Psychometric properties of the Arabic self-report version of the strengths and difficulties questionnaire

Abstract

Students in middle school tend to display emotional and behavioural difficulties (EBDs) compared to other forms of psychopathology. Early identification of EBDs is therefore a priority in order to prevent the chronic co-morbidity with other forms of psychopathology which may affect students’ academic achievement. Assessment of EBDs has been traditionally undertaken via proxy reporting; but psychometrically rigorous instruments are needed so that children and adolescents can report on their own emotions and behaviours. Such need increases in the Omani context given the lack of EBDs adequate assessment instruments. In the current study the factor structure of the Arabic self-report version of the Strengths and Difficulties Questionnaire (A-SDQ) was examined in a sample of 815 middle school students (mean age=14 years). The study tested the SDQ original five-factor model which received considerable empirical support. Responses on the A-SDQ were compared to responses obtained via proxy reports from teachers and parents through confirmatory factor analyses (CFAs). Results showed a reasonable fit for the three informant forms. Nevertheless, there were differences in item loadings across the three informant forms. Additionally, participants’ self-report responses were tested for invariance across gender. CFAs provided support to the invariance hypothesis for item loadings, indicating that the items were similarly valid indices of the five factors for males and females. Factor correlations, factor variances and item residuals were not invariant across gender. What this paper addsThis paper, to our knowledge, is the first research paper that provides empirical evidence on the Arabic self report version of the Strengths and Difficulties Questionnaire (A-SDQ) within a school-based sample. Though the focus was on the self-report version we collected data from the proxy versions for parents and teachers to provide additional evidence on the construct validity of the (A-SDQ) through cross informant data.