Aim:
- To test a group of adults with AS or HFA on the revised scale of the eyes test. This was to check whether the same deficits seen in the original study could be replicated.
- To test a sample of normal adults to see whether there was a negative correlation between the scores on the eyes test and their autism spectrum quotient.
- To test whether females scored better on the eyes test than males.
Method:
Design:
| Problems with original test | New design element |
| Forced between two response choices. Narrow range of scores. | Four response options, 36 pairs of eyes rather than 25. Individual differences can be examined better. |
| Basic and complex mental states – some too easy, some too hard | Only complex mental states. |
| Some eyes were easily matched because of eye direction (ignoring or noticing) | These pairs were deleted |
| More female’s eyes used | An equal amount of male and female pairs of eyes. |
| Choice of two responses were always ‘semantic opposites’ , made it easy | ‘Semantic opposites’ were removed and the ‘foil choices’ (wrong ones) were more similar to correct answers. |
| May have been comprehension problems with the choice of words used | A glossary of all terms was given. |
‘Correct’ and ‘foil’ words were chosen by the authors of the study and piloted on eight judges of equal sex – for the words to be used, five judges had to agree.
Data from group 2 and 3 were merged – did not differ in performance.
50% of the group had to get correct word, no more than 25% to pick the foil, in order to be included in new eye test.
36/40 eye pairs passed.
Independent variable: the factor of autism
Dependent variable: self-report scores
Quasi (natural) experiment with self-report.
Participants:
Four groups:
- 15 males formally diagnosed with either AS or HFA – volunteers through UK national autistic society magazines or support groups.
- 122 normal adults recruited from adult community and education classes in Exeter or public library in Cambridge.
- 103 normal adults (53 males, 50 females), undergraduates at Cambridge university (71 in sciences, 32 in others) – high IQ.
- 14 random adults – matched IQ with 1st group.
Procedure:
Each participant completed the eye test – individually in a quiet room.
Group 1 – judge the gender of each image.
Group 1,3,4 – questionnaire to measure their AQ.
All to read through glossary and ask if they were unsure – could revisit it at any time.
Results:
| Group | Eye test | AQ |
| AS/HFA adults | 21.9 | 34.4 |
| General Population | 26.2 | N/A |
| Students | 28.0 | 18.3 |
| Matched | 30.9 | 18.9 |
- The AS/HFA groups scored significantly worse than the other groups in the eye test.
- Females usually scored better on eye test rather than males.
- AS/HFA scored higher on the AQ.
- Correlation between AQ and eye test was negative.
Conclusion:
The revised version of the eye test could still discriminate between AS/HFA adults and controls from different sections of society as the findings were the same as before.
The new eye test seemed to overcome initial problems of the original version – it validates as a useful test to identify impairments in social intelligence in normally intelligent adults.
Strengths:
- Quantitative data – the data was quantitative, making more reliable as it could not be affected by subjectivity of researchers.
- Replicable and reliable
- High control
- Useful to humans
Weaknesses:
- Internal Validity
- Ecological validity and mundane realism
- Extraneous variables
CIE A Level Psychology 