The study for the validity and the reliability of the childhood autism rating scale- Turkish version

Objective: There are many diagnostic instruments used in the assessment of the pervasive developmental disorders. The Childhood Autism Rating Scale (CARS) is a valid and reliable scale that consists of 15 items which is used to diagnose autism and to differentiate children that have developmental handicaps who do not happen to have autistic syndrome from the ones that have autism. It is commonly used in the diagnosis and the screening of pervasive developmental disorders. In our country, by making the translation and the retranslation of the scale to form the Turkish version, Sucuoglu and colleagues have evaluated the validity and the reliability of the scale using internal consistency (?) and item analysis (?) methods. The aim of this study is to widen the validity and the reliability analysis of The Childhood Autism Rating Scale of which its internal consistency, content validity (?) and discriminative quality for the end groups in the sample were already assessed by Sucuoglu and colleagues. As for this reason, the reliability, test- retest consistency, the comparison of the scale with Autism Behavioural Checklist ( ABC) and Clinical Global Impression ( CGI) is carried out between the evaluators in a much larger sample size.Method: The study is carried out with a sample that consists of 96 children and adolescents in total aged between 4-18 yrs old, of which 48 diagnosed with pervasive developmental disorder(PDD) and 48 diagnosed with mental retardation (MR) according to the DSM-IV TR criteria.While the sociodemographic data form, The Childhood Autism Rating Scale, Autism Behavioral Checklist (ABC) and Clinical Global Impression (CGI) were applied to the PDD cases, the sociodemographic data form, The Childhood Autism Rating Scale and Autism Behavioral Checklist (ABC) were administered to the control group. Evaluations were performed by two independent clinicians using the video records of the same interview. As for the test-retest measurement, 48 Pervasive Developmental Disorder cases were evaluated by the same clinician in 60+/- 10 day intervals.Results: The findings within our study reflect definite evidences for the validity and the reliability of CARS.When the reliability results are examined, the Cronbach alpha coefficient was found to be 0.95.The intraclass correlation coefficient value that supports generalizability was found to be 0.96. 46 PDD cases were evaluated independently by two independent clinicians for the measurement of reliability among the evaluators, and the correlation coefficient was found to be 0.98. To measure the test-retest consistency, 48 children were evaluated by the same clinician within 60+/-10 day intervals and the correlation coefficient was found to be 0.98.Within the validity scope, construction validity (?), measurement validity (?) and discriminative validity (?) were examined. In the factor analysis that was made to assess construction validity, all items of the scale were correlated with each other significantly, except the 14th item that questioned the consistency of mental functions. To assess the measurement validity, CARS total scores were compared to the CGI scores and ABC total scores; the correlation coefficients for each one were found to be 0,873 and 0,567, respectively. The scale was shown to have a good degree of discrimination (p<0.001) between the autism and the mental retardation group.Conclusion: The Childhood Autism Rating Scale carries the qualities of a standardized measurement instrument. It is considered that the Childhood Autism Rating Scale may contribute to the reasearches carried out in the pervasive developmental disorder field as well as to the clinical practice.Keywords: Autism, Pervasive Developmental Disorder, Childhood Autism Rating Scale