Is providing informal care a path to meaningful and satisfying ageing?

Almost every population around the world is currently ageing, due to the demographic transition from high fertility and high mortality to low fertility and low mortality (Timonen 2008). Active ageing is both a concept and a policy that has been developed in response to these rapid changes in the age structures of populations in Europe and elsewhere. The active ageing approach supports several types of activities including caregiving in order to maximise the quality of life (QoL) of older people. The consequences of all of these activities are expected to be predominantly positive; no variations in their aspects or impacts have been distinguished by many official documents (EU Council 2012; European Commission 2013; Marsillas et al. 2017). Still, it can be supposed that, for example, the provision of care is often exhausting or dictated by circumstances. Caregiving has the potential to raise life satisfaction among older people, but only under certain conditions (Moen et al. 1995). Some of the most important conditions are mentioned in this paper and incorporated into the analysis.

This study investigates which dimensions of subjective QoL are connected to the provision of informal care and how this connection depends on the intensity, multiplicity, and voluntariness of the caregiving. Data from the Survey of Health, Ageing and Retirement in Europe (SHARE) are utilised to deal with this topic. The analysis, based on the data from 14 European countries, provides several contributions to existing research. First, the indicators of the provided care reflect all of the types of care provided to all potential recipients instead of just one type of care, and they also reflect the intensity and multiplicity of care. Second, the cross-national nature of the data makes it possible to control for the availability of formal care and its variations across Europe. Finally, some specific feelings that are important in later life (loneliness, meaningfulness of life, and overload) are examined and compared to a general and multidimensional measurement of QoL used in most studies.

The basic assumption of the active ageing theory is the notion that a higher level of participation leads to increased personal health and well-being (Marsillas et al. 2017; Walker 2009). This notion, supporting the performance of more roles in later life, is highly compatible with the main idea of the theory of role accumulation. This theory argues that role accumulation has mostly positive outcomes (Sieber 1974). The beneficial effect of additional roles is supported by some empirical research. In particular, studies focused on the effect on life satisfaction of provision of care at older ages have found a positive effect of caregiving compared to non-caregiving (De Jong Gierveld and Dykstra 2008; Potočnik and Sonnentag 2013).

Not all authors agree with the beneficial effect of additional roles for older adults. According to the theory of role strain, involvement in multiple roles leads to conflicts in demands or obligations among competing roles under normal conditions (Goode 1960). Role strain for older caregivers is well-documented in empirical studies (Colombo et al. 2011; Mui 1995; Reid and Hardy 1999). It is evident from the results of these studies that the effect of caregiving (as well as many other roles) depends on the context. Moen et al. (1995) and Penning (1998) reach this same conclusion in their studies.

The effect of providing care on QoL in older carers indeed depends on specific conditions. This topic has not been systematically examined, but it seems that caregiving has more positive outcomes when it is less intensive (Broe et al. 1999; Broese van Groenou et al.2013; Colombo et al. 2011) and more voluntary (Broese van Groenou et al.2013; Lowenstein et al. 2008), while the effect of multiplicity of care remains unclear (Chassin et al. 2010; De Jong Gierveld and Dykstra 2008). The next section explains these factors in more detail.

Several studies illustrate that intensive care has a detrimental effect on the consequences of caring (Broe et al. 1999; Broese van Groenou et al.2013; Colombo et al. 2011). A direct explanation for this effect is the fact that more intensive caregiving consumes more time and energy (Broese van Groenou et al.2013). Additionally, more intensive provision of care is often dictated by circumstances of less accessible formal care (Igel et al. 2009) or of fewer family members accessible for sharing family responsibilities (Broese van Groenou et al.2013). These findings explain why more intensive care is associated with a higher burden and other negative consequences. Occasional care does not seem to significantly change subjective QoL (Broe et al. 1999; Colombo et al. 2011), and thus, this paper investigates the effect of intensity by comparing the effect of regular care performed about every week or more often to the effect of very intensive care performed about daily.

The multiplicity of care has an unclear impact on QoL. Multiple caregiving is beneficial according to De Jong Gierveld and Dykstra (2008) and Pines et al. (2011), but harmful according to studies by Chassin et al. (2010) and Remennick (1999). The findings seem to imply that the nature of the effect of providing care to multiple recipients is connected to the intensity of the care provided (De Jong Gierveld and Dykstra 2008).

Unfortunately, most of the cited studies do not mention a precise minimum intensity of care provided by those respondents defined as ‘caregivers’. This intensity, in various studies, seems to range from at least one hour per week (Chassin et al. 2010), to co-residence with children and at least a few hours of help to parents per week (Pines et al. 2011), to very intensive care for more family members (Remennick 1999). Hence, previous inconsistencies among studies can be tentatively explained by stating that multiple caregiving is beneficial only to a certain level of intensity of care. This paper is interested in the regular provision of care; accordingly, the effect of multiple caregiving is expected to be negative.

The effects of both intensity of care and multiplicity of care indicate similar characteristics of care, and thus, it might be redundant to investigate both. However, it seems that multiple caregiving is of a different quality than merely the sum of its frequencies, because more roles often exist in subjectively perceived tension or conflict (Boudiny 2013; Coverman 1989; Goode 1960), and a negative effect of multiple caregiving on health has been found even if the total amount of provided care is controlled (Chassin et al. 2010). Hence, this paper aims to examine the caregiving situation in its complexity by elaborating on both intensity of care and the number of caregiving relationships. Further, a broad definition of care provided within or outside the household to all possible family and nonfamily members is employed, so no caregiving activity is omitted in the analysis.

Another factor of the effect of caregiving on its outcome is the relationship between the caregiver and the care receiver, which has been found by other studies (Broese van Groenou et al.2013; De Jong Gierveld and Dykstra 2008). This study is not able to include this factor into its already complex models due to high collinearity with other variables. This step is legitimate, as the paper examines the overall effect of caregiving as an activity promoted by active ageing policy, and thus the analysis needs to control only variables confounding this effect. Still, the relationship to care receiver is the important predictor of caregiving outcomes for specific individuals.

The last important factor of care addressed is voluntariness of care. Voluntary care means that a caregiver can manipulate role obligations and have some alternative options for satisfying them. The opposite is the case under the synonyms of involuntary (Schmid et al. 2012; Szinovacz 2008), normative (Lowenstein et al. 2008), or obligatory (Igel et al. 2009) care. Involuntary care exists mostly in situations in which formal care is not available and a caregiver is forced to provide (often intensive) practical help to someone close. Involuntary care could be a consequence of a lack of caregivers in the family (Broese van Groenou et al.2013) or insufficient access to formal care. Involuntary care is more prevalent in countries with less available social services and stronger familial norms, in which individuals are more often pushed by circumstances into a more demanding provision of care (Albertini et al. 2007; Igel et al. 2009).

The voluntariness of provided care seems to have a positive effect on care outcomes (Broese van Groenou et al.2013; Lowenstein et al. 2008). However, an individual-level indicator of this concept is not available in the utilised dataset. Therefore, the voluntariness of provided care is indicated by the macro-level factor of availability of formal care, since a higher availability of formal care in a country creates an alternative for informal caregivers, and provision of care is thus more voluntary (Albertini et al. 2007; Igel et al. 2009; Motel-Klingebiel, Tesch-Roemer, and Von Kondratowitz 2005).

A general aim of many policy programmes focused on older people – including active ageing policy programmes – is to increase the QoL of the target population (Walker 2009). However, QoL is not easy to define, as it has been conceptualised and measured in many ways. Even with a focus on the subjective assessment of life in general, many concepts, such as subjective QoL, life satisfaction, well-being, psychological distress, happiness, mood, and affect, have been interpreted in very similar ways and used almost as synonyms (Katz 2009; Mcilvane, Ajrouch, and Antonucci 2007; Roll and Litwin 2013). All of these concepts are general, vague, and potentially encompass several domains or dimensions (Bowling 2005).

This paper argues that, besides the extensive existing research examining life satisfaction generally, it is useful to research three specific domains of subjective QoL in older ages. First, De Jong Gierveld and Dykstra (2008) have found that (multiple) care provision substantially reduces loneliness, which is a very prevalent phenomenon in older ages (Shiovitz-Ezra 2013), strongly related to depression (Tsai et al. 2013). Second, meaningfulness of life is another important part of well-being and other desired outcomes of later life (Bauer and Park 2010; Boyle et al. 2009; van Kessel 2013; Krause 2009; Nygren et al. 2005). The effect of caregiving on the meaningfulness of life may be positive, as providing care seems to be one of the possible answers to existential issues in late mid-life (Pines et al. 2011), and it enhances meaningfulness of life by fulfilling social roles and personal ethics (Di Novi et al. 2015). Third, previous research has found an association of intensive care with higher strain and burden (Broese van Groenou et al.2013; Colombo et al. 2011; Reid and Hardy 1999), which may indicate a negative association between caregiving and overload corresponding to role strain theory (Goode 1960). Fourth, the effect of caregiving on life satisfaction as a vaguer indicator of QoL is compared to effects on the three specific domains. Generally, the effect of providing regular care on loneliness, meaningfulness of life, and overload is expected to differ from the predominantly not significant (Baydar and Brooks-Gunn 1998; Siegrist and Wahrendorf 2009) or negative (Broe et al. 1999; Colombo et al. 2011; Ekwall et al. 2004) effect on subjective QoL in its general and multidimensional conceptualizations.

This study examines the association between caregiving and three dimensions of subjective QoL. The paper hypothesises that the regular provision of care reduces loneliness and increases meaningfulness of life (role enhancement hypothesis) but also increases perceived overload (role strain hypothesis), while there is no effect of caregiving on life satisfaction. Furthermore, it is expected that the effect of caregiving on all three aspects of subjective QoL is more negative/less positive in situations in which the caregiver provides more intensive care and in which the caregiver provides care to more recipients. This paper also hypothesises that the effect of caregiving is more positive/less negative in countries with higher availability of formal care, because the provision of care is more voluntary and optional under this condition.

This paper utilises data from the Survey of Health, Ageing and Retirement in Europe (SHARE). SHARE is a panel cross-national project in which data are collected by computer-assisted personal interviews (Börsch-Supan 2017; Börsch-Supan et al. 2013). This paper uses data from wave 5, which enables the analysis of recent data collected from large national samples from various parts of Europe at a single time point (Stolz 2015). Data from all 14 European countries participating in wave 5 (Austria, Belgium, Czechia, Denmark, Estonia, France, Italy, Germany, Luxembourg, Netherlands, Slovenia, Spain, Sweden, and Switzerland) are analysed in this study. Response rate in wave 5 varies between 34 and 61 percent for individual countries (Bergmann et al. 2017).

Descriptive statistics, including a list of the countries and their sample sizes, are included in Table 1. About four percent of respondents from the sample have been dropped due to missing values and the spouses of original respondents under the age of 50 have also been omitted from the analysis. Then, respondents aged 90+, representing approximately one percent of the sample, have been excluded to avoid distortion by extreme values. If there are more respondents for one couple, one person called the ‘family respondent’ is defined as the respondent who was interviewed first and answered the modules about care and relatives on behalf of her/his household. Every household in this study is represented by the family respondent. Hence, people from households with more than one respondent aged 50+ are not overrepresented in the final sample.

The entire sample consists of 40,748 respondents after deletions. The sample size for each country can be reconstructed from the size of the overall sample and the proportions of the countries in the percentages displayed in Table 1 – the number of respondents per country varies from 1160 respondents for Luxembourg to 3770 respondents for Belgium. Other variables have plausible distributions and quite comprehensible values. The distribution of age is positively skewed among 51 and 90 years with a mean of 67 and a median of 66. Most of the independent variables are measured at the level of individuals, but one country-level variable and one cross-level interaction are also included in the analysis. Therefore, individuals and countries constitute two levels in a multilevel data structure. The only variable at the level of countries – the percentage of employees in social services – has been standardised, and thus does not have comprehensible values. Nevertheless, the highest ratio of these employees is predictably in Sweden, the Netherlands, and Denmark, while Estonia, Slovenia, and Czechia have the lowest values.

Three single-item variables are employed to indicate the three specific dimensions of subjective QoL and life satisfaction. Loneliness is measured by the question ‘How much of the time do you feel lonely? Often/Some of the time/Hardly ever or never’. This single-item measurement of loneliness is a valid measurement of loneliness, which highly correlates with the 20-item original UCLA scale of loneliness (Aartsen and Jylhä 2011; Park et al. 2013; Theeke 2009). Questions about the meaningfulness of life and overload are items from the CASP-12 scale measuring subjective QoL at older ages; items from this scale have been used repeatedly as indicators of specific concepts (Jansen et al. 2010; Towers et al. 2015). The item ‘I feel that my life has meaning’ indicates meaningfulness of life and a negative evaluation of the item ‘I can do the things that I want to do’ indicates overload in the analysis. Both items have options ‘Often/Sometimes/Not often/Never’. Finally, the life satisfaction is measured by the question ‘On a scale from 0 to 10 where 0 means completely dissatisfied and 10 means completely satisfied, how satisfied are you with your life?’. The original scales for indicators of loneliness and meaningfulness have been reversed, so higher values indicate higher loneliness, higher meaningfulness, and higher overload in the analysis.

The provision of care is measured by two different variables, which have been constructed from the SHARE module ‘Social support’, which examines the provision and intensity of caregiving to parents and parents-in-law, spouses, children, grandchildren, and many other relatives, as well as to friends and other nonrelatives. Three questions on caregiving are used to construct this broadly defined indicator of care provision: 1. ‘Which family member from outside the household, friend or neighbour have you helped in the last twelve months?’ with activities like ‘dressing, bathing or showering, eating (…), home repairs, gardening, transportation, shopping, household chores or help with paperwork (…)’; 2. ‘During the last twelve months, have you regularly or occasionally looked after your grandchild without the presence of the parents?’ combined with ‘Which of your children is the parent of the grandchild you have looked after?’; and 3. ‘Is there someone living in this household whom you have helped regularly during the last twelve months with personal care, such as washing, getting out of bed, or dressing?’

Each of three questions is followed by request to specify the intensity of activity. These questions measuring the number and intensity of caregiving relationships have been converted to an indicator of care provided about every week or more often (values are 0, 1, 2, 3, or 4 and more indicating a number of caregiving relationships of this intensity) and care provided about daily (values 0, 1, 2, or 3 and more indicating a number of caregiving relationships). These two explanatory variables directly determine the effect of care multiplicity by their values and indirectly obtain the effect of intensity of care by comparing their coefficients.

The two variables of caregiving described above indicate the frequency and multiplicity of care, while the percentage of employees in health and social services indicating the availability of formal care at the national level (OECD 2017) indirectly indicates the voluntariness of care on the individual level. This variable is ‘the best measurable public substitute for practical support between generations’ (Brandt and Deindl 2013: 240) and apparently also the most frequently used one (Brandt and Deindl 2013; Igel et al. 2009; Schmid et al. 2012), as it focuses on time transfers instead of financial ones. The only macro-level variable has been standardised to make fixed effects more easily interpretable.

All of the models in the analysis contain the cross-level interaction between the availability of formal care and the care provided by respondents to find the differences in the effect of caregiving by the availability of formal care. The provision of care in the interaction is treated as a continuous variable, even though the main effect is treated as categorical. The linearity of the interaction has been checked. This step makes the results easily interpretable and more appropriate for testing the hypotheses without disadvantaging the models (Francoeur 2011; Pasta 2013).

Various continuous and categorical control variables are used in the analysis, and almost all of them are characteristics of respondents measured at the first level of the data structure. Most of the continuous controls are measures of family structure: number of surviving parents, number of siblings, number of children, and number of grandchildren. These four variables are top-coded – approximately one percent of the highest and rarest values are merged into one category to eliminate outliers. The last measure of family structure is a binary variable partner lives in the same household (yes/no). Family members potentially increase the demands of caregiving, but they can also lower the care burden for the respondent through their provision of care (Di Novi et al. 2015). Just two other continuous control variables remain on the list: age and age squared of the respondent. All continuous variables have been centred on their mean to make the results easily interpretable.

The remaining categorical controls are the following characteristics of the respondent: sex (male/female), subjective health status (the category ‘unhealthy’ consists of the categories poor and fair, the category ‘healthy’ from the options good, very good, and excellent; this dichotomisation is widely used and does not affect results, see e.g. Contoyannis and Jones 2004; Di Novi et al. 2015; Piko 2007), employment status (not working, working part-time, working full-time), and finally, the educational level of the respondent (the ISCED scale is divided into three categories 0–1, 2–4, and 5–6; this categorisation is the most meaningful, according to the International Labour Office 2012).

More macro-level variables than the indicator of the availability of formal care could have been added to the analysis as control variables to make use of the multilevel data structure. However, the most relevant macro-factors are closely interrelated, and multicollinearity appears to represent a serious issue. For instance, the percentage of employees in social services correlates 0.64 with GDP per capita, −0.79 with the strength of caregiving norms, and −0.84 with the level of familialism. None of these or the other macro-factors, such as Gini coefficient, employment rate over 55, social expenditures, social participation in later life, and legal obligation of intergenerational care, improve the models or produce significant findings in combination with the percentage of employees in social services.

The analysis uses multilevel ordinal logistic regression for loneliness, meaningfulness of life, and overload and multilevel linear regression for life satisfaction. This study works with the large sample, and thus uses only significance levels p < 0.01 and p < 0.001. The models for the three domains of QoL and life satisfaction predicted by caregiving provided about every week or more often and by all control variables are presented in Table 2. The effect of caregiving is beneficial with all coefficients for the meaningfulness of life and most of the coefficients for loneliness and overload being significant. Furthermore, the models contain an approximately linear effect of caregiving illustrating the beneficial effect of multiple provision of care, though some coefficients for multiple caregiving are not significant due to the low number of cases. The strongest beneficial effect of caregiving is the effect on the meaningfulness of life, followed by the effects on loneliness and overload in descending order, although the differences are small. In contrast, none of the coefficients of care activities is significant in the model for life satisfaction.

Most of the parameters in the four models are similar for loneliness, meaningfulness of life, overload, and life satisfaction. The age of the respondent is related positively to a higher QoL with a reversal of that effect in very old age. Women feel lonelier than men, but perceive their lives as more meaningful and satisfactory; they do not differ in overload from men. Generally, beneficial effects are indicated by better health, higher education, active employment, and a partner in the household, with the last factor being especially important in reducing loneliness. Other coefficients of the variables indicating family structure are mostly not significant, except for a positive effect of a number of children on the meaningfulness of life.

The effects of special interest are represented by the coefficients of the macro-factors, and especially the coefficients of the cross-level interaction. The higher proportion of employees in health and social services increases meaningfulness of life together with life satisfaction and reduces overload. Then, the only significant coefficient of the cross-level interaction is present in the model for loneliness. An increase in the accessibility of formal care for one standard deviation decreases the beneficial effect of caregiving on loneliness by 0.051, which makes the coefficient −0.032 for one caregiving relationship under this condition. Hence, older adults generally experience higher meaningfulness of life, higher life satisfaction, and less overload in countries with more developed welfare services, while the effect of caregiving is more negative for loneliness in countries with higher availability of formal care. Therefore, more voluntary care is in a sense more harmful for one effect of caregiving activities, while it is not important for others.

The same three models for loneliness, meaningfulness of life, overload, and life satisfaction as dependent variables, but with the number of care activities performed about daily as the main explanatory variable, are presented in Table 3. None of the beneficial effects of caregiving from the previous models appears in this multivariate analysis, which means that very intensive provision of care does not improve QoL in these indicators. The only exception from this conclusion is the coefficient −0.125 for one caregiving activity in the model for life satisfaction, but this effect does not work for more activities. Despite this disappearance of all effects, the effect of multiple caregiving on loneliness and meaningfulness of life remains approximately linear, though even coefficients 0.331 and 0.337 are not significant due to a very small prevalence of multiple caregiving.

The nature of other predictors of QoL indicators is basically the same as for the models with care activities performed about every week or more often. The subjective QoL is enhanced by higher age (nonlinear effect), education, active employment, health, partnership, availability of formal care, and, in terms of the meaningfulness of life, by higher numbers of children. Again, women perceive higher loneliness, meaningfulness, and life satisfaction than men.

The effect of formal care accessibility remains beneficial for two domains of QoL – meaningfulness of life and overload – and for life satisfaction as vaguer and multidimensional indicator. Regarding the parameters of cross-level interaction, the effect of caregiving on loneliness remains less pronounced in countries with a higher proportion of employees in social services, which is again the only significant cross-level interaction. In sum, a stronger welfare state has a rather harmful effect in the association between care provided daily and loneliness with higher reduction of loneliness by caregiving in countries with less available formal care.

This study has investigated the effect of caregiving on three domains of QoL among older people. This paper contributes to this frequently researched topic in the following ways: (a) care for all possible types of recipients is considered; (b) the frequency, multiplicity, and voluntariness of caregiving are included in the analysis as characteristics of care, mediating its outcomes; (c) the effect on specific domains of QoL relevant for later life is compared to the effect on life satisfaction in general; and (d) the effect of macro-context is investigated in the analysis through the availability of formal care as a precondition for voluntary provision of care. The paper uses data from SHARE and draws the following conclusions.

First, the effect of regular caregiving on the three domains of QoL is generally beneficial. The multivariate association between provision of care and desirable outcomes in terms of reduced loneliness, more meaningfulness in life, and lower overload is positive. It is not possible to empirically test a causal relationship here, but the theory and previous research claim that this association mostly works in terms of the caregiving influencing subjective QoL (Colombo et al. 2011; De Jong Gierveld and Dykstra 2008). In contrast, the study did not find any effect of caregiving on life satisfaction in general. Most of these findings are according to expectations, but the fact that the provision of care reduces overload contrasts with the assumption of the theory of role strain (Colombo et al. 2011; Goode 1960). The results indicate that the theory of role accumulation is more suitable for explaining caregiving in older ages than the theory of role strain.

Second, the multivariate association between the provision of care and QoL depends on characteristics of the provided care. The effect of care provided about every week or more often is strongly beneficial, while the effect of less prevalent activity – daily caregiving – is not significant. Moreover, the enhancing effect is more pronounced for the provision of care to more recipients – more care activities result in a more beneficial effect. This finding is in accordance with the study by De Jong Gierveld and Dykstra (2008), which interprets lower loneliness among people supporting more generations in terms of altruistic theory. The enhancing effects of multiple caregiving in this study could be interpreted in terms of maintaining more close contacts, performing meaningful activities, and following cultural norms (Neuberger and Haberkern 2014).

Third, the effect of caregiving on loneliness is consistently less enhancing under the conditions of more available formal care. It is possible that a person providing very intensive care in the context of available formal care may differ from the majority in terms of cultural norms, social characteristics or availability of the social system (Fokkema et al. 2008; Igel and Szydlik 2011; Neuberger and Haberkern 2014). The fact that the effect of accessibility of formal care on QoL is consistently beneficial may support the usefulness of a strong welfare state. In any case, more research on the complex interactions of micro- and macro-factors is needed.

This study has two limitations and two possible future directions of the research, which stem from the limitations. The first limitation concerns the omission of the caregiver-care receiver relationship information from the analysis, which would require to include additional concepts and many new statistical models and is out of the scope of this paper. Daily care for a spouse has different consequences than daily babysitting of a grandchild (Broese van Groenou et al.2013; De Jong Gierveld and Dykstra 2008), so this factor is the important predictor of the individual outcome of caregiving. However, this article has aimed at testing the overall effect of caregiving as a set of possible roles among older adults, as the active ageing approach defines the consequences as generally positive (European Commission 2013; Zaidi 2015). Furthermore, the analysis has already proven to be context-sensitive in relation to other factors – caregiving intensity, multiplicity, and voluntariness, family structure, etc. These findings thus may be utilised by future research focusing more on the caregiving relationship than on other contextual factors.

The second limitation originates from the fact that it is not possible to indicate the level of voluntariness of care more directly than via an external macro-level variable with the use of SHARE data. It is not possible to determine empirically how well this macro-factor indicates the voluntariness of informal care. This study at least outlines some possible effects of more voluntary care, which may increase loneliness through the elimination of some incentives to meet other people. However, this and the other findings and ideas of this paper must be tested with other methods and data sources in order to further validate them.

This paper uses data from SHARE Wave 5 (DOI: 10.6103/SHARE.w5.100), see Börsch-Supan et al. (2013) for methodological details. The SHARE data collection has been primarily funded by the European Commission through FP5 (QLK6-CT-2001-00360), FP6 (SHARE-I3: RII-CT-2006-062193, COMPARE: CIT5-CT-2005-028857, SHARELIFE: CIT4-CT-2006-028812) and FP7 (SHARE-PREP: N°211909, SHARE-LEAP: N°227822, SHARE M4: N°261982). Additional funding from the German Ministry of Education and Research, the U.S. National Institute on Aging (U01_AG09740-13S2, P01_AG005842, P01_AG08291, P30_AG12815, R21_AG025169, Y1-AG-4553-01, IAG_BSR06-11, OGHA_04-064) and from various national funding sources is gratefully acknowledged (see www.share-project.org).

No potential conflict of interest was reported by the author.

Martin Lakomý is a Ph.D. student of sociology at the Faculty of Social Studies of Masaryk University. He uses quantitative methods to research population ageing, changes in fertility behaviour, and their connection to value changes.

Martin Lakomýhttp://orcid.org/0000-0003-4625-4463