Measurement Invariance Across Youth And Parent Sdq Ratings?

The internal structure of the Strengths and Difficulties Questionnaire (SOD) has been a controversial issue. The first purpose of this study was, therefore, to analyze the internal structure of the parent and youth versions of the SDQ. Using data from the first repeat survey of the KiGGS study, we found that a five-factor model with correlated error variances between items with similar content fits well with the data for parent and youth SDQ scores. However, to compare the scores across parent and youth version, the 5 latent variables need to have the same meaning for parents and youths. Hence, in a second step. measurement invariance across both SDQ versions was tested. Using multigroup confirmatory factor analysis we found that the SDQ measures the same constructs across parents and youths (configural invariance). However, scalar invariance was only observed for 19 out of 25 items. Consequently, any differences between parents and youths on the latent variables could be caused either by a lack of measurement invariance between the two respondent versions or by a real difference across parents and youths. Therefore, latent mean differences between parent and self-report versions should be interpreted with caution.