Premium
Factor structure of the Childhood Autism Rating Scale as per DSM‐5
Author(s) -
Park EunYoung,
Kim Joungmin
Publication year - 2016
Publication title -
pediatrics international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.49
H-Index - 63
eISSN - 1442-200X
pISSN - 1328-8067
DOI - 10.1111/ped.12770
Subject(s) - confirmatory factor analysis , dsm 5 , autism spectrum disorder , autism , rating scale , childhood autism rating scale , reliability (semiconductor) , clinical psychology , medicine , scale (ratio) , psychometrics , psychiatry , psychology , structural equation modeling , developmental psychology , computer science , machine learning , power (physics) , physics , quantum mechanics
Abstract Background The DSM‐5 recently proposed new diagnostic criteria for autism spectrum disorder (ASD). Although many new or updated tools have been developed since the DSM‐IV was published in 1994, the Childhood Autism Rating Scale (CARS) has been used consistently in ASD diagnosis and research due to its technical adequacy, cost‐effectiveness, and practicality. Additionally, items in the CARS did not alter following the release of the revised DSM‐IV because the CARS factor structure was found to be consistent with the revised criteria after factor analysis. For that reason, in this study confirmatory factor analysis was used to identify the factor structure of the CARS. Method Participants (n = 150) consisted of children with an ASD diagnosis or who met the criteria for broader autism or emotional/behavior disorder with comorbid disorders such as attention‐deficit hyperactivity disorder, bipolar disorder, intellectual or developmental disabilities. Previous studies used one‐, two‐, and four‐factor models, all of which we examined to confirm the best‐fit model on confirmatory factor analysis. Results Appropriate comparative fit indices and root mean square errors were obtained for all four models. The two‐factor model, based on DSM‐5 criteria, was the most valid and reliable. The inter‐item consistency of the CARS was 0.926 and demonstrated adequate reliability, thereby supporting the validity and reliability of the two‐factor model of CARS. Conclusions Although CARS was developed prior to the introduction of DSM‐5, its psychometric properties, conceptual relevance, and flexible administration procedures support its continued role as a screening device in the diagnostic decision‐making process.