Survey programmes accumulate an ever-growing treasure trove of data. However, studies often use different questionnaire items to measure identical concepts. This makes data difficult to compare and to combine. Comparing and combining survey data, however, creates great opportunities: time series can be extended, interesting subpopulations can be explored with adequate (combined) sample sizes, and research results can be better secured with integrative meta-analyses. The “Harmonized Variables” Service will help to unlock this potential by making measurement instruments and the data collected through them more comparable in three ways:
Comparability through standardised questionnaire items
Researchers are recommended to use standardised questionnaire items when constructing their questionnaires. The Demographic Standards provide a suitable basis for measuring socio-demographic characteristics in German surveys. Building on this and using other established constructs, a RatSWD working group has developed a standard questionnaire to collect socio-demographic and crisis-related variables [in German]. For some socio-demographic characteristics, additional short versions are suggested that are tested and compatible with the original versions, or more in-depth follow-up questions. In addition, this questionnaire includes questions with particular relevance in social crises, for example on trust in political institutions, social cohesion or resilience. Researchers can use these instruments directly into their questionnaires.
In addition, for some socio-demographic characteristics, detailed documentation based on the Demographic Standards and the RatSWD Standard Questionnaire, supplemented by in-depth content on the respective characteristic, is published in the Open Access Repository for Social and Behavioural Measurement Instruments, ZIS (in German).
Comparability through standard socio-demographic variables
In addition to standard socio-demographic questionnaire items, standard socio-demographic variables were developed and partly validated to allow making variables comparable after data collection. They specify uniform metadata such as variable names and coding schemes. If, for whatever reason, a study cannot measure a socio-demographic characteristic using the above-mentioned standard items, researchers should ensure during questionnaire development (i.e. ex-ante) that the corresponding standard variable can be derived from the data collected. In addition, RDCs can apply these standard variables to the microdata they curate (i.e. ex-post) and make them centrally available to data users. In this way, comparability across surveys can be achieved at least indirectly. These standard variables will be documented in the open access repository for social and behavioural science measurement instruments, ZIS.
Comparability through harmonisation of latent constructs
The harmonization database QuestionLink offers recoding scripts for selected latent constructs (e.g., political interest or general life-satisfaction). These allow the harmonization of all data on those constructs collected by the covered survey programs across all measurement instruments. Additionally, we facilitate the reuse of our methodological innovation and automatization technology for RDCs and researchers with educational materials and an R-package.
- QuestionLink is currently online with four harmonised constructs: https://www.gesis.org/en/services/processing-and-analyzing-data/data-harmonization/question-link
- Report with an overview on the collection of socio-demographic variables in large German surveys: https://zenodo.org/record/6810973 [in German]
- Open access repository for social and behavioural science measurement instruments ZIS: http://zis.gesis.org/ [in German]
- Demographic Standards: https://www.destatis.de/DE/Methoden/Demografische-Regionale-Standards/textbaustein-demografische-standards.html [in German]
- RatSWD standard questionnaire with socio-demographic and crisis-related variables.: https://www.konsortswd.de/publikation/standardfragen-soziodemographische-und-krisenbezogene-variablen/ [in German]