The challenge of shortening an existing well-validated instrument: Methodological procedures for constructing a short German version of the SDQ I

Year: 2012

Author: Arens, Katrin, Hasselhorn, Marcus, Craven, Rhonda, Yeung, Alexander

Type of paper: Abstract refereed


Purpose: The Self Description Questionnaire I (SDQ I; Marsh, 1990) is one of the most extensively used and empirically validated instruments for measuring the multidimensional self-concept of preadolescents. However, the length of this instrument (consisting of 76 items) may prevent researchers and practitioners from utilising this instrument frequently. The aim of the study was to create an economical and yet psychometrically sound German version of the SDQ I by applying a series of steps to ensure the structure and psychometric properties of the original SDQ I instrument are maintained.
Method: A full German version of the SDQ I was generated, which strictly followed the English original instrument using a backward translation process. The full German instrument was validated by applying reliability estimates and confirmatory factor analyses using a sample of 1958 German students attending grades 3 to 6. As the full German version was found to be a reliable and valid for measuring German preadolescent students' multidimensional self-concept, it was used for deriving a shortened version by selecting the three best loading items for each self-concept factor. In order to test the equivalence between the full and short versions, the same tests of reliability and validity used for the long version were used again for the short version.
Results: The factor structures and psychometric properties of the long and short versions were similar both evincing the multidimensionality of self-concept. Both the long and short versions were further scrutinised with external validity criteria. Both versions demonstrated logical relations to external validity criteria including school grades and another self-concept instrument. To further confirm the theoretical and practical vigour of the instruments, they were used to replicate models depicted in the self-concept literature. Prominent models of self-concept research such as the twofold multidimensional structure of academic self-concept and the internal/external frame of reference (I/E) model could be replicated by both the long and short German versions of the SDQ I. To ensure the applicability of the new instruments to various samples, further tests demonstrated strict measurement invariance across grade and gender
Conclusion: Using the SDQ I as an example, this presentation provides an overview of the methodological procedures that are required for designing and validating a translated shortened version of a well validated instrument without compromising psychometric rigour.