‘The fourth commandment effect’: church attendance and intergenerational support in late parent–child relationships

Because of increased life expectancy, intergenerational relations are becoming more and more important for the well-being of family members. Parents and their children spend longer years of shared lives and represent an enduring source of mutual support (Bengtson 2001). On the one hand, parents support their adult children, providing housing, financial transfers or help in grandchild care (Albertini et al. 2017; Brandt and Deindl 2013). On the other hand, children become a fundamental source of assistance when parents are in need (Künemund and Rein 1999). Although older people enjoy good health conditions for long parts of their lives, the ageing process inevitably increases adult children’s commitments to provide in-kind support to their older parents (Merz et al. 2008; Gautun and Bratt 2017). In this light, a large and growing body of research focuses on individual factors affecting adult children’s propensity to provide support and care to their older parents. However, with some relevant exceptions (Myers 2004), few studies have examined whether religious values and practices are associated with supportive behaviors in late parent–child relationships.

Over the half past century, religious participation has declined in many Western countries (Biolcati-Rinaldi and Vezzoni 2013; Crockett and Voas 2006). The widespread process of secularization has led toward a minor involvement of younger generations in religious institutions and to a greater individualization of religious practices. Some scholars suggest that the diminishing influence of institutional religious messages along with processes of individualization is related to lower pro-family attitudes and weakening relations between parents and their adult children in Western societies (Norris and Inglehart 2007; Komter and Vollebergh 2002). As a matter of fact, the cultivation and maintenance of strong intergenerational familial bonds is a fundamental teaching transmitted by religious institutions.

In the present study, we examine the role of church attendance as a potential factor influencing the provision of support from children to older parents in the Italian context. The country is characterized by a homogeneous religious market where Catholics constitute about 90% of the total population (Introvigne and Stark 2005; Vezzoni and Biolcati-Rinaldi 2015). Given that Catholics are explicitly obliged to honor the Lord’s Day by attending the Sunday ritual, religious practices and values are often seen as two sides of the same coin. On the one hand, church attendance is an indicator of religious values and norms which encourage members to adopt altruistic behaviors and prioritize the good of the family over individual gratification. On the other, religious institutions and communities foster traditional beliefs and practices regarding family life, by providing formal messages, seminars and teachings on family bonds as well as informal activities and assistance to families in stress.

Italy represents an interesting case in point: is one of the oldest countries in the world, and public policies provide few alternatives to outsource family responsibilities (Nazio and Saraceno 2013; Ferrera 1996). Italian sub-protective welfare state is characterized by low level of investment on elderly support services, which in turn increases the responsibility for adult-children to provide help and care to their parents ‘by default’ (Daatland and Lowenstein 2005; Saraceno and Keck 2010). Besides, in ‘strong family ties’ settings, the family is historically the main institution for protecting vulnerable individuals (Reher 1998; Dykstra and Fokkema 2011; Glaser et al. 2004) and family obligations toward parents are generally widespread among the population, especially when parents are in serious need (Kalmijn and Saraceno 2008). In addition, the presence of the Vatican City contributes to exert an influence on national mass media, public schools and Italian governments (Vignoli and Salvini 2014). Religious messages permeate several aspects of social life, arguably reinforcing a sense of belonging to Catholic community and promoting social cohesion within the family. The absence of a strong protective welfare state along with a culture of strong family ties and the Catholic religion may lead, for different reasons, Italian adult children to help their parents.

Religiosity influences several aspects of family life, such as union formation (Corijn and Klijzing 2001; Xu 2005), sexual and reproductive behavior (Caltabiano et al. 2006; Wilcox and Dew 2016), parenthood decisions (Zhang 2008) and socialization practices (Mahoney 2010). As Mahoney (2010) illustrates in her exhaustive analysis of the literature, religious values and practices affect and structure all stages of family life since the very beginning. In the U.S. churchgoing mothers tend to establish warmer relationships with their young children, and religious fathers devote a greater amount of time to them, compared to their non-religious counterparts (Wilcox 1998, 2002). The evidence suggests a higher involvement of churchgoing parents in children’s lives, regardless of their specific religious affiliation (Petts and Knoester 2007). Positive aspects of religious participation seem also to be carried over into later parent-adult child relationships. For example, Kalmijn and Dykstra (2006) find that in the Netherlands, parents who attend Catholic rituals are more likely to maintain frequent face-to-face contacts with their adult children compared to non-religious people. In a similar vein, considering a three-generation perspective, King and Elder (1999) point out that social and private dimensions of religiousness including the participation to religious services are positively correlated with grandparents’ involvement in childcare.

Concurrently, religious affiliation and participation affect the way in which adult children relate to their older parents. We distinguish four main mechanisms to explain how church attendance may promote intergenerational family support. Firstly, it is well known that attendance to religious services is an indicator of people’s orientations, including values regarding parent–child bonds and family relationships (Mahoney 2010; Norris and Inglehart 2007). Religious beliefs and values toward the family constitute a stock of internalized social values that may foster traditional family behaviors and feelings of solidarity between parents and their children (Silverstein et al. 2012). Churchgoers may be more likely to assist their older parents, because of their values regarding parental devotion, filial obligation and gratitude.

Secondly, the frequency of ritual attendance may play a role in reinforcing religious messages and recommendations. Religious institutions often transmit messages about parental respect and devotion through church teachings, sermons, and publications. As Silverstein et al. (2012) argue, the imperative to honor one’s mother and father can be traced in all religious tractates. The extent to which adult children are exposed to religious messages about parental devotion varies according to their involvement in church rituals and religious organizations. Attending services is a means through which adult children interiorize religious teachings such as parental respect, pro-family attitudes and the importance of maintaining intergenerational solidarity (Gans, Silverstein and Lowerstein 2009; Myers 2004).

Thirdly, religious institutions often provide services and programs to assist persons and families in need. These organizations tend to instil obligations of providing support and stress the importance of altruism and help outside and within the family (King and Elder 1999). Previous research consistently shows that altruism promotes intergenerational family transfers and thus encourages adult children to provide care assistance and support to their older parents (Albertini et al. 2007; Kohli and Künemund 2003).

Fourthly, as Lim and Putnam (2010) argue, regular churchgoers tend to create homogeneous networks that reinforce their religious identity. A strong sense of belonging to religious community may then strengthen religious beliefs of love, gratitude, forgiveness, along with prescriptions of filial respect and devotion. Interactions with coreligionist may also foster social norms regarding intergenerational bonds. The norm that one should love and respect one’s parents may be translated into stronger feelings of responsibility and obligation toward parents. Previous research findings indicate that in many Western countries religious adult children report having stronger feelings of filial obligation than the non-religious (Daatland and Herlofson 2003; Dykstra and Fokkema 2012; Gans et al. 2009; Killian and Ganong 2002; Liefbroer and Mulder 2006). It is unclear, however, whether filial responsibility makes churchgoing adult children more prone to provide support to older parents, compared to their non-catholic counterparts. In fact, Myers (2004) highlights that in the U.S. church attendance per se is not significantly related to intergenerational assistance, while the effect of religious denomination is marked only among religious congruent parent–child relationships.

While religious practices may not be related to supportive behaviors between non-co-resident family members, churchgoers may be more likely to postpone residential independence and returning home. Some studies underline the importance of religion in terms of leaving home decision and marriage particularly in Catholic countries (Corijn 1999). Because of their family orientations, Catholic adult children may leave the family nest at an older age than non-Catholic children and/or may welcome their older parents moving home when necessary. Shelter is an important form of support that Italian family members traditionally prefer over inter-household transfers (Albertini et al. 2017; Tosi 2017a, 2017b). Although this type of support may flow in both directions, i.e. form children to parents and vice versa, we assume that leaving or returning home is mainly a child’s decision related to his/her beliefs.

Churchgoers and non-religious individuals are often characterized by opposite attitudes. In general, the former tend to adopt conservative views on several aspects of family life, hence promoting traditional spousal roles within the household. The latter, conversely, are usually more libertarian and individualistic. Some scholars have coined the term ‘religious familism’ for indicating how in the U.S. religious people tend to defend a view of the heterosexual married couples with children as the central unity of societal order (Mahoney 2010). According to this view, the family is often perceived as the ‘natural’ source of help for older people (Daatland and Herlofson 2003). These conservative attitudes toward the family may be related to parental respect and may encourage adult children to maintain high levels of intergenerational solidarity, including geographically close relationships, frequent contacts and exchange of various forms of help with their parents (Silverstein and Bengtson 1997). Regular churchgoers may be more likely to have a traditional view of family life, which includes the idea of the provision of support as a behavioral expression of filial responsibility as well as traditional attitudes toward gender roles.

Religious beliefs and practices are usually formed within the family of origin during childhood and adolescence. Both religious practices and family obligations are often transmitted through the channel of the family of origin and tend to persist throughout the life course (Bengtson et al. 2009; King et al. 2006; King and Elder 1999). Religious family members are likely to attend the same church rituals and are engaged in religious celebrations. Church activities and rituals offer the opportunity for family members to share experiences, thus fostering family norms about the importance of maintaining strong family ties (Kalmijn and Dykstra 2006). Therefore, the association between religious service attendance and intergenerational support may reflect people’s family oriented socialization and the quality of their relationships. As previously mentioned, Myers (2004) finds support to this idea, showing that parent–child consensus on religious values as a sign of a successful socialization promotes high levels of assistance between generations. However, the effect of religious similarity in parent–child relationships can be debated. For example, Liefbroer and Mulder (2006) point out that in the Netherlands individuals’ religious affiliation is much more relevant than parent–child religious congruence in explaining filial responsibilities toward parents.

Also gender differences in religion and intergenerational relationships have been traced back to socialization practices. Some studies indicate that women have a greater commitment to religion partly because of their socialization toward expressive values, such as conflict resolution, that are congruent with religious emphases (De Vaus and McAllister 1987; Francis 1997). Female socialization also induces a greater involvement in family life. Women, as kin-keepers in the family, are typically more involved in managing intergenerational relationships and providing support (Kalmijn 2007; Silverstein et al. 2002; Tosi 2017a; Tosi and Gähler 2016). Gender-specific pattern in religious commitment and family relationships could then suggest that church attendance affects intergenerational support at greater extent for daughters than for sons.

In light of the results from previous studies on religious participation and parent–child relations, we formulate the following hypotheses. Firstly, we argue that the frequency of church attendance exposes adult children to prescriptions about parental devotion and pro-family messages as well as feelings of altruism and belonging in a community where family bonds are valued. Thus, we expect to find that regular churchgoers are more likely to live with and provide support to parents than not-Catholic people do, with occasional churchgoers falling somewhere in between (Hypothesis 1).

Secondly, we attempt to partly account for individual traditionalism considering individual orientations about marriage and divorce. Notably, traditional family orientations include attitudes toward marriage, divorce and other aspects of family life that may only partly be connected to the intergenerational exchange of practical support. Hence, our second hypothesis states as follows: the association between church attendance and intergenerational co-residence and support from children to parents is partly explained by traditional attitudes toward the family (Hypothesis 2).

Thirdly, churchgoers may be more likely to have close relationships with parents and their overall family network, compared to individuals who never attend religious rituals. Close relationships with other relatives may be an indicator of orientations toward the extended family network (Kalmijn and Dykstra 2006). Hence, we expect that the association between church attendance and intergenerational co-residence and support from children to parents is partly explained by having close relationships in the family network (Hypothesis 3).

The empirical analysis is based on the survey Family and Social Subject (FSS) that took place in 2003. This is the Italian version of Generation and Gender Survey (GGS) and is a five-year module of the Multipurpose survey conducted by Italian National Statistical Office (Istat). The survey is representative of the Italian population and involves more than 19,000 families (almost 50,000 individuals). The database contains detailed information about parental characteristics and intergenerational relations including proximity, contact frequency and support exchange.

The total sample is restricted to individuals who had at least one of the two parents aged 64 or older to investigate late parent–child relationships (mean equal to 75, see Table 1). We also exclude respondents and their parents who are born or live abroad (3.3%), since foreigners and people living abroad may be less affected by Italian religious context. Data on parents are gathered separately for mothers and fathers which allows us to consider two dyads for each respondent when both parents are alive. Adopting the parent–child dyad as unit of analysis makes it possible to examine a broader set of variables and parents’ characteristics than conventional analyses based on respondents. For example, adult children may have different relationships with their mothers and fathers when divorced or in case of declining health conditions of one of the two parents. The final sample can count on 7,157 sons and 7,818 daughters, and 10,098 parent-son dyads and 10,930 parent-daughter dyads.

In order to construct our dependent variables, we consider the question asking whether the respondent has provided support to someone living outside the household in the last four weeks. Among respondents who provided support, further questions are asked regarding the type of the most important support and to whom it was provided. Our measure of intergenerational support refers to the most important help provided by the respondent to his/her father and/or mother. It is important to note that adult children might provide multiple forms of support to multiple receivers without considering the support given to parents as the most important one. However, we are confident to include all transfers of time in our analysis. Indeed, as it is possible to notice from Table 1, the diffusion of intergenerational support from adult children to older parents is in line with previous findings on other databases (Albertini et al. 2007). Among the original ten forms of support, we distinguish between practical support and personal assistance. The former includes household help, keeping company, help with paperwork, groceries and shopping. The latter refers to health service (e.g. medical injections), nursing and caregiving (e.g. dressing, washing, spoon-feeding). From the list, we exclude economic support (0.6%), since monetary transfers from adult children to old parents are relatively rare in Western societies (Brandt and Deindl 2013). We include a further category concerning co-resident adult children, since living together is an important form of support that family members can provide to each other.

The main independent variable is the frequency of church attendance. We distinguish between three categories: regular churchgoers who attend the mass at least once per week; occasional churchgoers who go to church less than once per week; and non-catholic people who never attend church rituals. The choice of these three categories derives from the necessity to identify those who follow the Catholic precept of attending the Sunday ritual (regular churchgoers) and religious individuals who disregard it. Moreover, weekly attendance can be considered as a reliable measure of individual’s adherence to religious precepts (Ballarino and Vezzoni 2012). Unfortunately, the questionnaire does not provide any information about religious denomination. However, after excluding foreigners and those living abroad the distribution of our independent variable is consistent with the findings of previous research showing that about 90% of the Italian population belongs to the Catholic religion (Vezzoni and Biolcati-Rinaldi 2015). Church attendance thus refers to the attendance to Catholic rituals.

As for traditionalism, the questionnaire contains different questions asking respondents whether they totally agree (1) or disagree (5) with regard to several claims. We run a Principal Component Factor analysis (PCF) in order to extract a latent dimension and we construct a measure of traditionalism using the four variables associated with the first component (Cronbach’s alpha = 0.75). Our index of traditionalism includes the following items: ‘Marriage is an outdated institution’, ‘People can cohabit without marriage’, ‘Women can have a child even without a stable relation’ and ‘Unhappy couples can interrupt their marriage even when they have children’. We sum all the items together in the final index ranging from 4 (less traditional attitudes) to 20 (more traditional attitudes).

Furthermore, we consider an indicator of family network as a possible mediator that may be associated with a religious attendance and the likelihood of providing support in later life. Respondents are asked whether they are in a close relationship with other relatives they can count on. This is a dummy variable indicating the quality of the relationships with the extended family network.

We also include a range of socio-demographic variables that are known to be correlated with church attendance and intergenerational support from adult children to older parents: age, region of residence (living in the south), number of siblings, educational level, employment status, marital status (living with a partner), and the presence of children aged 7 or less in the household. We also include a set of parental characteristics: sex, age, education, marital status, and health conditions. Parents’ poor health refers to having (severe or not) limitations in daily activities. Parents’ age, marital status and health conditions refer to needs of support that are often included as predictors of intergenerational time transfers (Evandrou et al. 2016).

After a brief description of the sample characteristics, we present results from multinomial logit regression models on the likelihood of providing care assistance and practical support or living in the same household. Given the presence of both father–child and mother–child relationships for some respondents, we use clustered standard errors to take into account the correlation between dyads constituted by the same individual. In addition, we employ regression models for adult sons and daughters separately. Women generally occupy the role of ‘kin-keeper’ in the family and are more likely to provide intergenerational support and care to older parents (Rossi and Rossi 1990).

We employ separated models assuming that the link between church attendance and intergenerational support may be driven by gender-specific mechanisms. Additional analyses, available upon request, are also undertaken to test gender differences and the results showed no significant interactions between church attendance and adult children’s sex. Our empirical analysis is developed through two steps: in the first model, we test whether a significant association between church attendance and intergenerational support exists (Hypothesis 1), by distinguishing transfers of practical support from those of personal assistance. We hence introduce the two indicators of traditional family attitudes and extended family relationships to examine their roles as mediators (Hypothesis 2 and 3). Since the comparison of coefficients across nested models can lead to misleading interpretations, we adopt the method developed by Karlson et al. (2012) for the decomposition of direct and indirect effects in the context of nonlinear probability models. The KHB method allows understanding whether changes in coefficients across multinomial models are due to the introduction of mediating variables by calculating a correction factor that allows comparison between the full model and the reduced model By adopting this strategy, it is then possible to evaluate the role of traditionalism and the extended family network in explaining the association between church attendance and intergenerational support.

Table 1 shows that 5% of Italian adult daughters provide personal assistance to their older parents, while only 3% of sons provide such assistance. Similarly, 7.6% of adult daughters provide practical support to older parents, while 6.5% of sons report giving personal support to them. Intergenerational co-residence is, on the contrary, more widespread among sons (about 10%) than among daughters (6.8%) (Chi^2 = 144; p-value < 0.001). Italian adult daughters are more prone to attend church rituals: about 39% of women report attending church rituals at least weekly, while this proportion is 23% for men (Chi^2 = 692; p-value < 0.001). Those who never attend church rituals are 7.6% among adult daughters and 14.2% among adult sons. 44.7% of sons and 48% of daughters highly value extended family relationships, while the average score of traditionalism is equal to 12 in a scale ranging from 4 to 20.

There are other relevant gender differences: 22% of adult sons do not have a job, whereas more than 50% of adult daughters do not have an active position in the labor market (Chi^2 > 1000; p-value < 0.001). This is hardly surprising, since in Italy the female participation to the labor-force is one of the lowest among OECD countries (Esping-Andersen 2012).

Tables 2 and 3 present results from multinomial logistic regression models predicting the likelihood of living with or providing practical support and personal assistance to older parents aged 65 or over. Model 1a for adult sons indicates that non-catholic adult sons have less propensity to provide practical support compared to occasional churchgoers. Regular churchgoers are more likely than occasional churchgoers to provide support to their parents, although the coefficient is only marginally significant. Similarly, among adult daughters (Table 3), non-catholic are less likely to provide support than regular churchgoers, while occasional churchgoers fall somewhere in between. The frequency of attendance to Catholic rituals is positively associated with a higher likelihood of providing intergenerational support. In line with our first hypothesis, church attendance tends to foster intergenerational support from adult children to parents, even when controlling for children’s and parents’ characteristics that per se have an important influence on supportive behavior. For instance, parents’ characteristics such as age, marital and health status are significantly associated with adult children’s propensity to provide practical support.

However, church attendance is not associated with sons’ and daughters’ propensity to provide personal assistance (Tables 2 and 3). Consistent with previous research, the likelihood of giving personal assistance to parents is associated mainly with indicators of parents’ need, including parents’ physical health. With regard to intergenerational co-residence, we found that non-catholic adult sons are less prone to live with their parents than occasional and regular churchgoers. Among daughters attending church rituals with a regular basis is positively associated with the likelihood of living with older parents. Although children’s needs are usually more important than parents’ ones in predicting intergenerational co-residence (Albertini et al. 2017; Smits et al. 2010; Tosi and Grundy 2018), we find that parents’ poor health and widowhood are associated with living under the same roof.

In Tables 4 and 5 we introduce traditional family attitudes and orientations toward the extended family network in the analysis. The association between church attendance and supportive behavior declines from 0.20 to 0.19 for regular churchgoers and from −0.32 to −0.31 for non-Catholic sons (Table 4). Among adult daughters, the coefficients related to church attendance change from 0.25 to 0.22 and from −0.60 to −0.57 for regular churchgoers and non-Catholics respectively (Table 5). To understand whether the association between church attendance and intergenerational support is due to these two factors, we adopt the KHB method. Although maintaining close relationships with other relatives is associated with practical support, this factor plays only a marginal role in mediating the association between church attendance and the likelihood of providing practical support. Its mediating capacity is not significant and equal to 0.5% and 2.5% for regular churchgoing and non-Catholic sons respectively. Among daughters, the reduction of the coefficient concerning regular churchgoers is equal to 12.8%, and it does not reach the significant level; while it is 5.0% and marginally significant (at 90% level) when considering non-Catholics.

Regarding intergenerational co-residence, regular churchgoers are more likely to live with parents than occasional churchgoers and the non-catholic, partly because of their propensity of having traditional family orientations. Among non-Catholic adult sons, about 15% of this association is due to traditional attitudes. The reduction of the coefficient related to regular churchgoers is not estimated because it is close to zero even when traditional family attitudes are excluded from the model. Among daughters, having traditional family attitudes has a significant role in reducing the relationship between church involvement and children’s support to parents. This association is reduced by 29% when considering regular attendance to church rituals and 18.2% when analysing non-Catholic daughters. The KHB method indicates that this decrease in the coefficients is significant at 95% level.

Previous research has argued that religious values and pro-family attitudes constitute latent forms of solidarity that trigger manifest supportive behaviors between generations (Silverstein and Conroy 2008; Silverstein et al. 2012). Also, religious practices tend to have benefits for the quality of parent–child relationships and the amount of time that parents devote to their adolescent children (Wilcox 1998, 2002). The findings presented here extend this literature to intergenerational support in late parent–child relations. Our results indicate that church attendance is associated with the provision of practical support to older parents in Italy. This is in line with previous findings focusing on the effect of religiosity on filial obligations (Daatland and Herlofson 2003; Dykstra and Fokkema 2012; Gans et al. 2009).

We also provide new evidence that might shed further light on the phenomenon in a threefold manner. First, the distinction we operate between ‘strong’ and ‘light’ forms of assistance is an important one, because it highlights how church attendance is consistently associated with the latter whilst not the former. Burdening forms of care which usually require a daily commitment (i.e. bathing, cooking and dressing) are not related to children’s religious involvement. It seems that church attendance and other individual characteristics partly lose their influences on intergenerational support when analysing the provision of care. On the contrary, less intensive forms of support (i.e. paperwork, keeping company, shopping) seem to be affected by personal willingness and individual orientations, including religious practices. We suggest that attending the mass socializes adult children, and in particular daughters, to some widely acknowledged Catholic principles, among which the importance of maintaining positive family relations and especially ‘to honour thy father and thy mother’ (i.e. the fourth commandment). Our interpretation is that regular churchgoers are more likely to be exposed to religious messages that promote altruism and filial devotion, which in turn encourage supportive behaviors. Other concomitant explanations are based on values, network and interactions that regular churchgoers may develop in the context of their religious communities.

Second, we attempt to account for individual attitudes toward the family, in order to understand whether they play a role in mediating the association between church attendance and the likelihood of providing support to old parents. Possibly, religious people may be more likely to assist their old parents partly because of their tendency to hold traditional views about the family. However, we find no support to this hypothesis. Our results show that traditional family attitudes are not associated with the provision of practical support and personal assistance from adult children to parents, although traditionally oriented adult children are more likely to live with their parents compared to those with less traditional attitudes. The results show that the association between religion church attendance and intergenerational co-residence is partly explained by traditional attitudes toward the family.

Third, we consider the idea that a family oriented socialization affects both church attendance and intergenerational exchanges. To partly account for this process, we include an indicator of family network in the analysis. The coefficients related to church attendance change only marginally when we consider adult children’s propensity to value the extended family network, although valuing extended family relationships is associated with supportive behaviors.

There are some limitations in this study. Due to data limitation, we have focused on church attendance instead of considering the complex concept of religiosity, which may further describe the phenomenon under examination. The results of this study are, therefore, limited to adult children who comply or not with Catholic precepts of church attendance. In addition, our measure of traditional attitudes solely includes indicators of marriage and divorce, without considering the broader set of family attitudes and norms. Unfortunately, due to data availability, we could not construct a comprehensive measure of traditionalism, family norms and attitudes that could take into account heterogeneous aspects of family life. In a similar vein, our mediating factors are proxies of parent–child relationship quality and family norms that can be developed during childhood and adolescence. On the basis of an extended literature, we assume that pro-family norms are instilled early in life, although the cross-sectional structure of the data does not allow us to establish any temporal order between variables.

Another limitation should be taken under serious consideration. Differently from Myers’ research (2004), we cannot observe the religiosity level of the parents, thus impeding the investigation of religious congruence within the dyad. However, the literature addressing the role of religious congruence between parents and their adult children reveals mixed results. Liefbroer and Mulder (2006), for example, suggest that the effect of socialization processes along with parent–child religious congruence tends to lose its influence over the life course. It could be possible, therefore, that parents’ religiosity plays a marginal role also in Italy where almost all older people define themselves as Catholic believers.

Nonetheless, our findings support the idea that adult daughters who attend church rituals on a weekly basis are more likely to provide less intensive forms of support to parents than the not-religious, with occasional churchgoers falling somewhere in between. This suggests that the provision of practical support is not only affected by the adherence to Catholic precepts concerning the attendance of rituals and parental devotion, but also by a gradual exposure toward religious messages. In Italy, Catholic practices tend to promote supportive behaviors in late parent-daughter relationships, even accounting for traditional attitudes and family oriented socialization.

No potential conflict of interest was reported by the authors.

Marco Tosi is a Research Officer at the London School of Economics and Political Science. He currently works on the FAMHEALTH project led by Prof. Emily Grundy. His research focuses on family dynamics, intergenerational relationships and mental health among older people in Europe. Before joining the LSE, He received a Ph.D. degree in Sociology and Social Research from the University of Trento. His recent papers have been published in Social Science and Medicine, Ageing and Mental Health, and Demographic Research.

Filippo Oncini is research fellow in Sociology and Social Research at the University of Trento. His research interests lie in the fields of consumption, health, and cultural sociology, with attention to social inequalities and food consumption. He is currently working on project PLATEFORMS, which aims to produce in-depth knowledge on how food practices in the home are affected by innovations in provisioning platforms.