Mäki, Miika; Hägglund, Anna E ika; Ro ki ch, Anna; Kula hinal, Sangi a; My skylä,
Mikko
A icle — Published Ve sion
S able Ma i al His o ies P edic Happiness and Heal h
Ac oss Educa ional G oups
Eu opean Jou nal o Popula ion
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Mäki, Miika; Hägglund, Anna E ika; Ro ki ch, Anna; Kula hinal, Sangi a; My skylä,
Mikko (2025) : S able Ma i al His o ies P edic Happiness and Heal h Ac oss Educa ional G oups,
Eu opean Jou nal o Popula ion, ISSN 1572-9885, Sp inge Ne he lands, Do d ech , Vol. 41, Iss. 1,
h ps://doi.o g/10.1007/s10680-025-09733-x
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/330779
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
Eu opean Jou nal o Popula ion (2025) 41:12
h ps://doi.o g/10.1007/s10680-025-09733-x
ORIGINAL RESEARCH
S able Ma i al His o ies P edic Happiness andHeal h
Ac oss Educa ional G oups
MiikaMäki1,2,6,7 · AnnaE ikaHägglund1,3· AnnaRo ki ch1 ·
Sangi aKula hinal4· MikkoMy skylä5,6,7
Recei ed: 15 Decembe 2023 / Accep ed: 23 Feb ua y 2025 / Published online: 13 May 2025
© The Au ho (s) 2025, co ec ed publica ion 2025
Abs ac
Couple ela ions a e a key de e minan o men al and physical well-being in old age.
Howe e , we do no know how he ad an ages and disad an ages associa ed wi h
pa ne ship his o ies a y be ween socioeconomic g oups. We c ea e ela ionship
his o y ypologies o he coho s 1945-1957 using he Su ey o Heal h, Ageing,
and Re i emen in Eu ope and examine, o he i s ime, how ela ionship his o-
ies ela e o mul iple indica o s o well-being by educa ional a ainmen . The esul s
show ha s able ma iages p edic g ea e well-being, compa ed o single and less
s able pa ne ship his o ies. The posi i e ou comes a e simila ac oss all educa-
ional g oups. Those wi h lowe educa ion who ha e di o ced expe ience e en lowe
well-being in old age. The in e ac ion analyses sugges ha indi iduals wi h ewe
esou ces could su e mo e om losing a pa ne . The indings unde sco e ha cu -
en and pas oman ic ela ionships a e linked o well-being in old age and help
policymake s iden i y ulne able subg oups among he ageing popula ion.
Keywo ds Pa ne ship his o y· Resou ce subs i u ion· Cumula i e disad an age·
Heal h· Quali y o li e
1 In oduc ion
Roman ic couple ela ions a e among he mos in ima e, las ing and impo an
ela ionships in ou li es and a co ne s one o emo ional, social, and economic
well-being (Wängq is e al., 2016; Luyckx e al., 2014). P e ious esea ch has
exhaus i ely shown ha ha ing a pa ne is posi i ely associa ed wi h a mul i ude
o well-being ou comes in la e li e, such as li e sa is ac ion, quali y o li e, heal h,
mo bidi y, and mo ali y (Wong & Wai e, 2015; Han e al., 2014; Hol -Luns ad
e al., 2008; Coombs, 1991; Manzoli e al., 2007; Wai e, 1995; Wang e al., 2022;
Rene lo & Mamelund, 2012). Indeed, ha ing a spouse is a guably one o he single
mos impo an ac o s o heal hy ageing (Wood e al., 2009). A he same ime,
Ex ended au ho in o ma ion a ailable on he las page o he a icle
M.Mäki e al.
12 Page 2 o 54
he p opo ions o single, di o ced, o ema ied elde ly ha e inc eased (Che lin,
2017), and ela ionship biog aphies ha e become mo e di e se. I is c ucial o iden-
i y wha ypes o pa ne ship his o ies a e ela ed o inc eased ulne abili y, and o
wha ex en lacking o losing a pa ne can be compensa ed o .
Ha ing a pa ne as such may no imp o e well-being. Ins ead, he associa ion
be ween pa ne ships and well-being could eme ge om he pa ne ship his o y as
he social, economic and heal h ad an ages o pa ne ships accumula e o e ime (c .
DiP e e & Ei ich, 2006). Exis ing li e a u e, howe e , ends o examine ela ionships
ei he as single e en s, such as ma iage o di o ce, o depic hem a he s a ically
h ough he cu en pa ne ship s a us (Sassle , 2010; Schü z, 2019; Gumà e al.,
2014). So a , a ew s udies ha e concep ualised pa ne ships as ajec o ies (Zim-
me mann & Hameis e , 2019; Robe son e al., 2018; Jung, 2023; O’ lahe y e al.,
2016). These sugges ha aspec s o pa ne his o ies p edic well-being beyond he
cu en s a us, and ha u he insigh s in o well-being can be gained by o ming a-
jec o ies (Pe e s & Lie b oe , 1997; Jung, 2023; Zimme mann & Hameis e , 2019).
A i s con ibu ion o his pape is ha we concep ualise ela ionships as ajec o ies
and see whe he hey a e linked wi h well-being.
Se e al schola s, including Becke (1973), ha e s essed ha he gains om ma -
iage a e likely o depend on indi idual ai s, such as a ac i eness, in elligence,
and educa ion. These di e en gains may a ise because o homogamy, highe ma i al
sa is ac ion, and mo e ul illing amily li e and ies. So a , schola s ha e no consid-
e ed educa ion in g ea e de ail. Ins ead, educa ional a ainmen has been ea ed as
a con ol a he han a s a i ying ea u e (Boyce e al., 2016; Han e al., 2014; Wong
& Wai e, 2015). To ou knowledge, only one s udy in ecen decades has examined
whe he he associa ion be ween ma i al s a us and well-being indica o s a ies by
educa ion: Øien-Ødegaa d e al. (2021) ound ha low educa ion inc eased he sui-
cide isk among sepa a ed.
The e a e se e al easons o assume ha he gains om a s able pa ne ship his-
o y ollow an educa ional g adien . On he one hand, highly educa ed indi iduals
migh p o i mo e om ha ing a li elong pa ne han hose wi h a lowe educa ion
due o lowe con lic (Woszidlo & Seg in, 2013), mo e ul illing amily li e (Rob-
e s & Dunba 2011; Ma golis & My skylä 2011) and he possible compounding
e ec o di e en ypes o esou ces ha bo h educa ion and a s able ela ionship
p o ide (Zajaco a & Law ence, 2018; Wood e al., 2009). This p ocess would imply
esou ce mul iplica ion. On he o he hand, hose wi h a lowe educa ional le el may
ely mo e hea ily on a spouse, as s able pa ne ships compensa e o a lack o o he
esou ces. Pu di e en ly, spousal loss migh ha e mo e de imen al consequences
o hose wi h ewe esou ces (DiP e e & Ei ich, 2006). The la e p ocess, in u n,
sugges s esou ce subs i u ion.
In his s udy, we analyse wo such dimensions ha ha e no been s udied join ly
be o e: whe he pa icula ela ionship his o ies p edic well-being and whe he he
well-being associa ions o hose his o ies a y by educa ion. We in es iga e whe he
oman ic ela ionship his o ies a e associa ed wi h well-being o men and women
and how he associa ions di e be ween educa ional g oups o No he n and Wes e n
Eu opeans in hi d age, a s age in one’s li e be ween middle age and old age (Gillea d
& Higgs, 2007). Speci ically, we a e in e es ed in signs o esou ce mul iplica ion o
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 3 o 54 12
subs i u ion: is he combina ion o ha ing a ce ain educa ional le el and ela ion-
ship ajec o y linked wi h e en highe o lowe well-being ou comes?
Empi ically, we c ea e pa ne ship his o y ypologies wi h sequence analysis
using e ospec i e li e his o y in e iew da a om he Su ey o Heal h, Ageing,
and Re i emen in Eu ope (SHARE) om 13 Eu opean coun ies. We measu e well-
being wi h subjec i e heal h and li e sa is ac ion and conduc obus ness checks wi h
g ip s eng h and CASP-12 quali y-o -li e indica o . Ou analyses con ol o se e al
con ounding measu es including childhood condi ions, income, and numbe o chil-
d en. The esul s a e u he co obo a ed by a con ounde analysis.
1.1 Why a e Rela ionship S a us andWell‑Being Linked?
Empi ical e idence o he in e connec edness o ma i al s a us and well-being is
conclusi e. Cu en ma i al s a us and he quali y o couple ela ionships a e gene -
ally linked o well-being (see e iew in Wood e al., 2009). These associa ions seem
o p e ail in di e en socie ies: Diene e al. (2000) s udied he ela ionship be ween
ma i al s a us and subjec i e well-being in 42 coun ies and ound ha di e ences
in e ec sizes a e negligible. Ve bakel (2012) concludes wi h a simila s udy design
ha "no ma i e clima e appea s o ha dly a ec well-being gaps be ween pa ne -
ship s a uses". Thus, we expec pa ne ships o suppo well-being simila ly ac oss
ou s udy popula ion. Bu wha explains he consis ency o hese indings? P e ious
esea ch poin s o h ee mechanisms, namely selec ion in o ma iage, ma iage as
a key social ole, and he e e yday social suppo o li ing wi h a spouse (Dush &
Ama o, 2005; A e e e al., 2013).
Fi s , ma iage i sel migh no make people happy and heal hy. Ins ead, happy
and heal hy people could sel -selec in o ma iage (Koball e al., 2010). Al hough
some selec ion e ec s a e a play (see Goldman, 1993, o a heo e ical o e iew),
he majo i y o longi udinal s udies sugges ha selec ion alone does no accoun o
all he posi i e e ec s o ela ionships on well-being (Ho wi z e al., 1996; Wood
e al., 2009; Fu & Noguchi, 2018; G o e & Helliwell, 2019). Ye no all ag ee
(Mas ekaasa, 1992; Ludwig & B üde l, 2018), highligh ing ha esul s may a y
wi h me hodological design.
Second, he s uc u al symbolic in e ac ionis pe spec i e a gues ha oles wi h a
high le el o commi men esul in high le els o iden i y and sel -wo h (S yke &
S a ham 1985, as ci ed in Dush & Ama o 2005). In e nalising he ole o a wi e o
husband (o pa en ) shapes bo h he beha iou and sel -pe cep ion, and consequen ly
enhances well-being, al hough his depends on hei commi men o he gi en iden-
i ies (S yke , 1959). The ac ha ew wan o li e wi hou a long- e m oman ic
pa ne e en in weal hy, indi idualised, and libe al socie ies (Kon ula 2016; Ro -
ki ch 2020, pp. 40-41) demons a es he high alue a ached o couple ela ions.
Thi d, ma iage ends o p o ide social suppo and egula ion beyond any o he
o m o ela ionship (Ross, 1995; Coombs, 1991; Sco , 2000). In high-income
socie ies, unions a e he p ima y household uni o b ead-winning, consump ion,
and in imacy. Pa icula ly in he ageing popula ion, he ypical al e na i e o li ing
wi h a oman ic pa ne is o li e alone (Rind uss & Vandenheu el, 2022; Becke ,
M.Mäki e al.
12 Page 4 o 54
1991; Ro ki ch, 2020). Indi iduals who a e a ached o social ne wo ks a e heal hie
and li e longe (House e al., 1988); hose wi h a spouse ha e someone in he same
household o sha e he joys and so ows o li e in he long e m, which single adul s
o en lack.
1.2 Rela ionships andWell‑Being omaLi e Cou se Pe spec i e
Well-being in old age e lec s combined expe iences o e he en i e li e cou se, no
jus he p esen momen . S udying sequences o ela ionship s a u es, o ela ionship
his o ies, could e eal dynamics ha a e concealed in analyses ocusing on ela ion-
ship s a uses. The ocus on sequences aligns wi h li e cou se heo y: Indi idual ou -
comes a e he esul o p e ious ajec o ies and e en s. These e ol e wi h ime, a e
embedded in social s uc u es, and o m pa e ns o social s a i ica ion (Elde e al.,
2003; Diewald & Maye , 2009).
Empi ical esea ch poin s o he alue o analysing ela ionship ajec o ies. Fo
ins ance, ma iage dissolu ion has long- e m, nega i e implica ions on well-being
and heal h, which pe sis e en among hose who ema y (Hughes & Wai e, 2009).
Co espondingly, uns able pa ne ships, mul iple ela ionship ansi ions, and long-
e m singlehood a e associa ed wi h highe le els o dep ession and s ess and lowe
social and emo ional suppo (Jung, 2023; Zimme mann & Hameis e , 2019). Pa -
ne ship ajec o ies cha ac e ised by s able unions ypically display he highes le -
els o psychological and physical well-being (O’ lahe y e al., 2016; Tambellini
e al., 2024). The associa ion be ween pa ne ship his o ies and well-being is likely
o inc ease wi h age, since inequali ies in heal h, esou ces, and subjec i e well-
being end o accumula e as indi iduals g ow olde (K a z & Pa zina, 2020; Ross &
Wu, 1996; P us, 2007; Headey, 2010). Finding a new pa ne , howe e , can a leas
pa ly compensa e o mo e uns able pa ne ship ajec o ies (Jung, 2023).
O e all, his p ocess aligns wi h he amewo k o cumula i e ad an age/dis-
ad an age, sugges ing ha nega i e o posi i e implica ions in one aspec o li e
inc ease o e ime (Danne e , 2003). As ela ionship s a uses a e associa ed wi h
li es yles, heal h beha iou , social suppo , and ne wo ks, well-being—o lack
he eo —is likely o spill o e o o he domains and accumula e as indi iduals age.
Melo e al. (2019) goes as a as claiming ha he ageing p ocess as such is a esul
o accumula ed dis/ad an ages o e he yea s om di e en sphe es o li e.
In his s udy, we use cumula i e dis/ad an age o e e o he compound well-
being e ec s o ela ionship ajec o ies o e ime. Addi ionally, we hypo hesise how
educa ional a ainmen could po en ially ampli y o mi iga e cumula i e dis/ad an-
age (see Sec . 1.4).
1.3 Gende Di e ences
Empi ical esea ch highligh s ha he implica ions o oman ic ela ionships di e
by sex. In pa icula , men seem o bene i signi ican ly om being in a pa ne ship
(Coombs, 1991; Me sä-Simola & Ma ikainen, 2013); long- e m singlehood sub-
s an ially lowe s men’s well-being (Zimme mann & Hameis e , 2019; Jung, 2023).
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 5 o 54 12
Fo women, by con as , he associa ion be ween pa ne ships and well-being
appea s weake (O’ lahe y e al., 2016; Zimme mann & Hameis e , 2019). A i s
glance, his seems coun e in ui i e, as women ace g ea e economic disad an ages
ollowing union dissolu ion han men (see Mo elmans, 2020, o a e iew). Fo
ins ance, women’s sala y g ow h a es decline mo e sha ply, lea ing hem a a long-
e m economic disad an age (Raz-Yu o ich, 2013). Addi ionally, women a e mo e
likely o ake on p ima y childca e esponsibili ies a e a b eakup, which com-
pounds hei inancial s ain.
Despi e hese economic disad an ages, se e al ac o s could explain why ela-
ionship b eakups end o ha e a g ea e de imen al e ec on men’s well-being, pa -
icula ly in la e li e. Women a e mo e likely o engage in beha iou s ha p omo e
heal h and o en in luence hei pa ne s o do he same (Reczek e al., 2016; Umbe -
son, 1992; Wes maas e al., 2002; Lewis & Bu e ield, 2007). Men, on he o he
hand, end o ha e smalle and less emo ionally suppo i e social ne wo ks in la e
li e (Aj ouch e al., 2005). While single mo he s bea a hea ie inancial bu den a e
union dissolu ion, hey o en main ain s onge bonds wi h hei child en and ecei e
mo e ca e in old age. Fa he s who do no li e wi h hei child en a e a b eakup may
ins ead pay a highe social p ice. (Mo elmans, 2020)
Fu he mo e, women o en expe ience g ea e psychological dis ess om s aying
in ma iages wi h low ela ionship quali y (Bulanda e al., 2016; B own & W igh ,
2017; Ca & U z, 2020). This may make hem mo e inclined han men o end an
unsa is ac o y ela ionship and be e equipped o cope wi h li e wi hou a signi ican
o he (Rosen eld, 2018).
1.4 Educa ional Resou ce Mul iplica ion o Subs i u ion
S able unions a e likely o inc ease well-being, while uns able his o ies and long-
e m singlehood lowe i . Howe e , he ou comes o ela ionship his o ies migh ol-
low an educa ional g adien . Educa ion ancho s he accumula ion o human, social,
and pe sonal capi al (O’Rand, 2001), and is posi i ely associa ed wi h a mul i ude o
well-being ou comes (see Zajaco a & Law ence, 2018, o e iew). Theo e ically, i
emains unclea whe he high- o low-educa ed bene i mo e om s able unions. In
he ollowing, we a gue ha wo mu ually exclusi e pa e ns a e possible: esou ce
mul iplica ion and esou ce subs i u ion.
On he one hand, high educa ion migh s eng hen he posi i e associa ion o s a-
ble ma i al his o ies on well-being. S able ela ionships allow o he accumula ion
o esou ces, and highly educa ed ha e mo e esou ces o sus ain long- e m ela-
ionships o begin wi h: hey a e likely o pa ne wi h simila ly posi ioned indi-
iduals (Scha z, 2010), ecei e mo e suppo om he kin ne wo k (Robe s &
Dunba , 2011), and a e be e equipped o deal wi h con lic (Woszidlo & Seg in,
2013). These ac o s con ibu e o longe and happie unions among highly edu-
ca ed (Ta akol e al., 2017). S able unions, in u n, p o ide a pla o m o esou ce
accumula ion.
Highly educa ed men and women could also p o i mo e om ma i al s abili y.
Fo example, combined inancial esou ces (Jung, 2023), spousal emo ional suppo ,
M.Mäki e al.
12 Page 6 o 54
ul illing ela ionships wi h kin-ne wo k (Robe s & Dunba , 2011), o he joys o
aising child en could be mo e p onounced among highly educa ed indi iduals
(Ma golis & My skylä, 2011). Likewise, since bo h educa ion and s able ma iages
p edic heal h and li e sa is ac ion (Zajaco a & Law ence, 2018; Wood e al., 2009),
hei in e play could esul in mul iplica i e well-being ou comes. Thus, we assume
esou ce mul iplica ion i high educa ion coupled wi h s able pa ne ship his o ies
mu ually ein o ce well-being (Ross & Mi owsky, 2006).
Al e na i ely, low-educa ed indi iduals migh p o i mo e om s able pa ne ship
ajec o ies han highly educa ed. As low-educa ed indi iduals gene ally ha e ewe
esou ces (Ross & Mi owsky, 2010), hei well-being migh ely mo e on s able
pa ne ship his o ies. The loss o spousal inancial and emo ional suppo is likely
o in luence well-being mo e s ongly in his g oup han among he highly educa ed,
who can bu e he nega i e consequences o uns able ajec o ies and li ing alone
wi h o he social, economic, and psychological esou ces (Fe a o e al., 2009). This
pa e n would sugges esou ce subs i u ion (Mi owsky& Ross, 2005): s able union
his o ies would ha e s onge bene i s, o con e sely, single o agmen ed union
his o ies s onge disad an ages among low-educa ed indi iduals, as hey ha e ewe
esou ces o esilience.
Theo e ically, esou ce subs i u ion and mul iplica ion should no co-exis : ei he
he bene i s o a s able pa ne ship his o y a e s onge o hose wi h highe edu-
ca ion ( esou ce mul iplica ion), he bene i s o a s able pa ne ship his o y a e
s onge o hose wi h lowe educa ion ( esou ce subs i u ion), o he e a e no edu-
ca ional in e ac ion e ec s (c . Ross & Mi owsky, 2006).
1.5 Hypo heses
Based on p e ious esea ch, we o mula e a se o hypo heses. Fi s , ollowing
cumula i e ad an age and disad an age, we assume ha s able ma i al ela ionship
his o ies a e associa ed wi h highe well-being a e age 60 (1a), while his o ies
cha ac e ised by ins abili y o long- e m singlehood a e associa ed wi h lowe well-
being (1b). In addi ion, we expec (1c) mo e obus and consis en esul s o men,
bu possibly also some associa ions o women.
We an icipa e ha he link be ween pa ne ship his o ies and well-being could
a y by educa ion, bu he di ec ion is no clea . Gi en he mu ual exclusi i y o
esou ce subs i u ion and mul iplica ion, and he lack o conclusi e indings in p e-
ious esea ch, we emain agnos ic as o which, i ei he , will be suppo ed by ou
s udy. I he educa ional g adien is in line wi h esou ce mul iplica ion, we expec
high educa ion o p edic e en s onge posi i e associa ions be ween s able pa -
ne ship his o ies and well-being (2a). Educa ion would hen ampli y he cumula i e
ad an age o ha ing a s able ma i al his o y as he highly educa ed p o i mo e om
amily li e. I esou ce subs i u ion occu s, we expec ha low educa ion p edic s
s onge nega i e well-being ou comes o li elong single and uns able pa ne ship
his o ies (2b). In such case, he low educa ed could nei he u ilise educa ional no
spousal esou ces, compa ed o hose wi h high educa ion—who can bu e uns able
pa ne ship ajec o ies wi h esou ces a ached o educa ion.
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 7 o 54 12
2 Da a andMe hods
Li e his o ies we e gene a ed wi h he help o he SHARELIFE in e iews o he
Su ey o Heal h, Ageing, and Re i emen in Eu ope (Bö sch-Supan & Be gmann,
2019).1 SHARE is a mic o-panel da a in as uc u e ha co e s households wi h a
leas one membe o e 50 yea s o age in all EU coun ies, Swi ze land, and Is ael.
The su ey collec s bo h panel and e ospec i e li e cou se da a. The e ospec i e
SHARELIFE pseudo-panel in e iews we e conduc ed in 2008 and 2017 (wa e 3
and 7, espec i ely). Responden s we e asked o epo , amongs o he s, on hei
childhood ci cums ances as well as hei pas pa ne s, including all cohabi a ional,
ma i al, and da ing ela ionships. I he li e his o y in e iews we e conduc ed in
2008, any changes in ela ionship s a uses we e upda ed wi h subsequen panel
in e iew da a.
Ou analyses a e based on No he n and Wes e n Eu ope as de ined by Uni ed
Na ions (UNSD, 1999). The sample was es ic ed o hese geog aphical a eas
o ensu e ha ou s udy popula ion would come om oughly simila cul u es in
e ms o amily o ma ion and amily li e pa e ns (Klüsene , 2015). As a obus -
ness check, we examined whe he pa e ns a ied be ween wel a e s a es and cul-
u es wi hin No he n and Wes e n Eu ope by s a i ying he in e en ial analyses o
wel a e egimes ollowing he classi ica ion scheme in Eikemo e al. (2008): No dic
(Sweden, Denma k, Finland), Bal ic (Es onia, La ia, Li huania), and Bisma ckian
(Aus ia, Belgium, F ance, Ge many, Luxembou g, Ne he lands, Swi ze land) wel-
a e egimes.
We selec ed esponden s bo n be ween 1945 and 1957 o wo easons: hey we e
a leas 60 yea s old in 2017 when he da a we e collec ed and we e all pa o he
baby boome gene a ion (Ba el & Rehe , 2013). This gene a ion o med unions a a
ime when he p e alence o ma iage was a i s peak. Ye , his is also a gene a ion
when p ema i al cohabi a ion and non-ma i al childbea ing s a ed o become mo e
common, s a ing om No he n Eu ope. (Pe elli-Ha is & Amos, 2015; Klüsene ,
2015). The inal sample size o analysis was 18,256 indi iduals. In o de o ob ain
an adequa e sample size, all couples we e he e osexual.
He e, we ca ego ise pa ne ship s a uses in o unma ied, i s and highe -o de
ma iages, da ing, cohabi a ion, di o ce, and widowhood. This akes di e ences
be ween pa ne ed and unpa ne ed indi iduals mo e exhaus i ely in o conside a ion
(Næss e al., 2021; Pinqua , 2003): Al hough cohabi a ion and ma iage inc eas-
ingly esemble each o he bo h legally and socially (Che lin, 2020; Pe elli-Ha is
& Amos, 2015), he e is s ill a clea di e ence in e ms o commi men , longe i y,
and symbolic impo ance (Ba low e al., 2001; Che lin, 2004; Raul , 2019; B own
& W igh , 2017). Simila ly, i s - and highe -o de ma iages di e in e ms o s a-
bili y, ela ionship sa is ac ion, and demog aphic makeup (Zahl-Olsen e al., 2019;
1 This pape uses da a om SHARE Wa es 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 Bö sch-Supan e al. (2013) o me hodological
de ails.
M.Mäki e al.
12 Page 8 o 54
Hägglund e al., 2021; Boo h & Edwa ds, 1992). Thus, la e li e ou comes could
di e as well. We also ha e in o ma ion on da ing ela ionships. In SHARELIFE,
da ing is de ined as a oman ic ela ionship in which couples do no li e a he same
add ess mos o he ime. Da ing he e includes hose who ha e a oman ic ela ion-
ship be o e mo ing in oge he , as well as hose who pe manen ly a e in a li ing
apa oge he (LAT) ela ionship (Lewin, 2017). Such ela ionships end o be mo e
lexible and less commi ed han ma iage o cohabi a ion (Régnie -Loilie , 2016;
Duncan & Phillips, 2010). In ou ope a ionalisa ion, an indi idual is da ing i hey
en e a oman ic ela ionship wi hou co- esidence, i espec i e o p e ious s a us.
In a i s s ep, we c ea e ela ionship his o y ypologies. By pa ne ship his o y,
we mean he o de ed se ies o ela ionship s a uses om adolescence o old age
(c . Co nwell, 2015, p. 21).2 The pa ne ship ajec o ies we e c ea ed by sequence
analysis o agglome a i e hie a chical clus e ing (see Co nwell, 2015, o a heo-
e ical o e iew). We i s c ea ed a dis ance ma ix wi h he dynamic hamming
me hod. The subs i u ion cos s o wo s a es a posi ion p a e calcula ed as ollows:
whe e
sp(A,B)
is he subs i u ion cos a posi ion3 p and
Xp
is he s a e4 a he
p h
posi ion. The me hod assesses bo h he p obabili y o being in s a e A and s a e B
a posi ion p as well as he ansi ion p obabili ies om A o B a posi ion p and
ice e sa. This means ha bo h iming and o de a e aken in o accoun . (Co nwell,
2015, p.128) Thus he subs i u ion cos s a e calcula ed au oma ically based on how
common a ce ain ansi ion is a a gi en age. Fo example, a ansi ion om ma -
iage o widowhood will be gi en a much highe cos a he age o 20 han 60. We
chose o use dynamic hamming me hod due o he ac ha he subs i u ion cos s a e
de i ed empi ically om he da a. We s ill ound e y simila clus e s wi h dynamic
hamming and di e en a ian s o op imal ma ching.
The dis ance ma ix be ween he indi idual ajec o ies was u he analysed
using Wa d’s agglome a i e clus e ing me hod o o m clus e s (Wa d, 1963). The
clus e s a e ob ained by minimising he wi hin-clus e sum o squa es, and hence
p oduce g oups wi h simila his o ies. We also expe imen ed wi h o he algo i hms
bu he clus e ing al e na i es wi h Wa d’s me hod made he mos sense concep u-
ally. I is calcula ed as ollows:
(1)
s
p(A,B)=4−
(
p (Xp=A
|
Xp−1=B)+p (Xp=B
|
Xp−1=A
)
+p (X
p+1
=A
|
X
p
=B)+p (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 he e ms pa ne ship his o ies and pa ne ship ajec o ies in e changeably h oughou his
a icle.
3 Ages 15 o 60.
4 Singlehood, i s ma iages, highe -o de ma iages, da ing, cohabi a ion, di o ce, o widowhood.
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 15 o 54 12
e u ned o singlehood in a ew yea s, o had se e al sho ela ionships (see Fig.9
o indi idual li e cou ses).
O e all, he clus e s esemble each o he in demog aphic cha ac e is ics
(see Table 2), al hough he e a e a ew excep ions: men we e o e - ep esen ed
in he singlehood (5) and women in he di o ce (3) clus e . While educa ional
di e ences we e gene ally insubs an ial, men wi h s able ela ionship ajec-
o ies ended o be mo e educa ed. Fo women, he opposi e pa e n p e ailed
(see Fig.7). The dis ibu ion o ela ionship clus e s did no a y much ac oss
Fig. 1 Pa ne ship his o y clus e s o Wes e n and No he n Eu opeans bo n in 1945–1957 a he age o
15–60
M.Mäki e al.
12 Page 16 o 54
Table 2 Desc ip i e s a is ics by clus e
1.Ma iage 2.Rema iage 3.Di o ce 4.Se ial cohabi a ion 5.Single O e all
(N=12135) (N=1960) (N=1720) (N=1189) (N=1206) (N=18210)
Yea o bi h
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)
Gende
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 %)
Educa ion
Highe 4768 (39.3 %) 756 (38.6 %) 664 (38.6 %) 461 (38.8 %) 475 (39.4 %) 7124 (39.1 %)
Seconda y 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%)
Coun y
Aus ia 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 %)
Denma k 1084 (8.9 %) 235 (12.0 %) 113 (6.6 %) 154 (13.0 %) 86 (7.1 %) 1672 (9.2 %)
Es onia 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 %)
F ance 1277 (10.5 %) 156 (8.0 %) 186 (10.8 %) 117 (9.8 %) 144 (11.9 %) 1880 (10.3 %)
Ge many 1363 (11.2 %) 257 (13.1 %) 121 (7.0 %) 61 (5.1 %) 105 (8.7 %) 1907 (10.5 %)
La ia 405 (3.3 %) 84 (4.3 %) 90 (5.2 %) 31 (2.6 %) 31 (2.6 %) 641 (3.5 %)
Li huania 533 (4.4 %) 68 (3.5 %) 92 (5.3 %) 28 (2.4 %) 30 (2.5 %) 751 (4.1 %)
Luxembou g 479 (3.9 %) 48 (2.4 %) 46 (2.7 %) 12 (1.0 %) 37 (3.1 %) 622 (3.4 %)
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 17 o 54 12
Table 2 (con inued)
1.Ma iage 2.Rema iage 3.Di o ce 4.Se ial cohabi a ion 5.Single O e all
(N=12135) (N=1960) (N=1720) (N=1189) (N=1206) (N=18210)
Ne he lands 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 %)
Swi ze land 821 (6.8 %) 135 (6.9 %) 137 (8.0 %) 72 (6.1 %) 121 (10.0 %) 1286 (7.1 %)
M.Mäki e al.
12 Page 18 o 54
coun ies. One s able ma i al ajec o ies we e by a he mos common, and sin-
gle o se ial cohabi a ion ajec o ies we e he a es . Fu he inspec ion by mul-
inomial eg ession e ealed ha No he n Eu opean esponden s we e mo e
likely o belong o he se ial cohabi a ion clus e . No o he dis inc pa e ns we e
obse ed (see Table3 and Fig.8).
Fig. 2 Linea eg ession o subjec i e heal h and li e sa is ac ion wi hou in e ac ion e ec s (Models 1,2
& 3). No e: Ma iage (c1) is he e e ence ca ego y
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 19 o 54 12
3.2 Well‑being Ou comes
Nex , we analyse he associa ion be ween pa ne ship his o ies and wo measu e-
men s o well-being, namely subjec i e heal h and li e sa is ac ion. The esul s o
main e ec s (models 1-3) a e depic ed in Fig.2, whe e he ma iage (1) clus e is
ou e e ence ca ego y. As all non-bina y a iables we e s anda dised, an e ec size
o one would imply a di e ence o one s anda d de ia ion in he esponse a iable
compa ed wi h he e e ence ca ego y. The whi e ci cles and squa es ep esen poin
es ima es, while he hick and na ow ba s he 95% and 99% con idence in e als,
espec i ely.
O e all, he esul s a e in line wi h ou hypo heses in ha ajec o ies cha ac-
e ised by s able ma i al unions (clus e 1) we e associa ed wi h highe subjec i e
heal h and highe li e sa is ac ion o men and women (hypo hesis 1a). Adul s whose
ajec o ies we e cha ac e ised by singlehood (5) and di o ce (3), in u n, expe i-
enced he lowes well-being in old age (hypo hesis 1b). Ou esul s also highligh
ha ema ied olde adul s (2) did no display subs an ially lowe le els o well-
being han hose in hei i s ma i al unions. Mo e han one ma iages we e associ-
a ed wi h lowe assessmen s o subjec i e heal h only among women, bu di e ences
in e ec sizes we e modes . Finally, ajec o ies cha ac e ised by se ial cohabi a ion
(4) we e associa ed wi h lowe le els o well-being. These pa e ns emained a e
exogenous con ols (model 2) and e en in he o e -con olled model (model 3). (Fo
ull models, see Tables4, 5, 6, and 7)
O e all, he gene al pa e n was simila o bo h sexes, al hough he associa-
ion be ween singlehood and li e sa is ac ion appea s s onge o men. Robus ness
checks wi h al e na i e well-being measu es sugges ed ha he associa ions seemed
o hold mo e consis en ly o men (hypo hesis 1c). Findings a e desc ibed in g ea e
de ail below and in Fig.10).
Finally, we assess whe he he associa ion be ween well-being and pa ne ship is
con ingen on educa ion. He e, we con as p edic ions based on esou ce subs i u-
ion and mul iplica ion: we ask whe he educa ion ampli ies ad an ages a ached o
his o ies wi h one s able ma iage o , con e sely, whe he s able ma i al unions mi i-
ga e he lowe well-being o low-educa ed indi iduals. Figu es3, 4, and 11 display
he p edic ed alues o all educa ion and clus e combina ions by sex condi ional on
he espec i e con ols.
The esul s show ha he nega i e associa ion be ween he di o ce clus e and
subjec i e heal h was s onge among he basic and seconda y educa ed. Among
women, we ind simila in e ac ion e ec o li e sa is ac ion, albei no as s ong.
This means ha he gap in well-being be ween hose in hei i s ma i al union and
hose who did no ema y a e di o ce is la ge among low-educa ed indi idu-
als han highly educa ed indi iduals. These ends we e mos p ominen in model
4, which con olled o age. The pa e ns emained h oughou models 5 and 6,
al hough no as clea ly. We belie e ha his is mainly due o sample size educ-
ion: i is no so much he smalle di e ence in poin es ima es bu he wide con-
idence in e als ha seem o dispel he pa e ns as mo e con ols a e in oduced.
This should no su p ise us: as Gelman (2018) has demons a ed, one would need as
much as 16 imes mo e da a o es ima e in e ac ion e ec s han he main e ec .
M.Mäki e al.
12 Page 20 o 54
These esul s poin o an educa ional g adien in line wi h esou ce subs i u ion
(hypo hesis 2b). Those who do no ha e he esou ces a ached o highe educa ion
appea o be mo e elian on he esou ces ha yea s in ma iages can p o ide: a-
jec o ies cha ac e ised by di o ce wi hou ema iage esul in p onounced nega i e
well-being associa ions o hose wi h less educa ion. High educa ion, in u n, did
no p edic addi ional well-being o hose wi h s able ma i al his o ies. Ou esul s
Fig. 3 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 4)
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 21 o 54 12
do no acco dingly suppo esou ce mul iplica ion (hypo hesis 2a). This also indi-
ca es ha s able ma iages p edic heal h and quali y o li e simila ly o all educa-
ional g oups.
The indings poin ou ha pa e ns a e mo e complex han heo e ically assumed.
In con as o di o ce, he nega i e associa ions o singlehood a e no mo e p o-
nounced among hose wi h basic o seconda y educa ion. Al hough he singlehood
Fig. 4 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 5)
M.Mäki e al.
12 Page 22 o 54
clus e (5) had he lowes well-being and s able ma iages (1) he highes , he edu-
ca ional g adien was cons an wi hin bo h clus e s. The educa ional g adien also
seems o a y by sex: while se ial cohabi a ion (4) appea ed o be disad an ageous
o all men, he associa ion a ied among women by educa ion. Highly educa ed
women wi h se ial cohabi a ion a ed bes ou o all g oups. Those wi h basic o sec-
onda y educa ion expe ienced equally nega i e well-being ollowing hei complex
pa ne ship his o ies as hose in he pe manen singlehood (5) clus e .
3.3 Robus ness Checks
The esul s we e simila when measu ed wi h g ip s eng h and CASP-12 (Figs.10,
12). G ip s eng h de ia ed om o he esponse a iables: e ec sizes we e la ge
o men and smalle o women, o whom he e ec sizes we e almos ze o. The
di ec ion o he associa ion was mos ly as expec ed. S ill, he nea disappea ance o
associa ion in g ip s eng h implies ha he ela ionship be ween pa ne ship and
well-being, o heal h, is mo e obus o men. G ip s eng h o men in he clus e
cha ac e ised by singlehood was he s onges nega i e associa ion ha we ound in
his s udy: a ound 0.4 s anda d de ia ions lowe sco e han men in he s able i s
ma iage clus e . Those wi h basic educa ion had e en lowe sco es o bo h o g ip
s eng h and CASP-12. O he han hose, we did no ind suppo o in e ac ion
e ec s o esou ce subs i u ion among hose in he single clus e .
As subjec i e heal h and li e sa is ac ion we e measu ed on 5- and 11-poin
scales, we also es ima ed each model using o dinal eg ession. The esul s we e in
line wi h he OLS ( esul s omi ed o b e i y).
Es ima ing how an unobse ed con ounde would change he associa ions u -
he co obo a ed ou indings. The E- alues in Fig.13 a e mainly a ound 1.5 and
2.3, meaning ha , on he isk a io scale, an unmeasu ed con ounde associa ed
bo h wi h pa ne ship his o ies and la e li e ou comes would ha e o ha e a 1.5 o
2.0- old e ec in o de o explain away he associa ion. The ema iage (2) clus e
among men and se ial cohabi a ion (4) clus e among women we e excep ions: e en
a small con ounde would ha e been enough o nulli y he indings.
The wel a e egime-speci ic analyses ell essen ially he same s o y as he
agg ega ed one, bu he e a e some nuances: highly educa ed women in he se ial
cohabi a ion clus e (4) in he Bal ic wel a e egime did no display any posi i e
well-being in e ac ion e ec s. The end was s onges in Scandina ian coun ies,
bu dis inguishable in he Bisma ckian wel a e egime as well. In addi ion, he well-
being associa ions o he ema iage clus e (2) we e somewha inconsis en in he
egion-speci ic analyses, which is p obably bes explained by small e ec sizes and
dec eased sample sizes (see Fig.14 h ough Fig.19). I is also in e es ing ha he
no dic men in he di o ce (3) clus e had ela i ely much lowe well-being han men
in he o he wel a e egimes.
In o de o check ha he esul s we e consis en ac oss ou s udy coho s, we also
spli he analysis o esponden s bo n be ween 1945 and 1950 and 1951 and 1957.
The esul s be ween he wo age g oups esembled one ano he . The only no able
di e ence was ha he hose who belonged o he di o ce clus e in he younge
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 23 o 54 12
coho s, had sligh ly lowe well-being compa ed wi h hose in he ma iage clus e
wi hin espec i e coho s, ega dless which measu e o well-being we used. Also
men in he olde coho who belonged o he single clus e had lowe g ip s eng h in
compa ison wi h he ma iage clus e , han hei younge coun e pa s ( esul s a ail-
able upon eques ).
When he analyses we e un wi h cu en ela ionship s a us ins ead o clus e
membe ship, he associa ions we e almos iden ical. Men in he di o ce clus e had
a ound 0.2 s anda d de ia ions lowe subjec i e heal h compa ed o men in ma -
iage clus e han did men whose cu en s a us was di o ced compa ed o hose who
we e cu en ly ma ied. Wi hin he se ial cohabi a ion clus e , hose who we e da -
ing gene ally had be e well-being han he ones ha cohabi ed (Figs.20, 21). The
widowed and di o ced had simila well-being.
In sum, he esul s we e mainly co obo a ed wi h al e na i e esponse a iables;
o dinal eg ession p oduced simila esul s; he esul s we e simila ac oss di e -
en wel a e egimes; sensi i i y analysis e ealed ha he associa ions we e a he
esis an o po en ial unmeasu ed con ounding, especially in clus e s wi h ex ensi e
pe iods wi hou a pa ne ; s a i ying he analyses by coho g oups led o simila
ou comes, and exchanging clus e membe ship o cu en ela ionship in oduced
only nuances o he bigge pic u e.
4 Discussion andConclusion
We in es iga ed, o he i s ime, how pa ne ship his o ies a e linked wi h heal h
and quali y o li e in old age by educa ional g oups. Ou indings show ha s able
ma i al his o ies a e s eadily associa ed wi h well-being in old age ac oss all educa-
ional g oups as well as ac oss di e en wel a e egimes in No he n and Wes e n
Eu ope. In e ac ion analyses u he indica e ha lowe educa ion oge he wi h a-
jec o ies cha ac e ised by di o ce wi hou epa ne ing p edic p onounced ad e se
well-being in old age, especially among men.
Roman ic ela ionships seem o ma e o indi iduals’ well-being h oughou he
whole li e cou se, which suppo s he heo ies o cumula i e ad an age and disad-
an age. The gene al end was ha he less a ached indi iduals we e om one s a-
ble ma i al union, he less happy and heal hy hey we e. The indings also p o ided
a mo e nuanced iew on hose in pa ne ships: indi iduals wi h s able second o
highe -o de ma iages did no consis en ly ha e lowe well-being. This inding di -
e s om p e ious esea ch (Zahl-Olsen e al., 2019; Boo h & Edwa ds, 1992) ha
p o ided a s a ke di e ence be ween ajec o ies cha ac e ised by ema iages and
hose o i s ma iages.
As opposed o ea lie li e a u e (Zimme mann & Hameis e , 2019; Jung, 2023),
we did no ind subs an ial di e ences be ween analyses conduc ed wi h ela ion-
ship his o y clus e s o cu en ela ionship s a us. One he one hand, ou esul s
sugges ha when ela ionship his o y in o ma ion is no a ailable, cu en s a us
is a decen p oxy o ela ionship his o y. The cu en s a us o en e lec s he li e
cou se, especially o s able i s ma iages and li elong singlehood. On he o he
hand, we also ound some nuances using he ajec o y ypology app oach. Those
M.Mäki e al.
12 Page 24 o 54
who we e cu en ly da ing had simila well-being wi h hose who we e cu en ly
ma ied. Bu when looking a he ajec o ies whe e da ing is commonplace, he
well-being is simila o hose who a e unpa ne ed. Simila ly, p esen ing ajec o ies
show long- e m complexi ies: he ac ha cohabi a ional ela ionships end no o
be as long las ing as ma i al unions se es as a easonable explana ion as o why li e
cou ses cha ac e ised by cohabi a ion do no come wi h simila well-being associa-
ions as hose cha ac e ised by ma iage. In addi ion, men in a clus e cha ac e ised
by di o ce had clea ly lowe subjec i e heal h han men whose cu en s a us was
di o ced, which could also indica e ha s udying ajec o ies gi es some addi ional
in o ma ion han jus s udying he cu en s a e.
One subs an ial con ibu ion o he pape was i s di e en ia ion be ween esou ce
mul iplica ion and subs i u ion. Ou esul s sugges ha hose wi h ewe esou ces
migh ha e in e io oppo uni ies o compensa e o li e cou se e en s ha po en-
ially impede well-being. Con e sely, he highly educa ed could also be less ul-
ne able o uns able pa ne ship ajec o ies (Mi owsky & Ross, 2005). E idence o
esou ce subs i u ion was s onge among men, which could indica e ha men a e
mo e elian on spousal suppo han women in la e li e (Aj ouch e al., 2005).
Ou esul s also poin ed o pa e ns ha did no ollow om he chosen heo e ical
amewo k: highly educa ed women, whose ela ionship ajec o ies we e cha ac-
e ised by epea ed cohabi ing unions, expe ienced highe le els o well-being han
hose in i s ma iage clus e . I is possible ha women in ou s udy coho who
mo ed om one cohabi a ion spell o ano he we e a dis inc g oup wi h esou ces
and cha ac e is ics di e en om he gene al popula ion. A i ing a his kind o
ajec o y migh also ha e been mo e o a conscious choice han o o he s in he
clus e . As associa ions did no disappea a e con ols, i is possible ha he mech-
anisms be ween se ial cohabi a ion and well-being could be di e en o highly
educa ed women, especially in he Scandina ian coun ies, as demons a ed by ou
obus ness checks.
A pseudo-panel has he un o una e ea u e o excluding he mos ulne able seg-
men s o popula ion. No only a e hose who a e be e o and in ela ionships mo e
likely o pa icipa e, bu hose who ha e passed away a e excluded by design. As a
esul , he esou ce subs i u ion ha we obse ed he e a e mos likely o be s onge
in eali y (c . axiom 5 in Fe a o e al., 2009). Al hough we did measu e he e ec o
an unmeasu ed con ounde , i would ha e been easie o iden i y any con ounde s i
he panel was no e ospec i e. Unmeasu ed ai s could ha e included, o ins ance,
ela ionship quali y, pe sonal cha ac e is ics, con en men wi h ela ions wi h own
child en and nea kin, and mu ual iends. A ecen s udy using SHARE da a om
Finland ound ha ea ly and s able ma i al unions had highe ela ionship quali y,
sugges ing a p oxima e mechanism o he bene icial associa ions be ween pa ne -
ship his o y and well-being epo ed he e (Tambellini e al., 2024).
We chose o agg ega e ou analysis o Wes e n and No he n Eu ope as we
we e in e es ed in he gene al associa ions o union his o ies as opposed o coun y-
speci ic a ia ions. The cu en li e a u e does no suppo he hypo hesis ha he
con ibu ion o ela ionship s a us (o his o ies) would ha e no able c oss-na ional
a ia ion (Diene e al., 2000; Ve bakel, 2012). Empi ically, we ound no dis inc
pa e ns in he coun y dis ibu ions by he clus e s, and he well-being associa ions
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 31 o 54 12
Fig. 10 Linea eg ession o g ip s eng h and CASP-12 wi hou in e ac ion e ec s (Models 1,2 & 3).
No e: Ma iage (c1) is he e e ence ca ego y. No e: Ma iage (c1) is he e e ence ca ego y
M.Mäki e al.
12 Page 32 o 54
Fig. 11 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 6)
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 33 o 54 12
Fig. 12 P edic ed alues o linea eg ession o g ip s eng h and CASP-12 wi h educa ional in e ac ion
e ec s (Model 4)
M.Mäki e al.
12 Page 34 o 54
Fig. 13 E- alues o es ima e he magni ude ha an unmeasu ed con ounde would ha e o ha e in o de
o explain away he associa ion. No e: Ma iage (c1) is he e e ence ca ego y
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 35 o 54 12
Fig. 14 Linea eg ession o subjec i e heal h and li e sa is ac ion wi hou in e ac ion e ec s (Models
1,2 & 3)—Bisma ckian wel a e egime. No e: Ma iage (c1) is he e e ence ca ego y
M.Mäki e al.
12 Page 36 o 54
Fig. 15 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 4)—Bisma ckian wel a e egime
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 37 o 54 12
Fig. 16 Linea eg ession o subjec i e heal h and li e sa is ac ion wi hou in e ac ion e ec s (Models
1,2 & 3)—No dic wel a e egime. No e: Ma iage (c1) is he e e ence ca ego y
M.Mäki e al.
12 Page 38 o 54
Fig. 17 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 4) - No dic wel a e egime
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 39 o 54 12
Fig. 18 Linea eg ession o subjec i e heal h and li e sa is ac ion wi hou in e ac ion e ec s (Models
1,2 & 3)—Bal ic wel a e egime. No e: Ma iage (c1) is he e e ence ca ego y
M.Mäki e al.
12 Page 40 o 54
Fig. 19 P edic ed alues o linea eg ession o subjec i e heal h and li e sa is ac ion wi h educa ional
in e ac ion e ec s (Model 4)—Bal ic wel a e egime
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 47 o 54 12
Table 7 Linea eg ession o
li e sa is ac ion (women)
∗∗∗ p
<
0.001
;
∗∗ p
<
0.01
;
∗p
<
0.05
Model 1 Model 2 Model 3
(In e cep ) 0.15
−0.50∗∗
−0.46∗
(0.18) (0.18) (0.18)
2.Rema iage
−0.07∗
−0.08∗
−0.06
(0.03) (0.03) (0.03)
3.Di o ce
−0.30∗∗∗
−0.30∗∗∗
−0.25∗∗∗
(0.03) (0.03) (0.03)
4.Se ial cohabi a ion
−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)
Expe ienced hunge as a mino
−0.29∗∗∗
−0.27∗∗∗
(0.08) (0.08)
Fea u es a childhood home
0.17∗∗∗
0.10∗∗∗
(0.01) (0.02)
Books a childhood home
0.04∗∗
0.03∗∗
(0.01) (0.01)
Rooms a childhood home 0.04 0.01
(0.03) (0.02)
No li ing wi h bo h biological
pa en s when en
−0.07∗
−0.05
(0.03) (0.03)
1 child ( e 2+)
−0.11∗∗∗
(0.03)
0 child en ( e 2+)
−0.01
(0.04)
Seconda y educa ion ( e Te ia y)
−0.03
(0.02)
Basic educa ion ( e Te ia y)
−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 e al.
12 Page 48 o 54
Acknowledgemen s This esea ch was suppo ed by he Academy o Finland as pa o he Resea ch
P ojec Lo eAge, decision numbe 317808, and by he S a egic Resea ch Council (SRC): FLUX and
Ne Resilience conso ia, decision numbe s: 345130, 345131, 345184, 345183, 364374,364375; by he
Na ional Ins i u e on Aging (R01AG075208); by g an s o he Max Planck - Uni e si y o Helsinki Cen e
om he Max Planck Socie y (5714240218), Jane and Aa os E kko Founda ion (210046), Facul y o
Social Sciences a he Uni e si y o Helsinki(77204227), and Ci ies o Helsinki, Van aa and Espoo; and
he Eu opean Union (ERC Syne gy, BIOSFER, 101071773). Views and opinions exp essed a e, howe e ,
hose o he au ho s only and do no necessa ily e lec hose o he Eu opean Union o he Eu opean
Resea ch Council. Nei he he Eu opean Union no he g an ing au ho i y can be held esponsible o
hem.
We wish o hank And eas Weiland o his guidance wi h sequence analysis; Anneli Mie inen o
sha ing he knowledge and unde s anding abou he de elopmen o cohabi a ion in he No dics and
Wes e n Eu ope; CSC - IT Cen e o Science, Finland, o compu a ional esou ces; Jesse Ha ison, Kylli
Ek and o he CSC s a o hei assis ance wi h supe compu e s; and he Finnish Minis y o Educa ion
o unding CSC, and Tiina Helamaa wi h li e a u e sea ch.
Au ho con ibu ions M. Mäki conduc ed all in e en ial da a analyses, w o e he i s d a o he manu-
sc ip and inalised he manusc ip a e he co-w i ing p ocess. A.E. Hägglund cons uc ed he li e his-
o ies and combined mos o he a iables oge he , and M. Mäki did some inal ouches. A.E. Hägglund
also edi ed he manusc ip and helped o s eamline he s o yline. M. My skylä p o ided equen guid-
ance a e e y s age o he manusc ip p epa a ion and helped o e ise he manusc ip . A. Ro ki ch helped
in he heo y de elopmen , commen ed, e ised pa s o he heo y sec ions, secu ed da a collec ion, and
esea ch unding. S. Kula hinal assis ed wi h he s a is ical analyses and helped w i ing, especially he
me hods sec ion. All au ho s app o ed he inal manusc ip .
Funding Open access unding p o ided by Uni e si y o Helsinki (including Helsinki Uni e si y Cen al
Hospi al).
Da a a ailabili y The da a used in his a icle a e a ailable ee-o -cha ge o egis e ed esea che s ia
he SHARE p ojec websi e: h ps:// sha e- e ic. eu/ da a/ da a- access.
Decla a ions
Con lic o in e es The au ho s ha e no ele an inancial o non- inancial in e es s o disclose.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License,
which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as
you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Com-
mons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his a icle
a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line o he
ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is
no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission
di ec ly om he copy igh holde . To iew a copy o his licence, isi h p:// c ea i eco mmons. o g/ licen
ses/ by/4. 0/.
Re e ences
Ag es i, A. (2015). Founda ions o linea and gene alized linea models. Hoboken, New Je sey:
Wiley.
Aj ouch, K. J., Blandon, A. Y., & An onucci, T. C. (2005). Social ne wo ks among men and women: The
e ec s o age and socioeconomic s a us. The Jou nals o Ge on ology: Se ies B, 60(6), S311–S317.
A e e , S. L., A gys, L. M., & So kin, J. (2013). In sickness and in heal h: An examina ion o ela ion-
ship s a us and heal h using da a om he Canadian Na ional Public Heal h Su ey. Re iew o Eco-
nomics o he Household, 11(4), 599–633.
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 49 o 54 12
Ba low, A., Duncan, S., James, G. & Pa k, A. (2001). Jus a piece o pape ? Ma iage and cohabi a ion.
B i ish Social A i udes
Ba el, J. V., & Rehe , D. S. (2013). The baby boom and i s causes: Wha we know and wha we need o
know. Popula ion and De elopmen Re iew, 39(2), 257–288.
Becke , G. S. (1973). A heo y o ma iage: Pa I. Jou nal o Poli ical Economy, 81(4), 813–846.
Becke , G. S. (1991). A ea ise on he amily (Enla ged). Ha a d Uni e si y P ess.
Bohannon, R. W. (2019). G ip s eng h: An indispensable bioma ke o olde adul s. Clinical In e en-
ions in Aging, 14, 1681–1691.
Boo h, A., & Edwa ds, J. N. (1992). S a ing o e : Why ema iages a e mo e uns able. Jou nal o Family
Issues, 13(2), 179–194.
Bo ens ein, M., Hedges, L. V., Higgins, J. P., & Ro hs ein, H. R. (2021). In oduc ion o me a-analysis.
John Wiley & Sons.
Bo a -Besson, C., Ryse , V.-A., & Gonçal es, J. (2015). An e alua ion o he CASP-12 scale used in he
su ey o heal h, ageing and e i emen in Eu ope (SHARE) o measu e quali y o li e among peo-
ple aged 50+. (FORS wo king pape se ies 2015-4). Uni e si y o Lausanne.
Boyce, C. J., Wood, A. M., & Fe guson, E. (2016). Fo be e o o wo se: The mode a ing e ec s o pe -
sonali y on he ma iage-li e sa is ac ion link. Pe sonali y and Indi idual Di e ences, 97, 61–66.
B own, S. L., & W igh , M. R. (2017). Ma iage, cohabi a ion, and di o ce in la e li e. Inno a ion in
Aging, 1(2), igx015.
Bulanda, J. R., B own, J. S., & Yamashi a, T. (2016). Ma i al quali y, ma i al dissolu ion, and mo ali y
isk du ing he la e li e cou se. Social Science & Medicine, 165, 119–127.
an Buu en, S., & G oo huis-Oudshoo n, K. (2011). Mice: Mul i a ia e impu a ion by chained equa ions
in R. Jou nal o S a is ical So wa e, 45(3), 1–67.
Bö sch-Supan, A., & Be gmann, M. (2019). SHARE Wa e 7 me hodology: Panel inno a ions and li e
his o ies. Max Planck Ins i u e o Social Law and Social Policy, MEA.
Bö sch-Supan, A., B and , M., Hunkle , C., Kneip, T., Ko bmache , J., Mal e , F., Schaan, B., S uck, S.,
& Zube , S. (2013). Da a esou ce p o ile: The su ey o heal h, ageing and e i emen in Eu ope
(SHARE). In e na ional Jou nal o Epidemiology, 42(4), 992–1001.
CSC. (2021). Puh i: An en i onmen o high pe o mance compu ing. CSC - IT Cen e o Science L d.,
Technical epo .
Ca , D., & U z, R. L. (2020). Families in la e li e: A decade in e iew. Jou nal o Ma iage and Family,
82(1), 346–363.
Che lin, A. J. (2004). The deins i u ionaliza ion o Ame ican ma iage. Jou nal o Ma iage and Family,
66(4), 848–861.
Che lin, A. J. (2017). In oduc ion o he special collec ion on sepa a ion, di o ce, epa ne ing, and
ema iage a ound he wo ld. Demog aphic Resea ch, 37, 1275–1296.
Che lin, A. J. (2020). Deg ees o change: An assessmen o he deins i u ionaliza ion o ma iage hesis.
Jou nal o Ma iage and Family, 82(1), 62–80.
Coombs, R. H. (1991). Ma i al s a us and pe sonal well-being: A li e a u e e iew. Family Rela ions,
40(1), 97–102.
Co nwell, B. (2015). Social sequence analysis. Camb idge Uni e si y P ess.
Co i, G., & Sche e , S. (2021). Ma ing ma ke and dynamics o union o ma ion. Eu opean Jou nal o
Popula ion, 37(4–5), 851–876.
Danne e , D. (2003). Cumula i e ad an age/disad an age and he li e cou se: C oss- e ilizing age and
social science heo y. The Jou nals o Ge on ology: Se ies B, 58(6), S327–S337.
DiP e e, T. A., & Ei ich, G. M. (2006). Cumula i e ad an age as a mechanism o inequali y: A e iew o
heo e ical and empi ical de elopmen s. Annual Re iew o Sociology, 32, 271–297.
Diene , E., Gohm, C. L., Suh, E., & Oishi, S. (2000). Simila i y o he ela ions be ween ma i al s a us
and subjec i e well-being ac oss cul u es. Jou nal o C oss-Cul u al Psychology, 31(4), 419–436.
Diewald, M., & Maye , K. U. (2009). The sociology o he li e cou se and li e span psychology: In e-
g a ed pa adigm o complemen ing pa hways? Ad ances in Li e Cou se Resea ch, 14(1), 5–14.
Ding, P., & Vande Weele, T. J. (2016). Sensi i i y analysis wi hou assump ions. Epidemiology, 27(3),
368–377.
Duncan, S., & Phillips, M. (2010). People who li e apa oge he (LATs) - How di e en a e hey? The
Sociological Re iew, 58(1), 112–134.
M.Mäki e al.
12 Page 50 o 54
Dush, C. M. K., & Ama o, P. R. (2005). Consequences o ela ionship s a us and quali y o subjec i e
well-being. Jou nal o Social and Pe sonal Rela ionships, 22(5), 607–627.
Eikemo, T. A., Huisman, M., Bamb a, C., & Kuns , A. E. (2008). Heal h inequali ies acco ding o edu-
ca ional le el in di e en wel a e egimes: A compa ison o 23 Eu opean coun ies. Sociology o
Heal h & Illness, 30(4), 565–582.
Elde , G. H., Johnson, M. K., & C osnoe, R. (2003). The eme gence and de elopmen o li e cou se
heo y. In J. T. Mo ime & M. J. Shanahan (Eds.), Handbook o he li e cou se, handbooks o soci-
ology and social esea ch (pp. 3–19). Sp inge .
Fe a o, K., Shippee, T., & Scha e , M. (2009). Cumula i e inequali y heo y o esea ch on aging and
he li e cou se. In V. L. Beng son, M. Sil e s ein, N. M. Pu ney, & D. Gans (Eds.), Handbook o
heo ies o aging (pp. 413–433). Sp inge .
Fische , J. A. (2009). Subjec i e well-being as wel a e measu e: Concep s and me hodology. OECD,
Technical Repo MPRA.
Fu, R., & Noguchi, H. (2018). Does he posi i e ela ionship be ween heal h and ma iage e lec p o ec-
ion o selec ion? E idence om middle-aged and elde ly Japanese. Re iew o Economics o he
Household, 16(4), 1003–1016.
Gabadinho, A., Ri scha d, G., Mülle , N. S., & S ude , M. (2011). Analyzing and isualizing s a e
sequences in R wi h T aMineR. Jou nal o S a is ical So wa e, 40, 1–37.
Gansne , E. R., & No h, S. C. (2000). An open g aph isualiza ion sys em and i s applica ions o
so wa e enginee ing. So wa e - P ac ice and Expe ience, 30(11), 1203–1233.
Gelman, A. (2018) You need 16 imes he sample size o es ima e an in e ac ion han o es ima e a
main e ec . S a is ical Modeling, Causal In e ence, and Social Science
Gillea d, C., & Higgs, P. (2007). The hi d age and he baby boome s. In e na ional Jou nal o Ageing
and La e Li e, 2(2), 13–30.
Goldman, N. (1993). Ma iage selec ion and mo ali y pa e ns: In e ences and allacies. Demog aphy,
30(2), 189–208.
G o e , S., & Helliwell, J. F. (2019). How’s li e a home? New e idence on ma iage and he se poin o
happiness. Jou nal o Happiness S udies, 20(2), 373–390.
Guma, J., Cáma a, A. D., & T e iño, R. (2015). The ela ionship be ween heal h and pa ne ship his o y
in adul hood: Insigh s h ough e ospec i e in o ma ion om Eu opeans aged 50 and o e . Eu o-
pean Jou nal o Ageing, 12, 71–9.
Han, K.-T., Pa k, E.-C., Kim, J.-H., Kim, S. J., & Pa k, S. (2014). Is ma i al s a us associa ed wi h quali y
o li e? Heal h and Quali y o Li e Ou comes, 12(1), 109.
Hasselblad, V., & Hedges, L. V. (1995). Me a-analysis o sc eening and diagnos ic es s. Psychologi-
cal Bulle in, 117(1), 167–178.
Ha a i, E., & Mazzonna, F. (2015). Can we us olde people’s s a emen s on hei childhood ci cum-
s ances? E idence om SHARELIFE. Eu opean Jou nal o Popula ion, 31(3), 233–257.
Headey, B. (2010). The se poin heo y o well-being has se ious laws: On he e e o a scien i ic
e olu ion? Social Indica o s Resea ch, 97(1), 7–21.
Hla ac, M. (2018). S a gaze : well- o ma ed eg ession and summa y s a is ics ables. Cen al Eu o-
pean Labou S udies Ins i u e (CELSI), B a isla a, Slo akia
Hol -Luns ad, J., Bi mingham, W., & Jones, B. Q. (2008).Is he e some hing unique abou ma iage? The
ela i e impac o ma i al s a us, ela ionship quali y, and ne wo k social suppo on ambula o y
blood p essu e and men al heal h. Annals o Beha io al Medicine, 35(2), 239–244.
Ho wi z, A. V., Whi e, H. R., & Howell-Whi e, S. (1996). Becoming ma ied and men al heal h: A longi-
udinal s udy o a coho o young adul s. Jou nal o Ma iage and he Family, 58(4), 895.
House, J. S., Umbe son, D., & Landis, K. R. (1988). S uc u es and p ocesses o social suppo . Annual
Re iew o Sociology, 14, 293–318.
Hughes, M. E., & Wai e, L. J. (2009). Ma i al biog aphy and heal h a mid-li e. Jou nal o Heal h and
Social Beha io , 50(3), 344–358.
Hägglund, A., So sa, T., Danielsbacka, M., Tanskanen, A., & Ro ki ch, A. (2021). Vanhenee kuin
hy ä iini: yli 50- uo iaiden pa isuh ee Suomessa. Pe heba o me i
Jenkinson, C. (2013). Measu ing heal h and medical ou comes. Rou ledge.
Jung, J. (2023). Pa ne ship ajec o ies and hei consequences o e he li e cou se. E idence om
he Ge man li e s udy. Ad ances in Li e Cou se Resea ch, 55, 100525.
Kau man, L. & Rousseeuw, P. J. (1990). Finding g oups in da a: an in oduc ion o clus e analysis.
Wiley se ies in p obabili y and ma hema ical s a is ics, applied p obabili y and s a is ics. Wiley
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 51 o 54 12
Klüsene , S. (2015). Spa ial a ia ion in non-ma i al e ili y ac oss Eu ope in he wen ie h and
wen y- i s cen u ies: Recen ends, pe sis ence o he pas , and po en ial u u e pa hways. The
His o y o he Family, 20(4), 593–628.
Koball, H. L., Moiduddin, E., Hende son, J., Goesling, B., & Besculides, M. (2010). Wha do we
know abou he link be ween ma iage and heal h? Jou nal o Family Issues, 31(8), 1019–1040.
Kon ula, O. (2016). Lemmen Paula. Seksuaalinen hy in oin i pa isuhdeonnen a aimena. amily
ba ome e 2016, Popula ion Resea ch Ins i u e, Helsinki
K a z, F., & Pa zina, A. (2020). Endogenous selec ion bias and cumula i e inequali y o e he li e cou se:
E idence om educa ional inequali y in subjec i e well-being. Eu opean Sociological Re iew,
36(3), 333–350.
Len h, R.V. (2021). Emmeans: Es ima ed ma ginal means, aka leas -squa es means.
Les haeghe, R., & an de Kaa, K. (1986). Twee demog a ische ansi ies? In Be olking : G oei en k imp.
Van Loghum Sla e us.
Lewin, A. C. (2017). Heal h and ela ionship quali y la e in li e: A compa ison o li ing apa
oge he (LAT), i s ma iages, ema iages, and cohabi a ion. Jou nal o Family Issues, 38(12),
1754–1774.
Lewis, M. A., & Bu e ield, R. M. (2007). Social con ol in ma i al ela ionships: E ec o one’s pa ne
on heal h beha io s. Jou nal o Applied Social Psychology, 37(2), 298–319.
Long, J. A. (2024). j ools: Analysis and p esen a ion o social scien i ic da a. Jou nal o Open Sou ce
So wa e, 9(101), 6610.
Ludwig, V., & B üde l, J. (2018). Is he e a male ma i al wage p emium? New e idence om he uni ed
s a es. Ame ican Sociological Re iew, 83(4), 744–770.
Luyckx, K., Sei ge-K enke, I., Schwa z, S. J., C oce i, E., & Klims a, T. A. (2014). Iden i y con igu a-
ions ac oss lo e and wo k in eme ging adul s in oman ic ela ionships. Jou nal o Applied De el-
opmen al Psychology, 35(3), 192–203.
Manzoli, L., Villa i, P., M Pi one, G., & Boccia, A. (2007). Ma i al s a us and mo ali y in he elde ly:
A sys ema ic e iew and me a-analysis. Social Science & Medicine, 64(1), 77–94.
Ma golis, R., & My skylä, M. (2011). A global pe spec i e on happiness and e ili y. Popula ion and
De elopmen Re iew, 37(1), 29–56.
Mas ekaasa, A. (1992). Ma iage and psychological well-being: Some e idence on selec ion in o ma -
iage. Jou nal o Ma iage and he Family, 54(4), 901.
Mea a, E. R., Richa ds, S., & Cu le , D. M. (2008). The gap ge s bigge : Changes in mo ali y and li e
expec ancy, by educa ion, 1981–2000. Heal h A ai s, 27(2), 350–360.
Melo, S., Guedes, J., & Mendes, S. (2019). Theo y o cumula i e disad an age/ad an age. In D. Gu & M.
E. Dup e (Eds.), Encyclopedia o ge on ology and popula ion aging (pp. 1–8). Sp inge .
Me sä-Simola, N., & Ma ikainen, P. (2013). The sho - e m and long- e m e ec s o di o ce on mo al-
i y isk in a la ge Finnish coho , 1990–2003. Popula ion S udies, 67(1), 97–110.
Mi owsky, J., & Ross, C. E. (2005). Educa ion, cumula i e ad an age, and heal h. Ageing In e na ional,
30(1), 27–62.
Mo elmans, D. (2020). Economic consequences o di o ce: A e iew. In Pa en al li e cou ses a e sepa-
a ion and di o ce in Eu ope, (pp. 23–41). Sp inge .
Næss, E. O., Mehlum, L., & Qin, P. (2021). Ma i al s a us and suicide isk: Tempo al e ec o ma i-
al b eakdown and con ex ual di e ence by socioeconomic s a us. SSM - Popula ion Heal h, 15,
100853.
O’Rand, A. M. (2001). S a i ica ion and he li e cou se: The o ms o li e-cou se capi al and hei
in e ela ionships. In Handbook o aging and he social sciences (pp. 197–213). Academic
P ess.
O’Flahe y, M., Bax e , J., Haynes, M., & Tu ell, G. (2016). The amily li e cou se and heal h: Pa ne -
ship, e ili y his o ies, and la e -li e physical heal h ajec o ies in aus alia. Demog aphy, Sil e
Sp ing, 53(3), 777–804.
Pe elli-Ha is, B., & Amos, M. (2015). Changes in pa ne ship pa e ns ac oss he li e cou se: An exami-
na ion o 14 coun ies in Eu ope and he Uni ed S a es. Demog aphic Resea ch, 33(6), 145–178.
Pe e s, A., & Lie b oe , A. C. (1997). Beyond ma i al s a us: Pa ne his o y and well-being in old age.
Jou nal o Ma iage and Family, 59(3), 687–699.
Pinqua , M. (2003). Loneliness in ma ied, widowed, di o ced, and ne e -ma ied olde adul s. Jou nal
o Social and Pe sonal Rela ionships, 20(1), 31–53.
P us, S. G. (2007). Age, SES, and heal h: A popula ion le el analysis o heal h inequali ies o e he
li ecou se. Sociology o Heal h & Illness, 29(2), 275–296.
M.Mäki e al.
12 Page 52 o 54
R Co e Team. (2021). R: A language and en i onmen o s a is ical compu ing. R Founda ion o
S a is ical Compu ing.
Raul , W. (2019). Is he ci il solida i y pac he u u e o ma iage? The se e al meanings o he F ench
ci il union. In e na ional Jou nal o Law, Policy and he Family, 33(2), 139–159.
Raz-Yu o ich, L. (2013). Di o ce penal y o di o ce p emium? A longi udinal analysis o he con-
sequences o di o ce o men’s and women’s economic ac i i y. Eu opean Sociological Re iew,
29(2), 373–385.
Reczek, C., Pud o ska, T., Ca , D., Thomee , M. B., & Umbe son, D. (2016). Ma i al his o ies and
hea y alcohol use among olde adul s. Jou nal o Heal h and Social Beha io , 57(1), 77–96.
Rene lo , A., & Mamelund, S.-E. (2012). The associa ion be ween ma i al s a us and psychological well-
being in No way. Eu opean Sociological Re iew, 28, 355–365.
Rich, B. (2021). Table1: Tables o desc ip i e s a is ics in HTML.
Rind uss, R. R., & Vandenheu el, A. (2022). Cohabi a ion: A p ecu so o ma iage o an al e na i e o
being single? Popula ion and De elopmen Re iew, 16(4), 25.
Robe son, P. N. E., No ona, J. C., Lenge , K. A., & Olms ead, S. B. (2018). How do ela ionship s abili y
and quali y a ec wellbeing?: Roman ic ela ionship ajec o ies, dep essi e symp oms, and li e
sa is ac ion ac oss 30 yea s. Jou nal o Child and Family S udies, 27(7), 2171–2184.
Robe s, S. G. B., & Dunba , R. I. M. (2011). Communica ion in social ne wo ks: E ec s o kinship, ne -
wo k size, and emo ional closeness. Pe sonal Rela ionships, 18(3), 439–452.
Rosen eld, M. J. (2018). Who wan s he b eakup? Gende and b eakup in he e osexual couples. In D. F.
Alwin, D. H. Felmlee, & D. A. K eage (Eds.), Social ne wo ks and he li e cou se: In eg a ing he
de elopmen o human li es and social ela ional ne wo ks (pp. 221–243). Sp inge .
Ross, C. E. (1995). Reconcep ualizing ma i al s a us as a con inuum o social a achmen . Jou nal o
Ma iage and he Family, 57(1), 129.
Ross, C. E., & Mi owsky, J. (2006). Sex di e ences in he e ec o educa ion on dep ession: Resou ce
mul iplica ion o esou ce subs i u ion? Social Science & Medicine, 63(5), 1400–1413.
Ross, C. E., & Mi owsky, J. (2010). Gende and he heal h bene i s o educa ion. The Sociological Qua -
e ly, 51(1), 1–19.
Ross, C. E., & Wu, C.-L. (1996). Educa ion, age, and he cumula i e ad an age in heal h. Jou nal o
Heal h and Social Beha io , 37(1), 104.
Ro ki ch, A. (2020). Wha a e couples made o ? Union o ma ion in high-income socie ies. In O. Bu ge ,
R. Lee, & R. Sea (Eds.), Human e olu iona y demog aphy. Open Book Publishe s.
Régnie -Loilie , A. (2016). Pa ne ship ajec o ies o people in s able non-cohabi ing ela ionships in
F ance. Demog aphic Resea ch, 35, 1169–1212.
Sassle , S. (2010). Pa ne ing ac oss he li e cou se: Sex, ela ionships, and ma e selec ion. Jou nal o
Ma iage and Family, 72(3), 557–575.
Scha z, C. R. (2010). Pa hways o educa ional homogamy in ma i al and cohabi ing unions. Demog a-
phy, 47(3), 735–753.
Schü z, J. (2019). Ma i al biog aphy and heal h in old age: Insigh s om Eu opean su ey da a
Sco , E. S. (2000). Social no ms and he legal egula ion o ma iage. Vi ginia Law Re iew. h ps:// doi.
o g/ 10. 2307/ 10738 33
S yke , S. (1959). Symbolic in e ac ion as an app oach o amily esea ch. Ma iage and Family Li ing,
21(2), 111.
S ude , M. (2013). Weigh edClus e lib a y manual: A p ac ical guide o c ea ing ypologies o ajec o-
ies in he social sciences wi h R. Technical epo , LIVES Wo king Pape s 24.
Tambellini, E., Danielsbacka, M., & Ro ki ch, A. (2024). Bo h pa ne ship his o y and cu en ela ion-
ship quali y a e associa ed wi h li e sa is ac ion in old age. Resea ch on Aging. h ps:// doi. o g/ 10.
1177/ 01640 27524 13092 55
Ta akol, Z., Nikbakh Nas abadi, A., Behboodi Moghadam, Z., Salehiniya, H., & Rezaei, E. (2017). A
e iew o he ac o s associa ed wi h ma i al sa is ac ion. Galen Medical Jou nal, 6(3), e641.
UNSD. (1999). Uni ed na ions geoscheme o eu ope. h ps:// uns a s. un. o g/ unsd/ me ho dology/ m49/.
Umbe son, D. (1992). Gende , ma i al s a us and he social con ol o heal h beha io . Social Science &
Medicine, 34(8), 907–917.
Vande Weele, T. J. (2017). On a squa e- oo ans o ma ion o he odds a io o a common ou come.
Epidemiology, 28(6), e58–e60.
Vande Weele, T. J., & Ding, P. (2017). Sensi i i y analysis in obse a ional esea ch: In oducing he
e- alue. Annals o In e nal Medicine, 167(4), 268–274.
S able Ma i al His o ies P edic Happiness andHeal h Ac oss… Page 53 o 54 12
Ve bakel, E. (2012). Subjec i e well-being by pa ne ship s a us and i s dependence on he no ma i e
clima e. Eu opean Jou nal o Popula ion / Re ue Eu opéenne de Démog aphie, 28(2), 205–232.
Wai e, L. J. (1995). Does ma iage ma e ? Demog aphy, 32(4), 483–507.
Wang, D., Li, D., Mish a, S. R., Lim, C., Dai, X., Chen, S., & Xu, X. (2022). Associa ion be ween ma i al
ela ionship and mul imo bidi y in middle-aged adul s: A longi udinal s udy ac oss he US, UK,
Eu ope, and China. Ma u i as, 155, 32–39.
Wa d, J. H. (1963). Hie a chical g ouping o op imize an objec i e unc ion. Jou nal o he Ame ican
S a is ical Associa ion, 58(301), 236–244.
Wes maas, J. L., Wild, T. C., & Fe ence, R. (2002). E ec s o gende in social con ol o smoking cessa-
ion. Heal h Psychology, 21(4), 368.
Wickham, H. (2016). Ggplo 2: Elegan g aphics o da a analysis. Sp inge -Ve lag.
Wong, J. S., & Wai e, L. J. (2015). Ma iage, social ne wo ks, and heal h a olde ages. Jou nal o Popu-
la ion Ageing, 8(1–2), 7–25.
Wood, R. G., A ella , S., & Goesling, B. (2009). The e ec s o ma iage on heal h: A syn hesis o ecen
esea ch e idence. No a Science
Woszidlo, A., & Seg in, C. (2013). Nega i e a ec i i y and educa ional a ainmen as p edic o s o new-
lyweds’ p oblem sol ing communica ion and ma i al quali y. The Jou nal o Psychology, 147(1),
49–73.
Wängq is , M., Ca lsson, J., an de Lee, M., & F isén, A. (2016). Iden i y de elopmen and oman ic
ela ionships in he la e wen ies. Iden i y, 16(1), 24–44.
Zahl-Olsen, R., Thuen, F., & Espehaug, B. (2019). Di o ce and ema iage in no way: A p ospec i e
coho s udy be ween 1981 and 2013. Jou nal o Di o ce & Rema iage, 60(8), 600–611.
Zajaco a, A., & Law ence, E. M. (2018). The ela ionship be ween educa ion and heal h: Reducing dis-
pa i ies h ough a con ex ual app oach. Annual Re iew o Public Heal h, 39, 273–289.
Zimme mann, O., & Hameis e , N. (2019). S able cohabi a ional unions inc ease quali y o li e: Re o-
spec i e analysis o pa ne ship his o ies also e eals gende di e ences. Demog aphic Resea ch,
40, 657–692.
Øien-Ødegaa d, C., Hauge, L. J., & Rene lo , A. (2021). Ma i al s a us, educa ional a ainmen , and sui-
cide isk: A No wegian egis e -based popula ion s udy. Popula ion Heal h Me ics, 19(1), 33.
M.Mäki e al.
12 Page 54 o 54
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps
and ins i u ional a ilia ions.
Au ho s and A ilia ions
MiikaMäki1,2,6,7 · AnnaE ikaHägglund1,3· AnnaRo ki ch1 ·
Sangi aKula hinal4· MikkoMy skylä5,6,7
* Miika Mäki
miika.maki@ aes olii o. i
Anna E ika Hägglund
[email p o ec ed]
Anna Ro ki ch
anna. o ki ch@ aes olii o. i
Sangi a Kula hinal
sangi a.k[email p o ec ed]
Mikko My skylä
my skyla@demog .mpg.de
1 Popula ion Resea ch Ins i u e a heFamily Fede a ion o Finland, Helsinki, Finland
2 Cen e o Social Da a Science, Uni e si y o Helsinki, Helsinki, Finland
3 Uni e si y o Tu ku, Tu ku, Finland
4 Depa men o Ma hema ics andS a is ics, Uni e si y o Helsinki, Helsinki, Finland
5 Max Planck Ins i u e o Demog aphic Resea ch, Ros ock, Ge many
6 Max Planck – Uni e si y o Helsinki Cen e o Social Inequali ies inPopula ion Heal h,
Helsinki, Finland
7 Helsinki Ins i u e o Demog aphy andPopula ion Heal h, Uni e si y o Helsinki, Helsinki,
Finland
