Putting data in the driver’s seat
(Jul 20, 2017)Abstract
Various metadata systems for different sections of the data management lifecycle (e.g. questionnaire development, data preparation, documentation, data dissemination) are in use at institutions dealing with survey research. Some of these metadata systems make use of evolving metadata standards (such as DDI or SDMX), some others are developed independently as custom-tailored solutions. Most of them have one idea in common: Structured metadata, stored in relational databases, make it possible to have one single source of information for data on data.
With the increasing availability of metadata systems, their usage as a reference tool—e. g. for researchers looking for specific variables or questionnaire developers drawing on questions from other surveys—becomes more common. In this session we want to discuss uses of structured metadata that go beyond their passive reference function.
Since structured metadata are machine readable by definition, we are interested in exploring how and at which points in the data management lifecycle we can put metadata to use in a more active role. This may be as a means of automatically generating human readable questionnaires, automated plausibility checks during fieldwork, recoding raw survey data from the field and probably in numerous other ways. In order to implement data-driven data management processes, other sources of information come into play: for example paradata or sampling frame data can potentially be used in the same manner to enhance survey data management and gain the same benefits.
Papers presented in the session should thus focus on examples of the active use of such structured information. We would like to learn about your experiences with implementing data-driven routines as part of the data management process. The session will also provide room to discuss how much automation in the data management lifecycle is feasible and/or desirable.