Do imputed education histories provide satisfactory results in fertility analysis in the Western German context?

Zabel: Imputed education histories and fertility analysis in the western German context

underestimation of the baseline. If in later cohorts more women obtain master
craftswomen’s degrees, using the date they received that degree might lead to a major
overestimation of the risk of first birth for the category ‘vocational degree,’ since
people who do not have a child and go on to gain a master craftswoman’s degree are
selectively taken out of the ‘vocational degree’ category in the imputed histories as
delineated above.

To summarize, it seems one reason deviations in the estimates for the vocational
degree category in the imputed, as compared to the original, histories are so small is
that gaps between successive vocational training spells are not very large. Although
23% of the people in this sample obtained more than one vocational degree (a
supplementary master craftswoman’s degree or an additional basic vocational degree),
gaps between vocational training spells were not large enough to lead to extensive
misallocation of exposure time. Not many people seem to obtain second vocational
degrees very late in their life course. In addition, in this cohort, it was very uncommon
to obtain a university degree after a vocational degree (only 4% of the respondents did
so). In countries where access to university education is less restrictive, imputations
using only the highest educational degree might lead to greater distortions in estimates
of first birth risks for lower educational degrees that people held before obtaining their
university degree. In addition, only a comparatively small proportion of this cohort held
a master craftswoman’s degree in addition to a basic vocational degree (8%).
Distortions of estimates based on imputed histories could grow larger for later cohorts if
more women go on to obtain this higher level vocational degree.¹⁰

Reentries into or continuations of education after respondents already obtained
their highest degree also did not prove to cause much bias when using the imputed
histories. The greater the amount of time respondents spend enrolled in education after
they are considered to have already obtained their highest degree, the greater the extent
of downward bias to be expected for those highest degree categories. Particularly for
the imputation that used the first date respondents obtained their highest degree, a

¹⁰ A further problem for imputations is that not everyone is enrolled in vocational training full-time before
obtaining a master craftswoman’s degree. Many are enrolled part-time while employed. If in later cohorts
more women go on to obtain a master craftswoman’s degree, this too would make our imputation method
problematic, since we always assume that people are enrolled in education full-time up until the point they
obtain their highest degree.

The original histories used here cannot give much insight as to the extent of part-time enrollment among those
aspiring to a master craftswoman’s degree. This is because educational enrollment was recorded somewhat
imprecisely for this group. For some who obtained a master craftswoman’s degree, enrollment was recorded,
while for others, it was not clear whether they were enrolled full-time or not and when their training spell
began. We considered those for whom we had no information on enrollment not to be enrolled in our
preparations of the original histories. This is not likely to have strongly affected the results, since the group of
people who obtained a master craftswoman’s degree is altogether quite small, and the imprecision only
applies to a fraction of them.

http://www.demographic-research.org

<<   19   20   21   22   23   24   25   >>

More intriguing information