International Meeting for Autism Research (London, May 15-17, 2008): Reading the Mind in the Eyes: test-retest reliability of a Swedish version

Reading the Mind in the Eyes: test-retest reliability of a Swedish version

Thursday, May 15, 2008
Champagne Terrace/Bordeaux (Novotel London West)
M. Hallerbäck , Department of Child and Adolescent psychiatry, County Council of Värmland, Karlstad, Sweden
T. Lugnegård , Department of Adult Habilitation, County Council of Värmland, Karlstad, Sweden
F. Hjärthag , Department of Psychology, Karlstad University, Karlstad, Sweden
C. Gillberg , Department of Child and Adolescent Psychiatry, Göteborg University, Göteborg, Sweden
Background: The “Reading the Mind in the Eyes Test” revised version (Eyes Test), was developed by Simon Baron-Cohen and his group. It is considered to be an advanced theory of mind test, sensitive to “mind reading” ability in adults with normal intelligence. The Eyes Test has been used in studying theory of mind difficulties in autism spectrum disorders. Although the Eyes Test has been translated to several languages and is widely spread, no study of test-retest reliability has yet been published.

Objectives:

Examine test-retest reliability of a Swedish version of the Eyes Test.

Methods: A Swedish translation of the child version of the Eyes Test consisting of 24 items was used. Participants were 58 students (33 female and 25 male) at the University of Karlstad. The test was given twice at an interval of three weeks.  To examine the test-retest reliability we used the Bland Altman method to measure the limits of agreement. This is a statistical method designed to examine the agreement between two test methods or the repeatability of a method. Using this statistical method it is possible to both visualize and calculate the agreement between two test times.

Results: The limits of agreement was ± 4.3

Conclusions: When the Eyes Test is used, either in research or clinical practice, one has to take into account that an obtained test score must be regarded as an approximation. A test score variation in the range of ±4 (out of 24 possible) is to be expected for the same individual. However, there was no indication of learning effects when the test was repeated.

See more of: Cognition Posters 1
See more of: Poster Presentations