512 -overall -
pattern of fertility ignoring all covariates except
a person's sex and birth cohort. In this first stage we
examine survival probabilities, age-specific transition
rates, and, for higher order births, duration-specific
transition rates. The goal here is primarily to describe
basic patterns of fertility of native German men and women
in different birth cohorts, but in more detail than
aggregate data permit. In the second stage we estimate
models of birth rates at different parities as a function of
selected covariates that describe a person's social
situation and that are hypothesized to affect these rates.
This paper reports findings of the first stage of research;
results of the second stage are given in a forthcoming
companion paper.
THE DATA
Analyses of individual dynamics of fertility require a
fairly unusual form of data: longitudinal data on fertility
of a sample of individuals with known characteristics. The
data we analyze were gathered in the 1981-1982 German Life
History Survey (GLHS) sponsored by Sonderforschungsbereich 3
and the Max-Planck-Institut fuer Bildungsforschung under the
The respondents were
leadership of Karl Ulrich Mayer.
selected through a sampling plan that was designed to give
a
representative picture of native Germans living in the
Federal Republic of Germany in 1981-1982 who had been born
in one of three periods: 1929-1931, 1939-1941, and
1949-1951.3
We refer to these as the 1930, 1940, and 1950
cohorts for short. The total sample size is 2171, roughly
half men and half women.
Each respondent was personally interviewed using a
questionnaire that focused especially on the individual's
previous life history. It collected detailed information on
both parents and siblings of respondents, and on their own
schooling, vocational training, employment, residential
history, marital history, and fertility history. Limited
