Psychometrics of the Chinese Mandarin Version of the Social Responsiveness Scale (SRS) in a Taiwanese Population of Children Aged 6-8 Years

Background: Autism Spectrum Disorders (ASDs) represent one end of a larger spectrum of quantitative impairment that is continuously distributed in thegeneral population.  The Social Responsiveness Scale (SRS) is an instrument that characterizes quantitative impairments in social awareness, cognition, communication, motivation, and repetitive behavior/restricted interests that define ASD, and provides a more subtle characterization of individual symptoms than using traditional classification systems.  Reliability and validity of the tool need to be established before its adaption in non-Western populations. 

Objectives: To report reliabilities and validities of the Chinese Mandarin Version of the SRS in a large population of school children aged 6-8-years, and in children who are at high risk for ASD. 

Methods: Caregiver-reported SRS data were collected by an epidemiologic autism study recently conducted in PingTung, Taiwan.  Raw scores of the total SRS and five subscales (social awareness, social cognition, social communication, social motivation, and autistic mannerisms) were used for this analysis.  As recommended in the literature, a raw score of >=70 in males and >=65 in females is a cut-point that provides evidence for the presence of an ASD.  Based on these cut-points, we defined the clinical (high risk for ASD) vs. non-clinical groups separately for males and females.  Reliability was assessed using alpha, while validity was examined by factor analyses.  Both reliability and validity analyses were carried out in the clinical group (high risk for ASD) and the full study sample (clinical and non-clinical combined). 

Results: Because SRS score is sex specific, only data from participants who completed the SRS and whose child’s sex is known are included in the analysis: 1384 males and 1507 females.  Of those, 357 children (172 males and 185 females) met the recommended clinical cut-off and are classified as the clinical group.  Reliability Alphas for SRS full scale, social awareness, social cognition, social communication, social motivation, and autistic mannerisms are 0.94, 0.54, 0.74, 0.85, 0.69, and 0.87, respectively, in full study sample (n=2891); and are 0.83, 0.10, 0.25, 0.64, 0.43, and 0.79, respectively, in the clinical group.  Results from exploratory factor analysis showed factors loaded differently between the clinical group and the full study sample. 

Conclusions: Reliabilities of the SRS full scale in both the full study sample and the clinical group are excellent and good.  However, reliabilities of the subscales are overall poor in the clinical group, even given the sufficient sample size (n=357).  Possible explanation on the different loaded factors for the full study sample and the clinical group will be further discussed.