Assessing and addressing measurement equivalence in cross-cultural surveys 2
|Convenor||Dr Gijs Van Houten (Eurofound )|
|Coordinator 1||Dr Milos Kankaras (Eurofound)|
Unlike variable-centered measures, validity and stability of typologies have been rarely studied. Previously, we developed a value typology of the European population using the fourth round of the European Social Survey. The current paper tests the stability of this typology by extending the study to three time points. It was found to hold partial invariance. The reliability of the value classes was supported by stability of country probabilities of the classes and its correlations with the economic development. The results imply that the value classification of Europeans is not ad hoc, but reflects the natural structure of European societies.
The Strength and Difficulties Questionnaire (SDQ) is widely used for measuring dimensions of externalising and internalising behaviours in children and adolescents. It is hypothesised to measure five subscales: Conduct, Hyperactivity, Peer relations, Emotional problems and Prosocial behaviour. However, explorations of factor structure have failed to validate the five dimensions. Additionally, there has been very limited testing of invariance of the proposed SDQ subscales over time. In the current study, we will aim to explore factor structure of the SDQ and test whether or not both metric and scalar invariance over time is given.
The paper is devoted to the assessment of measurement equivalence in cross-cultural surveys. In the report, I address problems of equivalence we faced in the research project “Alienation and Group-Focused Enmity in European Context ” (Ekaterina Lytkina, Andreas Zick) in seven European countries. I will dwell on the results of multigroup latent class analysis as well as show MIMIC models where separate analysis was carried out within each country. Methodological questions as well as substantive results that we got from equivalence assessment will be taken in consideration.
Trust in institutions has been measured among European citizens by European Social Survey during three time periods, 2008, 2010 and 2012 for 8 countries. European societies have witnessed the presence and consequences of crisis, affecting with different intensity and different ways their views on the political system and institutions. We consider trust in institutions a complex theoretical construct treated as latent variable using CFA. ESS data will be used to test whether indicators for trust in institutions hold across periods and countries, affected differently by the crisis. Measurement invariance in Multiple Group Confirmatory Analysis will be used.