84
education (in particular associated with a university-level education) in
combination with an accumulation of cultural capital contributes to a
higher integration of non-kin ties. By contrast, individuals with lower
levels of education are more confined to kinship ties. These trends remain
the same even after the inclusion of family-related attitudes into the
regression model. The higher the score in the openness to new family
forms, the lower the proportion of kinship ties (Model 2c), revealing a
strong correspondence between a less institutionalized and traditional
view of family life and the permeability of the network boundaries
towards the inclusion of ties beyond the kinship system.
Model 3 in Table 3.7 shows the main predictors of the proportion of
present or past co-resident alters. In Model 3a, we can see the significant
effect of country, with Switzerland standing out as the country where
networks are composed of a lower proportion of present or past co-
residents and Lithuania as the country where networks are composed of
the highest proportion of present or past co-resident alters. This effect
remains significant across the blocks. Model 3b also shows the significant
role of birth-cohort and education. Individuals with lower education
show a higher proportion of present or past co-resident alters, and indi-
viduals with higher education show a lower proportion of co-resident
alters. Also, those individuals born in the younger cohort include a lower
proportion of present or past co-resident alters when compared to those
from the older cohort. This effect of birth-cohort can be explained by the
higher inclusion of non-kin who probably have never lived with ego in
the same household, but also a lower influence of the co-residence prin-
ciple. Actually, the significant role of the normative context indicators
supports this argument. In Model 3c, we can see that individuals who are
more child-centred show a higher proportion of present or past co-
resident members; and those who are more open to new family forms
show a lower proportion of network members with a shared co-residence
history. This means that an attitudinal profile decentred from traditional
principles linked to parenthood and a more inclusive vision of family
beyond institutionalized normative models are associated with affinity
mechanisms that go beyond the co-residence principles.
Model 4 in Table 3.7 reveals the role of the different factors on the
proportion of female alters. First, we can see that neither country nor nor-
R. Gouveia et al.