Education and Middle-Class Society in Imperial Austria, 1848-1918

Appendix B: Statistical Methods

APP~IX .Statistical Methods Sampling Methods he data on occupational and class origins for the matriculated univ sity .and technical colleges students in Vienna and Prague derive samples of the· manuscript registration records preserved in the archives of the individual institutions (see Sources for a listing). Where -the institutions used catalogs, they recorded the individual student registrations in volumes in alphabetical order by the student's last name. Where they used individual registration forms (Nationalen), those forms were bound together in alphabetical order by the student's last name. Samples large enough for statistical analysis with reasonabl ~ confidence intervals have been drawn ~ystematically by taking every nth (second, third, fourth,. etc.) page in each cata10g or every nth Nationale in the bound volumes, extending all the way through the alphabet. Nonmatriculated students have been excluded,· since they were not subject to all the normal admissions· requirements. and were not eligible to take the principal state examinations or degrees. Strictly speaking, the samples of student registrations from the catalogs · are "cluster samples" since each page in a catalog included a number of individual registrations~ Nonetheless, for statistical purposes the samples from the catalogs and Nationalen have all been treated as direct random samples since for both kinds of records a student with a last name beginning with any letter· in the alphabet had· an equal chance of inclusion and, with respect to the registrations recorded in the catalogs, listing on a particular page had no connection with any of the variables being measured. Many scholars prefer ,to draw samples from a run of records covering a period of years in order to avoid the possibility of temporary circumstances that might skew the data for a particular year or semester. Konrad H. jarausch, for instance, drew a systematic sample from the Bonn University records for the· whole period 1865~1914 (see jarausch 1979, 1982). That approach, however, makes it difficult to develop any reliable time series analysis unless one draws a very large sample running through the whole period and/or presents temporal comparisons for whole decades, as jarausch does. In this study 293 294 APPENDIX B limitations on access to the manuscript registrations and the sheer volume of the materials dictated selecting the four discrete semesters. Within their respective eras, the years 1879-80, 1899-1900, and 190910 do not appear to have been atypical for Austrian demography or the educational system over the whole period. Austrian university enrollments in the 1850s, however, were erratic. Winter 1859-60 may show some of the aftereffects of the "Hungry Forties," the economic conjuncture of 1857-58, and the Austrian mobilization for the war in northern Italy that occurred in the spring of 1859. John E. Craig points out that students who studied longer and thus registered more often during their academic careers than did others can introduce a bias into social analyses based on registration records (Craig 1983, 226-27). Craig argues that analyses should be based on the total numbers of matriculations during any given period of time rather than the number of registrations so as to correct for the effects of some students remaining in·school longer or changes in the length of academic programs. Statistics on the number of matriculations for each occupational or class category, however, cannot be easily derived from the records of registrations for each semester found in the catalogs and Nationalen of the Austrian institutions. Any attempt to adjust mathematically the number of registrations in order to derive statistics on· matriculations might produce anapproximatio'n that is no more reliable than the uncorrected registration statistics. Uncorrected registration statistics have therefore been used throughout this study. Occupational and Class Categories The classifications used here generally follow the occupational and class categories for nineteenth-century Germany and Austria offered in Hubbard 1977 and Hubbard and Jarausch 1979. As in census returns , the occupations of fathers or guardians reported in the student registration records were often vague, but no alternative source is readily available that would offer more precise information. One cannot be,sure of how large an enterprise or property was held by someone described" as a business owners (Kaufmann in German or obchodnik in Czech) or a farmer/peasant (Landwirt or Bauer in German , sedlak or rolnik in Czech). Also, the status and general qualifications are often unclear for many described as employees and "private . Statistical Methods 295 officials." For practical purposes in this study...