H567/02 Section A - Baron-Cohen et al. (1997): autism in adults
Baron-Cohen and colleagues developed an adult test of inferring mental states from eyes. Learn why the comparison and control tasks matter, what the results show, and where the conclusions stop.
Why test advanced theory of mind?
Theory of mind is the ability to attribute mental states, such as beliefs, intentions or feelings, to oneself and others and use them to understand behaviour.
An adult might pass a simple false-belief test designed for young children while still finding subtle social interpretation difficult. This is a ceiling effect: a task is too easy to distinguish higher levels of ability. Building on Happé's Strange Stories work, Baron-Cohen et al. aimed to test advanced mental-state inference in adults previously described as having high-functioning autism or Asperger syndrome.
The researchers predicted that the autism/AS group would perform less well than both an adult comparison group and a clinical comparison group with Tourette syndrome. They also examined male–female differences within the non-autistic comparison group. The historical labels describe the 1997 sample; they should not be treated as a ranking of people's worth or a complete account of current autism classification.
The study relates to understanding disorders by testing whether a specific social-cognitive difficulty can coexist with intelligence in the typical or higher range. It belongs to individual differences because it compares patterns of psychological performance between groups. Inferring a mental state from a photograph is a narrow ability: it does not measure every aspect of empathy, and lower performance does not establish absence of care for others.
The groups and quasi-experimental design
The study compared pre-existing groups using an independent measures quasi-experiment; participants were not randomly assigned a diagnosis.
The grouping variable was autism/AS, no reported psychiatric condition, or Tourette syndrome. The principal dependent measure was the number of correct responses on the Eyes Task, out of 25. A controlled task does not make the grouping variable experimentally manipulated, so background differences remain possible explanations.
| Group | Number and sex composition | Recruitment |
|---|---|---|
| Autism/AS | 16: four with high-functioning autism and 12 with AS; 13 male, three female | Clinical sources and an advertisement in the National Autistic Society magazine |
| Non-autistic comparison | 50: 25 male, 25 female | A Cambridge community participant panel, excluding university members |
| Tourette syndrome | Ten: eight male, two female | A tertiary referral centre in London |
Groups were age-matched. Clinical participants had intelligence in the typical range and passed first- and second-order false-belief tasks. The autism/AS participants had verbal and performance IQ above 85. IQ was not measured in the 50-person comparison group; it was assumed to be in the typical range. Therefore avoid saying that every person's measured IQ was matched.
Tourette syndrome provided a clinical control: another childhood-onset condition involving possible disruption of schooling and relationships. If that group performed well, an Eyes Task difficulty could not simply be attributed to having any developmental clinical condition. This comparison reduces one alternative explanation, but does not remove every difference between the groups. Small clinical groups, voluntary/clinical recruitment and their male-heavy composition limit representativeness.
The Eyes Task and checks on its validity
The 1997 Eyes Task used 25 black-and-white photographs showing only the eye region of male and female faces, with a forced choice between two mental-state words for each photograph.
Photographs were standardised to the same size and region. Each was shown for three seconds. Participants selected the word that best described what the person was thinking or feeling, with one mark per correct choice. Targets and opposite-meaning foils were developed using judges, with agreement checked by independent judges. Standard stimuli and scoring support repeatability, although agreement on a label does not guarantee access to the photographed person's actual mind. Do not confuse this original 25-item, two-option task with the later revised test.
Tasks were presented in random order, and people were tested individually in a quiet room at home, a clinic or the Cambridge laboratory. Random order reduces systematic practice or fatigue effects across tasks; it does not mean diagnoses were randomly allocated.
The gender-recognition control used the same eye photographs but asked for male/female classification, checking whether a low Eyes score might instead reflect difficulty extracting non-mental-state information from that region. The basic-emotion control used six whole-face photographs displaying emotions such as happiness and fear, checking basic emotion recognition. The primary paper describes these as controls administered to the autism/AS group; reported control comparisons showed no relevant deficit. The clinical groups' Strange Stories performance, from linked work, provided another check: a similar difficulty on a different advanced mental-state task strengthens the interpretation.
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account