Marriage Advantages in Perinatal Health: Evidence of Marriage Selection or Marriage Protection?
| DOI | http://doi.org/10.1111/jomf.12257 |
| Date | 01 February 2016 |
| Published date | 01 February 2016 |
| Author | Jennifer B. Kane |
J B. K University of California, Irvine
Marriage Advantages in Perinatal Health: Evidence
of Marriage Selection or Marriage Protection?
Marriage is a social tie associated with health
advantages for adults and their children, as
lower rates of preterm birth and low birth weight
are observed among married women. In this
study the author tested 2 competing hypotheses
explaining this marriage advantage—marriage
protection versus marriage selection—using a
sample of recent births to single, cohabiting,
and married women from the National Survey
of Family Growth,2006–2010. Propensity score
matching and xed effects regression results
demonstrated support for marriage selection, as
a rich set of early life selection factors account
for all of the cohabiting–married disparity and
part of the single–married disparity. Subsequent
analyses demonstrated that prenatal smoking
mediates the adjusted single–married dispar-
ity in birth weight, lending some support for
the marriage protection perspective. The study’s
ndings sharpen our understanding of why and
how marriage matters for child well-being and
provide insight into pre-conception and prena-
tal factors describing intergenerational trans-
missions of inequality via birth weight.
Marriage is a social tie associated with health
advantages for adults and their children (Repetti,
Taylor, & Seeman, 2002; Umberson & Mon-
tez, 2010; Waite, 1995), provided the union is
Department of Sociology, Universityof California, 4171
Social Sciences Plaza A, Irvine, CA 92697
(jbkane@uci.edu).
Key Words: child well-being, cohabitation, health dispari-
ties, infants, marriage, pregnancy.
generally supportive and not extremely stress-
ful (Kiecolt-Glaser & Newton, 2001; Umber-
son, Williams, Powers, Liu, & Needham, 2006).
Health advantages are also widely documented
for infants of married mothers: Lower rates of
problematic health outcomes such as low birth
weight (weighing less than 51
∕
2pounds at birth)
and preterm birth (delivery before 37 weeks of
gestation) have been reported among married
(vs. unmarried) mothers (Buckles & Price, 2013;
Martin, Hamilton, Osterman, Curtin, & Math-
ews, 2013).
At debate is whether marriage advantages in
health are confounded by selection factors—a
marriage selection hypothesis—or whether mar-
riage causes women to adopt heathier attitudes
or behaviors that translate to better health—a
marriage protection hypothesis. In the case of
perinatal health, the answer is unclear, in part
because nearly all studies to date have not con-
sidered both explanations.
We also know very little about advantages
in perinatal health for infants born to married
women relative to infants born to cohabiting and
single women. Past studies have largely con-
trasted birth outcomes between married women
and all unmarried women, despite clear dif-
ferences between cohabitation and singlehood.
Among the studies that have distinguished
between cohabitation and singlehood (Bird,
Chandra, Bennett, & Harvey, 2000; Fomby,
2011; Sullivan, Raley, Hummer, & Schiefelbein,
2012), most examined pre-1990 births and/or
did not directly measure cohabitation. The
widespread, dramatic increase in cohabitation
and nonmarital childbearing, along with marked
212 Journal of Marriage and Family 78 (February 2016): 212–229
DOI:10.1111/jomf.12257
Marriage Advantages in Perinatal Health 213
compositional change among subgroups of
women married, cohabiting, and single at birth,
since the 1990s (Cherlin, 2010; Seltzer, 2000;
Smock & Greenland, 2010) establishes the
importance of understanding population-level
perinatal health disparities across all three
groups and within recent years. This new
knowledge will not only clarify how pervasive
these marriage advantages are but will also
enrich our understanding of the broader impli-
cations of cohabitation for child well-being, a
topic of substantial interest among marriage and
family scholars.
Furthermore, nearly all marriage and peri-
natal health studies have measured marital
status at the time of birth. This is likely due
to data limitations; however, I argue below
that, in pursuit of capturing the true effect of
marriage on perinatal health, it is preferable
to measure marital status at conception, given
that (a) at least some of marriage’s protective
effect, as it has been described in past work, is
assumed to be exerted during pregnancy and
(b) changes in union status between the time
when couples learn of a pregnancy and the
time of birth are not uncommon (Bachu, 1999;
Rackin & Gibson-Davis, 2012). Establishing
the ideal stage at which to capture union status
has important implications for how researchers
conceptualize and model exposure to marriage.
Last, research in this area has yet to identify
a rich set of selection factors that confound
the association between marriage and perina-
tal health. It is well known that the early life
environment inuences women’s risk of marital
childbearing (see, e.g., Amato et al., 2008), but
whether and how the early life environment
inuences birth weight is not well established.
Health scholars have recently articulated the
importance of identifying pre-conception fac-
tors affecting perinatal health, yet most studies
continue to focus on the prenatal period—or on
the 12 months leading up to conception—to the
exclusion of prior life events (Johnson et al.,
2006; van Dyck, 2010). This is partly due to data
limitations, given that birth certicate data are
commonly used and do not contain this informa-
tion. But this is also due to statistical limitations,
as xed effects (FE) models are commonly used
to account for selection, and this method dif-
ferences out, rather than reveals, early life
selection factors. However, new data now exist
that include a rich set of pre-conception factors
in addition to marital status and birth weight.
In this study I also used a different statistical
approach, propensity score matching (PSM),
that can reveal key pre-conception factors. This
rich description is critical for expanding our
knowledge of pre-conception risk factors. This
information can also inform a growing literature
in the social sciences that pinpoints birth weight
as one specic mechanism that can transmit
inequality across generations from mothers
to children (Case, Lubotsky, & Paxson, 2002;
Case & Paxson, 2006; Kane, 2015). The rich
description of early life selection factors would
add new knowledge clarifying the rst portion
of this mechanism—specic ways in which
maternal disadvantage contributes to adverse
birth outcomes.
This study goes beyond past research by rig-
orously assessing the extent to which both mar-
riage protection and marriage selection explain
marriage advantages in perinatal health using
contemporary, population-level data (National
Survey of Family Growth, 2006–2010). Mar-
riage advantages were contrasted across three
relationship statuses (married, cohabiting, sin-
gle) at the time of conception and assessed using
two statistical strategies that account for unob-
served heterogeneity (PSM and FE models).
Furthermore, in this study I sought to identify
a rich set of pre-conception factors confound-
ing the marriage advantage (see Figure 1). The
study ndings will reveal whether and how mar-
riage matters for a critical indicator of child
well-being and provide insight into early life and
prenatal factors that describe how inequality is
transmitted across generations via birth weight.
B
Marriage Selection Versus Marriage Protection
Two competing hypotheses have been devel-
oped to explain marriage advantages in health.
The rst, marriage selection, posits that selec-
tion factors, such as greater socioeconomic sta-
tus or better health, place individuals at higher
risk of being married and of exhibiting health
advantages (Goldman, 1993; Haas, 2006, 2008).
Accounting for the endogeneity of marriage is
therefore necessary to produce an unbiased esti-
mate of the effect of marriage on health. This
explanation is plausible in the case of perinatal
health because mothers with lower human and
nancial capital are less likely to be married at
childbirth (Seltzer, 2000) and exhibit higher risk
of adverse birth outcomes (Martin et al., 2013).
To continue reading
Request your trial