Stable Marital Histories Predict Happiness and Health Across Educational Groups

Mäki, Miika; Hägglund, Anna Erika; Rotkirch, Anna; Kulathinal, Sangita; Myrskylä, Mikko Article — Published Version Stable Marital Histories Predict Happiness and Health Across Educational Groups European Journal of Population Provided in Cooperation with: Springer Nature Suggested Citation: Mäki, Miika; Hägglund, Anna Erika; Rotkirch, Anna; Kulathinal, Sangita; Myrskylä, Mikko (2025) : Stable Marital Histories Predict Happiness and Health Across Educational Groups, European Journal of Population, ISSN 1572-9885, Springer Netherlands, Dordrecht, Vol. 41, Iss. 1, https://doi.org/10.1007/s10680-025-09733-x This Version is available at: https://hdl.handle.net/10419/330779 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) European Journal of Population (2025) 41:12 https://doi.org/10.1007/s10680-025-09733-x ORIGINAL RESEARCH Stable Marital Histories Predict Happiness andHealth Across Educational Groups MiikaMäki1,2,6,7 · AnnaErikaHägglund1,3· AnnaRotkirch1 · SangitaKulathinal4· MikkoMyrskylä5,6,7 Received: 15 December 2023 / Accepted: 23 February 2025 / Published online: 13 May 2025 © The Author(s) 2025, corrected publication 2025 Abstract Couple relations are a key determinant of mental and physical well-being in old age. However, we do not know how the advantages and disadvantages associated with partnership histories vary between socioeconomic groups. We create relationship history typologies for the cohorts 1945-1957 using the Survey of Health, Ageing, and Retirement in Europe and examine, for the first time, how relationship histories relate to multiple indicators of well-being by educational attainment. The results show that stable marriages predict greater well-being, compared to single and less stable partnership histories. The positive outcomes are similar across all educational groups. Those with lower education who have divorced experience even lower well-being in old age. The interaction analyses suggest that individuals with fewer resources could suffer more from losing a partner. The findings underscore that current and past romantic relationships are linked to well-being in old age and help policymakers identify vulnerable subgroups among the ageing population. Keywords Partnership history· Resource substitution· Cumulative disadvantage· Health· Quality of life 1 Introduction Romantic couple relations are among the most intimate, lasting and important relationships in our lives and a cornerstone of emotional, social, and economic well-being (Wängqvist et al., 2016; Luyckx et al., 2014). Previous research has exhaustively shown that having a partner is positively associated with a multitude of well-being outcomes in later life, such as life satisfaction, quality of life, health, morbidity, and mortality (Wong & Waite, 2015; Han et al., 2014; Holt-Lunstad etal., 2008; Coombs, 1991; Manzoli etal., 2007; Waite, 1995; Wang etal., 2022; Reneflot & Mamelund, 2012). Indeed, having a spouse is arguably one of the single most important factors for healthy ageing (Wood etal., 2009). At the same time, Extended author information available on the last page of the article M.Mäki et al. 12 Page 2 of 54 the proportions of single, divorced, or remarried elderly have increased (Cherlin, 2017), and relationship biographies have become more diverse. It is crucial to identify what types of partnership histories are related to increased vulnerability, and to what extent lacking or losing a partner can be compensated for. Having a partner as such may not improve well-being. Instead, the association between partnerships and well-being could emerge from the partnership history as the social, economic and health advantages of partnerships accumulate over time (cf. DiPrete & Eirich, 2006). Existing literature, however, tends to examine relationships either as single events, such as marriage or divorce, or depict them rather statically through the current partnership status (Sassler, 2010; Schütz, 2019; Gumà et al., 2014). So far, a few studies have conceptualised partnerships as trajectories (Zimmermann & Hameister, 2019; Roberson etal., 2018; Jung, 2023; O’flaherty etal., 2016). These suggest that aspects of partner histories predict well-being beyond the current status, and that further insights into well-being can be gained by forming trajectories (Peters & Liefbroer, 1997; Jung, 2023; Zimmermann & Hameister, 2019). A first contribution of this paper is that we conceptualise relationships as trajectories and see whether they are linked with well-being. Several scholars, including Becker (1973), have stressed that the gains from marriage are likely to depend on individual traits, such as attractiveness, intelligence, and education. These different gains may arise because of homogamy, higher marital satisfaction, and more fulfilling family life and ties. So far, scholars have not considered education in greater detail. Instead, educational attainment has been treated as a control rather than a stratifying feature (Boyce etal., 2016; Han etal., 2014; Wong & Waite, 2015). To our knowledge, only one study in recent decades has examined whether the association between marital status and well-being indicators varies by education: Øien-Ødegaard etal. (2021) found that low education increased the suicide risk among separated. There are several reasons to assume that the gains from a stable partnership history follow an educational gradient. On the one hand, highly educated individuals might profit more from having a lifelong partner than those with a lower education due to lower conflict (Woszidlo & Segrin, 2013), more fulfilling family life (Roberts & Dunbar 2011; Margolis & Myrskylä 2011) and the possible compounding effect of different types of resources that both education and a stable relationship provide (Zajacova & Lawrence, 2018; Wood etal., 2009). This process would imply resource multiplication. On the other hand, those with a lower educational level may rely more heavily on a spouse, as stable partnerships compensate for a lack of other resources. Put differently, spousal loss might have more detrimental consequences for those with fewer resources (DiPrete & Eirich, 2006). The latter process, in turn, suggests resource substitution. In this study, we analyse two such dimensions that have not been studied jointly before: whether particular relationship histories predict well-being and whether the well-being associations of those histories vary by education. We investigate whether romantic relationship histories are associated with well-being for men and women and how the associations differ between educational groups of Northern and Western Europeans in third age, a stage in one’s life between middle age and old age (Gilleard & Higgs, 2007). Specifically, we are interested in signs of resource multiplication or Stable Marital Histories Predict Happiness andHealth Across… Page 3 of 54 12 substitution: is the combination of having a certain educational level and relationship trajectory linked with even higher or lower well-being outcomes? Empirically, we create partnership history typologies with sequence analysis using retrospective life history interview data from the Survey of Health, Ageing, and Retirement in Europe (SHARE) from 13 European countries. We measure wellbeing with subjective health and life satisfaction and conduct robustness checks with grip strength and CASP-12 quality-of-life indicator. Our analyses control for several confounding measures including childhood conditions, income, and number of children. The results are further corroborated by a confounder analysis. 1.1 Why are Relationship Status andWell‑Being Linked? Empirical evidence for the interconnectedness of marital status and well-being is conclusive. Current marital status and the quality of couple relationships are generally linked to well-being (see review in Wood etal., 2009). These associations seem to prevail in different societies: Diener etal. (2000) studied the relationship between marital status and subjective well-being in 42 countries and found that differences in effect sizes are negligible. Verbakel (2012) concludes with a similar study design that "normative climate appears to hardly affect well-being gaps between partnership statuses". Thus, we expect partnerships to support well-being similarly across our study population. But what explains the consistency of these findings? Previous research points to three mechanisms, namely selection into marriage, marriage as a key social role, and the everyday social support of living with a spouse (Dush & Amato, 2005; Averett etal., 2013). First, marriage itself might not make people happy and healthy. Instead, happy and healthy people could self-select into marriage (Koball etal., 2010). Although some selection effects are at play (see Goldman, 1993, for a theoretical overview), the majority of longitudinal studies suggest that selection alone does not account for all the positive effects of relationships on well-being (Horwitz etal., 1996; Wood et al., 2009; Fu & Noguchi, 2018; Grover & Helliwell, 2019). Yet not all agree (Mastekaasa, 1992; Ludwig & Brüderl, 2018), highlighting that results may vary with methodological design. Second, the structural symbolic interactionist perspective argues that roles with a high level of commitment result in high levels of identity and self-worth (Stryker & Statham 1985, as cited in Dush & Amato 2005). Internalising the role of a wife or husband (or parent) shapes both the behaviour and self-perception, and consequently enhances well-being, although this depends on their commitment to the given identities (Stryker, 1959). The fact that few want to live without a long-term romantic partner even in wealthy, individualised, and liberal societies (Kontula 2016; Rotkirch 2020, pp. 40-41) demonstrates the high value attached to couple relations. Third, marriage tends to provide social support and regulation beyond any other form of relationship (Ross, 1995; Coombs, 1991; Scott, 2000). In high-income societies, unions are the primary household unit for bread-winning, consumption, and intimacy. Particularly in the ageing population, the typical alternative to living with a romantic partner is to live alone (Rindfuss & Vandenheuvel, 2022; Becker, M.Mäki et al. 12 Page 4 of 54 1991; Rotkirch, 2020). Individuals who are attached to social networks are healthier and live longer (House etal., 1988); those with a spouse have someone in the same household to share the joys and sorrows of life in the long term, which single adults often lack. 1.2 Relationships andWell‑Being fromaLife Course Perspective Well-being in old age reflects combined experiences over the entire life course, not just the present moment. Studying sequences of relationship statutes, or relationship histories, could reveal dynamics that are concealed in analyses focusing on relationship statuses. The focus on sequences aligns with life course theory: Individual outcomes are the result of previous trajectories and events. These evolve with time, are embedded in social structures, and form patterns of social stratification (Elder etal., 2003; Diewald & Mayer, 2009). Empirical research points to the value of analysing relationship trajectories. For instance, marriage dissolution has long-term, negative implications on well-being and health, which persist even among those who remarry (Hughes & Waite, 2009). Correspondingly, unstable partnerships, multiple relationship transitions, and longterm singlehood are associated with higher levels of depression and stress and lower social and emotional support (Jung, 2023; Zimmermann & Hameister, 2019). Partnership trajectories characterised by stable unions typically display the highest levels of psychological and physical well-being (O’flaherty et al., 2016; Tambellini etal., 2024). The association between partnership histories and well-being is likely to increase with age, since inequalities in health, resources, and subjective wellbeing tend to accumulate as individuals grow older (Kratz & Patzina, 2020; Ross & Wu, 1996; Prus, 2007; Headey, 2010). Finding a new partner, however, can at least partly compensate for more unstable partnership trajectories (Jung, 2023). Overall, this process aligns with the framework of cumulative advantage/disadvantage, suggesting that negative or positive implications in one aspect of life increase over time (Dannefer, 2003). As relationship statuses are associated with lifestyles, health behaviour, social support, and networks, well-being—or lack thereof—is likely to spill over to other domains and accumulate as individuals age. Melo etal. (2019) goes as far as claiming that the ageing process as such is a result of accumulated dis/advantages over the years from different spheres of life. In this study, we use cumulative dis/advantage to refer to the compound wellbeing effects of relationship trajectories over time. Additionally, we hypothesise how educational attainment could potentially amplify or mitigate cumulative dis/advantage (see Sect. 1.4). 1.3 Gender Differences Empirical research highlights that the implications of romantic relationships differ by sex. In particular, men seem to benefit significantly from being in a partnership (Coombs, 1991; Metsä-Simola & Martikainen, 2013); long-term singlehood substantially lowers men’s well-being (Zimmermann & Hameister, 2019; Jung, 2023). Stable Marital Histories Predict Happiness andHealth Across… Page 5 of 54 12 For women, by contrast, the association between partnerships and well-being appears weaker (O’flaherty etal., 2016; Zimmermann & Hameister, 2019). At first glance, this seems counterintuitive, as women face greater economic disadvantages following union dissolution than men (see Mortelmans, 2020, for a review). For instance, women’s salary growth rates decline more sharply, leaving them at a longterm economic disadvantage (Raz-Yurovich, 2013). Additionally, women are more likely to take on primary childcare responsibilities after a breakup, which compounds their financial strain. Despite these economic disadvantages, several factors could explain why relationship breakups tend to have a greater detrimental effect on men’s well-being, particularly in later life. Women are more likely to engage in behaviours that promote health and often influence their partners to do the same (Reczek etal., 2016; Umberson, 1992; Westmaas etal., 2002; Lewis & Butterfield, 2007). Men, on the other hand, tend to have smaller and less emotionally supportive social networks in later life (Ajrouch etal., 2005). While single mothers bear a heavier financial burden after union dissolution, they often maintain stronger bonds with their children and receive more care in old age. Fathers who do not live with their children after a breakup may instead pay a higher social price. (Mortelmans, 2020) Furthermore, women often experience greater psychological distress from staying in marriages with low relationship quality (Bulanda etal., 2016; Brown & Wright, 2017; Carr & Utz, 2020). This may make them more inclined than men to end an unsatisfactory relationship and better equipped to cope with life without a significant other (Rosenfeld, 2018). 1.4 Educational Resource Multiplication orSubstitution Stable unions are likely to increase well-being, while unstable histories and longterm singlehood lower it. However, the outcomes of relationship histories might follow an educational gradient. Education anchors the accumulation of human, social, and personal capital (O’Rand, 2001), and is positively associated with a multitude of well-being outcomes (see Zajacova & Lawrence, 2018, for review). Theoretically, it remains unclear whether highor low-educated benefit more from stable unions. In the following, we argue that two mutually exclusive patterns are possible: resource multiplication and resource substitution. On the one hand, high education might strengthen the positive association of stable marital histories on well-being. Stable relationships allow for the accumulation of resources, and highly educated have more resources to sustain long-term relationships to begin with: they are likely to partner with similarly positioned individuals (Schartz, 2010), receive more support from the kin network (Roberts & Dunbar, 2011), and are better equipped to deal with conflict (Woszidlo & Segrin, 2013). These factors contribute to longer and happier unions among highly educated (Tavakol etal., 2017). Stable unions, in turn, provide a platform for resource accumulation. Highly educated men and women could also profit more from marital stability. For example, combined financial resources (Jung, 2023), spousal emotional support, M.Mäki et al. 12 Page 6 of 54 fulfilling relationships with kin-network (Roberts & Dunbar, 2011), or the joys of raising children could be more pronounced among highly educated individuals (Margolis & Myrskylä, 2011). Likewise, since both education and stable marriages predict health and life satisfaction (Zajacova & Lawrence, 2018; Wood etal., 2009), their interplay could result in multiplicative well-being outcomes. Thus, we assume resource multiplication if high education coupled with stable partnership histories mutually reinforce well-being (Ross & Mirowsky, 2006). Alternatively, low-educated individuals might profit more from stable partnership trajectories than highly educated. As low-educated individuals generally have fewer resources (Ross & Mirowsky, 2010), their well-being might rely more on stable partnership histories. The loss of spousal financial and emotional support is likely to influence well-being more strongly in this group than among the highly educated, who can buffer the negative consequences of unstable trajectories and living alone with other social, economic, and psychological resources (Ferraro etal., 2009). This pattern would suggest resource substitution (Mirowsky& Ross, 2005): stable union histories would have stronger benefits, or conversely, single or fragmented union histories stronger disadvantages among low-educated individuals, as they have fewer resources of resilience. Theoretically, resource substitution and multiplication should not co-exist: either the benefits of a stable partnership history are stronger for those with higher education (resource multiplication), the benefits of a stable partnership history are stronger for those with lower education (resource substitution), or there are no educational interaction effects (cf. Ross & Mirowsky, 2006). 1.5 Hypotheses Based on previous research, we formulate a set of hypotheses. First, following cumulative advantage and disadvantage, we assume that stable marital relationship histories are associated with higher well-being after age 60 (1a), while histories characterised by instability or long-term singlehood are associated with lower wellbeing (1b). In addition, we expect (1c) more robust and consistent results for men, but possibly also some associations for women. We anticipate that the link between partnership histories and well-being could vary by education, but the direction is not clear. Given the mutual exclusivity of resource substitution and multiplication, and the lack of conclusive findings in previous research, we remain agnostic as to which, if either, will be supported by our study. If the educational gradient is in line with resource multiplication, we expect high education to predict even stronger positive associations between stable partnership histories and well-being (2a). Education would then amplify the cumulative advantage of having a stable marital history as the highly educated profit more from family life. If resource substitution occurs, we expect that low education predicts stronger negative well-being outcomes for lifelong single and unstable partnership histories (2b). In such case, the low educated could neither utilise educational nor spousal resources, compared to those with high education—who can buffer unstable partnership trajectories with resources attached to education. Stable Marital Histories Predict Happiness andHealth Across… Page 7 of 54 12 2 Data andMethods Life histories were generated with the help of the SHARELIFE interviews of the Survey of Health, Ageing, and Retirement in Europe (Börsch-Supan & Bergmann, 2019).1 SHARE is a micro-panel data infrastructure that covers households with at least one member over 50 years of age in all EU countries, Switzerland, and Israel. The survey collects both panel and retrospective life course data. The retrospective SHARELIFE pseudo-panel interviews were conducted in 2008 and 2017 (wave 3 and 7, respectively). Respondents were asked to report, amongst others, on their childhood circumstances as well as their past partners, including all cohabitational, marital, and dating relationships. If the life history interviews were conducted in 2008, any changes in relationship statuses were updated with subsequent panel interview data. Our analyses are based on Northern and Western Europe as defined by United Nations (UNSD, 1999). The sample was restricted to these geographical areas to ensure that our study population would come from roughly similar cultures in terms of family formation and family life patterns (Klüsener, 2015). As a robustness check, we examined whether patterns varied between welfare states and cultures within Northern and Western Europe by stratifying the inferential analyses for welfare regimes following the classification scheme in Eikemo etal. (2008): Nordic (Sweden, Denmark, Finland), Baltic (Estonia, Latvia, Lithuania), and Bismarckian (Austria, Belgium, France, Germany, Luxembourg, Netherlands, Switzerland) welfare regimes. We selected respondents born between 1945 and 1957 for two reasons: they were at least 60 years old in 2017 when the data were collected and were all part of the baby boomer generation (Bavel & Reher, 2013). This generation formed unions at a time when the prevalence of marriage was at its peak. Yet, this is also a generation when premarital cohabitation and non-marital childbearing started to become more common, starting from Northern Europe. (Perelli-Harris & Amos, 2015; Klüsener, 2015). The final sample size for analysis was 18,256 individuals. In order to obtain an adequate sample size, all couples were heterosexual. Here, we categorise partnership statuses into unmarried, first and higher-order marriages, dating, cohabitation, divorce, and widowhood. This takes differences between partnered and unpartnered individuals more exhaustively into consideration (Næss etal., 2021; Pinquart, 2003): Although cohabitation and marriage increasingly resemble each other both legally and socially (Cherlin, 2020; Perelli-Harris & Amos, 2015), there is still a clear difference in terms of commitment, longevity, and symbolic importance (Barlow etal., 2001; Cherlin, 2004; Rault, 2019; Brown & Wright, 2017). Similarly, firstand higher-order marriages differ in terms of stability, relationship satisfaction, and demographic makeup (Zahl-Olsen etal., 2019; 1 This paper uses data from SHARE Waves 1, 2, 3, 4, 5, 6, and 7 (DOIs: 10.6103/SHARE.w1.710, 10.6103/SHARE.w2.710, 10.6103/SHARE.w3.710, 10.6103/SHARE.w4.710, 10.6103/SHARE.w5.710, 10.6103/SHARE.w6.710, 10.6103/SHARE.w7.711). See Börsch-Supan etal. (2013) for methodological details. M.Mäki et al. 12 Page 8 of 54 Hägglund etal., 2021; Booth & Edwards, 1992). Thus, later life outcomes could differ as well. We also have information on dating relationships. In SHARELIFE, dating is defined as a romantic relationship in which couples do not live at the same address most of the time. Dating here includes those who have a romantic relationship before moving in together, as well as those who permanently are in a living apart together (LAT) relationship (Lewin, 2017). Such relationships tend to be more flexible and less committed than marriage or cohabitation (Régnier-Loilier, 2016; Duncan & Phillips, 2010). In our operationalisation, an individual is dating if they enter a romantic relationship without co-residence, irrespective of previous status. In a first step, we create relationship history typologies. By partnership history, we mean the ordered series of relationship statuses from adolescence to old age (cf. Cornwell, 2015, p. 21).2 The partnership trajectories were created by sequence analysis or agglomerative hierarchical clustering (see Cornwell, 2015, for a theoretical overview). We first created a distance matrix with the dynamic hamming method. The substitution costs of two states at position p are calculated as follows: where sp(A,B) is the substitution cost at position3 p and Xp is the state4 at the pth position. The method assesses both the probability of being in state A and state B at position p as well as the transition probabilities from A to B at position p and vice versa. This means that both timing and order are taken into account. (Cornwell, 2015, p.128) Thus the substitution costs are calculated automatically based on how common a certain transition is at a given age. For example, a transition from marriage to widowhood will be given a much higher cost at the age of 20 than 60. We chose to use dynamic hamming method due to the fact that the substitution costs are derived empirically from the data. We still found very similar clusters with dynamic hamming and different variants of optimal matching. The distance matrix between the individual trajectories was further analysed using Ward’s agglomerative clustering method to form clusters (Ward, 1963). The clusters are obtained by minimising the within-cluster sum of squares, and hence produce groups with similar histories. We also experimented with other algorithms but the clustering alternatives with Ward’s method made the most sense conceptually. It is calculated as follows: (1) s p(A,B)=4− ( pr(Xp=A | Xp−1=B)+pr(Xp=B | Xp−1=A ) +pr(X p+1 =A | X p =B)+pr(X p+1 =B | X p =A) ) (2) Δ( J,K)= ∑ i∈J∪K‖ ‖ xi−mJ∪K‖ ‖ 2 − ∑ i∈J‖ ‖ xi−mJ‖ ‖ 2 − ∑ i∈K‖ ‖ xi−mK‖ ‖ 2 = nJnK n J +n K‖ ‖ mJ−mK ‖ ‖ 2, 2 We use the terms partnership histories and partnership trajectories interchangeably throughout this article. 3 Ages 15 to 60. 4 Singlehood, first marriages, higher-order marriages, dating, cohabitation, divorce, or widowhood. Stable Marital Histories Predict Happiness andHealth Across… Page 15 of 54 12 returned to singlehood in a few years, or had several short relationships (see Fig.9 for individual life courses). Overall, the clusters resemble each other in demographic characteristics (see Table 2), although there are a few exceptions: men were over-represented in the singlehood (5) and women in the divorce (3) cluster. While educational differences were generally insubstantial, men with stable relationship trajectories tended to be more educated. For women, the opposite pattern prevailed (see Fig.7). The distribution of relationship clusters did not vary much across Fig. 1 Partnership history clusters for Western and Northern Europeans born in 1945–1957 at the age of 15–60 M.Mäki et al. 12 Page 16 of 54 Table 2 Descriptive statistics by cluster 1.Marriage 2.Remarriage 3.Divorce 4.Serial cohabitation 5.Single Overall (N=12135) (N=1960) (N=1720) (N=1189) (N=1206) (N=18210) Year of birth Mean (SD) 1950.6 (3.6) 1950.7 (3.6) 1951.0 (3.6) 1951.4 (3.6) 1951.0 (3.7) 1950.8 (3.6) Gender Male 5610 (46.2 %) 865 (44.1 %) 550 (32.0 %) 532 (44.7 %) 646 (53.6 %) 8203 (45.0 %) Female 6525 (53.8 %) 1095 (55.9 %) 1170 (68.0 %) 657 (55.3 %) 560 (46.4 %) 10007 (55.0 %) Education Higher 4768 (39.3 %) 756 (38.6 %) 664 (38.6 %) 461 (38.8 %) 475 (39.4 %) 7124 (39.1 %) Secondary 4555 (37.5 %) 812 (41.4 %) 659 (38.3 %) 433 (36.4 %) 415 (34.4 %) 6874 (37.7 %) Basic 2690 (22.2 %) 383 (19.5 %) 381 (22.2 %) 283 (23.8 %) 294 (24.4 %) 4031 (22.1 %) Missing 122 (1.0%) 9 (0.5%) 16 (0.9%) 12 (1.0%) 22 (1.8%) 181 (1.0%) Income Mean (SD) 2676.5 (14396.5) 2603.4 (15796.4) 2496.8 (7764.3) 2941.1 (9428.9) 2418.8 (5347.4) 2651.3 (13337.5) Missing 775 (6.4%) 87 (4.4%) 61 (3.5%) 60 (5.0%) 54 (4.5%) 1037 (5.7%) Country Austria 992 (8.2 %) 149 (7.6 %) 165 (9.6 %) 86 (7.2 %) 105 (8.7 %) 1497 (8.2 %) Belgium 1642 (13.5 %) 231 (11.8 %) 271 (15.8 %) 163 (13.7 %) 144 (11.9 %) 2451 (13.5 %) Denmark 1084 (8.9 %) 235 (12.0 %) 113 (6.6 %) 154 (13.0 %) 86 (7.1 %) 1672 (9.2 %) Estonia 1273 (10.5 %) 283 (14.4 %) 201 (11.7 %) 167 (14.0 %) 127 (10.5 %) 2051 (11.3 %) Finland 607 (5.0 %) 71 (3.6 %) 72 (4.2 %) 80 (6.7 %) 66 (5.5 %) 896 (4.9 %) France 1277 (10.5 %) 156 (8.0 %) 186 (10.8 %) 117 (9.8 %) 144 (11.9 %) 1880 (10.3 %) Germany 1363 (11.2 %) 257 (13.1 %) 121 (7.0 %) 61 (5.1 %) 105 (8.7 %) 1907 (10.5 %) Latvia 405 (3.3 %) 84 (4.3 %) 90 (5.2 %) 31 (2.6 %) 31 (2.6 %) 641 (3.5 %) Lithuania 533 (4.4 %) 68 (3.5 %) 92 (5.3 %) 28 (2.4 %) 30 (2.5 %) 751 (4.1 %) Luxembourg 479 (3.9 %) 48 (2.4 %) 46 (2.7 %) 12 (1.0 %) 37 (3.1 %) 622 (3.4 %) Stable Marital Histories Predict Happiness andHealth Across… Page 17 of 54 12 Table 2 (continued) 1.Marriage 2.Remarriage 3.Divorce 4.Serial cohabitation 5.Single Overall (N=12135) (N=1960) (N=1720) (N=1189) (N=1206) (N=18210) Netherlands 647 (5.3 %) 68 (3.5 %) 60 (3.5 %) 29 (2.4 %) 76 (6.3 %) 880 (4.8 %) Sweden 1012 (8.3 %) 175 (8.9 %) 166 (9.7 %) 189 (15.9 %) 134 (11.1 %) 1676 (9.2 %) Switzerland 821 (6.8 %) 135 (6.9 %) 137 (8.0 %) 72 (6.1 %) 121 (10.0 %) 1286 (7.1 %) M.Mäki et al. 12 Page 18 of 54 countries. One stable marital trajectories were by far the most common, and single or serial cohabitation trajectories were the rarest. Further inspection by multinomial regression revealed that Northern European respondents were more likely to belong to the serial cohabitation cluster. No other distinct patterns were observed (see Table3 and Fig.8). Fig. 2 Linear regression for subjective health and life satisfaction without interaction effects (Models 1,2 & 3). Note: Marriage (c1) is the reference category Stable Marital Histories Predict Happiness andHealth Across… Page 19 of 54 12 3.2 Well‑being Outcomes Next, we analyse the association between partnership histories and two measurements of well-being, namely subjective health and life satisfaction. The results for main effects (models 1-3) are depicted in Fig.2, where the marriage (1) cluster is our reference category. As all non-binary variables were standardised, an effect size of one would imply a difference of one standard deviation in the response variable compared with the reference category. The white circles and squares represent point estimates, while the thick and narrow bars the 95% and 99% confidence intervals, respectively. Overall, the results are in line with our hypotheses in that trajectories characterised by stable marital unions (cluster 1) were associated with higher subjective health and higher life satisfaction for men and women (hypothesis 1a). Adults whose trajectories were characterised by singlehood (5) and divorce (3), in turn, experienced the lowest well-being in old age (hypothesis 1b). Our results also highlight that remarried older adults (2) did not display substantially lower levels of wellbeing than those in their first marital unions. More than one marriages were associated with lower assessments of subjective health only among women, but differences in effect sizes were modest. Finally, trajectories characterised by serial cohabitation (4) were associated with lower levels of well-being. These patterns remained after exogenous controls (model 2) and even in the over-controlled model (model 3). (For full models, see Tables4, 5, 6, and 7) Overall, the general pattern was similar for both sexes, although the association between singlehood and life satisfaction appears stronger for men. Robustness checks with alternative well-being measures suggested that the associations seemed to hold more consistently for men (hypothesis 1c). Findings are described in greater detail below and in Fig.10). Finally, we assess whether the association between well-being and partnership is contingent on education. Here, we contrast predictions based on resource substitution and multiplication: we ask whether education amplifies advantages attached to histories with one stable marriage or, conversely, whether stable marital unions mitigate the lower well-being of low-educated individuals. Figures3, 4, and 11 display the predicted values of all education and cluster combinations by sex conditional on the respective controls. The results show that the negative association between the divorce cluster and subjective health was stronger among the basic and secondary educated. Among women, we find similar interaction effect for life satisfaction, albeit not as strong. This means that the gap in well-being between those in their first marital union and those who did not remarry after divorce is larger among low-educated individuals than highly educated individuals. These trends were most prominent in model 4, which controlled for age. The patterns remained throughout models 5 and 6, although not as clearly. We believe that this is mainly due to sample size reduction: it is not so much the smaller difference in point estimates but the wider confidence intervals that seem to dispel the patterns as more controls are introduced. This should not surprise us: as Gelman (2018) has demonstrated, one would need as much as 16 times more data to estimate interaction effects than the main effect. M.Mäki et al. 12 Page 20 of 54 These results point to an educational gradient in line with resource substitution (hypothesis 2b). Those who do not have the resources attached to higher education appear to be more reliant on the resources that years in marriages can provide: trajectories characterised by divorce without remarriage result in pronounced negative well-being associations for those with less education. High education, in turn, did not predict additional well-being for those with stable marital histories. Our results Fig. 3 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 4) Stable Marital Histories Predict Happiness andHealth Across… Page 21 of 54 12 do not accordingly support resource multiplication (hypothesis 2a). This also indicates that stable marriages predict health and quality of life similarly for all educational groups. The findings point out that patterns are more complex than theoretically assumed. In contrast to divorce, the negative associations of singlehood are not more pronounced among those with basic or secondary education. Although the singlehood Fig. 4 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 5) M.Mäki et al. 12 Page 22 of 54 cluster (5) had the lowest well-being and stable marriages (1) the highest, the educational gradient was constant within both clusters. The educational gradient also seems to vary by sex: while serial cohabitation (4) appeared to be disadvantageous for all men, the association varied among women by education. Highly educated women with serial cohabitation fared best out of all groups. Those with basic or secondary education experienced equally negative well-being following their complex partnership histories as those in the permanent singlehood (5) cluster. 3.3 Robustness Checks The results were similar when measured with grip strength and CASP-12 (Figs.10, 12). Grip strength deviated from other response variables: effect sizes were larger for men and smaller for women, for whom the effect sizes were almost zero. The direction of the association was mostly as expected. Still, the near disappearance of association in grip strength implies that the relationship between partnership and well-being, or health, is more robust for men. Grip strength for men in the cluster characterised by singlehood was the strongest negative association that we found in this study: around 0.4 standard deviations lower score than men in the stable first marriage cluster. Those with basic education had even lower scores for both for grip strength and CASP-12. Other than those, we did not find support for interaction effects or resource substitution among those in the single cluster. As subjective health and life satisfaction were measured on 5and 11-point scales, we also estimated each model using ordinal regression. The results were in line with the OLS (results omitted for brevity). Estimating how an unobserved confounder would change the associations further corroborated our findings. The E-values in Fig.13 are mainly around 1.5 and 2.3, meaning that, on the risk ratio scale, an unmeasured confounder associated both with partnership histories and later life outcomes would have to have a 1.5 to 2.0-fold effect in order to explain away the association. The remarriage (2) cluster among men and serial cohabitation (4) cluster among women were exceptions: even a small confounder would have been enough to nullify the findings. The welfare regime-specific analyses tell essentially the same story as the aggregated one, but there are some nuances: highly educated women in the serial cohabitation cluster (4) in the Baltic welfare regime did not display any positive well-being interaction effects. The trend was strongest in Scandinavian countries, but distinguishable in the Bismarckian welfare regime as well. In addition, the wellbeing associations of the remarriage cluster (2) were somewhat inconsistent in the region-specific analyses, which is probably best explained by small effect sizes and decreased sample sizes (see Fig.14 through Fig.19). It is also interesting that the nordic men in the divorce (3) cluster had relatively much lower well-being than men in the other welfare regimes. In order to check that the results were consistent across our study cohorts, we also split the analysis for respondents born between 1945 and 1950 and 1951 and 1957. The results between the two age groups resembled one another. The only notable difference was that the those who belonged to the divorce cluster in the younger Stable Marital Histories Predict Happiness andHealth Across… Page 23 of 54 12 cohorts, had slightly lower well-being compared with those in the marriage cluster within respective cohorts, regardless which measure of well-being we used. Also men in the older cohort who belonged to the single cluster had lower grip strength in comparison with the marriage cluster, than their younger counterparts (results available upon request). When the analyses were run with current relationship status instead of cluster membership, the associations were almost identical. Men in the divorce cluster had around 0.2 standard deviations lower subjective health compared to men in marriage cluster than did men whose current status was divorced compared to those who were currently married. Within the serial cohabitation cluster, those who were dating generally had better well-being than the ones that cohabited (Figs.20, 21). The widowed and divorced had similar well-being. In sum, the results were mainly corroborated with alternative response variables; ordinal regression produced similar results; the results were similar across different welfare regimes; sensitivity analysis revealed that the associations were rather resistant to potential unmeasured confounding, especially in clusters with extensive periods without a partner; stratifying the analyses by cohort groups led to similar outcomes, and exchanging cluster membership for current relationship introduced only nuances to the bigger picture. 4 Discussion andConclusion We investigated, for the first time, how partnership histories are linked with health and quality of life in old age by educational groups. Our findings show that stable marital histories are steadily associated with well-being in old age across all educational groups as well as across different welfare regimes in Northern and Western Europe. Interaction analyses further indicate that lower education together with trajectories characterised by divorce without repartnering predict pronounced adverse well-being in old age, especially among men. Romantic relationships seem to matter for individuals’ well-being throughout the whole life course, which supports the theories of cumulative advantage and disadvantage. The general trend was that the less attached individuals were from one stable marital union, the less happy and healthy they were. The findings also provided a more nuanced view on those in partnerships: individuals with stable second or higher-order marriages did not consistently have lower well-being. This finding differs from previous research (Zahl-Olsen etal., 2019; Booth & Edwards, 1992) that provided a starker difference between trajectories characterised by remarriages and those of first marriages. As opposed to earlier literature (Zimmermann & Hameister, 2019; Jung, 2023), we did not find substantial differences between analyses conducted with relationship history clusters or current relationship status. One the one hand, our results suggest that when relationship history information is not available, current status is a decent proxy for relationship history. The current status often reflects the life course, especially for stable first marriages and lifelong singlehood. On the other hand, we also found some nuances using the trajectory typology approach. Those M.Mäki et al. 12 Page 24 of 54 who were currently dating had similar well-being with those who were currently married. But when looking at the trajectories where dating is commonplace, the well-being is similar to those who are unpartnered. Similarly, presenting trajectories show long-term complexities: the fact that cohabitational relationships tend not to be as long lasting as marital unions serves as a reasonable explanation as to why life courses characterised by cohabitation do not come with similar well-being associations as those characterised by marriage. In addition, men in a cluster characterised by divorce had clearly lower subjective health than men whose current status was divorced, which could also indicate that studying trajectories gives some additional information than just studying the current state. One substantial contribution of the paper was its differentiation between resource multiplication and substitution. Our results suggest that those with fewer resources might have inferior opportunities to compensate for life course events that potentially impede well-being. Conversely, the highly educated could also be less vulnerable to unstable partnership trajectories (Mirowsky & Ross, 2005). Evidence for resource substitution was stronger among men, which could indicate that men are more reliant on spousal support than women in later life (Ajrouch etal., 2005). Our results also pointed to patterns that did not follow from the chosen theoretical framework: highly educated women, whose relationship trajectories were characterised by repeated cohabiting unions, experienced higher levels of well-being than those in first marriage cluster. It is possible that women in our study cohort who moved from one cohabitation spell to another were a distinct group with resources and characteristics different from the general population. Arriving at this kind of trajectory might also have been more of a conscious choice than for others in the cluster. As associations did not disappear after controls, it is possible that the mechanisms between serial cohabitation and well-being could be different for highly educated women, especially in the Scandinavian countries, as demonstrated by our robustness checks. A pseudo-panel has the unfortunate feature of excluding the most vulnerable segments of population. Not only are those who are better off and in relationships more likely to participate, but those who have passed away are excluded by design. As a result, the resource substitution that we observed here are most likely to be stronger in reality (cf. axiom 5 in Ferraro etal., 2009). Although we did measure the effect of an unmeasured confounder, it would have been easier to identify any confounders if the panel was not retrospective. Unmeasured traits could have included, for instance, relationship quality, personal characteristics, contentment with relations with own children and near kin, and mutual friends. A recent study using SHARE data from Finland found that early and stable marital unions had higher relationship quality, suggesting a proximate mechanism for the beneficial associations between partnership history and well-being reported here (Tambellini etal., 2024). We chose to aggregate our analysis for Western and Northern Europe as we were interested in the general associations of union histories as opposed to countryspecific variations. The current literature does not support the hypothesis that the contribution of relationship status (or histories) would have notable cross-national variation (Diener etal., 2000; Verbakel, 2012). Empirically, we found no distinct patterns in the country distributions by the clusters, and the well-being associations Stable Marital Histories Predict Happiness andHealth Across… Page 31 of 54 12 Fig. 10 Linear regression for grip strength and CASP-12 without interaction effects (Models 1,2 & 3). Note: Marriage (c1) is the reference category. Note: Marriage (c1) is the reference category M.Mäki et al. 12 Page 32 of 54 Fig. 11 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 6) Stable Marital Histories Predict Happiness andHealth Across… Page 33 of 54 12 Fig. 12 Predicted values of linear regression for grip strength and CASP-12 with educational interaction effects (Model 4) M.Mäki et al. 12 Page 34 of 54 Fig. 13 E-values to estimate the magnitude that an unmeasured confounder would have to have in order to explain away the association. Note: Marriage (c1) is the reference category Stable Marital Histories Predict Happiness andHealth Across… Page 35 of 54 12 Fig. 14 Linear regression for subjective health and life satisfaction without interaction effects (Models 1,2 & 3)—Bismarckian welfare regime. Note: Marriage (c1) is the reference category M.Mäki et al. 12 Page 36 of 54 Fig. 15 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 4)—Bismarckian welfare regime Stable Marital Histories Predict Happiness andHealth Across… Page 37 of 54 12 Fig. 16 Linear regression for subjective health and life satisfaction without interaction effects (Models 1,2 & 3)—Nordic welfare regime. Note: Marriage (c1) is the reference category M.Mäki et al. 12 Page 38 of 54 Fig. 17 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 4) - Nordic welfare regime Stable Marital Histories Predict Happiness andHealth Across… Page 39 of 54 12 Fig. 18 Linear regression for subjective health and life satisfaction without interaction effects (Models 1,2 & 3)—Baltic welfare regime. Note: Marriage (c1) is the reference category M.Mäki et al. 12 Page 40 of 54 Fig. 19 Predicted values of linear regression for subjective health and life satisfaction with educational interaction effects (Model 4)—Baltic welfare regime Stable Marital Histories Predict Happiness andHealth Across… Page 47 of 54 12 Table 7 Linear regression for life satisfaction (women) ∗∗∗ p < 0.001 ; ∗∗ p < 0.01 ; ∗p < 0.05 Model 1 Model 2 Model 3 (Intercept) 0.15 −0.50∗∗ −0.46∗ (0.18) (0.18) (0.18) 2.Remarriage −0.07∗ −0.08∗ −0.06 (0.03) (0.03) (0.03) 3.Divorce −0.30∗∗∗ −0.30∗∗∗ −0.25∗∗∗ (0.03) (0.03) (0.03) 4.Serial cohabitation −0.10∗ −0.11∗∗ −0.09∗ (0.04) (0.04) (0.04) 5.Single −0.19∗∗∗ −0.21∗∗∗ −0.18∗∗ (0.04) (0.04) (0.05) Age −0.00 0.01∗∗ 0.01∗∗ (0.00) (0.00) (0.00) Experienced hunger as a minor −0.29∗∗∗ −0.27∗∗∗ (0.08) (0.08) Features at childhood home 0.17∗∗∗ 0.10∗∗∗ (0.01) (0.02) Books at childhood home 0.04∗∗ 0.03∗∗ (0.01) (0.01) Rooms at childhood home 0.04 0.01 (0.03) (0.02) Not living with both biological parents when ten −0.07∗ −0.05 (0.03) (0.03) 1 child (ref 2+) −0.11∗∗∗ (0.03) 0 children (ref 2+) −0.01 (0.04) Secondary education (ref Tertiary) −0.03 (0.02) Basic education (ref Tertiary) −0.01 (0.03) Log income 0.21∗∗∗ (0.03) nimp 6 6 6 nobs 10007 10007 10007 R2 0.01 0.06 0.10 Adj. R2 0.01 0.06 0.10 M.Mäki et al. 12 Page 48 of 54 Acknowledgements This research was supported by the Academy of Finland as part of the Research Project LoveAge, decision number 317808, and by the Strategic Research Council (SRC): FLUX and NetResilience consortia, decision numbers: 345130, 345131, 345184, 345183, 364374,364375; by the National Institute on Aging (R01AG075208); by grants to the Max Planck - University of Helsinki Center from the Max Planck Society (5714240218), Jane and Aatos Erkko Foundation (210046), Faculty of Social Sciences at the University of Helsinki(77204227), and Cities of Helsinki, Vantaa and Espoo; and the European Union (ERC Synergy, BIOSFER, 101071773). Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. We wish to thank Andreas Weiland for his guidance with sequence analysis; Anneli Miettinen for sharing her knowledge and understanding about the development of cohabitation in the Nordics and Western Europe; CSC - IT Center for Science, Finland, for computational resources; Jesse Harrison, Kylli Ek and other CSC staff for their assistance with supercomputers; and the Finnish Ministry for Education for funding CSC, and Tiina Helamaa with literature search. Author contributions M. Mäki conducted all inferential data analyses, wrote the first draft of the manuscript and finalised the manuscript after the co-writing process. A.E. Hägglund constructed the life histories and combined most of the variables together, and M. Mäki did some final touches. A.E. Hägglund also edited the manuscript and helped to streamline the storyline. M. Myrskylä provided frequent guidance at every stage of the manuscript preparation and helped to revise the manuscript. A. Rotkirch helped in the theory development, commented, revised parts of the theory sections, secured data collection, and research funding. S. Kulathinal assisted with the statistical analyses and helped writing, especially the methods section. All authors approved the final manuscript. Funding Open access funding provided by University of Helsinki (including Helsinki University Central Hospital). Data availability The data used in this article are available free-of-charge for registered researchers via the SHARE project website: https:// shareeric. eu/ data/ dataaccess. Declarations Conflict of interest The authors have no relevant financial or non-financial interests 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:// creat iveco mmons. org/ licen ses/ by/4. 0/. References Agresti, A. (2015). Foundations of linear and generalized linear models. Hoboken, New Jersey: Wiley. Ajrouch, K. J., Blandon, A. Y., & Antonucci, T. C. (2005). Social networks among men and women: The effects of age and socioeconomic status. The Journals of Gerontology: Series B, 60(6), S311–S317. Averett, S. L., Argys, L. M., & Sorkin, J. (2013). In sickness and in health: An examination of relationship status and health using data from the Canadian National Public Health Survey. Review of Economics of the Household, 11(4), 599–633. Stable Marital Histories Predict Happiness andHealth Across… Page 49 of 54 12 Barlow, A., Duncan, S., James, G. & Park, A. (2001). Just a piece of paper? Marriage and cohabitation. British Social Attitudes Bavel, J. V., & Reher, D. S. (2013). The baby boom and its causes: What we know and what we need to know. Population and Development Review, 39(2), 257–288. Becker, G. S. (1973). A theory of marriage: Part I. Journal of Political Economy, 81(4), 813–846. Becker, G. S. (1991). A treatise on the family (Enlarged). Harvard University Press. Bohannon, R. W. (2019). Grip strength: An indispensable biomarker for older adults. Clinical Interventions in Aging, 14, 1681–1691. Booth, A., & Edwards, J. N. (1992). Starting over: Why remarriages are more unstable. Journal of Family Issues, 13(2), 179–194. Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2021). Introduction to meta-analysis. John Wiley & Sons. Borrat-Besson, C., Ryser, V.-A., & Gonçalves, J. (2015). An evaluation of the CASP-12 scale used in the survey of health, ageing and retirement in Europe (SHARE) to measure quality of life among people aged 50+. (FORS working paper series 2015-4). University of Lausanne. Boyce, C. J., Wood, A. M., & Ferguson, E. (2016). For better or for worse: The moderating effects of personality on the marriage-life satisfaction link. Personality and Individual Differences, 97, 61–66. Brown, S. L., & Wright, M. R. (2017). Marriage, cohabitation, and divorce in later life. Innovation in Aging, 1(2), igx015. Bulanda, J. R., Brown, J. S., & Yamashita, T. (2016). Marital quality, marital dissolution, and mortality risk during the later life course. Social Science & Medicine, 165, 119–127. van Buuren, S., & Groothuis-Oudshoorn, K. (2011). Mice: Multivariate imputation by chained equations in R. Journal of Statistical Software, 45(3), 1–67. Börsch-Supan, A., & Bergmann, M. (2019). SHARE Wave 7 methodology: Panel innovations and life histories. Max Planck Institute for Social Law and Social Policy, MEA. Börsch-Supan, A., Brandt, M., Hunkler, C., Kneip, T., Korbmacher, J., Malter, F., Schaan, B., Stuck, S., & Zuber, S. (2013). Data resource profile: The survey of health, ageing and retirement in Europe (SHARE). International Journal of Epidemiology, 42(4), 992–1001. CSC. (2021). Puhti: An environment for high performance computing. CSC - IT Center for Science Ltd., Technical report. Carr, D., & Utz, R. L. (2020). Families in later life: A decade in review. Journal of Marriage and Family, 82(1), 346–363. Cherlin, A. J. (2004). The deinstitutionalization of American marriage. Journal of Marriage and Family, 66(4), 848–861. Cherlin, A. J. (2017). Introduction to the special collection on separation, divorce, repartnering, and remarriage around the world. Demographic Research, 37, 1275–1296. Cherlin, A. J. (2020). Degrees of change: An assessment of the deinstitutionalization of marriage thesis. Journal of Marriage and Family, 82(1), 62–80. Coombs, R. H. (1991). Marital status and personal well-being: A literature review. Family Relations, 40(1), 97–102. Cornwell, B. (2015). Social sequence analysis. Cambridge University Press. Corti, G., & Scherer, S. (2021). Mating market and dynamics of union formation. European Journal of Population, 37(4–5), 851–876. Dannefer, D. (2003). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social science theory. The Journals of Gerontology: Series B, 58(6), S327–S337. DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advantage as a mechanism for inequality: A review of theoretical and empirical developments. Annual Review of Sociology, 32, 271–297. Diener, E., Gohm, C. L., Suh, E., & Oishi, S. (2000). Similarity of the relations between marital status and subjective well-being across cultures. Journal of Cross-Cultural Psychology, 31(4), 419–436. Diewald, M., & Mayer, K. U. (2009). The sociology of the life course and life span psychology: Integrated paradigm or complementing pathways? Advances in Life Course Research, 14(1), 5–14. Ding, P., & VanderWeele, T. J. (2016). Sensitivity analysis without assumptions. Epidemiology, 27(3), 368–377. Duncan, S., & Phillips, M. (2010). People who live apart together (LATs) - How different are they? The Sociological Review, 58(1), 112–134. M.Mäki et al. 12 Page 50 of 54 Dush, C. M. K., & Amato, P. R. (2005). Consequences of relationship status and quality for subjective well-being. Journal of Social and Personal Relationships, 22(5), 607–627. Eikemo, T. A., Huisman, M., Bambra, C., & Kunst, A. E. (2008). Health inequalities according to educational level in different welfare regimes: A comparison of 23 European countries. Sociology of Health & Illness, 30(4), 565–582. Elder, G. H., Johnson, M. K., & Crosnoe, R. (2003). The emergence and development of life course theory. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of the life course, handbooks of sociology and social research (pp. 3–19). Springer. Ferraro, K., Shippee, T., & Schafer, M. (2009). Cumulative inequality theory for research on aging and the life course. In V. L. Bengtson, M. Silverstein, N. M. Putney, & D. Gans (Eds.), Handbook of theories of aging (pp. 413–433). Springer. Fischer, J. A. (2009). Subjective well-being as welfare measure: Concepts and methodology. OECD, Technical Report MPRA. Fu, R., & Noguchi, H. (2018). Does the positive relationship between health and marriage reflect protection or selection? Evidence from middle-aged and elderly Japanese. Review of Economics of the Household, 16(4), 1003–1016. Gabadinho, A., Ritschard, G., Müller, N. S., & Studer, M. (2011). Analyzing and visualizing state sequences in R with TraMineR. Journal of Statistical Software, 40, 1–37. Gansner, E. R., & North, S. C. (2000). An open graph visualization system and its applications to software engineering. Software - Practice and Experience, 30(11), 1203–1233. Gelman, A. (2018) You need 16 times the sample size to estimate an interaction than to estimate a main effect. Statistical Modeling, Causal Inference, and Social Science Gilleard, C., & Higgs, P. (2007). The third age and the baby boomers. International Journal of Ageing and Later Life, 2(2), 13–30. Goldman, N. (1993). Marriage selection and mortality patterns: Inferences and fallacies. Demography, 30(2), 189–208. Grover, S., & Helliwell, J. F. (2019). How’s life at home? New evidence on marriage and the set point for happiness. Journal of Happiness Studies, 20(2), 373–390. Guma, J., Cámara, A. D., & Treviño, R. (2015). The relationship between health and partnership history in adulthood: Insights through retrospective information from Europeans aged 50 and over. European Journal of Ageing, 12, 71–9. Han, K.-T., Park, E.-C., Kim, J.-H., Kim, S. J., & Park, S. (2014). Is marital status associated with quality of life? Health and Quality of Life Outcomes, 12(1), 109. Hasselblad, V., & Hedges, L. V. (1995). Meta-analysis of screening and diagnostic tests. Psychological Bulletin, 117(1), 167–178. Havari, E., & Mazzonna, F. (2015). Can we trust older people’s statements on their childhood circumstances? Evidence from SHARELIFE. European Journal of Population, 31(3), 233–257. Headey, B. (2010). The set point theory of well-being has serious flaws: On the eve of a scientific revolution? Social Indicators Research, 97(1), 7–21. Hlavac, M. (2018). Stargazer: well-formatted regression and summary statistics tables. Central European Labour Studies Institute (CELSI), Bratislava, Slovakia Holt-Lunstad, J., Birmingham, W., & Jones, B. Q. (2008).Is there something unique about marriage? The relative impact of marital status, relationship quality, and network social support on ambulatory blood pressure and mental health. Annals of Behavioral Medicine, 35(2), 239–244. Horwitz, A. V., White, H. R., & Howell-White, S. (1996). Becoming married and mental health: A longitudinal study of a cohort of young adults. Journal of Marriage and the Family, 58(4), 895. House, J. S., Umberson, D., & Landis, K. R. (1988). Structures and processes of social support. Annual Review of Sociology, 14, 293–318. Hughes, M. E., & Waite, L. J. (2009). Marital biography and health at mid-life. Journal of Health and Social Behavior, 50(3), 344–358. Hägglund, A., Sorsa, T., Danielsbacka, M., Tanskanen, A., & Rotkirch, A. (2021). Vanhenee kuin hyvä viini: yli 50-vuotiaiden parisuhteet Suomessa. Perhebaro metri Jenkinson, C. (2013). Measuring health and medical outcomes. Routledge. Jung, J. (2023). Partnership trajectories and their consequences over the life course. Evidence from the German life study. Advances in Life Course Research, 55, 100525. Kaufman, L. & Rousseeuw, P. J. (1990). Finding groups in data: an introduction to cluster analysis. Wiley series in probability and mathematical statistics, applied probability and statistics. Wiley Stable Marital Histories Predict Happiness andHealth Across… Page 51 of 54 12 Klüsener, S. (2015). Spatial variation in non-marital fertility across Europe in the twentieth and twenty-first centuries: Recent trends, persistence of the past, and potential future pathways. The History of the Family, 20(4), 593–628. Koball, H. L., Moiduddin, E., Henderson, J., Goesling, B., & Besculides, M. (2010). What do we know about the link between marriage and health? Journal of Family Issues, 31(8), 1019–1040. Kontula, O. (2016). Lemmen Paula. Seksuaalinen hyvinvointi parisuhdeonnen avaimena. family barometer 2016, Population Research Institute, Helsinki Kratz, F., & Patzina, A. (2020). Endogenous selection bias and cumulative inequality over the life course: Evidence from educational inequality in subjective well-being. European Sociological Review, 36(3), 333–350. Lenth, R.V. (2021). Emmeans: Estimated marginal means, aka least-squares means. Lesthaeghe, R., & van de Kaa, K. (1986). Twee demografische transities? In Bevolking : Groei en krimp. Van Loghum Slaterus. Lewin, A. C. (2017). Health and relationship quality later in life: A comparison of living apart together (LAT), first marriages, remarriages, and cohabitation. Journal of Family Issues, 38(12), 1754–1774. Lewis, M. A., & Butterfield, R. M. (2007). Social control in marital relationships: Effect of one’s partner on health behaviors. Journal of Applied Social Psychology, 37(2), 298–319. Long, J. A. (2024). jtools: Analysis and presentation of social scientific data. Journal of Open Source Software, 9(101), 6610. Ludwig, V., & Brüderl, J. (2018). Is there a male marital wage premium? New evidence from the united states. American Sociological Review, 83(4), 744–770. Luyckx, K., Seiffge-Krenke, I., Schwartz, S. J., Crocetti, E., & Klimstra, T. A. (2014). Identity configurations across love and work in emerging adults in romantic relationships. Journal of Applied Developmental Psychology, 35(3), 192–203. Manzoli, L., Villari, P., M Pirone, G., & Boccia, A. (2007). Marital status and mortality in the elderly: A systematic review and meta-analysis. Social Science & Medicine, 64(1), 77–94. Margolis, R., & Myrskylä, M. (2011). A global perspective on happiness and fertility. Population and Development Review, 37(1), 29–56. Mastekaasa, A. (1992). Marriage and psychological well-being: Some evidence on selection into marriage. Journal of Marriage and the Family, 54(4), 901. Meara, E. R., Richards, S., & Cutler, D. M. (2008). The gap gets bigger: Changes in mortality and life expectancy, by education, 1981–2000. Health Affairs, 27(2), 350–360. Melo, S., Guedes, J., & Mendes, S. (2019). Theory of cumulative disadvantage/advantage. In D. Gu & M. E. Dupre (Eds.), Encyclopedia of gerontology and population aging (pp. 1–8). Springer. Metsä-Simola, N., & Martikainen, P. (2013). The short-term and long-term effects of divorce on mortality risk in a large Finnish cohort, 1990–2003. Population Studies, 67(1), 97–110. Mirowsky, J., & Ross, C. E. (2005). Education, cumulative advantage, and health. Ageing International, 30(1), 27–62. Mortelmans, D. (2020). Economic consequences of divorce: A review. In Parental life courses after separation and divorce in Europe, (pp. 23–41). Springer. Næss, E. O., Mehlum, L., & Qin, P. (2021). Marital status and suicide risk: Temporal effect of marital breakdown and contextual difference by socioeconomic status. SSM - Population Health, 15, 100853. O’Rand, A. M. (2001). Stratification and the life course: The forms of life-course capital and their interrelationships. In Handbook of aging and the social sciences (pp. 197–213). Academic Press. O’Flaherty, M., Baxter, J., Haynes, M., & Turrell, G. (2016). The family life course and health: Partnership, fertility histories, and later-life physical health trajectories in australia. Demography, Silver Spring, 53(3), 777–804. Perelli-Harris, B., & Amos, M. (2015). Changes in partnership patterns across the life course: An examination of 14 countries in Europe and the United States. Demographic Research, 33(6), 145–178. Peters, A., & Liefbroer, A. C. (1997). Beyond marital status: Partner history and well-being in old age. Journal of Marriage and Family, 59(3), 687–699. Pinquart, M. (2003). Loneliness in married, widowed, divorced, and never-married older adults. Journal of Social and Personal Relationships, 20(1), 31–53. Prus, S. G. (2007). Age, SES, and health: A population level analysis of health inequalities over the lifecourse. Sociology of Health & Illness, 29(2), 275–296. M.Mäki et al. 12 Page 52 of 54 R Core Team. (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Rault, W. (2019). Is the civil solidarity pact the future of marriage? The several meanings of the French civil union. International Journal of Law, Policy and the Family, 33(2), 139–159. Raz-Yurovich, L. (2013). Divorce penalty or divorce premium? A longitudinal analysis of the consequences of divorce for men’s and women’s economic activity. European Sociological Review, 29(2), 373–385. Reczek, C., Pudrovska, T., Carr, D., Thomeer, M. B., & Umberson, D. (2016). Marital histories and heavy alcohol use among older adults. Journal of Health and Social Behavior, 57(1), 77–96. Reneflot, A., & Mamelund, S.-E. (2012). The association between marital status and psychological wellbeing in Norway. European Sociological Review, 28, 355–365. Rich, B. (2021). Table1: Tables of descriptive statistics in HTML. Rindfuss, R. R., & Vandenheuvel, A. (2022). Cohabitation: A precursor to marriage or an alternative to being single? Population and Development Review, 16(4), 25. Roberson, P. N. E., Norona, J. C., Lenger, K. A., & Olmstead, S. B. (2018). How do relationship stability and quality affect wellbeing?: Romantic relationship trajectories, depressive symptoms, and life satisfaction across 30 years. Journal of Child and Family Studies, 27(7), 2171–2184. Roberts, S. G. B., & Dunbar, R. I. M. (2011). Communication in social networks: Effects of kinship, network size, and emotional closeness. Personal Relationships, 18(3), 439–452. Rosenfeld, M. J. (2018). Who wants the breakup? Gender and breakup in heterosexual couples. In D. F. Alwin, D. H. Felmlee, & D. A. Kreager (Eds.), Social networks and the life course: Integrating the development of human lives and social relational networks (pp. 221–243). Springer. Ross, C. E. (1995). Reconceptualizing marital status as a continuum of social attachment. Journal of Marriage and the Family, 57(1), 129. Ross, C. E., & Mirowsky, J. (2006). Sex differences in the effect of education on depression: Resource multiplication or resource substitution? Social Science & Medicine, 63(5), 1400–1413. Ross, C. E., & Mirowsky, J. (2010). Gender and the health benefits of education. The Sociological Quarterly, 51(1), 1–19. Ross, C. E., & Wu, C.-L. (1996). Education, age, and the cumulative advantage in health. Journal of Health and Social Behavior, 37(1), 104. Rotkirch, A. (2020). What are couples made of? Union formation in high-income societies. In O. Burger, R. Lee, & R. Sear (Eds.), Human evolutionary demography. Open Book Publishers. Régnier-Loilier, A. (2016). Partnership trajectories of people in stable non-cohabiting relationships in France. Demographic Research, 35, 1169–1212. Sassler, S. (2010). Partnering across the life course: Sex, relationships, and mate selection. Journal of Marriage and Family, 72(3), 557–575. Schartz, C. R. (2010). Pathways to educational homogamy in marital and cohabiting unions. Demography, 47(3), 735–753. Schütz, J. (2019). Marital biography and health in old age: Insights from European survey data Scott, E. S. (2000). Social norms and the legal regulation of marriage. Virginia Law Review. https:// doi. org/ 10. 2307/ 10738 33 Stryker, S. (1959). Symbolic interaction as an approach to family research. Marriage and Family Living, 21(2), 111. Studer, M. (2013). WeightedCluster library manual: A practical guide to creating typologies of trajectories in the social sciences with R. Technical report, LIVES Working Papers 24. Tambellini, E., Danielsbacka, M., & Rotkirch, A. (2024). Both partnership history and current relationship quality are associated with life satisfaction in old age. Research on Aging. https:// doi. org/ 10. 1177/ 01640 27524 13092 55 Tavakol, Z., Nikbakht Nasrabadi, A., Behboodi Moghadam, Z., Salehiniya, H., & Rezaei, E. (2017). A review of the factors associated with marital satisfaction. Galen Medical Journal, 6(3), e641. UNSD. (1999). United nations geoscheme for europe. https:// unsta ts. un. org/ unsd/ metho dology/ m49/. Umberson, D. (1992). Gender, marital status and the social control of health behavior. Social Science & Medicine, 34(8), 907–917. VanderWeele, T. J. (2017). On a square-root transformation of the odds ratio for a common outcome. Epidemiology, 28(6), e58–e60. VanderWeele, T. J., & Ding, P. (2017). Sensitivity analysis in observational research: Introducing the e-value. Annals of Internal Medicine, 167(4), 268–274. Stable Marital Histories Predict Happiness andHealth Across… Page 53 of 54 12 Verbakel, E. (2012). Subjective well-being by partnership status and its dependence on the normative climate. European Journal of Population / Revue Européenne de Démographie, 28(2), 205–232. Waite, L. J. (1995). Does marriage matter? Demography, 32(4), 483–507. Wang, D., Li, D., Mishra, S. R., Lim, C., Dai, X., Chen, S., & Xu, X. (2022). Association between marital relationship and multimorbidity in middle-aged adults: A longitudinal study across the US, UK, Europe, and China. Maturitas, 155, 32–39. Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58(301), 236–244. Westmaas, J. L., Wild, T. C., & Ferrence, R. (2002). Effects of gender in social control of smoking cessation. Health Psychology, 21(4), 368. Wickham, H. (2016). Ggplot2: Elegant graphics for data analysis. Springer-Verlag. Wong, J. S., & Waite, L. J. (2015). Marriage, social networks, and health at older ages. Journal of Population Ageing, 8(1–2), 7–25. Wood, R. G., Avellar, S., & Goesling, B. (2009). The effects of marriage on health: A synthesis of recent research evidence. Nova Science Woszidlo, A., & Segrin, C. (2013). Negative affectivity and educational attainment as predictors of newlyweds’ problem solving communication and marital quality. The Journal of Psychology, 147(1), 49–73. Wängqvist, M., Carlsson, J., van der Lee, M., & Frisén, A. (2016). Identity development and romantic relationships in the late twenties. Identity, 16(1), 24–44. Zahl-Olsen, R., Thuen, F., & Espehaug, B. (2019). Divorce and remarriage in norway: A prospective cohort study between 1981 and 2013. Journal of Divorce & Remarriage, 60(8), 600–611. Zajacova, A., & Lawrence, E. M. (2018). The relationship between education and health: Reducing disparities through a contextual approach. Annual Review of Public Health, 39, 273–289. Zimmermann, O., & Hameister, N. (2019). Stable cohabitational unions increase quality of life: Retrospective analysis of partnership histories also reveals gender differences. Demographic Research, 40, 657–692. Øien-Ødegaard, C., Hauge, L. J., & Reneflot, A. (2021). Marital status, educational attainment, and suicide risk: A Norwegian register-based population study. Population Health Metrics, 19(1), 33. M.Mäki et al. 12 Page 54 of 54 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Authors and Affiliations MiikaMäki1,2,6,7 · AnnaErikaHägglund1,3· AnnaRotkirch1 · SangitaKulathinal4· MikkoMyrskylä5,6,7 * Miika Mäki [email protected] Anna Erika Hägglund [email protected] Anna Rotkirch [email protected] Sangita Kulathinal sangita.k[email protected] Mikko Myrskylä [email protected] 1 Population Research Institute attheFamily Federation ofFinland, Helsinki, Finland 2 Center forSocial Data Science, University ofHelsinki, Helsinki, Finland 3 University ofTurku, Turku, Finland 4 Department ofMathematics andStatistics, University ofHelsinki, Helsinki, Finland 5 Max Planck Institute forDemographic Research, Rostock, Germany 6 Max Planck – University ofHelsinki Center forSocial Inequalities inPopulation Health, Helsinki, Finland 7 Helsinki Institute forDemography andPopulation Health, University ofHelsinki, Helsinki, Finland