ESRA 2027 sessions by theme

Back to Overview of Sessions

Session Details

Comparability is one of the foundations of survey-based inference, especially for large cross-national infrastructures such as ESS, EVS, WVS, or ISSP. Such programmes have made measurement invariance (MI) testing a standard precondition for valid comparison, particularly for attitudinal and value constructs, where cultural, linguistic, religious, and normative context can shape not only response levels but the very meaning of an item. In comparative research more broadly, whether using one of these programmes directly or simply items drawn from them, an instrument is often taken to capture the same underlying construct across the countries, cultural or linguistic groups, time points, modes, or subpopulations being compared, with this assumption left untested more often than not.
The same logic extends to surveys, national, regional, or targeting a specific population, that adopt items from the above mentioned large infrastructures on the assumption that prior testing within the source programme is sufficient to guarantee their validity for a different population or setting. Whether borrowed items hold under such conditions is seldom examined. Where invariance is tested, across any of these settings, full equivalence frequently fails to hold. In practice, this non-equivalence tends to be set aside, since comparability remains a core objective of the research, yet it cannot be treated as a minor technical caveat.
We invite contributions addressing measurement (non-)equivalence in any of these forms, including cross-survey and cross-population applications of common items, and methodological or applied work diagnosing where and why instruments fail to travel across contexts. Contributions are welcome regardless of analytical approach or strategy employed, and regardless of substantive domain or item(s) focused on, provided the paper's central concern is whether, and how, measurement holds across the compared groups, waves, surveys, or contexts. Contributions focusing on values and attitudes will be particularly welcome.