Abstract
This article argues that the concept of equivalence is one of the most important methodological aspects of valid and reliable measurement in cross-national survey research. The important topic of survey measure equivalence has not been systematically in Czech social science publications to date and this article hopes to address this gap in the literature.
Consequently, the two main goals of this article are (1) to acquaint the reader with techniques that are used to find questions that are interpreted in the same way across countries before data collection and (2) to describe the testing and evaluation of measurement indicators’ equivalence or comparability after data collection. This study presents cognitive approaches to “good” question wording practices, best translation practices and the application of both ‘emic’ (culture specific) and ‘etic’ (culture universal) approaches to survey question design.
After data collection a range of statistic techniques are usually employed ranging from basic statistics such as the mean to advanced approaches such as multi-group structural equation modelling, multilevel modelling, latent class modelling and Item Response Theory. This article describes some of these techniques in the context of measurement equivalence and its associated research literature.