A Method for Universal Screening of Social Challenges in Elementary School Students

Background:  The current prevalence of ASD is 1 in 68, which is based on eight year old children (CDC, 2014).  Yet, the National Children’s Health Survey, which included 17,000 households with children between 6 and 16 years, indicated rates of 1 in 50  (Blumberg et al., 2013).  Thus, many individuals are diagnosed with ASD after early childhood. Further, a substantial literature indicates that social challenges in elementary school students are associated with limited academic performance (DiPerna & Elliott, 1999; Wentzel, 1991, 1993).  However, schools do not perform systematic, general screening for social challenges.  Universal screening of elementary school children may help to detect children with milder forms of ASD and lead to improved functioning. This report describes the empirical development of the Social Challenges Screening Questionnaire (SCSQ), a brief questionnaire to identify elementary school students at risk for ASD or other social disorders.

Objectives:  To empirically develop and test the SCSQ, which is appropriate for universal screening in elementary schools.

Methods:  The development phase included 30 participants (28 males) diagnosed with ASD after  age 6 years (ages 6 to 54 years; M = 19.55, SD = 10.93) and one parent informant..  Diagnosis of ASD was confirmed with ADI-R, ADOS, Vineland 2, and Stanford-Binet 5.  Participants for the validation phase were 35 teachers from eight charter schools who completed the SCSQ on 549 general education students (ages 8 to 11 years, M = 9.51, SD = 0.65).  Students with scores in the upper 2.5% of the distribution (n = 15) and a comparison sample of students (n = 35) from the lower end of the distribution, were selected for follow-up assessment with the Social Responsiveness Scale 2. The subsample (n = 50) consisted of 33 males and 17 females (ages 8 to 11 years of age; M = 9.42, SD = 0.67).    

Results: Scores on the SCSQ ranged from 0 to 14 (M = 1.38, SD = 2.42) and every item was endorsed by this sample.  Cronbach’s alpha was .85.   Inter-item correlations ranged from .008 to .638 and the deletion of any one item did not change reliability substantively in either direction (range =.82 to .85).  In the subsample of 50 students, total scores on the SRS-2 ranged from 2 to 167 (M = 60.22, SD = 41.13) and were significantly correlated (r = .87, p < .01) with the SCSQ.  An ROC curve was applied to compare the SCSQ to an SRS-2 dichotomy (0 = unaffected, 1 = affected) which indicated an optimal threshold score of 6 to yield a sensitivity of .94 and a specificity of .88.  

Conclusions:  Just as screening for vision and hearing impairments can lead to a referral to a specialist, the SCSQ should lead to a referral for formal evaluation for ASD.  Then, students with mild impairments can be identified early within schools and enrolled in appropriate evidence-based intervention programs.  The combined effect of universal screening, skilled evaluation, and early intervention may help to improve academic and long term functioning for students with social challenges.