A systematic review of evidence for the psychometric properties of the Strengths and Difficulties Questionnaire. (January 2016)

• Record Type: Journal Article
• Authors: Kersten, Paula; Czuba, Karol; McPherson, Kathryn; Dudley, Margaret; Elder, Hinemoa; Tauroa, Robyn; Vandal, Alain
• Is Part Of: International journal of behavioral development. Volume 40:Number 1(2016:Jan.)
• Page Start: 64
• Page End: 75
• Publication Date: 2016-01
• Subjects: Reliability -- SDQ -- Strengths and Difficulties Questionnaire -- validity; Developmental psychology -- Periodicals; Behaviorism (Psychology) -- Periodicals; 150
• DOI: 10.1177/0165025415570647
• Languages: English
• ISSNs: 0165-0254

Abstract

This article synthesized evidence for the validity and reliability of the Strengths and Difficulties Questionnaire in children aged 3–5 years. A systematic review using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guidelines was carried out. Study quality was rated using the Consensus-based Standards for the Selection of Health Measurement Instruments. In total, 41 studies were included (56 manuscripts). Two studies examined content and cultural validity, revealing issues with some questions. Six studies discussed language validations with changes to some wording recommended. There was good evidence for discriminative validity (Area Under the Curve ≥ 0.80), convergent validity (weighted average correlation coefficients ≥ 0.50, except for the Prosocial scale), and the 5-factor structural validity. There was limited support for discriminant validity. Sensitivity was below 70% and specificity above 70% in most studies that examined this. Internal consistency of the total difficulty scale was good (weighted average Chronbach's alpha parents' and teachers' version 0.79 and 0.82) but weaker for other subscales (weighted average parents' and teachers' range 0.49–0.69 and 0.69–0.83). Inter-rater reliability between parents was moderate (correlation coefficients range 0.42–0.64) and between teachers strong (range 0.59–0.81). Cross-informant consistency was weak to moderate (weighted average correlation coefficients range 0.25–0.45). Test-retest reliability was mostly inadequate. In conclusion, the lack of evidence for cultural validity, criterion validity and test-retest reliability should be addressed given wide-spread implementation of the tool in routine clinical practice. The moderate level of consistency between different informants indicate that an assessment of a pre-schooler should not rely on a single informant.