Parent-Teacher Concordance On the Social Responsiveness Scale for Children with a Putative Diagnosis of ASD

Background: The Social Responsiveness Scale (SRS) was developed as a screener and diagnostic aid for individuals on the autism spectrum.  While it was originally developed as a parent-report instrument, Constantino et al. (2007) published a study confirming its validity as a teacher report measure.  This validation study included a sample of 577 children recruited from the Washington University School of Medicine and the Autism Genetic Resource Exchange.  Parent-teacher responses were highly concordant (r=0.72); high scores by both teacher and parent increased the diagnostic accuracy of the SRS (Constantino et al., 2007).  Generalizability of the findings may be limited, however, due to the lack of socio-economic diversity in their sample and the need for replication.

Objectives: To examine parent-teacher concordance on the SRS with a racially and economically diverse sample of children in autism support classrooms in a large, urban setting, and child and family characteristics associated with concordance. 

Methods: The sample included 120 students enrolled in 38 kindergarten-second grade autism support classrooms in a large, urban school district.  All subjects received a primary diagnosis of autism and an autism support placement from a school district psychologist prior to enrollment in this study.  Subjects were administered the Autism Diagnosis Observation Schedule (ADOS) as a confirmatory measure.  The Social Responsiveness Scale (SRS) was completed by parents and teachers at two points during the school year.

Results: Data analyses are ongoing.Preliminary results suggest no significant correlation between teacher and parent report on the SRS raw scores (r=0.29, p=0.01).Further analysis suggests that 15% of the sample does not meet research criteria for an ASD. For children meeting criteria, the correlation between parent and teacher scores is moderate, but highly significant (r=0.34, p=0.008).For children not meeting criteria, the correlation between the two ratings were very strong, although not statistically significant (r=0.84, p=0.81). In this group, parents consistently rated their children as more impaired than teachers did. Analyses to be presented will further address factors associated with parent-teacher agreement.

Conclusions: Among children meeting research criteria for ASD, there was good parent-teacher agreement on the SRS in this diverse, primarily low SES group, suggesting the broad utility of the SRS. Among children not meeting research criteria for ASD, agreement was poor, with parents consistently making more severe ratings. Among other reasons, discordance may be due to parents’ concerns about loss of services or misattribution of symptoms.