Psychometric Analysis of the Social Communication Questionnaire Using an Item-Response Theory Framework: Implications for the Use of the Lifetime and Current Forms

Abstract

The Social Communication Questionnaire (SCQ) was developed as a screener of Autism Spectrum Disorder (ASD). To date, the majority of the SCQ utility studies focused on its external validity (e.g., ROC curve analyses), but very few have addressed the internal validity issues. With samples consisting of 2,134 individuals available from the National Database for Autism Research (NDAR), the current study examined the factor structure, item-level characteristics, and measurement equivalence of the SCQ forms (i.e., Lifetime form and Current form) using both the classical true score theory and the Item Response Theory (IRT). While our findings indicate sufficient psychometric properties of the SCQ Lifetime form, measurement issues emerged with respect to the SCQ Current form. These issues include lower internal consistencies, a weaker factor structure, lower item discriminations, significant pseudo-guessing effects, and subscale-level measurement bias. Thus, we caution researchers and clinicians about the use of the SCQ Current form. In particular, it seems inappropriate to use the Current form as an alternative to the Lifetime form among children younger than 5 years old or under other special situations (e.g., teacher-report data), although such practices were advised by the publisher of the SCQ. Instead, we recommend modifying the wording of the Lifetime form items rather than switching to the Current form where a 3-month timeframe is specified for responding to SCQ items. Future studies may consider investigating the association between the temporality of certain behaviors and the individual's potential for being diagnosed with ASD, as well as the age neutrality of the SCQ.

The Social Communication Questionnaire for adults with intellectual disability: SCQ‐AID

“…Further research is needed to determine if rephrasing these items would be appropriate (e.g., using the word “peers”). A similar idea was suggested by Wei, Chesnut, Barnard‐Brak, and Richman [], who found that certain items might differentially distinguish males and females because of gender stereotypical wording in items (e.g., “spinning the wheels of a car” or “soft toy”), which might influence the caregiver's response based on the gender of the child. Gender‐neutral phrasing of items might help to overcome some differences.…”

Extremely Preterm Born Children at Very High Risk for Developing Autism Spectrum Disorder

“…Total scores were compared with the established cut-off to screen for ASD (i.e., 15). External validity of the SCQ as a firstlevel screen for ASD in at-risk samples was demonstrated (sensitivity .88 -.96, specificity .72 -.80) (38,39) and sufficient internal validity was also established (40). Also the Social Responsiveness Scale (SRS) (41,42), a 65-item questionnaire, was administered.…”

An exploration of autism‐specific and non‐autism‐specific measures of anxiety symptomatology in school‐aged autistic children

“…It is noted that this screening tool provides an assessment of lifetime, rather than current, autistic characteristics; in future research it may be useful to employ a measure of current autistic characteristics rather than lifetime traits, to ensure that participants' current levels of anxious symptomatology are being compared to current, and not historical, autistic characteristics. Having said this, concerns have previously been raised regarding the psychometric properties of the SCQ Current Version (Wei et al, ); therefore, future studies would likely benefit from the inclusion of a gold‐standard autism diagnostic assessment.…”

“…It is well-researched and validated as a screening assessment for autism. While some concerns have been identified regarding the psychometric properties of the SCQ Current Version, the SCQ Lifetime Version is free of any such concerns (Wei, Chesnut, Barnard-Brak, & Richman, 2014). Although the recommended cut-off score on the SCQ for this age group is 15, a cut-off score of 11 has been recommended in younger age groups (Lee, David, Rusyniak, Landa, & Newschaffer, 2007).…”