Harmonized Variables

Survey programs are creating ever-growing volumes of valuable data. To fully unlock their potential, these data must be made findable, accessible, interoperable, and re-usable (in line with the FAIR principles). Content interoperability in particular is a major hurdle for survey data. Different instruments are used to measure identical concepts. This renders survey data less comparable and less combinable. Combining survey data, however, creates great opportunities: it could help extending time series, exploring interesting subpopulations with adequate (combined) sample sizes, and making research results more robust and reproducible with integrative meta-analyses.

The Harmonised Variables Service will help to unlock this potential by increasing the comparability of measurement instruments and the data collected through them. For the measurement of socio-demographic characteristics, best practice guidelines for data producers on questionnaire design will be developed on the basis of the demographic standards. On the other hand, specifications for standard variables will be developed for research data centers and data users, which will use existing classifications, if available. These standard variables allow to make existing data on different characteristics comparable (i.e., ex- post), or ex-ante, i.e., before a survey, to ensure that the resulting data can be transformed into comparable variables after the survey if standardized survey instruments cannot be used.

For use on latent constructs (e.g., attitudes, interests, or values), we will create QuestionLink, a harmonisation database. It will allow users to efficiently harmonise data for selected constructs from all participating surveys and across all the used measurement instruments. In addition, we will develop accessible education and training materials on QuestionLink as well as on methodological innovations in ex-post harmonisation of latent constructs in general.