Parent Reports of Social Functioning Provide Evidence of Criterion-Validity for the Social Responsiveness Scale, Second Edition in a High Functioning Sample of Youth with ASD

Background: The Social Responsiveness Scale, Second Edition (SRS-2; Constantino, 2012) is a parent/teacher-report measure of social impairment associated with Autism Spectrum Disorder (ASD) and is widely used as a screening tool to aid in clinical diagnosis.  The SRS has demonstrated strong predictive and concurrent validity.  However, we know little about criterion-related validity for the SRS-2. 

Objectives: In this study we examined the relations between SRS-2 scores and parent reports of peer relations, social skills, and leadership, as well as child report of interpersonal relations in a sample of high functioning youth with ASD.  It was hypothesized that higher scores on these scales (indicating better social functioning) would be associated with lower scores on the SRS (less social impairment) in this sample. 

Methods: Data from 60 youth with high functioning autism (i.e., IQ >70) between the ages of 7 and 17 was analyzed.  SRS-2 raw score data was correlated with data from a brief parent questionnaire inquiring about the quality of their child’s peer relationships, and also with scores from three subscales of the Behaviour Assessment System for Children, Second Edition (BASC-2): Social Skills and Leadership (both parent-rated), and Interpersonal Relations (child-rated).  An ordinary least squares multiple regression analysis was also used to examine the predictive relationships between these variables and SRS-2 scores.

Results: Results showed that SRS-2 raw scores were significantly negatively correlated with parent reports of peer relations (r=-.532, p<.000), number of acquaintances the child had (r=-.259, p<.05), number of close friends (r=-.265, p<.05), amount of time in a year spent with a close friend (r=-.270, p<.05), parent rated social skills (r=-.516, p<.000), and parent rated leadership (r= -.662, p<.000).  However, child report of interpersonal relations was not found to be significantly correlated with SRS-2 scores.  Next, a multiple regression analysis was conducted controlling for age and IQ in block one of the regression model, and predictor variables entered in block two.  Ordinary least squares regression analyses show that the variables in Model 1 (age and IQ) did not explain a statistically significant proportion of the variance in SRS-2 scores (R2=.012, F=.316, p=.730>.05).  Model 2 which included all the predictor variables accounted for an additional 55% of the variance in SRS-2 scores (ΔR2=.55, ΔF= 8.244, p=.000).  T tests show that parent reported peer relations (β=-.312, t=-2.225, p=.03<.05) and leadership scores (β= -.509, t= -3.28, p=.02<.05) have the only statistically significant regression coefficients in the model.

Conclusions:  Our prediction that higher scores on parent reports of social functioning would be associated with lower scores on the SRS-2 was confirmed.  However, the same was not true for child report of interpersonal relations.  These findings provide evidence of criterion-related validity for the SRS-2 among high functioning youth with ASD, at least with regard to parent reports of social functioning.