Encoding and Linking Swedish Censuses (SweCens)

Project description

Sweden is the home of several extraordinary longitudinal population databases built on the country’s excellent registries dating back to the 18th century. Still, there are registries in past times that are rarely used for research and that hold remarkable openings for international
high quality research.

The SweCens infrastructure promises to improve possibilities for national and international research on social history, historical demography and beyond. Professor of history and director of Minnesota Population Center, Stephen Ruggles highlights in his support letter that: “Swedish demographic data is of high international interest because of the extraordinary quality of enumeration and the length of time they have been collected. Releasing the data in the same format as other international censuses adds to the value of the Swedish data by facilitating both cross-national comparison, and research with Swedish data by international
researchers.”

Scientific aims
In this project we aim to prepare, harmonize and release the Swedish census of 1890 to the national and international research community. In addition we also aim to link the 1890 and 1900 censuses testing and evaluating different methods of record linkage on the Swedish censuses. The work with encoding the data will be accomplished using an international standard format set by the Minnesota Population Center (MPC) at University of Minnesota,USA (http://www.pop.umn.edu/about-mpc ). The MPC is an important collaborator and supporter of our application. The planning grant is constructed as a pilot study that will improve our practical and theoretical knowledge in encoding and linking Swedish censuses. The experiences from the pilot study will be incorporated in a full Research Infrastructure application in 2012 that will include all censuses from 1860 to 1910.

The MPC is the world leading organization for preserving and disseminating historical census data and our collaboration ensure quality and compability and lower barriers for international research (www.. Since 2001 the MPC encompasses the North Atlantic Population Project (NAPP) that is a machine-readable database of the complete censuses of Canada (1881), Great Britain (1881), Iceland (1870, 1880, 1901), Norway (1865, 1900), Sweden (1900), and the United States (1880). The Swedish 1900 census has been available for researchers through the MPC website since October 2008 at www.nappdata.org. The Swedish 1890 census already exists in digitized form but the available data need to be standardized and coded in the same way as the other NAPP censuses, particularly with regard
to
a) geographic information, birthplaces and residence, occupation and ethnicity
b) verify string classifications carried out in each of the other NAPP participating
countries,
c) document cross-national differences in enumeration procedures.

With the support of the planning grant we will be able to proceed and work for a full Research Infrastructure application that aims at encoding, linking and releasing the Swedish censuses from 1860 to 1910 to the international research community. Data sets can be retrieved for research purposes only and free of charge.

In 2008, the 1900 Swedish census was successfully prepared and delivered to the MPC and included in the database of the North Atlantic Population Project (NAPP). Although this was an achievement in itself, adding more censuses will enable us to work on methods of linking census data. Swedish historic demographic data are known for its high quality and many scientific fields, nationally and internationally make use of the databases available for research. Especially the historical databases built on parish records, by the Demographic Data Base (DDB) and the Scanian Demographic Database (SDD), are unmatched by other nations databases concerning depth of generations and information that can be retrieved (see information at http://www.ddb.umu.se/ddb-english/database/the-database-popum and
http://www.ed.lu.se/EN/databases/sdd.asp .) The DDB and SDD databases follow individuals longitudinally in selected parishes, year by year, and important aspects of life events such as births, deaths, marriages etc, can be carefully monitored. The censuses, in contrast, cover the entire Swedish population and give a cross-section at a certain point in time.

The SweCens as a complement to other historical databases in Sweden.
The database TABVERK run by the DDB contains all the information about the population in all Swedish parishes during the period 1749 to 1859, as reported by the clergymen in large forms of tables (Tabellverket) to the Tabellkommissionen in Stockholm. The SweCens project pick up where TABVERK ends (1860) and provide researchers with more than 150 years of national statistics of Sweden. Currently three Swedish censuses have been completely registered by the SVAR-department at the National Archives of Sweden: 1880, 1890 and 1900. Also work has been started by SVAR to register the censuses of 1860, 1870 and 1910 and tests on the 1930 census have been made.

We anticipate both national and international research on the SweCens,
The censuses are a complete count of the Swedish population. In addition to regional or local longitudinal studies of individual data it gives researchers a chance to make a comprehensive survey of the entire population. Because censuses are based on nominative extracts from the ecclesiastical registers it can be linked to datasets from the DDB, SDD or the Rotemansarkivet at Stockholm City Archive, thereby offering possibilities to investigate tendencies and processes.

There are few comparisons between Sweden and the international community based on historical population statistics. One of the main reasons is that the high quality of the Swedish parish registers cannot be met by most other countries because they lack longitudinal individual data. However, as NAPP assigns uniform codes across all the censuses and brings relevant documentation into a coherent form to facilitate analysis international comparisons of social and economic change can be carried out. Given the high quality of the Swedish census and possibilities to link it to DDB, the SweCens can work on evaluating and controlling for different methods of record
linkage.

Moreover, the SweCens opens up for international scholars to do demographic research on Swedish material, thereby strengthening the field of the historical demography of Sweden.

Significance
The nine censuses that currently are available at the NAPP-database collectively comprise our richest source of quantitative information on the population of the North Atlantic world in the late nineteenth century. Sweden has so far provided its 1900 census and if more censuses are added it would multiply the alternatives for successful research and open up excellent possibilities for international comparisons and comparative analysis of human behaviour, social and economic change.

This kind of research can hardly be accomplished if data is not standardized and made more coherent. NAPP collaborators have harmonized the record layouts, coding schemes and documentation for the different censuses and assigned uniform codes across all the censuses.We anticipate that the SweCens project will have a long lasting effect on Swedish research in the humanities and social sciences. It opens up for Swedish researchers to make international comparisons and for international scholars to do demographic research on Swedish material, thereby strengthening the field of the historical demography of Sweden.

National and International Collaboration.
The SweCens project is collaboration between six different institutions, embracing two universities, the Stockholm City Archives and the Swedish National Archives. Within the National Archives the SVAR-department is responsible for providing the digitized censuses and the IT-department is responsible for the NAPP encoding. Several members of the team in SweCens took part in the work that led to the deliverance of the 1900 census in 2008. Mats Danielsson, Head of Research Archives of Umeå University led the group, Johan Gidlöf of the Stockholm City Archives was responsible for coding geographic variables, Dr. Mats Hayen of the Stockholm City Archives was responsible for coding work variables, Mats Berggren of the National Archives was responsible for coding other variables and project coordination and Dr. Per Axelsson, Centre for Sami Research, Umeå University coded ethnicity. All of the above, except Dr. Mats Hayen, are still involved in the current application but new expertise has been added. Professor Christer Lundh, Department of Economic History University of Gothenburg and Docent Sören Edvinsson, Demographic Database, Umeå University that both have published extensively within the field of social history and
historical demography and own significant experience from working with demographic databases. Professor Lundh and Docent Edvinsson will take part in evaluating the coding and linking and have an active role in the two scheduled projects. In Janary 2011, several members were added to the core team: Martin Dribe, Lunds University, Maria Wisselgren, Umeå University, Maria Larsson, Umeå University, Carl Szabad, National Archives.The project is led by Dr. Per Axelsson.

The SweCens project will also have close cooperation with our American colleagues at the Minnesota Population Center at University of Minnesota, USA led by professor Steven Ruggles, the Norwegian Historical Data Centre (NHDR) at the University of Tromsö and also
the Max Planck Institute for Demographic Research (MPIDR) in Rostock.