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

different levels of university degrees. However, wherever there is a differentiation
between master’s and bachelor’s degrees, this may provide problems for imputations. If
questionnaires only ask for respondents’ highest degree, this would mean we would
often only have the date they obtained their master’s degree. Then, we might risk
overestimating first birth risks for the ‘bachelor’s degree’ category. This corresponds to
the problems found for cases when we only have the last date respondents received a
vocational degree, described above. If enrollment in studies preparing for a master’s
degree is more common for those who have not already become mothers,
overestimations of risks of first birth for the bachelor’s degree category could likewise
occur.

4.3 Differences in estimates for school degrees between the imputed and original
histories

As mentioned earlier, the greatest deviation in the estimates using the imputed
compared to the original histories applies to the category ‘lower secondary degree.’
This is a school degree that is generally received after the ninth or tenth grade. For
respondents who do not subsequently go on to obtain a vocational degree, this is their
highest degree. In the imputed histories, they are the only ones who contribute exposure
time to the category ‘lower secondary degree.’ In the original histories though, other
respondents also contribute exposure time to this category. Respondents who do not go
directly from school to vocational training, but wait for some time before starting their
vocational training also contribute exposure time to the ‘lower secondary degree’
category during the gap between school and vocational training. As we can see in Table
A1a, of the exposure time originating from the ‘lower secondary degree’ category, only
75% is correctly allocated to that category in the imputed histories. People who held
that degree in the gap between school and vocational training or before reenrolling in
school were imputed to have been in vocational training or in school the whole time.
However, if people have a child after finishing school, this is likely to affect the
probability of their enrolling in vocational training. Those who do not have a child may
be more likely to enroll. If this is the case, people who do not have a child during their
‘lower secondary degree’ spell would be selectively taken out of the ‘lower secondary
degree’ category in the imputed histories, since they are more likely to go on to
vocational training and to be imputed as having been enrolled the entire time after
finishing school. For them, the gap during which they held a lower secondary degree
would be neglected. Indeed, the number of births removed from the ‘lower secondary
degree’ category is disproportionately low compared to the amount of exposure time
allocated away from that category in the imputed histories (Table A1a and Table A1c).

http://www.demographic-research.org

<<   21   22   23   24   25   26   27   >>

More intriguing information