Who becomes a Grandparent – And When? Education and Fertility in Multigenerational Perspective

Full text

Who becomes a Grandparent – And When?
Education and Fertility in Multigenerational Perspective
Jan Skopek* & Thomas Leopold**
(Working paper, June 2016)
* Corresponding author
Trinity College Dublin, Department of Sociology
3 College Green, Dublin 2, Dublin
Ireland
+353 (87) 179 6238
skopekj@tcd.ie
** University of Amsterdam
Postbus 15508
1001 NA Amsterdam
The Netherlands
Abstract
This study examines educational differences in the demographic transition to grandparenthood.
Our data from the German Ageing Survey (N = 2,434 men and women born 1933-1948) offer
rich life history data on fertility across three family generations, allowing us to study
educational gradients not only in the transition to grandparenthood, but also in transitions to
higher parities of grandchildren. We analyze educational differences (a) in the chance of
becoming a grandparent, and (b) in the timing of grandparenthood. For the chance of
grandparenthood, we find a strong educational gradient only among West German women:
Three quarters of the lower educated become grandmothers, compared to only half of the higher
educated. These differences emerged primarily as a consequence of the high levels of
childlessness among highly educated West German women. For the timing of grandparenthood,
educational differences were larger compared to the timing of parenthood, and larger in the
West compared to the East of Germany. In West Germany, differences between lower and
higher educated people in the median age at grandparenthood amounted to 6 years among
grandmothers and 7 years among grandfathers, at least doubling the educational differences
observed for the median age at parenthood. Looking at higher parities, we find no substantial
differences by education in the chance and timing of further grandchildren. Our study adds a
demographic perspective on grandparenthood to a growing body of research being concerned
with its consequences.
Keywords: grandparenthood; fertility; education; gender; East and West Germany

Introduction
The importance of grandparenthood has been recognized in various research fields. These
include the salience of the grandparent role in older age (Landry-Meyer 2004; Szinovacz 1998;
Troll 1985); the rising importance of multigenerational family bonds (Bengtson 2001); the
contribution of grandparents to child care and other forms of intergenerational solidarity
(Aassve et al. 2012; Gray 2005; Hank and Buber 2009; Presser 1989; Saraceno 2011; Thomese
and Liefbroer 2013); the consequences of being a grandparent for individual-level outcomes
such as health and work (Arpino and Bordone 2015; Bavel and Winter 2013; Christiansen 2014;
Ku et al. 2012; Leopold and Skopek 2014; Minkler and Fuller-Thomson 1999); and the effects
of grandparents on social mobility (Chan and Boliver 2013; Chiang and Park 2014; Mare 2011;
Modin et al. 2012; Modin and Fritzell 2009; Warren and Hauser 1997).
Despite their merits, these studies share an important limitation: they condition on
grandparenthood. This is reflected in common strategies of sample selection. Analysts restrict
their studies to three-generational structures, selecting either grandparents (G1), their children
(G2), or their grandchildren (G3) as a starting point of the analysis. Although this approach is
adequate to address many research questions, it implicitly treats grandparenthood as an
exogenous factor – a phenomenon that is taken as given rather than as a process to be explored.
Recent demographic research has addressed this issue, investigating the demography of
grandparenthood in modern societies. These studies showed considerable heterogeneity in
transitions to grandparenthood across Europe and the US (Leopold and Skopek 2015a), and
strong delays in transitions to grandparenthood in both in Germany (Leopold and Skopek
2015b) and in Canada (Margolis 2016; Margolis and Iciaszczyk 2015).
Despite these recent advances, the demographic understanding of grandparenthood
remains limited. A notable gap of knowledge concerns educational differences in
grandparenthood. In light of educational gaps in the timing and level of fertility, it seems

obvious that the experience of grandparenthood is stratified. Importantly, in transitions to
grandparenthood, educational differences may be larger because they emerge as a combined
outcome of fertility in successive family generations. This suggests not only that higher
educated people become grandparents later in life, but also that they are less likely to experience
the transition at all.
The present study explored the demographic linkages between education and
grandparenthood. Specifically, we examined the role of education with regard to the question
of (a) who becomes a grandparent?, and (b) when do people become grandparents? These
questions are relevant to various strands of current grandparent research, given that the chance
of becoming a grandparent and the timing of this transition are vital factors to determine the
meaning, performance, and consequences of the grandparent role.
By studying the link between education and grandparenthood in Germany, our study takes
a comparative perspective. The German context presents an interesting setting to study
educational differences in the demography of grandparenthood because the legacy of the East-
West divide still offers sharp contrasts in various important aspects, including the timing and
level of fertility, kinship structure, career trajectories, and gender differences. We draw on data
from the German Ageing Survey, which offers rich life history data on fertility across three
family generations. Our sample is representative of Germans born between 1933 and 1948 and
includes information about grandchildren in high detail. These data allowed us not only to
examine educational gradients in the transition to grandparenthood, but also in transitions to
higher parities of grandchildren – an aspect that has not been addressed in previous
demographic research.
Background
Studying grandparenthood as a demographic transition requires conceptual clarification. First,
given that grandparenthood is the result of fertility processes in subsequent generations, it is

conditional on the first-order process of parenthood. Second, not only the age at first birth, but
also the timing and spacing of subsequent births and the total number of children shape the
chances of becoming a grandparent at different ages. Third, grandparenthood is a three-
generational event – a “counter-transition” (Hagestad and Burton 1986) that is not self-initiated,
but contingent on one’s children becoming parents. All of these aspects need to be considered
when studying education and grandparenthood in demographic perspective.
Education, fertility, and grandparenthood
Although the link between educational attainment and parenthood is well-studied, no study has
drawn the link to grandparenthood. At least three mechanisms operate in stratifying the
transition to grandparenthood along educational lines. First, educational enrolment and
attainment structure the timing of fertility. Higher educational investments are associated with
more time spent in educational institutions, which go along with delayed transitions to marriage
and parenthood (Blossfeld and Huinink 1991; Oppenheimer 1988). Educational expansion in
the second half of 20th century was associated with an ideational change (Surkyn and
Lesthaeghe 2004) that has been identified as a driving force in the postponement of fertility in
modern and post-modern societies (Billari et al. 2006).
Second, the timing of fertility usually depends on joint decisions in unions (Corijn et al.
1996). Unions are not formed randomly but assortatively along educational lines (Mare 1991).
As absolute rates of educational homogamy have been rising in all modern societies (Blossfeld
and Timm 2003), educational differences in fertility have been amplified through assortative
postponement. Third, parental resources shape children’s educational opportunities and life
chances through various paths of intergenerational transmission (Breen and Jonsson 2005).
Depending on the degree of intergenerational mobility in a society, children of higher educated
parents attain higher educational levels, giving rise to the first and second mechanism. From
the perspective of prospective grandparents, education entails first-order consequences for their

own fertility and second-order consequences for their children’s fertility through mechanisms
of social reproduction.
Taken together, these considerations lead us to expect, first, that the timing and
prevalence grandparenthood is stratified by educational level: higher educated people are less
likely to become grandparents, and if they do, they experience grandparenthood later in life;
second, given that grandparenthood reflects combined fertility processes over two generations,
we expected that educational differences in these characteristics are larger for grandparenthood
than for parenthood.
Gender differences
In view of the factors already stipulated, we also expect gender differences in the linkage
between education and grandparenthood. In the course of educational expansion, women
increased their educational participation at higher rates compared to men, more recently even
surpassing men in education (Blossfeld et al. 2015; DiPrete and Buchmann 2013). This trend
has not only increased the opportunity costs of children, but also prolonged stays education and
training institutions while postponing family and childbearing decisions (Blossfeld and Huinink
1991; Lappegård and Rønsen 2005). A relevant trend in the German context is the rise of
childlessness among higher educated women (Andersson et al. 2008). As a result, an
educational gap in childlessness has emerged among women but not among men, although
higher-educated men have also delayed their fertility (Schmitt and Winkelmann 2005).
Consequently, we expected the demographic linkage between education and grandparenthood
to be gendered: the chance of becoming a grandmother should depend more strongly on
educational level than the chance of becoming a grandfather.

East and West Germany
Studying the transition to grandparenthood in Germany requires attention to the socio-historical
context that has differentially shaped the fertility of East and West Germans over several
decades. Fertility regimes differed strongly between the East (former German Democratic
Republic, GDR) and the West of Germany. As a result of pro-natalistic socialist policy, the
former GDR was characterized by higher fertility and earlier parenthood (Kreyenfeld 2003,
2004). These historical differences between still bear an important legacy for contemporary
patterns of grandparenthood in Germany (Leopold and Skopek 2015b).
The same applies to education and women’s labor force participation. The planning
economy of the GDR was characterized by high educational enrolment and strong integration
of women into the labor market. Even though access to tertiary education and employment was
largely regulated by planning concerns, gender differences in enrolment were much smaller
compared to West Germany. As a result, there were hardly any gender differences in
employment rates (Matysiak and Steinmetz 2008). Moreover, for women in the pro-natalistic
regime of the GDR, combining work and family was the norm proliferated by extensive
provision of early childcare, very much in contrast to the traditional and familialistic system of
West Germany (Nazio and Blossfeld 2003). Accordingly, the timing of fertility in East
Germany was less strongly linked to the process of educational attainment. In view of these
differences, we expected educational differences in grandparenthood to be weaker and less
gendered in East Germany.
Data and Methods
Our empirical analysis was based on data from the third wave of the German Ageing Survey
(DEAS), conducted in 2008. The DEAS is a large-scale survey representative of older adults
aged 40 to 85 living in private households in Germany (Engstler and Motel-Klingebiel 2010).
In the third wave of the DEAS, comprehensive retrospective information about fertility and

grandparenthood has been collected. These data have already been used for earlier studies on
the transition to grandparenthood (Engstler and Menning 2005; Leopold and Skopek 2015a,
2015b). A major advantage of the DEAS data is that information is available not only on the
birth dates of all children, but also on the dates of birth of all grandchildren. The latter is
important for identifying not only the transition to grandparenthood, but also transitions to
higher parities of grandchildren. We restricted our analysis to data from the cross-sectional
sample of 6,205 first-time respondents recruited in 2008.
Our sample restrictions were largely modelled after those applied by Leopold and Skopek
(2015b), selecting individuals aged 60 to 75 at the interview who lived in East or West Germany
before reunification (birth cohorts 1933 to 1948; N = 2,473). An important difference is that we
focused not only on parents (i.e., those “art risk” of grandparenthood), but also included
childless persons in our analysis. For those who reported to have children, we included only
data on biological children and grandchildren born to those children. We dropped a small
number of cases that had missing or implausible data on own fertility, fertility of children,
educational level, or part of Germany (East or West). After all restrictions, our analytical sample
consisted of 2,434 individuals.
The DEAS provides data on highest educational attainment coded on the basis of the
ISCED scheme. Due to limited case numbers within more differentiated educational categories,
we dichotomized educational level into higher (tertiary level, ISCED 5-6) versus lower
education (non-tertiary, ISCED 1-4). This distinction captures meaningful differences with
regard to the purposes of our study: In the East and particularly in the West German context,
tertiary education was most strongly associated with the timing of childbearing, as it prolonged
the qualification process into the mid-twenties (Kreyenfeld 2004).
=== TABLE 1 ===

Table 1 provides a descriptive overview of our sample on different indicators of fertility.
Data are shown separately by part of Germany (East or West), gender, and educational level
(lower or higher). For the cohorts under study, we observed higher levels of tertiary education
for East Germans (about 10 percentage points higher share for both men and women).
Childlessness was more prevalent in West Germany, especially among higher-educated
women: almost 28% of higher educated women were childless at the interview date, compared
to about 13% of women with non-tertiary education. Conditioning on parenthood, however, we
observed no East-West differences in the average number of children.
To study educational differences in the demography of grandparenthood, we used a
population-based approach. In a first step, we estimated the chance of becoming a grandparent,
measured by the probability of grandparenthood at different ages separately for men and women
from East and West Germany. Non-parametric Kaplan-Meier estimates accounted for right-
censoring in the data. We calculated the cumulative probability function F(t) by inverting the
survival function S(t), representing the estimated proportion of individuals having experienced
grandparenthood at age t. Differences between educational groups were assessed in absolute
and relative terms. Absolute differences express the difference of the group-specific expectation
of grandparenthood at age t (educational level indicated in superscript): F(t)Higher – F(t)Lower.
Given that absolute differences in cumulative probability inevitably decline as the prevalence
of grandparenthood increases with age, we also studied relative differences by assessing the
factor difference in the cumulative odds of grandparenthood – the rate of grandparents versus
non-grandparents at age t: !""#$%& ' ( )$%&
*$%&'. We assessed educational differences in
terms of the ratio of education-specific odds:
!+ % ( )$%&,-./0
*$%&,-./0 1*$%&2345/0
)$%&2345/0
Compared to differences in probabilities, the odds ratio (OR) has the advantage of
providing a measure for the association between education and grandparenthood that is

independent of prevalence. It indicates the factor difference in the ratios of grandparents versus
non-grandparents up to time t for the group of lower educated compared to the group of higher
educated individuals. This measure allowed us to compare the educational stratification in
grandparenthood at various ages. Because the odds ratio measures the association between two
categorical variables independent of their marginal distributions, it also facilitates a comparison
of age-specific associations between educational level and fertility outcomes that result from
parenthood and grandparenthood, respectively.
In a second step we examined the link between education and higher parities of
grandparenthood. For these analyses we used the rich DEAS data on further grandchildren. We
examined three parity transitions – zero to one (parenthood until the birth of the first
grandchild), one to two (birth of the first grandchild until birth of the second grandchild), and
two to three (birth of the second grandchild until birth of the third grandchild). When
interpreting the results on higher parities, it is important to consider that samples get smaller
and increasingly selective towards earlier grandparenthood due to conditioning on previous
events.
In a third step, we examined the timing of grandparenthood. Similar to previous research
(Leopold and Skopek 2015a, 2015b; Margolis 2016), we conducted these analyses for the
subsample of parents being “at risk” of experiencing grandparenthood (N = 2,120). We assessed
the typical timing of grandparenthood by estimating the median age at the transition (i.e., the
age at which 50 percent of the risk population had experienced grandparenthood). For our
purposes, the main interest was in how median ages differed by educational level. Again, we
compared median ages at grandparenthood with median ages at parenthood.
We conducted all analyses separately by East and West Germany, gender, and educational
level. We did not use design weights because these weights were constructed in the same way
in which we stratified our analysis (part of Germany, gender, and education). Additional
analyses showed that our results were not affected by the use of weights.

Results
Education and the chance of grandparenthood: Who becomes a grandparent?
Kaplan-Meier estimates for rates of parenthood and grandparenthood at different ages are
presented in Figure 1. Table 2 (men) and Table 3 (women) show estimates at different ages
additionally including statistical tests and odds ratios. Estimates for grandparenthood are shown
in 5-yearly intervals, starting at age 40 and up to age 70. We chose these age bounds because
only few events occurred before age 40 and after age 70. We calculated all rates separately by
gender and educational level.
=== FIGURE 1 ===
=== TABLE 2 ===
=== TABLE 3 ===
A strong educational gradient in the chance of becoming a grandparent emerged among
women in West Germany (Table 3 and Figure 1): Among lower educated women, estimates
showed that approximately three in four (74 %) had become a grandmother by the age of 70.
Among higher educated women of the same cohorts, only one in two (48 %) had experienced
this transition by the age of 70. Moreover, the shape of the probability curve (Figure 1) indicates
that this share is unlikely to rise far above the 50 % mark, suggesting that half of higher educated
women of our study cohorts in West Germany never become grandmothers. The odds ratios
illustrated these educational differences in relative terms: the ratio of grandmothers per non-
grandmother in the population of lower educated women was three times higher than the
respective ratio for the population of higher educated women. More detailed analysis of
grandmotherhood at different ages show that educational gaps in West Germany were largest
at younger ages, decreased thereafter, but remained large even towards the end of the process.
The plots for parenthood show that part of this gap in grandmotherhood was due to
educational differences in parenthood, in particular in terms of childlessness. However, even

when conditioning the sample on mothers, differences in the chances of becoming a
grandmother in West Germany remained substantial: about 84% of lower educated mothers
were grandmothers at age 70, compared to only 68% of higher educated mothers (about a 2.5
factor difference in the odds of becoming a grandmother among mothers). High chi-square
values obtained from a global log-rank test for the equality of survivor functions indicated
significant educational differences both for rates of parenthood and for rates of
grandparenthood among women in West Germany. These findings support the expectation that
the educational gradients in fertility are magnified in processes of grandparenthood.
The findings for all remaining groups – East German women and men as well as West
German men – looked very different. The curves of lower educated and higher educated
individuals ultimately converged in each of these groups, indicating that differences in
prevalence were only temporary, emerging from differential timing of fertility in successive
generations. Among East German men, the higher educated were slightly more likely to become
a grandfather, whereas the reverse was true among West German men. None of these estimated
differences was different from zero at conventional levels of statistical significance.
Next we examined educational differences in transitions to higher parities of
grandchildren. Figure 2 illustrates Kaplan-Meier estimates for parities 1 to 3. Table 4 provides
estimates, statistical tests and odds ratios evaluated at two-yearly intervals for higher-order
parities. Results were largely consistent across all groups considered in the analysis. More than
75 percent of parents experienced the transition to grandparenthood (parity 1) within 40 years
after the birth of their first child. Approximately three quarters of those who became a
grandparent could expect a second grandchild within the next ten years (parity 2); among those
who had two grandchildren, approximately two thirds could expect a third grandchild within
the next ten years. In these higher-parity processes, we found only slight differences with
respect to education, gender, or part of Germany. Taken together, our findings show that
educational gradients in the chance of becoming a grandparent are largely limited to the initial

transition to grandparenthood among women in West Germany. In this group, however,
differences are substantial.
=== FIGURE 2 ===
=== TABLE 4 ===
Education and the timing of grandparenthood: When do people become grandparents?
In a second step of the analysis, we turned our attention from educational differences in the
chances of becoming a grandparent to educational differences in the timing of grandparenthood.
In these analyses, we restricted the sample to parents in order to estimate age at
grandparenthood among those who were at risk of experiencing the transition. Consistent with
previous studies, we calculated the typical age at grandparenthood by estimating the median
age – the age at which 50 % of the population at risk had experienced the event. Figure 3
presents the results for median ages (including confidence intervals). For ease of interpretation
we again included results for the timing of parenthood along with those for the timing of
grandparenthood.
=== FIGURE 3 ===
The results corroborate our expectation that the timing of grandparenthood is more
strongly stratified by education than the timing of parenthood. Half of higher educated West
German mothers were grandmothers at an estimated age of 61, compared to an estimated age
of 55 among lower educated mothers. This six-year gap in median age amounts to twice the
three-year gap found for age at motherhood. Similarly, among West German men the seven-
year gap in median age at grandfatherhood was much larger than the two-year gap in

fatherhood. In East Germany, educational gradients in the median age at grandparenthood were
less pronounced and closer to the age gap in parenthood, particularly among men. In a final
step of the analysis, we inspected median durations in years since previous fertility events for
different parities of grandparenthood (Table 5). In contrast to the differences observed for the
timing of the initial transition to grandparenthood, we found no substantial differences in the
median duration to higher parities.
Discussion
This study is the first to examine educational differences in the demographic transition to
grandparenthood. We focused on educational gradients in the chance of becoming a
grandparent and in the timing of grandparenthood. Rich data about grandchildren available in
the German Ageing Survey allowed us to study these aspects not only for the initial transition
to grandparenthood, but also for higher parities. In doing so, our study fills gaps of demographic
knowledge about one of the major role transitions of older age.
We expected, first, that higher educated people are not only less likely to become
grandparents, but also to experience grandparenthood later in life. Moreover, we expected that
educational differences in these characteristics are more pronounced for grandparenthood than
for parenthood. In view of long-standing differences across East and West Germany with regard
to family policies, education and labor market attachment of women, and fertility behavior, we
further hypothesized that these educational differences in the demography of grandparenthood
are less pronounced and less gendered in East Germany.
Our results provide qualified support for these expectations. For the chance of becoming
a grandparent, our most prominent finding concerned West German women. In this group, only
half of the higher educated became grandmothers – compared to three quarters of the lower
educated. Put into a larger life course context, this means that lower educated women’s chances
of becoming a grandmother were about equal to higher educated women’s chances of becoming

a mother. Complementary analyses of transitions to parenthood showed that these differences
emerged primarily as a consequence of childlessness among highly educated West German
women in the birth cohorts under study.
In contrast to the sharp educational gradient observed for West German women, we
found no educational differences in the chance of becoming a grandparent up to age 70 for all
remaining groups – East German women, West German men, and East German. Moreover, we
observed no substantial educational differences for higher-parity grandparenthood in any
subgroup studied. Across all groups, approximately three quarters of first-time grandparents
experienced the birth of a second grandchild; among these, two in three had a third grandchild
before our observation window was censored at the age of 70.
For the timing of grandparenthood, the analyses confirmed our expectations that
educational differences were larger compared to the timing of parenthood, and larger in the
West compared to the East of Germany. In West Germany, differences between lower and
higher educated people in the median age at grandparenthood amounted to six years among
grandmothers and seven years among grandfathers, at least doubling the educational differences
observed for the median age at parenthood. These findings illustrate how fertility delays are
magnified in multigenerational processes of demographic reproduction. They also demonstrate
how the legacy of historical differences between East and West Germany still shapes
contemporary transitions to grandparenthood. In the former GDR, higher fertility levels over
longer periods of time resulted not only in earlier parenthood and grandparenthood, but also in
a weaker linkage between educational attainment and fertility.
Similar to the findings on the chance of becoming a grandparent, we found no
educational differences in the timing of grandparenthood for higher-parity transitions. The
absence of notable differences beyond the initial transition to grandparenthood illustrates the
selectivity associated with conditioning on grandparenthood.

An important implication of this is that educational differences in the chances of
grandparenthood are implicitly canceled out of the analysis if samples a reduced to three-
generational structures – a common strategy in current grandparent research. Some analysts of
“grandparent effects” on social mobility have concluded, for example, that higher educated
grandparents improve their grandchildren’s chances for school success above and beyond the
contribution of parents (Modin et al. 2012). Yet, this focus on social reproduction disregards
educational differences in demographic reproduction. In this regard, our findings on the chances
of becoming a grandparent showed that higher educated women in West Germany were much
less likely to have grandchildren on which to pass on their cultural and economic capital. More
generally, we recommend that studies on the social reproduction of attributes such as education
also account for the demographic reproduction of individuals who carry those attributes,
especially when looking at three-generational structures.
Our descriptive findings on educational differences in the timing of grandparenthood
also have implications for other areas of current grandparent research. One provocative
hypothesis to test empirically, for example, is whether grandparenthood is “a blessing for the
rich and a curse for the poor.” This hypothesis derives from the idea that among those in higher
social positions, passages into the grandparent role are experienced later in life and in a possibly
less demanding life course context. These grandparents might reap the benefits of an enriching,
affirming, and rewarding role, leading to a more active lifestyle yet not overburdening them or
interfering with other activities. Among the disadvantaged, in contrast, grandparenthood at
younger ages is more common. Earlier transitions, in turn, might induce stresses and role
overload, thus potentially entailing adverse effects on well-being, mental health, and physical
health (Musil et al. 2011). In this regard, empirical research has documented that younger
grandparents are not only generally more likely to start and continue caregiving (Luo et al.
2012), but also that their support coincides more frequently with family crises and more often
involves co-residence with the grandchild as well as highly demanding grandparent roles as

primary or extensive secondary caregivers (Burton 1996; Goodman and Silverstein 2002).
Concurrent parenting responsibilities faced by younger grandparents might further add to these
pressures. An example in this regard is the uneven distribution of teenage pregnancies across
social positions (Miller and Benson 2001). Overall, these considerations suggest that not only
the timing of grandparenthood but also the life course conditions surrounding this transition to
are socially stratified, potentially inducing and exacerbating cumulative advantages and
disadvantages in older age. As we have looked only at educational differences in the typical
age at grandparenthood, further research should examine how these differences intersect with
constraints and opportunities in other life domains to shape the meaning, performance, and
consequences of the grandparent role
Finally, our findings illustrate the importance of societal, demographic, and historical
context in the study of fertility across multiple generations. Even within unified Germany,
differences between East and West were substantial not only in the chance and in the timing of
grandparenthood, but also in how these aspects were stratified along educational lines. In light
of the large international variation in found in recent research Leopold and Skopek (2015b),
this indicates the potential for future comparative research on the linkage between education
and fertility in multigenerational perspective.

References
Aassve, A., Arpino, B., & Goisis, A. (2012). Grandparenting and mothers’ labour force
participation: A comparative analysis using the Generations and Gender Survey.
Demographic Research, 27, 53–84. doi:10.4054/DemRes.2012.27.3
Andersson, G., Rønsen, M., Knudsen, L., Lappegård, T., Neyer, G., Skrede, K., et al. (2008).
Cohort Fertility Patterns in the Nordic Countries. Stockholm.
Arpino, B., & Bordone, V. (2015). Regular provision of grandchild care and participation in
social activities (No. 42). RECSM Working Paper.
http://www.upf.edu/survey/_pdf/RECSM_wp042.pdf
Bavel, J. Van, & Winter, T. De. (2013). Becoming a Grandparent and Early Retirement in
Europe. European Sociological Review, 29(6), 1295–1308.
Bengtson, V. L. (2001). Beyond the Nuclear Family: The Increasing Importance of
Multigenerational Bonds. Journal of Marriage and Family, 63(1), 1–16.
doi:10.1111/j.1741-3737.2001.00001.x
Billari, F. C., Liefbroer, A. C., & Philipov, D. (2006). The postponement of childbearing in
Europe: Driving forces and implications. Vienna Yearbook of Population Research, 1–17.
doi:10.1553/populationyearbook2006s1
Blossfeld, H.-P., & Huinink, J. (1991). Human Capital Investments or Norms of Role
Transition? How Women’s Schooling and Career Affect the Process of Family Formation.
American Journal of Sociology, 97(1), 143–168. doi:10.1086/229743
Blossfeld, H.-P., Skopek, J., Triventi, M., & Buchholz, S. (Eds.). (2015). Gender, Education
and Employment. An International Comparison of School-to-Work Transitions.
Celtenham, UK: Edward Elgar.
Blossfeld, H.-P., & Timm, A. (2003). Who Marries Whom? Educational Systems as Marriage
Markets in Modern Societies. Dordrecht: Kluwer.
Breen, R., & Jonsson, J. O. (2005). Inequality of Opportunity in Comparative Perspective:

Recent Research on Educational Attainment and Social Mobility. Annual Review of
Sociology, 31(1), 223–243. doi:10.1146/annurev.soc.31.041304.122232
Burton, L. M. (1996). Age Norms, the Timing of Family Role Transitions, and Intergenerational
Caregiving Among Aging African American Women. The Gerontologist, 36(2), 199–208.
doi:10.1093/geront/36.2.199
Chan, T. W., & Boliver, V. (2013). The Grandparents Effect in Social Mobility: Evidence from
British Birth Cohort Studies. American Sociological Review, 78(4), 662–678.
doi:10.1177/0003122413489130
Chiang, Y.-L., & Park, H. (2014). Do grandparents matter? A multigenerational perspective on
educational attainment in Taiwan. Social Science Research, 1.
doi:10.1016/j.ssresearch.2014.09.013
Christiansen, S. G. (2014). The association between grandparenthood and mortality. Social
science & medicine (1982), 118, 89–96. doi:10.1016/j.socscimed.2014.07.061
Corijn, M., Liefbroer, A., & Gierveld, J. D. J. (1996). It takes two to tango, doesn’t it? The
influence of couple characteristics on the timing of the birth of the first child. Journal of
Marriage and Family, 58, 117–126. doi:10.2307/353381
DiPrete, T. A., & Buchmann, C. (2013). The Rise of Women - The Growing Gender Gap in
Education and What It Means for American Schools. New York: Russell Sage Foundation.
Engstler, H., & Menning, S. (2005). Transition to grandparenthood in Germany: Historical
change in the prevalence, age and duration of grandparenthood. 7th European
sociological association conference.
Engstler, H., & Motel-Klingebiel, A. (2010). Datengrundlagen und Methoden des Deutschen
Alterssurveys (DEAS). In A. Motel-Klingebiel, S. Wurm, & C. Tesch-Römer (Eds.),
Altern im Wandel. Befunde des Deutschen Alterssurveys (DEAS). (pp. 34–60). Stuttgart:
W. Kohlhammer.
Goodman, C., & Silverstein, M. (2002). Grandmothers raising grandchildren: family structure

and well-being in culturally diverse families. The Gerontologist, 42(5), 676–89.
Gray, A. (2005). The Changing Availability of Grandparents as Carers and its Implications for
Childcare Policy in the UK. Journal of Social Policy, 34(4), 557–577.
Hagestad, G. O., & Burton, L. M. (1986). Grandparenthood, Life Context, and Family
Development. American Behavioral Scientist, 29(4), 471–484.
doi:10.1177/000276486029004008
Hank, K., & Buber, I. (2009). Grandparents Caring for their Grandchildren: Findings From the
2004 Survey of Health, Ageing, and Retirement in Europe. Journal of Family Issues,
30(1), 53–73. doi:10.1177/0192513X08322627
Kreyenfeld, M. (2003). Crisis or Adaptation–Reconsidered: A Comparison of East and West
German Fertility Patterns in the First Six Years after the “Wende.” European Journal of
Population, 19, 303–329.
Kreyenfeld, M. (2004). Fertility Decisions in the FRG and GDR: An Analysis with Data from
the German Fertility and Family Survey. Demographic Research, (Special Collection 3:
Article 11), 275–318. doi:10.4054/DemRes.2004.S3.11
Ku, L. J. E., Stearns, S. C., Van Houtven, C. H., & Holmes, G. M. (2012). The health effects
of caregiving by grandparents in Taiwan: An instrumental variable estimation. Review of
Economics of the Household, 10(2012), 521–540. doi:10.1007/s11150-012-9154-9
Landry-Meyer, L. (2004). An Exploration of the Grandparent Caregiver Role. Journal of
Family Issues, 25(8), 1005–1025. doi:10.1177/0192513X04265955
Lappegård, T., & Rønsen, M. (2005). The Multifaceted Impact of Education on Entry into
Motherhood. European Journal of Population, 21(1), 31–49. doi:10.1007/s
Leopold, T., & Skopek, J. (2014). Gender and the Division of Labor in Older Couples: How
European Grandparents Share Market Work and Childcare. Social Forces, 93(September),
63–91. doi:10.1093/sf/sou061
Leopold, T., & Skopek, J. (2015a). The Demography of Grandparenthood: An International

Profile. Social Forces, 94(2), 801–832. doi:10.1093/sf/sov066
Leopold, T., & Skopek, J. (2015b). The Delay of Grandparenthood: A Cohort Comparison in
East and West Germany. Journal of Marriage and Family, 77(2), 441–460.
doi:10.1111/jomf.12169
Luo, Y., LaPierre, T. A., Hughes, M. E., & Waite, L. J. (2012). Grandparents Providing Care
to Grandchildren: A Population-Based Study of Continuity and Change. Journal of Family
Issues, 33(9), 1143–1167. doi:10.1177/0192513X12438685
Mare, R. D. (1991). Five Decades of Educational Assortative Mating. American Sociological
Review1, 56(1), 15–32.
Mare, R. D. (2011). A multigenerational view of inequality. Demography, 48(1), 1–23.
doi:10.1007/s13524-011-0014-7
Margolis, R. (2016). The Changing Demography of Grandparenthood. Journal of Marriage
and Family, (Article first published online, 14 MAR 2016). doi:10.1111/jomf.12286
Margolis, R., & Iciaszczyk, N. (2015). The changing health of Canadian grandparents.
Canadian Studies in Populationan, 42(3–4), 63–76.
Matysiak, A., & Steinmetz, S. (2008). Finding Their Way? Female Employment Patterns in
West Germany, East Germany, and Poland. European Sociological Review, 24(3), 331–
345. doi:10.1093/esr/jcn007
Miller, B. C., & Benson, B. (2001). Family Relationships and Adolescent Pregnancy Risk: A
Research Synthesis. Developmental Review, 21(1), 1–38. doi:10.1006/drev.2000.0513
Minkler, M., & Fuller-Thomson, E. (1999). The Health of Grandparents Raising Grandchildren.
American Journal of Public Health, 89(9), 1384–1389. doi:10.3928/00989134-20140219-
01
Modin, B., Erikson, R., & Vagero, D. (2012). Intergenerational Continuity in School
Performance: Do Grandparents Matter? European Sociological Review, 29(4), 858–870.
Modin, B., & Fritzell, J. (2009). The long arm of the family: Are parental and grandparental

earnings related to young men’s body mass index and cognitive ability? International
Journal of Epidemiology, 38(3), 733–744. doi:10.1093/ije/dyp001
Musil, C. M., Gordon, N. L., Warner, C. B., Zauszniewski, J. a, Standing, T., & Wykle, M.
(2011). Grandmothers and caregiving to grandchildren: continuity, change, and outcomes
over 24 months. The Gerontologist, 51(1), 86–100. doi:10.1093/geront/gnq061
Nazio, T., & Blossfeld, H. (2003). The Diffusion of Cohabitation among Young Women in
West Germany , East Germany and Italy. European Journal of Population, 19, 47–82.
doi:10.1023/A:1022192608963
Oppenheimer, V. K. (1988). A Theory of Marriage Timing. American Journal of Sociology,
94, 563. doi:10.1086/229030
Presser, H. B. (1989). Can We Make Time for Children? The Economy, Work Schedules, and
Child Care: Population Association of America, 1989 Presidential Address. Demography,
26(4), 523. doi:10.2307/2061256
Saraceno, C. (2011). Childcare needs and childcare policies: A multidimensional issue. Current
Sociology, 59(1), 78–96. doi:10.1177/0011392110385971
Schmitt, C., & Winkelmann, U. (2005). Wer bleibt kinderlos? Sozialstrukturelle Daten zur
Kinderlosigkeit von Frauen und Männern (No. 473). DIW-Diskussionspapiere.
Surkyn, J., & Lesthaeghe, R. (2004). Value orientation and the Second Demographic Transition
(SDT) in northern, western and southern Europe: An update. Demographic Research,
10(SUPPL. 3), 45–86. doi:10.4054/DemRes.2004.S3.3
Szinovacz, M. E. (1998). Grandparent Research: Past, Present, and Future. In M. E. Szinovacz
(Ed.), Handbook on Grandparenthood (pp. 1–20).
Thomese, F., & Liefbroer, A. C. (2013). Child Care and Child Births: The Role of Grandparents
in the Netherlands. Journal of Marriage and Family, 75(April), 403–421.
doi:10.1111/jomf.12005
Troll, L. E. (1985). The Contingencies of Grandparenting. In V. L. Bengston & J. F. Robertson

(Eds.), Grandparenthood (pp. 135–149). Beverly Hills: Sage.
Warren, J. R., & Hauser, R. M. (1997). Social Stratification across Three Generations: New
Evidence from the Wisconsin Longitudinal Study. American Sociological Review, 62(4),
561. doi:10.2307/2657426

Table 1 Descriptive statistics on the total sample, parents, and grandparents by part of
Germany, gender, and educational level
West
East
Women
Men
Women
Men
Sample
Lower
Higher
Lower
Higher
Lower
Higher
Lower
Higher
Total
(N=2,434)
Year of birth
1940.2
(4.4)
1941.8
(4.5)
1939.3
(4.3)
1940.2
(4.3)
1939.8
(4.3)
1941.4
(4.3)
1939.4
(4.4)
1940.0
(4.1)
Childless (%)
12.4
27.9
13.3
14.0
7.9
9.1
14.5
9.1
Grandparent (%)
70.5
42.9
66.7
56.1
78.9
74.4
72.3
76.4
N
596
140
528
314
280
121
235
220
Parents
(N=2,120)
Average number
of children
2.2
(1.2)
2.1
(0.9)
2.2
(1.1)
2.1
(0.9)
2.2
(1.2)
2.0
(0.9)
2.2
(1.1)
2.1
(1.1)
Average age
at first birth
24.2
(4.1)
28.4
(5.1)
27.1
(5.0)
28.6
(5.6)
23.0
(3.9)
23.9
(3.3)
25.2
(4.7)
26.3
(4.6)
Grandparent (%)
80.5
59.4
76.9
65.2
85.7
81.8
84.6
84.0
N
522
101
458
270
258
110
201
200
Grandparents
(N=1,657)
Average number
of grandchildren
3.4
(2.5)
2.8
(1.9)
3.2
(2.3)
3.2
(2.1)
3.6
(2.5)
2.9
(1.8)
3.1
(1.9)
3.0
(1.8)
N
420
60
352
176
221
90
170
168
Notes: Men and women born from 1933 to 1948. Parents and grandparents are sub-samples of the total sample.
Standard deviation in parentheses. Lower = educational level less than ISCED 5 (non tertiary); Higher =
educational level ISCED 5 or 6 (tertiary). Data: DEAS (German Ageing Survey, 2008).

Table 2 Men’s cumulative probability of being father and grandfather at different ages by part of Germany and
educational level
East
West
Fathers
Grandfathers
Fathers
Grandfathers
At age
Lower
Higher
OR
Lower
Higher
OR
Lower
Higher
OR
Lower
Higher
OR
20
9
5
2.06†
–
3
2
1.93
–
25
51
45
1.30
–
36
24
1.82*
–
30
76
76
0.99
–
69
61
1.43*
–
35
83
86
0.82
–
83
79
1.29
–
40
85
90
0.61†
2
1
2.37
85
83
1.16
1
0
2.39
45
85
91
0.57†
26
16
1.75*
86
84
1.16
6
2
3.53*
50
86
91
0.59†
48
40
1.39†
87
86
1.09
20
8
2.85*
55
86
91
0.59†
63
56
1.29
87
86
1.09
39
21
2.38*
60
86
91
0.59†
69
70
0.97
87
86
1.09
56
36
2.34*
65
86
91
0.59†
72
75
0.86
87
86
1.06
64
51
1.69*
70
86
91
0.59†
73
79
0.74
87
86
1.06
69
63
1.30
c
2
0.04
0.14
4.73*
17.26*
Notes: Estimated probabilities expressed in percentages. Values obtained by Kaplan-Meier estimation. OR = odds ratio – factor by which
odds of grandparenthood (grandparents/non-grandparents) of lower educated differ from those of higher educated; local statistical test of
proportions based on standard errors obtained from Greenwood’s approximation.
c
2 of a global statistical test of equality of survivor
functions (log-rank test). Lower = educational level less than ISCED 5 (non tertiary); Higher = educational level ISCED 5 or 6 (tertiary).
† p <.10, * p < .05.

Table 3 Women’s cumulative probability of being mother and grandmother at different ages by part of Germany
and educational level
East
West
Mothers
Grandmothers
Mothers
Grandmothers
At age
Lower
Higher
OR
Lower
Higher
OR
Lower
Higher
OR
Lower
Higher
OR
20
25
15
1.94*
–
15
1
12.27*
–
25
75
64
1.62*
–
59
23
4.94*
–
30
89
86
1.27
–
81
48
4.76*
–
35
91
91
0.98
–
87
66
3.26*
–
40
91
91
1.07
9
3
2.74*
87
70
2.98*
3
0
n.e.
45
92
91
1.17
36
22
1.93*
88
72
2.72*
13
1
9.93*
50
92
91
1.17
59
43
1.90*
88
72
2.72*
29
4
9.13*
55
92
91
1.17
70
64
1.29
88
72
2.72*
46
16
4.63*
60
92
91
1.17
77
70
1.40
88
72
2.72*
60
34
2.88*
65
92
91
1.17
79
76
1.21
88
72
2.72*
70
43
3.02*
70
92
91
1.17
79
76
1.21
88
72
2.72*
74
48
3.07*
c
2
3.16†
4.14*
48.58*
38.34*
Notes: Estimated probabilities expressed in percentages. Values obtained by Kaplan-Meier estimation. OR = odds ratio – factor by which
odds of grandparenthood (grandparents/non-grandparents) of lower educated differ from those of higher educated; local statistical test of
proportions based on standard errors obtained from Greenwood’s approximation. n.e. = not estimable.
c
2 of a global statistical test of
equality of survivor functions (log-rank test). Lower = educational level less than ISCED 5 (non tertiary); Higher = educational level ISCED
5 or 6 (tertiary).
† p <.10, * p < .05.

Table 4 Probability of having a second and third grandchild conditional on lower parities by gender, part of
Germany and educational level
East
Parity 2
1st grandchild → 2nd grandchild
Parity 3
2nd grandchild → 3rd grandchild
Men
Lower
(n=170,
e=134)
Higher
(n=168,
e=134)
OR
Lower
(n=134,
e=92)
Higher
(n=133,
e=96)
OR
Years since
birth of a grandchild
2
33
34
0.95
33
34
0.97
4
58
56
1.11
52
54
0.91
6
71
68
1.14
60
61
0.94
8
74
73
1.03
65
67
0.92
10
77
79
0.93
68
68
1.00
12
78
83
0.75
70
75
0.79
c
2
0.05
0.64
Women
Lower
(n=221,
e=179)
Higher
(n=90,
e=65)
OR
Lower
(n=177,
e=129)
Higher
(n=65,
e=42)
OR
Years since
birth of a grandchild
2
34
29
1.27
35
31
1.20
4
57
56
1.04
55
44
1.58
6
70
62
1.42
65
54
1.57
8
77
65
1.82*
68
56
1.67
10
81
68
1.97*
69
62
1.42
12
82
71
1.83†
72
67
1.29
c
2
2.02
1.26
West
Parity 2
1st grandchild → 2nd grandchild
Parity 3
2nd grandchild → 3rd grandchild
Men
Lower
(n=352,
e=279)
Higher
(n=176,
e=140)
OR
Lower
(n=278,
e=178)
Higher
(n=140,
e=87)
OR
Years since
birth of a grandchild
2
35
49
0.56*
36
37
0.95
4
65
76
0.58*
52
51
1.03
6
74
82
0.60*
58
62
0.83
8
79
86
0.58*
63
68
0.80
10
82
89
0.57†
66
71
0.81
12
85
90
0.58
68
71
0.89
c
2
9.11*
0.34
Women
Lower
(n=419,
e=344)
Higher
(n=60,
e=47)
OR
Lower
(n=344,
e=231)
Higher
(n=47,
e=25)
OR
Years since
birth of a grandchild
2
38
43
0.81
31
44
0.57
4
64
65
0.98
49
50
0.96
6
78
75
1.16
58
55
1.13
8
82
82
0.99
65
60
1.23
10
84
85
0.96
70
74
0.83
12
86
85
1.13
72
74
0.92
c
2
0.25
0.07
Notes: Estimated probabilities expressed in percentages. Values obtained by Kaplan-Meier estimation. n = number of persons
at risk; e = number of events. Due to few cases of invalid data on birth dates of higher parity grandchildren, sample size n of
parity 3 does not always equal the number of events in parity 2 sample. OR = factor difference in odds of lower compared
with higher educated, local statistical test of proportions based on standard errors obtained from Greenwood’s approximation.
c
2 of a global statistical test of equality of survivor functions (log-rank test). Lower = educational level less than ISCED 5
(non tertiary); Higher = educational level ISCED 5 or 6 (tertiary).
† p <.10, * p < .05.

Table 5 Median time to parity level (in years) by gender, part of Germany, and educational
level
Parity 1
Parity 2
Parity 3
From birth of
1st child
1st grandchild
2nd grandchild
to birth of
1st grandchild
2nd grandchild
3rd grandchild
West
East
West
East
West
East
Mothers
Lower
31
25
3
4
5
4
Higher
33
27
3
3
4
6
Total
31
25
3
4
5
4
Fathers
Lower
29
24
3
3
4
4
Higher
33
26
3
4
4
4
Total
31
25
3
4
4
4
Notes: Calculations of median durations based on Kaplan-Meier estimation. Lower = educational level less than
ISCED 5 (non tertiary); Higher = educational level ISCED 5 or 6 (tertiary).

Fig. 1 Probability of parenthood and grandparenthood by gender, part of Germany, and education. Cumulated probability functions obtained by
Kaplan-Meier estimation
Grandfatherhood
Fatherhood
West
East
West
East
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
15 20 25 30 35 40 45 50 55 60 65 70 75
Age
Lower educated
Higher educated
Men
Grandmotherhood
Motherhood
West
East
West
East
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
15 20 25 30 35 40 45 50 55 60 65 70 75
Age
Lower educated
Higher educated
Women

Fig. 2 Sequential transition to first time, second time, and third time grandparenthood by gender, part of Germany and educational level. Cumulated
probability functions obtained by Kaplan-Meier estimation
Parity 2 grandchild
(time since birth of 1st grandchild)
Parity 3 grandchild
(time since birth of 2nd grandchild)
Parity 1 grandchild
(time since birth of first child)
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
0 5 10 15 20 25 30 35 40 45
Time since fertility event (years)
Lower educated
Higher educated
East Germany - Women
Parity 2 grandchild
(time since birth of 1st grandchild)
Parity 3 grandchild
(time since birth of 2nd grandchild)
Parity 1 grandchild
(time since birth of first child)
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
0 5 10 15 20 25 30 35 40 45
Time since fertility event (years)
Lower educated
Higher educated
East Germany - Men
Parity 2 grandchild
(time since birth of 1st grandchild)
Parity 3 grandchild
(time since birth of 2nd grandchild)
Parity 1 grandchild
(time since birth of first child)
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
0 5 10 15 20 25 30 35 40 45
Time since fertility event (years)
Lower educated
Higher educated
West Germany - Women
Parity 2 grandchild
(time since birth of 1st grandchild)
Parity 3 grandchild
(time since birth of 2nd grandchild)
Parity 1 grandchild
(time since birth of first child)
0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1
Estimated probability
0 5 10 15 20 25 30 35 40 45
Time since fertility event (years)
Lower educated
Higher educated
West Germany - Men

Fig. 3 Educational differences in median age at grandparenthood. Calculations of median ages and 95% confidence intervals based on Kaplan-
Meier estimation for parents
49
47
56
55
52
51
63
61
24
22
26
24
26
23
28
27
West
East
Women
Men
Women
Men
Lower
Higher
Lower
Higher
Lower
Higher
Lower
Higher
15 20 25 30 35 40 45 50 55 60 65
Median age
First child First grandchild