Marital Biography and Well-Being in Later Life: the Role of Remarriage
MARITAL BIOGRAPHY AND WELL-BEING IN LATER LIFE: THE ROLE OF REMARRIAGE, DISRUPTION PATHWAYS, AND DURATION ON HEALTH, PARENT- CHILD CONTACT, AND AMBIVALENCE TOWARD CHILDREN
Anna Marie Hammersmith
A Dissertation
Submitted to the Graduate College of Bowling Green State University in partial fulfillment of the requirments for the degree of
DOCTOR OF PHILOSOPHY
August 2018
Committee:
I-Fen Lin, Advisor
Sudershan Jetley Graduate Faculty Representative
Susan L. Brown
Karen Benjamin Guzzo
Wendy D. Manning © 2018
Anna Marie Hammersmith
All Rights Reserved iii
ABSTRACT
I-Fen Lin, Advisor
In 2015, nearly 30% of individuals aged 50 and older had two or more marriages
compared to 19% in 1980, indicating that a growing share of older adults either divorced or were
widowed during early or midlife and later remarried. Although widowhood remains a common
exit from marriage in later life, divorce to people over 50 is also on the rise. Despite increasingly
complex marital biographies of older adults, few researchers have examined differences between
disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions) and whether
remarriage is associated with fewer costs of marital disruption. It is also unclear whether
duration remarried or unmarried relates to better or worse health after different disruption
pathways. Using the 1992-2014 Health and Retirement Study, I investigate the associations of
different disruption pathways, subsequent remarriage relative to being unmarried, and duration
remarried or unmarried with older adults’ mental and physical health, contact with children, and
ambivalence toward children. I also account for gender differences as the health and parent-child
ties of men and women often differ in later life. This dissertation underscores the need to pay
attention to older people with multiple disruptions, as they are often disadvantaged in health and parent-child relationships relative to older adults with one divorce or widowhood. The findings regarding the role of remarriage for each well-being outcome are mixed. Remarriage is beneficial for the mental health of men relative to being unmarried after any type of disruption, and for the physical health of divorced women. Although remarriage relates to more frequent parent-child
contact for divorced men, remarriage relates to less contact among women after one widowhood
or multiple disruptions. Remarriage also links to greater ambivalence among men after multiple disruptions. Duration also matters, but not uniformly across outcomes. Remarried men after iv multiple disruptions have worse mental health with more years remarried men after one divorce, indicating that duration remarried after multiple disruptions links to poorer mental health than duration remarried after one divorce. Although men with multiple disruptions have less contact with children than widowed men, additional years remarried yield more contact for men with multiple disruptions than for men with one widowhood. Moreover, women who remarry after widowhood have less contact with children than their unmarried counterparts, but each additional year remarried after widowhood is associated with more contact, suggesting that these remarried women are able to rebuild ties with children over time. In sum, my dissertation highlights the utility of employing different disruption pathways, subsequent remarriage, and duration remarried or unmarried to capture the increasing complexity of marital biographies among older adults and to clarify its associations with multiple dimensions of well-being.
v
ACKNOWLEDGEMENTS
There are so many people for whom I have had insurmountable appreciation as I have worked toward completing my degree. I have an endless amount of gratitude for my wonderful mentor and advisor, Dr. I-Fen Lin. It is amazing to reflect on how much I have learned from you since you became my mentor nearly six year ago. You have not only taught me how to be a curious and conscientious researcher, but from you I have also learned how to be steadfastly persistent. You have been a great academic model, but you have also taught me a lot about being a caring mentor. I appreciate the number of tears you have let me shed in your office when circumstances have been tough. Thank you for always helping me achieve my goals and shaping me into the scholar that I am today.
I am also incredibly grateful for my relationship with Dr. Susan L. Brown as I pursued my degree. Over my years of working with you on the gray divorce grant project, I have learned how staying curious and attentive can lead to new and innovative research ideas. I would not be the researcher I am today without these experiences.
Dr. Karen Guzzo, Dr. Wendy Manning, and Dr. Sudershan Jetley thank you for your dedication to my success as I finished my dissertation. I also want to express my appreciation for all that I have learned from both Dr. Guzzo and Dr. Manning as I have worked toward my degree. You have both served as great academic models.
I also owe a wealth of gratitude for Dr. Anita Waters and Dr. Fareeda Griffith for assisting me with independent research throughout my undergraduate years at Denison
University. You both encouraged me to pursue my graduate degree and have continued to be supportive beyond my graduation from Denison. I’d also like to thank Dr. Veronica Gonzalez- vi
Lopez and Dr. Ruth Toulson for being exceptional educators and models for me inside and
outside of the classroom.
Mom, you are a magical unicorn. Thank you for your constant inspiration and support.
Not only did you empower me to pursue my dreams from a young age, you have continued to
model how to be an exemplary scholar, mother, and human. I would not be where I am or who I
am without you.
Dad, thank you for your endless love and support. It has meant the world to me that you
often would come to Bowling Green (or Denison) to take me to dinner when you recognized that
I was especially stressed with work or cat illnesses. You have always been there when I have
needed you the most without any hesitation.
Renee Hammersmith, I have so much gratitude for all of your care and compassion as I
have worked toward finishing my degree. I appreciate all of the visits, text messages, and care
packages that were always well-timed for when I have needed them the most.
Bob Hust, I have much appreciation for the role you have played in my life. Not only have you been a sounding board as I have planned my future, but you have shown enormous empathy when I have faced difficulties in completing my degree. I also appreciate your humor and the fun times we have had in Cincinnati and at Lake Erie that have always managed to take my mind off of work.
Lori Daiello, thank you for serving as a model of what a successful researcher (and human) looks like. You have been such a great example of what women with careers in science can achieve. vii
Erik, Amanda, Colton, and David Hammersmith, thank you for all of the love and laughs throughout over these past years. Shout out to Colton and David for reminding me what it means to continue to be both mystified and curious about how the world works.
I also want to thank the rest of my extended family members for their love and encouragement. I am so lucky you have you all in my life.
To the MVR (most valuable roommate) Jessica Ziegler, thanks for continuing to text me daily even after you moved to Iowa (even if it is just us exchanging pictures of cats). I have numerous friends who have been my cheerleaders as I have worked toward my degree: Rachel
Bohn, Meryl Duff, Sarah Burns, Meghan Ulsh, Courtney Usmani, Olivia Bilyk, Lindsey Cooper,
Marissa Landeis, Kasey Eickmeyer, Andie Krieg, Bart Stykes, Vanessa Lang, Paul Hemez,
Jordan Galluzzo, Marta Alvira-Hammond, Jen Copp, and Emily Schondelmyer. After listing each of your names, I realize how fortunate I have been to accrue such an awesome group of friends.
I also want to acknowledge the love and support that was given to me throughout my educational journey who are no longer with me. Barbara and Edward Daiello, Joanne and
Clifford Hammersmith, and Justin Ziegler- I feel your love around me constantly.
Dorian Slaybod, thank you for being my champion through all of this. You have been a bright light in my life these past few years and I cannot imagine this journey without you. You have helped me maintain my focus during the most difficult of times and I am not sure I could have finished my degree without your unrelenting love and support. I am pretty sure the coming years are going to be pretty cool with you by my side. viii
Finally, even though they cannot read, I’d like to thank Moo and Dot for being the smallest, softest gentlemen who are always willing to offer ample distraction through play, snuggles, or climbing everything in my apartment. ix
TABLE OF CONTENTS
Page
CHAPTER 1: INTRODUCTION ...... 1
CHAPTER 2: MARITAL BIOGRAPHY AND HEALTH IN LATE LIFE ...... 8
Marital Resource, Stress, and Cumulative (Dis)Advantage Models ...... 10
Marital Status and Health...... 12
Remarriage, Disruption Pathways, and Health ...... 14
Duration Remarried, Duration Unmarried, and Health ...... 16
Variation by Gender ...... 18
The Present Study ...... 19
Demographic Characteristics ...... 22
Socioeconomic Resources ...... 22
Method ...... 23
Dependent Variables ...... 24
Depressive Symptoms ...... 24
Chronic Conditions ...... 24
Focal Independent Variables ...... 24
Marital Status ...... 24
Disruption Pathways ...... 24
Duration Remarried or Unmarried ...... 25
Covariates ...... 25
Analytic Strategy ...... 25
Descriptive Results ...... 26 x
Multivariate Results ...... 28
Marital Status and Health...... 28
Remarriage, Disruption Pathways, and Health ...... 29
Duration Remarried, Duration Unmarried, and Health ...... 31
Discussion ...... 32
CHAPTER 3: MARITAL BIOGRAPHY AND PARENT-CHILD CONTACT IN
LATER LIFE ...... 43
Intergenerational Solidarity ...... 45
Marital Status and Parent-Child Contact ...... 46
Remarriage, Disruption Pathways, and Parent-Child Contact ...... 47
Duration Remarried, Duration Unmarried, and Parent-Child Contact ...... 50
Variation by Gender ...... 51
The Present Study ...... 53
Demographic Characteristics ...... 54
Socioeconomic Resources ...... 55
Health ...... 55
Children’s Characteristics ...... 56
Method ...... 56
Dependent Variable ...... 57
Contact with children ...... 57
Focal Independent Variables ...... 58
Marital Status ...... 58
Disruption Pathways ...... 58 xi
Duration Remarried or Unmarried ...... 58
Covariates ...... 58
Analytic Strategy ...... 59
Descriptive Results ...... 60
Multivariate Results ...... 62
Marital Status and Parent-Child Contact ...... 62
Remarriage, Disruption Pathways, and Parent-Child Contact ...... 63
Duration Remarried, Duration Unmarried, and Parent-Child Contact ...... 65
Discussion ...... 66
CHAPTER 4: MARITAL BIOGRAPHY AND PARENT’S AMBIVALENCE
TOWARD CHILDREN IN LATER LIFE ...... 80
Intergenerational Ambivalence ...... 83
Marital Status and Ambivalence ...... 84
Remarriage, Disruption Pathways, and Ambivalence ...... 85
Duration Remarried, Duration Unmarried, and Ambivalence ...... 88
Variation by Gender ...... 90
The Present Study ...... 92
Demographic Characteristics ...... 94
Socioeconomic Resources ...... 94
Health ...... 95
Children’s Characteristics ...... 95
Method ...... 95
Dependent Variable ...... 97 xii
Ambivalence ...... 97
Focal Independent Variables ...... 97
Marital Status ...... 97
Disruption Pathways ...... 97
Duration Remarried or Unmarried ...... 98
Covariates ...... 98
Analytic Strategy ...... 99
Descriptive Results ...... 100
Multivariate Results ...... 102
Marital Status and Ambivalence ...... 102
Remarriage, Disruption Pathways, and Ambivalence ...... 104
Duration Remarried, Duration Unmarried, and Ambivalence ...... 103
Discussion ...... 104
CHAPTER 5: DISCUSSION ...... 116
Key Findings and Contributions ...... 118
Limitations ...... 127
Future Directions ...... 129
Conclusion ...... 132
REFERENCES ...... 138
xiii
LIST OF TABLES
TABLES Page
2.1 Weighted Means (Standard Deviations) and Percentages by Gender ...... 39
2.2 Coefficients (Standard Errors) from the Weighted OLS Regression of
Depressive Symptoms and Chronic Conditions on Marital Status ...... 40
2.3 Coefficients (Standard Errors) from the Weighted OLS Regression of
Depressive Symptoms and Chronic Conditions on Marital Status and
Disruption Pathway ...... 41
2.4 Coefficients (Standard Errors) from the Weighted OLS Regression of
Depressive Symptoms and Chronic Conditions on Marital Status, Marital
Disruption, and Duration ...... 42
3.1 Weighted Means (Standard Deviations) and Percentages by Gender ...... 73
3.2 Coefficients (Standard Errors) from the Weighted Multilevel Ordered
Logistic Regression Model of Marital Status on Contact with Children ...... 75
3.3 Coefficients (Standard Errors) from the Weighted Multilevel Ordered Logistic
Regression Model of Marital Status, and Disruption Pathway on Contact with
Children ...... 76
3.4 Coefficients (Standard Errors) from the Weighted Multilevel Ordered Logistic
Regression Model of Marital Status, Disruption Pathway, and Duration on
Contact with Children ...... 78
4.1 Weighted Means (Standard Deviations) and Percentages by Gender ...... 109
4.2 Coefficients (Standard Errors) from the Weighted OLS Regression Model of
Marital Status on Ambivalence toward Children ...... 111 xiv
4.3 Coefficients (Standard Errors) from the Weighted OLS Regression Model of
Marital Status and Disruption Pathway on Ambivalence toward Children ...... 112
4.4 Coefficients (Standard Errors) from the Weighted OLS Regression Model of
Marital Status on Ambivalence toward Children ...... 114
5.1 Summary of Results for Marital Biography and Depressive Symptoms/Chronic Conditions
by Gender ...... 135
5.2 Summary of Results for Marital Biography and Parent-Child Contact by Gender
...... 136
5.3 Summary of Results for Marital Biography and Ambivalence by Gender ...... 137
1
CHAPTER 1: INTRODUCTION
Older adulthood in the United States has been marked by notable shifts in marriage during recent decades. In the past, many people shared the narrative of entering into a marriage and remaining in that union until the death of one spouse (Cherlin, 2010). However, dissolution of a first marriage through widowhood is no longer the modal pathway out of marriage. Today, a greater share of all marriages involving someone over 50 are remarriages (Lin, Brown, & Cupka, in press). In 1980, 19% of older married adults were in higher order marriages and by 2015 the share had climbed to 30%, indicating that many people experienced either divorce or widowhood during early or midlife followed by a subsequent marriage. Marital disruption is not only common in earlier life, but widowhood and divorce increasingly punctuate the lives of older people. Widowhood remains a common exit from marriage in later life (Brown, 2017; Manning
& Brown, 2011). Divorce to people over 50– called gray divorce— is also on the rise. In 1990, only 1 in 10 divorces involved someone over 50, but by 2010, this number had risen to 1 in 4, driven in part by the dissolution of higher order marriages (Brown & Lin, 2012).
Some older people who are widowed or divorced go on to remarry (Brown, Lin,
Hammersmith, & Wright, in press). It is crucial to account for remarriages in earlier and later life, especially as remarriages are more vulnerable to disruption than first marriages (Brown &
Lin, 2012; Sweeney, 2002), leading to the possibility of more instances of marital disruption over the life course. Higher order marriages are shorter in duration than first marriages, also contributing to greater marital instability (Wu & Brown, 2016). Further, a growing number of older people are unmarried whether through never marrying, divorce, or widowhood (Manning
& Brown, 2011; Lin & Brown, 2012). Understanding the extent to which such marital biography characteristics are associated with well-being in later life is critical, especially as nearly 21% of 2
the United States is projected to be 65 and older by 2030 (Cherlin, 2010; Older Americans: Key
Indicators of Well-being, 2016).
Marital biography is an increasingly common strategy implemented by researchers to
capture marital instability over the life course. Marital biography encapsulates current marital
status, prior disruption, duration, age at first marriage, and sequencing (Zhang, Lin, & Yu, 2016).
This dissertation will focus on current marital status, marital disruption, and duration as these
factors are most pertinent in studies of well-being in later life (Hughes & Waite, 2009; Zhang et
al., 2016). Although prior studies have examined how aspects of marital biography are associated
with a host of well-being indicators, research to date has neglected to fully consider whether
remarriage is negatively related to the costs of different disruption pathways (i.e., one divorce,
one widowhood, and multiple disruptions) as they relate to mental and physical health, parent-
child contact, and ambivalence toward children. Moreover, few studies have accounted for the role of duration either remarried or unmarried after different disruption pathways for well-being.
This dissertation focuses on older adults’ well-being manifested in three ways: mental and physical health, parent-child contact, and parent’s ambivalence toward children. It is important to pay attention to the relationship between marital biography and various aspects of well-being in later life for several reasons. First, poor mental and physical health may positively relate to the amount of support people need as they grow older. Along with this, contact with children may be positively associated with support availability for older people and lower ambivalence is likely indicative of older peoples’ willingness to ask social network members, like their adult children, for assistance (Bengston & Roberts, 1991; Mancini & Blieszner, 1989).
Second, well-being in later life is associated with events that occurred from earlier life all the way into later life (Dannefer, 2003; Elder, 1994; O’Rand, 1996), including those related to 3
marital biography. Divorce, widowhood, or multiple disruptions are associated with changes in
health or parent-child relationships. Divorce is often associated with poorer health and parent-
child ties while widowhood is disruptive for health but supportive of parent-child relationships
(Connidis, 2010; Daatland, 2007; Ha & Ingersoll-Dayton, 2008; Pezzin & Schone, 1999).
Remarriage likely plays a role when considering well-being in later life. On the one hand,
remarriage relates to recovery of health-promoting resources following widowhood or divorce
(Barrett, 2000; Hughes & Waite, 2009). On the other hand, remarriage further weakens parent-
child ties, and thus, is associated with less parent-child contact and greater ambivalence toward
children (Kalmijn, 2007, 2013, 2015; Noël-Miller, 2013; Pezzin & Schone, 1999).
Duration either remarried or unmarried should also be accounted for when investigating
well-being in later life. In this dissertation, duration reflects the accumulation of years in the
current marital status and how the dimension of time is related to well-being. Specifically,
accumulating years in a disrupted status could positively link to long-term financial and
emotional stress, especially among individuals after multiple disruptions who face added strain
from more than one widowhood or divorce (Williams & Umberson, 2004; Zhang et al., 2016).
Alternatively, years unmarried could allow older people to adjust to divorce or widowhood
(Williams & Umberson, 2004). Relative to being unmarried, duration remarried may allow for
accumulation of resources that protect health, however, duration remarried could contribute to
increasingly weakened parent-child ties, ultimately leading to less contact and more ambivalence
(Dykstra & Fokkema, 2011; Kalmijn, 2007, 2015; Umberson, 1992). Thus, incorporating
dimensions like current marital status, marital disruption, and duration can help researchers better understand the relationship between accumulated transitions and instability in marital
biography over the life course and well-being in later life. 4
Informed by prior work focused on the association between marital biography and well-
being, this dissertation aims to examine whether remarriage following different disruption pathways and duration remarried or unmarried relate to: (1) mental and physical health, (2) parent-child contact, and (3) parent’s ambivalence toward children. Finally, in addition to investigating the aforementioned aspects of well-being, each chapter will consider gender differences in the association between marital biography and health, parent-child contact, or ambivalence. The experiences of men and women are often diverse in later life especially as they relate to health and parent-child relationships (Daatland, 2007; Kalmijn, 2007; Umberson, 1992;
Zhang et al., 2016). Thus, it is crucial to investigate how marital biographies of men and women separately relate to their respective well-being.
To meet the aforementioned dissertation goals, this project draws on data from the 1992 through 2014 waves of the Health and Retirement Study. The first analytic chapter focuses on mental and physical health, operationalized as depressive symptoms and chronic conditions.
Entrance into marriage promotes the accumulation of health-supporting resources, whereas marital disruption is positively associated with stress that jeopardizes health (Barrett, 2000;
Umberson, 1992; Zhang et al., 2016). The benefits of marriage and costs of disruption can accumulate over time, either supporting or eroding health (Barrett, 2000; Marks & Lambert,
1998; Williams & Umberson, 2004; Zhang & Hayward, 2006). Many studies conclude that relative to being unmarried, remarriage supports health, although not to the same extent as a first marriage (Barrett, 2000; Dupre & Meadows, 2007). Further, other researchers have investigated disruption pathways (i.e., one divorce, one widowhood, and multiple disruptions), but have neglected to compared different disruption pathways to one another among the remarried and unmarried (Hughes & Waite, 2009). Thus, prior studies have not considered how remarriage 5
relative to being unmarried is associated with fewer financial and emotional costs of different disruption pathways even though individuals after multiple disruptions likely face compounded stress relative to those after one divorce or one widowhood. This chapter also investigates duration remarried or unmarried as years remarried may be associated with better or worse health relative to being unmarried. However, the benefits of remarriage or costs of being unmarried over time could vary based on different disruption pathways. Thus, I account for the role of remarriage after different disruption pathways and duration remarried or unmarried on mental and physical health.
The second analytic chapter considers the relationship between marital biography and parent-child contact. Parent-child contact is a fundamental pillar of the intergenerational solidarity framework as contact between parents and children indicates support avenue strength for older people (Bengston & Roberts, 1991; Mancini & Blieszner, 1989; Silverstein &
Bengston, 1997). Still, parent-child contact is not necessarily stable over time, and has been linked to particular marital biography changes such that contact is positively related to widowhood, but negatively associated with divorce or remarriage (Daatland, 2007; Kalmijn,
2007, 2015; Roan & Raley, 1996). Despite evidence from prior research, it remains unknown whether different disruption pathways and subsequent remarriage take a cumulative toll on parent-child contact. It is also unclear whether duration remarried or unmarried after different disruption pathways are positively or negatively associated with parent-child contact. On the one hand, relative to years unmarried, years remarried may allow time for the parent to rebuild relationships with children following marital disruption. On the other hand, duration remarried could be associated with weakened ties with children relative to being unmarried as remarriage necessitates a renegotiation of parent-child ties (Kalmijn, 2007). Therefore, this chapter adds to 6
current research by investigating the role of remarriage following different disruption pathways
as well as duration remarried or unmarried on parent-child contact.
The final analytic chapter examines the association between marital biography and
parent’s ambivalence toward children. Although parents and children share close bonds over the
life course, parent-child ties often change following parent’s marital transitions like widowhood,
divorce, or remarriage (Connidis, 2010; Ward, Deane, & Spitze, 2014). Thus, parents may feel
more or less ambivalence—the comingling of positivity and negativity—after different marital
biography events. Parent-child ties become less ambivalent following widowhood as
bereavement unites parents and children in mutual grief (Ha & Ingersoll-Dayton, 2008). Divorce
and remarriage may lead to greater ambivalence as both necessitate renegotiation of parent-child
ties, suggesting parents and children remain close, but experience more conflict (Connidis,
2010). This chapter builds upon past research by investigating how remarriage relative to being
unmarried after different disruption pathways is associated with parent’s ambivalence. Some
researchers have examined the role of widowhood, divorce, and subsequent remarriage on
parent-child ties, but few have incorporated ambivalence. This omission is notable especially in light of the negative association between ambivalence and health in later life (Lendon,
Silverstein, & Giarusso, 2014; Willson, Shuey, Elder, & Wickrama, 2006). This project also builds upon prior work through considering how duration remarried or unmarried following different disruption pathways is associated with parent’s ambivalence toward children. Years remarried may be positively related to parent-child conflict, thereby contributing to a positive relationship between duration remarried and ambivalence relative to the unmarried. Conversely, years remarried may allow for parents and children to rebuild their relationship, through which they likely feel less ambivalence. 7
In each chapter, I pay attention to gender differences. Women often assume the role of
nurturer in romantic partnerships, providing care and monitoring their spouse’s health behaviors
(Scommegna, 2016; Umberson 1992). Women are also described as kin-keepers, and thus, they
maintain the social networks of themselves and their spouse (Rosenthal, 1985; Zhang et al.,
2016). Therefore, instability in marital biographies of men may relate to poorer health, less parent-child contact, and greater ambivalence relative to women. For this reason, I factor in gender differences for each well-being outcome.
The key contributions of this dissertation to current literature on marital biography are twofold. First, this research extends prior work by considering remarriage after different disruption pathways for a host of well-being outcomes. Prior work has yet to investigate how remarriage is positively or negatively linked to well-being consequences related to different disruption pathways. Moreover, although prior research has examined duration in the current marital status, few studies have accounted for how duration remarried or unmarried after different disruption pathways relates to mental and physical health, parent-child contact, and ambivalence toward children. Among older people with different disruption pathways, it is crucial to understand whether years remarried relative to being unmarried are a detriment or a benefit to health and parent-child ties. Given a growing population of older adults with increasingly complex marital biographies (Cherlin, 2010; Older Americans: Key Indicators of
Well-being, 2016), it is pivotal to unpack the influence of remarriage, different marital disruption pathways, and duration remarried or unmarried on the well-being of older people. Through this, researchers can better devise policy initiatives and strategies to support aging populations.
8
CHAPTER 2: MARITAL BIOGRAPHY AND HEALTH IN LATER LIFE
Older adulthood has been accompanied by noteworthy demographic change in recent
decades, especially regarding mid to late-life marital disruption and remarriage. In the past,
people would enter marital unions and remained married until the death of one spouse (Cherlin,
2009). Today higher order marriages are more common. In 2015, nearly 30% of married people over 50 were in remarriages compared to 19% in 1980 (Lin, et al., in press). Thus, the share of
adults entering mid to late-life with at least one marital disruption is growing, and many enter a
subsequent remarriage. In fact, more than 20 million older adults are in a higher order marriage
(Lin et al., in press). Recent studies show the prevalence of remarriage, but remaining
unmarried— through divorce, widowhood, or never marrying— is also commonplace (Manning
& Brown, 2011; Lin & Brown, 2012).
To accommodate shifting demographic trends, scholars have looked beyond current marital status by examining the entire marital biography. Marital biography encompasses current marital status, marital disruption, duration, timing, and sequencing (Brown & Wright, 2017;
Hughes & Waite, 2009; Zhang et al., 2016). Not all of these components are equal and findings from prior studies suggest that current marital status, marital disruption, and duration are the most salient predictors of health in later life (Barrett, 2000; Dupre & Meadows, 2007; Hughes &
Waite, 2009; Lillard & Waite, 1995; Zhang, 2006; Zhang & Hayward, 2006).
Scholarly work examining the relationship between current marital status and health shows marriage is associated with better health as marriage is associated with a host of protective resources (Umberson, 1992). In contrast, being unmarried—especially through divorce or widowhood—is detrimental for well-being as marital disruption is associated with financial, emotional, and social stressors that erode health (Liu & Umberson, 2008; Umberson, 1992). 9
Although studies show disruption is harmful for health, scholars have also found remarriage after
disruption promotes health through offering some of the benefits of marriage, even though these
are not as strong as the first marital union (Barrett, 2000; Dupre & Meadows, 2007; Marks &
Lambert, 1998; Williams & Umberson, 2004). Despite this, research has not fully considered
whether older adults reap fewer benefit from remarriage after multiple disruptions compared to
just one disruption (Hughes & Waite, 2009; McFarland, Hayward, & Brown, 2013; Reczek,
Pudrovska, Carr, Thomeer, & Umberson, 2016; Zhang, 2006; Zhang & Hayward, 2006).
Moreover, prior work has demonstrated the importance of duration in current status (Barrett,
2000; Lillard & Waite, 1995; Zhang & Hayward, 2006), yet only a handful of studies examine
how the benefits of duration remarried or the detriments of duration unmarried as the result of
marital disruption may vary depending on different disruption pathways (i.e., one divorce, one
widowhood, or multiple disruptions).
Using Health and Retirement Study data, I propose three research questions. First, among
those who experience marital disruption, is remarriage related to fewer negative consequences of
different disruption pathways on mental and physical health? Second, is duration remarried or
unmarried after disruption related to better or worse health following different disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions)? Finally, are there gender differences in the relationship between remarriage and the costs associated with marital disruption as well as in whether duration remarried or unmarried after disruption is associated
with better or worse health outcomes following different disruption pathways?
This chapter makes several contributions. Studies on the relationship between marital biography and health have focused on marital biography for particular segments of the adult life course (Reczek et al., 2016; Zhang & Hayward, 2006) or on particular birth cohorts (Henretta, 10
2010; Hughes & Waite, 2009; Zhang, 2006). This chapter extends prior work by evaluating
marital biography for the entire adult life course to capture marital characteristics for all cohorts
aged 50 and older. Recent demographic shifts, like the rise of gray divorce and the prevalence of
remarriage suggest evaluating the entire marital biography can provide new insight regarding
health in later life (Brown & Lin, 2012; Brown & Wright, 2017; Hemez & Brown, 2016).
Second, I extend prior work by examining whether remarriage and duration relate to fewer health
problems related to different disruption pathways. Scholars have examined how remarriage
compares to first marriage, but have largely ignored comparisons of remarried and unmarried
groups with different disruption pathways (Dupre & Meadows, 2009; Hughes & Waite, 2009;
Zhang 2006; Zhang & Hayward, 2006; Zhang et al., 2016). Further, studies have investigated
duration, but neglected to pay attention to the role of years remarried or unmarried after different disruption pathways. Last, I incorporate gender, adding to knowledge of how remarriage after disruption and duration remarried or unmarried variably link to men’s and women’s health
Marital Resource, Stress, and Cumulative (Dis)Advantage Models
Prior studies have best understood the relationship between marital biography and health through two theoretical perspectives: the marital resource model and the stress model. The marital resource model suggests marriage promotes health through financial, emotional, and social resources (Carr & Springer, 2010; Hughes & Waite, 2009; Umberson, 1992; Zhang et al.,
2016). The married are more financially stable than the unmarried (Umberson, 1992; Wilmoth &
Koso, 2002), as married couples are able to pool their resources and reap the institutional benefits of marriage (Becker, 1981; Cherlin, 2009). Married couples are also less socially isolated than the unmarried and receive more support from their spouse, friends, family, and children (Albertini & Garriga, 2011; Umberson & Montez, 2010). Marital partners are another 11 health-promoting resource, especially as many married—whether in a first marriage or a remarriage—people report their spouse monitors their health behaviors through encouraging doctor’s visits and healthy habits (Liu & Umberson, 2008; Umberson, 1992).
In contrast, the stress model posits disruption—through divorce or widowhood— is one of the most stressful events experienced over the life course (Umberson, Thomeer, & Williams,
2013; Zhang et al., 2016). Disruption creates financial, social, and emotional challenges. Divorce and widowhood are detrimental for the financial well-being of older people, although divorce is associated with more financial difficulties than widowhood (Carr & Springer, 2010; Lin et al.,
2017; Wilmoth & Koso, 2002). Disruption interferes with social support, and this is especially true among the divorced. Divorced and widowed older adults lose social support provided by their former spouse, moreover, the divorced suffer from poorer relationships with other members of their social network, like children (Albertini & Garriga, 2011; Kalmijn, 2013; Noël-Miller,
2013). Some of the divorced and widowed go on to remarry (Brown et al., in press). Although not as beneficial for health as a first marriage, remarriage is associated with fewer health costs related to disruption, but this may vary depending on whether the disruption was through divorce, widowhood, or multiple disruptions (Barrett, 2000; Dupre & Meadows, 2007; Marks &
Lambert, 1998; Williams & Umberson, 2004).
The marital resource model and the stress model are useful for explaining the benefits of marriage coupled with negative consequences of disruption, but the cumulative advantage/disadvantage perspective can be instrumental when considering the role of duration on health in later life (Zhang et al., 2016). Variation in marital biography is accrued over the entire adult life course. Thus, cumulative advantage/disadvantage perspective allows for a fuller understanding of whether duration remarried after disruption relates to better health outcomes 12 and if duration unmarried after disruption is associated with better or worse health (Dannefer,
2003; O’Rand, 1996; Zhang et al., 2016). It is possible years in a remarriage are advantageous as each year may be associated with less stress following a disruption. Similarly, duration unmarried after disruption may be related to better health as stressors related to disruption could dissipate with years unmarried. However, duration unmarried could also be associated with poorer health over time as the strain associated with disruption accumulates with years in status
(Barrett, 2000; Hughes & Waite, 2009; Zhang & Hayward, 2006).
Despite the accumulation of benefits that relate to marriage as well as consequences associated with divorce or widowhood, researchers cannot ignore the role of selection. That is, individuals with better health and greater financial resources are more likely to marry than their counterparts who have fewer of these advantages (Amato, 2010). Similarly, those who are in better health and who are wealthier are also more likely to stay married while individuals with fewer resources may be more likely to experience divorce or widowhood (Amato, 2010; Johnson
& Wu, 2002). Thus, older people with one divorce, one widowhood, or multiple disruptions may have had poorer health prior to their disruption. Correspondingly, entering a remarriage relative to remaining unmarried could be associated with already having better health and greater resources prior to that remarriage. Nonetheless, some studies have accounted for selection by controlling for prior health status and still found that divorce is negatively associated with health whereas gaining a partner is related to better health (Blekesaune, 2008; Wu & Johnson, 2002;
Wade & Pevalin, 2004).
Marital Status and Health
Marital status shares a notable link to health (Barrett, 2000; Brown, Bulanda, & Lee,
2005; Hughes & Waite, 2009; Liu, 2012; Liu & Johnson, 2009; Liu & Umberson, 2008; Rendall, 13
Weden, Favreault, & Waldron, 2011; Umberson, 1992). In general, partnereds report better health when compared to all unmarried older adults (Brown & Wright, 2017; Carr & Springer,
2010; Hughes & Waite, 2009; Wright, 2017); however, there is variation among the partnered.
First married individuals are healthier than remarrieds or cohabitors due to their greater financial, emotional, and social resources (Umberson, 1992). Research has been limited concerning the health of older cohabitors compared to other groups, but recent work suggests that the remarried and cohabitors enjoy similar health (Brown & Wright, 2017; Wright, 2017). Although remarriage does not provide the same health benefits as first marriage, the remarried still reap some financial, social, and emotional benefits from their subsequent marriage, yielding a health advantage relative to the unmarried (Barrett, 2000; Cherlin, 2004; Wilmoth & Koso, 2002).
Among the unmarried, the health of never marrieds is similar to the partnered and better than the divorced and widowed as the never married have not experienced stress associated with marital disruption (Hughes & Waite, 2009; Liu & Johnson, 2009; Liu & Umberson, 2008). The widowed report better health than the divorced as they preserve more financial resources and stronger ties to their social network than the divorced (Ha, 2008; Lin, 2008; Lin, Brown, & Hammersmith,
2017). Last, prior work shows mental health is more sensitive to status changes, like those related to current marital status than physical health. Specifically, there are long-term physical health benefits associated with occupying a particular marital status, making physical health less sensitive to status change. Conversely, mental health fluctuates more in light of a status change, as transitions in marital status often encompass an emotional reaction, which is subject to quicker changes than physical health responses (Hughes & Waite, 2009; Marks & Lambert, 1998;
Simon, 2002).
14
Remarriage, Disruption Pathways, and Health
A greater share of married older adults are in a higher order marriage than in past decades
(Lin et al., in press). Investigating current marital status alone masks growing complexity in
marital biographies of older people by not considering remarriage and different disruption
pathways. Prior studies indicate remarriage is beneficial for health (Barrett, 2000; Marks &
Lambert, 1998; Williams & Umberson, 2004), but this line of research is limited as it does not study whether remarriage is associated with fewer negative health consequences following divorce, widowhood, or multiple disruptions. Similarly, Reczek and her colleagues (2016) examined alcohol use among an array of marital groups. They found remarriage was differentially associated with alcohol consumption among men and women such that remarriage was negatively related to men’s alcohol usage, but positively related to women’s. Despite the contributions of this work, this study did not factor in different disruption pathways prior to remarriage.
McFarland and his colleagues (2013) as well as Dupre and Meadows (2007) distinguished number of disruptions prior to remarriage. McFarland (2013) et al. did not find an association between number of disruptions prior to remarriage and health. However, Dupre and
Meadows (2007) showed multiple disruptions harm health, especially among men. Even though these studies account for the role of multiple disruptions on health, they are limited in a couple of ways. First, these studies grouped all disruptions into the same category, thus, not distinguishing between one divorce and one widowhood. Second, these studies did not consider how different disruption pathways compare to one another (e.g., how multiple disruptions compare to one divorce or one widowhood). 15
A few other researchers have examined both number of disruptions as well as different disruption pathways among the remarried, and those that do primarily compare individuals who have remarried to those in a first marriage (Hughes & Waite, 2009; Zhang, 2006). These studies showed disruption pathway matters such that multiple marital disruptions are detrimental for overall health relative to individuals in a first marriage. This line of research adds to knowledge of how remarriage and disruption pathways compare to those in a first marriage, but it is limited by not drawing comparisons between remarried and unmarried among older adults with different disruption pathways (Zhang et al., 2016). Barrett (2000) compared remarrieds and unmarrieds with a variety of disruption pathways and showed that remarriage provided some health benefits compared to remaining unmarried. Moreover, disruption type did not play a notable moderating role in the relationship between current marital status and mental health. Even though the findings from this study make a contribution to current research, this study drew on a wide age range (from 30 to 95), making it unclear whether the same patterns would be observed among older adults.
Mental and physical health outcomes may be differentially related to remarriage or remaining unmarried following different disruption pathways. Prior researchers have shown that advantages or disadvantages, such as a lack of financial or health related resources, are associated with transitions such as remarriage or marital disruption will be more strongly linked to changes in physical rather than mental health (Hughes & Waite, 2009; Singer & Ryff, 1999).
This is because physical health conditions develop more slowly and reflect one’s life history of accumulated resources (or conversely, experience the loss of resources) (Hayward, Crimmins,
Miles, & Yang, 2000). Thus, remarriage may confer resources that promote physical health 16 whereas marital disruption likely relates to loss of resources, which is detrimental for physical health.
In sum, based on marital biography research that considers the relationship between remarriage, marital disruptions, and health, I propose several hypotheses. Remarriage is beneficial for health and well-being as remarriage is associated with the accumulation of financial and emotional resources and allows for a new avenue of support. I expect that following marital disruption, the remarried will report better health than the unmarried regardless of disruption pathway (Barrett, 2000; Wilmoth & Koso, 2002). Experiencing multiple disruptions creates greater emotional and financial strain (Williams & Umberson, 2004). Thus, it is likely after multiple disruptions, older adults will report more depressive symptoms and chronic conditions relative to individuals after either one widowhood or one divorce. Moreover, individuals with one widowhood will be advantaged relative to the divorced. Although stressful, widowhood leaves the bereaved spouse with more financial and social resources that are protective of health (Ha, 2008; Lin, 2008; Lin et al., 2017). Last, I expect remarriage after marital disruption will be more strongly associated with physical rather than mental health conditions.
Duration Remarried, Duration Unmarried, and Health
Solely examining marital status obscures the role of duration, which could be associated with better or worse health with years following disruption. Lillard and Waite (1995) found that duration in the first marriage was beneficial for overall health, but this study ignored duration remarried or unmarried. Factoring in duration remarried is critical, as years remarried may allow older people to accrue more resources related to overall health. Moreover, duration unmarried 17
likely relates accumulate stress and loss of resources or alternatively, is associated with time to
recover from strain associated with disruption.
Zhang and Hayward (2006) examined duration in current marital status and found that
duration divorced or widowhood had a small, but positive association with the incidence of
cardiovascular disease. Although this study discerned the role of years divorced or widowed, this
work neglected to look at duration following multiple disruptions, which could introduce more
stress as years accumulate. Barrett (2000) investigated the link between duration remarried and
unmarried while also factoring in various disruption pathways. She found differences in anxiety
among the widowed with different disruption pathways such that twice widowed individuals
proved more resilient and their mental health recovered over time relative to individuals after one
divorce or one widowhood. Despite the contributions of this study, it encompassed a much wider
age range (ages 30 to 95), thus, including may younger adults for whom widowhood would be
less common than for older adults.
Duration remarried or unmarried following disruption may also exhibit variable
associations with mental versus physical health. Physical ailments develop slowly over a long
period of time often due to the loss of resources that are protective of health (Hughes & Waite,
2009; Singer & Ryff, 1999). Thus, duration remarried following disruption may permit time to
accrue resources that are protective of physical compared to mental health. In a similar fashion,
on the one hand, duration unmarried after disruption may allow for recovery of resources that
promote physical health versus mental health. On the other hand, duration unmarried could be
associated with poorer physical health relative to mental health over time as older adults continue to experience the loss of resources related to marital disruption. 18
Building upon prior work that examines the role of duration remarried or unmarried
following marital disruptions on health, I put forth several predictions. First, duration remarried
after marital disruption may allow for recovery of financial resources as well as social and
emotional support, thereby promoting good health. Second, duration unmarried after disruption
could be associated with better or worse health depending on the disruption pathway (Williams
& Umberson, 2004). On the one hand, financial, social, and emotional stress may be
compounded with duration unmarried, thus, harming health. On the other hand, some evidence
shows that duration unmarried after widowhood is associated with better mental health (Barrett,
2000). Even if greater duration unmarried mitigates the stressors associated with marital
disruption, I still expect years remarried will be advantageous over years unmarried, as
remarriage is associated with accumulating more resources relative to remaining unmarried
(Barrett, 2000; Marks & Lambert, 1998; Williams & Umberson, 2004). It is likely individuals after multiple disruptions will be slower to recover than those who have had one divorce or one
widowhood, as additional disruptions are likely linked to greater financial, social, and emotional
strain compared to one divorce or one widowhood. Finally, I predict duration remarried or
unmarried after disruption will have a stronger relationship with the development of physical
ailments than mental health conditions.
Variation by Gender
The benefits of marriage are more supportive of men’s than women’s health. Women are
often nurturers in romantic relationships such that women monitor men’s health behaviors and
tend to serve as caregivers for their ailing husbands (Scommegna, 2009; Umberson, 1992).
Women typically bring more social resources to a marriage than do men, which also promote
health and well-being of men relative to women (Zhang et al., 2016). Furthermore, researchers 19
have shown that men benefit more from remarriage than do women, especially when considering
health in later life (Gardner & Oswald, 2004; Johnson, Backlund, Sorlie, & Loveless, 2000;
Reczek et al., 2016), whereas remaining unmarried is less detrimental for women than men
(Amato & Rodgers, 1997; Hu & Goldstein, 1990).
The relationship between marital disruption and health among men and women is less
clear. Some studies have found marital disruption, particularly through divorce, is more
detrimental for women than men (Brockmann & Klein, 2004; Dupre, Beck & Meadows, 2009;
Dupre & Meadows, 2007). However, other studies show health risks for men are greater than
those for women after disruption (Reczek et al., 2016; Shor, Roelfs, Bugyi, & Schwartz, 2016;
Williams & Umberson, 2004). Thus, based on this conflicting evidence, it is unclear whether
different disruption pathways and subsequent remarriage will matter more for the health of men
compared to women.
I expect remarriage will be associated with fewer consequences of marital disruption for men relative to women, especially as men tend to benefit more from the institution of marriage than their female counterparts (Reczek et al., 2016; Scommegna, 2016; Umberson, 1992; Zhang et al., 2016). However, given conflicting evidence regarding the relationship between men and women’s prior marital disruptions and health, it is unclear whether different disruption pathways will be more consequential for the health of men compared to women. Finally, due to limited past research, I make no prediction about whether duration remarried or unmarried after different disruption pathways will be more consequential for the health of women relative to men.
The Present Study
Marriage and partnership are beneficial for health (Liu & Umberson, 2008; Umberson,
1992; Zhang et al., 2016). At the same time, being unmarried, especially divorced or widowed, is 20
associated with poorer health (Williams & Umberson, 2004). Incorporating marital biography
extends this knowledge to look at the role of other factors related to marriage when considering
health in later life, like marital disruption pathways and duration. Marital disruption relates to
stress, which harms health (Dupre & Meadows, 2007; Hughes & Waite, 2009; Williams &
Umberson, 2004). However, a number of older adults remarry, and studies show remarriage
promotes health although not to the same extent as a first marriage (Barrett, 2000; Brown et al.,
in press; Dupre & Meadows, 2007; Hughes & Waite, 2009; Marks & Lambert, 1998; McFarland
et al., 2013; Reczek et al., 2016; Williams & Umberson, 2004). We know little about how
remarriage relates to negative health consequences, especially when considering different
disruption pathways through either one divorce, one widowhood, or multiple disruptions. Prior
work has also accounted for how current marital status duration relates to health (Barrett, 2000;
Dupre & Meadows, 2007; Lillard & Waite, 1995). Still, prior research has not looked into the
role of duration remarried after disruption for mitigating health problems associated after one divorce, one widowhood, or multiple disruptions. Similarly, research has not examined whether duration unmarried may be positively or negatively related to health outcomes following different disruption pathways.
To fill these gaps in the literature, I propose the following hypotheses to answer the three research questions presented in this chapter. First, among those who experience marital disruption, is remarriage associated with fewer consequences following disruption and what is the role of different disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions)? Second, is duration remarried related to fewer health consequences among individuals with different disruption pathways? Similarly, is duration unmarried after disruption associated with greater or fewer health costs related to different disruption pathways? Next, are 21
there gender differences in whether remarriage is related to fewer negative consequences of
marital disruption and in whether duration remarried or unmarried is associated with better or
worse health outcomes related to disruption?
Hypothesis 1a: Given the same disruption pathway, older adults who remarry after
disruption will report better mental and physical health than those who remain unmarried
following disruption.
Hypothesis 1b: Among the remarried and unmarried, older adults after one widowhood
will have the best health compared to other disruption pathways. Moreover, older adults after one
divorce will report better health than individuals after multiple disruptions.
Hypothesis 2a: Duration remarried will be advantageous relative to duration unmarried regardless of disruption pathway, while it is unclear whether duration unmarried will be deleterious or beneficial for health.
Hypothesis 2b: Whether duration is spent remarried or unmarried, older adults after multiple disruptions will have worse health with more years in status than older adults after one divorce or one widowhood.
Hypothesis 3a: Remarriage will be associated with fewer costs of disruption for men than for women. It is unclear whether different disruption pathways will be more consequential for the health of men relative to women (through one divorce, one widowhood, or multiple disruptions).
Hypothesis 3b: The relationship between duration remarried or unmarried after different disruption pathways for men relative to women remains unclear, and thus, no prediction is made in either direction. 22
Hypothesis 4: Current marital status will relate more strongly to depressive symptoms whereas being remarried or unmarried following different disruption pathways and duration remarried or unmarried will be tied more strongly to chronic conditions.
In this study, I accounted for several covariates that may confound the association between marital biography and health such as demographic characteristics and socioeconomic resources.
Demographic Characteristics
Age yields more time to experience marital instability or to accrue years in one’s current marital status (Cherlin, 2009; Goldstein & Kenney, 2001) and there are age differences in the association of marital status with health (Dannefer, 2003; Liu, 2012). Blacks report worse mental and physical health than Whites. Compared to Whites, Hispanics have lower mortality rates and better health, whereas the health of Blacks is poor, comparatively (Brown, O’Rand, & Adkins,
2012; Franks, Gold, & Fiscella, 2003). Blacks are less likely to marry than Whites (Manlove et al., 2017). When married, Whites reap greater health benefits from marriage than do Blacks
(Broman, 1993; Bulanda & Brown, 2007). Blacks are more likely to experience disruption, resulting in more marital instability than Whites (Heaton, 2002). Whereas Blacks endure more marital instability, Hispanics have similar marital stability as Whites (Amato, 2010; Bramlett &
Mosher, 2002).
Socioeconomic Resources
The well-educated enjoy more health-promoting resources than the less-educated (Franks et al., 2003; Ross & Wu, 1995). Education also relates to partnership stability (de Graaf &
Kalmijn, 2006; Heaton, 2002); higher education is associated with greater stability whereas less education is related to a higher incidence of marital disruption (Amato, 2010; Heaton, 2002). 23
Similarly, greater wealth translates into better mental and physical health as wealth encourages access to health-promoting resources (Franks et al., 2003; Miech & Shanahan, 2000). The wealthy also enjoy greater marital stability (Heaton, 2002). The wealthy report fewer marital disruptions and accrue more years in marriage than people with fewer socioeconomic resources
(Amato, 2010; Heaton, 2002). Moreover, different disruption pathways bear variable relationships with wealth. Divorce or widowhood are each associated with financial strain from which older people have difficulty recovering, although the widowed are often better off than the divorced (Hungerford, 2001; Lin et al., 2017). Finally, employment relates to better health.
People in worse health are more likely to retire than individuals in better health (Flippen &
Tienda, 2000). Employment also relates to marital biography. Retirement often brings about marital conflict, and, thus, exiting the labor force could be associated with marital disruption, although this has not received consistent support in the literature (Dew & Yorgasen, 2010; Lin et al., 2017).
Method
I employed data from the Health and Retirement Study (HRS). The HRS sample is a nationally representative, continuous cohort of participants that commenced in 1992. To be eligible for the HRS, respondents must be noninstitutionalized at baseline. Respondents are reinterviewed every two years. Every six years, the HRS incorporates a new cohort of older adults aged 51 to 56 to replenish the sample and maintain representativeness. The response rates for the HRS are relatively high; baseline rates hover around 70-82% and increase to roughly 90% or higher for follow-up interviews. The HRS oversamples of Blacks, Hispanics, and Floridians.
The data used to construct marital biography were from the 1992 through 2014 waves, whereas current marital status and the health measures as well as other covariates were from the 24
2014 core survey. I began sample selection by employing a file synthesizing marital information for all respondents from 1992 to 2014. The initial HRS sample included 37,495 respondents. I selected participants in the 2014 interview, yielding 18,748 respondents. I restricted respondent’s age to 50 and older (n = 18,332) in 2014. I removed nursing home residents, yielding 17,272 older adults. Last, I removed seven cases with missing information on gender and 205 cases in which marital biography cannot be determined, leaving an analytic sample of 17,060 respondents, of whom 7,127 were men and 9,933 were women.
Dependent Variables
Depressive Symptoms. Mental health was captured through depressive symptoms. I summed eight items in which the respondent reported whether they experienced the following symptoms during the week prior to the interview: feeling depressed, everything was an effort, restless sleep, unhappiness, loneliness, not enjoying life, sadness, or unable to get going (0 = No,
1 = Yes). Depressive symptoms ranged from 0 to 8 and the inter-item reliability was .80.
Chronic Conditions. Physical health measured the respondent’s number of chronic conditions as diagnosed by a physician, including psychiatric illness, diabetes, heart disease, cancer, lung disease, hypertension, stroke, and arthritis (1 = Yes, 0 = No). Chronic conditions ranged from 0 to 8.
Focal Independent Variables
Marital Status. Marital status included six categories: married, remarried (reference category), cohabiting, divorced, widowed, or never married in 2014.
Disruption Pathways. Marital disruptions were used to construct pathways through which individuals arrived at their current marital status. This measure included the following categories: first married, remarried after one divorce (reference category), remarried after one 25
widowhood, remarried after multiple disruptions, cohabiting, unmarried after one divorce,
unmarried after one widowhood, unmarried after multiple disruptions, and never married.
Among respondents with multiple disruptions, 63% had two or more divorces, 23% had one divorce and one widowhood, 5% had two or more widowhoods, and 9% reported three disruptions that included both divorces and widowhoods (e.g., two divorces and one widowhood). Respondents who were remarried after multiple disruptions made up a little over
5% of the full sample whereas the unmarried after multiple disruptions represented about 9% of
the full sample.
Duration Remarried or Unmarried. Duration was a continuous measure that captured
years currently remarried or unmarried following the most recent marital disruption.
Covariates
Age was measured using three categorical variables coded into age 50 to 64, ages 65 to
84 (reference category), and ages 85 and older. Race and ethnicity comprised a categorical variable indicating whether the respondent identified as White (reference category), Black,
Hispanic, or of other races. Education was coded into four categories: less than high school, high
school (reference category), some college, and college or more. The distribution of household
assets was skewed so I employed a series of five categories to capture household wealth: in debt,
$0 to $50,000 (reference category), $50,001 to $100,000, $100,001 to $250,000, and $250,001 or
more. Work measured whether the respondent was in the labor force or not. Number of children
was a continuous measure of the respondent’s living children.
Analytic Strategy
I conducted two sets of analyses. First, I reported weighted descriptive statistics for
depressive symptoms and chronic conditions, each component of marital biography, and all other 26
covariates for men and women separately. Second, I ran a series of three weighted ordinary least
squares regression models of (1) depressive symptoms and (2) chronic conditions on various
components of marital biography. The first model examined the relationship between marital
status and health. The second model added to this by accounting for remarriage and different
disruption pathways. Next, I added in duration remarried or unmarried to the model. Interactions
between duration and each marital status category were included in the model although only
relevant comparisons were displayed in Table 2.4 (complete results available upon request). I
also tested for gender differences in all models by estimating a pooled model with interactions
between gender and each marital biography characteristic. Finally, I estimated models that
standardized both chronic conditions and depressive symptoms to compare the magnitude of the
coefficients in these two models. Missing values ranged from less than 1% on chronic conditions
to about 7% on duration in the current marital status. Missing data were imputed using multiple
imputation by creating ten replicates (Acock, 2005), and each model was weighted to correct for
unequal probability of selection and sample attrition (Ofstedal, Weir, Chen, & Wagner, 2011).
Descriptive Results
Table 2.1 displays weighted descriptive statistics. The results show on average, women
reported more depressive symptoms in the week prior to the interview (1.49 versus 1.16) and a
greater number of chronic conditions (2.22 versus 2.09) than men. Although the majority of the
sample was partnered—close to three-quarters of men and almost 57% of women, more men were partnered than women whether this was in a first marriage, remarriage, or cohabitation.
More women were unpartnered than men—16% of women were divorced versus about 13% of men. One-fifth of women were widowed versus only 6% of men. 27
More men than women were remarried after one divorce (about 68% compared to a little less than 61%). However, a greater share of women were remarried after multiple disruptions than men— only about a quarter of remarried men reported multiple disruptions versus close to one-third of remarried women. When unmarried, more men reported one divorce or multiple disruptions than women. Nearly 41% of unmarried men had one divorce versus a little less than
27% of women and 36% of unmarried men reported multiple disruptions versus 33% of women.
More women had one widowhood versus their male counterparts. Only about a quarter of unmarried men were widowers versus about two-fifths of unmarried widowed women. Men had also spent more years in the current marital status than women (29.6 years versus 28.5 years).
In the analytic sample, men were younger than women, on average. More men fell between the ages of 50 and 64 than women. About half of men were 50 to 64 versus around 45% of women. A greater percentage of women were 85 or older with about 7% of women in this category versus a little less than 5% of men. A greater percentage of men were White than women (78% versus 77%), whereas a greater percentage of women were Black (11% versus
9%). Both men and women had about three children. Men were more educated than women on average, with nearly one third of men reporting a college degree or higher versus a little over one quarter of women. Men had accrued more assets than women. More men than women had
$250,000 or more—nearly one-third of men reported $250,000 or more in assets versus slightly fewer than 29% of women. A greater percentage of men than women were employed. Less than a third of women were working whereas nearly two-fifths of men were in the workforce.
28
Multivariate Results
Marital Status and Health
Table 2.2 considers the relationship between current marital status and depressive symptoms as well as chronic conditions for men and women separately. Consistent with prior literature, I found first married men reported fewer depressive symptoms than men in all other marital statuses with the exception of remarried and cohabiting men (i.e., superscript refers to differences between the first married and “a” the divorced, “b” the widowed, and “c” the never married). Moreover, men in a remarriage reported a similar number of depressive symptoms as cohabiting men. Among unmarried men, the divorced, widowed, and never married reported more depressive symptoms than remarried men. All groups of unmarried men shared similar levels of depressive symptoms.
Women in a first marriage reported fewer depressive symptoms relative to all other marital status groups (i.e., superscript refers to differences between the first married and “a” the cohabitors, “b” the divorced, “c” the widowed, and “d” the never married). Remarried women were less depressed than their divorced and widowed counterparts, but there was no difference between remarried women’s depressive symptoms and those for cohabiting or never married women. Moreover, there were no differences between divorced, widowed, and never married women, suggesting there was no variation among unmarried women. In this model, I found no differences in depressive symptoms by marital status group for men relative to women.
There were no differences in chronic conditions for men occupying different marital statuses. Among women, the first married reported fewer chronic conditions than their divorced and widowed counterparts whereas there were no differences between the first married compared to remarried, cohabiting, and never married women (i.e., superscript denotes differences between 29
the first married and “a” the divorced as well as “b” the widowed). The only other significant
difference that arose among women was remarried women reported fewer chronic conditions
than the currently widowed. There were no differences between groups of unmarried women.
When comparing men and women, the only significant difference that emerged for marital status showed never-married men reported fewer chronic conditions than never married women
(denoted by underlined coefficients).
Remarriage, Disruption Pathways, and Health
In Table 2.3, I examined whether different disruption pathways and subsequent remarriage were associated with mental or physical health separately for men and women. I expected the remarried would be less depressed than their unmarried counterparts, regardless of disruption pathway (Hypothesis 1a). Consistent with my prediction, I found men who were unmarried after one divorce, one widowhood, or multiple disruptions reported greater depressive symptoms than remarried men experiencing the corresponding disruption pathway (i.e., superscript refers to differences between the remarrieds and unmarrieds for “a” the widoweds and “b” individuals with multiple disruptions). I also predicted among either unmarried or remarried older adults, those after one widowhood would fare best followed by individuals after one divorce, and finally, those after multiple disruptions would report the worst health
(Hypothesis 1b). However, there were no differences among unmarried or remarried older men of differing disruption pathways.
For women, although I expected the remarried to report fewer depressive symptoms than the unmarried, I did not find support for this prediction (Hypothesis 1a). However, I did find differences between different disruption pathways among women (Hypothesis 1b). Unmarried women after multiple disruptions experienced more depressive symptoms than unmarried 30
women after either one divorce or one widowhood (i.e., superscript refers to differences between
the unmarried after multiple disruptions and “c” the unmarried after one divorce and “b” the
unmarried after one widowhood). Moreover, remarriage after multiple disruptions was linked to
greater depressive symptoms than for remarriage after one divorce for women. I also expected remarriage and disruption pathway to be associated with worse health for men relative to women
(Hypothesis 3a). One finding emerged that supported this prediction (denoted by underlined coefficients). Men who were unmarried after one divorce reported more depressive symptoms
than unmarried women after one divorce (0.48 versus 0.18).
For chronic conditions, my predictions regarding the benefits of remarriage relative to
being unmarried (Hypothesis 1a) as well as differences between disruption pathways
(Hypothesis 1b) were unsupported for men as there were no significant associations between
chronic conditions, marital status, and disruption pathways. For women, the unmarried after one
divorce had more chronic conditions than their divorced counterparts who had remarried
(Hypothesis 1a). As for differences among various disruption pathways, among remarried
women, those who remarried after multiple disruptions had more chronic conditions than their
remarried counterparts after one divorce (Hypothesis 1b). Although I expected women to be
advantaged over men when considering remarriage and different disruption pathways
(Hypothesis 3a), this prediction was not supported. In fact, men who were unmarried after one
divorce had a slight reduction in number of chronic conditions (-0.03), whereas for women,
being unmarried following one divorce was positively associated with chronic conditions (0.17)
(denoted by underlined coefficients).
31
Duration Remarried, Duration Unmarried, and Health
Table 2.4 adds duration remarried or unmarried to examine whether years spent remarried or unmarried following different disruption pathways were associated with greater or fewer depressive symptoms and chronic conditions for men and women. I hypothesized that relative to unmarried groups with the same disruption pathway, years remarried would be associated with better health (Hypothesis 2a). However, my findings did not support the hypothesis for either depressive symptoms or chronic conditions. I also expected that for different disruption pathways, those after multiple disruptions would be disadvantaged relative to those after one widowhood or one divorce (Hypothesis 2b). This prediction was supported for men’s depressive symptoms. Among remarried men, each additional year remarried after multiple disruptions was associated with greater depressive symptoms (0.02) compared to duration remarried after one divorce.
All additional covariates included in the analyses remained relatively consistent across the models for men and women. As shown in Table 2.4, younger and less educated men reported greater depressive symptoms than their respective counterparts. Men with greater wealth, and who were currently employed reported fewer depressive symptoms than their respective counterparts. For chronic conditions, older men reported a greater number of chronic conditions while younger men reported fewer (relative to men aged 65 to 84). Hispanic men had fewer chronic conditions than White men. Having a college education or higher, wealth, and current employment are inversely related to chronic conditions for men.
As shown in Table 2.4, younger women (ages 50 to 64), Hispanic women and women of
other racial and ethnic backgrounds were more depressed than their respective counterparts.
Women who had less than a high school diploma reported greater depressive symptoms whereas 32
women with college or more education, greater wealth, and who were currently employed had
fewer depressive symptoms. Women of younger ages reported fewer chronic conditions, while
older women reported more chronic conditions (compared to women 65 to 84). Similar to men,
Hispanic women reported fewer chronic conditions versus White women. Having a higher
education for women was negatively associated with number of chronic conditions. Greater
wealth and employment also relate to fewer chronic conditions among women.
Last, I standardized depressive symptoms and chronic conditions and re-estimated each of the models (Hypothesis 4). The results from the standardized models were inconsistent with
my prediction. Across all of the models, when standardizing the two health outcomes, I found the
effect sizes were larger for the associations of marital biography and depressive symptoms than
they were with chronic conditions (results not shown but available upon request).
Discussion
Marital status is linked to health in later life such that the married report better health compared to the unmarried (Liu & Umberson, 2008; Umberson, 1992). Marriage provides
financial, social, and emotional resources that promote health relative to the unmarried who often
face many stressors that erode overall health (Umberson, 1992; Williams & Umberson, 2004).
Studies of marital biography extend work on current marital status by factoring in the role of
marital characteristics like marital disruption and duration. The goal of this chapter is to investigate whether remarriage is associated with fewer consequences of different disruption pathways and to also examine the role of duration remarried or unmarried after marital disruption. By accounting for remarriage after different disruption pathways as well as the function of duration remarried or unmarried after disruption, researchers can better understand how marital biography is associated with physical and mental health. 33
When considering current marital status, my findings mostly confirm prior studies. For the most part, first marrieds were less depressed than all other marital status groups, which is likely attributable to resources related to being continuously married compared to being unmarried (Liu & Umberson, 2008). Remarried persons’ and cohabitors’ depressive symptoms and chronic conditions did not differ from one another, a finding that received support from other studies (Wright, 2017). Finally, I did not find substantial variation among groups of unmarried older adults.
For chronic conditions, some of my findings were also consistent with prior research. For instance, widows had more chronic conditions than remarried and first married women.
Moreover, the divorced women reported more chronic conditions than first married women,
reinforcing the idea that the married (either in first marriages and remarriages) garner resources
from the institution of marriage, promoting physical health (Hughes & Waite, 2009; Umberson,
1992; Zhang et al., 2016). However, these findings were not mirrored in the results for men, as
there were no differences in chronic conditions among marital status groups. As for gender, one
contradictory finding regarding chronic conditions emerged. Contrary to the results for
depressive symptoms, I found that women who were unmarried after one divorce reported more
chronic conditions than their unmarried male counterparts with one divorce. This finding could
be reminiscent of recent work that has shown that women’s onset of illness was associated with
an increase in the risk of divorce (Karraker & Latham, 2015).
Consistent with my predictions regarding the advantage of the remarried relative to the
unmarried regardless of the disruption pathway, I found remarried men were less depressed than
all unmarried groups, no matter what type of disruption they had experienced (i.e., one divorce,
one widowhood, or multiple disruptions) (Hypothesis 1a). I also found remarried women after 34
one divorce reported fewer chronic conditions than their unmarried counterparts. These findings
suggest remarriage provides health-promoting resources that allow a mental health recovery for
men and a physical health recovery for women after marital disruption.
As for the prediction regarding differences between disruption pathways (Hypothesis 1b),
I found disruption pathway mattered for women’s depressive symptoms and chronic conditions, but was not significantly associated with men’s. Remarried women after one divorce reported fewer depressive symptoms and chronic conditions than their remarried counterparts after multiple disruptions. In a similar vein, unmarried women after one divorce or widowhood also reported fewer depressive symptoms than unmarried women after multiple disruptions. These findings indicate among women, different disruption pathways are important. Specifically, having more than one disruption seemingly spurs additional costs that are more difficult to recover from relative to experiencing one divorce or one widowhood for women.
The results did not show a positive relationship between years remarried and physical health relative to spending years unmarried (Hypothesis 2a). There was also no relationship between chronic conditions and duration after different disruption pathways regardless of whether these years were spent remarried or unmarried (Hypothesis 2b). Taken together, these findings indicate that duration did not bear any significant tie to chronic conditions in this dissertation. Additionally, I did not find support for Hypothesis 2a when examining depressive symptoms for men or women. The lack of findings that support this prediction suggests that additional years spent remarried relative to years spent unmarried are not associated with a health advantage. Although remarriage likely provides resources which potentially accumulate over time and offer health benefits, years spent unmarried may also allow an individual to recover 35 from the stress of a marital disruption, which then would also be related to better health. In sum, there may be few differences between one year remarried relative to one spent unmarried.
Consistent with Hypothesis 2b, duration remarried after multiple disruptions was positively related to depressive symptoms compared to duration remarried following one divorce for men. Recovering after more than one marital disruption may be more difficult due to the additional financial, social, and emotional consequences of disruption even among men who remarry (Kalmijn, 2017). Furthermore, the benefits of marriage or remarriage require time to accrue health-promoting resources (Umberson, 1992). Older adults who have had at least one marital disruption— especially divorce— are more likely to eventually have another disruption in the future (Brown & Lin, 2012), interrupting time to accrue any benefits from remarriage.
Additionally, although remarriage provides financial, social, and emotional resources, a remarriage is still not as beneficial as remaining in a stable first marriage (Barrett, 2000; Dupre
& Meadows, 2007; Marks & Lambert, 1998; Reczek et al., 2016; Williams & Umberson, 2004).
A recent study suggests the cost of disruption is three times as high as entry into marriage
(Kalmijn, 2017), implying even though remarriage could be associated with fewer consequences of disruption, the costs of divorce cannot be entirely erased. Despite this finding for men, this prediction did not receive support among women. This suggests that additional years following any disruption pathway did not differ significantly from one another when investigating mental health.
Prior work suggests there are gender differences in the relationship between marital biography and health, such that remarriage is more beneficial for men’s health than women’s
(Reczek et al., 2016). However, the findings from this chapter did not yield substantial gender variation. There were only two gender differences that arose from the analyses. First, consistent 36
with my predictions, men who were unmarried after one divorce reported more depressive symptoms than women who were unmarried after one divorce (Hypothesis 3a). However, women who were unmarried following one divorce reported greater chronic conditions than their male counterparts. The few gender differences emerging from this chapter are supported by some prior work demonstrating that men often encounter more negative health consequences related to being unmarried that do women (Amato & Rodgers, 1997; Hu & Goldstein, 1990). Still, this was the only finding that emerged regarding gender differences, suggesting that there is not much of variation in the relationship between marital biography and health for men and women. The lack of significant differences between men and women is congruent with research suggesting the health of men and women converge with age (Manzoli, Villari, Pirone, & Boccia, 2007).
Finally, despite arguments that current marital status would be related more to mental health outcomes, like depressive symptoms whereas marital disruption and duration would be tied more to physical health, like chronic conditions, my findings did not lend support to this speculation. Indeed, I found depressive symptoms to be more strongly tied to chronic conditions in all of the models (Hypothesis 4). In general, the majority of the predictions surrounding chronic conditions were unsupported. It is plausible that there was less variation in the measure for chronic conditions than depressive symptoms in the HRS. Once a respondent reports a chronic condition in the survey, they will always be assumed to have this condition in subsequent waves and thus, respondents can gain chronic ailments, but they cannot recover from them.
Moreover, conclusions from prior work on marital biography and health lend support to these null findings. Among a group of ever-disrupted older people, Hughes and Waite (2009) found that neither (1) being remarried nor (2) having more than one disruption were significantly associated with number of chronic conditions. Thus, taken together, the findings from both this 37
dissertation and prior work suggest that these marital characteristics may be less salient for chronic conditions relative to other health indicators, like depressive symptoms.
This chapter has several other limitations that warrant future research. First, due to small
sample sizes, I cannot account for cohabitation after different disruption pathways. Thus, as the
number of older cohabitors continues to increase (Brown & Hemez, 2016), future work should
examine whether subsequent cohabitation is associated with greater or fewer costs of one
divorce, one widowhood, or multiple disruptions relative to entering a remarriage. Second, this
chapter does not establish causal ordering between marital biography and health. Marital
disruption has negative effects on health, but health declines could also lead to marital disruption
or could deter remarriage.
Nevertheless, this chapter makes a contribution to literature by investigating the
relationship between marital biography and health. Today, many older adults experience either
divorce or widowhood (Brown & Lin, 2012; Manning & Brown, 2011; Lin & Brown, 2012) and
some even experience multiple disruptions. This chapter shows that it is important to unpack
different disruption pathways to see whether one divorce, one widowhood, or multiple
disruptions are associated with similar health disadvantages. Specifically, in most of the
analyses, older people after multiple disruptions were at a health disadvantage while there were
fewer differences between the divorced and widowed.
Some of these people with prior marital disruptions go on to remarry; nearly 30% of all
marriages to people 50 and older are to individuals in higher order marriages (Lin et al., in
press). Moreover, studies show remarriage provides some of the health benefits as first marriage
(Barrett, 2000; Dupre & Meadows, 2007; Marks & Lambert, 1998; Williams & Umberson, 2004;
Reczek et al., 2016). In alignment with this prior work, my chapter makes another significant 38 contribution to the current literature by showing that remarriage is beneficial for the health of older people. More specifically, the findings from this chapter show remarriage, particularly for men, related to fewer health consequences that are often associated with divorce, widowhood, or multiple disruptions. Still, remarriage is not associated with a complete health recovery from disruption, especially among those who have had multiple disruptions, as evidenced by the decline in the health of remarried men after multiple disruptions relative to men after one divorce with each additional year remarried. With increasing marital instability in later life, it is likely that the number of older people with more than one marital disruption will grow, which may jeopardize the health of these older adults. Even if these older adults do remarry—although the majority of them will not (Brown et al., in press)—remarriage may not completely ameliorate the health consequences of different disruption pathways
In sum, multiple disruptions are particularly harmful for health and duration remarried cannot ameliorate the negative consequences of marital disruptions. Given these findings, it is crucial for policymakers and practitioners to consider new strategies to support the health of older people who likely lack social, financial, and emotional support due to marital disruption.
Moreover, in light of the conclusions presented here, it is essential to investigate other factors, like parent-child contact, that may relate not only to the well-being of older people, but also their availability of social support, particularly for older adults with poor health (Dykstra & de Jong
Gierveld, 2006; Lawton, Silverstein, & Bengston, 1994; Mancini & Blieszner, 1989). In the next chapter, I will take a closer look at the association of marital biography and parent-child contact in later life. 39
Table 2.1 Weighted Means (Standard Deviations) and Percentages by Gender Men Women Health Depressive symptoms 1.16 (1.8) 1.49 (3.8) *** Chronic conditions 2.09 (1.5) 2.22 (3.0) *** Marital biography First marriage 45.3 36.1 *** Remarriage 24.8 17.5 *** After one divorce 67.7 60.9 *** After one widowhood 6.5 8.5 After multiple disruptions 25.8 30.6 * Cohabiting 4.7 3.1 *** Divorced 12.5 16.4 *** Widowed 6.1 20.4 *** Unmarried after one divorce 40.9 26.6 ** Unmarried after one widowhood 23.4 40.3 *** Unmarried after multiple disruptions 35.7 33.1 *** Never married 6.4 6.5 Duration in current status 29.56 (18.1) 28.47 (36.1) *** Demographic characteristics Age 50 to 64 48.6 45.4 ** 65 to 84 46.7 47.2 85 and older 4.7 7.4 *** Racial and ethnic background White 78.4 76.7 * Black 9.3 10.7 ** Hispanic 8.6 9.1 Other race 3.7 3.5 Number of children 2.83 (1.9) 2.89 (3.9) Socioeconomic resources Education Less than high school 12.3 13.6 * High school 30.0 33.3 ** Some college 25.6 26.5 College or more 32.1 26.6 *** Wealth In debt 8.2 8.5 $1 to $50,000 (ref) 36.6 40.6 *** $50,001 to $100,000 9.1 8.8 $100,001 to $250,000 13.7 13.2 $250,001+ 32.4 28.9 *** Currently employed 39.1 30.7 *** Weighted Percentages 46.07 53.93 Unweighted N 7,127 9,933 * p < .05; ** p < .01; *** p < .001
40
Table 2.2 Coefficients (Standard Errors) from the Weighted OLS Regression of Depressive Symptoms and Chronic Conditions on Marital Status Depressive Symptoms Chronic Conditions Men Women Men Women Marital biography First marriage - 0.06 (.07) abc - 0.23 (.07) **abcd 0.02 (.06) - 0.03 (.06) ab Remarriage (ref) ------Cohabiting - 0.03 (.12) 0.11 (.18) a 0.04 (.09) 0.02 (.11) Divorced 0.56 (.11) ***a 0.21 (.09) *b 0.04 (.09) 0.13 (.07) a Widowed 0.54 (.10) ***b 0.30 (.09) **c 0.10 (.09) 0.16 (.06) *b Never married 0.50 (.16) **c 0.20 (.16) d - 0.11 (.12) 0.12 (.12) Demographic characteristics Age 50 to 64 0.42 (.08) *** 0.57 (.07) *** - 0.46 (.05) *** - 0.38 (.06) *** 65 to 84 (ref) ------85 and older - 0.01 (.11) - 0.10 (.08) 0.22 (.08) ** 0.13 (.05) * Racial and ethnic background White (ref) ------Black 0.13 (.11) 0.02 (.09) - 0.08 (.05) 0.08 (.05) Hispanic 0.13 (.11) 0.30 (.10) ** - 0.32 (.07) *** - 0.26 (.07) ** Other race 0.30 (.06) 0.43 (.14) ** - 0.12 (.11) - 0.07 (.11) Number of children - 0.00 (.02) - 0.02 (.01) 0.02 (.01) 0.01 (.01) Socioeconomic resources Education Less than high school 0.27 (.11) * 0.62 (.10) *** 0.04 (.07) 0.23 (.06) *** High school (ref) ------Some college - 0.12 (.07) - 0.10 (.07) - 0.04 (.07) - 0.10 (.04) * College or more - 0.24 (.06) *** - 0.31 (.06) - 0.26 (.05) *** - 0.24 (.05) *** Wealth In debt 0.30 (.14) 0.65 (.11) *** 0.45 (.11) *** 0.36 (.08) *** $1 to $50,000 (ref) ------$50,001 to $100,000 - 0.28 (.07) *** - 0.16 (.10) 0.08 (.07) - 0.34 (.07) *** $100,001 to $250,000 - 0.33 (.06) *** - 0.44 (.08) *** - 0.09 (.06) - 0.41 (.06) *** $250,001+ - 0.37 (.07) *** - 0.42 (.07) *** - 0.31 (.06) *** - 0.58 (.05) *** Currently employed - 0.67 (.07) *** - 0.77 (.07) *** - 0.90 (.05) *** - 0.72 (.05) *** Constant 1.31 (.11) *** 1.61 (.11) *** 2.78 (.10) *** 2.81 (.08) *** N 7,127 9,933 7,127 9,933 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing same superscript letter denote significant differences at p < .05 Underlined coefficients denote significant differences between men and women at p < .05
41
Table 2.3 Coefficients (Standard Errors) from the Weighted OLS Regression of Depressive Symptoms and Chronic Conditions on Marital Status and Disruption Pathway Depressive Symptoms Chronic Conditions Men Women Men Women Marital biography First marriage 0.01 (.07) - 0.15 (.08) 0.05 (.06) 0.03 (.07) Remarriage After one divorce (ref) ------After one widowhood - 0.02 (.12) a 0.06 (.22) 0.00 (.11) - 0.00 (.17) After multiple disruptions 0.22 (.11) b 0.25 (.11) * 0.10 (.09) 0.20 (.10) * Cohabiting 0.02 (.12) 0.18 (.17) 0.06 (.09) 0.08 (.10) Unmarried After one divorce 0.48 (.13) *** 0.18 (.10) c - 0.03 (.10) 0.17 (.08) * After one widowhood 0.71 (.11) a 0.28 (.08) d 0.15 (.09) 0.19 (.07) After multiple disruptions 0.67 (.14) b 0.52 (.10) cd 0.18 (.11) 0.24 (.08) Never married 0.54 (.17) 0.26 (.17) - 0.09 (.12) 0.18 (.13) Demographic characteristics Age 50 to 64 0.43 (.08) *** 0.55 (.07) *** - 0.45 (.05) *** - 0.38 (.05) *** 65 to 84 (ref) ------85 and older - 0.04 (.11) - 0.05 (.07) 0.21 (.08) * 0.15 (.05) ** Racial and ethnic background White (ref) ------Black 0.14 (.11) 0.04 (.09) - 0.07 (.05) 0.09 (.05) Hispanic 0.14 (.11) 0.32 (.10) ** - 0.31 (.07) *** - 0.25 (.07) ** Other race 0.30 (.15) 0.45 (.14) ** - 0.12 (.11) - 0.06 (.11) Number of children - 0.00 (.02) - 0.02 (.01) 0.01 (.01) 0.01 (.01) Socioeconomic resources Education Less than high school 0.26 (.11) * 0.62 (.10) *** 0.05 (.07) - 0.23 (.06) *** High school (ref) ------Some college - 0.14 (.07) - 0.10 (.07) - 0.04 (.07) - 0.10 (.04) * College or more - 0.24 (.06) *** - 0.31 (.06) - 0.26 (.05) *** - 0.24 (.05) *** Wealth In debt 0.30 (.14) * 0.63 (.11) *** 0.44 (.11) *** 0.36 (.08) *** $1 to $50,000 (ref) ------$50,001 to $100,000 - 0.28 (.07) *** - 0.15 (.10) 0.08 (.07) - 0.33 (.08) *** $100,001 to $250,000 - 0.33 (.06) *** - 0.43 (.08) *** - 0.09 (.06) - 0.41 (.06) *** $250,001+ - 0.37 (.07) *** - 0.41 (.07) *** - 0.31 (.06) *** - 0.57 (.05) *** Currently employed - 0.67 (.07) *** - 0.76 (.07) *** - 0.90 (.01) *** - 0.72 (.05) *** Constant 1.27 (.12) *** 1.54 (.12) *** 2.77 (.10) *** 2.76 (.08) *** N 7,127 9,933 7,127 9,933 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing same superscript letter denote significant differences at p < .05 Underlined coefficients denote significant differences between men and women at p < .05
42
Table 2.4 Coefficients (Standard Errors) from the Weighted OLS Regression of Depressive Symptoms and Chronic Conditions on Marital Status, Marital Disruption, and Duration Depressive Symptoms Chronic Conditions Men Women Men Women Marital biography First marriage 0.18 (.22) - 0.39 (.27) - 0.21 (.15) - 0.00 (.21) Remarriage After one divorce (ref) ------After one widowhood - 0.22 (.29) - 0.05 (.37) 0.18 (.21) 0.27 (.30) After multiple disruptions - 0.20 (.20) - 0.02 (.27) - 0.02 (.15) 0.10 (.18) Cohabiting - 0.04 (.17) - 0.29 (.27) 0.19 (.15) 0.12 (.22) Unmarried (divorced or widowed) After one divorce 0.47 (.27) 0.04 (.26) - 0.02 (.17) 0.13 (.21) After one widowhood 0.43 (.20) * 0.28 (.25) 0.21 (.16) 0.30 (.17) After multiple disruptions 0.41 (.26) 0.54 (.28) 0.18 (.18) 0.31 (.18) Never married 0.69 (.19) *** 0.34 (.23) 0.14 (.14) 0.12 (.16) Duration 0.01 (.01) - 0.01 (.01) - 0.00 (.00) 0.00 (.01) Remarried after one divorce (ref) - - - - Remarried after one widowhood 0.01 (.01) 0.00 (.01) - 0.01 (.01) - 0.01 (.01) Remarried after multiple disruptions 0.02 (.01) * 0.01 (.01) 0.01 (.01) 0.01 (.01) Unmarried after one divorce - 0.00 (.01) 0.00 (.01) - 0.00 (.00) 0.00 (.01) Unmarried after one widowhood 0.01 (.01) - 0.00 (.01) 0.00 (.00) - 0.00 (.01) Unmarried after multiple disruptions 0.01 (.02) - 0.01 (.01) 0.00 (.00) - 0.00 (.01) Demographic characteristics Age 50 to 64 0.36 (.09) *** 0.55 (.08) *** - 0.40 (.06) *** - 0.35 (.06) *** 65 to 84 (ref) ------85 and older 0.02 (.12) - 0.02 (.07) 0.17 (.08) * 0.14 (.05) ** Racial and ethnic background White (ref) ------Black 0.12 (.11) 0.06 (.10) - 0.06 (.06) 0.09 (.05) Hispanic 0.14 (.11) 0.33 (.11) ** - 0.31 (.07) *** - 0.24 (.07) ** Other race 0.27 (.15) 0.45 (.14) ** - 0.10 (.12) - 0.05 (.11) Number of children - 0.00 (.02) - 0.02 (.01) 0.01 (.01) 0.01 (.01) Socioeconomic resources Education Less than high school 0.26 (.10) * 0.63 (.10) *** 0.05 (.07) 0.23 (.06) *** High school (ref) ------Some college - 0.14 (.07) * - 0.09 (.07) - 0.03 (.06) - 0.09 (.04) * College or more - 0.26 (.06) *** - 0.31 (.06) *** - 0.17 (.05) *** - 0.05 (.11) *** Wealth In debt 0.29 (.14) 0.63 (.11) 0.44 (.11) *** 0.35 (.08) *** $1 to $50,000 (ref) ------$50,001 to $100,000 - 0.27 (.07) *** 0.15 (.10) *** 0.08 (.07) - 0.33 (.08) *** $100,001 to $250,000 - 0.32 (.06) *** - 0.44 (.08) *** - 0.09 (.05) - 0.41 (.06) *** $250,001+ - 0.37 (.07) *** - 0.42 (.07) *** - 0.30 (.06) *** - 0.57 (.05) *** Currently employed - 0.69 (.07) *** - 0.76 (.07) *** - 0.87 (.05) *** - 0.71 (.05) *** Constant 1.51 (.18) *** 1.70 (.25) *** 2.67 (.15) *** 2.62 (.19) *** N 7,127 9,933 7,127 9,933 * p < .05; ** p < .01; *** p < .001
43
CHAPTER 3: MARITAL BIOGRAPHY AND PARENT-CHILD CONTACT IN LATER LIFE
Older adults today experience a wealth of demographic change, especially among marital
behaviors. In the past, people would marry and remain in that union until the death of one spouse
(Cherlin, 2009), however today, this pathway is less normative. In 2015, nearly 30% of married
people age 50 and older were in a higher order marriage compared to less than one-fifth in 1980
(Lin et al., in press), suggesting a number of older people were divorced or widowed in early,
mid, or later life. Although many older people experience widowhood, gray divorce—or divorce
to people over age 50—is on the rise (Brown & Lin, 2012). Moreover, a number of these older
people who have divorced or widowed go on to remarry (Brown et al., in press). All of these
trends lead to an accumulation of marital transitions by the time people enter mid to later life.
Marital biography is an innovative way to capture this rise in marital instability and encompasses current marital status, marital disruption, duration, timing, and sequencing (Zhang et al., 2016).
Even though each of these components are useful when examining how marital biography relates to parent-child contact in later life, current marital status, marital disruption, and duration are particularly salient for investigating the well-being of older people, including contact with adult children (Albertini & Garriga, 2011; Daatland, 2007; Kalmijn, 2007, 2013, 2015; Noël-Miller,
2013; Shapiro, 2003; Ward et al., 2013).
Parents’ bonds with biological children are one of the most enduring social ties situated within the life course (Seltzer & Bianchi, 2013). Parents provide support to children as they grow older and assist children as they enter adulthood. As parents age and face health declines, adult children provide financial, emotional, and time assistance (Silverstein & Giarrusso, 2010).
Parent-child contact, including frequency of emails, telephone calls, and face-to-face interaction, is a pillar of intergenerational solidarity (Bengston & Oyama, 2010; Silverstein & Bengston, 44
1997). Contact reduces loneliness among older adults and promotes affection between parents
and children (Dykstra & de Jong Gierveld, 2006; Lawton et al., 1994; Mancini & Blieszner,
1989). Parent-child ties persist over time, but are subject to change, especially after shifts in the
parent’s life, like those related to marital biography (Connidis, 2010; Hammersmith, in press).
Prior studies of marital characteristics and parent-child contact show stable marriage and widowhood are positively related to parent-child contact whereas divorce and remarriage are negatively associated with contact (Albertini & Garriga, 2011; Daatland, 2007; Kalmijn, 2007,
2013, 2015; Noël-Miller, 2013; Pezzin & Schone, 1999; Roan & Raley, 1994; Shapiro, 2003).
Nonetheless, only a handful of studies have considered how parent’s disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions) and subsequent remarriage could take a cumulative toll on parent-child contact. Further, fewer studies have investigated the role of duration remarried or unmarried after disruption (Ward et al., 2013). The parent-child relationship changes over time (Birditt, Miller, Fingerman, & Lefkowitz, 2009), and thus, it is
likely duration remarried or unmarried after different disruption pathways could be variably
linked to parent-child contact.
Using Health and Retirement Study data from 1992 to 2012, I address three research
questions. First, how does remarriage after disruption relate to contact with biological children and what is the role of disruption pathway (i.e., one divorce, one widowhood, or multiple disruptions)? Second, is duration remarried or unmarried after different disruption pathways associated with parent-child contact? Finally, does remarriage after different disruption pathways and duration remarried or unmarried relate to contact with children differently across gender?
There are several contributions of this chapter. First, I incorporate not only current marital status, but also marital disruption and duration remarried or unmarried. Research has 45 mostly focused on singular dimensions of marital biography when assessing parent-child contact, like divorce or repartnership, finding that these marital transitions are negatively related to contact with one’s biological children (Albertini & Garriga, 2011; Daatland, 2007; Kalmijn,
2007, 2013; Roan & Raley, 1994; Shapiro, 2003). Still, we know little about whether remarriage has different associations with parent-child relationships following different disruption pathways.
There is also limited research about how duration remarried or unmarried following divorce, widowhood, or multiple disruptions relates to parent-child contact. Finally, this chapter also extends prior work through investigating gender differences. The contact men and women share with children varies by current marital status, and thus, remarriage following disruption and duration remarried or unmarried may differentially relate to contact women share with children relative to men.
Intergenerational Solidarity
Intergenerational solidarity describes closeness that exists within family ties across generations, including relationships between parents and children (Bengston & Oyama, 2010;
Silverstein & Bengston, 1997). Parents and children’s lives are interwoven over the life course; parents raise children until they reach adulthood and maintain close ties with children after their offspring establish independent households (Rossi & Rossi, 1990). Likewise, children reach out to parents as they navigate adulthood (Silverstein & Bengston, 1997). Intergenerational solidarity is multifaceted and consists of six dimensions (Bengston & Roberts, 1991; Silverstein &
Bengston, 1997). Structural solidarity examines proximity between parents and children.
Affectual solidarity captures emotional closeness between parents and children, whereas consensual solidarity measures their shared opinions. Responsibility or obligations toward parents is examined by normative solidarity and functional solidarity captures help as well as 46 support exchanged between parents and children. The sixth dimension and focus of this chapter accounts for associational solidarity or contact between parents and children (Bengston &
Harootyan, 1994; Bengston & Silverstein, 1997). Associational solidarity is an integral component of intergenerational ties, especially as research indicates normative, affectual, and associational solidarity are interrelated, such that normative solidarity links to greater affectual solidarity, which relates to higher associational solidarity (Bengston & Roberts, 1991).
Associational solidarity is a fundamental indicator of relationship strength and older adults’ potential support avenues (Bengston & Oyama, 2010; Mancini & Blieszner, 1989; Silverstein &
Bengston, 1997).
Contact between parents and children typically remains strong as parents and children grow older (Umberson, 1992). Yet, changes in the parent’s life, like those related to marital biography are associated with parent-child contact (Connidis, 2010; Ward et al., 2013). For instance, certain marital statuses encourage closeness between parents and children, like continuous marriage and widowhood, and thus, promote parent-child contact (Guiaux, van
Tilberg, & Broese van Groenou, 2007; Roan & Raley, 1996; Waite & Harrison, 1992; Ward et al., 2009). Others, like divorce, remarriage, and never marrying foster weaker parent-child relationships, which relate to less contact between parents and children (Albertini & Garriga,
2011; Daatland, 2011; Dykstra & Gierveld, 2006; Kalmijn, 2007, 2013, 2015).
Marital Status and Parent-Child Contact
Older adults in a first marriage share more contact with children than any other marital status group, apart from the widowed (Dykstra & de Jong Gierveld, 2006; Kalmijn, 2013; Roan
& Raley, 1996; Shapiro, 2003). Marriage fosters stability in parent-child ties, which facilitates contact (Kalmijn, 2013; Pezzin, Pollak, & Schone, 2008). Among other partnered groups, 47
remarrieds and cohabitors report similar contact frequency with children (Noël-Miller, 2013;
Wright, 2017). When comparing the remarried to the unmarried, it is important to distinguish by
unmarried status as the widowed report more contact with children than the remarried, the
divorced, and never married. Children rally around parents after the death of a spouse, promoting
contact (Ha, 2008; Roan & Raley, 1996). When comparing the divorced with the remarried, prior
work shows even though divorce lowers contact relative to marriage, remarriage jeopardizes parent-child contact even further, especially for fathers (Kalmijn, 2007; 2015; Noël-Miller, 2013;
Pezzin & Schone, 1999). Parent-child contact among the never married has received limited attention due to fewer never-married older adults with children until recent increases in nonmarital fertility (Center for Disease Control and Prevention, 2009). Nevertheless, never marrieds are more socially isolated from social network members—like children—in later life than other marital status groups, and thus, share less contact with children than the widowed, divorced, and remarried (Dykstra & de Jong Gierveld, 2006).
Remarriage, Disruption Pathways, and Parent-Child Contact
A growing number of older people experience at least one marital disruption through divorce and many still experience marital disruption through widowhood (Brown & Lin, 2012;
Manning & Brown, 2011). Further, some divorced or widowed older people go on to remarry
(Brown et al., in press). Although prior studies have investigated how marital events like divorce, widowhood, and remarriage separately relate to contact frequency between parents and children, only a few studies have focused on remarriage after different disruption pathways. For instance,
some studies have shown that divorce is negatively associated with contact with children,
especially among fathers (Albertini & Garriga, 2011; Aquilino, 1994; Daatland, 2007; Kalmijn,
2013; Shapiro, 2003) whereas other researchers discover that widowhood yields similarity in 48 contact compared to continuous marriage (Roan & Raley, 1996). These studies are limited as they did not factor in the role of a subsequent remarriage or the importance of different disruptions pathways as they compare to one another on frequency of parent-child contact.
A few studies improve upon this work by incorporating not only how divorce relates to parent-child contact, but also the relationship between subsequent remarriage and parent-child contact. These studies uncovered that not only is divorce related to less frequent parent-child contact, but also remarriage lessens contact to an even greater extent, particularly for fathers
(Dykstra & Fokkema, 2011; Kalmijn, 2007, 2015; Noël-Miller, 2013). Although findings from these studies render noteworthy conclusions about how divorce and remarriage link to parent- child contact, these studies did not consider disruption pathways beyond divorce and how remarriage after different disruption pathways may be positively or negatively associated with contact.
Moreover, a handful of studies move beyond examining solely divorce and subsequent repartnership by accounting for disruption pathways through both divorce and widowhood as well as subsequent repartnership. Yet, these studies did not consider marital history to determine whether men or women had prior divorces or widowhoods before the most recent disruption.
Therefore, these studies did not factor in the role of multiple disruptions when investigating parent-child contact (Kalmijn, 2007, 2015). Pezzin and Schone (1999) extended these studies to look at contact for the widowed, divorced, and repartnered. They discovered relative to the widowed, the divorced shared less contact with children, and contact further deteriorated with repartnership. Similarly, Ward and colleagues (2014) examined contact of parents who remarry or remain unmarried following divorce and widowhood. This study advanced prior work by accounting for whether the parent remarried after their divorce or widowhood. They found 49
divorce related to less parent-child contact, but they found no association for widowhood or for parents who remarry following either type of disruption. Despite these advancements to research on parent-child contact, this study is limited in several key ways. First, contact was
operationalized as visitation, which is a less common event, especially among parents and
children who live far from one another. Second, this study did not factor in the role of multiple
disruptions, which appear increasingly important when investigating well-being in later life
(Dupre & Meadows, 2007). Finally, this study encompassed a wider age range of respondents
(ages 30 to 89), and thus, it is unclear whether the same findings would be unearthed in a sample
of older adults.
To fill in these gaps, I propose several predictions about remarriage or remaining
unmarried following disruption. Regardless of disruption pathway, older adults who remarry will
have less contact with biological children. This is because remarriage tends to disrupts
relationship norms between parents and children as parents and children have to renegotiate their
relationships with one another (Kalmijn, 2015). Further, it is probable that individuals after one
widowhood will share more contact with children than older adults after one divorce or multiple
disruptions. Widowhood is a stressful, but unifying event for the surviving spouse and children
such that parents and children share more contact after widowhood (Guiaux et al., 2007; Roan &
Raley, 1996). Divorce, however, can disrupt family norms and create strain, and thereby will be
negatively associated with parent-child contact (Daatland, 2007; Nakonezny, Rodgers, &
Nussbaum, 2003). Similarly, although there is no research on multiple disruptions, reporting
multiple disruptions will likely be associated with less parent-child contact compared to one
widowhood or one divorce as multiple disruptions may create more turmoil in the parent-child
relationship. 50
Duration Remarried, Duration Unmarried, and Parent-Child Contact
Accounting for disruption pathways and subsequent remarriage will offer more insight into the relationship between marital biography and parent-child contact. Yet, it is critical to incorporate duration remarried or unmarried when investigating parent-child contact, especially since parent-child relationships evolve over time, and thus, it is likely these ties continue to fluctuate with years after the parent’s marital disruption or remarriage (Connidis, 2010).
Nonetheless, the extent to which duration remarried or unmarried following different disruption pathways relates to parent-child contact remains unknown. Moreover, the few studies that do consider the role of duration remarried or unmarried following disruption have primarily focused on widowhood, obscuring the potential relationship between parent-child contact and duration remarried or unmarried through one divorce or multiple disruptions.
Guiaux et al. (2007) examined relationships between parents and children following bereavement. They found although contact between parents and children increased following widowhood, after about two and a half years widowed, parent-child contact started to decline.
This study yields noteworthy findings about the role of duration following disruption, but it is limited in several aspects. First, this study only examined marital disruption through widowhood.
Widowhood tends to promote contact between parents and children (Roan & Raley, 1996) whereas divorce (and likely multiple disruptions) lowers parent-child contact (Daatland, 2007;
Kalmijn, 2007, 2013, 2015; Shapiro, 2003). Thus, the relationship between duration following one divorce or multiple disruptions and contact with children likely differs from widowhood.
Second, this study did not account for the role of duration remarried after widowhood. A number of older people choose to remarry after divorce or widowhood (Brown et al., in press), and it is important to understand the implications of years in a remarriage on parent’s contact with 51 children. Even though prior studies have not examined duration remarried versus unmarried, it is likely additional years remarried following different disruption pathways will yield less contact with children than years unmarried. This is because remarriage often strains parent-child ties and necessitates renegotiation of the parent-child relationship (Kalmijn, 2015; Nakonezny et al.,
2003), which could link to less contact relative to duration unmarried.
To better understand the role of duration remarried or unmarried following various disruption pathways, I propose several predictions. Remarriage disrupts bonds between parents and children (Kalmijn, 2015; Nakonezny, et al., 2003), and thus, contact after remarriage may be lower and each year remarried may be associated with less contact with children than for the unmarried. Additional years remarried will be associated with less contact than years unmarried.
For disruption pathways, regardless of remarried or unmarried status, it is likely that additional years widowed will be positively associated with parent-child contact relative to years divorced or following multiple disruptions. Even though widowed older adults will have greater contact with children when considering disruption pathway alone (Roan & Raley, 1996), as parents and children adjust to the bereavement, their contact with one another may decline more quickly relative to older adults after one divorce or multiple disruptions. Moreover, parents who have had one divorce or multiple disruptions, will likely have less contact with children due to disrupted relationship norms brought on by divorce or the strain of multiple disruptions. However, with additional years, parents and children will likely improve their relationships, thus, recover contact more quickly with children relative to parents after one widowhood.
Variation by Gender
Marital biography is influential for contact women and men share with children, but prior evidence indicates marital biography is more important when assessing men’s contact with 52 children versus women’s (Aquilino, 1994; de Graaf & Fokkema, 2007; Kalmijn, 2013, 2015;
Kaufman & Uhlenberg, 1998; Noël-Miller, 2013; Rosenthal, 1985; Shapiro, 2003; Silverstein &
Bengston, 1997). Women are described as kin-keepers, meaning they have closer relationships with children than do men (Kalmijn, 2013; Rosenthal, 1985). Moreover, women often report more contact than men regardless of the marital status they occupy (Albertini & Garriga, 2011;
Kalmijn, 2013, 2015; Pezzin & Schone, 1999; Shapiro, 2003).
Studies show divorce relates to less contact between parents and their children relative to most other marital status groups and this relationship is particularly salient for men (Albertini &
Garriga, 2011; Kalmijn, 2007, 2013; Pezzin & Schone, 1999; Shapiro, 2003). Although widowed parents share more contact with children, some evidence indicates men do not report as much contact with children as do women following bereavement perhaps due to the loss of the female kin-keeper (Kalmijn, 2007). Further, men who remarry after divorce are disadvantaged in contact with children relative to women in the same marital status (Kalmijn, 2007; Pezzin & Schone,
1999). Thus, I hypothesize that women will have more frequent contact with children than men regardless of the disruption pathway or whether the parent enters a remarriage following disruption.
The relationship between duration remarried and unmarried after different disruption pathways and parent-child contact for men relative to women has received little attention in current literature. Given evidence that marital status, remarriage, and marital disruptions matter more for the men’s relationships with children than women’s, it is likely the role of duration remarried or unmarried after any disruption pathway will operate similarly. Specifically, I predict duration remarried or unmarried following any disruption pathway will be associated with less frequent contact between men and their children relative to women. 53
The Present Study
Parent-child bonds endure over the entire life course, and parent-child contact is a
fundamental component of the intergenerational solidarity framework (Bengston & Oyama,
2010; Silverstein & Bengston, 1997). Ties between parents and children shift over time,
especially following changes in the parent’s marital status (Connidis, 2010; Ward et al., 2014),
which likely relates to parent-child contact. For instance, first married and widowed parents share more contact with children than divorced, remarried, cohabiting, and never married parents
(Albertini & Garriga, 2011; Kalmijn, 2013, 2015, Pezzin & Schone, 1999; Shapiro, 2003). Still, we know little about the role of remarriage following different disruption pathways when studying parent-child contact in later life. Moreover, few studies examine the role of duration remarried or unmarried following different disruption pathways. Additional years in status likely
allow parents to repair relationships, which may be positively related to contact with children or
if these ties erode further, associated with less contact.
To fill these gaps in the literature, I propose the following research questions. First, is
remarriage after disruption related to less contact with biological children and what is the role of
disruption pathway (i.e., one divorce, one widowhood, or multiple disruptions)? Second, is
duration remarried or unmarried after different disruption pathways positively or negatively
associated with parent-child contact? Finally, what is the role of remarriage after different
disruption pathways and does duration remarried or unmarried yield more or less contact for
women relative to men? I propose the following hypotheses to answer each research question.
Hypothesis 1a: Given the same disruption pathway, older adults who remarry will have
less contact with children than unmarried older adults. 54
Hypothesis 1b: Among remarried and unmarried older adults, those after one widowhood
will have the most contact with children, followed by parents after one divorce, and least by
parents after multiple disruptions.
Hypothesis 2a: Duration unmarried will be linked to higher frequency in contact with
children than duration remarried, no matter the disruption pathway.
Hypothesis 2b: With additional years in status, parents who have one divorce or multiple disruptions will likely report greater contact with children with each additional year in status relative to parents after one widowhood.
Hypothesis 3a: Disruption through one widowhood, one divorce, or multiple disruptions will be more negatively associated with the contact men who share with children relative to
women. Additionally, remarriage will be more detrimental for men than women, regardless of
disruption pathway.
Hypothesis 3b: Duration remarried and unmarried will be more detrimental for contact
between men and children compared to women regardless of the disruption pathway.
In this chapter, I also accounted for several covariates that may confound the association
between marital biography and parent-child contact, including demographic characteristics,
socioeconomic resources, health, and children’s characteristics.
Demographic Characteristics
As parents grow older, they report more contact with children because aging parents
often have smaller social networks (Cartensen, 1992; Umberson, 1992). Older age also yields
time to accrue marital transitions, or conversely, longer duration in the current marital status
(Cherlin, 2009; Goldstein & Kenney, 2001). Minorities have more cohesive family networks,
and thus, share more contact with children than do Whites (Lawton et al., 1994; Umberson, 55
1992). In addition, Whites and Hispanics have similar rates of marriage and marital stability
(Amato, 2010; Bramlett & Mosher, 2002). In comparison, Blacks are less likely to marry than
Whites and Hispanics, and when they do marry, they have a greater risk of marital disruption
(Heaton, 2002; Manlove et al., 2012).
Socioeconomic Resources
Education tends to be inversely related to contact with children. Families with highly educated members are often more spread out geographically, which relates to less contact
(Lawton et al., 1994). The highly educated also have more stable marital biographies than the less educated as they delay marriage and remain in stable unions (de Graaf & Kalmijn, 2006;
Heaton, 2002; Manning et al., 2014). Comparatively, the less educated have a higher incidence of marital disruption (Amato, 2010; Heaton, 2002). Financial resources (e.g., assets) are positively related to education, and thus, more wealth implies parents and children live farther from one another, potentially lowering contact. The wealthy also enjoy more marital stability
(Heaton, 2002), although greater marital stability positively relates to wealth as well.
Specifically, researchers have demonstrated financial costs of marital disruption, especially for the divorced (Hungerford, 2001; Lin et al., 2017). Contact with children tends to be positively related to retirement (Damman & van Duijn, 2016). Retirement contributes to marital strain, leading to potential disruption, which fosters marital biography instability for retirees, although this has not received consistent support in current studies (Lin et al., 2017; Dew & Yorgasen,
2010).
Health
Parents’ poor health relates to more contact with children as offspring reach out to parents who are facing health problems or parents may ask their adult children for more support 56
(Ha, Khang, & Choi, 2017). In addition, the link between marital biography and health could be
reciprocal. Marital biographies that include many transitions like both divorces and widowhoods
are associated with poorer health (Brockmann & Klein, 2004; Dupre et al., 2009; Dupre &
Meadows, 2007; Hughes & Waite, 2009; McFarland et al., 2013), but poor health increases the
likelihood of marital disruption (Karraker & Latham, 2015), despite some conflicting evidence
suggesting there is no such association (Lin et al., in press).
Children’s Characteristics
Daughters are more likely to stay in frequent contact with parents than sons (Umberson,
1992). Older adult children have more competing family demands than younger adult children, like employment, partners, and childcare, which are likely associated with less parent-child contact (Marks, 1998; Sarkisian & Gerstel, 2008). Children who are partnered or employed report less contact with children as partnership and employment add competing demands to children’s lives, interfering with contact (Noël-Miller, 2013; Sarkisian & Gerstel, 2008). Finally, parents tend to report more contact with children from the most recent or current marriage versus a prior union, a phenomenon known as swapping families (Manning & Smock, 2000).
Method
Data came from the Health and Retirement Study (HRS), spanning from 1992 to 2012 and the HRS RAND Family File from 2012. The HRS includes a nationally representative sample of adults age 50 and older. To be eligible, respondents must be noninstitutionalized at baseline and live in a household. Respondents are reinterviewed every two years. Every six years, the HRS incorporates a new cohort of older adults aged 51 to 56 to replenish the sample and maintain representativeness. Response rates for the HRS are relatively high. Baseline rates 57 hover around 70-82%, increasing to about roughly 90% or higher for follow-up interviews. The
HRS oversamples Blacks, Hispanics, and respondents living in Florida.
Marital biography measures were taken from the 1992 through 2012 waves of the HRS.
Children’s characteristics come from the RAND Family File to use information on each of the respondent’s biological children. Current marital status, parent-child contact, and other covariates were from the 2012 core survey. I began sample selection by compiling a file with marriage information from 1992 to 2012, including marital history. The initial HRS sample included 37,495 respondents. First, I chose only respondents in the 2012 HRS survey (n =
20,554). Next, I removed respondents under 50 in 2012, yielding 19,906 respondents. I excluded nursing home residents, leaving 18,829 respondents. I omitted respondents without valid information on children in the synthesized RAND Family File, resulting in 17,572 respondents. I excluded respondents without living, biological children over 17 years of age (n = 1,206). Last, I omitted respondents with missing values on contact with children, yielding a final sample size of
15,569 respondents with 43,842 children. Of the respondents, 9,076 were women and 6,493 were men.
Dependent Variable
Contact with Children. Frequency of contact with each child came from a question asking about contact in person, by phone, or mail. Contact with children was coded as: never (coded 1), less than once a month (coded 2), more than once a month, but less than once a week (coded 3), once or twice a week (coded 4), and three times a week or more (coded 5). The HRS did not query respondents about frequency of contact with children who live in the household, so parents were assumed to have daily contact with resident children (coded 5).
58
Focal independent variables
Marital Status. Marital status included six categories: married, remarried (reference category), cohabiting, divorced, widowed, or never married in 2012.
Disruption Pathways. Marital disruption was used to construct pathways through which individuals arrived at their current marital status. This measure included the following categories: first married, remarried after one divorce (reference category), remarried after one widowhood, remarried after multiple disruptions, cohabiting, unmarried after one divorce, unmarried after one widowhood, unmarried after multiple disruptions, and never married. For those after multiple disruptions, 63% had two or more divorces, 22% had one divorce and one widowhood, 5% had two or more widowhoods, and 10% reported three disruptions including both divorces and widowhoods (e.g., two divorces and one widowhood). Respondents who were remarried after multiple disruptions made up a little over 5% of the full sample whereas the unmarried after multiple disruptions composed about 9% of the full sample.
Duration Remarried or Unmarried. Duration was a continuous measure that captured years in the current remarried or unmarried state following the most recent marital disruption.
Covariates
Age was measured using three categorical variables coded into ages 50 to 64, ages 65 to
84 (reference category), and ages 85 and older. Race and ethnicity comprised a categorical variable indicating whether the respondent identified as White (reference category), Black,
Hispanic, or of other races. Education was coded into four categories: less than high school, high school (reference category), some college, and college or more. The distribution of household assets was highly skewed, and thus, I employed a series of five categories to capture household wealth: in debt, $0 to $50,000 (reference category), $50,001-$100,000, $100,001-$250,000, and 59
$250,001 or more. An indicator of work status measured whether the respondent was currently employed full or part-time (1 = Yes, 0 = No). Mental health was captured by depressive symptoms, which summed eight items in which the respondent reported symptoms during the prior week: feeling depressed, everything was an effort, restless sleep, unhappiness, loneliness, not enjoying life, sadness, or unable to get going. Depressive symptoms ranged from 0 to 8 and the inter-item reliability was .80. Physical health was a continuous measure summing the number of the respondent’s chronic conditions diagnosed by a physician and included psychiatric illness, diabetes, heart disease, cancer, lung disease, hypertension, stroke, and arthritis. Chronic conditions ranged from 0 to 8.
Several covariates were used to capture characteristics of the respondent’s biological children for whom they reported contact frequency. Children’s gender was coded 1 for sons and
0 for daughters. Children’s age was measured in years. Partnership status measured whether children were married/partnered (coded 1) or unmarried (coded 0). I accounted for whether the child was working at least 30 hours a week (coded 1) or not (coded 0). Finally, I included an indicator of whether the child was from the current marriage or the most recently dissolved union
(coded 1) or from a prior union (coded 0).
Analytic Strategy
I conducted two sets of analyses. First, I displayed descriptive statistics by gender for all variables included in the analyses. Second, I employed three ordered logit regressions to assess the relationship between marital biography and parent-child contact. I used multilevel models as many respondents had more than one biological child. Biological children (Level 1) were clustered within respondents (Level 2). Multilevel modeling is appropriate as it accounts for dependence of observations, as children from the same family are likely more similar to each 60 other than children from another family (Hedeker, 2008). The first model included current marital status to evaluate how marital status alone relates to parent-child contact. The second model examined disruption pathways as well as current marital status to determine the role of disruption and remarriage on parent-child contact. Finally, I estimated a model including duration, current marital status, and disruption pathways to examine the role of duration remarried or unmarried on contact with children. Interactions between duration and each marital status category were included in the model although only relevant comparisons were shown in
Table 3.4 (complete results available upon request). Each of the multivariate models were estimated separately for men and women. This was done by pooling a model that included gender interactions with the components of marital biography. Missing values ranged from less than 1% on children’s partnership status to about 8% on whether the child was from the most recent union. Missing values were imputed using the mean for continuous variables and mode for categorical variables due difficult model convergence when using data imputed through multiple imputation. Each model was weighted to correct for unequal probability of selection and sample attrition (Ofstedal et al., 2011).
Descriptive Results
Table 3.1 displays the weighted descriptive statistics for all variables included in the analyses for men and women. Most parents had contact with children at least three times a week, and, women had more frequent contact with children than men (57% of women versus 48% of men had contact with children at least three times a week). Over half of all men (about 51%) were in a first marriage compared to 38% of women. More men than women were remarried—a little more than a quarter of men were in a remarriage compared to 17% of women. Among the remarried, about 67% of men were remarried after one divorce whereas close to 61% of women 61
occupied this status. Similar shares of men and women were remarried after one widowhood (8%
of men and about 9% of women). Slightly greater than one-quarter of men were remarried after
multiple disruptions compared to 30% of women. Fewer women than men were cohabitors (2% versus 4%). More women than men were divorced (about 16% of women and 13% of men).
Around a quarter of women were widowed whereas fewer than 7% of men reported being
widowers. Among all unmarrieds, more men were unmarried after one divorce than women
(37% versus 24%). More women than men were unmarried through widowhood— about 46% of
women versus a little over one quarter of men. Nearly 37% of men were unmarried through
multiple disruptions, whereas only about 30% of women occupied this status. Finally, less than
1% of men were never married versus a little over 2% of women. Men had spent slightly longer
in their current marital status than women—about 31 years versus 28 years for women.
Men in the sample were younger than women on average. More men fell between the
ages of 50 and 64 than women (50% of men versus 49% of women) whereas more women were
aged 85 or older relative to men (7% of women versus 5% of men). A greater percentage of men
in the sample were White than women (75% of men versus 73% of women) whereas more
women were Black (12% of women versus 11% of men). Men were more educated than women,
on average. A greater percentage of men had a college degree than women (30% versus 20%).
Men reported more wealth than women. A greater percentage men reported being in the top
wealth category of $250,001 or greater assets than women (29% of men versus 24% of women).
Fewer women were working than men—nearly 28% of women were currently employed
whereas 38% of men were working. Women reported worse physical and mental health than
men. Women had about 2.4 chronic conditions on average compared to 2.3 for men. Women 62
experienced about 1.5 depressive symptoms in the week prior to the interview whereas men
reported only 1.1 symptoms.
Among children, about half were sons. A little less than two-thirds of women and men’s children were currently partnered. Men reported children who were about 38 years old on average whereas women had children who are about 41 years old. Around 80% of men and women had children who were currently employed. Finally, about 68% of children were from men’s most recent unions whereas a little under 72% of children were from women’s most recent unions.
Multivariate Results
Marital Status and Parent-Child Contact
Table 3.2 considers how current marital status is related to parent-child contact. Among
men, the first married reported greater contact with children compared to all other marital status
groups, including the remarried (i.e., superscript refers to differences between the first married
and “a” the divorced, “b” the widowed, and “c” the never married). However, remarried men did
not report greater contact with children than cohabitors. The widowed reported more contact
with children than the remarried while the divorced had less contact with children relative to the remarried. As predicted, never married men reported less contact than the remarried. Among unmarried older men, the divorced and never married had less contact with children than their widowed counterparts (superscript “e” refers to the difference between the divorced and widowed while “f” denotes the difference between the widowed and never married).
As for women, the findings also mirror conclusions from prior research. The first married reported more contact with children than remarried, cohabiting, and divorced women
(superscript denotes differences between the first married and “a” cohabitors and “b” the 63
divorced). Cohabiting women also had less contact with children relative to the remarried. For
remarried women relative to unmarried women, the widowed reported more contact with
children than remarried women. There were no differences between remarried women and the
divorced or never married. Among unmarried women, widowed women reported more contact
with children than divorced and never married women (superscript refers to differences between
the widowed and “c” the divorced and “d” the never married). I also tested differences between men and women. First married and widowed men reported more contact with children than their female counterparts whereas divorced, remarried, and never married women reported greater contact with children than their male counterparts (differences between men and women were denoted by underlined coefficients or table note).
Remarriage, Disruption Pathways, and Parent-Child Contact
Table 3.3 adds to the prior models by examining whether different disruption pathways and subsequent remarriage were associated with parent-child contact for men and women. In my
first hypothesis, I expected given the same disruption pathway, the remarried would have less
contact with children than the unmarried (Hypothesis 1a). Contrary to this hypothesis, compared
to men who remarried after divorce, those who remained unmarried after one divorce had less
contact with children. I also hypothesized that among either remarried or unmarried older adults,
those who had one widowhood would report the greatest contact, followed by one divorce, and
finally, least by parents who experienced multiple disruptions (Hypothesis 1b). Supporting my
hypothesis, men who remarried after one widowhood had more contact with children than
remarried men after one divorce. Moreover, those who remarried after multiple disruptions had
less contact with children than the remarried after one widowhood or after one divorce
(difference denoted by superscript “a”). Also consistent with my predictions, among unmarried 64 men, widowed men had more contact with children than divorced men and men who had experienced multiple disruptions (superscript refers to differences between unmarried men with one widowhood and “b” one divorce and “d” multiple disruptions). However, unmarried men who have experienced multiple disruptions reported more contact than divorced men (difference indicated by superscript “c”).
For women, as hypothesized, being unmarried following either one widowhood or multiple disruptions yielded more contact with children than their remarried counterparts
(Hypothesis 1a) (superscript denotes differences between the remarried and unmarried for “b” the widowed and “c” individuals with multiple disruptions). As anticipated, compared to remarriage after one divorce or after one widowhood, women who remarried after multiple disruptions had less contact with children (Hypothesis 1b) (difference denoted by superscript
“a”). Among the unmarried, women after one divorce or multiple disruptions had less contact than women who were unmarried after widowhood (superscript denotes differences between the unmarried after one widowhood and “d” after one divorce and “e” after multiple disruptions). I predicted all disruption pathways for men would be more negatively associated with parent-child contact relative to women (Hypothesis 3a). When testing this hypothesis, two differences between men and women emerged (gender difference indicated by underlined coefficients or table note). As expected, remarried men with one divorce had less contact with children relative to their female counterparts with one divorce. Moreover, in alignment with my hypothesis, men who remained unmarried after one divorce reported less contact with children compared to women occupying the same status.
65
Duration Remarried, Duration Unmarried, and Parent-Child Contact
Finally, Table 3.4 accounts for duration remarried or unmarried to see whether parent-
child relationships recover with years remarried or unmarried following different disruption
pathways. I expected duration unmarried to be positively linked to parent-child contact relative
to duration remarried, regardless of the disruption pathway (Hypothesis 2a). I did not find support for this prediction for men. However, for women, contrary to my expectation, I found that compared to women who remarried following widowhood, those who remained widowed
had less contact with children with longer duration unmarried (denoted by superscript “d”). I also
hypothesized that additional years either remarried or unmarried with one prior divorce or
multiple disruptions, parents will be able to recover contact with children relative to those after
one widowhood (Hypothesis 2b). Supporting this prediction, among remarried men, each additional year of remarriage after widowhood yields less contact with children relative to remarried men who have had multiple disruptions (indicated by superscript “d”). For women, no significant findings regarding duration among different disruption pathways emerged.
Across all of the models, the relationship between other covariates and parent-child contact remained relatively consistent. Based on the Table 3.4, relative to men aged 65 to 84, men who were 85 or older had more contact with children. Black men reported less contact with children. Having less than a high school education was also inversely related to contact between men and children than men with a high school education. Men reported less contact with sons and children who were currently partnered, employed, and older whereas men reported more contact with children from their most recent union relative to their respective counterparts.
As for women, those who belonged in the 85 and older age category in addition to employed women had more contact with children. Women who were Black or Hispanic had 66
more frequent contact with children than White women. College education was negatively
associated with women’s contact with children. Women had less contact with sons, partnered children, older children, or who were currently employed whereas women reported more contact with children who were from their most recent union.
Discussion
Parent’s ties with children are crucial over the life course. Parents tend to be primary providers of support to children from birth through adulthood (Bengston & Oyama, 2010;
Silverstein & Bengston, 1997). As parents grow older, they come to rely on children for financial, emotional, and social support (Rossi & Rossi, 1990; Silverstein & Bengston, 1997).
Despite the strength of bonds between parents and children, transitions in the parent’s life, like those related to marital transitions, can create change in the parent-child tie (Connidis, 2010;
Hammersmith, in press), altering contact with children (Albertini & Garriga, 2011; Daatland,
2007; Kalmijn, 2007, 2013, 2015; Noël-Miller, 2013; Shapiro, 2003; Ward, et al., 2014). Parent- child contact promotes parental well-being and is indicative of strength of support avenues for older people (Bengston & Roberts, 1991; Dykstra & de Jong Gierveld, 2006; Lawton et al.,
1994).
The findings from this chapter indicate the remarried reported less contact with children
than the first married and widowed, consistent with prior research (Dykstra & de Jong Gierveld,
2006; Kalmijn, 2013; Pezzin et al., 2008; Roan & Raley, 1996; Shapiro, 2003). Moreover, for the most part, the divorced and never married had less contact with children than the widowed, which aligns with research suggesting parents and children come together following widowhood, which promotes contact (Roan & Raley, 1996). Interestingly, the findings also indicated that remarried women reported more contact with children than cohabiting women. Further, divorced 67
men reported less contact with children than remarried men. This finding is somewhat surprising
given prior studies showing remarriage is negatively related to contact with adult children
(Kalmijn, 2007, 2015; Noël-Miller, 2013; Pezzin et al., 2008). Still, this model only considers
current marital status, not disruption pathways. Therefore, this model did not account for
differences among men who had remarried following divorce or widowhood. The findings
regarding current marital status also indicate divorced and never married men had less contact
with children than women whereas widowed and first married men reported more contact
relative to their female counterparts. This finding is somewhat consistent with prior studies suggesting men tend to have less contact with children relative to women.
This chapter also accounts for remarriage after different disruption pathways to add to our understanding of parent-child ties in later life. Specifically, I expected that remarriage would be associated with less contact with children relative to the unmarried, no matter the disruption pathway (Hypothesis 1a). Among men, the results showed that remarriage mattered, however, not in the direction that I hypothesized. Men who remarried after one divorce had more contact with children than men who remained unmarried after one divorce. This finding is counterintuitive as prior work suggests fathers who repartner after divorce have less contact with children than divorced fathers who remain unpartnered (Kalmijn, 2007; Noël-Miller, 2013).
However, these prior studies did distinguish between other disruption pathways, specifically fathers who experience multiple disruptions. Thus, it is possible men who remarry after one divorce report more contact with children than their unmarried counterparts, because they gain a partner who prompts them to strengthen bonds with children through acting as a kin-keeper
(Rosenthal, 1985). Remarriage relative to being unmarried also mattered among women. As predicted, women who remarried after one widowhood or after multiple disruptions reported less 68
contact with children than their unmarried counterparts. It is possible that when women remarry, parent-child ties become especially strained as parents and children renegotiate their changing relationships, yielding less contact between mothers and their children (de Jong Gierveld &
Merz, 2013).
Disruption pathways also matter, as posited in Hypothesis 1b. Whether remarried or unmarried, widowed men and women often had more contact with their children compared to those after one divorce or after multiple disruptions. Generally, men and women after multiple disruptions had less contact with children relative to the widowed and the divorced. These findings are supported by literature that shows the widowed tend to benefit from more contact with children than other marital status groups (Roan & Raley, 1996). There was one notable inconsistency. Unmarried men who had multiple disruptions were more likely to have greater contact with children relative to unmarried men after one divorce. This finding is perhaps explained by the fact that for most unmarried men (and women) who have experienced multiple disruptions, close to half had at least one widowhood (about 45%), and widoweds tend to have more frequent contact with children than other marital status groups (Roan & Raley, 1996).
Moreover, I found unmarried men after one divorce had less contact with their children
compared to their female counterparts. This finding is supported by prior work indicating divorce
is more harmful for men’s ties with children than women’s (Daatland, 2007; Kalmijn, 2007;
Shapiro, 2003).
This chapter adds to the current body of research by demonstrating duration remarried or
unmarried after different disruption pathways is an important factor when explaining parent-
child contact. In alignment with my predictions, duration remarried after multiple disruptions
was positively associated with men’s contact with children relative to remarried men after one 69
widowhood (Hypothesis 2b). This finding suggests even though remarried men who have had
multiple disruptions had less contact with children relative to men with one widowhood when
just considering their disruption pathway, duration allows these men after multiple disruptions to rebuild bonds with children leading to more contact. Despite this, Hypothesis 2b was not supported for women, suggesting that no matter the disruption pathway, additional years in a particular status were not associated with variation in women’s contact with their children.
I also expected duration remarried to be associated with less contact with children relative to the unmarried with any disruption pathway (Hypothesis 2a). In contrast to my prediction, duration remarried after widowhood among women related to more contact with children over time relative to years unmarried after one widowhood. Thus, when examining remarriage and disruption on their own (Hypothesis 1a) without factoring in duration, remarried women after one widowhood had less contact with children than their unmarried counterparts.
These findings suggest women may be able to rebuild bonds with children as duration remarried increases, which was positively related to contact with children relative to their unmarried counterparts. Remarriage is often accompanied by a period of adjustment for both the parent and their children, and thus, additional years remarried may allow for parents and children to reconcile their new relationship norms (de Jong Gierveld & Merz, 2013; Nakonezny et al.,
2003), thereby increasing contact. For men, there were no differences in whether years were spent unmarried or remarried following different pathways of disruption. This may be attributable to strained relationships that men report with children regardless of whether they have experienced a marital disruption or a repartnership (Kalmijn, 2007, 2013; Noël-Miller,
2013). 70
This chapter has several limitations. First, I could not account for cohabitation following
different disruption pathways due to small sample sizes of cohabitors in the HRS. Future
research should incorporate cohabitation after different disruption pathways if data permit, as we
know little about how marital histories of cohabitors relate to parent-child contact in later life.
Second, the analyses relied solely on parent’s reports of contact with children. Like any dyadic
relationship, there are two sides of every story, and it is unclear whether children would agree
with the parent’s assessment of the relationship (Lin & Wu, 2018; Shapiro, 2004). Finally, I did
not account for stepchildren, but few stepchildren remain in contact with stepparents after
separating from the child’s biological parent (Noël-Miller, 2013).
Nevertheless, this dissertation advances current research by deepening our knowledge of
the relationship between marital biography and parent’s contact with their biological children.
Specifically, this work contributes to current literature by showing the importance of accounting for remarriage, different disruption pathways—specifically, multiple disruptions—as well as
duration. Interestingly, remarriage did not operate completely as anticipated. Whereas unmarried
women after one widowhood or multiple disruptions had more contact with children than their
remarried counterparts, surprisingly, unmarried men after one divorce had less contact than
remarried men after one divorce. These findings suggest remarriage may not be uniformly
associated with less contact with adult children. As for disruption pathway, this chapter adds to
current literature by showing that older people after multiple disruptions almost always reported
less contact with children relative to either the divorced or widowed.
Finally, this research shows that accounting for how parent-child contact may be more or
less frequent with additional years remarried or unmarried contributes to our understanding of
the relationship between marital biography and parent-child contact. Although remarried women 71
had less contact with children than their unmarried counterparts when considering disruption
pathway alone, factoring in duration changes our understanding of this relationship. Specifically,
following one widowhood for women, remarriage can improve parent-child contact with
additional years remarried. Moreover, for men, when accounting for disruption pathway alone,
the remarried following multiple disruptions reported less contact than their male counterparts
after one widowhood. When factoring in duration, I found additional years remarried following
multiple disruptions yielded greater contact with children relative to remarried men after one widowhood. This finding indicates that even though older people after multiple disruptions face poorer well-being outcomes, they are likely able to rebuild their ties with children, who then may provide social support as they age (Bengston & Roberts, 1991; Dupre & Meadows, 2007).
Finally, consistent with prior work, the findings from this chapter regarding gender differences also suggest that unmarried or remarried men after one divorce had less contact with children than did their female counterparts (Daatland, 2007; Shapiro, 2003). However, these were the only findings that supported the predictions regarding differences between men and women, suggesting that perhaps gender is less important when studying the association of marital biography with well-being in later life (Manzoli et al., 2007).
In light of these findings and growing instability associated with marital biography in later life, it is essential to continue to understand how aspects of the parent-child bond relate to remarriage, different disruption pathways, as well as duration remarried or unmarried. Thus, this chapter informs researchers, policymakers, and practitioners on ways they can work toward identifying vulnerable groups of older people who may lack social support resources as well as work toward devising relevant interventions to well-being in later life. Moreover, parent-child contact is indicative of affection between parents and their children, which could relate to 72 positivity that parents feel toward children (Bengston & Roberts, 1991; Lawton et al., 1994).
However, frequent contact could also allow for more occurrences of conflict between parents and children, relating to negativity toward one’s offspring. The comingling of positivity and negativity often characterizes parent-child relationships, which are better known as feelings of ambivalence (Lüscher & Pillemer, 1998). Thus, the next chapter examines the relationship between remarriage, different disruption pathways, and duration with parent’s ambivalence toward children. 73
Table 3.1 Weighted Means (Standard Deviations) and Percentages by Gender Men Women Contact with children Never 4.63 2.35 *** Less than 1 time a month 11.35 6.96 *** Less than 1 time a week, more than 1 time a month 9.74 7.20 *** Once or twice a week 26.59 27.03 *** At least three times a week 47.69 56.46 *** Marital biography First marriage 50.89 38.21 *** Remarriage 25.65 16.50 *** After one divorce 66.90 60.85 ** After one widowhood 7.60 9.21 After multiple disruptions 25.50 29.94 * Cohabiting 3.47 2.43 *** Divorced 12.46 16.12 *** Widowed 6.69 24.47 *** Unmarried after one divorce 36.53 23.87 ** Unmarried after one widowhood 26.83 46.02 *** Unmarried after multiple disruptions 36.64 30.11 *** Never married 0.84 2.27 *** Duration in current status 30.86(16.8) 28.09(17.2) *** Demographic characteristics Age 50 to 64 50.42 48.99 65 to 84 44.70 43.87 85 and older 4.88 7.14 *** Racial and ethnic background White 75.27 73.24 * Black 10.62 11.97 *** Hispanic 10.40 11.70 Other race 3.71 3.09 * Socioeconomic resources Education Less than high school 18.77 22.4 * High school 28.20 33.88 *** Some college 22.91 23.93 * College or more 30.12 19.79 *** Wealth In debt 8.63 8.92 $1 to $50,000 38.50 44.92 *** $50,001 to $100,000 9.99 8.96 $100,001 to $250,000 13.93 13.20 $250,001+ 28.95 24.00 *** Currently employed 37.60 27.66 *** Health Chronic conditions 2.26(1.5) 2.36(1.5) *** Depressive symptoms 1.09(1.8) 1.53(2.1) *** Weighted percentages 44.75 55.25 Number of parents 6,493 9,076 74
Table 3.1 continued Men Women Children's characteristics Son 50.71 49.27 * Partnered 61.44 64.64 Age 37.92(11.0) 41.41(11.7) *** Full or part-time work 81.36 78.34 *** From recent union 68.44 71.54 Number of children 17,869 25,373 * p < .05; ** p < .01; *** p < .001 75
Table 3.2 Coefficients (Standard Errors) from the Weighted Multilevel Ordered Logistic Regression Model of Marital Status on Contact with Children Men Women Marital biography First marriage 1.04 (.11) ***abcd 0.52 (.09) ***ab Remarriage (ref) -- -- Cohabiting - 0.24 (.21) a - 0.39 (.16) a* Divorced - 0.83 (.14) ***be 0.14 (.10) bc Widowed 0.61 (.15) ***cef 0.56 (.09) ***cd Never married - 1.04 (.33) **df 0.14 (.23) d Demographic characteristics Age 50 to 64 - 0.16 (.10) - 0.04 (.08) 65 to 84 (ref) -- -- 85 and older 0.42 (.11) *** 0.34 (.09) *** Racial and ethnic background White (ref) -- -- Black - 0.35 (.11) ** 0.29 (.08) *** Hispanic 0.10 (.14) 0.21 (.09) * Other race 0.25 (.19) - 0.00 (.16) Socioeconomic resources Education Less than high school - 0.24 (.11) * - 0.08 (.07) High school (ref) -- -- Some college 0.07 (.10) - 0.02 (.07) College or more - 0.08 (.09) - 0.31 (.07) *** Wealth In debt - 0.11 (.14) - 0.07 (.10) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.08 (.12) - 0.02 (.09) $100,001 to $250,000 - 0.03 (.11) 0.06 (.08) $250,001+ 0.08 (.09) 0.00 (.07) Currently employed 0.16 (.09) 0.19 (.08) * Health Depressive symptoms - 0.04 (.02) - 0.02 (.01) Chronic conditions 0.03 (.03) - 0.02 (.02) Children's characteristics Son - 0.38 (.04) *** - 0.75 (.04) *** Partnered - 0.25 (.05) *** - 0.52 (.04) *** Age - 0.06 (.00) *** - 0.05 (.00) *** Full or part-time work - 0.45 (.06) *** - 0.34 (.05) *** From recent union 1.40 (.10) *** 0.28 (.07) *** Constant 2.52 (.16) 1.67 (.09) Number of Respondent (Level 2) 6,493 9,076 Number of Children (Level 1) 17,869 25,373 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing same superscript letter denote significant differences in the same model at p < .05 Underlined coefficients denote significant differences between men and women at p < .05 Remarried men and women are significantly different at p < .05
76
Table 3.3 Coefficients (Standard Errors) from the Weighted Multilevel Ordered Logistic Regression Model of Marital Status, and Disruption Pathway on Contact with Children Men Women Marital biography First marriage 0.91 (.11) 0.46 (.11) Remarriage After one divorce (ref) -- -- After one widowhood 0.49 (.17) **a 0.12 (.17) ab After multiple disruptions - 0.64 (.16) ***a - 0.38 (.14) **ac Cohabiting - 0.36 (.21) - 0.49 (.16) Unmarried After one divorce - 1.09 (.18) **bc - 0.03 (.13) d After one widowhood 0.49 (.17) bd 0.62 (.11) bde After multiple disruptions - 0.55 (.17) cd 0.19 (.11) ce Never married - 1.18 (.33) 0.08 (.24) Demographic characteristics Age 50 to 64 - 0.16 (.10) - 0.04 (.08) 65 to 84 (ref) -- -- 85 and older 0.42 (.12) *** 0.29 (.09) ** Racial and ethnic background White (ref) -- -- Black - 0.36 (.11) ** 0.28 (.08) *** Hispanic 0.09 (.14) 0.19 (.09) * Other race 0.22 (.18) - 0.01 (.16) Socioeconomic resources Education -- -- Less than high school - 0.21 (.11) * - 0.08 (.07) High school (ref) -- -- Some college 0.06 (.10) - 0.02 (.07) College or more - 0.07 (.09) - 0.31 (.07) *** Wealth In debt - 0.12 (.14) - 0.07 (.10) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.07 (.12) - 0.01 (.09) $100,001 to $250,000 - 0.03 (.11) 0.06 (.08) $250,001+ 0.05 (.09) - 0.01 (.07) Currently employed 0.16 (.09) 0.19 (.08) * Health Depressive symptoms - 0.04 (.02) - 0.02 (.01) Chronic conditions 0.03 (.03) - 0.02 (.02)
77
Table 3.3. continued Men Women Children's characteristics Son - 0.38 (.04) *** - 0.76 (.04) *** Partnered - 0.25 (.05) *** - 0.52 (.04) *** Age - 0.06 (.00) *** - 0.05 (.00) *** Full or part-time work - 0.45 (.06) *** - 0.34 (.05) *** From recent union 1.42 (.10) *** 0.23 (.08) ** Constant 2.50 (.16) 1.66 (.09) Number of Respondent (Level 2) 6,493 9,076 Number of Children (Level 1) 17,869 25,373 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing same superscript letter denote significant differences in the same model at p < .05 Underlined coefficients denote significant differences between men and women at p < .05 Men who are remarried after one divorce and women are significantly different at p < .05
78
Table 3.4 Coefficients (Standard Errors) from the Weighted Multilevel Ordered Logistic Regression Model of Marital Status, Disruption Pathway, and Duration on Contact with Children Men Women Marital biography First marriage 1.31 (.30) 0.83 (.29) Remarriage After one divorce (ref) -- -- After one widowhood 0.79 (.31) *a - 0.07 (.33) a After multiple disruptions - 0.95 (.29) **a - 0.10 (.29) Cohabiting 0.43 (.33) 0.07 (.30) Unmarried After one divorce - 1.09 (.36) b 0.22 (.29) c After one widowhood 0.40 (.29) bc 0.85 (.24) ***abc After multiple disruptions - 0.48 (.27) c 0.29 (.23) b Never married - 1.44 (.38) *** 0.20 (.31) Duration 0.01 (.01) 0.01 (.01) Remarried after one divorce (ref) -- -- Remarried after one widowhood - 0.02 (.01) d 0.01 (.01) d Remarried after multiple disruptions 0.02 (.01) d - 0.01 (.01) Unmarried after one divorce - 0.01 (.01) - 0.01 (.01) Unmarried after one widowhood - 0.03 (.01) - 0.01 (.01) d Unmarried after multiple disruptions - 0.01 (.01) - 0.00 (.01) Demographic characteristics Age 50 to 64 - 0.19 (.10) - 0.05 (.09) 65 to 84 (ref) -- -- 85 and older 0.51 (.12) *** 0.31 (.09) *** Racial and ethnic background White (ref) -- -- Black - 0.29 (.11) * 0.30 (.08) *** Hispanic 0.11 (.14) 0.20 (.09) * Other race 0.24 (.18) - 0.01 (.16) Socioeconomic resources Education Less than high school - 0.22 (.11) ** - 0.07 (.07) High school (ref) -- -- Some college 0.04 (.14) - 0.02 (.07) College or more - 0.10 (.09) - 0.32 (.07) *** Wealth In debt - 0.10 (.14) - 0.08 (.10) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.08 (.14) - 0.01 (.09) $100,001 to $250,000 - 0.02 (.11) 0.06 (.08) $250,001+ 0.07 (.09) - 0.01 (.07) Currently employed 0.16 (.09) 0.19 (.08) * Health Depressive symptoms - 0.04 (.02) - 0.02 (.01) Chronic conditions 0.03 (.03) - 0.02 (.02)
79
Table 3.4 continued Men Women Children's characteristics Son - 0.38 (.04) *** - 0.76 (.04) *** Partnered - 0.25 (.05) *** - 0.52 (.04) *** Age - 0.06 (.00) *** - 0.05 (.00) *** Full or part-time work - 0.45 (.06) *** - 0.34 (.05) *** From recent union 1.59 (.11) *** 0.26 (.08) ** Constant 2.48 (.16) 1.65 (.09) Number of Respondent (Level 2) 6,493 9,076 Number of Children (Level 1) 17,869 25,373 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing same superscript letter denote significant differences in the same model at p < .05
80
CHAPTER 4: MARITAL BIOGRAPHY AND PARENT’S AMBIVALENCE TOWARD
CHILDREN IN LATER LIFE
Older adulthood in the United States is changing, especially as it relates to marriage trends. In the past, it was normative to marry and stay married until the death of one spouse
(Cherlin, 2009). However, this marital pathway has become less common. Today, higher order marriages are much more common among older people. In fact, in 1980, only 19% of married older adults were in a remarriage, however, this number had risen to nearly 30% in 2015 (Lin et al., in press). This rise in remarried older adults suggests a greater number of older people experience divorce or widowhood before reaching or during their later years. Along with this surge in marital instability during early and mid-life, marital trends are shifting in later life, too.
Gray divorce—divorce involving people 50 and older—is on the rise. In 1990, only 1 in 10 divorces was attributed to someone age 50 or older, whereas this had risen to 1 in 4 in 2010
(Brown & Lin, 2012). These shifting marital trends exemplify greater marital instability in later life. To capture this growing instability associated with marriage, researchers have increasingly studied marital biography. Marital biography accounts for not only current marital status, but also marital disruption, duration, timing, and sequencing (Zhang et al., 2016). All components of marital biography may be useful for gaining a fuller understanding of parent’s ambivalence toward children in later life, however, this chapter will focus on current marital status, marital disruption, and duration, as these factors appear most prevalent for parent-child ties (Albertini &
Garriga, 2011; Daatland, 2007; Ha & Ingersoll-Dayton, 2008; Hammersmith, in press; Kaufman
& Uhlenberg, 1998).
Parents and children share a close, intimate bond over the life course (Bengston &
Oyama, 2010; Birditt et al., 2009; Seltzer & Bianchi, 2013). Even though parent-child bonds are 81 strong over the life course, they are not static over time (Birditt et al., 2009). Parent’s divorce, widowhood, or remarriage, can bring about conflict or closeness between parents and children
(Barrett & Lynch, 1999; Daatland, 2007; de Jong Gierveld & Merz, 2013; Eggebeen, 1992; Ha,
Carr, Utz, & Nesse, 2006; Kaufman & Uhlenberg, 1998; Pezzin & Schone, 1999; Roan & Raley,
1996; Wijckmans & van Bavel, 2013). Relationships that are high in conflict and closeness are described as ambivalent as there are contradictions in the way parents view ties with their children (Lüscher & Pillemer, 1998). Ambivalence is associated with greater stress and strain for parents, which in turn, relates to poorer well-being (Fingerman, Pitzer, Lefkowitz, Birditt, &
Mroczek, 2008; Lendon, Silverstein, & Giarusso, 2014; Willson et al., 2006). Thus, understanding how marital biographies relate to parent’s ambivalence toward children is fundamental for deepening our knowledge of the well-being of older adults.
Research examining the relationship between marital biography and parent-child relationships has shown divorce is associated with increasingly negative parent-child interaction, thus, greater ambivalence (Albertini & Garriga, 2011; Connidis, 2010; Hammersmith, in press;
Kaufman & Uhlenberg, 1998; Pezzin & Schone, 1999). Whereas divorce is associated with negative feelings regarding parent-child bonds, widowhood decreases parent’s ambivalence (Ha
& Ingersoll-Dayton, 2008). Moreover, other studies have found repartnering following divorce or widowhood weakens parent-child ties (de Jong Gierveld & Merz, 2013; de Jong Gierveld &
Peeters, 2003; Kalmijn, 2013; Pezzin & Schone, 1999). Still, no studies that have focused on ambivalence to date have considered whether the individual has experienced multiple disruptions as well as a subsequent remarriage. This chapter advances prior work through examining how remarriage—relative to being unmarried— relates to ambivalence after different disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions). Few studies account for 82
duration remarried or unmarried after widowhood, divorce, or multiple disruptions. Those that
do have found the negative effect of widowhood on ambivalence wears off over time, whereas
the positive effect of divorce on ambivalence deteriorates with years after divorce (Connidis,
2010; Guiaux et al., 2007; Ha & Ingersoll-Dayton, 2008). Still, researchers have yet to examine
comparisons between duration in remarried or unmarried statuses after different disruption
pathways.
I employ data from the 1992-2012 Health and Retirement Study to address three primary
research questions. First, how is remarriage after disruption related to parent’s ambivalence
toward children and what is the role of different disruption pathways? Second, what is the
relationship between duration remarried or unmarried following different disruption pathways
and ambivalence toward children? Finally, does remarriage after different disruption pathways
and duration remarried or unmarried following disruption relate to more or less ambivalence for men compared to women?
This chapter adds to current literature in several ways. First, this chapter accounts for not only current marital status, but also the role of disruption and duration when investigating parent’s ambivalence toward children. Most prior studies have solely examined the role of particular marital statuses such as widowhood on parent’s ambivalence toward children (Ha &
Ingersoll-Dayton, 2008). The current study extends beyond past research to consider how marital biography relates to parent’s ambivalence. Prior work has mostly focused on parent’s quality of ties with children or support received from children, but the extent to which marital biography
relates to parent’s ambivalence remains unclear. This is a noteworthy gap in the literature,
especially as researchers have identified that ambivalence is positively correlated with the health
and well-being of older people (Lendon et al., 2014; Willson et al., 2006). Finally, this chapter 83
deepens our understanding of how the association between marital biography and ambivalence
differs between men and women. Men’s and women’s relationships with children vary in later
life, making it essential to learn whether marital characteristics like current marital status,
remarriage following different disruption pathways, and duration remarried or unmarried are
differentially associated with men’s and women’s ambivalence toward children.
Intergenerational Ambivalence
Parent-child relationships are one of the most enduring social ties cultivated over the life
course (Birditt et al., 2009). Even though parents and children remain close as parents grow older and as children move through childhood to adulthood, these bonds are not static (Birditt et al.,
2009). In fact, intergenerational ties, like those between parents and children, are inherently structured to create emotional and attitudinal contradictions, often described as feelings of ambivalence (Lüscher & Pillemer, 1998). Whereas the intergenerational solidarity framework
emphasizes the positive, consensual nature of intergenerational ties, and conflict theory focuses
on stressors and problems that weaken relationships, the ambivalence framework accounts for
the positive and negative to bridge these two theoretical perspectives (Lüscher & Pillemer,
1998).
Intergenerational ambivalence is characterized by simultaneous conflict and closeness
between parents and children. Conflict and closeness are not merely opposite feelings. Rather,
high conflict and closeness can coexist when describing the same parent-child relationship. It is
crucial to acknowledge conflict and closeness simultaneously as intermingled positive and
negative feelings often require parents to constantly renegotiate relationships with children,
especially in the face of life course changes, like those pertaining to marital biography. When
parents describe ties with children as high in conflict and closeness, this yields a relationship 84
characterized by ambivalence (Lüscher & Pillemer, 1998). Ambivalence is a key component of
parent-child ties in later life as contradictory feelings toward people with whom one shares close
ties are potentially harmful for well-being (Lendon et al., 2014; Willson et al., 2006).
Changes in the parent’s life, like those related to marriage, are associated with shifts in
conflict or closeness between parents and children, thus, relating to greater ambivalence
(Connidis, 2010; Kaufman & Uhlenberg, 1998; Willson et al., 2006). Specifically, some marital
statuses encourage closeness and reduce conflict between parents and children, such as
widowhood or remaining in a first marriage, suggesting the parent-child tie is less ambivalent.
However, other marital statuses generate conflict between parents and children, such as divorce
or remarriage. In such circumstances when conflict is high, but parents still feel close to children,
the relationship is described as more ambivalent (Connidis, 2010; de Jong Gierveld & Merz,
2013; Schenk & Dykstra, 2012).
Marital Status and Ambivalence
Parent’s marriage is associated with less ambivalent parent-child ties (Pillemer, Munsch,
Fuller-Rowell, Riffin, & Suitor, 2012). This suggests parents in a first marriage have
relationships with children that are high in closeness and low in conflict, and thus, lower in
ambivalence than parents of other marital statuses, like divorceds and cohabitors. Divorce leads
to deterioration of parent-child relationships through creating conflict (Connidis, 2010; Daatland,
2007). Thus, although parents may still feel positively toward children, heightened negative
parent-child interaction within the tie will lead to greater ambivalence (Daatland, 2007;
Furstenberg, Hoffman, & Shrestha, 1995; Hammersmith, in press; Kaufman & Uhlenberg, 1998;
Pezzin & Schone, 1999; Umberson, 1992). Cohabiting parents report lower quality relationships
with children (Wright, 2017). Children may not approve of parent’s cohabitation, which sparks 85 conflict thereby heightening ambivalence. Widowhood often draws parents and children together, diminishing conflict while also increasing closeness (Ha, 2008), which lowers overall ambivalence compared to divorceds and cohabitors. Prior work indicates that remarriage relates to lower quality parent-child bonds, as entering another marriage disrupts the parent-child tie through necessitating renegotiation of relationships (de Jong Gierveld & Merz, 2013; de Jong
Gierveld & Peeters, 2003; Schenk & Dykstra, 2012). Thus, following a remarriage, conflict between parents and children is likely higher, while parents continue to report feeling close to children (de Jong Gierveld & Merz, 2013). The never married tend to be more disadvantaged in support receipt from informal social network members than other marital groups (Barrett &
Lynch, 1999; Dykstra & de Jong Gierveld, 2006; Keith, Kim, & Schafer, 2000). Thus, conflict may be low between never marrieds and their children, but closeness is likely also low, indicating less ambivalence than other groups like divorceds and cohabitors.
Remarriage, Disruption Pathways, and Ambivalence
Assessing current marital status provides valuable insight about parent’s relationships with children. However, accounting for parent’s subsequent remarriage and disruption pathway will provide a better understanding of how parent-child ties are associated with complex marital transitions in later life. Specifically, more married older people enter later life in a remarriage than in the past (Lin et al., in press), which shows greater marital instability in early to mid-adult life. Moreover, although widowhood remains a prevalent disruption pathway among older people
(Manning & Brown, 2011), divorce in later life is an increasingly common exit from marriage
(Brown & Lin, 2012). Some studies have investigated how aspects of marital biography like remarriage, widowhood, or divorce relate to parent-child relationships. Still, few have directly 86
examined how remarriage after different disruption pathways is associated with parent’s ambivalence.
Some studies have investigated the relationship between divorce and parent-child relationships, although not specifically regarding parent’s ambivalence toward children.
Kaufman and Uhlenberg (1998) found that parent’s divorce deteriorates parent-child relationship quality. Similarly, Hammersmith (in press) discovered that older parent’s divorce was related to parent’s greater negativity toward children. However, these studies are limited in several ways.
They did not consider how divorce compares to other disruption pathways. They also did not look at the role of remarriage after disruption relative to being unmarried. Finally, these studies show divorce is harmful for parent-child relationships, but the extent to which these findings apply to parent’s ambivalence remains unknown.
Another study extended work that solely looked at the association of divorce and parent- child ties by incorporating not only the role of marital disruption through divorce, but also subsequent repartnering when considering aspects of the parent-child relationship. Pezzin and
Schone (1999) found divorce was negatively related to support transfers between parents and children and repartnering reduced exchanges even further. Still, this study did not account for other disruption pathways (i.e., one widowhood or multiple disruptions). Moreover, this study operationalized parent-child relationships through contact and support, but it did not examine parent’s ambivalence toward children.
A few other studies have extended this line of work by examining repartnering among not only divorced older adults, but also among widowed older people. De Jong Gierveld and Merz
(2013) conducted analyses using qualitative data and revealed that among divorced and widowed parents, repartnering often made children feel threatened and spurred the renegotiation of parent- 87 child ties. Although their findings depicted the stress brought on by repartnering following divorce or widowhood, this study did not explicitly investigate the role of multiple disruptions and did not consider parent’s ambivalence toward children. Similarly, two quantitative studies also examined repartnership among the formerly divorced and widowed. Schenk and Dykstra
(2012) and de Jong Gierveld and Peeters (2003) discovered that repartnering was seemingly associated with more conflict and less positivity between parents and children than divorce and widowhood. However, these studies are limited as they did not factor in remarriage after different disruption pathways. Moreover, these studies used outcomes related to relationship quality and support, thus, it remains uncertain if the conclusions would be similar when examining parent’s ambivalence toward children.
There are only a handful of studies that have looked beyond parent-child relationship quality or contact to focus specifically on ambivalence. Those studies that do examine the relationship between ambivalence and marital characteristics have solely focused on transitions out or marriage like divorce or widowhood. For instance, using qualitative data, Connidis (2010) studied how parent’s divorce links to parent’s ambivalence and concluded that divorce was positively associated with parent’s ambivalence. Finally, Ha and Ingersoll-Dayton (2008) incorporated ambivalence into their work to examine reports made by widowed older adults.
They found that relative to the married, the widowed reported less ambivalence toward children.
Although these studies extend prior work by considering the contradictory nature of parent-child ties through incorporating ambivalence, they are limited in several ways. First, each of these studies only accounted for one disruption pathway—either through one divorce (Connidis, 2010) or one widowhood (Ha & Ingersoll-Dayton, 2008)—rather than investigating how these 88 disruption pathways may compare to one another. Further, neither study accounted for multiple disruptions or the role of a remarriage following different disruption pathways.
I propose several predictions to address the gaps in current research. First, I predict remarriage will be positively associated with ambivalence relative to remaining unmarried if the disruption pathway is the same. Remarriage relates to deterioration of parent-child bonds through creating ambiguous relationship norms and generating conflict, and thus, remarriage will likely be positively associated with ambivalence (de Jong Gierveld & Merz, 2013; Pezzin & Schone,
2008; Schenk & Dykstra, 2012). I also expect disruption through widowhood will relate to less ambivalence toward children than one divorce or multiple disruptions as widowed parents tend to report more closeness and less conflict with offspring (Ha & Ingersoll-Dayton, 2008). Along with this, even though there is limited research on the well-being of older adults after multiple disruptions I predict that older adults with multiple disruptions will feel greater ambivalence toward children relative to parents with one divorce or widowhood. However, I predict parents after one divorce will report less ambivalence than parents after multiple disruptions as multiple disruptions may be associated with more conflict beyond one divorce.
Duration Remarried, Duration Unmarried, and Ambivalence
Incorporating remarriage following different disruption pathways builds upon prior work on parent’s marital transitions and ambivalence toward children. Nevertheless, the role of duration remarried or unmarried following different disruption pathways remains largely unknown. Parent’s relationships with children are constantly evolving, especially in the wake of changes in the parent’s life (Hammersmith, in press; Kaufman & Uhlenberg, 1998). It is likely the parent-child bond continues to evolve as parents spend years either divorced, widowed, or remarried (Connidis, 2010; Ha & Ingersoll-Dayton, 2008; Kaufman & Uhlenberg, 1998). 89
Despite the changing nature of parent-child ties over time and the salience of parent’s marital
status when considering parent-child relationships, few studies examine duration when assessing
the parent-child tie. Studies that have incorporated how duration links to parent-child ties have
mostly focused on bereavement. Thus, it is unclear the extent to which duration after different
disruption pathways plays a role in parent’s ambivalence toward children.
Prior work has investigated the role of time in a particular marital status on parent-child
relationships. For instance, Guiaux et al. (2007) looked at support that parents received from
children following widowhood. They found support receipt was high immediately following the
spouse’s death, although this tapered off about two and a half years after bereavement. Although
it makes important contributions to current research, this study did not consider disruption
pathways or remarriage. These researchers also did not account for parent’s ambivalence toward
children.
Building upon this work, there have been several studies that have looked past parent-
child relationship quality and focused specifically on the relationship between marital status
duration and ambivalence toward children. Ha and Ingersoll-Dayton (2008) showed that widowed parents reported less ambivalence relative to the married, and similar to Guiaux et al.
(2007), they found the negative association of relationship quality after widowhood tapered off after about two years bereaved. Connidis (2010) conducted a qualitative study of ambivalence after an older parent’s divorce. Although divorce generated more ambivalence between parents and children relative to the married, length of time divorced allowed parents and children to renegotiate their relationships, which yielded less ambivalence. Even though these studies shed light on the relationship between widowhood or divorce and ambivalence, it remains unclear how different disruption pathways as well as remarriage may relate to parent’s ambivalence. 90
Based on the limited prior research on duration remarried or unmarried and parent’s
ambivalence toward children, I make the following predictions. First, I hypothesize given the
same disruption pathway, duration remarried will be associated with lower ambivalence relative
to the unmarried. Although remarriage disrupts relationship norms and can create relationship
strain, parents and children likely adjust to the parent’s remarriage over the years, and thus, ambivalence will decline among the remarried over time. Second, I expect regardless of whether older adults remarry or remain unmarried, duration following one divorce or multiple disruptions will be more negatively associated with ambivalence relative to parents after one widowhood.
Prior work indicates the widowed experience less ambivalence toward children, but this association declines over time (Ha & Ingersoll-Dayton, 2008). However, for the divorced or those after multiple disruptions, older adults and children likely have time to adjust to the parent’s change in marital status as time accumulates since the disruption. Thus, older adults who divorced or have multiple disruptions may feel less ambivalent toward children compared to the widowed as duration since disruption increases.
Variation by Gender
Men’s and women’s marital biographies are important for relationship quality with children— specifically ambivalence (Amato, Rezac, & Booth, 1995; Aquilino, 1994; Booth &
Amato, 1994; Daatland, 2007; Kalmijn, 2015; Pillemer, Munsch, Fuller-Rowell, Riffin, &
Suitor, 2012; Ward et al., 2009). Women tend to share closer, more intimate bonds with children over the life course and are often described as kin-keepers (Rosenthal, 1985). Women report more positivity toward children than men whereas their negativity is equivalent to that of male counterparts (Ward et al., 2009). Moreover, married parents report less ambivalence than those 91
in other marital statuses, although this association is stronger for men than for women (Pillemer et al., 2012).
When considering parent’s remarriage following marital disruption, it is likely remarriage is more consequential for men’s ambivalence than women’s. Specifically, Kalmijn (2015) found repartnering after divorce was more detrimental for men’s relationships with children than for women’s. Disruption pathway also appears to be more salient for men’s relationships with children relative to women’s. Daatland (2007) and Kalmijn (2007) found divorce had a more negative relationship with men’s ties with children than for women. Moreover, Aquilino (1994) discovered widowhood negatively related to men’s relationship quality with children, but this was not true for women. Although these studies did not explicitly look at ambivalence, based on this evidence, I predict that men will experience more ambivalence toward children than women when they remarry after disruption. Further, I hypothesize men will report more ambivalence than women regardless of disruption pathway.
Prior work has not yet considered the role of duration remarried or unmarried after different disruption pathways on men’s or women’s ambivalence toward children. However, based on prior work suggesting marital characteristics are more influential for men’s relationships with children than women’s, I first hypothesize that disruption through one widowhood, one divorce, or multiple disruptions will be associated greater ambivalence felt by men relative to their female counterparts. Relatedly, remarriage will be more consequential for men’s greater ambivalence than for women’s, no matter the disruption pathway. Second, duration remarried or unmarried will relate to more ambivalence felt by men relative to women, regardless of the disruption pathway.
92
The Present Study
Parent-child relationships are a source of mutual support for parents and children
throughout the life course (Bengston & Oyama, 2010; Silverstein & Bengston, 1997; Silverstein
& Giarusso, 2010). Yet, these bonds are not static (Bengston & Oyama, 2010; Birditt et al.,
2009). For instance, changes in the parent’s life, like those which relate to their marital
biography, alter the parent-child relationship (Connidis, 2010; Hammersmith, in press; Kaufman
& Uhlenberg, 1998). Specifically, marital transitions like widowhood, divorce, and remarriage
may be associated with fluctuations in parent’s ambivalence toward children. Studying parent’s
ambivalence is important, as greater ambivalence is associated with poorer parental well-being
(Fingerman et al., 2008; Lendon et al., 2014; Wilson et al., 2006); further, ambivalence also
relates to support exchanged between parents and children such that more ambivalence is
associated with fewer parent-child transfers (Connidis, 2010). Although some researchers have demonstrated the relationship between parent’s marital biography and ambivalence toward children (Connidis, 2010; Ha & Ingersoll-Dayton, 2008), research remains limited in two ways.
First, even though studies have examined ambivalence after divorce or widowhood, researchers have not considered how remarriage compared to remaining unmarried following different disruption pathways (i.e., one divorce, one widowhood, or multiple disruptions) may be uniquely associated with ambivalence toward children. Further, scholars have not accounted for how parent’s ambivalence toward children may vary with duration remarried or unmarried after different disruption pathways.
This chapter puts forth three research questions to fill these research gaps. First, how is remarriage following disruption linked to parent’s ambivalence toward children and how does this vary by different disruption pathways? Second, what is the relationship between duration 93
remarried or unmarried following different disruption pathways and parent’s ambivalence toward
children? Finally, does remarriage after different disruption pathways and duration remarried or
unmarried following disruption relate to more or less ambivalence for men versus women? I
propose the following hypotheses.
Hypothesis 1a: When the disruption pathway is the same, remarriage will be positively
related to ambivalence relative to staying unmarried.
Hypothesis 1b: Regardless of whether older adults remarry or stay unmarried, parents
after one widowhood will be least ambivalent, followed by parents after one divorce, and then
multiple disruptions.
Hypothesis 2a: Duration remarried will relate to less ambivalence with each additional
year remarried relative to duration unmarried, given older adults experience the same disruption
pathway.
Hypothesis 2b: Regardless of whether duration following disruption is spent remarried or unmarried, years following one divorce or multiple disruptions will be related to a smaller
increase in ambivalence with each additional year in status relative to parents after one
widowhood.
Hypothesis 3a: Disruption through one widowhood, one divorce, or multiple disruptions
will link to greater ambivalence felt by men relative to women. Similarly, remarriage will be
more consequential for men’s greater ambivalence than for women’s, regardless of disruption
pathway.
Hypothesis 3b: Duration remarried or unmarried will be associated with greater
ambivalence felt by men relative to women, regardless of the disruption pathway. 94
In this chapter, I also accounted for several covariates that may confound the association between marital biography and parent’s ambivalence toward children, such as demographic characteristics, socioeconomic resources, health, and children’s characteristics.
Demographic Characteristics
As parents age, they report more satisfaction with social relationships, suggesting older
parents feel less ambivalence toward children than younger parents (Carstensen, 1992;
Umberson, 1992). Older age coincides with time to accumulate different marital transitions, or
contrastingly, to accrue time in the current marital status (Cherlin, 2009; Goldstein & Kenney,
2001). Non-Whites engage in more support exchanges than Whites, suggesting more closeness,
and thus, less ambivalence in Non-Whites compared to White families (Coleman & Ganong,
2008). Whites and Hispanics also have similar rates of marriage and comparable marital stability
(Amato, 2010; Bramlett & Mosher, 2002), whereas Blacks are less likely to marry and more
likely to report marital instability or disruption (Heaton, 2002; Manlove et al., 2012).
Socioeconomic Resources
Parents with higher education have more resources, which enhance support exchanges,
and influence parent-child closeness (Birditt et al., 2009). The highly educated report greater
marital stability whereas the less educated have a higher incidence of marital disruption (Amato,
2010; Heaton, 2002). Wealth may also lower ambivalence by enabling parents to provide for
themselves as well as children, thereby reducing conflict. In addition, the wealthy experience
more marital stability than their poorer counterparts (Heaton, 2002). Employment allows parents
access to more resources. Thus, employed parents have more resources for themselves and to
share with children, supporting parent-child tie, and leading to less ambivalence than parents
who are not currently working (Thomas, 2011). Retirement is often related to marital strain, and 95
thus, could be positively associated with the likelihood of marital disruption (Dew & Yorgasen,
2010).
Health
Poor mental and physical health promote closeness between parents and children because
children reach out to parents who are in poor health more often (Ha et al., 2017). Nevertheless,
poor health can lead to parent’s perceptions of being burdensome to children, likely increasingly
conflict, and thus, ambivalence (Kiecolt, Blieszner, & Salva, 2011; Thomas, 2011). The
relationship between health and marital biography appears to be reciprocal. Marital instability is associated with poor health (Brockmann & Klein, 2004; Dupre & Meadows, 2007; Dupre et al.,
2009; Hughes & Waite, 2009; McFarland et al., 2013), and poor health can destabilize marriages, making them more vulnerable to disruption (Karraker & Latham, 2015).
Children’s Characteristics
Having older children links to less ambivalence, as relationships often improve as parents and children age (Cartensen, 1992). Parents who have more contact with children report feeling closer to children, although frequent contact can also increase instances of conflict (Ward,
Deane, & Spitze, 2014). Thus, it is unclear whether contact influences more or less ambivalence toward children. Parent-child coresidence and living nearby one another (within 10 miles) promote parent-child closeness, but could also spur conflict, possibly increasing or decreasing ambivalence (Aquilino & Supple, 1991). Finally, number of children relates to parent’s positive perceptions of ties with children (Ward et al., 2009).
Method
I employed data from the Health and Retirement Study (HRS) spanning from 1992 to
2010/2012. The HRS is a nationally representative sample of older adults. To be eligible, 96
respondents must be noninstitutionalized at baseline and live within a household. Respondents
are interviewed separately and are reinterviewed every two years. Every six years, the HRS
incorporates a new cohort (called a refresher sample) of older adults aged 51 to 56 to replenish
the study’s sample and preserve representativeness. The HRS has relatively high response rates,
with baseline rates hovering around 70-82% and rising to close to 90% or higher for follow-up
interviews. The HRS includes oversamples of Blacks, Hispanics, and Floridians. I used the HRS
Psychosocial Survey, a leave-behind questionnaire that queries respondents about their lifestyle
and relationships. To reduce respondent burden, the Psychosocial Survey is administered to only
half the sample every four years. By combining the 2010 and 2012 waves, I was able to
investigate how parent’s marital biography was related to ambivalence toward children.
Marital biography measures came from the HRS from 1992 to 2010/2012, whereas
current marital status, ambivalence, and other covariates were from the 2010 or 2012 surveys. I
began sample selection by pooling a file that included all HRS respondent’s information from
1992 to 2010/2012. I limited the sample to respondents who returned the 2010 or 2012
Psychosocial Survey (n = 15,571). Next, I excluded respondents under 50 in 2010 or 2012, leaving 15,094 respondents. I removed nursing home residents in 2010 or 2012, yielding a sample of 14,572 older people. I kept only respondents with valid information on children in the
RAND Family File to preserve information on the respondent’s children, leaving 13,589
respondents. I retained only respondents who had at least one biological child (n = 12, 899). In
order to examine only respondents with at least one adult child, I removed respondents who only
reported children under 18 in either 2010 or 2012 leaving 11,983 respondents. Finally, I omitted
576 respondents who were missing on any of the measures used to construct ambivalence, 97 yielding a sample of 11,407 respondents. Of the respondents, 6,768 were women and 4,639 were men.
Dependent Variable
Ambivalence. Parent’s ambivalence was measured by self-evaluated positive and negative feelings toward all children as an aggregate. Positive feelings included three questions: whether respondent’s children understand them, whether respondents can rely on children, and how much respondents can open up to children about worries. Negative feelings included three questions: whether children criticize them, get on their nerves, and let them down. The response categories for all measures were: not at all (coded 1), a little (coded 2), some (coded 3), or a lot (coded 4).
For calculating ambivalence, I used the respondent’s average score for positive and negative items. Ambivalence from 2010 and 2012 was calculated based on the following formula called the Griffin measure, which has been widely used in studies on ambivalence (Thompson, Zanna,
& Griffin, 1995):
Ambivalence= (positive+negative)/ 2 - |positive-negative| + 1.5
Ambivalence scores ranged from 0 to 13.5, with a larger value indicating higher ambivalence and a smaller value indicating lower ambivalence.
Focal Independent Variables
Marital Status. Marital status included six categories: married, remarried (reference group), cohabiting, divorced, widowed, or never married in 2010/2012.
Disruption Pathways. Prior disruptions were used to construct pathways through which individuals arrived at their current marital status. This measure included the following categories: first married, remarried after one divorce (reference category), remarried after one widowhood, remarried after multiple disruptions, cohabiting, unmarried after one divorce, 98
unmarried after one widowhood, unmarried after multiple disruptions, and never married. For
respondents after multiple disruptions, 65% have two or more divorces only, 22% have one
divorce and one widowhood, 6% have two or more widowhoods only, and 7% report three
disruptions that included a constellation of divorces and widowhoods (e.g., two divorces and one
widowhood). Respondents who were remarried after multiple disruptions made up a little over
5% of the full sample whereas the unmarried after multiple disruptions represented about 9% of
the full sample.
Duration Remarried or Unmarried. Duration was a continuous measure that captured
years currently remarried or unmarried after the most recent marital disruption.
Covariates
Several factors were included in the analyses that may confound the association between
marital biography and ambivalence toward children. Age was measured using three categorical
variables coded into age 50 to 64, ages 65 to 84 (reference category), and ages 85 and older.
Race and ethnicity comprised a categorical variable indicating whether the respondent identified
as White (reference category), Black, Hispanic, or of other races. Education was coded into four categories: less than high school, high school (reference category), some college, and college or higher. The distribution of household assets was highly skewed, and thus, I employed a series of five categories to capture household wealth: in debt, $0 to $50,000 (reference category),
$50,001-$100,000, $100,001-$250,000, and $250,001 or more. Work status was measured by whether the respondent was currently employed. Depressive symptoms summed of eight items in which the respondent reported whether they experienced certain symptoms during the week prior to the interview: feeling depressed, everything was an effort, restless sleep, unhappiness, loneliness, not enjoying life, sadness, or unable to get going. Depressive symptoms ranged from 99
0 to 8 and the inter-item reliability equaled .80. Physical health was a continuous measure
summing the number of chronic conditions diagnosed by a physician, including psychiatric
illness, diabetes, heart disease, cancer, lung disease, hypertension, stroke, and arthritis. Chronic
conditions ranged from 0 to 8.
To construct children’s characteristics, I aggregated all information provided about the
respondent’s children, since ambivalence was asked about all children rather than about each
individual child. Age was a continuous variable that captured the average age of all adult
children. Parent-child contact included the following categories: never (coded 1), less than once
a month (coded 2), more than once a month, but less than once a week (coded 3), once or twice a
week (coded 4), and three times a week or more (coded 5). The HRS did not query respondents
about frequency of contact with children who lived in their household, so parents were assumed
to have daily contact with resident children (coded 5). Parent-child proximity indicated whether
the respondent had at least one child who was resident, whereas children living within 10 miles
measured parent’s and children’s close proximity to one another. Last, I accounted for
respondent’s total number of children. All covariates were from 2010 or 2012, depending on the
respondent’s Psychosocial Survey wave.
Analytic Strategy
The proposed research questions are answered through two analyses. First, I presented
descriptive statistics for men and women for all variables included in the analyses. Second, I
employed ordinary least squares regressions to determine the association between marital biography and ambivalence. The first model accounted for how current marital status relates to ambivalence toward children. The second model included current marital status as well as the relationship between different disruption pathways and parent’s ambivalence. In the third model, 100
I included current marital status, different disruption pathways, as well as duration remarried or
unmarried following disruption. Interactions between duration and each marital status category
were included in the model although only relevant comparisons were shown in Table 4.4
(complete results available upon request). I specified the models separately by gender and tested
for differences between men and women. A pooled model including gender interactions with
each of the marital biography indicators were included to examine gender differences. Missing
data ranged from less than 1% for race/ethnicity to about 5% for proximity of children. Missing
values were imputed using the multiple imputation procedure by creating ten replicates (Acock,
2005). Finally, the analyses were weighted to correct for unequal probability of selection and
sample attrition (Ofstedal et al., 2011).
Descriptive Results
Table 4.1 includes the weighted descriptive results for all variables included in the
analyses for men and women. On average, men and women reported similar levels of
ambivalence toward children. The majority of men and women were in a first marriage. Over
half of men were married compared to nearly two-fifths of women. A little over one-quarter of
men were in a remarriage versus a little less than one-fifth of women. Few men or women were cohabiting. About 4% of men were cohabitors compared to 3% of women. Over 12% of men were divorced compared to 17% of women. More women than men were widowed—over one-
fifth of women were widowed compared to only 5% of men. Less than 1% of men or women in
the sample reported never married. I found a greater percentage of men were remarried after one
divorce than women (70% versus 61%) while more women were remarried after multiple
disruptions than men (29% versus 23%). There were no differences between men and women
who remarried following one widowhood—slightly less than 10% of women compared to 7% of 101
men. Among the unmarried, more men had one prior divorce compared to women (43% versus
27%). More women than men were unmarried following one widowhood—over two-fifths of
women compared to one-fifth of men. Finally, more men than women were unmarried after multiple disruptions (35% versus 32%). Men, on average, spent a longer time in their current
marital status than did women (29 years versus 27 years).
Women were older than men on average. A greater percentage of women were classified
as ages 85 or older relative to men (6% versus 4%). Compared to men, more women were Black
(10% versus 8%). In general, men were more educated than women. Over 55% of men had some
college or higher whereas only 49% of women had the same education. Women had fewer
average assets than men on average. One-fifth of women had wealth totaling between $1 and
$50,000 whereas 16% of men reported wealth within this range. About one-fifth of men had
wealth between $100,001 and $250,000 compared to about 19% of women. At the top wealth
category, a little over 46% of men fell into this range compared to 43% of women. More women
than men were employed. Two-fifths of men were currently working compared to one-third of
women. Women reported poorer physical and mental health than men. Women had an average of
2.04 chronic conditions relative to 1.95 for men and for depressive symptoms, women reported
1.60 symptoms in the week prior to the interview versus 1.29 symptoms among men. Contact
with children was slightly higher for women than men (4.13 versus 4.03, translating into contact
about once per week). The average age of children was slightly higher for women than for
men— 41 years for women compared to 37 years for men. Finally, a greater number of women had children living nearby than men (55% versus 51%).
102
Multivariate Results
Marital Status and Ambivalence
Table 4.2 investigates the relationship between current marital status and parent’s ambivalence toward children. Contradictory to evidence from prior work, men in a first marriage were more ambivalent than both divorced and widowed men although there were no differences between first married and remarried men. Relative to remarried men, divorced and widowed men all reported less ambivalence toward children. I also did not find any differences between remarried and cohabiting men. Moreover, no differences emerged between the different groups of unmarried men.
Among women, only one finding appeared as significant, and it was consistent with prior work. Women who were remarried reported slightly greater ambivalence toward children than widowed women. When comparing men and women, several significant associations emerged.
Cohabiting, divorced, and widowed men reported less ambivalence than their female counterparts who occupied the same marital status.
Remarriage, Disruption Pathways, and Ambivalence
Table 4.3 examines whether different disruption pathways and subsequent remarriage are associated with parent’s ambivalence toward children. I predicted that regardless of the disruption pathway, the remarried would be more ambivalent than the unmarried (Hypothesis
1a). Partially, in alignment with my hypothesis, I found remarried men after multiple disruptions were more ambivalent than their unmarried counterparts (0.08 versus 0.43) (difference denoted by superscript “a”). I also expected that regardless of whether older adults remarried or were unmarried, the widowed would be the least ambivalent followed by the divorced and then those 103 after multiple disruptions (Hypothesis 1b). However, there were no differences among remarried and unmarried men with different disruption pathways.
Among women, I found no significant differences between the remarried and unmarried
(Hypothesis 1a). My prediction regarding variation in ambivalence based on different disruption pathways was also not supported among women (Hypothesis 1b). Further, I hypothesized that remarriage and disruption pathways would be tied to greater ambivalence among men relative to women (Hypothesis 3a). However, this prediction was not supported and when I compared men and women’s ambivalence toward children. Only one contradictory result emerged (gender difference denoted by underlined coefficients). For men, being unmarried after divorce linked to significantly less ambivalence than for women who occupied the same status.
Duration Remarried, Duration Unmarried, and Ambivalence
I had several predictions regarding the role of duration remarried or unmarried following different disruption pathways, results from which are shown in Table 4.4. Specifically, I expected that duration remarried would yield less ambivalence relative to duration unmarried
(Hypothesis 2a) and for disruption pathways, older adults after one divorce or multiple disruptions would experience smaller increases in ambivalence with years in status relative to their counterparts after one widowhood (Hypothesis 2b). I also predicted that duration remarried or unmarried would be related to greater ambivalence felt by men relative to women (Hypothesis
3b). However, none of these predictions were supported.
The other covariates included in the model did not vary greatly between the multivariate models. In Table 4.4, for men, more frequent contact with children was associated with less ambivalence toward children. Number of chronic conditions and depressive symptoms, average age of children, resident children, as well as any children living within ten miles were all 104
positively associated with ambivalence. Among women, older age (ages 85 and older), less than a high school diploma, frequent contact with children, and average age of children were each associated with less ambivalence toward children. The number of depressive symptoms and chronic conditions as well as resident children related to greater ambivalence.
Discussion
Relationships between parents and children are one of the most important social ties situated within the life course (Bengston & Oyama, 2010). Despite the close ties most parents share with children throughout life, these relationships are not immutable. Conflict and closeness between parents and children often fluctuates as a result of the parent’s own life changes, like those related to marital transitions. Divorce, widowhood, and remarriage are associated with parent’s feelings toward children, like ambivalence, through instilling more closeness or creating conflict (Connidis, 2010; Ha & Ingersoll-Dayton, 2008; Lüscher & Pillemer, 1998). Considering how parent’s ambivalence toward children varies based on marital biography in later life is important as a number of prior researchers have linked parent’s greater ambivalence to poorer well-being (Fingerman et al., 2008; Lendon et al., 2014; Willson et al., 2006).
This chapter indicates that among men in particular, the first married were more ambivalent than their divorced and widowed counterparts. These findings are seemingly inconsistent with prior literature that shows widowed and first married parents tend to have better quality relationships with children, and thus, less ambivalence (Ha & Ingersoll-Dayton, 2008;
Pillemer et al., 2012). However, men tend to have smaller social networks later in life than their female counterparts as women tend to assume the kin-keeper role and maintain social networks within these relationships (Rosenthal, 1985). Thus, when men become divorced or widowed, they lose the partner that played the biggest role in managing social relationships. Since men 105
may have fewer people to rely on in their social network, it is possible men will reflect more positively overall on their ties with children. Moreover, women are often more invested in their relationships with children, and therefore, they may react more strongly to positive and negative aspects of the parent-child bond, thereby yielding greater ambivalence for women than that
which is reported by men (Connidis & McMullin, 2002). The findings for men also show that
remarried men are more ambivalent than their divorced and widowed counterparts. This is
consistent with prior work that suggests remarriage disrupts family norms, which could erode
parent-child bonds and thus, relate to greater ambivalence toward children (de Jong Gierveld &
Merz, 2013; de Jong Gierveld & Peeters, 2003; Schenk & Dykstra, 2012).
Among women, I found that the widowed were less ambivalent than remarried women,
which is in line with prior research showing that widowhood supports parent-child ties (Ha &
Ingersoll-Dayton, 2008). Moreover, I found compared to being remarried, cohabiting, divorced,
and widowed men were less ambivalent than their female counterparts. This is inconsistent with
research suggesting marital characteristics are more salient for men’s ties with children
compared women (Amato et al., 1995; Aquilino, 1994; Booth & Amato, 1994; Daatland, 2007;
Kalmijn, 2015; Pillemer et al., 2012; Ward et al., 2009). Again, it is possible that men who no
longer have a marital partner have more limited social networks than their female counterparts
(Rosenthal 1985; Umberson, 1992), suggesting men reflect more positively on their relationships
with children than women.
I expected that the remarried would be more ambivalent than unmarried older adults
regardless of disruption pathway (Hypothesis 1a). One interesting finding emerged that supports
this prediction. Men who remarried following multiple disruptions were more ambivalent
towards children than their unmarried counterparts. Greater ambivalence felt by the remarried 106
compared to unmarried men suggests that remarriage after multiple disruptions (most commonly
two or more divorces), further strain relationships with children. Moreover, there were no
differences among remarried or unmarried men with different disruption pathways (Hypothesis
1b).
The results from this chapter did not reveal any significant associations for women
regarding remarriage relative to being unmarried (Hypothesis 1a) or for different disruption
pathways (Hypothesis 1b). Alternatively, I did find unmarried men after one divorce reported
less ambivalence than their female counterparts, again suggesting being unmarried after one
divorce is more salient for reducing men’s ambivalence relative to women’s (Hypothesis 3a).
When incorporating duration remarried or unmarried following different disruption pathways, I did not find any association between years spent in status and ambivalence toward children
(Hypothesis 2a and Hypothesis 2b). This suggests parent’s ambivalence toward children remain relatively stable following a disruption and a possible subsequent remarriage.
This chapter has several limitations. First, I was only able to account for parent’s perceptions of relationships with children. Children’s reports of the relationship may not match their parent’s evaluations (Shapiro, 2004). Future work should consider both members of the parent-child dyad to better understand the role of concordance in reports of ambivalence.
Second, ambivalence in this chapter reflected what was felt by parents toward all children rather than individual children. Thus, I could not determine whether the respondent’s reports were driven by a specific child. If possible, future research should assess how parent’s ambivalence varies for each child. Finally, due to a small sample size of cohabitors, I was unable to measure disruption pathways leading up to a cohabitation. Future studies should assess whether cohabitation operates similar to remarriage when examining parent’s ambivalence. 107
Even so, this research makes a contribution to the current body of literature that examines parent-child relationships in later life. More specifically, this chapter shows that, for the most part, remarriage following different pathways of disruption was not strongly linked to parent’s ambivalence toward children. Nonetheless, it is crucial to deepen our knowledge of the role of remarriage following disruption, especially as older people experience increasing marital instability in later life (Brown, 2017; Brown et al., in press; Lin et al., in press). Although the findings from this chapter are few, those that arose from the analyses suggest that this relationship seemingly matters more for men’s ambivalence toward children than for women’s.
Particularly, remarried men who have reported multiple disruptions are at risk of experiencing
more ambivalence toward children than unmarried men after multiple disruptions. This finding
yields important implications. Greater ambivalence may further strain ties, jeopardizing a potential support avenue with one’s children. Moreover, ambivalence is associated with poorer well-being, which could make this group of men who have had multiple disruptions even more
vulnerable in later life relative to other groups of men as well as women (Lendon et al., 2014;
Wilson et al., 2006).
Duration remarried or unmarried after different disruption pathways was inconsequential
for parent’s ambivalence toward children. This finding can be taken in two ways. On the one
hand, this suggests that bonds between parents and children may not deteriorate with years in any
disrupted status, regardless of whether the parent has remarried. However, this finding also
implies relationships were not associated with any gains following marital disruption and a
subsequent remarriage, which could be harmful for parent-child relationships and the overall
well-being of parents. 108
In sum, this chapter sheds light on how marital biography relates to parent’s ambivalence
toward children in later life. Interestingly, different disruption pathways, remarriage, and
duration remarried or unmarried seemingly mattered less for parent’s ambivalence than in the
chapters focused on health and parent-child contact. Although one finding regarding remarried
men after multiple disruptions emerged, it is possible that marital biography is less important when studying parent’s ambivalence relative to other well-being outcomes. Nevertheless, understanding the relationship between remarriage, different disruption pathways as well as duration remarried or unmarried and ambivalence is a valuable endeavor, especially as children are one of the most common sources of support to aging parents other than the parent’s spouse
(Scommegna, 2016; Wolff & Kasper, 2006). If instability in marital biographies strains parent- child ties such that it negatively relates to parent’s feelings toward children, this may threaten potential support avenues when parents are in need. By gaining a better understanding of whether marital biography is associated with parent-child ambivalence, researchers can identify vulnerable groups of older people who likely need more assistance as they grow older. 109
Table 4.1 Weighted Means (Standard Deviations) and Percentages by Gender Men Women Ambivalence 2.35(2.7) 2.27(5.5) Marital biography First marriage 51.70 39.82 *** Remarriage 26.01 18.06 *** After one divorce 69.82 61.33 *** After one widowhood 7.00 9.94 After multiple disruptions 23.18 28.73 * Cohabiting 4.39 2.98 *** Divorced 12.25 17.26 *** Widowed 5.15 20.90 *** Unmarried after one divorce 43.39 27.25 *** Unmarried after one widowhood 22.07 41.12 *** Unmarried after multiple disruptions 34.54 31.63 *** Never married 0.50 0.98 Duration in current status 29.13(16.9) 27.13(33.9) Demographic characteristics Age 50 to 64 52.80 51.78 65 to 84 42.96 41.82 85 and older 4.24 6.40 *** Racial and ethnic background White 80.34 78.74 Black 8.07 10.17 *** Hispanic 8.37 8.15 Other race 3.22 2.94 Socioeconomic resources Education Less than high school 12.17 13.96 * High school 32.38 36.75 *** Some college 24.73 26.95 * College or more 30.72 22.34 *** Wealth In debt 6.50 7.39 $1 to $50,000 16.34 20.68 *** $50,001 to $100,000 10.62 10.75 $100,001 to $250,000 20.17 18.73 * $250,001+ 46.37 42.45 *** Currently employed 40.96 33.71 *** Health Chronic conditions 1.95(1.5) 2.04(2.9) ** Depressive symptoms 1.29(1.8) 1.60(3.8) ***
110
Table 4.1 continued Children's characteristics Contact with children 4.01(.95) 4.13(1.8) *** Average age of children 37.36(10.7) 40.85(22.6) *** Any stepchildren 22.27 20.81 Any resident children 28.35 28.41 Any children within 10 miles 51.31 55.15 *** Number of children 3.22 3.25 Weighted percentages 44.65 55.35 Unweighted N 4,639 6,768 * p < .05; ** p < .01; *** p < .001 111
Table 4.2 Coefficients (Standard Errors) from the Weighted OLS Regression Model of Marital Status on Ambivalence toward Children Men Women Marital biography First marriage - 0.04 (.15) ab - 0.23 (.14) Remarriage (ref) -- -- Cohabiting - 0.46 (.28) 0.21 (.25) Divorced - 0.38 (.15) *a - 0.10 (.19) Widowed - 0.45 (.22) *b - 0.29 (.13) * Never married 0.74 (1.0) 0.25 (.43) Demographic characteristics Age 50 to 64 0.06 (.14) 0.20 (.12) 65 to 84 (ref) -- -- 85 and older - 0.33 (.20) - 0.42 (.18) * Racial and ethnic background White (ref) -- -- Black 0.20 (.15) 0.23 (.12) Hispanic - 0.10 (.13) 0.00 (.15) Other race 0.55 (.28) - 0.48 (.29) Socioeconomic resources Education Less than high school - 0.07 (.18) - 0.26 (.11) * High school (ref) -- -- Some college 0.02 (.14) 0.06 (.11) College or more - 0.24 (.13) - 0.05 (.09) Wealth In debt 0.33 (.23) - 0.05 (.21) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.17 (.25) - 0.03 (.17) $100,001 to $250,000 0.31 (.19) - 0.01 (.15) $250,001+ 0.24 (.21) 0.04 (.14) Currently employed 0.13 (.14) 0.02 (.11) Health Depressive symptoms 0.11 (.03) *** 0.19 (.02) *** Chronic conditions 0.12 (.04) ** 0.17 (.03) *** Children's characteristics Contact with children - 0.22 (.06) *** - 0.12 (.05) * Average age of children - 0.03 (.01) * - 0.03 (.01) Any resident children 0.84 (.12) *** 0.68 (.12) *** Any children within 10 miles 0.18 (.08) * 0.10 (.09) Number of children - 0.02 (.03) - 0.02 (.03) Constant 3.60 (.49) *** 3.04 (.39) *** N 4,639 6,768 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing the same superscript letters denote significant differences in the same model at p < .05 Underlined coefficients denote significant differences between men and women at p < .05
112
Table 4.3 Coefficients (Standard Errors) from the Weighted OLS Regression Model of Marital Status and Disruption Pathway on Ambivalence toward Children Men Women Marital biography First marriage - 0.01 (.17) - 0.25 (.17) Remarriage After one divorce (ref) -- -- After one widowhood 0.05 (.31) - 0.11 (.32) After multiple disruptions 0.08 (.21) a - 0.06 (.22) Cohabiting - 0.43 (.30) 0.18 (.24) Unmarried After one divorce - 0.36 (.19) - 0.04 (.23) After one widowhood - 0.31 (.26) - 0.27 (.19) After multiple disruptions - 0.43 (.22) a - 0.30 (.16) Never married 0.76 (1.0) 0.23 (.45) Demographic characteristics Age 50 to 64 0.06 (.15) 0.21 (.12) 65 to 84 (ref) -- -- 85 and older 0.55 (.28) - 0.45 (.29) Racial and ethnic background White (ref) -- -- Black 0.20 (.15) 0.22 (.12) Hispanic - 0.10 (.13) - 0.00 (.15) Other race 0.55 (.28) - 0.49 (.29) Socioeconomic resources Education Less than high school - 0.07 (.18) - 0.26 (.10) High school (ref) -- -- Some college 0.02 (.14) 0.06 (.11) College or more - 0.24 (.13) - 0.05 (.09) Wealth In debt 0.34 (.23) - 0.04 (.21) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.17 (.24) - 0.03 (.17) $100,001 to $250,000 0.31 (.19) - 0.01 (.15) $250,001+ 0.23 (.21) 0.04 (.13) Currently employed 0.13 (.14) 0.02 (.10) Health Depressive symptoms 0.11 (.03) *** 0.19 (.02) *** Chronic conditions 0.12 (.04) ** 0.17 (.03) ***
113
Table 4.3 continued Children's characteristics Contact with children - 0.22 (.06) *** - 0.12 (.05) *** Average age of children - 0.03 (.03) - 0.03 (.03) Any resident children 0.84 (.12) *** 0.68 (.12) *** Any children within 10 miles 0.18 (.09) * 0.10 (.09) Number of children - 0.02 (.03) - 0.01 (.03) Constant 3.62 (.50) *** 3.07 (.41) *** N 4,639 6,768 * p < .05; ** p < .01; *** p < .001 Note: Coefficients sharing the same superscript letters denote significant differences in the same model at p < .05 Underlined coefficients denote significant differences between men and women at p < .05
114
Table 4.4 Coefficients (Standard Errors) from the Weighted OLS Regression Model of Marital Status, Disruption Pathway, and Duration on Ambivalence toward Children Men Women Marital biography First marriage 0.72 (.43) 0.10 (.41) Remarriage After one divorce (ref) -- -- After one widowhood 0.07 (.54) 0.02 (.71) After multiple disruptions 0.15 (.47) - 0.52 (.35) Cohabiting - 0.57 (.50) - 0.19 (.45) Unmarried After one divorce - 0.19 (.35) - 0.11 (.40) After one widowhood - 0.41 (.43) - 0.56 (.32) After multiple disruptions - 0.52 (.41) - 0.56 (.32) Never married 0.80 (1.0) 0.16 (.52) Duration 0.00 (.01) 0.01 (.01) Remarried after one divorce (ref) -- -- Remarried after one widowhood - 0.01 (.02) - 0.01 (.02) Remarried after multiple disruptions - 0.00 (.02) 0.03 (.04) Unmarried after one divorce - 0.01 (.02) 0.00 (.02) Unmarried after one widowhood - 0.01 (.02) 0.01 (.01) Unmarried after multiple disruptions 0.00 (.02) 0.01 (.02) Demographic characteristics Age 50 to 64 0.01 (.15) 0.19 (.12) 65 to 84 (ref) -- -- 85 and older - 0.31 (.20) - 0.47 (.18) *** Racial and ethnic background White (ref) -- -- Black 0.21 (.15) 0.20 (.12) Hispanic - 0.11 (.13) 0.06 (.11) Other race 0.52 (.28) - 0.49 (.29) Socioeconomic resources Education Less than high school - 0.07 (.19) - 0.27 (.10) * High school (ref) -- -- Some college 0.02 (.14) 0.06 (.11) College or more - 0.25 (.13) 0.06 (.10) Wealth In debt 0.34 (.19) - 0.04 (.21) $1 to $50,000 (ref) -- -- $50,001 to $100,000 0.17 (.25) - 0.03 (.17) $100,001 to $250,000 0.30 (.19) - 0.01 (.15) $250,001+ 0.22 (.21) 0.04 (.13) Currently employed 0.10 (.14) 0.01 (.10) Health Depressive symptoms 0.11 (.14) *** 0.19 (.02) *** Chronic conditions 0.13 (.04) ** 0.17 (.03) ***
115
Table 4.4 continued Children's characteristics Contact with children - 0.22 (.21) *** - 0.13 (.05) * Average age of children 0.03 (.01) ** - 0.02 (.01) ** Any resident children 0.81 (.12) *** 0.67 (.12) *** Any children within 10 miles 0.19 (.09) * 0.10 (.09) Number of children - 0.02 (.03) - 0.01 (.03) Constant 3.45 (.59) *** 3.15 (.44) *** N 4,639 6,768 * p < .05; ** p < .01; *** p < .001
116
CHAPTER IV: DISCUSSION
Marital biographies of older people have evolved over several decades such that a greater share of individuals enter mid or late-life in a remarriage than in the past (Lin et al., in press).
Marital disruption is also increasingly common in later life with growth in the share of divorces
involving someone who is 50 or older coupled with the continued prevalence of widowhood as
an exit from marriage (Brown & Lin, 2012; Brown et al., in press; Manning & Brown, 2011).
Some older people who experience marital disruption through either divorce or widowhood go
on to remarry (Brown et al., in press). As remarriages are more likely to end in disruption than
first marriages (Brown & Lin, 2012; Sweeney, 2002), it is possible that a greater number of older
people will experience multiple disruptions in the coming decades. Further, higher order
marriages tend to be shorter in length than first marriages (Wu & Brown, 2016). In light of
greater instability in marital biographies of older people, this dissertation focused on the role of
remarriage, disruption pathways, and duration remarried or unmarried as they relate to mental
and physical health, parent-child contact, and parent’s ambivalence toward children.
Prior studies have considered how aspects of marital biography are related to well-being.
A common outcome used in studies of marital biography is health (Barrett, 2000; Dupre et al.,
2009; Dupre & Meadows, 2007; Dupre et al., 2009; Hughes & Waite, 2009; McFarland et al.,
2013; Reczek et al., 2016; Zhang & Hayward, 2006). As marital transitions are accumulated over the life course, certain aspects of marital biography have the potential to benefit or harm health
(Elder, 1994; O’Rand, 1996; Williams & Umberson, 2004; Umberson, 1992; Zhang et al., 2016).
Even with prior work demonstrating the relationship between marital biography and mental or physical health outcomes, this line of research is limited in several ways. Specifically, studies of marital biography have yet to examine whether remarriage relative to remaining unmarried is 117
associated with fewer consequences of a marital disruption (Barrett, 2000; Marks & Lambert,
1998; Umberson, 1992; Williams & Umberson, 2004; Zhang et al., 2016). Moreover, researchers
have not examined whether there are mental or physical health differences between older people
who have different disruption pathways (i.e., one divorce, one widowhood, or multiple
disruptions). Further, even though scholars have considered duration in different marital statuses,
few have accounted for duration remarried or unmarried following different disruption pathways
(Lillard & Waite, 1995; Zhang & Hayward, 2006). Studying duration remarried allows
researchers to understand whether additional years remarried are associated with fewer costs
following a marital disruption. Similarly, investigating duration unmarried can reveal whether
years unmarried following disruption are associated with better or worse health.
Parent-child contact is a fundamental pillar of the intergenerational solidarity framework.
Even though parents often have frequent contact with children throughout their lives, researchers
have shown marital transitions like divorce, widowhood, and remarriage are associated with
fluctuations in contact (Daatland, 2007; Kalmijn, 2007, 2013, 2015; Noël-Miller, 2013; Pezzin &
Schone, 1999; Roan & Raley, 1996). Still, few researchers have factored in the entire marital
biography when investigating parent-child contact. More specifically, research does not pay attention to how different disruption pathways and a subsequent remarriage could take cumulative toll on contact with children. Prior work also does not investigate duration remarried or unmarried after different disruption pathways. This is a notable omission, especially as parent-
child contact often changes over time. Thus, years remarried or unmarried may be associated
with rebuilding or erosion of parent-child ties, and consequently, more or less parent-child
contact. 118
Although parent-child bonds can be characterized by intergenerational solidarity, parent- child ties also tend to involve conflict, and thus, both positive and negative feelings could coexist within the same relationship (Lendon et al., 2014; Lüscher & Pillemer, 1998). Moreover, ambivalence toward children often varies over time and can depend on the parent’s marital biography (Connidis, 2010; Ward et al., 2013). Only a handful of studies have considered how parent’s marital transitions such as widowhood, divorce, and remarriage are associated with parent-child relationship quality. Extending prior work to account for the relationship between marital biography and ambivalence is crucial as recent scholars have noted that greater ambivalence is negatively associated with overall health and well-being in later life (Lendon et al., 2014; Willson et al., 2006). In the few studies that examine ambivalence, researchers only do so by examining specific types of marital disruption (i.e., one divorce or one widowhood) rather than comparing disruption pathways to one another. Thus, it is unclear how different disruption pathways and subsequent remarriage are associated with ambivalence toward children. Last, similar to parent-child contact, accounting for time may be an important layer when studying ambivalence, as parent’s feelings toward their children often change over the life course
(Bengston & Oyama, 2010; Birditt et al., 2009). Spending years remarried or unmarried after disruption could allow time to rebuild relationships with children, thereby reducing ambivalence.
Alternatively, parent-child ties could become more conflicted as years pass, which could be positively associated with ambivalence.
Key Findings and Contributions
The first analytic chapter assessed how remarriage and different disruption pathways were associated with mental and physical health. Prior work on marital biography finds that remaining in a first marriage is associated with better health outcomes relative to other marital 119
statuses (Hughes & Waite, 2009). Still, remarriage may be associated with fewer negative consequences of marital disruptions although not to the same extent as a first marriage (Barrett,
2000; Marks & Lambert, 1998; Williams & Umberson, 2004). This dissertation extends prior work by investigating whether remarriage relative to remaining unmarried is beneficial for different pathways of disruption (one divorce, one widowhood, or multiple disruptions). Among men, unmarrieds with any type of disruption pathway reported more depressive symptoms than their remarried counterparts. Moreover, unmarried women after one divorce reported more chronic conditions than their remarried counterparts. Thus, although remarriage does not offer the same benefits as a first marriage (Barrett, 2000; Dupre & Meadows, 2007; Marks & Lambert,
1998; Williams & Umberson, 2004), it appears that relative to remaining unmarried, remarriage following disruption may provide financial, social, and emotional resources that protect mental and physical health.
I found that there were health variations associated with different disruption pathways such that remarried women after one divorce reported fewer depressive symptoms and chronic conditions than remarried women after multiple disruptions. Similarly, unmarried women after
one divorce or one widowhood also reported fewer depressive symptoms than their unmarried
counterparts after multiple disruptions. Therefore, the stress from marital disruption may be
additive in nature such that experiencing multiple disruptions is linked to greater emotional,
financial, and social strain, resulting in poorer mental health for older women. Moreover, the
results suggest that even though remarriage was beneficial for the health of some groups,
remarriage was still associated with negative consequences of multiple disruptions relative to one
divorce. 120
I also considered the role of duration remarried or unmarried following disruption. I did not find that duration remarried provided a health advantage over years unmarried, given the same disruption pathway. However, I discovered that among men, each additional year in a remarriage was associated with greater depressive symptoms for those after multiple disruptions relative to those after one divorce. These findings suggest that older men who had experienced multiple disruptions were especially disadvantaged. Whereas I expected to find a health benefit associated with duration remarried, remarried men after multiple disruptions faced additional negative health consequences with each additional year in a remarriage. This contradictory finding could be reminiscent of recent work that demonstrates the costs of disruption are three times as great as entrance into a marriage (Kalmijn, 2017). Moreover, taken together, these findings show that remarriage is associated with fewer costs following marital disruption, however, the benefits of remarriage were not equivalent across different disruption pathways and health indicators.
I found conflicting results regarding gender variation in this chapter. Consistent with my expectations, unmarried men after one divorce reported greater depressive symptoms than their female counterparts. This finding is consistent with work that indicates that remaining unmarried after a marital disruption is more detrimental for men than for women (Amato & Rodgers, 1997;
Hu & Goldstein, 1990). However, in conflict with my prediction, women who were unmarried after one divorce reported more chronic conditions than men who occupied the same status, showing that marital disruption was not always more detrimental for men relative to their female counterparts.
Finally, I found that the magnitude of the coefficients in all of the models for depressive symptoms was greater than the magnitude of the coefficients for chronic conditions. This finding 121
is inconsistent with my prediction and implies that the relationship between remarriage,
disruptions pathways, and duration remarried or unmarried is stronger for mental rather than
physical health, which mostly mirrors prior work (Hughes & Waite, 2009).
The second analytic chapter examined the relationship between parent-child contact and remarriage following different disruption pathways. Based on prior research, I expected to find that remarriage would have a negative association with parent-child contact (Kalmijn, 2007,
2015; Noël-Miller, 2013; Pezzin & Schone, 1999). For men, I found the opposite. Remarried men after one divorce had more contact with children than unmarried men after one divorce.
Although this finding is somewhat surprising, it is possible men gain a new female partner who acts as a kin-keeper, thus, promoting contact with children. In accordance with my prediction, remarried women after one widowhood or multiple disruptions reported less contact with children than unmarried women with the same disruption pathway. Women who enter a remarriage likely face challenges as they work to renegotiate their ties with children, which could have a negative effect on contact with children.
As for disruption pathway, among remarried men, the widowed had the most contact with children, followed by men after one divorce, and finally, remarried men after multiple disruptions had the least amount of contact with children. These findings are consistent with prior research showing that the widowed have more contact with their children than the divorced.
Additionally, these results contribution to our understanding of how disruption pathways relate to the frequency of contact with children. Men’s multiple disruptions may increasingly strain parent-child ties, rendering less contact between these men and their children relative to women.
Among unmarried men, the results were similar, however, there was one notable difference. Unmarried men after one widowhood still reported the most contact with children 122
relative to unmarried men after one divorce or multiple disruptions. However, unmarried men after multiple disruptions had more contact with children than unmarried men after one divorce.
There may be an explanation for greater contact of unmarried men after multiple disruptions
relative to men after one divorce. Unmarried men after one divorce were nearly five years
younger on average than unmarried men after multiple disruptions in my sample and younger
age is positively related to forming new partnerships (Brown et al., in press). Even though these
men were not cohabitors or remarried, they still could have been partnered in some way. More
specifically, relative to unmarried men after multiple disruptions, unmarried men after one
divorce may be more likely to be in a dating or LAT (living apart together) relationship.
Although increasingly common among older people (Brown & Shinohara, 2013; Duncan &
Phillips, 2011; Wu, 2018), dating and LAT partnerships are not captured by the questions
currently included in the HRS. Unmarried men in a dating or LAT relationship may face strained
ties with children, which could be associated with less contact (de Jong Gierveld & Merz, 2003).
Similar to men, remarried women after multiple disruptions also had less contact with
children than their remarried counterparts after one divorce or widowhood. Finally, unmarried
women after one widowhood reported greater contact with children relative to unmarried women
after one divorce or multiple disruptions. This finding supports the notion that experiencing
divorce or multiple disruptions is likely associated with strained parent-child ties (Daatland,
2007) relative to the widowed who typically maintain close relationships with their offspring
(Ha, 2008; Ha & Ingersoll-Dayton, 2008; Roan & Raley, 1996).
Duration was also important for parent-child contact. Remarried men who had multiple
disruptions reported greater contact with children with each additional year remarried than their
counterparts who were remarried after one widowhood. When considering disruption pathway 123
alone, remarried men after multiple disruptions reported less frequent contact with children
relative to their remarried counterparts after one widowhood. By accounting for duration, I found
that additional years remarried for men who had multiple disruptions were associated with
greater contact with children relative to their widowed counterparts. Thus, although multiple
disruptions may have reduced frequency of parent-child contact, it appeared that these men were
able to rebuild their relationships with children over time.
Moreover, although remarried women after one widowhood had less contact with
children relative to their unmarried counterparts when considering disruption pathway alone,
factoring in duration in a remarriage showed that remarried women after one widowhood had
greater contact with children with each additional year widowed relative to unmarried women
after one widowhood. Remarriage was disruptive to women’s contact with children following widowhood, however, incorporating duration showed that these relationships seemed to improve with years in status. Also in alignment with my expectations, I found unmarried women who had one divorce had greater contact with children than men with the same status. Consistent with prior research, women tend to maintain stronger ties with children than men. Therefore, even though divorce is stressful, women may be able to draw on their closer ties with children to maintain more contact than their male counterparts.
In the final analytic chapter, I investigated the relationship between parent’s ambivalence toward children and remarriage following different disruption pathways. Fewer findings emerged regarding the relationship between remarriage, disruption pathways, duration, and ambivalence toward children relative to health and parent-child contact. Nonetheless, supporting one of my predictions, I found that remarried men after multiple disruptions reported greater ambivalence toward children than their unmarried counterparts. This finding implies that remarriage further 124
strains parent-child ties following multiple disruptions for men. However, no significant
associations between remarriage or disruption pathways arose for women, suggesting that
ambivalence did not vary depending on whether remarried or unmarried women arrived at their
current marital status through one divorce, one widowhood, or multiple disruptions. The analyses
also did not yield a significant association of duration remarried or unmarried among men or
women with different disruption pathways, showing that ambivalence was neither positively or
negatively associated with years in status. As for differences between men and women, one
significant association emerged and showed that unmarried men after one divorce reported less
ambivalence than their unmarried female counterparts with the same marital status. Although I
expected to find that women would be less ambivalent than men, it is possible that these
unexpected results emerged for two reasons. First, men often have smaller social circles in later
life, which may mean that men place a higher value on their ties with children, and thus, they may see their children in a more positive light than women. Second, women are often more invested in their children’s lives than men, and therefore, they may react more strongly positive and negative characteristics of their relationships with children, making them more ambivalence than men (Connidis & McMullin, 2002).
Taken together, these findings demonstrate that researchers should pay attention to older people who have experienced multiple disruptions. In most chapters, relative to their divorced and widowed counterparts, older adults after multiple disruptions seemed to be disadvantaged in their health and relationships with children. More specifically, the findings from this dissertation mostly showed that having multiple disruptions was associated with poorer health and less contact with children. This dissertation also deepens our understanding of the role of remarriage
relative to remaining unmarried following different disruption pathways. For health, remarriage 125
often provided a health advantage relative to remaining unmarried. Yet, when looking at parent-
child relationships, it was unclear whether remarriage was beneficial or not. For men, remarriage
after one divorce was associated with more contact with children relative to the unmarried
whereas for women, remarriage was associated with less parent-child contact among the
widowed and those after multiple disruptions relative to the unmarried with the same disruption pathway. Moreover, men who remarried following multiple disruptions reported more ambivalence than their unmarried counterparts. These conflicting findings suggest that remarriage supports certain aspects of well-being in later life whereas it is detrimental for others.
The findings regarding duration yielded meaningful, but conflicting conclusions. In
particular, among remarried men, those after multiple disruptions had poorer health when
considering disruption pathway alone, and when accounting for duration, their health
increasingly deteriorated over time relative to remarried men after one divorce. Despite this
finding, for parent-child contact, I found remarried men after multiple disruptions had less
frequent contact than the widowed when looking at disruption pathway alone, yet, with years in
status, additional years remarried following multiple disruptions for men were associated with
greater contact with children relative to years spent remarried following widowhood. In sum,
these findings from this dissertation indicate that the benefits of remarriage as well as duration
remarried or unmarried could operate differently depending on the well-being outcome, although
more research is needed before any substantive conclusions can be drawn.
Taking all of the findings from each chapter together as a collective, this dissertation
contributes to the wealth of literature on marital biography in several ways. First, the findings
from this dissertation demonstrate that different aspects of marital biography may matter more
than others depending on the outcome that is being considered. In the second chapter of this 126 dissertation, which focused on health, both remarriage and disruption pathways were associated with older adults’ health, although there was variation between the findings for men and women.
Specifically, remarriage was associated with men’s health, whereas disruption pathway emerged as the aspect of marital biography that was related to women’s health. When focusing on parent- child contact in the third chapter, for the most part, both remarriage and disruption pathways were significantly associated with the contact that parents shared with their children. In sum, the findings from chapters two and three suggest that both remarriage and disruption pathway are integral components of marital biography for understanding older adults’ well-being and relationships with their children.
In the second chapter, there appeared to be fewer differences between the divorced and widowed, whereas older people with multiple disruptions often had poorer health than the divorced and widowed. However, unlike in the second chapter where the divorced and widowed looked quite similar when examining health, important differences between the divorced and widowed emerged in the investigation of parent-child contact. Specifically, the widowed had the most contact with their children whereas the divorced and older people with multiple disruptions had less contact with children. Marital biography is complex, and researchers are often looking for ways to simplify its measurement. However, the findings from chapters two and three show that combining different disruption pathways (i.e., the divorced and widowed) may mask important differences in older adult well-being, depending on which outcome being considered.
Finally, the conclusions from this dissertation also suggest that perhaps the use of marital biography as it is defined here cannot be suitably applied to all well-being outcomes. For instance, health and parent-child contact were associated with various aspects of marital biography, but there were few findings that emerged from the investigation of the relationship 127
between marital biography and parent’s ambivalence. Perhaps marital status alone is more
sensitive to parents’ ambivalence toward their children, or perhaps a marital biography framework other than that focusing on remarriage, disruption pathway, and duration should be considered in studies of ambivalence. Still, more research is needed before any definite
conclusions can be drawn.
Limitations
There are several limitations to this research that should be addressed. First, in each
analytic chapter, I am unable fully explore how cohabitation following different disruption
pathways is associated with mental and physical health, parent-child contact, and parent’s
ambivalence toward children. The difficulties posed by this limitation are twofold. First, due to a
small sample size of cohabitors, I am unable to distinguish whether cohabitors arrived at their
current marital status through one divorce, one widowhood, multiple disruptions, or never
marrying. Similar to remarriage following different disruption pathways, it is possible that
cohabitation after different disruption pathways are likely associated with more or less well-
being advantages relative to others. For example, cohabitation with no prior marital union could
be more advantageous to the health of these cohabitors relative to those after one divorce, one
widowhood, or multiple disruptions as never married cohabitors would not have experienced the
stress associated with marital disruption (Liu & Umberson, 2008). Second, the HRS does not
collect information on cohabitation histories. Thus, among remarried or unmarried older people,
their disruption pathways do not include the disruption of any prior cohabitations. Therefore, it is
unclear whether the dissolution of a cohabitation would have a similar detrimental effect as other
types of marital disruption on the overall well-being of older people. 128
This research is also limited by the questions used to assess relationship status and history in the
HRS. Specifically, recent estimates show the number of daters and living apart together (LAT)
relationships are on the rise among older people (Brown & Shinohara, 2013; Duncan & Phillips,
2011; Wu, 2018). Despite increases in daters as well as LAT relationships, the HRS does not ask
questions to allow identification of older people who may fall into one of these relationship
categories. Thus, among the unmarried or never married in this study, some could be dating or in
a LAT relationship. On the one hand, daters or older adults in a LAT relationship likely reap health advantages associated with having a partner relative to their unmarried counterparts. On the other hand, daters or LATs could have strained relationships with their children as dating or repartnership necessitate the renegotiation of parent-child ties (de Jong Gierveld & Merz, 2013), which could lower parent-child contact and spur greater ambivalence toward children.
Although there are two sides to every dyadic relationship, the HRS only collects data from older people, thus, ignoring the perspective of the respondent’s children. Prior work has indicated that parent’s and children’s reports often differ from one another. Parents often report higher relationship quality and reflect more positively on the parent-child relationship than children whereas children report they have more contact with parents and provide a greater amount of emotional assistance (Rossi & Rossi, 1990; Shapiro, 2004). Thus, analyses relying on the reports of parents, like those using the HRS, could paint a more positive picture of the parent- child relationship than if children’s reports were also included. Specifically, the association between parent-child contact and parent’s ambivalence toward children with remarriage following different disruption pathways relies exclusively on the parent’s reports of contact and ambivalence. 129
Finally, this dissertation, specifically the final analytic chapter, is also limited by the
aggregate design of the questions used to measure ambivalence toward children. Thus, when
parents report on positivity and negativity felt toward children, they are giving a general
assessment of their feelings toward all of their children. Still, parent’s feelings of ambivalence
could be driven by one child in particular. Alternatively, a parent could feel high positivity
toward one of their children and high negativity toward another child, which would render a
greater score of ambivalence even though these feelings are directed toward different children.
Moreover, Fingerman, Cheng, Birditt, and Zarit (2011) found that parent’s well-being is mostly
driven by reports of children’s problems, yet there is no effect of children’s successes on parental
well-being. Thus, even having one problem child could drive parent’s feelings of negativity
toward all of their children even though they may be close to other children.
Future Directions
Each analytic chapter makes a contribution to our understanding of the relationship
between remarriage, different disruption pathways, and duration remarried or unmarried with
various aspects of well-being. However, there are still many avenues for future work to extent
the research questions investigated in this dissertation. The results of this dissertation show in some ways, remarriage is associated with fewer negative consequences of marital disruption, whereas relative to being unmarried, remarriage can be detrimental for other indicators of well- being. Recent studies have increasingly found that when examining well-being in later life, cohabitation and remarriage generally mirror one another among populations of older people
(Brown & Kawamura, 2010; Brown et al., 2012; King & Scott, 2005; Wright, 2017). Future work should examine whether cohabitation operates similarly to remarriage as it relates to 130
whether a subsequent cohabitation is associated with greater or fewer consequences of one divorce, one widowhood, or multiple disruptions in later life.
Given ample evidence highlighting the differences in the health and parent-child relationships of older people in later life, this dissertation accounts for the role of gender differences in each analytic chapter (Amato & Rodgers, 1997; Daatland, 2007; de Graaf &
Fokkema, 2007; Hu & Goldstein, 1990; Kalmijn, 2007; Pillemer et al., 2012; Reczek et al., 2016;
Silverstein & Bengston, 1997; Shor et al., 2012). However, future work should consider other differences, such as racial or ethnic variation (Zhang et al., 2016). As indicated by prior research, marriage does not operate similarly across racial and ethnic groups. More specifically, marriage is less beneficial for Blacks than for Whites as the institution of marriage provides fewer social and economic resources for Black than for Whites (Broman, 1993; Bulanda & Brown, 2007).
Moreover, fewer Blacks are married than Whites, and when married, Blacks are more likely to end their marriages in divorce than Whites (Amato, 2010; Heaton, 2002). Only a handful of studies have delved into racial and ethnic variation in the association between marital biography and health, and specifically focus on marital disruption through either divorce or widowhood
(Barrett, 2003; Elwert & Christakis, 2006). For instance, one study showed divorce was less detrimental for the mental health of Blacks, but had more physical health consequences for
Blacks than Whites (Barrett, 2003). Elwert & Christakis found that widowhood was associated
with greater risk of mortality among Whites, but not among Blacks. Given racial and ethnic
differences in marriage and marital disruption between Whites and Blacks, future work should
consider whether the relationship between remarriage, different disruption pathways, and
duration remarried or unmarried differs by race/ethnicity when investigating health, parent-child
contact, and parent’s ambivalence toward children. 131
Future work could also more holistically assess the role of age on marital biography and
well-being in later life. Age and duration are often collinear with one another, and therefore, I
employed age dummies in this dissertation research. Still, it is likely that there is more to be
known about the relationship between age, marital biography and well-being. For instance, on
the one hand, older adults who fall into the middle-age or young adult category have
accumulated fewer years during which they may experience disruption or remarriage relative to
their older counterparts. On the other hand, older adults who are middle-aged or young old may
have more complex marital biographies relative to their older counterparts as these younger
groups of older people have been through a period of significant change in marriage and families
in the past decades (Cherlin, 2010; Lin & Brown, 2012). Therefore, it would likely be of value
for researchers to further disentangle the role of age as well as cohort when studying marital
biography in later life.
Other research avenues could involve making methodological advancements to current
work on marital biography. Specifically, if data are available, future research should incorporate
dyadic data to measure the couple as the unit of analysis rather than the individual (Zhang et al.,
2016). The lives of romantic partners are intertwined with one another such that past and current
experiences of either spouse or partner likely have an impact on the other spouse/partner. For
example, the findings from the first analytic chapter indicated that having multiple disruptions
was associated with poorer mental health in later life. Building upon this, it will likely be important to examine whether the individual’s current spouse or partner has also experienced multiple disruptions. If both partners have had multiple disruptions, disadvantage associated with more than one marital disruption could be compounded, and thus, negatively relate to the well- being of both partners. 132
Future work should also consider the role of duration prior to remarriage following different disruption pathways. More specifically, if a respondent remarries following one divorce, it may be fruitful to assess the years they spent divorced prior to their entry into remarriage. Remarrying more quickly following a marital disruption may be beneficial for one’s overall health as prior work suggests that duration divorced or widowed may be somewhat detrimental for health (Zhang & Hayward, 2006). However, remarrying after a shorter duration divorced or widowed could be increasingly harmful for parent-child ties, as both marital disruption and remarriage are associated with an adjustment period for the parent-child relationship (Connidis, 2010; de Jong Gierveld & Merz, 2013; Kalmijn, 2007; Nakonezny et al.,
2003). Therefore, if a parent experiences both disruption and a remarriage in a short period of time, this could compound the difficulty of renegotiating the parent-child relationship.
Conclusion
Marital biographies of older people are increasingly complex such that a number of older people enter mid to later life in a remarriage (Lin et al., in press), and some go on to experience another divorce or widowhood and a subsequent remarriage (Brown & Lin, 2012; Brown et al., in press; Manning & Brown, 2011). Nevertheless, researchers have not accounted for differences between disruption pathways and whether remarriage relative to remaining unmarried is associated with fewer costs following marital disruption. Moreover, research has not fully considered whether duration remarried or unmarried is positively or negatively associated with health following different disruption pathways. The findings from this dissertation show that factoring in multiple disruptions contributes to our understanding of the relationship between marital biography and well-being. Whereas fewer differences arose between divorced and 133
widowed older people, those after multiple disruptions were disadvantaged most specifically in
their mental health and contact with children.
This dissertation also investigated whether remarriage relates to fewer consequences of a disruption through one divorce, one widowhood, or multiple disruptions. The findings are mixed with regards to the role of remarriage relative to being unmarried. Remarriage was associated with fewer mental health costs following all types of disruption for men, and the physical health of divorced women. However, for parent-child contact, remarriage was positively related to
contact for divorced men, but remarriage was negatively associated with women’s contact with
children relative to being unmarried. Last, remarriage after multiple disruptions was associated
with greater ambivalence among men relative to remaining unmarried. Taken together, these
findings suggest remarriage matters when examining well-being in later life, but, remarriage may
not operate uniformly across all well-being outcomes.
It also appears that duration cannot be overlooked when considering health and parent- child contact. Still, in both instances, duration operated differently for these two aspects of well- being. On the one hand, duration remarried after multiple disruptions was harmful for the health of men relative to men who were remarried after one divorce, implying that remarriage is not completely associated with erasing health consequences of multiple disruptions. In contrast,
years remarried related to more frequent contact with children for widowed women compared to
being unmarried, indicating a benefit of remarriage. Moreover, years remarried after multiple
disruptions were associated with greater contact for men relative to the remarried after one
widowhood, suggesting men may be able to recover somewhat following multiple disruptions.
In sum, findings from this dissertation add to current literature by underscoring the need
to account for different disruption pathways. Moreover, this dissertation highlights the 134
importance of conducting more research focused on the role of remarriage in supporting or
harming the well-being of older people relative to remaining unmarried. This work also
underscores the necessity of considering duration remarried or unmarried following disruption.
Marital disruption, particularly through multiple disruptions seem to be harmful for the well-
being of older people and remarriage is not uniformly associated with fewer costs of marital
disruption, nor is duration remarried or unmarried. As marital biographies of older people
become increasingly complex, we can anticipate that a greater number of older people will
experience marital disruption, and likely more will experience multiple disruptions. The
conclusions from this dissertation suggest that in light of changing marital biographies,
particularly a greater number of individuals who have experienced multiple disruptions, older people may be increasingly susceptible to poorer health and well-being. Thus, more work is needed to better understand the role of disruption pathways, subsequent remarriage, and duration for a host of well-being indicators. Furthermore, the findings from this dissertation can inform policymakers and practitioners of the most vulnerable groups of older people who will likely need additional social and financial resources to support their overall well-being as they age. 135
Table 5.1 Summary of Results for Marital Biography and Depressive Symptoms/Chronic Conditions by Gender Depressive Symptoms Chronic Conditions Men Women Men Women
Partially supported: Hypothesis 1a: Unmarrieds after one Remarried Supported Not supported Not supported divorce had more chronic versus conditions than remarried unmarried after one divorce
Partially supported: (1) Remarrieds after one divorce had fewer depressive symptoms Partially supported: Hypothesis 1b: than remarrieds after Remarrieds after one Different multiple disruptions divorce had fewer Not supported Not supported disruption (2) Unmarrieds after one chronic conditions than pathways divorce or one remarrieds after multiple widowhood had fewer disruptions depressive symptoms than unmarrieds after multiple disruptions
Hypothesis 2a: Duration Not supported Not supported Not supported Not supported remarried or unmarried
Partially supported: Duration remarried Hypothesis 2b: after one divorce was Duration after negatively related to different Not supported Not supported Not supported depressive symptoms disruption than duration pathways remarried after multiple disruptions
Hypothesis 3a: Partially supported: Contradicted: Gender Unmarried women Unmarried men after difference in after one divorce one divorce reported remarriage after reported fewer Not supported fewer chronic Not supported different depressive symptoms conditions than disruption than unmarried men unmarried women after pathways after one divorce one divorce
Hypothesis 3b: Gender difference in duration remarried or Not supported Not supported Not supported Not supported unmarried after different disruption pathways
Hypothesis 4: Contradicted: Depressive Contradicted: Magnitude of symptoms Magnitude of depressive depressive symptoms — — versus symptoms was greater was greater than chronic than chronic conditions chronic conditions conditions
136
Table 5.2 Summary of Results for Marital Biography and Parent-Child Contact by Gender Men Women Hypothesis 1a: Contradicted: Partially supported: Remarried Remarrieds after one divorce had Unmarried women after one widowhood or versus more contact than unmarrieds after multiple disruptions had more contact relative unmarried one divorce to their remarried counterparts
Hypothesis 1b: Partially supported: Partially supported: Different disruption (1) Remarrieds after one widowhood (1) Remarrieds after one widowhood or one pathways had the most contact, then remarrieds divorce had more contact than remarrieds after one divorce, and followed by after multiple disruptions remarried after multiple disruptions (2) Unmarrieds after one widowhood had (2) For unmarried men, widowed men more contact than unmarrieds with one had the most contact with children divorce or multiple disruptions relative to unmarried men with one divorce or multiple disruptions. Unmarried men with multiple disruptions had more contact than unmarrieds with one divorce
Hypothesis 2a: Not supported Contradicted: Duration remarried Remarried women after one widowhood had or unmarried more contact with each additional year remarried relative to unmarried women after one widowhood
Hypothesis 2b: Partially supported: Not supported Duration after different Remarried men with multiple disruption pathways disruptions had more contact with each additional year remarried relative to men who remarried after one widowhood
Hypothesis 3a: Partially supported: Not supported Gender difference in Unmarried women after one divorce remarriage after different had more contact than unmarried men disruption pathways after one divorce
Hypothesis 3b: Not supported Not supported Gender difference in duration remarried or unmarried after different disruption pathways 137
Table 5.3 Summary of Results for Marital Biography and Ambivalence by Gender Men Women Hypothesis 1a: Partially supported: Not supported Remarried Remarrieds after multiple disruptions had versus greater ambivalence than unmarrieds after unmarried multiple disruptions
Hypothesis 1b: Not supported Not supported Different disruption pathways
Hypothesis 2a: Not supported Not supported Duration remarried or unmarried
Hypothesis 2b: Duration Not supported Not supported after different disruption pathways
Hypothesis 3a: Contradicted: Not supported Gender difference in Unmarried women after one divorce remarriage after different reported more ambivalence than unmarried disruption pathways men after one divorce
Hypothesis 3b: Not supported Not supported Gender difference in duration remarried or unmarried after different disruption pathways
138
REFERENCES
Acock, A. C. (2005). Working with missing values. Journal of Marriage and Family, 67, 1012-
1028.
Albertini, M., & Garriga, A. (2011). The effect of divorce on parent-child contact: Evidence on
two declining effect hypotheses. European Societies, 13, 257-278.
Amato, P. R. (2010). Research on divorce: Continuing trends and new developments. Journal of
Marriage and Family, 72, 650-666.
Amato, P. R., & Booth, A. (1991). Consequences of parental divorce and marital unhappiness for
adult well-being. Social Forces, 69, 895-914.
Amato, P. R., Rezac S. J., & Booth, A. (1995). Helping between parents and young adult
offspring: The role of parental marital quality, divorce, and remarriage. Journal of
Marriage and Family, 57, 363-374.
Amato, P. R., & Rogers S. J. (1997). A longitudinal study of marital problems and subsequent
divorce. Journal of Marriage and Family, 59, 612-624.
Aquilino, W. S. (1994). Impact of childhood family disruption on young adults’ relationships
with parents. Journal of Marriage and Family, 56, 295-313.
Aquilino, W. S., & Supple, K. R. (1991). Parent-child relations and parent’s satisfaction with
living arrangements when adult children live at home. Journal of Marriage and Family,
53, 13-27.
Barrett, A. E. (2000). Marital trajectories and mental health. Journal of Health and Social
Behavior, 41, 451-464. 139
Barrett, A. E. (2003). Socioeconomic status and age identity: The role of dimensions of health in
the subjective construction of age. The Journals of Gerontology: Psychological Sciences
and Social Sciences, 52B, S101-S109.
Barrett, A. E., & Lynch, S. M. (1999). Caregiving networks of elderly persons: Variation by
marital status. The Gerontologist, 39, 695-704.
Becker, G. S. (1981). A treatise on the family. Cambridge, MA: Harvard University Press.
Bengtson, V. L., & Harootyan, R. A. (Eds.). (1994). Intergenerational linkages: Hidden
connections in American society. New York, NY: Springer.
Bengtson, V. L., & Oyama, P. S. (2010). Intergenerational solidarity and conflict. In M. Cruz-
Saco & S. Zelenov (Eds.), Intergenerational Solidarity (pp. 35-52). New York, NY:
Palgrave Macmillan US.
Bengston, V. L., & Roberts, R. E. L. (1991). Intergenerational solidarity in aging families: An
example of formal theory construction. Journal of Marriage and Family, 53, 856-870.
Birditt, K. S., Rott, L. M., & Fingerman, K. L. (2009). ‘If you can’t say something nice, don’t
say anything at all’: Coping with interpersonal tensions in the parent–child relationship
during adulthood. Journal of Family Psychology, 23, 769-778.
Blekesaune, M. (2008). Partnership transitions and mental distress: Investigating temporal order.
Journal of Marriage and Family, 70, 879-890.
Bramlett, M. D., & Mosher, W. D. (2002). Cohabitation, marriage, divorce, and remarriage in
the United States. Center for Disease Control and Prevention. Retrieved from:
http://www.cdc.gov/nchs/data/series/sr_23/sr23_022.pdf
Brockmann, H., & Klein, T. (2004). Love and death in Germany: The marital biography and its
effect on mortality. Journal of Marriage and Family, 66, 567-581. 140
Broman, C. L. (1993). Race differences in marital well-being. Journal of Marriage and Family,
55, 724-732.
Brown, S. L. (2017). Families in America. Berkeley, CA: University of California Press.
Brown, S. L., Bulanda J. R., & Lee, G. R. (2012). Transitions into and out
of cohabitation in later life. Journal of Marriage and Family, 74, 774-793.
Brown, S. L., Bulanda, J. R., & Lee, G. R. (2005). The significance of nonmarital cohabitation:
Marital status and mental health benefits among middle-aged and older adults. The
Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 60, S21-
S29.
Brown, S. L., & Kawamura, S. (2010). Relationship quality among cohabitors and
marrieds in older adulthood. Social Science Research, 39, 777-786.
Brown, S. L., & Lin, I.-F. (2012). The gray divorce revolution: Rising divorce among middle-
aged and older adults, 1990-2010. The Journals of Gerontology, Series B: Psychological
and Social Sciences, 67, 731-741.
Brown, S. L., Lin, I.-F., Hammersmith, A. M., & Wright, M. R. (forthcoming). Later life marital
dissolution and repartnership status: A national portrait. The Journals of Gerontology,
Series B: Psychological Sciences and Social Sciences, 00, 1-11.
Brown, S. L., & Shinohara, S. K. (2013). Dating relationships in older adulthood: A
national portrait. Journal of Marriage and Family, 75, 1194-1202
Brown, S. L., & Wright, M. R. (2017). Marriage, cohabitation, and divorce in later
life. Innovation in Aging, 1, 1-11. 141
Brown, T. H., O’Rand, A. M., & Adkins, D. E. (2012). Race-ethnicity and health trajectories:
Tests of three hypotheses across multiple groups and health outcomes. Journal of Health
and Social Behavior, 53, 359-377.
Bulanda, J. R., & Brown, S. L. (2007). Race-ethnic differences in marital quality and divorce.
Social Science Research, 36, 945-967.
Carr, D., & Springer, K. W. (2010). Advances in families and health research in
the 21st century. Journal of Marriage and Family 72, 743-761.
Cartensen, L. L. (1992). Social and emotional patterns in adulthood: Support for socioemotional
selectivity theory. Psychology and Aging, 7, 331-338.
Centers for Disease Control and Prevention. (2009, May). NCHS data brief: Changing patterns
of nonmarital childbearing in the United States (No. 18). Hyattsville, MD: National
Center for Health Statistics.
Cherlin, A. J. (2004). The deinstitutionalization of American marriage. Journal of Marriage and
Family, 66, 848-861.
Cherlin, A. J. (2009). The marriage-go-round: The state of marriage and the family in America
today. New York: Random House, Inc.
Cherlin, A. J. (2010). Demographic trends in the United States: A review of research in the
2000s. Journal of Marriage and Family, 72, 403-419.
Coleman, M., & Ganong, L. (2008). Normative beliefs about sharing housing with an older
family member. The International Journal of Aging and Human Development, 66, 49-72.
Connidis, I. A. (2010). Divorce and union dissolution: Reverberations across three generations.
Canadian Journal on Aging, 22, 353-368. 142
Connidis, I. A., & McMullin, J. A. (2002). Sociological ambivalence and family ties: A critical
perspective. Journal of Marriage and Family, 64, 558-567.
Daatland, S. O. (2007). Marital history and intergenerational solidarity: The impact of divorce
and unmarried cohabitation. Journal of Social Issues, 63, 809-825.
Damman, M., & van Duijn, R. (2016). Intergenerational support in the transition from work to
retirement. Work, Aging and Retirement, 3, 66-76.
Dannefer, D. (2003). Cumulative advantage/disadvantage, and the life course: Cross-
fertilizing age and social science knowledge. The Journals of Gerontology Series B:
Psychological Sciences and Social Sciences, 58, S327-S337.
De Graaf, P. M., & Fokkema, T. (2007). Contacts between divorced and non-divorced parents
and their adult children in the Netherlands: An investment perspective. European
Sociological Review, 23, 263-277.
De Graaf, P. M., & Kalmijn, M. (2006). Change and stability in the social determinants of
divorce: A comparison of marriage cohorts in the Netherlands. European Sociological
Review, 22, 561-572.
De Jong Gierveld, J., & Dykstra, P. A. (2002). The long-term rewards of parenting:
older adults’ marital history and the likelihood of receiving support from adult
children. Ageing International, 27, 49-69.
De Jong Gierveld, J., & Peeters, A. (2003). The interweaving of repartnered
older adults’ lives with their children and siblings. Ageing & Society, 23, 187-205.
De Jong, Gierveld, J., & Merz, E.-M. (2013). Parents’ partnership decision making after divorce
or widowhood: The role of (step)children. Journal of Marriage and Family, 75, 1098-
1113. 143
Dupre, M. E., Beck, A. N., & Meadows, S. O. (2009). Marital trajectories and mortality among
U.S. adults. American Journal of Epidemiology, 170, 546-555.
Dupre, M. E., & Meadows, S. O. (2007). Disaggregating the effects of marital trajectories on
health. Journal of Family Issues, 28, 623-652.
Duncan, S., & Phillips, M. (2011). People who live apart together (LATs): New
family form or just a stage? International Review of Sociology 21, 513-532.
Dykstra, P. A., & de Jong Gierveld, J. (2006). Gender and marital-history differences in
emotional and social loneliness among Dutch older adults. Canadian Journal on Aging,
23, 141-155.
Dykstra, P. A., & Fokkema, T. (2011). Relationships between parents and their adult children: A
West European typology of late-life families. Aging and Society, 31, 545-569.
Eggebeen, D. J. (1992). Family structure and intergenerational ties. Research on Aging, 14, 427-
447.
Elder, G. H. Jr. (1994). Time, human agency, and social change: Perspectives on the life course.
Social Psychology Quarterly, 57, 4-15.
Elwert, F., & Christakis, N. A. (2006). Widowhood and race. American Sociological Review, 71,
16-41.
Federal Interagency Forum on Aging-Related Statistics. (2016). Older Americans 2016: Key
indicators of well-being. Washington, DC: Federal Interagency Forum on Aging-Related
Statistics.
Ferraro, K. F., & Barresi, C. M. (1982). The impact of widowhood on the social relations of
older persons. Research on Aging, 4, 227-247. 144
Fingerman, K. L., Cheng, Y. P., Birditt, K., & Zarit, S. (2011). Only as happy as the least happy
child: Multiple grown children’s problems and successes and middle-aged parents’ well-
being. Journals of Gerontology, Series B: Psychological and Social Sciences, 67, 184-
193.
Fingerman, K. L., Pitzer, L., Lefkowitz, E. S., Birditt, K. S., & Mroczek D. (2008). Ambivalence
relationship qualities between adults and their parents: Implications for the well-being of
both parties. The Journal of Gerontology: Psychological Sciences and Social Sciences,
63, P362-P371.
Fingerman, K. L., Miller, L. M., Birditt, K. S., & Zarit, S. H. (2009). Giving to the good and the
needy: Parental support of grown children. Journal of Marriage and Family, 71, 1220-
1233.
Fingerman, K. L., VanderDrift, L. E., Dotterer, A. M., Birditt, K. S., & Zarit, S. H. (2009).
Support to aging parents and grown children in Black and White families. The
Gerontologist, 51, 441-452.
Flippen, C., & Tienda M. (2000). Pathways to retirement: Patterns of labor force participation
and labor market exit among the pre-retirement population by race, Hispanic origin, and
sex. The Journals of Gerontology: Psychological Sciences and Social Sciences, 55, S14-
S27.
Franks, P., Gold, M. R., & Fiscella, K. (2003). Sociodemographics, self-rated health, and
mortality in the U.S. Social Science & Medicine, 56, 2505-2514.
Furstenberg, F. F. Jr., Hoffman, S. D., & Shrestha, L. (1995). The effect of divorce on
intergenerational transfers: New evidence. Demography, 32, 319-333. 145
Ganong, L., & Coleman, M. (2006). Patterns of exchange and intergenerational responsibilities
after divorce and remarriage. Journal of Aging Studies, 20, 265-278.
Gardner, J., & Oswald, A. (2004). How is mortality affected by money, marriage, and stress?
Journal of Health Economics, 23, 1181-1207.
Goldstein, J. R., & Kenney, C. T. (2001). Marriage delayed or marriage forgone? New forecasts
of first marriage for U.S. women. American Sociological Review, 66, 506-519.
Guiaux, M., van Tilburg, T., & Broese van Groenou, M. (2007). Changes in contact and support
exchanges in personal networks after widowhood. Personal Relationships, 14, 457-473.
Ha, J.-H. (2008). Changes in support from confidants, children, and friends following
widowhood. Journal of Marriage and Family, 70, 306-318.
Ha, J.-H., Carr, D., Utz, R. L., & Nesse, R. (2006). Older adults’ perceptions of intergenerational
support after widowhood: How do men and women differ? Journal of Family Issues, 27,
3-30.
Ha, J.-H., & Ingersoll-Dayton, B. (2008). The effect of widowhood on intergenerational
ambivalence. Journals of Gerontology: Social Sciences, 63B, S49-S58.
Ha, J.-H., Kahng, S. K., & Choi, N. (2017). Reciprocal effects between health and social support
in older adults’ relationships with their children and friends. Research on Aging, 39, 300-
321.
Hammersmith. A. M. (In press). Life interrupted: Parents’ positivity and negativity toward
children following children’s and parents’ transitions later in life. The Gerontologist, 00,
00-00. 146
Hayward, M. D., Miles, T. P., Crimmins, E. M., & Yang, Y. (2000). The significance of
socioeconomic status in explaining the racial gap in chronic condition health conditions.
American Sociological Review, 65, 910-930.
Heaton, T. B. (2002). Factors contributing to increasing marital stability in the United States.
Journal of Family Issues, 23, 392-409.
Hedeker, D. (2008). Multilevel models for ordinal and nominal variables. In J. de Leeuw & E.
Meijer (Eds.), Handbook of Multilevel Models (pp. 237-274). New York, NY: Springer
Science+Business Media, LLC.
Hemez, P., & Brown, S. L. (2016). Cohabitation in middle and later life. Family Profiles, FP-16-
20. Bowling Green, OH: National Center for Marriage & Family Research. Retrieved
from: http://www.bgsu.edu/ncfmr/resources/data/family-profiles/hemez-brown-
cohabitation-middle-later-life-2014-fp-16-20.html.
Henretta, J. C. (2010). Lifetime marital history and mortality after age 50. Journal of Aging and
Health, 22, 1198-1212.
Hughes, M. E., & Waite, L. J. (2009). Marital biography and health at mid-life. Journal of
Health and Social Behavior, 50, 344-358.
Hungerford, T. L. (2001). The economic consequences of widowhood on elderly women in the
United States and Germany. The Gerontologist, 41, 103-110.
Hu, Y., & Goldman, N. (1990). Mortality differentials by marital status: An international
comparison. Demography, 27, 233-250.
Johnson, N. J., Backlund, E., Sorlie, P. D., & Loveless, C. A. (2000). Marital
status and mortality: The National Longitudinal Mortality Study. Annals of
Epidemiology, 10, 224-238. 147
Johnson, D. R., & Wu, J. (2002). An empirical test of crisis, social selection, and role
explanations of the relationship between marital disruption and psychological distress: A
pooled time-series analysis of four-wave panel data. Journal of Marriage and Family, 64,
211-224.
Kalmijn, M. (2007). Gender differences in the effects of divorce, widowhood, and remarriage on
intergenerational support: Does marriage protect fathers? Social Forces, 85, 1079-1104.
Kalmijn. M. (2013). Adult children’s relationships with married parents, divorced parents, and
stepparents: Biology, marriage, or residence. Journal of Marriage and Family, 75, 1181-
1193.
Kalmijn, M. (2013). Long-term effects of divorce on parent-child relationships: Within-family
comparisons of fathers and mothers. European Sociological Review, 29, 888-898.
Kalmijn, M. (2015). Relationships between fathers and adult children: The cumulative effects of
divorce and repartnering. Journal of Family Issues, 36, 737-759.
Kalmijn, M. (2017). The ambiguous link between marriage and health: A dynamic reanalysis of
loss and gain effects. Social Forces, 95, 1607-1636.
Kalmijn, M., & de Graaf, P. M. (2012). Life course changes of children and well-being of
parents. Journal of Marriage and Family, 74, 269-280.
Karraker, A., & Latham, K. (2015). In sickness and in health? Physical illness as a risk factor for
marital dissolution in later life. Journal of Health and Social Behavior, 56, 59-73.
Kaufman, G., & Uhlenberg, P. (1998). Effects of life course transitions on the quality of
relationships between adult children and their parents. Journal of Marriage and Family,
60, 924-938. 148
Keith, P. M., Kim, S., & Schafer, R. B. (2000). Informal ties of the unmarried in middle and later
life. Sociological Spectrum, 20, 221-238.
Kiecolt, J. K., Blieszner, R., & Salva, J. (2011). Long-term influences of intergenerational
ambivalence on midlife parents’ psychological well-being. Journal of Marriage and
Family, 73, 369-382.
King, V., & Scott. M. E. (2005). A comparison of cohabiting relationships among older and
younger adults. Journal of Marriage and Family, 67, 271-285.
Lawton, L., Silverstein, M., & Bengston, V. (1994). Affection, social contact, and geographic
distance between adult children and their parents. Journal of Marriage and Family, 56,
57-68.
Lee, G. R. (2015). The limits of marriage: Why getting married won’t solve all of our problems.
Lanham, MD: Lexington Books.
Lendon, J. P., Silverstein, M., & Giarrusso, R. (2014). Ambivalence in older parent-adult child
relationships: Mixed feelings, mixed measures. Journal of Marriage and Family, 76,
272-284.
Lillard, L. A., & Waite, L. J. (1995). ’Til death do us part: Marital disruption and mortality.
American Journal of Sociology, 100, 1131-1156.
Lin, I.-F. (2008). Consequences of parental divorce for adult children’s support of their frail
parents. Journal of Marriage and Family, 70, 113-128.
Lin, I.-F., & Brown, S. L. (2012). Unmarried boomers confront old age: A national portrait. The
Gerontologist, 52, 153-165.
Lin, I.-F., Brown, S. L., & Hammersmith, A. M. (2017). Marital biography, social security
receipt, and poverty. Research on Aging, 39, 86-110. 149
Lin, I.-F., & Wu, H.-S. (In press). Early-life parent-child relationships and adult children’s
support of unpartnered parents in later life. The Journals of Gerontology, Series B:
Psychological and Social Sciences, 00, 00-00.
Liu, H. (2012). Marital dissolution and self-rated health: Age trajectories and birth cohort
variations. Social Science & Medicine, 74, 1107-1116.
Liu, H., & Johnson, D. (2009). Till death do us part: Marital status and U.S. mortality trends,
1986-2000. Journal of Marriage and Family, 71, 1158-1173.
Liu, H., & Umberson, D. J. (2008). The times they are a changin’: Marital status and health
differentials from 1972 to 2003. Journal of Health and Social Behavior, 49, 239-253.
Lüscher, K., & Pillemer, K. (1998). Intergenerational ambivalence: A new approach to the study
of parent-child relations in later life. Journal of Marriage and Family, 60, 413-445.
Mancini, J. A., & Blieszner, R. (1989). Aging parents and adult children: Research themes in
intergenerational relations. Journal of Marriage and Family, 51, 275-290.
Manlove, J., Wildsmith, E., Ikramullah, E., Ryan, S., Holcombe, E., Scott, M., & Peterson, K.
(2012). Union transitions following the birth of a child to cohabiting parents. Population
Research and Policy Review, 31, 361-386.
Manning, W. D., & Brown, S. L. (2011). The demography of unions among older
Americans, 1980-present: A family change approach. In R. A. Settersten & J. L. Angel
(Eds.), Handbook of Sociology of Aging (pp. 193-210) New York: Springer.
Manning, W. D., & Smock, P. J. (2000). “Swapping” families: Serial parenting and economic
support for children. Journal of Marriage and Family, 62, 111-122.
Manzoli, L., Villari, P., Pirone, G. M., & Boccia, A. (2007). Marital status and mortality in the
elderly: A systematic review and meta-analysis. Social Science & Medicine, 64, 77-94. 150
Marks, N. (1998). Does it hurt to care? Caregiving, work-family conflict, and midlife well-being.
Journal of Marriage and Family, 60, 951-966.
Marks, N. F., & Lambert, J. D. (1998). Marital status continuity and change among young and
midlife adults: Longitudinal effects on psychological well-being. Journal of Family
Issues, 19, 652-686.
McEwen, B. S., & Stellar, E. (1993). Stress and the individual: Mechanisms leading to disease.
Archives of Internal Medicine, 153, 2093-2101.
McFarland, M. J., Hayward, M. D., & Brown, D. (2013). I’ve got you under my skin: Marital
biography and biological risk. Journal of Marriage and Family, 75, 363-380.
Miech, R. A., & Shanahan, M. J. (2000). Socioeconomic status and depression over the life
course. Journal of Health and Social Behavior, 41, 162-176.
Nakonezny, P. A., Rodgers, J. L., & Nussbaum, J. F. (2003). The effect of later-life parental
divorce on adult-child/older-parent solidarity: A test of the buffering hypothesis. Journal
of Applied Social Psychology, 33, 1153-1178.
Noël-Miller, C. M. (2013). Repartnering following divorce: Implications for older fathers’
relationships with their adult children. Journal of Marriage and Family, 75, 697-712.
Noël-Miller, C. M. (2013). Former stepparents’ contact with their stepchildren after midlife. The
Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 68,
S409-S419.
O’Rand, A. M. (1996). The precious and the precocious: Understanding cumulative disadvantage
and cumulative advantage over the life course. The Gerontologist, 36, 230-238. 151
Orbuch, T. L., Thronton, A., & Cancio, J. (2000). The impact of marital quality, divorce, and
remarriage on the relationships between parents and their children. Marriage and Family
Review, 29, 221-246.
Ofstedal, M. B., Weir, D. R., Chen, K. T., & Wagner, J. (2011). Updates to HRS sample weights
(HRS Report No. DR-013). Retrieved from University of Michigan, Health and
Retirement Study website: http://hrsonline.isr.umich.edu/sitedocs/userg/dr-013.pdf
Pezzin, L. E., Pollak, R. A., & Schone, B. S. (2008). Parental marital disruption, family type, and
transfers to elderly, disabled parents. Journal of Gerontology: Social Sciences, 63B,
S349-S358.
Pezzin, L. E., & Schone, B. S. (1999). Parental marital disruption and intergenerational transfers:
An analysis of lone elderly parents and their children. Demography, 36, 287-297.
Pillemer, K., Munsch, C. L., Fuller-Rowell, T., Riffin, C., & Suitor, J. J. (2012). Ambivalence
toward adult children: Differences between mothers and fathers. Journal of Marriage and
Family, 74, 1101-1113.
Reczek, C., Pudrovska, T., Carr, D., Thomeer, M. B., & Umberson, D. J. (2016). Marital
histories and heavy alcohol use among older adults. Journal of Health and Social
Behavior, 57, 77-96.
Rendall, M. S., Weden, M. M., Favreault, M. M., & Waldron, H. (2011). The protective effect of
marriage for survival: A review and update. Demography, 48, 481-506.
Roan, C. L., & Raley, R. K. (1996). Intergenerational coresidence and contact: A longitudinal
analysis of adult children’s response to their mother’s widowhood. Journal of Marriage
and Family, 58, 708-717. 152
Rosenthal, C. J. (1985). Kinkeeping in the familial division of labor. Journal of Marriage and
Family, 4, 965-974.
Rossi, A. S., & Rossi, P. H. (1990). Of human bonding: Parent-child relations across the life
course. New York, NY: Walter de Gruyter, Inc.
Sarkisian, N., & Gerstel, N. (2008). Till marriage do us part: Adult children’s relationships with
their parents. Journal of Marriage and Family, 70, 360-276.
Schenk, N., & Dykstra, P. A. (2012). Continuity and change in intergenerational family
relationships: An examination of shifts in relationship type over a three-year period.
Advances in Life Course Research, 17, 121-132.
Seltzer, J. A., & Bianchi, S. M. (2013). Demographic change and parent-child relationships in
adulthood. Annual Review of Sociology, 39, 275-290.
Shapiro, A. (2003). Later-life divorce and parent-adult child contact and proximity: A
longitudinal analysis. Journal of Family Issues, 24, 264-285.
Shapiro, A. (2004). Revisiting the generation gap: Exploring the relationships of parent/adult
child dyads. International Journal of Aging and Human Development, 58, 127-146.
Shor, E., Roelfs, D. J., Bugyi, P., & Schwartz, J. E. (2012). Meta-analysis of marital dissolution
and mortality: Reevaluating the intersection of gender and age. Social Science Medicine,
75, 1-35.
Silverstein, M., & Bengston, V. L. (1997). Intergenerational solidarity and the structure of adult
child-parent relationships in American families. American Journal of Sociology, 103,
429-460.
Silverstein, M., & Giarusso, R. (2010). Aging and family life: A decade in review. Journal of
Marriage and Family, 72, 1039-1058. 153
Simon, R. W. (2002). Revisiting the relationships among gender, marital status, and mental
health. American Journal of Sociology, 107, 1065-1096.
Singer, B., & Ryff, C. D. (1999). Hierarchies of life histories and associated health risks. Annals
of the New York Academy of Sciences, 896, 96-115.
Strohschein, L., McDonough, P., Monette, G., & Shao, Q. (2005). Marital transitions and mental
health: Are there gender differences in the short-term effects of marital status change?
Social Science & Medicine, 61, 2293-2303.
Sweeney, M. M. (2002). Remarriage and the nature of divorce: Does it matter which spouse
chose to leave? Journal of Family Issues, 23, 410-440.
Thomas, P. A. (2011). Trajectories of social engagement and limitations in late life. Journal of
Health and Social Behavior, 52, 430-443.
Thompson, M. M., Zanna, M. P., & Griffin, D. W. (1995). Let’s not be indifferent about
(attitudinal) ambivalence. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength:
Antecedents and consequences (pp. 361–386). Hillsdale, NJ: Erlbaum.
Umberson, D. J. (1992). Relationships between adult children and their parents: Psychological
consequences for both generations. Journal of Marriage and the Family, 54, 664-674.
Umberson, D. J., Thomeer, M. & Williams, K. (2013). Family status and mental health: Recent
advances and future directions. In C. S. Aneshensel, J. C. Phelan, & A. Bierman (Eds.)
Handbook of the sociology of mental health (pp. 405-431). Dordrecht, The Netherlands:
Springer.
Wade, T. J., & Pevalin, D. J. (2004). Marital transitions and mental health. Journal of Health and
Social Behavior, 45, 155-170. 154
Waite, L. J., & Harrison, S. C. (1992). Keeping in touch: How women in mid-life allocate social
contacts among kith and kin. Social Forces, 70, 637-654.
Ward, R. A. (2008). Multiple parent-adult child relations and well-being in middle and later life.
The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 63,
S239-S247.
Ward, R. A., Deane, G., & Spitze, G. (2014). Life-course changes and parent-adult child contact.
Research on Aging, 36, 568-602.
Ward, R. A., Spitze, G., & Deane, G. (2009). The more the merrier? Multiple parent-adult child
relations. Journal of Marriage and Family, 71, 161-173.
Wijckmans, B., & van Bavel, J. (2013). Divorce and adult children’s perceptions of family
obligations. Journal of Comparative Family Studies, 44, 291-310.
Williams, K., & Umberson, D. J. (2004). Marital status, marital transitions, and health: A
gendered life course perspective. Journal of Health and Social Behavior, 45, 81-98.
Willson, A. E., Shuey, K. M., Elder, G. H. Jr., & Wickrama, K. A. S. (2006). Ambivalence in
mother-child relations: A dyadic analysis. Social Psychology Quarterly, 69, 235-252.
Wilmoth, J., & Koso, G. (2002). Does marital history matter? Marital status and wealth
outcomes among preretirement older adults. Journal of Marriage and Family, 64, 254-
268.
Wolff, J. L., & Kasper, J. D. (2006). Caregivers of frail elders: Updating a national profile. The
Gerontologist, 46, 344-356.
Wright, M. R. (2017). Cohabitation among older adults: Well-being, relationships with adult
children, and perceptions of care availability (Unpublished doctoral dissertation).
Bowling Green State University, Bowling Green, Ohio. 155
Wright, M. R., & Brown, S. L. (2017). Psychological well-being among older adults: The role of
partnership status. Journal of Marriage and Family, 79, 833-849.
Wu, H. (2018, April). Older adults who live apart together (LAT): A comparison of unmarried,
dating, LAT, cohabitation, remarriage, and first marriage. Research presented at
Population Association of America, Denver.
Wu, H., & Brown, S. L. (2016). Long-term marriage among older adults. Family Profiles, FP-
16-08. Bowling Green, OH: National Center for Marriage & Family Research. Retrieved
from: https://www.bgsu.edu/ncfmr/resources/data/family-profiles/wu-brown-longterm-
marriage-older-adults-fp-16-08.html.
Yu, Y.-L. (2016). Marital biography and chronic disease progression in mid- and late life
(Unpublished doctoral dissertation). Michigan State University, Lansing, Michigan.
Zhang, Z. (2006). Marital history and the burden of cardiovascular disease in midlife. The
Gerontologist, 46, 266-270.
Zhang, Z., Liu, H., & Yu, Y. (2016). Marital biography and health in middle and late life. In J.
Bookwala (Ed.), Couple relationships in the middle and later years: Their nature,
complexity, and role in health and illness (pp. 199-218). Washington D.C.: American
Psychological Association.