The mental wellbeing of parents and the childless in a European context: The case of France and Sweden
Ellen Ellfolk Kenttä
Department of Sociology, Demography Unit (SUDA) Master’s Thesis in Demography
Supervisor: Sven Drefahl Co-supervisor: Johan Dahlberg
Abstract
A growing number of individuals in Europe and other developed regions of the world are remaining childless throughout their life course. These individuals will lack major social and caring components relative to parents and likely have other needs and issues as they age and as the proportion of childless continues to grow. It has previously been shown that partnered individuals and parents’ live longer and healthier lives, but studies into the mental health of childless individuals have been inconclusive. In the present study I investigate the relationship between mental wellbeing and childlessness in France and Sweden. The mental wellbeing of the respondents was measured using a short-form of a depression scale. The result of the study showed that the effects of childlessness were affected by country of residence, gender, parental status as well as marital status. Further, a poor relationship to one’s children had a negative impact on mental wellbeing. Individual attitude towards childlessness did not have an effect on mental wellbeing however, and contrary to previous studies and the first hypothesis of this study, being childless did not have a negative effect on mental wellbeing. The results are limited by the nature of the sample, where a more complete measure of the depression scale would have been beneficial. Future research could further investigate the significance of positive family
relationship compared to those with poor relationships and the childless. As the proportion of childless individuals continues to grow there will be ample research opportunities into their mental wellbeing.
Table of contents
Abstract ... 2
1. Introduction ... 4
2. Family units and social support ... 6
3. Gender, childlessness and selection ... 8
4. Changing effects of parenthood over the life course ... 10
5. Social and cultural context ... 11
France and Sweden ... 11
6. Mental wellbeing ... 12
7. Hypotheses ... 12
8. Method and Data ... 13
8.2 Sample ... 14
8.3 Dependent variable: Mental wellbeing ... 15
8.4 Main independent variable of interest: Parental status ... 15
8.5 Other independent variables ... 15
8.6 Descriptive statistics ... 17
9. Results ... 19
9.1 Results - Model A ... 20
9.2 Results – Model B ... 22
9.3 Results – Model C ... 24
10. Discussion and Conclusions ... 25
Acknowledgements ... 30
Bibliography ... 31
Appendix 1 – Distribution of dependent variable ... 35
1. Introduction
The family remains a major component of emotional support and social interaction for many individuals. Demographic developments in much of Europe and the rest of the Western world have increased the number of childless individuals and decreased union stability. Thus, family is becoming a component that an increasing number of individuals will lack. Social interaction is a key part of mental wellbeing for the vast majority. As number of childless individuals increases it will be necessary to determine if they have other needs than parents, particularly in respect to social interaction. Social interaction is a key part of mental wellbeing and those who lack family may also lack social interaction and therefore experience higher rates of depression and mental health issues. Particularly, having children would have a positive effect later in life due to the increased social contact compared to older and childless individuals. The aim of this thesis is to investigate the effects of childlessness on mental wellbeing depending on social context in France and Sweden. What follows is a general demographic background, and an overview of previous research in regards to family, gender, the life course, and the social and cultural context.
People are getting married later, having more children out of wedlock or postponing family formation, or avoid having children all together. This pattern is seen in all developed countries to different degrees, but it is most prevalent in Europe (Lesthaeghe, 2010). This demographic development is combined with aging European populations, will likely result in a growing number of elderly, childless individuals in the future. As an increasing number of individuals remain childless for the entire duration of their lives it will become important to establish what needs these individuals will have. These individuals will lack major family components that in old age would represent both a social safety net and large proportion of their social interaction on a regular basis (Bachrach, 1980). It is likely that childless individuals will have less social interaction as a result of their childlessness, especially in retirement ages, and experience greater social isolation than parents (Ibid). Depending on factors such as socio economic status and marital status, this may have potential to lead to detrimental mental health among older individuals (Beckman & Houser, 1982; Brown, 2000; Buber & Engelhardt, 2008).
The number of childless individuals has been steadily increasing in developed countries in the past several decades (Frejka & Sardon, 2006; Sobotka, 2016). In Europe we have generally seen an increase in individuals who by their own choosing or otherwise remain childless throughout their life course. This has been in combination with a continent wide aging of the population and general change in demographic behaviour, such as postponed marriages, increased divorce and lower fertility (Lesthaeghe, 2010; Goldscheider, Bernhardt, & Lappegård, 2015). While desired
family size and number of children has stayed relatively stable around two children, realised fertility has not met this desired level (Frejka & Sobotka, 2008; Goldstein, Lutz, & Testa, 2003;
Sobotka, 2016; Testa, 2007).
There are several components that may possibly affect mental health of parents and childless individuals. Childless individuals will have less social interaction, and they might experience considerable pressure and negativity from their peers and society if they choose to remain childless in a culture that has negative social norms in regards to childlessness (Koropeckyj-Cox, 2002). As was pointed out in their 2013 study on the subject, Huijts, Kraaykamp and
Subramanian found that social context will play a major factor in determining the effects of childlessness on mental health. Social context will also play a role in the different effects of childlessness between men and women found in the same study, as gender is a likely determinant of how individuals are affected by social context. Most previous research on the topic has been done on American data, often on younger individuals and often focusing on women or excluding men. However, men are also shaped by becoming parents or not and need to be considered to the same extent as women (Huijts et al., 2013).
With decreased levels of fertility in most of Europe we can expect the levels of childless individuals to increase in the coming years (Sobotka, 2016). It is estimated that as many as 10- 20% of women born after 1950 in Europe may remain childless throughout their lives (Rowland, 2007). The same study found a similar pattern in the United States, Australia and Japan. Children represent a major component of many individual’s social and everyday needs and care for aging parents. The growing proportion of childless individuals lack this safety net and social support of the family will require both social and physical care from the market or the state to a greater degree than parents. While countries with a large public sector such as Sweden has provided extensive elderly care at home and at residential care facilities, we are seeing a decrease in this provision in Sweden as well as other countries. Ulmanen and Szebehely, 2015, found that state provided elderly care had decreased substantially in Sweden and that there are significantly fewer beds in residential care facilities today. This has resulted in an increase in family provided support for the elderly, which is mainly given by daughters among individuals with lower education while individuals with higher levels of education are more likely to pay for private elderly care. The reduction in state provided elderly care has been remarkable in Sweden but the same pattern can be seen in much of Europe (Ibid). In combination with lower levels of fertility and increased childlessness the question of childless individual’s health and mental health needs to be extensively addressed.
2. Family units and social support
It has been argued that the nuclear family and its function have become less relevant as an institution in society as divorce rates have increased and the nuclear family has decreased its function and meaningfulness (Popenoe, 2016). However, lower levels of state provided elderly care is again reasserting the importance of the family for intergenerational transfers of care and support (Brandt, Haberkern, & Szydlik, 2009; Leitner, 2003; Silverstein, Gans, & Yang, 2006).
Fertility levels and family size are decreasing and partnerships are becoming increasingly unstable.
In a setting where fewer individuals are in stable long-term relationships, and where the number of siblings have decreased, the social and emotional support-factor of children will become increasingly important (Brandt et al., 2009). Higher life expectancy also contributes to the importance of family and inter-generational transfers, as more years are shared between the generations (Bengtson, 2001). Buber and Engelhardt, 2008, goes so far as to say that they expect mental health to deteriorate in Europe based on their findings in regards to childlessness and mental wellbeing. While the decline of the importance of the nuclear family has been discussed and researched we cannot overlook the remaining, and perhaps in certain areas growing,
importance of social support from family (Bengtson, 2001). With declining elderly care provision in most welfare states, the caring role of the family will become more pronounced, especially among poorer individuals with no access to paid care (Ulmanen, 2013; Ulmanen & Szebehely, 2015). As such there is an important socio economic aspect to any research on the topic because income determines ability to pay for care not provided by the state or the family.
It has been established in previous research that parents live longer than their childless
counterparts; they appear to live healthier and more social lives. While parenting is stressful and costly when children are young, there appears to be a significant pay off for elderly parents later in life (Modig, Talbäck, Torssander, & Ahlbom, 2016; Nomaguchi & Milkie, 2003). Children certainly represent higher access to social contact and safety nets in old age. Childless individuals have to rely on other forms of social connections, such as extended family and friends especially in the absence of a partner. While previous research into the health and mortality among parents and childless have found that children have a positive effect on life expectancy, research into what parental status means for the mental health has been inconclusive and sometimes contradictory.
While earlier studies (Glenn and Mclanahan, 1981; Ross, Mirowsky, and Goldsteen, 1990; and Zhang and Hayward, 2001), found little evidence for differences in mental health outcomes among parents and the childless, more recent studies have found evidence for a social and caring
effect of children on parental mental wellbeing later in life (Buber & Engelhardt, 2008; Huijts, Kraaykamp, & Subramanian, 2013; Zunzunegui, Béland, & Otero, 2001). In the 2016 study on mortality and childlessness, using Swedish registry data, Modig et al. found that childless men have a life expectancy that is two years shorter than for men with children. Zunzunegui et al.
(2001) found that low emotional support among the elderly in Spain had a negative impact on the respondent’s mental health. The study looked at both self-rated health and the mental health using the Center for Epidemiologic Studies Depression Scale (CES-D scale) and found that mental and general health were affected by social and emotional support given by family and children.
Another aspect of parenting is whether childbearing is done within a marital or cohabiting setting or as a single parent. As more children are born outside of the traditional marital setting today, marital status of the parents will become increasingly important. Single parents likely experience more stress and pressure than parents who have a partner. However, having children within a marriage could lead to pressure and stress between the partners. As noted by Umberson et al., (2010), there is also the possibility that not only childlessness but also the gender and relationship to the child will affect the mental wellbeing of the parent. Some studies suggest that the care given by daughters living close by in old age gives considerable mental wellbeing and social interaction, but that this relationship is not visible when it comes to sons, regardless of the living distance, or if the a daughter lives far away (Buber & Engelhardt, 2008). This effect would be caused by daughters more frequent contact with their elderly parents as well as norms regarding greater care-giving roles of daughters than of sons (Grundy & Read, 2012; Ulmanen, 2013).
While many individuals do not have children of their own, there is also an increase in individuals who have stepchildren due to increased divorce and new partners with children from previous relationships. Previous research has found little difference in the impact of having biological or stepchildren; therefore I do not distinguish between the two in this study. Zhang and Hayward (2001), found that there was virtually no difference in psychological impact of children between stepparents and biological parents. This would indicate that the mere presence of a child in a persons’ life has a positive effect, and that an individual will reap the social benefits of their partner’s children. However, this is also likely in part due to the increased social network that comes with a partner as well as possible stepchildren. It has been shown that partnership is a major determinant in mental health among the elderly (Brown, 2000; Buber & Engelhardt, 2008;
Ross et al., 1990). However, with increasingly unstable relationships, fewer individuals will retain the support of a partner throughout their life course (Mclanahan & Adams, 1987).
3. Gender, childlessness and selection
There are several factors that could potentially affect mental health outcomes of childless individuals depending on gender. Social policies, cultural gender inequality, attitudes as well as stereotypes are all factors that could potentially affect gender differences in mental health outcomes. Women have been found to be better at compensating for a lack of social support from children than men, with a wider social network within the family and among friends.
However, it has been found that they have more social contacts than childless women (P. A.
Dykstra, 2006; Grundy & Read, 2012). If the effect of childlessness is largely social, then the effect on mental health should be more pronounced among men than women. Whilst studies have found that partners are more important for the mental wellbeing of men than having
children, it has also been found that childless men have poorer family relationships (P. Dykstra &
Keizer, 2009). The study by Dysktra (2006), which looked at childlessness and social integration, showed that individuals who had never had children, and those who had outlived their children had smaller social networks to draw from in older life. Having children appears to not only provide social support from the children themselves, but also increase the social network in general, when it comes to friends and more distant family. Evidence also suggests that women and men react differently to the context of limited social networks, where women are more likely to compensate for a lack of children with other relations. However, as social norms are likely to affect men and women differently, this compensating effect might have less of an impact depending on the sex of the individual and the social context. In contrast to this assessment, in their 2008 study by Buber and Engelhardt, they found that never married men without a partner did not report higher levels of depressive symptoms, while their female counterparts did.
However, this finding differed across the 11 countries that were studied. Regardless, differing social norms across countries and gender are clear indicators of mental health (Copeland et al., 1999).
As such, previous research on the topic has been inconclusive on whether there are any differences between the mental effects of childlessness for men and women. Many of the previous findings have been done on data from the United States, where it seems that
childlessness has greater detrimental mental health effects for women than for men (Huijts et al., 2013; Zhang & Hayward, 2001). However, studies recent European studies, by Buber and Engelhardt (2008) and Huijts, Kraaykamp and Subramanian (2013), using Survey of Health, Ageing and Retirement in Europe (SHARE) and the European Social Survey (ESS) respectively, have both found that childlessness increases symptoms of mental distress and depression in men more than women. Social context and welfare state provision are likely factors behind this
difference. Strong financial prospects as well as stable mental health could mean that children have a more positive effect on individuals who have certain characteristics while they put additional stress and financial pressure on others (Umberson et al., 2010). Additional research needs to be done to properly distinguish whether poor initial mental health is the reason why people remain childless, but the evidence that has been found on the subject indicates that there is little or limited self-selection effects on childlessness and mental health (Huijts et al., 2013).
Instead there has been a greater focus on other explanations and mechanisms behind why there might be a negative effect of childlessness on mental health.
Since men appear to reap more benefits from marriage and parenthood, it is likely that there also remain gendered differences in the effect of selectivity. While uneducated men marry and have children to a lesser degree than men with higher levels of education, the opposite has been true for women (Modig et al., 2016). While this is changing, women with higher levels of education have historically been less likely to have children than their less educated counterparts.
Demographic developments over the past recent decades have however reversed this pattern, and today both highly educated men and women are the groups most likely to marry and have children (Jalovaara et al., 2017).
The problem of selectivity remains in most studies into the topic. Since both marriage and parenthood appears to have positive effects on longevity and physical as well as mental health it could be that individuals who go on to marry and have children have better health in general. It is also possible that married individuals and parents make healthier life style choices than the childless. There has so far been limited research done on the effect of self-selection regarding mental and physical health of the childless. Social context as well as self-selection determines who become parents (Buber & Engelhardt, 2008). Individuals with poor mental health will likely not become parents to the same extent as individuals who have no inclination towards mental health problems or depression. Giving care to children and receiving care from adult children can both have positive effects on the parents mental well-being, but socio-economic factors and
underlying mental health status can lead to a self-selection of educated individuals with high socio-economic status and strong mental health to be more likely to be married and have
children, further reaping the benefits of a family and the social connections which it includes. As men with lower levels of education, and to an increasing degree, women with lower levels of education are the groups least likely to have children, it is likely that selection is a factor. If individuals are not able to become parents due to either financial or health constraints, it is possible that these aspects either play a greater part in the individuals mental health than the lack
of children. Without comprehensive data on whether childlessness is voluntary or not, and reproductive health as well as socio-economic status it is difficult for any study to clearly
conclude the reason for differing mental health outcomes. While selection poses a problem there are indications that the selection into marriage and parenthood is taking on more similar
characteristics for men and women.
4. Changing effects of parenthood over the life course
In a life course perspective we would assume that the effect on mental health for parents and the childless differs during the course of an individuals life. Generally there are both costs and benefits associated with having children, small children infer high costs for their young parents in early adulthood while adult children can provide support and care to their aging parents
(Umberson et al., 2010). Financial as well as social pressure associated with raising smaller children could have adverse effects on the mental health of younger parents. However, later in life this should likely be replaced by the social effect of having children who can help you later in life. Beyond the mere financial and physical aspects of having children there is also the social aspect and in many countries there is a cultural norm that says children is part of the life course and a sign of fulfilment (Ibid). This norm varies depending on the values of the culture, and so does the need for social and physical care. As this paper looks at individuals who have likely finished their transitions to parenthood and so either already have at least one child or will remain childless, it is likely that the negative impact of having children will not be a factor.
Umberson et al. (2010) took a life course perspective on the issue as much of the previous research had been done on different types of parents. As previously mentioned, young parents might have adverse mental health effects due to having less socioeconomic resources as their children are young, and so children will have a negative effect on young parents mental health.
On the other hand, adult children will become an asset to their parents later in life, providing social support and social interaction (Ibid). It is likely that children provide psychological well- being to their parents, but that this association only holds for certain social groups and is likely not constant over the life-course (Mclanahan & Adams, 1987; Nomaguchi & Milkie, 2003). Much of the evidence suggests that young children living at home has a negative effect on wellbeing, especially for women, but that the negative effects diminish and reverse as the children grow up (Hansen, 2012). However, Nomaguchi and Milkie (2003), using data from the United States, found that even when children are young they have a positive effect on mothers’ mental health and symptoms of depression compared to the childless. Regardless, as this effect was only found
among mothers and not fathers, and since the study was carried out on American data, it is likely that social norms and culture plays a part in the effect on parenthood throughout the life course.
5. Social and cultural context
Societal context will play a major role, as several factors will differ across Europe. Cultural norms regarding gender roles, family and childlessness will affect differences between the sexes, how they feel regarding pressure to have children and how the family is expected to be (Fahlén, 2014).
Failure to live up to norms within the societal contexts can lead to adverse mental health results as outside pressure plays its role. Social institutions within the societal context will also play into the mental health effects of parents and childless individuals. Generous retirement policies could have a positive effect on the mental health of both parents and childless people while childcare provision and parental leave will only affect parents (Ibid).
The social context of the effects on mental health of being a parent is dependent on a number of factors. Firstly, the social support given to parents and to single individuals in retirement from the welfare system will affect the mental health of parents as well as childless individuals. In a welfare system with a familialistic approach to elder care, or where the norms strongly favour support given from the family rather than the state, parents and the childless will likely receive different levels of care (Brandt et al., 2009). A familialistic approach to elderly care, where the family and especially women, is by norms and policies expected to take a greater caring role for their elderly parents, maintains the importance of the family as a source for care in old age (Leitner, 2003; Schmid, Brandt, & Haberkern, 2012). Further, the cultural acceptance of childlessness and expectations of family formation being the cultural norm will play a vital role for both parents and those without children. In a society where childlessness is not accepted to a great extent or even frowned upon, individuals can both feel pressured into having children they did not want, and thus experience negative mental health effects, or feel that they are being judged if they despite the societal expectations choose to remain childless or are unable to have children (Hansen, 2012; Huijts et al., 2013).
France and Sweden
Both Sweden and France have reported low levels of childlessness compared to other western European countries (Kreyenfeld and Konietzka, 2016; Köppen, Mazuy and Toulemon, 2016).
Compared to the rest of Western Europe both Sweden and France continue to have high levels of fertility. In France, less than 15 % of women remain childless and this is often attributed to
state-support and family friendly policies. The family as an institution remains strong in France, while most men and women consider two to three children to be the ideal family size, less than 5
% thinks that childlessness is desirable (Toulemon, 2001, as cited in Köppen, Mazuy and Toulemon, 2016). Sweden is also known for family friendly policies with a special focus on gender equality and specifically for a focus on increasing female labour market participation (Duvander & Andersson, 2006). Similar to France, the results have been high labour force participation for both sexes and relatively high total fertility rate (TFR). In 2015 Frances’ TFR was slightly higher than that of Sweden at 2.0 and 1.8 per woman respectively (The World Bank).
With such similar fertility patterns Sweden and France a suitable for comparison due to their main differing feature being cultural norms and attitudes towards childlessness.
6. Mental wellbeing
Previous studies of the topic with a focus on the European situation have used the European Social Survey (ESS) as well as Survey of Health, Ageing and Retirement in Europe (SHARE).
Since results have been inconclusive in previous research it is of value to perform a similar study on a new set of data. While there are several ways of measuring mental health and wellbeing, the CES-D scale mentioned earlier will be employed in this paper. Both Zunzunegui et al. (2001) and Huijts, Kraaykamp and Subramanian (2013) among others, used a short form of the CES-D scale to measure self-reported mental wellbeing as an indicator for mental health. The nature of the GGP data makes a shortened version of the CES-D scale suitable as well as facilitating
comparability to previous and future research. The full CES-D scale employs a twenty-item list of questions and seven of these questions are also found in the GGP.
7. Hypotheses
Social interaction with family and friends has been shown to be of major importance for an individual’s social and mental wellbeing. As family and children represent major components for many individuals’ social interactions and care on a daily basis it is possible that a lack of these components have adverse effects on mental wellbeing, especially later in life (Buber &
Engelhardt, 2008; Huijts et al., 2013; Zunzunegui et al., 2001). While relationships remain an important aspect of one’s social life, increased divorce rates and union dissolution as well as longer life spans potentially mean that fewer individuals retain their partners later in life. Children could be a more stable source of social interaction and emotional care. As such the first
hypothesis to be tested in this paper is:
• Do childless individuals have worse mental wellbeing than parents?
Both Sweden and France continue to have relatively high levels of fertility, which is often
attributed to family friendly policies in both countries (Kreyenfeld and Konietzka, 2016; Köppen, Mazuy and Toulemon, 2016). If there is a difference in mental wellbeing in conjunction with childlessness it is likely that this effect is largely due to cultural norms and attitude towards childlessness as has been pointed out in previous research (Huijts et al., 2013). As French individuals have previously been shown to be more opposed to childlessness as a life choice, the second hypothesis to be tested is formulated as:
• Due to social context, is the negative effect of childlessness weaker in Sweden?
These will be set against the null hypothesis that childless individuals do not have worse mental health than parents and that there is no difference in reported mental health between the two countries.
To test these hypotheses I will control for twelve factors; country of birth, age, sex, education, partnership status, parental status, presence of a child in the household, satisfaction of
relationship to an individuals children, deceased child, presence of a daughter, attitude towards childlessness and attitude in regards to parents versus childless individuals right to jobs when jobs are scarce. These variables will be discussed in greater detail in the method and data section on page 15.
8. Method and Data
The data used for this study is the Gender and Generations Survey (GGS), which is collected and provided by the Gender and Generations Programme (GGP). The GGS provides micro level data and enables cross-national comparability. For France and Sweden, data Wave 1 is used as this Wave contained the necessary control variables and because Wave II has yet to be released for Sweden. The data for France was collected in 2005 and the Swedish data was collected in 2012. The aim of the GGP is to provide data for policy relevant research in the fields of economics, demography, sociology and other fields within the social sciences.
A linear regression model will be used to measure the effect of childlessness on mental wellbeing.
There are four groups of respondents in the sample, French men and women and Swedish men and women. In Model A, the country variable will be included in the model, while men and
women will be analysed as two separate groups. For the other two models, Model B and Model C, these groups will be looked at separately.
8.2 Sample
The data for France contained a total of 6 279 observations after excluding respondents younger than 40, of which 2 782 were men and 3 470 were women. For Sweden the original number of observations was 6 258, of which 3 019 were men and 3 239 were women. It was assumed that childlessness after the age of 40 was likely to be permanent. It has previously been found that the vast majority of men, 97%, have their first child before the age of 40 years (Dykstra & Keizer, 2009). While the age at which men and women have their first child is increasing it was assumed that individuals who had not had their first child at age 40 were not likely to have any, as such age 40 was chosen as the cut-off point for the sample.
Due to missing values the following observations have been excluded:
Variable Excluded
Parental status 60
CES-D 1 516
Education 5
Partnership status 5
Marital status 337
Child mortality 100
Satisfaction with
relationship 96
Attitude 199
Job attitude 33
Total excluded 2 351
Original sample 12 510
Final sample 10 159
Table 1 – Excluded variables and final sample
The response rate of the sample with row percentages for France and Sweden, as well as men and women can be seen in Table 2 below.
Male Female Total France 2 637
44.15% 3 336
55.85% 5 973 Sweden 1 955
46.7% 2 231
53.3% 4 186 Total 4 592
45.2% 5 567
54.8% 10 159
Table 2 – Frequency table for French and Swedish men and women
8.3 Dependent variable: Mental wellbeing
To measure the dependent variable mental wellbeing, the Center for Epidemiologic Studies Depression Scale (CES-D scale) will be employed. The CES-D scale is an often-used tool for measuring the mental wellbeing, while not being a method of diagnosing clinical depression (Radloff 1977). For the seven items on the CES-D scale included in the GGS the respondents were asked to state how many times in the past week that they had felt (i) “I felt I could not shake of the blues”, (ii) “I felt depressed”, (iii) “I thought my life had been a failure”, (iv) “I felt fearful”, (v) “I felt lonely”, (vi) “I had crying spells”, (vii) “I felt sad”. All questions were
answered on a scale of 1-4, where 1 indicated “seldom or never”, 2 indicated “sometimes”, 3 indicated “often” and 4 indicated “most of the time”. The CES-D scale was constructed using these questions by the author. Respondents was given an aggregated score based on their responses to these seven questions, giving a scale ranging from 0-21 where a score of a 21 indicates the highest level of mental distress. While this short-form of the CES-D has not been used with GGS data previously, it has successfully been employed in similar studies using other data (Van de Velde et al, 2010; Huijts, Kraaykamp and Subramanian, 2011). The mean CES-D scores for the different groups are presented in Table 3.
France
Sweden
CES-D Men Women Men Women
Mean 1.6674 3.4092 1.8215 2.5482 Sample
mean 2.4624
Table 3 – Mean CES-D score for sample groups and sample mean
While the reported CES-D scores are generally low in the sample it is clearly higher for women and especially so for French women compared to the other groups. The distribution of the dependent variable is presented in Figure 1 in Appendix 1.
8.4 Main independent variable of interest: Parental status
As the effects of parenthood is likely to differ between parents with co-residential children and parents who’s children have moved away, so called empty-nesters, the variable for parental status is a category variable. The three categories are 1=No children, 2=Empty nest and 3=Co-
residential child.
8.5 Other independent variables
The control variables to look at will be country of birth, age, sex, education, presence of a partner, marital status, presence of a child in the household, satisfaction of relationship to an
individuals children, deceased child, presence of a daughter, attitude towards childlessness and attitude in regards to parents versus childless individuals right to jobs when jobs are scarce.
While there will be separate models for France and Sweden, they will also be included as control variables in a separate model. The variable is coded as a dummy variable where 1=France and 2=Sweden. The age variable is categorical, and the responses range from 40 to 80 years for men and 40 to 79 years for women. The categories are 1=40-49 years of age, 2= 50-59 years of age, 3=60-69 years of age, and 70-80 years of age. The mean age for men is 57.46 years and 57.28 years for women. As education was reported as country specific it was recoded as a 3-category variable that was the same for the two countries to facilitate comparison. For both countries, 1=Primary schooling, 2=Secondary schooling, and 3=Tertiary schooling.
The variable Partnership status was coded as a categorical variable, were 1=Never married and single, 2=Never married with a partner, 3=Married, 4=Divorced/Widowed and single and 5=Divorced/Widowed with a partner. This categorical variable was constructed from two variables, denoting presence of a partner and marital status respectively. As it is likely that never married single individuals differ from the never married with a partner, these variables were separated.
The Parental status variable was coded as a binary variable, where 0=No children and 1=Has children. The Empty nest variable denotes whether an individual has a child living in their household. The variable is binary and coded as 0= Empty nest and 1=Co-residential child.
The variable for satisfaction with the relationship to non-residential children was given on a scale of 1-10 for each child of the respondents that did not live at home. To make the analysis simpler the mean response was divided into a categorical variable. The mean level of satisfaction was divided on the respondents’ number of children to give a scale of 1-10 where 10 were the most satisfied with the relationship. The scale of 1-10 was divided into a categorical variable with 5 categories. As there are individuals who do not have children and those who have children but none that have moved away from home, these groups are in separate categories. Respondents without children are coded as category 1 and individuals who do not have any children who have left the home are in category 5. As such category 1 indicates that the respondent has no children, 2= low levels of satisfaction with the relationship to non-resident children, 3=medium levels of satisfaction with non-resident children, 4=high levels of satisfaction with non-resident children, and finally 5=respondent only has children who are co-residential.
The child mortality variable was included to control for the experience of having had a child that was deceased, as this is very likely to be traumatic and affect mental wellbeing. The variable was coded as a binary variable where 0=No (had not experienced the death of a child) and 1=Yes (had experienced the death of a child).
The variable Daughter controlled for whether the respondent had at least one daughter, the variable was coded as binary where 0=No and 1=Yes.
The Child attitude variable gave the respondents opinion on the statement “A woman/man has to have children in order to be fulfilled” separately for men and women on a 1-5 scale. The variable was coded so as 1=Strongly disagree, 2=Disagree, 3=Neither agree nor disagree, 4=Agree, and 5=Strongly agree. The responses for both statements were combined into one attitude variable on the same scale of 1-5. As such a higher category denotes that the individual more strongly agrees with the assertion that “A woman/man has to have children in order to be fulfilled”.
The Job attitude variable gave the respondents opinion on the statement “When jobs are scarce, people with children more have right to job than childless” on a 1-5 scale. The response was given as 1=Strongly disagree, 2=Disagree, 3=Neither agree nor disagree, 4=Agree, and 5=Strongly agree.
8.6 Descriptive statistics
Frequencies with column percentages for the independent variables for French and Swedish men and women are presented in Table 4 below.
France Sweden
Men Women Men Women Sample Total
Education
Primary 1 002
38%
1 565 46.91%
179 9.16%
137 6.14%
2 883 28.38%
Secondary 1 089
41.3% 1 100
32.97% 1 218
62.3% 1 213
54.37% 4 620 45.48%
Tertiary 546
20.71% 671
20.11% 558
28.54% 881
39.49% 2 656 26.14%
Age
40-49 years 784
29.74%
1 008 30.22%
471 24.09%
608 27.25%
2 871 28.26%
50-59 years 780
29.58% 1 019
30.55% 516
26.39% 591
26.49% 2 906 28.61%
60-69 years 607
23.02% 713
21.37% 633
32.38% 651
29.18% 2 604 25.63%
70-80 years 466
17.67% 596
17.87% 335
17.14% 381
17.08% 1 778 17.5%
Partnership status
Never married and
single 304
11.53% 465
13.94% 157
8.03% 146
6.45% 1 072 10.55%
Never married with
partner 99
3.75% 108
3.24% 187
9.57% 230
10.31% 624 6.14%
Married 1 776
67.35%
1 819 54.53%
1 358 69.46%
1 417 63.51%
6 370 62.7%
Divorced/Widowed
and single 235
8.91% 640
19.18% 130
6.65% 313
14.03% 1 318 12.97%
Divorced/Widowed
with partner 223
8.46% 304
9.11% 123
6.29% 125
5.6% 775 7.63%
Parental status
No children 413
15.66% 421
12.62% 210
10.74% 155
6.95% 1 199 11.8%
Has children 2 224
84.34% 2 915
87.38% 1 745
89.26% 2 076
93.05% 8 960 88.2%
Co-residential child
Empty nester 1 776
67.35% 2 229
66.82% 1 221
62.46% 1 431
64.14% 6 657 65.53%
Co-residential child 861 32.65%
1 107 33.18%
734 37.54%
800 35.86%
3 502 34.47%
Satisfaction with relationship
1 – No children 413
15.66%
421 12.62%
210 10.74%
155 6.95%
1 199 11.8%
2 – Low satisfaction 111 4.21%
97 2.91%
38 1.95%
23 1.03%
269 2.64%
3 – Medium
satisfaction 469
17.79% 526
15.77% 272
13.91% 212
9.5% 1 479 14.56%
4 – High satisfaction 1 015 38.49%
1 455 43.62%
925 47.31%
1 262 56.57%
4 657 45.84%
5 – Only children
living at home 629
23.85% 837
25.09% 510
26.09% 579
25.95% 2 555 25.15%
Experienced the death of a child
No 2 489
94.39% 3 086
92.51% 1 919
98.16% 2 180
97.71% 9 674 95.23%
Yes 148
5.61% 250
7.49% 36
1.84% 51
2.29% 485 4.77%
Daughter
No 954
36.18%
1 149 34.44%
609 31.15%
599 26.85%
3 311 32.59%
Yes 1 683
63.82% 2 187
65.56% 1 346
68.85% 1 632
73.15% 6 848 67.41%
Child Attitude
1 – Strongly
Disagree 183
6.94% 331
9.92% 607
31.05% 1 005
45.05% 2 126 20.93%
2 155
5.88% 252
7.55% 604
30.9% 692
31.02% 1 703 16.76%
3 409
15.51% 517
15.5% 623
31.87% 477
21.38% 2 026 19.94%
4 643
24.38% 800
23.98% 102
5.22% 51
2.29% 1 596 15.71%
5 – Strongly Agree 1 247
47.29% 1 436
43.05% 19
0.97% 6
0.27% 2 708 26.66%
Job attitude
1 – Strongly
Disagree
549 20.82%
792 23.74%
404 20.66%
498 22.32%
2 243 22.08%
2 255
9.67% 332
9.95% 601
30.74% 633
28.37% 1 821 17.92%
3 473
17.94% 657
19.69% 715
36.57% 934
41.86% 2 779 27.36%
4 684 25.94%
839 25.15%
205 10.49%
143 6.41%
1 871 18.42%
5 – Strongly Agree 676
25.64% 716
21.46% 30
1.53% 23
1.03% 1 445 14.22%
Table 4 - Frequency table for independent variables
A factor worth noting on from the descriptive table is that the attitude variable is vastly different for the French and Swedish respondents. A majority of Swedish respondents Strongly disagree or Disagree with the statement that “A woman/man has to have children in order to be fulfilled”.
For French respondents the case is the opposite and a majority agrees with this statement.
9. Results
Three models are presented below; in model A the results are presented separately for men and women, while the country variable is present in the model. The results for model B and C are presented separately by country and sex.
9.1 Results - Model A
Model A has an interaction term between the country variable and the parental status variable to analyze how cultural context affects mental wellbeing in regards to parenthood.
Model A Men Women
Country#Parental status
Reference group: Sweden#No Children
France#No children -0.49 0.02
France#Has children -0.65*** -0.08
Sweden#Has children Omitted Omitted
Co-residential child
Reference group: Empty nest
Co-residential child 0.31* -0.08
Partnership status
Reference group: Never married & single
Never married with partner -1.41*** -0.99**
Divorced/widowed & single 0.6** -0.11
Divorced/widowed with partner -1.2*** -1.52***
Married -1.62*** -1.74***
Education
Reference group: Secondary
Primary 0.4** 0.24
Tertiary 0.04 -0.37**
Age
Reference group: 40-49 years
50-59 years 0.08 -0.1
60-69 years -0.22 -0.47*
70-80 years -0.18 -0.44*
Satisfaction with relationship
Reference group: 4
1 0.04 0.55
2 0.7** 3.26***
3 0.27* 1.09***
5 0.07 0.31
Experienced the death of a child
Reference group: No deceased child
Deceased child 0.49* 1.21***
Daughter
Reference group: No daughter
Daughter -0.11 0.36**
Child Attitude 0.05 0.12*
Job attitude 0.12** 0.17***
Constant 2.62*** 3.09***
Table 5 – Results for Model A
***P≤0.001, **P≤0.01, *P≤0.05
Model A includes the country variable inside the model in an interaction term with parental status (country#parental status). The effect of age in this model is not significant for men and only significant for women for the age groups 60-69 years and 70-80 years. The older age groups appear to have better mental wellbeing than the reference category, 40-49 years of age.
For the education variable, the effect is only significant for men in the category for Primary education. Compared to having Secondary education, Primary education has a negative effect on mental wellbeing for men. For women the effect is only significant for Tertiary education, which has a positive effect on mental wellbeing compared to only having Secondary education.
For the partnership status variable almost all groups are significantly and consistently better of in regards to mental wellbeing by having a partner, also in the case of having been divorced or widowed. Women appear to have a stronger positive effect on mental wellbeing from being married than just having a partner. Men also have a small but negative and significant effect on mental wellbeing by being divorced or widowed and being single.
For the interaction term the reference category is Sweden#No children. For women the interaction effect between country and parental status has no significant effects. For men the effect is significant for the category French men with children. Compared to Swedish men with no children, French men with children have better mental wellbeing. The co-residential child variable is only significant for men in model A. Having a co-residential child has a negative effect on mental wellbeing for men. For women this effect is not significant.
Satisfaction with the relationship to one’s children has a significant effect for men and women for the categories 2 and 3. This indicates that a lower level of satisfaction with the relationship to ones children has a negative effect on mental wellbeing. For the variables experienced the death of a child, daughter, child attitude and job attitude, the effects are all significant for women. The variables experienced the death of a child and job attitude is significant for men. Both men and women have worse mental wellbeing if they have experienced the death of a child.
The results indicate that there are greater differences between the genders in regards to
parenthood and mental wellbeing than there is between the two countries.. Social context does not appear to be a major indicator of mental wellbeing in the case of women. The results for men are not significant, but the positive relationship between CES-D and parenthood is not present.
