Demographic Research: Volume 21, Article 6
well as post-secondary degrees4. A person’s current degree refers to the highest degree
the person has attained up to that point in time. Post-secondary degrees are generally
considered to be higher than school degrees. Table 2 shows estimation results for this
model specification using both imputed and complete histories. These results will be
discussed in the next section.
As mentioned earlier, the imputations conducted here are based on the method
presented by Hoem and Kreyenfeld (2006). They develop a method of imputation using
the respondent’s highest level of education at interview, the date of attainment, the time
the respondent completed school, and the starting date of the respondent’s first job. The
present study uses their imputational method as a starting point. However, for the
results to be more generally applicable, in contrast to Hoem and Kreyenfeld (2006), we
do not use the date the respondent finished school, unless a school degree was the
highest degree ever attained, nor do we use the starting date of the respondent’s first
job, as this information cannot be expected to be generally available in data from
surveys focusing on fertility histories. The general method of imputation is to consider
respondents to be continuously enrolled in education up until the point in time at which
they attain their highest degree. The remainder of this section explains how this was
done in detail.
We perform five different imputations based on different assumptions about the
type of information available from surveys that do not include complete education
histories. When asking for the highest level of education a respondent has attained by
the time of interview and the date this level was attained, a questionnaire generally will
not specify whether to give the first or the last date a degree at this level was attained.
However, as pointed out in the previous section, this can make a difference. Therefore,
education histories are imputed alternately assuming that the questionnaire asked for the
first or the last date the highest degree was attained. This was done by taking from the
complete histories either the first or the last date a respondent received her highest
degree, and imputing the histories from there.
As pointed out in the previous section as well, it can make a difference for models
of the effect of education on first birth risks whether the questionnaire uses simple or
more differentiated education categories. Thus, different imputations are performed
assuming each case. In addition, a further imputation of education histories is
undertaken assuming that no information on the date a respondent attained her highest
degree is surveyed at all.
4 ‘Lower secondary degree’ is our translation for 'Hauptschulabschluβ,' ‘advanced lower secondary degree’
refers to ‘mittlere Reife,’ upper secondary degree to ‘Hochschulreife’ or ‘Fachhochschulreife,’ ‘vocational
degree’ to 'Ausbildungsabschluβ,' and ‘university or college degree’ refers to 'Hochschulabschluβ' or
'Fachhochschulabschluβ. ’ An upper secondary school degree is generally required to enroll at a university.
http://www.demographic-research.org
145