3
THE AVAILABILITY AND INTERGENERATIONAL
STRUCTURE OF FAMILY RELATIONSHIPS
Pearl Dykstra and Kees Knipscheer
For many decadesafter Parsons(1943) launchedhis nuclear family theory,
family researchwas preoccupiedwith the viability of the modern family.
Theorists assumedthat disintegrationof the family was an unavoidableresult
of industrializationandurbanization.ManciniandBlieszner( 1989)summarize
'scientists
the argumentasfollows:
suchasWirth, Park, andBurgessbelieved
that the diversity of urban life necessarilyweakenedprimary relationship
cohesionand that the accompanyingsocial and geographicmobility was not
compatiblewithextendedfamily relationships'(p. 278). Many studiesfocused
on the question of whether members of different family generationswere
becoming estrangedand isolated from one another, and more generally,
whether families were disintegrating. Mancini and Blieszner comment;
'Unfortunately, researchcontinuesto pursue this line of work, even though
the questionhaslong beenanswered'.Researchhasrepeatedlyconfirmed that
older adults are not alienatedfrom their families (Troll, Miller, & Atchley,
1979).
The present chapter examinesfamily relationshipsin the Netherlands. It is
organizedas follows. First, we describethe generaldemographicand social
changeswhich were evident in recent decadesand that have had an impact
on the multigenerationalstructure of families. Next, using data from the
NESTOR-LSN survey we describethe generationalstructure of later-life
families. In what follows, we look more closelyat the availabilityof different
types of family relationships(parents,siblings, children, grandchildrenand
greatgrandchildren),and analyseinteractionpatterns.Our analysesof the
levels of interaction with siblings and children specifically consider the
structure of the families in which theserelationshipsare embedded.
C.P.M. Knipscheer, J. de Jong Gierveld, T.G. van Tilburg, & P.A. Dykstra (Eds.), Living
arrangementsand social networks of older adults. Amsterdam: VU University Press, i,995.
38
Pearl Dykstra and Kees Knipscheer
Structural changes in family relationships
One of the major demographicchangesduring the 20th century has been the
genesis of the multigenerational family as a statistically normative
phenomenon.It is no longer uncommon that family membersof three, four
and five different generationsare alive simultaneously.This developmentis
of course linked with the increasesin life expectancysince the beginning of
this century.Estimatesindicatethat, of thoseborn in the Netherlandsin 1980,
about 30% had a living great grandmotherin l99I (Langeveld, 1985); of
those over the age of 60 in France, one third belongedto a four-generation
family (Paillat, Attias-Donfut, Clement& Delbes, 1989).
If one takesparent-child relationshipsas the key-linkagesor building blocks
of the multigenerationalfamily (Hagestad,1984), then multi-generational
familiesarebuilt up of multiple 'overlapping'parent-childrelationships
. These
overlappingrelationshipsaretheunits constitutingthetypical intergenerational
concern and commitment within the modern family.
The increasingproportions of multigenerationalrelationshipsare not the only
structural changevisible in late twentieth century families in industrialized
societies. We would like to draw attention to three other characteristicsof
aging families.
Verticalization. Not only has the number of generationswithin families
changed, but also the shape of the multigenerational family has become
different. The original pyramidical structure,where few generationsare alive
at the same time and where the members of the younger generations
outnumberthosein the older generations,hasbecomea vertical structurewith
severalgenerationsof similar sizesand a small top-generation.
The verticalization of the multigenerationalfamily is the result of two recent
contractionhastakenplacein connection
changes.First, anintra-generational
with the decreasingbirth rate in the twentieth century. There is a contraction
in the number of horizontal family links: fewer membersper generationin
a family. This trend is consistentacrossmostwesternindustriahzedcountries.
The secondchangewhich led toward the verticalizationof the family is the
large inter-generational extension that has taken place as the result of
decreasingmortality. There are more generationsper family. Bengtson,
Cutler, Mangen, and Marshall (1985) showedthat in the early eightiesin the
United States,38% of people aged 65 and over had families consistingof
t
!
Intergenerationalstructureof farnily relationships
39
three generations,and another 36% four generations.Starting with the
youngestgeneration,Uhlenberg(1980)demonstrated,for the United States,
that under 1900 mortality conditions,only one-fourthof the children would
haveall grandparentsalive at birth; by I976 it actuallywas almosttwo-thirds.
The probability of three or four grandparentsalive when the child was 15
years old increasedfrom 0.17 to 0.55 (Watkins, Menken, & Bongaarts,
1987).
Verticalizatron is also evident in the interactions among family members.
Interactionsacrossgenerationshaveincreased,while thosewithin generations
have decreasedin frequency and intensity (Cicirelli, 1982; White &"
Riedmann, 1992), changeswhich are of course linked with the changing
that the intensity
structure.Knipscheer(1980)demonstrated
intergenerational
of intragenerationalcontact within families dependsupon the availability of
intergenerationalfamily ties. Using data on the frequency of interaction,
intimacy, and the exchangeof instrumental support, he showed that older
siblings (65+) in the Netherlandsinteract infrequently unless one of the
siblings is childless.
Economic independencybetweengenerations. Severalstudieshavedescribed
the substantial changes in the dependency relationships between the
generationswithin the family (e.g. Bengtson& Treas, 1980; Cherlin &
Furstenberg,1985; Knipscheer,1986;De Regt, 1993).Parents'dependency
on their children has diminished with the introduction of general pension
systems, and the expansionof private pensionsand social services. The
customaryprinciple that one must look after one's parentshas lost much of
its impact, although there are large differences between countries in this
respect(Council of Europe, 1984; for the Netherlandssee De Regt, 1993;
for the United StatesseeCallahan,1985; Ruffin, 1984).Nevertheless,there
is ample evidence of economic interdependence,as for example in studies
reporting that parents continue to provide financial support to their adult
children (e.g. down payment on a house, the first car). In recent decades,
relationship quality has gained importance over financial and material
obligations as the basis for intergenerationalinteractions(Hagestad,1992;
independence
is reported
Knipscheer,1990).The desirefor intergenerational
in attitude surveys. For example, older adults' responsesto questions
pertaining to housing preferences and eare arrangements in the 1994
Population Policy AcceptanceSurvey conductedin the Netherlands(Moors,
Beets, & Van den Brekel, 1995) show a generalunwillingnessto become
dependentupon adult children. Younger respondents,when askedabout their
40
Pearl Dykstra and Kees Knipscheer
ideas about life in old age, also indicatea preferenceto continueliving on
their own and to be caredfor by professionalswhen the need arises.
Living arrangements.When Lasslet(1965) publishedhis famousstudy Zfte
world we have lost, his concern was primarily with the three generation
household. He identified a nostalgia for the earlier three generationfamily
in western European family ideology: all family members living under one
roof, in one household;a picture of family integrationand harmony. On the
basisof detailedstudiesof the householdsituationin England in the sixteenth
and seventeenthcentury, Lasslett and his team of historical demographers
concluded that the idealized family household was a myth (Wall, 1995).
Children who married and started a family of their establishedtheir own
households.Children who did not marry remainedwith their parents.Some
did not marry until after the death of their parents.The nostalgiafor the
idealizedthreegenerationhouseholdis ratherpersistent,however.According
'a
to Laslett,this persistenceindicates world we havelost syndrome'.He calls
it a syndrome becauseit keeps people under the spell of a family pattern,
which in the westernEuropeanworld has alwaysbeenmore of an exception
than a rule.
Industrialization and urbanization have changedthe family, and in people's
minds it has moved away from the earlier idealizedpattern. Sincethe second
rapidly (Kobrin, 1976;Wolf,
World War the sizeof householdshasdecreased
in fertility, but also to
to
the
decline
1990). This is attributed not only
-married
preferenceson the part of adult children
or not- and their elderly
parents to live on their own (Burch & Matthews, t987; Pampel, 1983).
Particularly striking is the increase in the number, both relatively and
absolutely,of older adults living alone. In 1960 in the Netherlands(Prins,
1990), nine per cent of males aged65 and over, and23% of women in that
age categorylived alone.In1971 the figureshad risento 10 and28%, and
in 1987they were 15 and4l%, for malesandfemalesrespectively.According
to recentestimates(De Beer, De Jong, & Visser, 1993),22% of malesand
44% of femalesaged60 and over are living alone.
The previously described structural changesin the family have important
implicationsfor family life (Bumpass,1990).Membersof the youngestand
the oldest generationsmay be 50 up to 80 years apart, most are not tied by
coÍïrmon economic interests, and neither are they situatedin the same local
coÍnmunity. Daily interaction is no longer the dominantform of communication. However, in general, frequent contactsand interactionshave been
Intergenerational structure of family relationships
41
maintained(Dooghe,1970;Knipscheer,1990;Mancini & Bliesznet,1989).
Most families are characterizedby regular visiting patterns.There is much
concern for and interest in each other's well-being. In case of need there
appearsto be a strongwillingnessto supporteachother. In conclusion,there
is less involvementin eachothersdaily life but a high degreeof reciprocal
concern. Though this general pattern of family life has been confirmed
repeatedly in research in different countries, the level of interaction and
reciprocitytendsto vary considerablyacrossfamilies.Somefamiliesaremore
cohesive,othersmore individualistic.In addition,the compositionof families
has become more diverse becauseof divorce and/or remarriageof family
members,becauseof a consciousdecisionnot to marry or to remainchildless
(Riley, 1983).
Design of the study
Though thegeneral trendsinfluencing family structureare well-known, there
is little information on their implications at the micro level. In this chapter
we will explore the variability in the intergenerationalcompositionof the
Dutch family networks. Up to now information on the structure of Dutch
family networkshasbeenscarcelyavailable(Langeveld,1985)and basedon
estimatesonly. The NESTOR-LSN study is the first to produce basic,
descriptiveinformation on family compositionin the Netherlands.
In this chapter, the respondentsserve as the anchor-pointsof the family
networks. As mentioned earlier, we will start with a description of the
generationalstructureof late-lifefamilies,using informationprovidedby the
respondents
aboutthe availabilityof parents,siblings,children,grandchildren
'Parents' are the respondent'snatural mother and
and great grandchildren.
'siblings', 'children', 'grandchildren' and
father only. The availability of
'great grandchildren'is not restrictedto blood ties, but alsoincludesadoptive
and step relationships.The reasonfor focusing on a broader category of
family relationshipsis that we are primarily interestedin the availability of
potentially supportive tres rather than fertility per se.
Next, we will considereachgenerationseparately.We will look at the number
of surviving family members of a particular generation. With regard to
siblings and adult children, we will also look at the levelsof interaction,and
residential proximity. More particularly, we are interestedto know the
42
Pearl Dykstra and Kees Knipscheer
number of siblings and adult children with whom contact is relatively
frequent, and the number who are living relatively nearby.
To determine the frequency of interaction, for each sibling and child still
'How
often are you in touch with him/her
alive, the question was asked:
(either face-to-face,on the telephoneor in writing?'. Eight answercategories
'1' never, '8'
were used,with scoresranging
to
daily. Geographicproximity
was measuredin terms of travelling time. The respondentwas asked, again
'How much time doesit take you to go
for eachsibling and child still alive,
and visit him/her (for the meansby which you usually travel)?'. Travelling
time was measuredin minutes.
Not only will we provide descriptiveinformation on the frequencyof contact
with siblingsand with children,but we will alsoreport the resultsof analyses
of determinantsof theseinteractions.We were interestedto find out to what
extent the frequency of contact among family members varies according to
the intergenerationalfamily structureinwhichthe relationshipsareembedded.
The following substantiveconsiderationsguidedthe analyses.The first is the
notion of apreferential hierarchy within the family, with partner relationships
being most preferred,followed by children,parents,and siblings(cf. Cantor,
1979). Frequent interaction with a particular type of relationship is assumed
to be less likely if family relationshipshigher in the hierarchy are present.
Thus for example,contactwith children is assumedto be lessfrequentamong
older adults with a partner than among those who are single. Likewise,
contact with siblings is assumedto be less frequent among elderly parents
than among the childless.
A second notion focuses on the special functions of parents, and more
particularly on parents as linking pins within the family. The assumptionis
that interactions among siblings in particular, are more frequent if parents
are still alive. One reasonis that the parentalhome often servesas the locale
for family-related rituals, with parents (mothers) as organizersof the social
gatherings. Another is that parentsthemselvesmay be the focus of contact
among siblings: communications are guided by concern about parental
welfare.
A third notion concernsfamily size, with the assumptionof greaterselectivity
in large families, which in turn is reflectedin lower levels of contact. On the
one hand, thosewith a relativelylargenumberof siblingsor a relativelylarge
'afford' to be
number of children can
selective.and thus choosefor lower
lntergenerational structure of family relationships
43
levels of interactions.On the other hand, limits on time and emotionalenergy
may lead to less intensive contact in large families.
Analyses of variance were conductedto examine whether the frequency of
contact with siblings was associatedwith (a) the existenceof parents, (b) the
number of surviving siblings, (c) the parentalstatusof the anchor-person,(d)
the partner status of the anchor-person,and (e) the partner status of the
'Partner
status' refers to the presenceor absenceof a cohabitant.
sibling.
Furthermore, we examined whether the frequency of contact with children
was associatedwith (a) the existenceof the anchor-person'sparents,(b) the
number of surviving siblings, (c) the number of surviving children, (d) the
parental status of the child, (e) the partner statusof the anchor-person,and
(0 the partner statusof the child.
Both analysescontrolledfor geographicproximity and the anchor-persons's
age becausethey are well-known determinantsof the frequency of contact
with network members. To avoid problems of interdependenceamong the
data of siblings and children from the samefamily, one sibling and one child
were selectedat random. In other words, the analyseswere performed on only
one sibling or only one child per family. That way family level influences
are not confounded with determinants at the dyadic level. Siblings and
children who were membersof the anchorperson'shouseholdwere excluded.
Respondents
ln1992, face-to-faceinterviews were conductedwith 4494respondents.They
constituted a stratified random sampleof men and women born in the years
1903 to 1937. The random sample was taken from the registers of 11
municipalities: the city of Amsterdamand two rural communitiesin the west,
one city and two rural communitiesin the south, and one city and four rural
coÍnmunitiesin the eastof the Netherlands.The responsewas 61.7 per cent.
The data were collectedby 88 interviewers.
was 72.8. Most were living in their own
The averageageof the respondents
(28.9%)
werenot marriedand lived alone,2582 (57.5%)lived
homes: 1298
with a partner, and 206 (4.6%) lived in another kind of multi-person
household.Finally , 35I (7 .8%) lived in an institutionof some sort, such as
a nursing home, a home for the aged,psychiatric hospital, or shelter for the
homeless.
44
Pearl Dykstra and Kees Knipscheer
Weights
The datapresentedin this chapterhavebeenweightedto correctfor selective
non-response,with the exceptionof thoseused in the multivariateanalyses.
Descriptivedata pertainingto the entire samplehave beenweightedin such
a way that they are representative
of the Dutch populationof older adults of
55 yearsof ageandover. Thesedataarecontrolledfor the over-representation
of the oldest respondents,and
the over-representation
of males. Descriptive
data pertaining to different age groups have been controlled for the overrepresentationof males in the oldest age groups. Within eachage category,
the proportions of males and femaleshave beenmade consistentwith those
at the national level.
Results
Generationalfamily structure
Complete information on the generationalstructure of their families, that is
on the existenceof parents,children, grandchildrenand great grandchildren,
is availablefor 3780 respondents.How many one-, two-, three-, four-, and
five-generationfamilies are there?The findings are summarizedin Table 3. L
Older adults are members of one-generationfamilies if their parents are no
longer alive and if they have no offspring of their own: ll.9% are in this
situation. Another 15.2% are membersof two-generationfamilies, further
Table3.1. Generational
family structure(n - 3780)
number of generations
two
parents
*
anchor
children
grandchildren
great grandchildren
per c ent age I I.9
*
*
four
three
*
*
*
*
*
*
+
*
*
*
*
*
*
*
{
anchor outlived his/her children
4 . 3 4 9 . 90 . 4
<
*
*
*
-
*
*
1 .8 1 3 .4
five
*
{
*
*
8.1 9.0 0.1
.
*
*
0.4
lntergenerational structure of family relationships
45
distinguishedaccordingto the I.8% where the anchor-personis a member
of the youngestgeneration(i.e. there is no offspring) and the 13.4% wherc
he or she belongs to the oldest generation(i.e. the parentsare no longer
alive).
54.6% belongto three-generationfamilies.Here the most frequentlyobserved
structure is the one with the anchor-personas the grandparent, and with
surviving children and grandchildren:it appliesto 49.9% of older adults.
Only a small proportion (0.4 %) arc in a situation where they, as oldest
members of their families, have outlived their children, but do have
grandchildren.Another 4.3% are membersof the middle generationof a
three-generationfamily.
17.8% are membersof pur-generation famrlies.In 9.0% of thesefamilies
the anchor-personis the oldestmemberand eachgenerationhas at leastone
representative.Only a small proportion (0.I%) arein a situationwhere they,
as oldestmembersof their families, haveoutlivedtheir children,but do have
grandchildrenand great grandchildren.Another 8.7 % are membersof fourgeneration families, where not they, but their parents are the great
grandparents.
Finally, 0.4% are membersofy've-generationfamtlies.In thesefamilies the
anchor-personis the memberof the secondgeneration.Apart from the older
adults' surviving parents,there are the older adults' children, grandchildren
and great grandchildren.
So far, entire families have been taken into consideration.In doing so, we
have looked only at whether or not they include members of a particular
generation. No attention has yet been paid to differences in the number of
family relationshipsof a particular generation.In what follows, we will
examinethe generationsseparately,andbroadenour focusto include siblings.
Parents
Given currentmortality patternsandthe agerangeof our sample,one expects
particularlythosein the youngestage
a considerablenumberof respondents,
groups, to still have one or two living parents.Data on the survivorshipof
parentsis availablefor 3780 respondents.Of these, 13.2% have a surviving
mother or father, and 1.7 % still havetwo survivingparents.Table -J.2shows
Pearl Dykstra and Kees Knipscheer
Table3.2. Availability of parentsby age, in % (n - 3789)
age of the respondents
one
both
55-59
60-64
65-69
70-74
75-79
32.6
5.4
18.1
1.5
8.5
0.5
4.r
r.3
0.2
0.3
80-84
85-89
0.9
0.2
0.3
the differences by age. Not surprisingly, the likelihood that either parent is
alive is greatestamong the younger respondents.Of those in the 55-59 years
of age category, approximately 33% have at leastone parent alive, and more
than five per cent still have two living parents.Among thosein the oldest age
categories(80-89), less than 1 % have at least one parent alive.
Siblings
Severalrecentstudieshaveconfirmed earlier findings aboutthe relatively low
level of interactionbetweensiblings. 'Most measuresof sibling solidarity are
negatively affectedby having adult children or a living parent, suggestingthat
vertical ties occupy center stage in the lives of most adults' (White &
Riedmann, 1992). What do the LSN-data show?
First, we look at the availabilityof brothersandsisters.For 4135 respondents
we have information on the number of siblings they ever had and the number
still surviving. There are large differencesin the number of living siblings:
l2.4To have no siblings, 20.4% have one, 17.0% have two, 14.8% have
three, I0.4% have four, 7.1% havefive, whereasI8.0% havesix or more.
Approximately one-fourth of thosewith no living siblings were only children;
they never had brothers or sisters. The previous figures include natural as
well as step and adoptivesiblings. A large majority of 93.2% have natural
brothers and sistersonly, while 4.2% haveno naturalsiblingsbut only step
or adoptivebrothers and sisters.
Additional information on the availability of siblings is presentedin Figure
3.1. For eachof sevenagecategories,it showsthe proportion of respondents
who have at least a particular number of siblings. This number varies from
one to eight or more. Regardlessof the number that is considered,the figure
Intergenerationalstructureof farnily relationships
7o
4l
Numbcr of siblings
100
70-74
Figure 3.1. Availabilityof siblingsby age (n : 4135)
showsa near linear declinein the availabilityof siblingswith increasingage.
The youngest respondentsare least likely to have a relatively small number
of siblings and most likely to havea relatively large numberof siblings. The
situation for the oldest respondentsis the opposite:they are most likely to
have relatively few siblingsand leastlikely to haverelativelymany siblings.
Next, we look at the level of interaction.Of those with living siblings and
for whom data on the frequency of interaction is available (n - 3323), a
considerablenumber are not in touch frequently with any of their siblings:
6I.4% havecontacton a lessthan weeklybasis,and 27.6% haveno contact
with any of their siblings monthly or more often. In contrast stand the
respondentswho have intensive contactswith a relatively large number of
siblings: 3.5% have four or more siblings with whom they interactweekly
or more often, and L4.2% havefour or more siblingswith whom they interact
monthly or more often. What information do we have on the geographic
dispersionof siblings?Of thosewith living siblings and for whom relevant
dataare available(n - 3243),,33 .0 % havenonewithin a 3O-minutetravelling
Pearl Dykstra and Kees Knipscheer
distance,24.I% haveone, 14.5% havetwo, 9.5% havethree,while 18.9%
have four or more siblings living at that distance.Included in thesefigures
are the number of respondentssharinga householdwith one or more siblings
(1.0% of all older adults).
To what extent do we see differences in the frequency of interaction with
siblings that are associatedwith the intergenerationalstructureof the families
in which theserelationshipsare embedded?The resultsof an ANOVA based
on the data from 3147 respondentsindicate that such differencesexist (see
Table 3 .3). Controlling for geographicproximity and the age of the anchorperson, the data show that contactwith siblings is more frequent if (a) there
are few surviving siblings in the family, (b) the older adult is childless,and
Table 3.3. Frequency of interaction (I : never to
B : daily) withsiblings in associationwith
the generational compositionof thefamily
(multiple classification analy sis)
deviationa
parents alive
no
yes
# siblings
t-2
3-4
> 5
children alive
no
yes
partner status anchor
single
cohabiting
partner status sibling
single
cohabiting
.02
2786
36r
-.01
.08
r463
885
799
.t7
.00
-.31
420
2727
.19
-.03
rt23
2024
.16
-.09
950
2r97
.23
-.10
. 11 * *
.05*
.07**
.08**
a deviationfrom the grandmean(3.99) adjustedfor the
covariatesand the other independentvariables
R2 : 65%
*p1.01,**p(.001.
Intergenerationalstructureof family relationships
49
(c) if either the older adult or the sibling is living without a partner.Whether
or not parentsare alive makesno difference.Generallyspeaking,the likelihood that contactswith siblings are activateddependson the availability of
other family ties. The fewer the alternatives,the greater the likelihood of
frequent interaction.
Children
Information on the number of ever-born children and the number still
surviving is availablefor 4196respondents.
More than 85 %have oneor more
living children. There is a large variation in the number of children; I0.7 %
haveonly one living child, while 27.2% havetwo, 19.0% havethree, l2.I%
lravefour, 7 .5% havefive, and 9.l% havesix or more living children.Most
parents (96.0%) have natural children only; 2.5% have either step and
adoptive children togetherwith natural offspring, and I.5% have only step
or adoptivechildren. The caÍegoryof thosewithout any children consistsof
respondentswho have always been childless(13.4%) and a small minority
of 0.9% who no longer have surviving children.
Figure 3.2 shows the differencesin the number of surviving children for
respondentsin the different age categories.The youngest(i.e. those under
the age of 65) are least likely to have relatively large families consistingof
three or more surviving children. Here we seeevidence(Verhoef, 1989) of
the relatively marked decline in the Netherlandsin the averagenumber of
births per woman for the birth cohorts 1925-1945,from 2.8 to 2.0 (for
postwar cohorts a further decline to approximately 1.7 is expected).The
oldest respondentsare most likely to be without any children, primarily
becauserates of childlessnessare highest among the oldest cohorts (which
in turn is linked with lower proportionsever marrying, Liefbroer & De Jong
Gierveld, 1995), and to a lesserextentdue to outsurvival.
How intensiveis parent-childcontact?For 3504parentswe haveinformation
on the frequency of contact with their offspring. Only a small number seem
to be socially isolatedfrom their offspring: I.4% arc not in touch monthly
or more often and 8.0% are not in touch weekly or more often with any of
their adult children. As Table -1.4shows,the respondentsare generallywell
embeddedin family life: over 50% have three or more children with whom
they interact monthly or more often, and over 40% have three or more
children with whom they interaëtat least weekly.
50
Pearl Dykstra and Kees Knipscheer
Number of children
60-64
65-69
70-74
Figure 3.2.
Availabílity
8GE4
75-79
of children by age (n :
4196)
Table 3.4. Availability of children with whom
interaction takes place regularly,
in%(n:3504)
# children
0
1
2
3
> 4
monthly
or more often
weekly
or more often
r.4
8.0
20.2
31.5
19.2
2T.T
t4.4
31.0
22.2
31.0
lntergenerationalstructureof tarnily lelationships
51
The data also indicate that relatively few older adults are geographically
isolatedfrom their offspring. Of thosewith living children and for whom data
on the frequencyof interactionis available(n - 3422), I4.l% do not have
any child living within a 30-minute travelling distance,24.5% have one,
27 .3% havetwo, 15.6% havethree,while L8.5%havefour childrenor more
living at that distance.
To what extent is the frequency of interaction with children associatedwith
the availability of family members? At the bivariate level, significant
differences in the frequency of contact with children and the generational
compositionof the family are found. Bivariate analysesshow that contactwith
children is more frequent if one or both of the anchorperson'sparentsare
still alive, if the anchorhasfewer siblings,if the childrenhavefewer siblings,
if the children are parentsthemselves,if the anchor-personis living with a
partner, and if the child is living without a partner.However, all differences
but the one for family size, are no longer significant once introducedin a
rnultivariateanalysis(seeTable3.5 for details).ANOVA-results (n : 287I)
indicatethat (apartfrom geographicproximity and the anchorperson'sage)
family size is the only significantpredictor of the frequencyof contactwith
children. The more siblings the child has, the less frequent the contact
between the child and the parent (anchor-person)tends to be. This finding
suggeststhat parentsand children in small families createspecialbondswith
a high level of social connectedness.
Grandchildren and great grandchildren
Information on the existence of grandchildren and great grandchildren is
available for 4137 respondents.68.8% have at least one grandchild, and
I0.4% have at least one great grandchild. Not surprisingly, the likelihood
of being a (great) grandparentis strongly linked with age. As Figure 3.3
shows,the proportionsof grandparentsincreasesteadilywith eachsuccessive
age category,but drop among the 80 and over category. That drop is linked
with the relativelyhigh rateof childlessness
in theparticularcohorts.Between
20 and 25% of women born in the period 1905-1915remainedchildless
(Liefbroer & Gierveld, 1995).Figure 3.3 also showsa steadyincreasein the
proportions of greatgrandparentswith increasingage. Of course,the older
one is, the greaterthe likelihood of having grandchildrenwho are in or past
their childbearingyears.
52
Pearl Dykstra and Kees Knipscheer
Table 3.5. Frequency of interaction (1 : never to
I : daily) with children in assocíation with
the generational compositíon of the family
(multiple classification analy sis)
deviationa
parentsalive
no
yes
# siblingsanchor
r-2
3-4
> 5
# siblings child
1
2-3
> 4
grandchildren
no
yes
partner status anchor
single
cohabiting
partner status child
single
cohabiting
.01
2574
36r
.00
-.04
1306
834
731
.05
.00
-.06
r205
1041
625
.25
-.07
-.37
818
2035
-.07
.03
97r
1900
-.01
.00
419
2452
.t2
-.02
.03
.15*
.03
.00
.03
deviation from the grand mean (6.12) adjusted for
the covariates and the other independentvariables
R2 : n.L%
p < .001.
While information on the frequencyof interactionand residentialproximity
is available for each surviving sibling and child, this is not the case for
grandchildren and great grandchildren. Information on interactions with
grandchildren was obtained for a subsample of the respondents only
(n : 936). Detaileddescriptionsof the importanceof grandparenthood
will
be consideredin future publications.
53
Intergenerationalstructureof family relationships
great grandchildren
5G.59
70-74
75-79
Figure3.3. Pressenceof grandchildren and great grandchildren by age (n : 4137)
Conclusion
Our data indicatethat Dutch older adults are most likely to be membersof
three-generationfamilies, and more precisely to belong to the oldest
generation:that of the grandparents,with eachgenerationrepresentedby at
least one surviving member. Almost half of the older adults are in that
situation. Five-generationfamilies are relatively rare; our data indicatethat
less than one per cent of older adults in the 55-89 age category in the
Netherlandsbelongto sucha family. It is not unlikely that a higherproportion
of five-generationfamilies would have been found if those aged 90 and over
had also been membersof the NESTOR-LSN sample.
The findings on the geographicdispersionof family memberscan be said to
be typically Dutch. The Netherlandsis a small, denselypopulated country,
54
Pearl Dykstra and Kees Knipscheer
and the Dutch are geographicallynot very mobile. Relativelyunique to the
Netherlandsis thereforethat the majority of older adultslive neartheir family
members.Closeto 40 % ofthosewith living siblingshaveone or two siblings
living within a 30-minute travelling distance, while 85 % of those with
surviving offspring have at least one child living within that distance. One
can perhaps characterize the circumstancesin the Netherlands as being
conduciveto high levels of family interaction.
Multigenerational families have becomemore prevalentin recentdecadesas
the result of increasinglongevity. More and more members of the oldest
generationsare surviving to advancedages.Fertility patternsalsoaffectfamily
structures: births are the advent of a new generation, and their timing
influences the likelihood that members of multiple generations are alive
simultaneously.Timing determineshow fat apartsuccessivegenerationsare
in terms of age, and thus the likelihood of co-survival.Recentestimatesfor
the Netherlandsindicatethat the impact of decreasedfertility on the likelihood
of becoming a grandparentwill be visible only in the extendedfuture (Prins,
1994). Of those born in the 1930's 84% are expectedto have at least one
grandchild,a figure which is consistentwith that for older cohorts.However,
25% of the 1965-birthcohortare estimatedto never have any grandchildren
if the fertility pattern of their children is similar to their own.
One of the outcomesof the increasesin longevity is that family membersare
spending more and more years together. As yet, the implications of colongevity are unclear. What is the significance for the quality of
intergenerationalties?Inwhat waysareintergenerationalcommitmentsshaped
by sharing a large number of years together?Some (e.g. Arling, 1976)
emphasizethe intergenerationaldifferencesinperspectivesandinterestswhich
are linked with having grown up in different time periods and occupying
different positionsin the life course.Others (e.g. Bengtsonet al., 1985)
emphasizethe intrafamilial similarities acrossgenerationsas regardssocial
values and attitudes.Do thesedifferencesand similaritieslose significance
as the members of successivegenerationsage or do they become more
poignant?KnipscheerandBevers(1985)havedrawn attentionto the strategies
of members of the oldest generationin nurturing reciprocal concernsas long
as possibleso as to postponethe unavoidableasymmetryin later years.
Our data allow only an indirect assessmentof the ways in which
intergenerationalcommitmentsmay havegrown, evolvedandbeenconfirmed
over the years, namely by means of an analysis of current frequency of
Intergenerationalstructureof family relationships
55
interaction.First, our resultsshow that keepingin touch is the norm among
parents and their adult children. Only a small number of older adults do not
interactwith one or more of their children at leastmonthly. Once again, the
results of empirical studiesconfirm that older adults are not alienatedfrom
their families. Secondly,thereare clear differencesin the frequencyof interandintragenerational
interaction.Older adultshavesubstantiallyhigherlevels
of interactionwith their childrenthan with their siblings.To a certainextent,
this finding is attributableto differencesin geographicproximity: children
are more likely to be householdmembers and to be living nearby than are
siblings. Nevertheless,we feel more is involved. In our view the relatively
high levels of parent-childinteractionsay somethingaboutthe quality of those
relationships.They underscorethe specialnature of the parent-childbond,
a sharedconcernabout eachother's progressand well-being.
The data on the levels of interaction with siblings support the idea of a
preferentialhierarchyof relationships.Contactswith siblingsare more likely
to be activated in the absenceof presumably 'more preferred' relationships
such as those with a partner and children. The childlessand those who are
single interact most frequently with siblings. We did not find evidencefor
the linking pin function of parents.In other words, contraryto expectations,
the frequency of contact with siblings did not vary betweenthose with and
those without surviving parents. Finally, strong differencesaccording to
family size were found: the frequency of contact is inversely related to the
number of surviving siblings.
It is often suggestedthat large and extensive families serve a socially
embracing function for their members. Our data, however, highlight the
cohesivefunctions of smaller families. Members of smaller families seem
most likely to keep in touch with one another.This is not only evident in
interactionsamongsiblings(as describedabove)but also in interactionswith
children. In fact, the sizeof the family of procreationwasthe only significant
determinantof the levelsof interactionwith children. Thesefindingsprovide
a different perspectiveon often heard lamentsabout declining family sizes:
for example, the notion that aging parentsrisk having insufficient sourcesof
support. There tends to be a one-sidedemphasison the vulnerability of the
situation of children from small families and on the burdens they face.
Lacking a wider circle of brothersand sisters,there are few optionsto share
supportivetasks.Elsewherewe haveshownthat the supportchildrenprovide
to older parentsis inversely relatedto family size (Dykstra & Knipscheer,
1993). Children with few siblings seemparticularly committedto remain in
56
Pearl Dykstra and Kees Knipscheer
close contactwith their parentsand to help them when the needarises.One
shouldbe carefulto basepredictionsaboutthe future situationof older adults
on demographicdata alone.
References
Arling, G. (1976). The elderly widow andher family, neighborsand friends.
Journal of Marriage and the Family, 38,757-768.
Beer, J. de, de Jong, 4., & Visser, H. (1993). Nationale
Huishoudensprognose 1993 [National household forecasts 1993].
Maandstatistiek van de Bevolking, 93/8, 13-23.
Bengtson,V.L., Cutler N.L., Mangen, D.J., & Marshall V.W. (1985).
Generations,cohortandrelationsbetweenagegroups.In R.H. Binstock
& E. Shanas(Eds), Handbook of aging and the social sciences(2nd
ed., pp. 304-339).New York: Van NostrandReinhold.
Bengtson,V.L. & Treas,J. (1980).The changingfamily contextof mental
healthand aging. In J.E. Birren & R.B. Sloane(Eds.),Handbookof
mentalhealthand aging (pp. a00-428).EnglewoodCliffs, NJ: Prentice
Hall.
Bumpass, L .L . ( 1990). What's happeningto the family? Interactionsbetween
demographicand institutionalchange.Demography,27, 483-498.
Burch, T.K. & Matthews,B.J. (1987). Householdformationin developed
societies. Population and DevelopmentReview, I 3, 495-511.
Callahan, J. (1985). What do children owe elderly parents?The Hasting
CenterReport, 15, 32-37.
Cantor, M.H. (1979). Neighborsand friends: An overlookedresourcein the
informal support system.Researchon Aging, 1,434-463.
Cherlin, A., Furstenberg, F.F. (1985). Styles and strategies of
grandparenting. In V.L. Bengtson & J.F. Robertson (Eds.),
(pp. 97-116).Beverly Hills, CA: Sage.
Grandparenthood
Cicirelli, V.G. (1982). Sibling influencethroughoutthe lifespan. In M.E.
Lamb & B. Sutton-Smith(Eds.), Sibling relationships:Their nature
and significance across the liftspan (pp 267-284). Hillsdale N.J:
Lawrence Erlbaum.
Council of Europe (1984). The social protection of the very old.
Strassbourg.
Dooghe, G. (1970). De structuur van het gezin en de sociale relaties van
bejaarden[Family structureandthe socialrelationshipsof olderadults].
Antwerpen: NederlandseBoekhandel.
Intergenerationalstructureof Íamily relationships
-$
I
57
Dykstra, P.A. & Knipscheer,C.P.M. (1993). Kindertal en door ouderen
verkregensteun [Family size and supportreceivedby adult children].
In M.M. Blom, Y. Kuin & H.J.F. Hendriks(Eds.), Ouderworden '93
(pp. 141-147).Utrechr: NIZV/.
Hagestad,G.O. (1984). Multi-generationalfamilies: Socialization,support
and strain. In V. Garms-Homolová,E. Hoerning & D. Schaeffer
(Eds.), Intergenerationalrelationships(pp. 105-ll4). Lewiston, NY:
Hogrefe.
Hagestad, G.O. (1992). Family networks in an ageing society: Some
reflectionsand explorations.In W. van den Heuvel, R. Illsley, A.
Jamieson& C.P.M. Knipscheer(Eds.), Opportunitiesand challenges
in an ageingsociety.Amsterdam:Royal Academyof Arts andSciences.
Knipscheer,C.P.M. (1980). Oude mensenen hun sociale omgeving.Een
studie van het primair sociaal netwerk [Older adults and their social
surroundings:A study of the primary social networkl. Den Haag:
VUGA.
Knipscheer,C.P.M. (1986). Family care of impairedelderly, possibilities
for innovations. In W. van den Heuvel & G. Schrijvers (Eds.),
Innovations in care for the elderly: European experiences(pp. I37165). Lochem: De Tijdstroom.
Knipscheer, C.P.M. (1990). Ouder worden,familie en intergenerationele
betrokkenheid [Aging, family and intergenerational commitment].
Assen: Van Gorcum.
Knipscheer,C.P.M. & Bevers,A. (1985).Older parentsand their middleagedchildren: Symmetry orasynrmetryintheir relationships. Canadian
Journal on Aging, 4, 145-159.
Kobrin, F.E. (1976). The fall in householdsize and the rise of the primary
individual in the United States.Demography,13,I27-I38.
Langeveld, H.M. (1985). Binding in vrijheid: Een studie naar toekomstige
gezinnen, relaties en hulpverlening. Den Haag: Sociaal Culturele
Studies6.
Laslett, P. (1965). The world we have lost. London: Methuen.
Liefbroer, A.C. &, Jong Gierveld, J. de (1995). Standardizationand
individualization:Thetransitionfromyouthto adulthoodamongcohorts
born between1903and 1965.In H. van denBrekel& F. Deven(Eds.),
Population and family in the low countries 1994 (pp. 5l-79).
Dordrecht: Kluwer.
Mancini, J.A. & Blieszner,R. (1989). Aging parentsand adult children:
Researchthemesin intergenerationalrelations. Journal of Marriag eand
the Family, 5l, 275-290.
58
Pearl Dykstra and Kees Knipscheer
Moors, H., Beets,G., & Brekel, H. van den (1995). Opvattingenover en
acceptatie van bevolkingsbeleid,1983-1990 [Opinions about and
acceptance
of populationpolicies, 1983-19901.
NIDI-report 41. Den
Haag: NederlandsInterdisciplinairDemografischInstituut.
Paillat, P., Attias-Donfut,C., Clement,F., & Delbes,C. (1989).Passages
de Ia vie active a la retraite. Paris: PUF Collection Politique
d'Auj ourd'hui.
Pampel,F.C. (1983).Changesin the propensityto live alone:Evidencefrom
consecutivecross-sectional
surveys,1960-197
6 . D emography,20,,433447.
Parsons,T. (1943). The kinship systemof contemporaryUnited States.In
T. Parsons(Ed.), Essaysin sociologicaltheory (pp. I77-196). New
York.
Prins, C.J.M. (1990). Sterkegroei éénpersoonshuishoudens
onderbejaarden
growth
one-person
households
among
the elderly].
[Rapid
Maandstatistiek van de Bevolking, 90/2, 10-11.
Prins, C.J.M. (1994). Kans op grootouderschapneemt pas op termdn af
[Probability of great-parenthoodnot down yet]. Maandstatistiekvan
de Bevolking, 94/1, 6-8.
Regt, A. de (1993). Geld en gezin: Financiëleen emotionelerelatiestussen
gezinsledenlMoney and family: Financial and emotional relationships
among family membersl. Meppel: Boom.
Riley, M.w. (1983). The family in an aging society: A matrix of larent
relationshrps.Journal of Family Issues,4, 439-454.
Ruffin, M.D. (1984). Contributionof the family to the economicsupportof
the elderly. Family EconomicReview,4, 1-11.
Troll, L.E., Miller, S.J., & Atchley R. C. (1979). Families in later life.
Belmont, CA: Wadsworth.
Uhlenberg, P. (1980). Death and the family. Journal of family history, 5,
313-320.
Verhoef, R. (1989). Geboorte[fertility]. Maandstatistiekvande Bevolking,
89/10,r4-r5.
Watkins, S.C., Menken, J.A., & Bongaarts,J. (1987). Demographic
foundationsof family change.AmericanSociologicalReview,52,346-
3s8.
White, L.K. & Riedmann(1992). Ties amongadult siblings. Social Forces,
71, 85-102.
Wolf, D.A. (1990). Householdpatternsof older women: Someinternational
comparisons.Researchon Aging, 12, 463-486.