Demolink Compared to Automatic and Manual Systems Record Linking Strategy

3. The Data Material and the Demolink System

In my project Demolink is used on data from censuses, church books and land registers from the parish of Asker and Bærum, situated 5-30 km west of Oslo. The sources and dates in the study are shown in Table 1. Its purpose is to study various aspects of demographic behavior in the past.


The choice of area was determined largely by the availability of sources. While the 1801 and 1865 censuses for Norway are nominal, the 1815 to 1855 censuses are normally only numerical. The Asker and Bærum parish, however, has quite good nominal drafts from the census takers in 1815, 1825 and 1835.

The period chosen was partly limited by sources already computerized, and the courtesy of public institutions involved in the project. The starting point was the 1801 census, already computerized at the University of Bergen in the 1970s. The final point was originally set to the census of 1865, available in machine readable form from RHD at the University of Tromsø. RHD also computerized the church records from 1814-1878 for this project. The censuses from 1815, 1825 and 1835 as well as the land registers from 1802-03 (manuscript), 1826, 1838, 1866 and 1888 (all printed) were computerized with the help of a graduate student.

The 1801 population in the parish was close to 4600 inhabitants, and in 1865 the number of inhabitants had risen to about 8400. The total number of individual event records was about 100,000. The Demolink system facilitates linkage of the records from all sources simultaneously by presenting all the individual event records, i.e. each time a person is mentioned in a source, in one individual event file. The number of entries in the event file was about 125,000 because people with more first names or more surnames were listed as many times as they had different names. Thus an Anne Kari Hansdatter would be listed both among the Anne Hansdatters and the Kari Hansdatters.

Neither the software, the data preparation, the production of files, nor the design of the system with its user interface, will be presented and discussed in this article. Only the most basic information needed for understanding the use of the system will be given. The computer's part is to calculate, (re-)organize, combine, check, store, retrieve and present the data, while the thinking is left to the historian.

Demolink displays the data principally in two windows, shown in Figure 1. The large, upper window, the individual event window, shows a part of the individual event file, which contains records with the first name, patronymic, residence, year of birth (either given or calculated), role in the event (such as groom, father, mother, baptized child, presence in census or land register). The lower left corner of the screen shows the source entry window, a window presenting complete information from the source entries. A useful pop-up window presenting linked life courses is not shown in the figure.

In the individual event window, those records selected by the historian as pertaining to the same person are shown in inverse video. The source entry window gives more detailed information from the sources. Fig. 1 shows three of the events in the life course of Gunnild Sørine Pedersdatter, in this case her baptism, presence in the 1835 census, as well as her marriage.

The individual event file can be presented in different ways according to which sorting criteria are used. The version mostly used was the one where the individual event records were sorted alphabetically after first name, patronymic, residence and year of event. Before sorting, the names were standardized to a code by a separate name standardization system called Foneq [Nygaard 1992]. The standardization is partly etymological, partly phonetic. The names in the sources, not the codes, appear on the screen, e.g. the first names Embret and Ingebrigt receive the same code, and they are sorted to the same place in the file even if the spelling is quite different. The linkage proper was done by selecting individual event records, and then storing them as linked individuals. Each individual received a distinct person number generated by the computer.


Demolink Compared to Automatic and Manual Systems Record Linking Strategy

Interactive Record Linkage: The Cumulative Construction of Life Courses
Eli Fure
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871