Appendix A. Data Sets, Samples, and

Definition of Variables

Samples

For the quantitative analysis in this volume I used data from the Micro-

census of Austria for 1982, 1988, and 1996 (collected by the Austrian

Statistical Office in Vienna) and the 1992 Hungarian Mobility Survey

done by the Central Statistical Office. The data sets contain representa-

tive samples of households for each year and in each country.

For Austria, I used different surveys for each time point I study in this

book. For Hungary, however, I used a single data set to simulate cohorts

for earlier periods. Using cross-sectional data for a longitudinal argument

warrants a long explanation, especially because the Hungarian Central

Statistical Office had conducted similar mobility surveys in 1982 and

1972 as well. There was, however, one important, and not quite innocent,

omission in those previous years: information on party membership was

not included in the questionnaire. For this reason, I was forced to limit my

usage to the 1992 data set, where this singularly important variable is

available. If the 1992 data set is weighted to produce a random sample of

the population, then it is fair to assume that it is also a reasonable repre-

sentation of the population in 1982. Of course, the 1992 survey will have

missed people who died in the ten-year period between 1982 and 1992 and

thus are not included in the sample. To alleviate this problem I limited my

selection to individuals between the ages of twenty to fifty for the analysis

in each year. Restricting the upper age limit reduces the influence of the

differential mortality rates in different social groups. Nevertheless, the

results for 1972 are more tenuous than those for 1982, and they should be