Zabel: Imputed education histories and fertility analysis in the western German context
craftswoman’s degrees, or between a standard university degree and a doctoral degree,
(or, in different contexts, also between a bachelor’s and a master’s degree), we would
be more likely to get the last date the respondent ever received a degree by asking for
their highest degree than if the survey only included simple, undifferentiated degree
categories. For the empirical analyses, separate imputations are also performed
assuming that the questionnaire uses a simple vs. a more differentiated list of degree
categories.
3. Data and method of imputation of education histories
For the empirical analyses, data from the project “Education, Training, and Occupation:
Life Courses of the 1964 and 1971 Birth Cohorts in West Germany” is used3. The data
includes retrospective information on many different realms of the life course. We used
data for the cohort born in 1964, since first birth histories for this cohort were
sufficiently advanced by the time of the interview in 1998/99. We excluded respondents
who were not born in West Germany, since our objective is to determine how well
education histories can be imputed in the context of the West German education system.
We conducted the study for female respondents only. After excluding a small number
of respondents with missing first birth information, this left a sample size of 641
respondents, who gave birth to 471 first children during the observation period.
Our method of analysis is event-history analysis, where the dependent variable is
the rate of transition to first birth. The date of first birth was backdated by nine months,
to better account for the influence of educational status at the point in time the decision
to have a first child was made. First, complete education histories were prepared using
the detailed information provided in the survey. We used information from the
vocational training record file as well as the school education record file. Next, several
different imputations were conducted based on different assumptions about information
available in surveys lacking complete education histories.
The specification of the education variable chosen differentiates between
enrollment in school education, vocational training, and university education. The idea
here is to account for the social context connected with the type of educational
institution people are attending. For those not enrolled in education, several different
levels of educational attainment were included, encompassing both school degrees as
3 The project “Education, Training, and Occupation: Life Courses of the 1964 and 1971 Birth Cohorts in West
Germany” is part of the German Life History Study (GLHS), conducted by the Max Planck Institute for
Human Development, Center for Sociology and the Study of the Life Course (Berlin) in cooperation with the
Institute for Employment Research (IAB, Nürnberg). For documentation, see Hillmert et al. (2004).
144
http://www.demographic-research.org