and 4 judges generated the target words e.g. TARGET= calm, FOIL = anxious.

The full set of mental state terms (and their foils) is shown below. Note that the mental state terms include both basic and complex mental states.

To check whether deficits on the Eyes Task were due to other factors, the researchers administered two control tasks to the participants in Group 1 (autism or Asperger syndrome) – the gender recognition task and the basic emotion recognition task.

The Gender Recognition Task involved looking at the same sets of eyes in the experimental task, but this time identifying the gender of the person in each photograph. This is a social judgement without involving mind reading, and allowed the researchers to check if any deficits on the Eyes Task could be attributed to general deficits in face perception, perceptual discrimination, or social perception. This also had a maximum score of 25.

The Basic Emotion Recognition Task (Emotion Task) involved judging photographs of full faces displaying the basic emotions. This was designed to check whether any deficits on the Eyes Task could be attributed to a deficit in basic emotion expression recognition. Six faces were used, testing the following basic emotions: happy, sad, angry, afraid, disgusted, and surprised.

In order to check the validity of the Eyes Task as a theory of mind task, participants in the two clinical groups (autism and Tourette syndrome) were also tested on Happe’s Strange Stories. This is known as concurrent validity, and is based on the assumption that two measures of the same thing should produce similar results. The Eyes Task, the Strange Stories and two control tasks were presented in random order to all subjects; this was to avoid order effects.

Participants were tested individually in quiet room in their own home, in the researchers’ clinic or in the laboratory at the university.

Data collection

Quantitative: number of correct responses out of 25 on the Eyes Task (also, number of correct responses in the control tests: gender of eyes and basic emotion test)

Qualitative: participants were asked to give reasons for how characters behaved in Happe’s Strange Stories Task


The Autistic adults were less likely to identify the Target foil than the Normal or Tourettes group.

Condition Mean score on the Eye Task
Adults with autism or Asperger syndrome 16.3
‘Normal’ adults 20.3
Adults with Tourette syndrome 20.4

Normal females were better at reading minds from eyes than normal males.

Condition Mean score on the Eye Task
‘Normal’ males 18.8
‘Normal’ females 21.8

On the Strange Stories Task none of the participants with Tourette syndrome made any mistakes whereas many of those participants with autism and Asperger syndrome had difficulties with this task.

On the Gender and Emotion Control Tasks, there were no differences between the groups.


Baron-Cohen concluded that the results of this study provide evidence for subtle ‘mindreading’ deficits in intelligent adults on the Autistic spectrum. Therefore the core deficit involved in autism is the lack of an advanced theory of mind.

The eye task is a ‘pure theory of mind test’ for adults because there is NO context (does not require an understanding of what the person whose eyes are shown is ‘doing’.)

Females have more advanced theory of mind skills than males, supporting the idea that the male brain predisposes some individuals to developing autism



+ Experiment has high control over variables such as intelligence, sex and developmental disorders. Therefore the findings are high in validity as the researchers ensured that the independent variable which is the characteristics of autism was causing the dependent variable that is performance on the Eye Task.

+ Experiment is high in reliability as the procedure was standardised in the way that every participant was tested in the same way, for example the way the photographs were presented.

+ Quasi-experiments tend to be high in ecological validity as the IV is naturally occurring therefore findings are more true to real life.

+ Collection of both quantitative and qualitative data allows for statistical comparisons between participant scores as well as rich insight into their decision making processes thus providing a more holistic understanding of the phenomena being studied.

+ The Eyes task is high in validity because the target terms also consisted of cognitive mental states, not simply emotions. This is therefore more than just an emotion perception test. Also, performance on the Eyes task was equivalent to performance on the Happe’s Strange Stories Task (so there was concurrent validity). Contrastingly, performance on the Eyes task was not equivalent to that on the control tasks suggesting that the poor performance by participants with autism was not due to a deficit in extracting social information from minimal cues.


– The ‘Eyes task’ lacks ecological validity and thus findings may not generalise to real life situations. In real life we have to interpret emotions from full faces, people’s body language, in real social situations as opposed to examining static images of eyes. Also, the artificial lab setting may have affected participant’s performance.

– The Eyes Task may lack validity. Because participant’s had only one of two mental state options to choose from and only 3 seconds to examine a static image of eyes and make their decisions in a lab context, the Eye Task may not be actually measuring theory of mind.

– The quantitative may have lacked reliability because there were different numbers of people in each group therefore the mean can become unreliable.

– Quasi-experiments tend to have less control over extraneous variables, for example any potential individual differences in the severity of diagnosed conditions, and therefore we cannot clearly establish cause and effect because the IV is not directly controlled.

– The sample is unrepresentative of the target population because the people with Autism had no other disorder or mental problems – most people with Autism have other problems as well as the Autism. Also the sample size was small which may not have masked the effect of individual differences meaning the findings may lack generalisability.

Click here for model answers and exam practice

Useful resources:

Section A Questions

Section B Questions

Multiple Choice Quiz

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>