Conferences

Several sites providing online access to data and/or metadata have come into practice (e.g. ESS/NSD, IQSS Dataverse, ASEP/JDC data bank, ICPSR, CCESD, badason, ZACAT, Statista). In this presentation the online service of SDA (Survey Documentation & Analysis – http://sda.berkeley.edu/ ) is focused. This set of programs provides documentation and web-based analysis of survey data. It is maintained by the Computer-assisted Survey Methods Program (CSM) at the University of California, Berkeley.

At this time, SDA is considered from two perspectives, those of a data service manager and of an empirical sociologist. My attention was drawn this site, because it offered straightforward access to some important data as GSS, American National Election Study, and Census Microdata. The facilities were used for various purposes as investigating metadata, data download, statistical analysis, and finally for the pilot implementation of special surveys into the system (ISSP and ALLBUS). Comparisons across time were done concerning religious beliefs and social pessimism.

Starting major tests with the integration of ALLBUS 1998 into SDA the facilities of looking into metadata were already rather satisfying, but later on some important improvements could be added (especially text search within a single study and across studies). This improved the relative position SDA formerly had in comparison with some similar systems.
Analysis of data has always been exceptionally quick. The access to the basic statistical procedures is pretty easy even for newcomers with only basic statistical and/or computer knowledge. Helpful documentation on statistics is online. Moreover, interested users are not restricted to the given original data, but may create and even store their own user-defined variables. New highlights at SDA concern safeguarding against disclosure risks and offering complex standard errors in some statistical analysis, which are two quite relevant implications for elaborate statistical analyses.
Unrestricted download of even large files in various formats is provided. Download of customized metadata and data sets turn out to be straightforward. On the other hand, there are also possibilities of controlling the access, if such protection is needed.

Of course, there would be challenges for future work (e.g. implementing more exploratory statistical procedures (e.g. PCA, LCA)). Nonetheless, as data manager and as researcher I do confirm that SDA has really proven to be useful online tool.