Exploratory factor analysis of the Gilliam Autism Rating Scale - second edition with a sample of students with autism spectrum disorders
Dua, Elissa H.
MetadataShow full item record
The Gilliam Autism Rating Scale - Second Edition (GARS-2; Gilliam, 2006) is a widely used screening tool that assists in the identification and diagnosis of autism spectrum disorders (ASDs). Although this assessment measure is commonly used during autism assessments, there has been little research on its psychometric properties. The purpose of this study was to examine the factor structure, internal consistency and screening sensitivity of GARS-2 ratings completed by special education teaching staff for a sample of students ( N = 216) with autism spectrum disorders. The exploratory factor analysis supported the retention of three factors labeled Stereotyped and Repetitive Behaviors, Social Avoidance and Withdrawal, and Atypical Language and Communication. These three factors were very similar to those found in prior research by Lecavalier (2005) on the original GARS and Volker et al. (under review) using the GARS-2 with a more mixed developmental disabilities sample. In the present study, internal consistency estimates met or exceeded standards for screening and were generally higher than those in previous studies. Using an Autism Index cut score of 85 yielded a screening sensitivity of .5879, while a cut score of 90 yielded a sensitivity of .4537. These sensitivity results were generally similar to those of prior studies involving the GARS or GARS-2. It is noteworthy that the three-factor solution retained in the present study matches very closely with those from two prior studies with different samples (i.e., Lecavalier, 2005; Volker et al. [under review])--suggesting the presence of a robust three-factor structure. However, this factor structure was different from the four-factor structure found by Pandolfi, Magyar and Dill (2010) using the GARS-2 standardization sample. Although this study provides some psychometric support for the use of the GARS-2 as a screening tool, the sensitivity estimates for the Autism Index suggest that the instrument results in a high percentage of false negative results for ASD. This significantly limits the GARS-2's utility and, along with the factor analytic results, suggests the need to substantially revise the instrument. Implications for instrument revision and future research are provided.