In order to provide research data, the IAB draws on the register data generated during the administrative process, which is collated in the data warehouse of the Federal Employment Agency’s statistics department. This includes information on the services provided under labour market policy, on measures such as careers advice and preparation, and on employment that is above the earnings threshold for social insurance contributions. The IAB adapts this information to meet the needs of research and develops it into research data. Additional data is used from quantitative studies and panels and from qualitative surveys of companies, individuals, households and job centres. In strict compliance with data protection provisions, a wide range of data sources such as company and personal data can be combined using record linkage methods. The resulting data products are made available internally for research and as a basis for policy advice. They are then processed further and made available across various channels in anonymised form to the national and international research community via the Federal Employment Agency’s Research Data Centre at the IAB. External data sources are also harnessed, most notably large, partly unstructured data (big data), which includes audio recordings of interviews taken with the permission of the interviewees.
In order to increase the quality and usability of data products for research, the IAB attaches great importance to quality assurance throughout the data lifecycle and supports this systematically through research projects. To this end, the IAB develops and evaluates new ways of collecting, correcting and analysing data using, for example, experimental methods, statistical models and, increasingly, artificial intelligence.
In the area of qualitative methods, for example, one area currently being explored concerns the extent to which graphic elements (such as diagrams) and photos help the respondents in surveys of families and children to express themselves. In the field of quantitative methods, the ‘push-to-web’ approach is being used in employee surveys for datasets including the Linked Personnel Panel. By encouraging respondents to take part in surveys online instead of by telephone, researchers can capitalise on the advantages of online surveys, as with the OPAL survey, for example. In another project, differences in data quality between random and non-random samples will be examined, and potential corrective methods developed and evaluated. The focus here is both on the selectivity of the samples and the prevalence of inattentive respondents. To date, the IAB’s research has mainly relied on random samples. Expanding this to include non-random samples such as open-access online surveys could be one way of countering the steadily declining number of people willing to take part in telephone surveys.
With its extensive data-based expertise, the IAB is advising on the drafting of the Register Modernisation and Research Data Act as part of the federal government’s data strategy.
Activities
- Publications on the topic: Research Data and Methods
- Projects on the topic: Research Data and Methods
Departments
- Research Data Centre
- Statistical Methods
- Research Department Establishments and Employment
- Research Department Migration and International Labour Studies
- Research Department Panel Study Labour Market and Social Security
- Service Department Data and IT-Management
- Working Group Data Quality
- Working Group Qualitative Methods
