Show Less
Restricted access

Psychometrically Relevant Differences between Source and Migrant Populations


Patrick Brzoska

Psychometrically relevant population differences may limit the transferability of research instruments between migrants and their source population and may contribute to a low performance of quantitative questionnaires. Based on a review of existing taxonomies, this book develops a comprehensive analytical framework of equivalence that can serve multiple purposes. It allows to examine psychometrically relevant population differences, it can assist in the re-adaptation of questionnaires and it is a valuable tool for cross-group comparative research. The application of the framework is illustrated through the examination of equivalence of the Revised Illness Perception Questionnaire (IPQ-R) between chronically ill Turks residing in Turkey and Turkish migrants residing in Germany.
Show Summary Details
Restricted access

8. Conclusion, outlook and recommendations for researchers


The (measurement) equivalence of questionnaires is usually discussed in a cross-cultural research context where it is regarded as a crucial prerequisite for standardization and comparability. The present investigation shows that questionnaires applied to populations of the same culture in different geographical locations who have been separated from each other for years and decades may also show substantive evidence of semantic and measurement non-equivalence. The present study continued the explorative work of Brzoska & Razum (2010). It revealed by means of a hypothesis-testing approach that the Revised Illness Perception Questionnaire (IPQ-R), the most widely used questionnaire for the assessment of illness perceptions, cannot be regarded invariant between Turks residing in Turkey and Turkish migrants residing in Germany in terms of its semantic and measurement properties. While a 34-item measurement model showed good factorial validity in Turks residing in Turkey, the same measurement model had a weak performance in Turkish migrants residing in Germany and needed substantial modification to reach acceptable model fit. Although partial configural invariance could be established, subsequent analyses identified several items with differential item functioning in both population groups.

What do these results mean for quantitative public health researchers working on the health of migrants? First of all and most importantly, the results show why questionnaires validated in source populations may show poor properties when administered to migrant populations: different forms of method and item bias that partially go back to language differences may contribute to an overall reduced structural validity of questionnaires. Thorough...

You are not authenticated to view the full text of this chapter or article.

This site requires a subscription or purchase to access the full text of books or journals.

Do you have any questions? Contact us.

Or login to access all content.