The UC Prussia Project

The Prussia Project was conceived by Patrick Galloway in the late 1980s and subsequently developed by Galloway, E. A. Hammel, and R. D. Lee, with support from NICHD (R01 HD25841). The documents on this web site consist of a list of publications on the research, a codebook for the data files from the Prussian Statistical Office, and the data files themselves. The following links provide access to the documents:

• Publications. (HTML format)
• Codebook. (TIFF format or Postscript format. TIFF files can be displayed by many graphics applications. Postscript files can be printed by many word processing programs.) The codebook describes the data format for each individual data file and also shows a photocopy of one page from the source document, with the original German row and column headings.
• Data files.(TEXT format, tab-delimited, can be read by most spreadsheet or statistical programs). This link will take you to the DATA directory and a listing of the files. Clicking on a filename will display it in your browser, and you can download it from the browser. With four exceptions these files are individual data files identifiable from their names (and from the codebook) by subject and year. For example, VIT1904.txt is the file of vital event records for the year 1904. The four exceptions are:
  1. WK3DATA. This is a subdirectory containing the original Lotus 1-2-3 data files. They are an alternative source format. See also ZIPPEDPRUSSIA.ZIP.
  2. ZIPPEDPRUSSIA.ZIP This is a zip file containing all of the individual wk3 data files. The readme file gives instructions on accessing ZIPPEDPRUSSIA.ZIP after downloading it. Accessing this file is a quick way to get the whole data set. Some other spreadsheet programs can read wk3 files directly.
  3. readme.txt This text file contains instructions on the ZIPPEDPRUSSIA.ZIP file.
  4. bonecave. This is a subdirectory containing the original floppy disk images. Leave it alone unless you want to crosscheck the data files.

The research required the scanning and conversion by OCR techniques of remarkably detailed Prussian statistical records c. 1875-1914 at a level of aggregation analogous to the census tract or county level in US censuses, and the establishment of a set of such geographical units that was consistent across the time span of the records. Much of the analysis involved the use of pooled cross-section and time series methods. The general conclusions of the research are that socalled cultural variables, proxied by language and religion, are important determinants of the level of fertility but not of the rate of fertility decline. On the other hand, structural and economic variables are powerful determinants of the rate of fertility decline and thus in the long run must be determinants of ultimate levels of fertility. This formulation, supported by sophisticated pooled time series analysis of data of impeccable quality, is an important correction to the conclusions of the Princeton European Fertility Project, which maintained that structural and economic variables were of little import. Details are to be found in the publications from the project.

In pursuit of cultural determinants of demographic behavior, the project also explored classification of small regions by folkloristic characteristics and by dialect. The abundant German survey data on folkloristic characteristics proved to have been contaminated by political correctness during the Nazi era and were thus unreliable as indicators of local folkloristic traits. The complexities of classifying small regions by dialect have not yet been resolved.

NOTE: Patrick Galloway was originally listed as a co-author of the administrative final report submitted to the funding agency at the conclusion of the Prussia project and previously posted on this website. Dr. Galloway disagrees with some of the statements in the final report and his name has been removed from the list of authors at his request. We regret any misunderstanding.