Kinkeeping Across Families: The Central Role of Mothers and Stepmothers in the Facilitation of Adult Intergenerational Ties

A kinkeeper is the person within the household that is involved in the management of family relationships, a position traditionally fulfilled by women. Due to the increased complexity of family life, which resulted from the rise in divorce and remarriage, the kinkeeper role might nowadays be particularly important but also more ambiguous. First, we examined differences in parental involvement in kinkeeping (buying presents, organizing outings, relaying family news, and discussing problems) along the lines of gender, family structure, and biological relatedness. Second, we explored whether the kinkeeping of parents and their partners is effective in the facilitation of intergenerational closeness with adult children. We used the OKiN survey, which includes information on kinkeeping in N = 746 intact, N = 982 mother-stepfather, and N = 1,010 father-stepmother families. Findings indicated a central facilitative role for mothers and stepmothers. Substantial gaps were found between mothers and fathers, married and divorced parents, and biological and stepparents with respect to the (variety of) kinkeeping in which these parents were involved. Yet, the contrasts of biological relatedness and family structure were also found to be gendered, as these gaps were smaller for mothers than fathers. Finally, an association was found between adult closeness with biological parents and the kinkeeping of the spouse, regardless of the nature of the relationship between the spouse and adult child. This implies that a stepmother can be just as effective as a married biological mother in facilitating the ties between a father and his biological children.

. In practice, kinkeeping remains imprecisely defined among household members and is not always seen as equivalent to physical household tasks ("invisible" to those within the household and long unrecognized by scholars, see Daminger, 2019). Alongside gendered perceptions of "emotional work" and strict norms surrounding mothering ideals, this ambiguity is expected to make kin work highly unequal in that mothers are generally the (only) ones involved (Di Leonardo, 1987). This gendered division in kin work, along with the implicit assumption that fathers do not participate, has rarely been tested (for exceptions, see Braithwaite, 1996 andBrown &DeRycke, 2010). Although the kinkeeper role might even be more ambiguous if certain parents within the household do not have a biological relation with the child (Gaunt, 2006;Weaver & Coleman, 2005), even fewer scholars have studied the kinkeeper role in the context of divorce and remarriage (Schmeeckle, 2007). In this article, we examine whether mothers, fathers, stepmothers, and stepfathers are involved in kinkeeping to empirically consider the presumed gaps in kin work according to gender, family structure, and biological relatedness. This provides us insights into how societal expectations are translated differently into the behaviors of mothers and fathers and tells us whether gender norms continue to shape differences in parent kinkeeping in the context of divorce and repartnering.
The second chief aim of this contribution is to capture the effectiveness of parents' kin work in terms of facilitating close parent-child bonds. A test of the effectiveness of kin work is relevant for biological fathers in particular (Marsiglio, 1993), especially if we take the increased diversity of family structures into account (Thomson, 2014). Existing literature, which compared kin contact across different family structures (e.g., married households, divorced households), indicates that, in some cases, the very existence of kin contact depends on the presence of a woman in the household, with kin contact being much lower in the households of divorced men (Kalmijn, 2007). Remarriage has been suggested to facilitate divorced fathers in the reconstitution of their contact with kin (Schmeeckle, 2007). We built upon these results by directly testing the relationship between kinkeeping and adult closeness with biological parents. More specifically, we examined whether parents can facilitate each other in their ties with any shared adult children, reflecting upon this association within married and repartnered households. This focus tells us about the potential of stepparents to be equally effective kinkeepers and thus, relevant facilitators of the closeness between separated parents and their adult biological children.
Our focus is on a specific societal and historical context. Similar to many western societies, the Netherlands has gone through several demographic developments since the 1960s which collectively altered the functioning of households (Thomson, 2014). As more and more women achieved high levels of education and the labor market participation of women increased rapidly, an overall trend toward more egalitarian gender role attitudes was also witnessed (Boehnke, 2011). At the same time, part-time workforce participation became highly popular among Dutch women and the unequal division of physical household labor, and childcare in particular, remained very prominent (Portegijs et al., 2021). Such contextual features naturally shape the normative expectations that exist around gender, family, and parenthood (Connidis, 2020), and thus influence how the current generation of Dutch adults and their parents and stepparents think about the rights and responsibilities surrounding the kinkeeper role.

Conceptualization of Kinkeeping
Kinkeeping includes all efforts of maintaining connections among family members (Di Leonardo, 1987;Gerstel & Gallagher, 1993;Rosenthal, 1985). It has famously been conceptualized by Rosenthal (1985) as an important but less frequently examined aspect of the division of household labor. The "emotion work" involved in maintaining family ties can be understood as part of the household tasks divided among partners (Hagestad, 1986;Rosenthal, 1985;Seery & Crowley, 2000). From that perspective, kinkeeping behaviors not only include the active facilitation of closer relationships between two family members, but also comprises a variety of familial responsibilities and activities enacted on behalf of the familial household (e.g., buying birthday presents, spreading news, representing the family, organizing outings). A combination of these activities being performed is theorized to increase family cohesion and solidarity and therefore, simply the presence of someone assuming a kinkeeper role is expected to increase the quality of all individual ties (Di Leonardo, 1987).
So far, most literature on kinkeeping has been dedicated to defining the concept and determining which practices it comprises. In the gerontological literature, various activities were mentioned in relation to kin work, such as visiting, telephoning, buying presents, sharing news, or providing emotional aid on behalf of the household (Adams, 1968;Aldous, 1967;Leach & Braithwaite, 1996). In other studies, which were more directly interested in how kinkeeping helps specific relationships, interviews were done to gather information about the work in which mothers engage to build and maintain the relationships between their children and the children's father. The most commonly reported behaviors in such studies were: relaying positive feelings, suggesting joint parent-child activities, organizing family events, and helping to resolve problems (e.g., Schmeeckle, 2007;Seery & Crowley, 2000). In general, the central commonality in previous conceptualizations is the idea that kinkeepers, while engaging in a combination of kin activities, make themselves an important communication link between others within (and sometimes beyond) the familial household and as such, facilitate the closeness of involved family relationships.
Who is a Kinkeeper? Rosenthal (1985) proposed that people generally identify one person within the household as fulfilling the position of kinkeeper, rather than viewing both partners as taking on the position simultaneously. Are certain parents more likely to take on the kinkeeper role? In the context of rising family complexity, this question evolves around three themes: the differences between (a) mothers versus fathers, between (b) married versus divorced parents, and between (c) biological versus stepparents within families (Ganong & Coleman, 2016;van Houdt et al., 2020;Weaver & Coleman, 2005).

Mothers Versus Fathers
The area of family work captured by the concept of kinkeeping is dominated by women (Brown & DeRycke, 2010;Leach & Braithwaite, 1996;Rosenthal, 1985). When dividing household labor, partners often only regard the physical component, while specific agreements are less often made about the cognitive component. There is evidence showing that, under these conditions, women take on a larger share of "emotion work" than men (Daminger, 2019), a dynamic that scholars have theorized to also apply to kinkeeping (Rosenthal, 1985;Di Leonardo, 1987). The reason for women to be the default person to take on kin work lies in the gendered expectations that exist about family life (see Di Leonardo, 1987). In relation to parenthood, kinkeeping behaviors are more compliant with the normative concept of traditional motherhood (Braverman, 1989;Russo, 1976). This aligns with the literature on the intensive mothering ethos, i.e., the parenting paradigm that a "good" mother should be highly involved in all parts of their children's lives (Johnston & Swanson, 2006). In fact, there is research demonstrating that both women and men perceive the woman as the one (that should be) most involved in managing kin relationships (Aronson, 1992). In practice, kinkeeping has indeed been reported to be performed mostly by the mother in the familial household (Brown & DeRycke, 2010;Leach & Braithwaite, 1996;Rosenthal, 1985). In sum, we hypothesize that: mothers will be involved in more kinkeeping behaviors than fathers, regardless of the family structure or biological relatedness (H1).

Continuously Married Parents Versus Repartnered Parents
The level of kinkeeping performed by a biological parent may in part depend on the family structure in which the parent is embedded. That is, there might be more need for parents' efforts to connect family members in divorced or repartnered households (Ganong & Coleman, 2016). Through divorce, parents not only dissolve a partnership but often also lose a source of mutual aid in maintaining close relationships with any shared children. At the same time, the value of kin work, such as organizing joint activities or discussing family issues, might become more pronounced due to the disruptive effects of parental divorce (Amato, 2010) or parental remarriage (King, 2009) on the kinship system. As a result, biological parents might feel more obligated to focus on kinkeeping after separation, in order to countervail these disruptive effects and compensate for the fact that they no longer have another biological parent within the familial household to provide additional kin work. A separated parent would then invest more time and effort in the facilitation of family relationships. In sum, divorced parents are expected to feel more need to facilitate a connection between their familial household and adult child than married parents. Based on this idea, we hypothesize that repartnered parents will be involved in more kinkeeping behaviors than married parents (H2a).

Gendered Effect of Family Structure
Given the gendered effects of divorce on parent-child ties, which entail that fathers less often receive custody and more often lose contact with their children after divorce (Amato, 2010), we need to consider the potential intersection between the role of family structure and gender. Here, the guiding expectation is that, because biological mothers are considered the "traditional kinkeepers", the relationships of adult children with fathers vis-à-vis mothers become more unstable outside of marriage (Ganong & Coleman, 2016). That is, married fathers will benefit from the kinkeeping role of the mother in their relationships with adult children, as the mother, for instance, relays positive messages or organizes visits at which the father is also present (Hagestad, 1986). In the context of divorce, a repartnered mother is expected to shift her role as kinkeeper to her new familial household. For instance, in some cases, mothers have been found to shift their facilitative efforts from the biological father to the stepfather and invest specific effort so that the new household can operate as a "nuclear" family (Marsiglio & Hinojosa, 2007). Repartnered fathers have lost their partnership with the "traditional kinkeeper", and compared to a repartnered mother, also face a greater risk of feel encouraged to engage in kinkeeping. After all, stepfathers are linked to a mother, the "traditional kinkeeper", whereas stepmothers are linked to the father, who is traditionally less involved (Schmeeckle, 2007). Stepmothers might therefore be more inclined than stepfathers to step into the kinkeeper role and engage in behaviors, such as buying birthday presents, relaying family news, and encouraging contact between the child and other parents, regardless of their social bond with the focal child (Braithwaite et al., 2017). If so, the gap in kinkeeping between biological parents and stepparents would be smaller for mothers than fathers (H3b). On the other hand, one should also consider the culturally constructed primacy of the biological-motherrole within the family system (Kalmijn et al., 2019;Weaver & Coleman, 2005). Because biological motherhood is generally perceived as a preferable base for childcare (Braverman, 1989;Gaunt, 2006), the normative boundaries to be involved in the child's life, let alone influence the child's family relationships, should be much higher for stepmothers than stepfathers (Ganong & Coleman, 1995). In other words, biological relatedness is expected to play a larger role for mothers than fathers. This implies that the gap between biological parents and stepparents in kinkeeping will be larger for mothers compared to fathers (H3c).

How Effective is a Kinkeeper?
Although scholars have extensively studied the communicative manners through which kinkeepers maintain relationships (Leach & Braithwaite, 1996;Braithwaite et al., 2017), the literature on their effectiveness is sparse and the empirical evidence is convincing but mostly indirect (i.e., see Leach & Braithwaite, 1996;Kalmijn et al., 2019;Rosenthal, 1985). In general, existing studies suggest that both the presence of a kinkeeper and the variety of kin activities in which the kinkeeper is involved are important for the facilitation of family solidarity (Leach & Braithwaite, 1996). Building on these studies, we explore whether parents' involvement in kinkeeping is effective. That is, does it facilitate the ties between the other parent in the familial household and the adult children?
This question is relevant for biological fathers in particular, especially in light of increasing family complexity (Thomson, 2014). As theorized above, the kinkeeping role is mainly fulfilled by female family members and therefore, mothers play an important mediating role in the relationships between fathers and their children (Di Leonardo, 1987). If couples divorce or if the mother passes away, the work of kinkeeping is assumed to be left undone (Di Leonardo, 1987). Although fatherhood is increasingly defined in terms of affection, with an "involved father" nowadays diminished parent-child closeness and contact as a result of parental separation or the entrance of a stepparent (Kalmijn, 2007). As a result, there is more need for repartnered fathers to increase their own involvement in kinkeeping. So, if parental separation indeed is an activator in terms of kin work, as we hypothesized above (H2a), this should apply to fathers more so than mothers. In organizing family outings or resolving family issues, for instance, remarried fathers are expected to have increased their involvement more than remarried mothers. In sum, we expect that the gap between married and repartnered parents in kinkeeping will be larger for fathers than mothers (H2b).

Biological Versus Stepparents
As many divorced parents start a household with a new partner after separation, the question is whether stepparents can also take on a position as kinkeeper and assist the ties between their stepchildren and other members of the familial household. Due to the symbolic power of biological relatedness, which may be rooted in evolutionary preferences (Anderson et al., 1999) but also in normative ideas about obligations towards biological kin (Rossi & Rossi, 1990), the bonds between biological parents and adult children are generally seen as the standard marker and perceived as more self-evident than the bonds between "social parents" and adult children (Arránz Becker et al., 2013;Kalmijn et al., 2019;Loehlin et al., 2010). Meanwhile, the norms on stepparenthood in particular are not clearly defined and therefore, there is more ambiguity about which behaviors are expected from or seen as appropriate for a stepparent (Cherlin, 1978). In short, cultural norms exist about the centrality of biological parents within kinship networks (Ganong & Coleman, 2016), which guides the expectation that biological parents' ought to invest in their ties with children, as well as the kinship of these children with other members of the familial household (Brown & DeRycke, 2010). Stepparents are perceived to carry less responsibility in that regard, a dynamic that may be driven by the social bond between the stepparent and child, but can also be attributed to the fact that responsibilities are less well defined for stepparents in particular. Therefore, we hypothesize that stepparents would be less involved in kinkeeping behaviors than biological parents (H3a).

Gendered Effect of Biological Relatedness
Theoretical arguments on the intersection of gender and biological relatedness provide two opposing expectations. On the one hand, kinkeeping is generally performed by the woman in a household and therefore, there might be a difference in the extent to which stepfathers and stepmothers from a new partner (Kalmijn, 2007). In short, this proposes that there is a gap between biological parents and stepparents in the effectiveness of their kin work.
On the other hand, interviews with adult stepchildren indicate that stepparents can also be effective in kin work, but the evidence that exists so far is based on qualitative data and relates to the involvement of stepmothers only. In these studies, stepmothers were reported to not only invest time in facilitating the relationship between the biological father and her own children, but also in facilitating the potentially strained relationship between the father and his biological children with a previous partner (Schmeeckle, 2007). In fact, occasionally, stepmothers are perceived as the "carpenters" of the family that work to mend, repair, or maintain relationships between otherwise uninvolved fathers and their biological children (Vinick & Lanspery, 2000;Schmeeckle, 2007). A stepmother involved in kin work was suggested to compensate for the fact that the biological father has lost a kinkeeper through divorce or widowhood. If stepparents can work as the "carpenters" of strained ties, this would be an indication that stepparents can be equally effective as biological parents in mending the ties between their partners and any involved children. In other words, these arguments suggest there to be no gap between biological parents and stepparents in effectiveness of their kin work.

Sample Description
We used the second wave of the OKiN survey (Ouders en Kinderen in Nederland; Parents and Children in the Netherlands, Kalmijn et al., 2018). The OKiN survey was initially collected in 2017 among a stratified-random sample of 25-to 45-year-old adults in the Netherlands. The sample included a large register-based oversample of adults who did not reside with both biological parents at the age of fifteen, resulting in a comparatively large number of respondents from divorced and repartnered families. The fieldwork was done by Statistics Netherlands under the conditions of the Dutch Personal Data Protection Act and the data collection was approved by the project's Ethics Advisor. Fieldwork yielded a response rate of 62%. In the second wave, questions from the first wave were repeated and additionally, a new battery of questions on inter-parental dynamics was added to the questionnaire. This battery included our variable of interest: parental kinkeeping.
The second wave, which was collected in 2020, was comprised of adults from the original sample who indicated in 2017 that they were open to future participation (82%). These participants were approached for being economically but also emotionally involved (Dermott & Miller, 2015;Lamb, 2013;Miller, 2010), the relationship between a father and his adult children still depends greatly on the kinkeeping role of a mother (Kalmijn, 2007). Losing this kinkeeper through divorce further emphasizes the disadvantaged position of separated fathers in maintaining positive ties with their children (Ganong & Coleman, 2016). A remaining question is therefore whether stepmothers (or stepparents in general) can compensate for the lost connection between the biological parent and the adult child. Following these ideas, we explore the effectiveness of kinkeeping across intact families, stepfather families, and stepmother families and focus on one contrast in particular: the effectiveness of stepparents' versus biological parents' kinkeeping. We question if repartnered biological parents can benefit from their partners' kinkeeping to the same degree as married biological parents. Such comparisons reveal the potential for stepparents to act as kinkeepers and contribute to family cohesion.

Biological Parents Versus Stepparents as Most Effective
The existing (mostly qualitative) literature on adult children suggests that stepparents sometimes step into the kinkeeping role (e.g., Schmeeckle, 2007). Yet, due to the limited number of studies on this topic, most empirical evidence is on the involvement of stepmothers (Schmeeckle, 2007;Vinick & Lanspery, 2000) and few scholars have empirically examined whether stepparents, when involved in kin work, are indeed effective in maintaining solidarity within a familial household. Based on earlier research, different expectations exist on the question of biological parents' versus stepparents' effectiveness as kinkeepers. Therefore, we decided to examine this question from an explorative viewpoint, but our thinking was still guided by several theoretical arguments.
In general, the arguments for stepparents' effectiveness as kinkeepers are twofold. On the one hand, stepparents are expected to be less effective than biological parents in mending the ties between their partners and any involved children. The underlying reason is that, due to the social bond between the stepparent and adult stepchild and the ambiguity surrounding step-parenthood, stepparents are less likely to be involved in kinkeeping. Also, when stepparents are involved, the stepchildren connected to the familial household may be less inclined to accept the interference of a stepparent in their relationships with other parent figures (Ganong & Coleman, 2016). For instance, empirical evidence suggests repartnered divorced parents and single divorced parents to be similar with respect to the quality of their ties with children, regardless of the potential kin work

Parental Kinkeeping
We used a new quantitative measure of parents' involvement in kinkeeping. The measure is based on past qualitative studies and consists of four items that reflect four main types of kinkeeping activities: (a) buying (birthday) presents on behalf of the household, (b) organizing family outings, (c) relaying family news/messages, and (d) discussing family problems. The adult children were asked to report which parent in their parental household is most involved in each of the activities. The answer categories ranged from 1 (almost always my father/new partner of my mother) to 5 (almost always my mother/new partner of my father), also including the response option not applicable (6). This type of measurement is in line with earlier measures of partners' relative involvement in household labor and was therefore, by design, only presented to respondents from two-parent households.
For each parent figure, we recoded the four items into dichotomous variables to reflect whether this specific parent was 1 (involved) or 0 (not involved) in this kinkeeping activity. We did so in the following manner: 0 (scores of 6, "neither parent does it" and scores of 1, 2 "other parent does it more/most") and 1 (scores of 3 "both parents do it" and scores of 4, 5 "parent does it more/most"). We calculated Computer-Assisted-Web-Interviewing (CAWI). Of the reapproached adults, N = 3,070 participated (a response rate of 58%). Non-response analyses showed that, although there was some panel attrition according to a few demographic and socioeconomic variables (e.g., no partner, unemployed), non-response was not associated with the quality of family relationships (i.e., parent-child closeness, parentchild contact, parent-parent conflict). From the complete sample, we analyzed the reports of adults whose parents are still together and alive (N = 746) and the reports of adults whose parents separated and who have at least one stepparent (N = 982 with a living mother and stepfather; N = 1,010 with a living father and stepmother). Adults whose divorced parents repartnered with a person of the same sex were thus not included in the sample. Note that, since the OKiN survey does not differentiate between partnerships formed through marriage and partnerships formed through cohabitation, both are included in our study. Accordingly, a stepparent was conceptualized as a new partner of the opposite sex with whom the biological parent has been married or cohabiting for at least two years. We analyzed the data with the parent figures as the unit of analysis (total of N = 5,476 parents), including 13.6% continuously married fathers, 13.6% continuously married mothers, 18.4% repartnered fathers, 17.9% repartnered mothers, 17.9% stepfathers, and 18.4% stepmothers. in earlier research on (adult) intergenerational relationships (e.g., Kalmijn et al., 2019;King, 2009;Loehlin et al., 2010;van Houdt et al., 2020).

Type of Parent
We compared the involvement and effectiveness of three parent types: married biological parents, separated and repartnered biological parents, and stepparents. In the analyses, we used the repartnered biological parents as reference category and included dummy variables to represent the other two categories. As mentioned earlier, we cannot make comparisons with respect to single biological parents, given that, due to the design of the questionnaire, kinkeeping was not measured with respect to single-parent households.

Controls
Various controls were included in the analyses (see Analytical Strategy for more details). To start, we measured three controls that have the potential to define the parent-child relationship currently: duration of co-residence, closeness in youth, contact frequency currently, and no contact currently. The duration of co-residence was measured as the total number of years the respondent co-resided with the parent during their upbringing (up to the age of 18). In addition, we used a 5-point Likert-scale to measure how close the adult was with their biological parent during the period of their upbringing (up to the age of 18), with answer categories ranging from 1 (very close) to 5 (not at all close). This item was recoded so that higher values reflect closer ties. The item about the frequency of face-to-face contact with a parent was measured on a scale from 1 (no contact) to 6 (weekly contact). This item was also recoded into a dichotomous variable capturing if there is any contact (0) or no contact at all (1). At certain steps in our analyses, we need to control for these variables to account for the fact that, on average, there are differences between the studied parent types in the amount of time they have been in the a sum score for these dummy variables to create a count measure per parent figure. This count measure captures the range of kinkeeping activities in which a parent is involved (running from 0 to 4). A descriptive overview of our kinkeeping measure by household structure and parent gender can be found in Table 2.
We decided to use a count measure for two reasons. First, the count measure can be created for each parent figure separately. This is preferable with respect to our research objectives, as it allows us to directly estimate potential contrasts between the different parents in terms of kinkeeping involvement and kinkeeping effectiveness. If we would use a relative measure, contrasts between parents would be inherent to the measure and therefore, more difficult to interpret. The separate count measures also allow us to run estimations while controlling for the kinkeeping involvement of another parent (e.g., when estimating the effectiveness of stepparent kinkeeping in addition to biological parent kinkeeping). Second, as we discussed in the theoretical framework, the effectiveness of kinkeeping has been theorized to rely on the kinkeeper being a communication link between family members due to their engagement in a combination of kinkeeping activities (Rosenthal, 1985). This adds substantive meaning to our count measure, as family cohesion would in theory depend on a range of kinkeeping activities being performed, rather than the division of the separate activities.

Parent-Child Closeness
The degree of emotional closeness between an adult and biological parent was measured on a five-point Likert scale ranging from 1 (very close) to 5 (not close at all) and reported on by the adult child. This item measure was recoded so that higher values referred to closer relationships, and the recoded measure was used in the analyses as an interval variable in a manner similar to earlier studies on parent-child closeness. In general, this single-item measure of the degree of emotional closeness is expected to have high face validity and has therefore frequently been used  (746) 100 (746) 100 (982) 100 (982) 100 (1010) 100 (1010) Note. The kinkeeping-count-variable was created by dichotomizing the four items: 0 (scores of 6, "neither parent does it" and scores of 1, 2 "other parent does it") and 1 (scores of 3 "both parents do it" and scores of 4, 5 "focal parent does it). For each parent in the household, a sum score was calculated to create the count measure.
Source: Ouders en Kinderen in Nederland (OKiN), Wave 1 and 2. This usually suggests that a separate process is going on that generates these excess zeroes. In this case, that makes sense theoretically, as there likely exists a separate process that makes certain types of parents more likely to not at all be involved in kinkeeping (which would naturally equal to zero). A ZIP model therefore essentially estimates two equations simultaneously: one is a logit estimation that predicts membership in a latent class representing that a parent is not at all involved in kinkeeping (i.e., the Inflate equation) and the other is a Poisson regression estimation that predicts the range of kinkeeping activities they perform (i.e., the Poisson equation). Note that, as our count measure is capped, there is no need for adjusting our standard errors for overdispersion. The hypotheses were tested using three ZIP models (Table 3). We started with a basic model with all control variables, including the individual-level controls, the household-level controls, and the controls on parent-child contact and co-residence (Model 1). Thereafter, we ran a model in which we added parent gender and all parent-type indicators (Model 2). Finally, we tested the interactions between all parent-type indicators and parent gender (Model 3). To get a better understanding, we also showcase for married mothers, married fathers, remarried mothers, remarried fathers, stepmothers, and stepfathers separately the average range of kinkeeping in which they are involved (Fig. 1).
Second, to explore the effectiveness of parents' kin work, a series of OLS regression models assessed the associations between the range of kinkeeping activities in which parents and their partners are involved and the degree of closeness between parents and their adult children (Table 4). As we want to capture whether stepparents can add to family cohesion to a similar degree as biological parents, we are primarily interested in the association between the kinkeeping of the partner and the closeness of the parent and biological children. The hypotheses were tested using two sets of models, which were estimated separately for closeness with biological mothers and closeness with biological fathers. We started with a model that tested whether the range of kinkeeping activities in which the partner is involved benefits parent-child closeness in addition to the parents' own kinkeeping involvement (Model 1 and 3). Thereafter, we added the interaction between kinkeeping and the dichotomous variable that captures whether the partner is a stepparent (as opposed to a biological parent; Model 2 and 4). Throughout the analyses, we control for three characteristics of the biological parent-child tie that have the potential to influence current closeness: closeness during the period of upbringing, the duration of co-residence, and no contact between the parent and adult child. In doing so, we provide a more conservative test that accounts for potential confounding influences that are linked to closer ties with biological parents, as well as increased parental involvement from the spouse. life of the reporting child, the number of years they have spent co-residing with the reporting child, and the opportunities they had to invest in and build a shared history with the reporting child. Such differences need to be considered when, for instance, trying to establish if there is a gap in kinkeeping involvement between biological and stepparents (Arránz Becker et al., 2013;Kalmijn et al., 2019).
In addition, we captured two household-level characteristics: partnership quality and the total number of siblings attached to the parental household. Partnership quality was measured by asking "How would you describe the partnership between your [parent] and [parent/new partner of parent]?" with answers running from 1 (very good) to 5 (very bad). This variable was recoded so that higher values refer to higher-quality partnerships. Moreover, we measured the total number of siblings the adult has via either parent within the household (also counting half-and stepsiblings). Siblings are thus biologically related to one or both parents and stepsiblings are only counted if they have at some point coresided with the respondent, but a sibling does not have to still be a resident in the familial household. Note that we did not create a control on the partnership type between the two parents, as the contrast between marriage and cohabitation is less pronounced in the context of Western Europe (Thomson, 2014) and has been found not to affect intergenerational ties (Kalmijn et al., 2019). An additional check showed that among our sample of children -born between 1971 and 1991 in the Netherlands -our main findings are indeed unaffected by the exclusion of partnership type from the models. Finally, we controlled for several individual-level characteristics: the age of the parent (25 to 87), the age of the adult child (25 to 45), and the gender of the adult child: male (0) or female (1).

Analytical Strategy
To test our hypotheses, reports of the adult children on married mothers, married fathers, repartnered mothers, repartnered fathers, stepmothers, and stepfathers were pooled. Because some cases in this pooled dataset are dependent in that certain parents ("alters") are nested within the same reporting adult children ("anchors"), the standard errors in all our models were adjusted using a VCE correction for clustering. We estimated two sets of models to test our hypotheses.
First, to examine parents' relative involvement in kinkeeping, Zero-point Inflated Poisson (ZIP) models were estimated to assess the differences between parent types in terms of kinkeeping involvement. This method is the most appropriate for our outcome, as it involves a count measure which is nonlinear in that we observe more zeroes than would be expected based on a Poisson distribution (for more information, see Lambert, 1992;Long & Freese, 2006). Note. * p < .05, ** p < .01, S.E. in brackets, estimated with a VCE adjustment for clustering. An additional check showed that these results hold if parent education is included as a control (data availability precludes its inclusion in the main analyses).
Source: Ouders en Kinderen in Nederland (OKiN), Wave 1 and 2 none of the kinkeeping behaviors. The coefficients of the Poisson portion can be interpreted as the effect on the number of kinkeeping behaviors in which a parent engages. For both equations, the model estimates display the gendered differences between continuously married parents and repartnered parents and between stepparents and biological parents in kinkeeping involvement, controlling for contact, contact frequency, and duration of co-residence.
The analyses confirm that there is an overall gender difference in kinkeeping involvement (Model 2): mothers are generally more involved in kinkeeping than fathers, regardless of the family structure in which the parents are embedded (i.e., married versus repartnered structure) or the nature of their relationship with the reporting adult child (i.e., biological parent versus stepparent). More specifically, the Inflate portion of the equation shows that the likelihood that mothers are completely uninvolved in kinkeeping is smaller in comparison to fathers: mothers are 90% (1-e − 2.29 ) less likely than fathers to be uninvolved (p < .01). Mothers are also involved in a larger range of kinkeeping activities than fathers. If we look at the Poisson equation, the number of kinkeeping behaviors is higher among mothers than fathers, with mothers performing about 56.8% (e 0.45 ) more types of kinkeeping behaviors than fathers (p < .01). These results are in line with H1. An additional check showed that the gender gap in involvement is consistent across the separate items used in our count measure, although minimal differences could be detected in terms of size: the gendered

Kinkeeping Involvement
To examine parental involvement in kin work, we performed a Zero-inflated Poisson (ZIP) model in which the Inflate equation displays the likelihood of no involvement in kin work and the Poisson equation displays the likelihood that an involved parents engages in a larger range of kinkeeping activities. The model is presented in Table 3. The coefficients of the Inflate portion can be interpreted as the effect on the likelihood of a parent being involved in Our results also confirmed that the parent's biological relatedness with the reporting adult child plays a role in the range of kinkeeping in which the parent engages. To illustrate, the Inflate portion shows that stepparents are approximately 4 times more likely than (repartnered) biological parents to be uninvolved in kinkeeping (e 1.64 ) and, if we look at the difference between stepparents versus biological parents in the Poisson portion, we see that the number of kinkeeping behaviors performed by stepparents is lower by a factor of 0.763 (e − 0.27 ). The latter means that, once involved, stepparents also perform 23.6% less types of kinkeeping behaviors than biological parents, holding all other variables in the model constant (p < .01). These results comply with H3a.
Moreover, the gap between biological and stepparents' kinkeeping was gendered, but only with respect to the question of whether a parent is involved in kinkeeping at all. The interaction model suggests that a stepfather is more likely to be uninvolved in kinkeeping than a repartnered biological father (with a factor of 6.82; e 1.92 ). The same contrast exists for stepmothers and biological mothers, but the gap is significantly smaller for mothers (1.99; e 2.01−1.01 ) compared to fathers (p < .01). This implies that, regardless of the less frequent involvement of stepparents compared to biological parents, stepmothers are still rather frequently involved in any kinkeeping (even more so than repartnered biological fathers, see Fig. 1). Among those involved, we see that stepfathers perform about 22.1% (1-e − 0.25 ) fewer types of kinkeeping behaviors than (repartnered) biological fathers, and stepmothers perform about 18.1% (e − 0.25+0.05 ) fewer types of kinkeeping behaviors than (repartnered) biological mothers, but the difference between the two gaps was not significant (p > .05). Based on these results, H3b was partly supported, whereas no evidence was found in support of H3c.

Kinkeeping Consequences
As the analyses on kinkeeping involvement established that there are differences in parental kinkeeping behaviors according to gender, family structure, and biological relatedness, we continued with the question of whether involvement in kin work facilitates the closeness adult children feel towards other members of the familial household. Table 4 shows the estimates of the OLS regression model testing whether the range of kinkeeping in which a biological parent is involved, as well as the range of kinkeeping in which the spouse is involved, relates to a higher degree of emotional closeness. The results are shown separately for closeness with biological mothers and closeness with biological fathers for exploratory purposes.
The analyses confirmed that parents' involvement in a larger range of kinkeeping behaviors is related to closer contrast is largest with respect to buying presents (item 1) and smallest with respect to discussing problems (item 4).
Moving to the overall impact of family structure (Model 2), the Inflate portion shows that the likelihood that a parent is not at all involved in kinkeeping is higher among continuously married biological parents compared to repartnered biological parents with a factor of 3.46 (e 1.24 ). A small but significant difference is also apparent with respect to the range of kinkeeping behaviors in which a parent engages. That is, the number of kinkeeping behaviors in which married parents engage is less than repartnered parents by a factor of 0.914 (e − 0.09 ). This suggests that continuously married parents display 8.6% fewer kinkeeping behaviors, holding all other variables in the model constant (p < .01). These results are in line with H2a and suggest that parents' embeddedness in a repartnered household might work as an additional motivation to engage in kin work, compensating for the disruptive effect of parental divorce or remarriage on the kinship system. However, the significant interactions with gender (Model 3) show that the main effect of family structure is strongly gendered and should thus, be interpreted separately for mothers and fathers.
In the interaction model, the main effect of family structure can now be interpreted as the difference between married and repartnered biological fathers. Continuously married fathers are more likely than repartnered fathers to be involved in none of the kinkeeping activities (by a factor of 5.1; e 1.63 ), and, once involved, also engage in a smaller range of kinkeeping behaviors compared to repartnered fathers. To illustrate, among the group of involved fathers, repartnered fathers perform 38.1% (1-e − 0.48 ) more types of kinkeeping behaviors than continuously married fathers (p < .01). A similar contrast does not exist between continuously married mothers and repartnered mothers, however. In fact, according to our analyses, a repartnered mother is actually more likely to be involved in none of the kinkeeping behaviors than a continuously married mother (difference of about 95%, 1-e 1. 63-4.67 ) and, once involved, there is no significant difference in the range of kinkeeping activities performed (e 0.47-0.48 ). In general, the gap between married and repartnered biological mothers is smaller than the gap between married and repartnered biological fathers (as visualized in Fig. 1), which supports H2b. Taken together, these findings are consistent with the idea that divorce works as an activator for fathers but not mothers. To get a better grasp of the implications, we also visualized the results to explore the average range of kin work per parent (see Fig. 1). This figure shows that, regardless of the finding that repartnered fathers are more involved in kin work than continuously married fathers, repartnered mothers are on average still engaged in a substantially larger variety of kinkeeping activities than repartnered fathers. insight into fathers' and mothers' involvement in kinkeeping and investigated the association between kin work and adult intergenerational closeness.
Our descriptive findings showed that, even though multiple parents can engage in kin work simultaneously, the kinkeeper role remains clearly defined across parent types in contemporary families, with specific characteristics making some parents more likely than others to be involved in adult children's lives as active kinkeepers. Based on our findings, three main conclusions can be drawn about parents' involvement in kinkeeping.
First, involvement in kin work is strongly gendered. Mothers more often engage in kinkeeping than fathers and, once involved, also engage in a larger variety of kinkeeping behaviors than fathers. The existing literature on kin work, which is focused primarily on married families, suggests that kinkeeping is performed exclusively by women. Our findings confirm that mothers are much more involved in kin work across all studied household structures. The found patterns illustrate how societal expectations on family life and parental responsibilities translate differently into the kinkeeping behaviors of women and men. More specifically, it suggests that the kin work of connecting family members and managing relationships is still perceived to align most with traditional notions of motherhood (Braverman, 1989). This also complies with qualitative research, which showed that people often justify (unequal) patterns in kinkeeping in reference to gendered working lives and gendered views of parental roles and identities (Aronson, 1992).
Second, separated biological parents were found to more often engage in (a larger variety of) kinkeeping than married biological parents, but this applied to fathers only. A repartnered father more often engages in kinkeeping and, once involved, performs a larger range of kinkeeping behaviors compared to continuously married fathers. In other words, biological fathers increase their involvement in kin work under the specific circumstances of separation and repartnering. The biological mother, the "traditional kinkeeper", is expected to relay her kin work to her new familial households after separation, leaving the ties between separated fathers and children particularly vulnerable (Schmeeckle, 2007). Our findings thus suggest that divorced fathers try to compensate for losing their partnership with the main kinkeeper, the biological mother, although not enough to perform an equal number of kin activities compared to any mother. Surprisingly, the opposite is true among mothers: repartnered mothers are more often than married mothers uninvolved in kinkeeping and, once involved, both types of mothers perform an equal range of kinkeeping activities. In short, whereas fathers are activated to perform a larger variety of kinkeeping after divorce, a similar activation was absent among mothers. Further inquiry is needed to fully relationships with adult children, with similar results for biological mothers (B = 0.16, p < .01) and biological fathers (B = 0.13, p < .01). Moreover, the results indicated that the spouse's involvement in a wider range of kin work is also significantly associated to closer ties with adult children, showing a small but significant positive association for biological mothers (B = 0.07, p < .01) and biological fathers (B = 0.09, p < .01). Important to note is that this positive association is in addition to the role of biological mothers' and biological fathers' own involvement in a larger range of kinkeeping. Hence, closeness with biological children may benefit from and, in some cases, even depend on the facilitative behaviors of the other parent figure within the familial household. Overall, the associations are small (additional stepwise models showed ~ 6.5% variance in father-child closeness and about ~ 3.1% variance in mother-child closeness is explained by kinkeeping) but significant, and rather interesting given that we focus on a sample of adult and independently living children.
Moreover, our results also showed to what extent the association with adult intergenerational closeness works differently in continuously married versus repartnered households. The interaction models (Model 2 and 4) tested whether the effectiveness of kinkeeping differs based on the nature of the relationship between the spouse and the adult child (see Table 4). Surprisingly, the interactions were found to be nonsignificant and, as the absent change in the R 2 indicated, added virtually no explained variance in adult closeness with biological mothers and fathers. In other words, the association with parent-child closeness was not substantially weaker when the partner had a social rather than biological relationship with the reporting adult child. The implication of this finding is that biological parents and stepparents can be equally effective in facilitating the ties between their adult children and partner. In other words, kinkeeping can be effective in maintaining a connection between an adult child and the parental household, regardless of the type of parent involved.

Discussion
The dyadic relationship between a parent and adult child may in part depend on the facilitative involvement of a spouse (Bildtgård et al., 2021). Such kinkeeping involvement is the work of bringing together and encouraging cohesion among family members (Rosenthal, 1985). Parental divorce disrupts this system, as parents not only lose a significant partnership but often also a source of mutual aid in maintaining close relationships with any shared adult children. Guided by the lack of empirical evidence on intergenerational kinkeeping patterns, especially within the context of divorce and remarriage, this research provided the potential to be successful facilitators of the closeness between separated parents and their adult children. This implies a particularly important role for stepmothers in the contact between non-resident fathers and their adult children. In short, both with respect to intergenerational patterns in involvement as well as consequences, gender differences persist and continue to shape parental kinkeeping within the context of divorce and repartnering.

Limitations and Future Research Directions
A few limitations of this study should be considered when interpreting our results. First, we use a count measure that captures the range of kin activities in which parents are involved and thus, does not capture intensity. Still, the count measure was preferable over the original relative measure, because it allowed for separate measures for all parents and complied best with earlier conceptualizations of kinkeepers as "communication links" (Rosenthal, 1985). Second, the data we use are cross-sectional, which means that, although we are able to control for specific aspects of the period of upbringing, we do not have information on possible variations in kinkeeping over time. Third, we cannot conclude with certainty what drives the found contrasts in kinkeeping involvement. An alternative that should be considered is the influence of the adult children themselves. Children are not passive bystanders in the development of family relationships but have the potential to encourage their parents to engage in kin work. To illustrate, the gap between married and separated fathers might in part be due to children expecting more involvement from separated fathers. More in-depth research on the role of child expectations would enrich the literature on adult intergenerational ties. Fourth, there is a risk of reverse causality with respect to results on the consequences of kinkeeping. After all, parents who have closer ties with their adult children may also want to more actively engage in activities that bring together and facilitate cohesion between these children and kin. An ideal alternative would be to use longitudinal information to establish the direction of the found association. This option is presently not available, but is highly encouraged in the future.
There are several ways that future studies could build upon our work. For one, the found tendency of divorced fathers to be more involved in kinkeeping involvement than married fathers (i.e., the compensation effect we hypothesized) was, due to the design of the questionnaire, solely based on repartnered fathers. Yet, the literature on "swapping families" suggests that divorced fathers often transfer the time and effort they invest in their family to their "new nest" once children are involved (Manning & Smock, 2000). If we would have compared married parents and divorced grasp these findings, although future works should consider the role of selection effects (i.e., those who select themselves into separation and remarriage may hold less traditional values on gender roles) and ceiling effects (i.e., the absent gap among mothers may be due to the already high involvement of married mothers, as those who would be activated are simply not able to further increase their involvement).
Third, parents are less involved in facilitating the kinship ties of adult children when there is no biological relationship. Our findings showed that stepparents were less often involved in kinkeeping than biological parents and, once involved, performed a more limited variety of kinkeeping behaviors. Situating these findings in the literature, we suggest that, due to normative ideas of biological versus social relations (Rossi & Rossi, 1990) and the ambiguity surrounding stepparenthood (Cherlin, 1978), stepparents feel less inclined and may experience barriers to engage in children's lives, or interfere in these children's ties with other kin. Interestingly, the likelihood for a stepparent to be involved in kinkeeping differed by gender: the barrier to become involved was higher for stepfathers (compared to biological fathers) than for stepmothers (compared to biological mothers). This suggests that, driven by their link with biological fathers, stepmothers are more prone than stepfathers to step into the kinkeeping role, in compliance with the idea of stepmothers as "carpenters" (Schmeeckle, 2007). Alternatively, women may behave more similarly across parental roles than men due to cultural expectations about the "natural" ability of all women to mother (Braverman, 1989;Gaunt, 2006;Weaver & Coleman, 2005). Either way, the smaller gap among mothers relative to fathers is surprising in light of the culturally constructed primacy of the biological mother role, which has previously been suggested to preclude the involvement of stepmothers in particular (Ganong & Coleman, 1995;van Houdt et al., 2020).
Finally, we examined parents' own as well as their spouse's involvement in kin work and found significant associations with parents' emotional closeness with adult children. Although the association with parents' own kinkeeping was larger in comparison, the spouses' kinkeeping was found to have an additive impact in facilitating closer parent-child ties. This finding is relevant for fathers' closeness with adult children in particular, especially within the context of divorce and remarriage. Due to gendered kinkeeping patterns, fathers' kinship ties are more likely to become unstable outside of marriage (Kalmijn, 2007). Surprisingly, we established that the found associations were not hampered in repartnered households, despite the fact that the spouse did not have a biological relationship with the adult child. Our findings therefore not only suggest that separated parents can increase their own intergenerational closeness by investing in kinkeeping, but also that stepparents have this may explain why the ties between biological fathers and children from lower socioeconomic backgrounds are particularly vulnerable post-separation.
Finally, we focused on a specific type of kin tie to test the consequences of kin work: adult intergenerational relationships. In taking this focus, we do not reflect on the influence of kinkeeping on ties with siblings or extended family. As such, we do not fully capture the wide range of "emotion work" needed to sustain a contemporary family (Gester & Gallagher, 1993). Among the kin ties that could provide an interesting starting point for future studies, adult children's relationships with elderly parents are particularly relevant, especially in light of the continuing debates on intergenerational "upward" solidarity in times of family complexity (Swartz, 2009). Notice that the items used to measure our kinkeeping construct are formulated in general terms, rather than referring to the mending of parent-child ties specifically. Future work could thus easily go beyond our work and connect kinkeeping to other types of family ties encompassing the kinship systems of adult children.

Practice Implications
The insights from our research have several implications for family practitioners but also for school counselors and health professionals in their guidance of families toward more positive family functioning and closer intergenerational bonds. First, our results could be used to educate families and family counselors about kinkeeping as a less often recognized form of household work. Although we do not provide any information on how parents experience kin work, research on cognitive labor imply a link with stressors that are also less visible and generally land on the plate of mothers (Daminger, 2019;Gerstel & Gallagher, 1993). Second, our results illustrate how gendered divisions of relationship maintenance make father-child contact and closeness greatly dependent on the kin work of a mother, creating specific challenges for fathers going through a divorce (Kalmijn, 2007). Third, our findings highlight that stepparents (especially stepmothers) are no "secondary actors" but actually have great potential to positively influence the kinship system. While we know from other qualitative work that stepmothers face significant challenges in navigating their role (Weaver & Coleman, 2005), we show that, when they are involved, stepmothers can be successful kinkeepers. On a more general level, this implies that practitioners that assist families post-divorce should consider influences beyond the traditional nuclear family in order to have complete picture and make sure that they do not leave certain opportunities for improvement untouched. single parents, the compensation effect among divorced fathers may have been even larger. After all, if a divorced father does shift his time investments to his children with a new partner, as suggested in the literature on "swapping families", this could hamper the tendency to increase their involvement toward the children he has with a previous partner. In addition, single fathers have no kinkeeping stepmother to rely upon (Schmeeckle, 2007). This comparison group would be interesting to include in future studies.
Although normative expectations about familial roles have been changing toward greater gender equalities over past decades, the parents and stepparents of the current generation of adults grew up and have raised their children within the cultural context of more traditional gender roles (Sigle-Rushton et al., 2013). The normative idea that mothers are the best caretakers of their children (the "motherhood mandate"; Hays, 1996;Russo, 1976), which is expected to underlie the found gendered contrasts, might therefore be more salient for the studied cohort. In future research, scholars should acknowledge how the found gaps in kin work might change over time alongside societal shifts in perceptions of gender and parenthood. Two trends are expected to have an influence in particular. First, although the emphasis has traditionally been on fathers as providers, intimacy between fathers and children has become more prominent in contemporary perceptions of fatherhood (Dermott & Miller, 2015;Lamb, 2013;Miller, 2010). Second, whereas maternal resident arrangements were still the norm when the studied cohort of adults was growing up, shared custody has become more common and residence arrangements have increased in variety (Poortman & van Gaalen, 2017). If future works replicate our questions for more recent cohorts, we can detect to what extent these trends are accompanied by different patterns in kinkeeping and adult intergenerational closeness.
Another important avenue for future research is to examine whether there is an educational gradient in the found contrasts in kinkeeping involvement with respect to gender, family structure, and biological relatedness. The trends toward more egalitarian gender roles, changing perceptions of fatherhood, and increased post-separation father involvement are all expected to be highly stratified (Poortman & van Gaalen, 2017). Although an additional check showed that our results remain unchanged when parents' educational attainment is controlled for, the found gaps could still differ in size if we compare them across different socioeconomic backgrounds. Fathering behaviors are not uniform, and have been found to depend on factors such as education, income and class (Edwards et al., 2009). One possibility that should be explored is that the gap between married and separated fathers in terms of kinkeeping involvement might be larger among lower compared to higher educated fathers. If so,

Conclusion
A few general conclusions can be drawn based upon this study, all of which have implications for the literature on gender dynamics within families. Traditional gender divides link women to the practice of caring for others, whereas they link men to the practice of providing for others through breadwinning and personal fulfillment (Johnston & Swanson, 2006;Sigle-Rushton et al., 2013). Recent demographic changes, in particular within the sphere of women's increased school and workforce engagement and more egalitarian divisions of (physical) household labor, suggest a weakening of these links. This contribution shows, however, that there remains a strong gender divide in the kin work of connecting family members and managing family relationships on behalf of the familial household, even within the context of divorce and remarriage. This further amplifies the idea of relationship maintenance as a somewhat "invisible" form of cognitive household labor that is mainly ascribed to the mother figure and emphasizes the more vulnerable position of fathers' kinship ties outside of marriage.

Conflicts of Interest
This study was conducted as part of a larger project on Family Complexity funded by the European Research Council under the Horizon 2020 program [ERC grand agreement no. 669334]. The OKiN was developed, designed, and executed by a collaborative team of researchers from the University of Amsterdam and Statistics Netherlands (CBS). The authors have no financial or non-financial conflicts of interest to disclose.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.