This study examined the structural validity of the teacher-report Strength and Difficulties Questionnaire (SDQ) in Norwegian preschoolers aged 3–6 years. We tested the original five-factor structure, the five-factor structure with two broader second-order factors, and a three-factor structure, all suggested in the literature. Since the positively worded items in SDQ have been shown to introduce noise, we also examined all three structures with a positive construal method factor for these items.
Preschool teachers from 43 preschools completed the SDQ questionnaire for 1,142 children [48% girls, mean age 4.3 (SD 0.9) years]. Confirmatory factor analyses were used to estimate and compare the six models. Measurement invariance was tested across sex (multi-group approach) and age (multiple-indicator multiple-cause approach).
The original five-factor structure of SDQ was supported, where the model fit improved when including a method factor for positively worded items. Both models showed scalar invariance across sex and age. The second-order and the three-factor structures were not supported.
We recommend using the original five-factor structure when using SDQ for both clinical and research purposes in young children and adding a method factor when using structural equation modeling.