An Evaluation of the Psychometric Properties of the Social Responsiveness Scale in Two Groups of Children in Iceland

Background: : In the context of a study of the genetics of autism a screening instrument was needed to assess symptoms in the relatives of probands.

Objectives: The reliability and construct validity of the Social Responsiveness Scale (SRS) was evaluated in two groups of Icelandic children.

Methods: The SRS, an instrument designed to assess autistic traits, consists of 65 items scored on a four-point Likert scale. The questionnaire was completed by the parents of 606 children aged 4-15 years, relatives of probands with autism and the parents of a random general population sample of 248 children aged 6-15 years. Median scores were compared by age and gender. In the group of relatives the familial aggregation of autistic traits was explored by considering both SRS high-scorers and individuals with autism as affected. Summary statistics of the relatives were compared with those of the general population sample. The relationships between SRS scores and scores on the Strengths and Difficulties Questionnaire (SDQ) were examined.

Results: Internal consistency measured by Cronbach´s alpha was .93-.96 according to group and gender. Median scores for males were higher than for females and the differences in scores between the group of relatives and the general population sample was in the expected direction. Autistic traits were elevated in closer relatives compared with more distant ones. An examination of the relationships of SRS total scores with subscale scores of the SDQ showed  positive correalations with the Peer problems subscale and negative correlations with the Prosocial behaviour scale.

Conclusions: The results lend support to the reliability and validity of the Icelandic version of the SRS.