Introduction

The presence and role of grandparents in European societies have increased remarkably in recent decades. Today the vast majority of elderly Europeans’ families span three or more generations, and grandparents have a growing role in the upbringing of their grandchildren and in affecting their well-being (Arber and Timonen 2012; Attias-Donfut and Segalen 1998; Grundy et al. 1999; Kohli et al. 2005; Szinovac 1998). Grandparents invest a considerable amount of time in looking after and providing care to their young grandchildren. Their contribution significantly eases parents’ work–private life reconciliation, support parents’ participation into paid labour market, and, in some contexts, it also affects the reproductive behaviour of their adult children (Aassve et al. 2012a; Albertini 2016; Bordone et al. 2017; Hank and Buber 2009; Thomese and Liefbroer 2013).

Grandparents are also beneficiaries of their own role. Although becoming primary caregivers to grandchildren—as in the case of skipped generation families—may have negative consequences on their well-being (Chen et al. 2015; Goodman and Silverstein 2006; Minkler et al. 1997; Musil and Ahmad 2002; Szinovacz et al. 1999), in general there is a positive relation between providing grandchild care and an individual’s psychological and physical health outcomes in later life (Di Gessa et al. 2016; Drew and Silverstein 2007; Grundy et al. 2012; Zhou et al. 2017). Grandparenting activities in regard to young grandchildren are correlated with better health status and self-perceived well-being, and a lower likelihood of depression. The positive relation between the grandparenting role and an individual’s well-being also extends to relations between grandparents and adolescent or adult grandchildren (Danielsbacka and Tanskanen 2016; Mahne and Huxhold 2015). These beneficial consequences are typically lacking for grandparents who fail to perform grandparenting activities and may even turn into negative effects when grandparents lose contact with grandchildren (Drew and Silverstein 2007; Jappens and van Bavel 2012; Neuberger and Haberkern 2014).

There are numerous factors affecting the intensity, type and quality of grandparent–grandchild relations and grandparenting. The divorce of the middle generation is one of the most influential ones. Previous studies have documented that, across different European societies, parental divorce generally brings about a weakening of grandparent–grandchild relations, particularly for paternal grandparents (Drew and Smith 1999; Jappens and van Bavel 2016; Mahne and Huxhold 2012; Westphal et al. 2015). It has also been suggested that the negative effect of parental divorce on grandparent–grandchild relations can be mainly attributed to the deterioration of relations between the non-resident parent and his/her children and, thus, to post-divorce residential and custodial arrangements (Jappens and van Bavel 2016; Westphal et al. 2015). Residential arrangements after family break-ups facilitate the preservation of intergenerational relationships in one lineage (bridging) and enable the custodial parent to control access to the other lineage (gatekeeping). However, the available empirical evidence on this relation is scant, and the role of face-to-face contacts with the non-resident parent is rarely addressed. Furthermore, previous research covers only a limited number of the different cultural and institutional contexts and family systems within Europe.

This paper considers the case of Italy—a society where the care provided by grandparents plays a pivotal role in the upbringing of small children and where parental separation and divorce is still a relatively rare phenomenon (Istat 2016)—and explores the extent to which the likelihood of providing grandparental care to a young child after the parents’ divorce is associated with the frequency of meetings between non-resident parents and their children.

Background

Grandchild–grandparents relations after parental divorce

Previous studies have documented that the presence and involvement of grandparents in the lives of their grandchildren have beneficial effects for both generations’ well-being. In particular, grandparents’ support helps young grandchildren to cope with stressful events and life course transitions, such as parental divorce. Thus, for example, in a study of English families, Gretchen et al. (2002) have found that closeness to maternal grandparents is associated with children’s better adjustment during family transitions, especially for children living with a single mother. Other studies have shown that in the first years after parental divorce contacts and closeness with grandparents reduce children’s psychological problems (Bridges et al. 2007; Doyle et al. 2010; Henderson et al. 2009; Lussier et al. 2002). At the same time, it has been reported that a loss of contact with grandchildren following parental divorce may be detrimental to grandparents’ mental health (Drew and Silverstein 2007).

Because young grandchildren have limited control over their contacts with grandparents, when divorce occurs their relationships are likely to be largely affected by the relation that they have with their parents. Studies of divorce have provided consistent evidence that marital dissolution has negative effects on the quality and intensity of the relationship between children and the non-resident parent—usually the father; additionally, this negative effect worsen when parents re-partner and persists even many decades after the divorce has taken place (Albertini and Garriga 2011; Daatland 2007; de Graaf and Fokkema 2007; De Jong Gieveld and Merz 2013; Kalmijn 2008, 2016; Noël-Miller 2013; Tosi and Gähler 2016). The negative effect of divorce on intergenerational relations extends beyond parent–child relations. A number of studies have documented that grandparents–grandchild relations are less intense when union dissolution occurs in the middle generation (Creasey 1993; Drew and Smith 1999; Hagestad 2006; Jappens and van Bavel 2016; Kemp 2007; Silverstein et al. 2003; Westphal et al. 2015). In particular, a weakening of grandparent–grandchild relations is found almost universally when considering paternal grandparents, while findings on maternal grandparents are mixed, especially when considering adult grandchildren (Cooney and Smith 1996; Ehrenberg and Smith 2003; Kruk and Hall 1995; Pillonel et al. 2013).

In sum, previous studies highlight that, although the loss of contact with grandparents after parental divorce can have negative consequences on the well-being of both grandchildren and grandparents, children of divorced parents tend to have less intense relations with and receive less support from their grandparents, and this almost uniformly applies to paternal grandparents. As we will discuss in the next section, this effect has been generally attributed to (a) post-divorce physical custody and residential arrangements, (b) the gatekeeping role of the resident parent vis-à-vis the ex-partner’s family and (c) the matrifocal character of parent–child relations in Western families.

Bridging and gatekeeping: the role of the middle generation

After marital dissolution, non-resident parents (generally the fathers) tend to have less intense and lower quality relations with their children. Grandparent–grandchild relations mirror this pattern: grandparents, especially paternal ones, are less likely to have contact with, be close to, and provide care to their grandchildren after parental divorce. Some scholars have suggested that the more pronounced negative effect on the paternal lineage could be explained by the matrifocal bias in kin relations in Western societies. Thus, parental divorce would simply contribute to perpetuating or even reinforcing such bias, also extending its effects to grandparent–grandchild relations (Chan and Elder 2000; Doyle et al. 2010; Monserud 2008). Other important social mechanisms that have been mentioned in the literature are the post-divorce residential arrangements and the gatekeeping role of the ex-partner. Various studies have found that the intensity of face-to-face contact with grandparents is almost entirely driven by the specific residential arrangement adopted after the union’s dissolution. Thus, while father-solo residence advantages the paternal lineage, mother-solo physical custody enhances relations with the maternal family. In other words, the role of parents in bridging the relation (and the provision of support) between grandparents and grandchildren is only played when a divorced parent has physical custody of the children, whereas the agency of the grandparents on the side of the non-custodial parent has a limited effect (Cooney and Smith 1996; Hilton and Macari 1998; Jappens and Van Bavel 2016; Monserud 2008; Timonen and Doyle 2012; Westphal et al. 2015). There is also evidence suggesting that ex-partners sometimes engage in gatekeeping behaviour, and thus control and reduce access to young children by former in-laws (Barranti 1985). In particular, mothers may withdraw their children from non-resident fathers when the relations between the ex-spouses are highly conflictual; in such cases, sporadic father–child contact moderates the negative consequences of inter-parental conflict on children’s well-being (Amato and Rezac 1994; Kalmijn 2016; Westphal et al. 2015). Conflicts between former spouses and obstruction of visits with the non-resident parent may, thus, reduce the frequency of meetings between grandparents and grandchildren (Arránz Becker and Steinbach 2012; Mueller and Elder 2003). All of these mechanisms indicate that, for grandparents, providing care to small grandchildren may be more difficult, less frequent, or even impossible when divorce occurs in the middle generation. Previous studies, however, have paid relatively little attention to analysing the extent to which grandparents’ involvement in childcare is correlated (or not) with the intensity of meetings between the non-resident parent and his/her young children, over and above the residential and custodial arrangements adopted after union dissolution.

In light of previous research findings, we can expect that a high frequency of meetings between the non-resident parent and his/her children is positively associated with grandparents’ involvement in looking after their young grandchildren. On the one hand, frequent meetings with the non-resident parent and care received from grandparent can be a response to increasing time-constraints faced by the resident parent after divorce. In other words, the positive association may be driven by the increased childcare support that parents need after divorce. On the other hand, as shown in the previous studies, non-conflictual relations between ex-spouses are generally associated with more frequent meetings between children and non-resident parents. In this context, both parents can be expected to act as bridges between grandparents and grandchildren and refrain from gatekeeping behaviour. This may improve intergenerational relations and facilitate the provision of support from grandparents. Finally, grandparents can themselves actively seek to maintain their relationship with grandchildren and thus encourage their son/daughter to stay in contact with their children after union dissolution (Timonen and Doyle 2012).

The Italian case

The study of grandparent–grandchild relations after parental divorce is particularly important in a social context like that of Italy, where performing the grandparent role is essential not only to improve both parents’ and grandchildren’s well-being, but also for the definition of the individual self in later life (Di Vita 2005; Dozza and Frabboni 2012; Saraceno 2011; Zanatta 2013). Although divorce was introduced comparatively late in Italy (in 1970) and its rate is below the European average, the quota of divorces involving young children has increased in recent years, up to 40.5% in 2015 (Istat 2016). The overall divorce rate has increased continuously in the past two decades, with a spike in 2015 following the changes in the legislation that eased and shortened the procedure for obtaining a divorce: there were 319.5 separations and 180.1 divorces per 1000 marriages in 2014 and 339.8 separations and 297.3 divorces per 1000 marriages in 2015 (Istat 2016).Footnote 1

Similarly to what has been documented for other European countries, in Italy union dissolution has important consequences in terms of children’s well-being and non-resident parent–child relations (Albertini and Dronkers 2009; Mazzucco and Meggiolaro 2014; Meggiolaro and Ongaro 2015; Todesco 2009). In particular, it has been shown that separated/divorced parents are at higher risk of losing contact with their children and that this effect not only lasts many decades after marital dissolution, but also extends to the relationship with the grandchildren. Moreover, it has been found that the negative effect is stronger for fathers than for mothers and that remarriage exacerbates the negative effect of marital dissolution (Albertini and Saraceno 2008).

Method

Data and sample

This study uses data from the last two waves of the Family and Social Subject Survey (FSS, Italian National Statistical Office: ISTAT) conducted in 2003 and 2009. The FSS is a cross-sectional survey representative of the entire Italian population (response rate is 85%) and contains information on 49,541 individuals (from 19,000 households) in 2003, and 43,850 individuals (from 18,000 households) in 2009. According to ISTAT sampling procedure, once a household is selected to participate to the survey, all its members are interviewed and one parent answers the questionnaire on behalf of children below age 14. The data, therefore, include individual records for people aged from 0 to 104 years. In our analyses, we combine the 2003 and 2009 data sets into a unique data set.

A specific section of the questionnaire, answered by the parents on behalf of their children, collects information on who regularly provides care to each child aged 0–13 living in the household—grandparents are included as potential providers of support. However, it is not possible to use the FSS to identify the specific grandparent(s) who provide childcare and thus distinguish between paternal and maternal grandparents, and between grandfathers and grandmothers.

Since our purpose is to examine grandparents’ involvement in childcare after marital dissolution in the middle generation, the sample selected for this study comprises only children below age 14 of divorced or separated parents (see note 1), and who had at least one grandparent alive at the time of the interview. The combined 2003 and 2009 individual-level data include only 782 children aged between 0 and 13 years who, at the moment of the interview, had at least one grandparent alive and whose parents were divorced or separated. A further selection criterion consists in excluding children whose parents were re-partnered (n = 99), because in this case it is not clear from the questionnaire whether the support received from grandparents was related to biological grandparents or step-grandparents. The final sample, therefore, consists of 683 young children living with their divorced or separated parents, among whom only a minority regularly lived with their fathers (n = 72). The survey does not collect information about legal and physical custody arrangements decided by the judge in the separation/divorce process. It is also worth noting that since more than one child from the same household can be included in the sample, all the following analyses are adjusted for the clustering of the units of analysis within the same household.

Measures

Grandparents’ involvement in providing care to children aged 13 or less was measured by the question: “When the child is not with parents or at school, who usually cares for him/her?”. The question comprises six (not mutually exclusive) answer categories: co-residing grandparents, non-coresiding grandparents, siblings, other relatives, friends or neighbours and babysitters. The dependent variable in the analysis, therefore, is a dummy variable that takes the value 1 when the child was looked after by his/her grandparents.

The main independent variable refers to contact frequency between the (grand)child and the non-resident parent. The frequency of contact is measured with a question about the amount of face-to-face meetings between the child and his/her parent. The six answer categories—ranging from daily to never—are collapsed into two: “more than weekly” and “weekly or less”. We run a number of sensitivity analyses changing the threshold of face-to-face contact frequency—for example using a three-category variable (daily, weekly and less than weekly)—and we found significant differences between daily or more than weekly meetings and the other categories. Next, there are two other relevant reasons for why we decided to use the threshold of weekly meetings: first, daily or almost daily meetings are clearly indicative of the willingness (or need) of the non-resident parent to maintain a relation with the young child that almost resembles co-residence. These parents are likely to have less conflictual relationships with the ex-partner and facilitate their own parents—children’s grandparents—to maintain relations with their offspring (Amato and Rezac 1994; Kalmijn 2016). Second, the same threshold has been used in the previous studies of intergenerational contacts in the Italian context, using it makes our results comparable with those studies (Bordone 2009; Tosi 2017).

As argued above, the resident parent’s need of support in looking after young children may be behind both the intensity of meetings with the non-resident parent and the care provided by the grandparents. Consequently, in the analyses, we control for variables known to be associated with those needs. First, we include the age of the child as a proxy for his/her care needs. Grandparents’ support may decline as children grow older, become more autonomous and begin to distance themselves from their grandparents (Geurts et al. 2009). We assume that, ceteris paribus, care needs tend to decrease with child’s age. We then control for a possible substitution and complementary role of other support providers by introducing three dummies taking value 1 if childcare was received from relatives, friends or neighbours, and babysitter, respectively. As noted above, these answer categories are not mutually exclusive and explicitly refer to alternative sources of support when the child is not with parents or at school. Another source of support that may affect the need to receive care from grandparents is primary school or formal childcare. Full-time education may reduce grandparents’ likelihood of becoming primary caregivers, but it may also increase their involvement in various forms of support, such as help with transportation and school-related activities. We consider a young child to be attending school if he or she was enrolled on a public or private educational programme. Information about the amount of time that young children spent at school is not available in the data set. Finally, we consider the employment status of the co-resident parent distinguishing not employed, in part-time employment, and in full-time employment, while no information is available on the non-coresident parent.

Other relevant control variables introduced in the multivariate analyses are the sex of the co-residing parent (mother or father), their educational level and age. Because highly educated parents typically have greater occupational opportunities, grandparenting may be particularly important in enabling these parents to work (Aassve et al. 2012b; Igel and Szydlik 2011). Working parents have less time to devote to their children and may require more support from grandparents. It is also important to assess the role of the number of siblings and that of living grandparents. The likelihood of receiving support from grandparents may be lower in larger families where grandchildren “compete” for grandparents’ time with their siblings (Oppelaar and Dykstra 2004). Conversely, the number of living grandparents can play a positive role: the more grandparents that children have, the higher is the probability of receiving assistance from at least one of them. We also include information on the distance between the place where the child usually lived and where the grandparents lived in our models. Residential proximity between generations tends to enable grandparenting; it shapes opportunities for exchange of support between grandparents and grandchildren, because living nearby makes it possible to give help in childcare more often (Uhlenberg and Hammill 1998). Since our dependent variable explicitly refers to usual providers, it is likely that young children receive assistance from grandparents who live in the same municipality. A separate category for co-resident grandparents captures their immediate availability as care providers within the household. This variable is related to the parents of the residential parent, while the data do not provide any indication about where the parents of the non-residential parent lived. Finally, we include information about the age of the child at the moment of the parents’ separation/divorce. In fact, grandparents may be more involved in the lives of those grandchildren who—being very young at the moment of union dissolution—need more support; this generational bonding at earlier child’s ages is likely to affect intergenerational relations later.

Analytical strategy

Our analytical strategy proceed as follows. First, we present descriptive results including summary statistics for our sample (Table 1). Second, we perform multivariate analyses (Table 2), by using logistic regression models, on the likelihood of receiving support in childcare from grandparents. To aid understanding of the magnitude of the coefficients, Fig. 1 presents the average marginal effects for each variable included in the fourth model of Table 2. This procedure is in line with what has been suggested by Mood (2010).

Table 1 Sample characteristics (N = 683)
Full size table
Table 2 Logistic regression models on the likelihood of receiving support in childcare from grandparents
Full size table
Fig. 1
figure 1

Average marginal effects on the likelihood of receiving care from grandparents

Full size image

In the first regression model, we only introduce the sex of the co-residing parent as a control variable; we then proceed by analysing whether the correlation between non-resident parent–child visits and grandparental care is mediated by the support needs of the resident parent. Thus, variables accounting for co-resident parent’s employment status, child’s age and the presence of other care providers are introduced in model 2. In the third step (model 3), we control for factors connected with care availability and constraints of the resident parent (i.e. age, educational level and the number of children in the household) and grandparents (number of grandparents alive and residential proximity between the child and his/her grandparents). In model 4, the age of child at time of separation/divorce is also included in the analysis. We also carried out a sensitivity analysis to check the robustness of our results by examining whether the correlation between grandparental role and the frequency of meetings with the non-resident parent varies according to the sex of the latter. The results (available upon request) showed a non-statistically significant interaction and are, thus, similar to those presented in Table 2.

Results

Descriptive results

Table 1 shows that 48% of young children from dissolved families received support from grandparents; about 21% received such assistance from other relatives, 7% from friends and 5% from babysitters. Grandparents, therefore, remain one of the main sources of care support to young children also after parental separation/divorce. The large majority of young children (88%) in our sample attended elementary school or preschool education, indicating that in most cases the grandparents’ role was complementary to that of formal education. More than half (55%) of young children from separated/divorced families met their non-coresident parents on a weekly basis or less often; less than half lived in a different municipality than the parents of the co-resident parent.

Multivariate results

Table 2 presents logistic regression models on the likelihood of receiving childcare assistance from grandparents among children from separated/divorced families. The first model shows that the association between the frequency of meetings with the non-resident parent and grandparents’ involvement in childcare is statistically significant. Young children who had less than weekly meetings with their non-resident parents were also less likely to receive childcare from their grandparents. This result provides evidence in support of the expectation that less meetings with non-resident parents hamper the relationship between grandparents and grandchildren beyond specific living and residential arrangements.Footnote 2

We include indicators of support needs in model 2. Children whose parents were not in full-time employment were less likely to receive care from grandparents. Having a babysitter was negatively associated with grandparents’ involvement in childcare. Grandparents’ involvement decreased at growing ages of the grandchild. After accounting for these associations, the coefficient related to the frequency of meetings with the non-resident parent changes only marginally from − 0.55 in model 1 to − 0.60 in model 2. Therefore, there is an association between face-to-face contacts and grandparents’ childcare provision net of parents’ support needs and the presence of other sources of support. Similarly, in model 3, the frequency of meetings with the non-resident parent is associated with the likelihood of receiving grandparental care, net of the effect of other relevant factors such as: the number of grandparents and siblings, the resident parent’s educational level and living distance from grandparents. In the fourth model, we analyse whether the association between the frequency of meetings and grandparents’ involvement in childcare is confounded by the child’s age at parental divorce or separation. The older was the child when he or she experienced the marital dissolution of their parents, the lower was the probability of receiving support from grandparents. However, having more than weekly meetings with non-residential parents was significantly associated with the likelihood of receiving care assistance from grandparents even after controlling for the timing of parental separation/divorce.

To better understand the size of these associations, Fig. 1 presents average marginal effects estimated from the fourth model in Table 2. The estimates indicate that meeting frequency between young children and non-resident parents is an important predictor of grandparents’ care provision, and the magnitude of this association is similar to that of the employment status of the resident parent. The probability of receiving support from grandparents was on average 10 percentage points higher for young children who meet their non-residential parents more than weekly than for those who have fewer meeting with them. Among other covariates, geographical distance from grandparents appears to be the most important factor in affecting the probability of receiving assistance in childcare. Figure 1 shows that the probability of receiving assistance from grandparents is about 25 percentage points higher for young children living near grandparents (in the same municipality) than for those living further away (in another municipality). Furthermore, this probability is 60 percentage points higher for young children living with grandparents than for those whose grandparents lived in another municipality.

Discussion

The role of grandparents in the life course of their grandchildren is growing in European societies. Separation and divorce in the middle family generation, however, can disrupt the relations between grandchildren and their grandparents. Previous studies on the topic have suggested that the loss of contact with grandparents is almost completely explained by the residential arrangements adopted after divorce (Jappens and van Bavel 2016; Westphal et al. 2015). Thus, children living with their mother tend to keep contact with maternal grandparents and lose contact with paternal ones, whereas the opposite applies when fathers have the physical custody of their young children. The resident parent can operate as a bridge in connecting grandparents and grandchildren from their own lineage or can assume the role of gatekeeper by exerting control over visits between his/her children and former in-laws. Our research examines the extent to which the intensity of meetings between a non-resident parent and his/her children can be associated with grandparents’ support in childcare, over and above post-divorce residential arrangements. The results show that children who have more than weekly face-to-face contact with their non-residential parent are also more likely to be looked after, on a regular basis, by their grandparents. Overall, these findings suggest that (almost) daily face-to-face contacts with non-residential parents operate as an efficient bridging mechanism in increasing grandparents’ involvement in childcare. There are, however, other underlying social mechanisms that could explain the observed positive association. Grandparents may play an active role in maintaining their relationships with grandchildren and encouraging their separated/divorced children to do the same. It may be that high-frequency contact with both the non-resident father and grandparents is an indication that the child lives in a “contact-intense” family (vs. a “contact-weak” family), and this characteristic of the family is preserved even after parental divorce. It is also possible that separated/divorced parents receive support from both the ex-partner and grandparents in looking after their children when they need it. Although we cannot rule out all possible indicators of need, the results presented here suggest that the association between parent–child meetings and grandparents’ provision of support persists after controlling for some of the most relevant indicators of childcare support needs.

The study has some limitations. First, because we have used FSS data, we are not able to distinguish the support received from maternal or parental grandparents, and from grandmothers or grandfathers. Second, the analysis is based on cross-sectional data; consequently, we are not able to observe changes—before and after separation or divorce—in the interaction between children and their grandparents. Third, measures of inter-parental conflict, before or after separation/divorce, are missing. Resident parents who have less conflictual relationships with ex-partners may facilitate the interaction between non-resident parents and their young children, as well as the contact between children and the grandparents from the ex-partner lineage. The intensity of meetings with the non-resident parent could be a proxy for the overall quality of family relations after separation/divorce. This would indicate that less conflictual separations have a positive influence on grandparent–grandchild relations. Therefore, we are unable to assess the extent to which the correlation between parent–child meetings and the likelihood of receiving grandparental care is explained by the bridging role of the non-residential parent, or the absence of a gatekeeping behaviour, or an increasing involvement of the grandparents by the residential parent. Finally, when analysing grandparent–grandchild relations we are only able to observe the provision of care to young grandchildren, while omitting from the analysis other important relational dimensions such as emotional support, meeting frequency and intimacy levels.

Despite these limitations, we found evidence that face-to-face contact frequency between a non-resident parent and their children is positively associated with grandparents’ supportive behaviour. Future research in the fields of gerontology and family sociology should consider the fact that the lack of grandparents’ support is not a necessary consequence of parental separation/divorce. Non-resident parents can play an important role in preserving the relation between their own parents and their children, thus benefitting all three family generations.