AUTHOR=Chen Meng , Peng Kanglong , Zhou Libing , Weng Xiaofang 

TITLE=Differential item functioning in the children autism rating scale first edition in children with autism spectrum disorder based on a machine learning approach

JOURNAL=Frontiers in Neurology

VOLUME=Volume 16 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2025.1648991

DOI=10.3389/fneur.2025.1648991

ISSN=1664-2295

ABSTRACT=PurposeOur study used Rasch Analysis to examine the psychometric properties of the Children Autism Rating Scale First Edition (CARS1) in children with autism spectrum disorder (ASD).MethodsThe Partial Credit Model (PCM) was used to test reliability and validity. The GPCMlasso Model was used to test the differential item functioning (DIF).ResultsThe response pattern of this sample showed acceptable fitness for the PCM. This analysis supports the unidimensionality assumption of the CARS1. Disordered category functions and DIF were found for all items in CARS1. Performance can be related to age group, gender, symptom classification, and autistic symptoms.ConclusionRasch analysis provides reliable evidence to support the clinical application of the CARS1. Some items may produce inaccurate measurements originating from unreasonable category structures. Differences in age group, sex, and symptom classification can be related to test performance and may lead to unnecessary bias. Hence, clinical applications may require additional consideration of population characteristics to draw reliable conclusions.