Socialized Choices - Labour Market Behaviour of Dutch Mothers
118
Independent variables: demographic and situational factors
To allow for objective background characteristics, several control variables are
included. These variables allow weighing the importance of attitudinal factors as
compared to objective characteristics. Moreover, the path analysis enables us to
distinguish between the (direct) influence of these background characteristics on
women’s labour market behaviour and the indirect influence through her work
preferences.
Educational attainment. Many studies have
shown that the higher the
educational level attained, the higher the labour market participation of women
(Merens et al., 2011). Higher educated women more frequently continue to work
after giving birth than low educated women, because the former earn a higher
wage, which allows them to pay for child-care facilities (Doorewaard et al., 2004,
p.11). Furthermore, it is known that women’s and men’s educations are positively
associated with egalitarianism (Kroska and Elman, 2009, p.373,), which might be
the result of exposure to ideas about equality or the
establishment of career-
oriented networks (Cunningham et al., 2005, p.887). Thus, the study expects
higher educated mothers to (prefer to) work more hours than lower educated
mothers.
40
Number and age of children. It is expected that the larger the number of
children, and the younger their age, the lower the mother’s (preferred) number of
work hours will be.
The age of youngest child is measured in three categories: 0
to 4 years, 5 to 8 years, and 9 to 12 years.
Age. Age may refer to the life phase as well as to the generation (birth cohort)
of the respondent. In a cross-sectional analysis it is not possible to distinguish
between age and cohort effects. Recent research has shown that the number of
hours women work after giving birth to their first child is higher for younger
generations than for older ones (Lut, Van Galen and Latten, 2010). Younger
mothers are expected to work more hours than older mothers. Yet, the older a
mother is, the
older her children tend to be, and so her caring tasks diminish.
Thus, the number of work hours older mothers prefer is expected to be higher
than the number younger mothers prefer. After a certain age, as mothers approach
retirement, their preferred labour market orientation may change again (Román et
al., 2007). To account for this possible non-linear relationship between age and
labour participation, the variable age squared (divided by 100) is included in the
analyses.
Partner. A partner can affect his spouse’s employment decisions in various
ways: his income, the number of hours he works, his
career perspectives, his
attitude towards her income and her career perspectives, and his acceptance or
40
The categories are: 1: primary school, 2: VMBO (intermediate secondary education (equivalent to
US: junior high school); 3: HAVO/WVO (higher secondary education/preparatory university; 4:
MBO (intermediate vocational education, equivalent to US junior college); 5: HBO (higher
vocational education, equivalent to US college); 6: WO (university).
Chapter 4 - The vital and stabilising role of work preferences
119
rejection of the male breadwinner model may all influence a mother’s labour
market preferences and behaviour (Hoffnung
and Williams, 2013; Kangas and
Rostgaard, 2007; Van Wel and Knijn, 2006). I expect that cohabiting mothers
work fewer hours than single mothers, since single women cannot financially rely
on a partner and, thus, have to work more hours in order to earn a living. And yet
the preferred number of work hours for single mothers might be lower because
they cannot share their caring tasks with a partner.
In line with microeconomic theory, the higher the income of the partner, the
fewer hours his wife works, because his income is sufficient to make ends meet.
41
Moreover, a strongly career-oriented husband, with high earnings, may work
long hours and leave the bulk of household responsibilities to his wife, which
may hinder her labour participation (Cloïn, 2010). However, a high income might
not necessarily reduce a mother’s
work preference, which might be higher than
one would expect in view of her partners’ income.
Religion. Religiosity is also expected to coincide with more traditional general
and personal gender attitudes, since most religions endorse traditional views on
the role of women (Bolzendahl and Meyers, 2004; Cunningham, 2001; De Valk,
2008; Kraaykamp, 2012; Thompson, 1991). I thus expect that religious mothers
have lower work preferences and more traditional general values and personal
attitudes. The respondents were asked how often, aside from special occasions
such as weddings and funerals, they attend religious gatherings nowadays: 1:
every day; 2:
more than once a week; 3: once a week; 4: at least once a week; 5:
only on special religious days; 6: less often; 7: never.
4.6 Results
Hypothesis 1:
A mother’s labour participation is largely determined by her preferred number
of work hours.
Participation decision
Initially, all relevant demographic variables and the number of preferred work
hours are included in the analysis, in order to examine which variables affect
whether a mother has a paid job or not. Remarkably, only the number of preferred
work hours exerts a direct significant effect on participation in paid work (beta
.673) (table 6). Work preference explains 45 per
cent of the variance of the
participation. In itself, this is not very surprising, but the impact of this variable is
so strong that the background characteristics of the mother, such as age,
educational level, the number of children, the presence and the income of a
41
Since we include the log of the income of the partner in the analysis, we imputed a value of zero
for the
log of zero income, which does not exist. To correct for this, we added a dummy variable
of one when the partner has no income.