Mental Well‐Being Differences in Cohabitation and Marriage: The Role of Childhood Selection

• DOI: http://doi.org/10.1111/jomf.12431
• Author: Marta Styrc,Brienna Perelli‐Harris
• Date: 01 February 2018
• Published date: 01 February 2018

B P-H University of Southampton

M S University of Southampton∗

Mental Well-Being Differences in Cohabitation and

Marriage: The Role of Childhood Selection

Prior studies have found that marriage benets

well-being, but cohabitation may providesimilar

benets. An analysis of the British Cohort Study

1970, a prospective survey following respon-

dents to age 42, examines whether partnerships

in general, and marriage in particular, inuence

mental well-being in midlife. Propensity score

matching indicates whether childhood char-

acteristics are a sufcient source of selection

to eliminate differences in well-being between

those living with and without a partner and

those cohabitating and married. The results

indicate that matching on childhood character-

istics does not eliminate advantages to living

with a partner; however, matching eliminates

differences between marriage and cohabitation

for men and women more likely to marry. On

School of Social Sciences, Bldg 58, Room 4013, University

of Southampton, SO17 1BJ, United Kingdom. (b.g.perelli-

harris@soton.ac.uk).

∗ESRC Centre for Population Change, Bldg 58, Room

2001, Faculty of Social, Human and Mathematical

Sciences, University of Southampton, SO17 1BJ, United

Kingdom.

© 2017 The Authors. Journal of Marriage and Familypub-

lished by WileyPeriodicals, Inc. on behalf of National Coun-

cil on Family Relations.

This is an open access article under the terms of the Creative

Commons Attribution License, which permits use, distribu-

tion and reproduction in any medium, provided the original

work is properly cited.

Key Words: childhood, cohabitation, marriage, mental

health.

the other hand, marriage may provide benets

to women less likely to marry unless they have

shared children and are in long-lasting part-

nerships. Hence, childhood selection attenuates

differences between cohabitation and marriage,

except for women less likely to marry.

Numerous studies have found that marriage ben-

ets mental well-being (for a review, see Waite

& Gallagher, 2000; Lamb, Lee, & DeMaris,

2003; Simon, 2002). The strength and persis-

tence of these ndings have led some policy

makers to call for programs that encourage

marriage. For example, pro-marriage policy

initiatives were pushed during the George W.

Bush administration in the United States (Bir

et al., 2012), and the current conservative UK

government has legislated tax breaks for mar-

ried couples (BBC News, 2015). Much of the

research underlying these initiatives, however,

has compared the married and unmarried, with-

out distinguishing between those who were

cohabiting or single (e.g., Hughes & Waite,

2009; Liu & Umberson, 2008; Waite & Gal-

lagher, 2000). Some studies have begun to

examine differences between cohabitation and

marriage with respect to mental health and

depressive symptoms (e.g., Brown, 2000; Lamb

et al., 2003; Musick & Bumpass, 2012), but most

of these studies use American cross-sectional

data and may not sufciently control for selec-

tion effects. Less is known about the relationship

between marriage and mental health in other

contexts, where cohabitation is widespread.

Journal of Marriage and Family 80 (February 2018): 239–255 239

DOI:10.1111/jomf.12431

240 Journal of Marriage and Family

Given the recent increase in cohabitation and its

changing meaning (Berrington, Perelli-Harris,

& Trevena, 2015; Perelli-Harris et al., 2014),

it is important to revisit whether partnerships

in general, and marriage in particular, con-

tinue to provide distinct benets to well-being,

especially for those who are less likely to marry.

In this study, we examine whether being in

a partnership and the type of the partnership—

marriage or cohabitation—are associated with

higher mental well-being in midlife in the United

Kingdom. In the United Kingdom, cohabitation

has become the normative pathway to union

formation: From 2004 to 2007, 80% of all

marriages started with premarital cohabitation,

and the duration of cohabiting unions has been

steadily increasing (Beaujouan & Ni Bhrolchain,

2011). Cohabitation has also become common

for childbearing: In 2012, 30% of all babies were

born to cohabiting mothers (Ofce for National

Statistics, 2013). The pervasiveness of cohabita-

tion, especially if it takes on much of the form

and function of marriage, suggests that cohab-

itors may have well-being similar to that of mar-

ried individuals.

Nonetheless, a large number of studies

across countries have shown distinct differences

between cohabitation and marriage with respect

to depressive symptoms in the United States

(Brown, 2000), relationship quality across

Europe (Wiik, Keizer, & Lappegård, 2012), and

life satisfaction in Europe (Soons & Kalmijn,

2009). Some of the studies that examined the

relationship between marriage and psycholog-

ical or subjective well-being employed a panel

design that starts in young adulthood, providing

insights into transitions during a short period

but not directly comparing long-term cohabiting

and marital unions and their effects in midlife

(Germany: Zimmerman & Easterlin, 2006;

the Netherlands: Soons, Liefbroer, & Kalmijn,

2009; United States: Musick & Bumpass, 2012).

For example, using xed effects models, Musick

and Bumpass (2012) found that in the United

States, transitions into cohabitation and mar-

riage have similar effects on psychological,

health, and social well-being, and any differ-

ences found are relatively small. Musick and

Bumpass (2012), however, did not examine

the consequences of cohabitation versus mar-

riage later in the life course, after the typical

postponement of marriage throughout young

adulthood and after the majority of childbearing.

In addition, although this and other studies using

xed effects models examined variation within

individuals over time, they did not compare

across individuals with different characteristics

that select people into cohabitation or marriage,

particularly drawing on selection mechanisms

occurring early in life, before the entrance

into adulthood. These studies did not exam-

ine whether marriage was likely to increase

well-being for those who were less likely to

marry, often those with disadvantaged back-

grounds and targeted by pro-marriage policies.

Our study uses propensity score matching to

investigate differences between marriage and

cohabitation in the United Kingdom. Although

union status changes during the life course and

cohabiting couples often marry (Perelli-Harris

& Lyons-Amos, 2015), we consider marriage

a “treatment,” because couples must ofcially

decide to marry and act on that decision. Our

propensity score matching (PSM) strategy

examines whether people with similar back-

ground characteristics are more likely to have

higher mental well-being scores if they marry.

This approach takes into account important

selection characteristics that predict both union

formation and well-being, but cannot control for

the respondents’ current situation. By matching

people with similar characteristics, however,

we can ascertain whether marriage provides

benets beyond living with someone as well as

the heterogeneity of treatment effects—whether

the effects of marriage differ for those with a

higher or lower propensity to marry based on

childhood characteristics. Our data, the British

Cohort Study, followed the 1970 birth cohort up

to age 42. This cohort experienced partnership

formation in the 1990s and 2000s, which is more

recent than many previous studies (e.g., Musick

& Bumpass, 2012). In addition, cohabitation

in midlife is relatively understudied despite

it becoming a more common phenomenon

in Britain (i.e., 20% of adults in our survey

were cohabiting at age 42). Given that we

are not interested in the timing of marriage

or cohabitation per se, the PSM approach is

appropriate for examining whether currently

being in a marriage matters. However, because

the duration of and investments in a union may

signal the positive benets of the partnership,

we also compare the type of union for those

who have been in long-lasting partnerships,

never experienced union separation, and have

had children together. In addition, we com-

pare results for men and women who may