How Do Coresidence and Relationship Quality Matter?

Midlife Parenthood and Wellbeing: How do Coresidence and Relationship Quality Matter?

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Zhe Zhang

Graduate Program in Sociology

The Ohio State University

2018

Dissertation Committee

Professor Corinne Reczek, Advisor

Professor Cynthia G. Colen

Professor Kristi Williams

Copyrighted by

Zhe Zhang

2018

Abstract

Parenthood, one of the most important and long-lasting social roles, shapes almost every aspect of individual wellbeing across the life course. Yet, how the varying contexts of raising young children into adulthood matter for individual wellbeing across midlife remains understudied. This research gap warrants immediate attention given the drastic changes in parenthood for the past 50 years. Postponement in childbearing, increasing investment on children, and adult children’s delayed departure from parental household all contribute to reshaping the parenthood landscape at midlife (age 40 to 60). In addition, midlife is a life course stage when population health differences further diverge – changes in midlife parenthood may escalate such health differences further. Drawing on a life course perspective and prior literature on family and health, my dissertation addresses this research gap by answering the following question – how do various parenthood contexts matter for midlife adults’ mental and physical wellbeing? Using data from

NLSY79 and appropriate quantitative methods, my three empirical chapters examine how two main aspects of parent-child ties – structure (e.g., coresidence) and content (e.g., relationship quality) are associated with adults’ mental and physical health across midlife.

The first empirical chapter examines how children’s life course stage and parent- child coresidence at midlife matters for individuals’ psychological wellbeing, and further tests how this relationship differs for men and women. Results from OLS regression

ii models suggest that mothers with only adult children in the household experienced higher levels of psychological distress than other midlife women, whereas coresidence with both minor and adult children might be conducive to mental wellbeing for both men and women at midlife. The second empirical chapter tests how three emerging mother-child coresidential biographies: the gone-for-good, the boomerangers, and the never-left, matter for mothers’ BMI trajectories. Results from growth curve modeling suggest that boomerangers’ mothers were at higher risk of gaining additional weight across midlife relative to mothers to the gone-for-good. In turn, mothers whose children never left the home experienced declining body weight over time relative to mothers whose children left “on time.” The last empirical chapter examined the association between intergenerational relationship quality and midlife mother’s mental wellbeing. Results from OLS regression models with lagged dependent variables suggest that midlife mothers in a collective ambivalent relationship with multiple adolescent and young adult children had more psychological distress than mothers in a uniformly close or a uniformly unclose relationship with all children. Mothers with at least one child reporting declining feelings toward mother also had worse mental health over time.

Taken together, the findings elaborate that increasing parenting demands and responsibilities to adolescent and coming of age children may take a toll on midlife parents’, especially midlife mother’s health. This dissertation provides research evidence to policy makers in helping and caring for a broader population of American families.

iii

Acknowledgments

The completion of this dissertation and the completion of this degree would not have been possible without the support and encouragement of my mentors, friends, and family. First, I would like to express my deepest gratitude to my advisor and committee chair, Corinne Reczek, for her unstinted support and constant encouragement. Through our weekly meetings and countless exchanges of manuscripts, Corinne is critical in helping me cultivate confidence and competence as a social scientist. She is the epitome of a scholar and teacher, whose strong belief that research and mentorship go hand-in- hand is something I want to emulate in the future. I would also like to express my heartfelt gratitude to my committee members Cynthia Colen and Kristi Williams for their valuable feedback on my work. Cindy’s constant emphasis on the “big picture” has brought me back to the drawing boards many times, and always for the better. Kristi’s seminal work on family and health has been my go-to source for inspiration to produce timely scholarly research that helps improve the wellbeing of families and push for policy changes. Finally, I would like to express my sincere appreciation to all my committee members for their unbelievable support through the job market season. I would not have been able to “keep trucking” during this brutal time without their support.

I would also like to thank my friends for their valuable friendship and support. I am particularly grateful to Li Zhang and Yanting Guo, who often remind me not to lose

iv sight of the “big picture” in life and to develop a greater belief in myself. I would also like to extend my gratitude to Li Xiao for reminding me not to dwell on the occasional setbacks and to embrace the challenges. My thanks also go to Fiona Lin, who is only a phone call away when I need her support and advice.

I have also been fortunate enough to receive support and encouragement from numerous fellow graduate students, additional faculty, and staff. Special thanks go to

Alexandra Kissling, Lauren Gebhardt-Kram, Yue Qian, Siqi Han, Claudia Buchmann,

John Casterline, Liana Sayer, Hui Zheng, Reanne Frank, Claire Kamp Dush, Hollie

Nyseth Brehm, Susan Pennington, Mary McKay, Jacob Tarrence, Paola Echave, Jasmine

Whiteside, Christopher Munn, Johnathan Dirlam, Bethany Boettner, Jill Morris, Amelia

Qi Li, Jill Yavorsky, Anna Muraveva, Aimee Yoon, Laura DeMarco, Erica Phillips,

Emma Bosley-Smith, Brandon Moore, Kait Smeraldo, and Guangyi Wang.

I gratefully acknowledge the financial support from a number of sources including the Coca-Cola – Critical Difference for Women Graduate Studies Grants for

Research on Women, Gender, and Gender Equity and the summer fellowship from the

Institute of Population Research (IPR). I would also like to thank IPR for the cozy cubicle and all the lovely conversations that happened in the IPR kitchen.

Finally, I would like to thank my parents, Huake Zhang and Qunni Chen, for being the relentless supporters of my pursuits. Their unconditional love and support are the true examples of how intergenerational support can serve as an important buffer against stress and improve the wellbeing of family members across the life course.

Vita

2011...... B.A. Sociology, Susquehanna University

2013...... M.A. Sociology, The Ohio State University

Publications

Qian, Yue, Claudia Buchmann, and Zhe Zhang. 2018. Gender differences in educational adaptation of immigrant-origin youth in the United States. Demographic Research, 38, 1155-1188.

Zhang, Zhe. 2017. “Division of Housework in Transitional Urban China” in Chinese Sociological Review, 49(3), 263-291. doi: http://dx.doi.org/10.1080/21620555.2017.1295809

Reczek, Corinne and Zhe Zhang. 2016. “Parent-child Relationships and Parent Psychological Distress: How do Social Support, Strain, Dissatisfaction, and Equity Matter?” Research on Aging 38(7), 742-766. doi: 10.1177/0164027515602315

Fields of Study

Major Field: Sociology

Table of Contents

Abstract ...... ii Acknowledgments...... iv Vita ...... vi List of Tables ...... ix List of Figures ...... x Chapter 1. Introduction ...... 1 References ...... 5 Chapter 2. Midlife Parental Context and Mental Health ...... 6 Background ...... 8 Parenthood context and psychological wellbeing at midlife ...... 8 Methods...... 15 Data ...... 15 Measures ...... 16 Analytic Strategy ...... 18 Results ...... 19 Descriptive Results ...... 19 Multivariate Results ...... 20 Discussion ...... 21 References ...... 25 Chapter 3. Intergenerational Coresidence and Mothers’ Body Weight at Midlife ...... 37 Background ...... 38 Intergenerational Coresidence and Mother’s Body Weight: A Life Course Perspective ...... 38 The Present Study ...... 46 Data and Methods ...... 46 Description of the Data ...... 46 Description of the Measures ...... 48 Analytical Approach ...... 51 vii

Results ...... 52 Descriptive Findings ...... 52 Multilevel Regression Results ...... 53 Discussion ...... 56 References ...... 62 Chapter 4. Mother-child relationship quality and mental health at midlife ...... 79 Theoretical and Empirical Background ...... 81 Midlife mothers’ and intergenerational relationship quality mental wellbeing ...... 81 Rethinking intergenerational relationship quality: One mother and multiple children ...... 83 Intergenerational Relationship Quality and Mother’s Mental Health: a longitudinal study design ...... 85 Other Factors Associated with Intergenerational Ties and Mental Health ...... 86 The present study ...... 87 Data and Methods ...... 88 Data ...... 88 Measures ...... 89 Analytical Strategy ...... 90 Results ...... 91 Descriptive Results ...... 91 Multivariate Results ...... 92 Discussion ...... 94 References ...... 98 Chapter 5. Conclusion ...... 110 References ...... 115 Complete References ...... 118 Appendix. Additional Tables ...... 138

viii

List of Tables

Table 1. Sociodemographic and Health Characteristics of the Sample (Unweighted, Imputed). 32

Table 2. Results from OLS Regression Models Predicting Psychological Distress at Age 40 ... 34

Table 3. Results from OLS Regression Models Predicting Psychological Distress at Age 50 ... 35

Table 4. Sociodemographic and Health Characteristics of the Sample (Imputed) ...... 72

Table 5. Results from Multilevel (Growth Curve) Regression Models with Random Coefficients Predicting Mom's BMI by Young Adult's Residential Biography, 1994-2014 ...... 75

Table 6. Sociodemographic and Health Characteristics of the Sample (Unweighted, Imputed) ...... 106

Table 7. Results from OLS Regression Models Predicting Psychological Distress at Wave 2...... 107

Table 8. Results from OLS Regression Models Predicting Psychological Distress at Wave 2...... 108

Table 9. Results from OLS Regression Models Predicting Psychological Distress at Age 40...... 138

Table 10. Results from OLS Regression Models Predicting Psychological Distress at Wave 2 ...... 140

Table 11. Results from OLS Regression Models Predicting Psychological Distress at Wave 2 ...... 141

List of Figures

Figure 1. Predicted Psychological Distress by the Composition of Coresident Children.

...... 36

Figure 2. Predicted BMI by Coresidential Pattern (Table 5- Model 1) ...... 78

Figure 3. Predicted BMI by Coresidential Pattern (Table 5- Model 3) ...... 78

Chapter 1. Introduction

As one of the most important and long-lasting social roles, parenthood, is endowed with profound social expectations that shape almost every aspect of individual wellbeing across the life course (Umberson, Pudrovska, & Reczek, 2010). Yet, most of the existing literature on parenthood and well-being either emphasizes the immediate health effects of parenting young children, or focuses on older parents’ wellbeing in relationship to their ties with adult children and care arrangements. A less developed body of research examines how the varying contexts of raising young children into adulthood influence parental wellbeing across midlife (Umberson et al., 2010; Fingerman, 2017). This research gap needs immediate attention given the dramatic changes in midlife parenthood over the past 50 years. Midlife parents today are more likely to postpone childbearing compared to their parents’ generation (Eickmeyer, 2016). And, overall, parental investment in children in terms of time and economic resources has increased; parents continue to offer substantial economic, emotional, and residential assistance to young adult children (Fingerman et al., 2012). These changes have significant effects on what parenthood looks like at midlife (age 40 to 60), including family composition and parenting responsibilities.

Notably, midlife is also a time when population health differences begin to diverge more dramatically (Read & Gorman, 2011; Palloni et al. 2009). Thus, increasing health disparities at midlife may in part be due to differences in parenthood contexts at midlife — this dissertation tests this relationship.

Overall, my dissertation asks the question: how are different parenthood contexts related to adults’ mental and physical health at midlife? Drawing on a life course perspective and prior research on family and health, this dissertation examines how two main aspects of parent-child ties – structure (e.g., coresidence) and content (e.g., emotional support) are linked with adults’ health at midlife. In all chapters, I use data from NLSY79, a US nationally representative and ongoing longitudinal dataset that uniquely includes data from both midlife adults and their children, and appropriate analytical methods to address the research questions. Below, I introduce my dissertation, composed of three empirical essays, in more detail.

The first empirical chapter examines the association between parenthood heterogeneity and mental wellbeing at midlife. In tandem with the recent demographic trends including the childbearing postponement and adult children’s delayed departure from home, parenthood at midlife has become increasingly diverse. Such heterogeneity has implications for health, as parenting minor children and adult children are found to have different ramifications for parental wellbeing including mental health. Drawing on a gendered life course approach, this chapter tackles two questions: (1) how do parenthood contexts in terms of children’s life course stage and parent-child coresidence at midlife matter for individuals’ psychological wellbeing?; and (2) how does this relationship differ for men and women? According to results generated from OLS regression models and fixed effects models, mothers to the coresidential adult children were at risk of elevated psychological distress, whereas the coresidence with both minor and adult children appears to experience a mental health benefit. This chapter contributes to existing literature by depicting the landscape as well as the mental health effects of midlife parenthood among the late baby boomers.

My dissertation next follows children into adulthood, using longitudinal data from both mothers and their children in NLSY79 to track mothers’ body weight alongside one key aspect of motherhood: mother-child coresidential patterns. The living arrangements of young adults have drawn widespread public attention, with studies showing that young adults now live with parents for longer period of time, tend to boomerang back to the parental home after the initial departure, and rely on parents for both economic and emotional resources. Most research on this topic focuses on the effects of adult children’s coresidential patterns on children’s well-being.

Yet, how these processes matter for mothers is virtually unknown. To address this research gap, the second empirical chapter tests how three emerging mother-child coresidential biographies: the gone-for-good, the boomerangers, and the never-left, matter for mothers’ BMI trajectories.

Results from growth curve modeling show that boomerangers’ mothers are at higher risk of gaining additional weight across midlife relative to mothers who have children who left and never returned over the same period. In turn, mothers whose children never left the home experienced declining body weight over time relative to mothers whose children left “on time”.

The findings illuminate how young adults’ coming-of-age experiences have rippling effects on maternal wellbeing beyond the traditionally defined core parenting years.

Notably, while the first two chapters focus on the structural part of the parent-child ties – coresidence, the third empirical chapter of my dissertation examines the link between parent- child relationship quality and mothers’ mental health at midlife. The majority of existing work has found that intergenerational support, strain, and ambivalence are linked with older parents’ wellbeing. Much less research examines parents at midlife, despite the fact that parenthood at this life course stage can be particularly demanding as adolescent and emerging young adult

3 children are undergoing various life transitions, redefining their social relationships, which can be connected with heightened conflicts and changes in parent-child support. Additionally, current work has not sufficiently addressed how the relationship with multiple children is linked with midlife parents’ mental wellbeing. Using a sample of midlife mothers from NLSY79, this study examines how adolescent and young adult children’s reports of relationship quality with their mother, categorized by uniformly close, ambivalent, and uniformly unclose, are associated with mother’s mental health at age 50. Models from OLS regression with lagged dependent variables find that midlife mothers in an ambivalent relationship with adolescent and young adult children had the highest level of psychological distress among the three groups. Midlife mothers with at least one child reporting declining feelings toward mother were also set up for worse mental health over time. This chapter demonstrates the relational nature of family relationships and highlights the importance of incorporating accounts from multiple family members, with further implications for improving both midlife mother’s and family’s wellbeing.

In sum, my dissertation extends prior research on intergenerational relationship and adults’ wellbeing at midlife. The results elucidate that the link between parenthood and wellbeing extends beyond the core parenting years (i.e., parenting young children) and is manifested in both mental and physical health outcomes. The findings deepen our understanding about midlife parents, especially mothers, who experience increasing parenting demands and responsibilities to adolescent and coming of age children compared to their parents’ generation.

This study provides important policy implications to help parents and families in need — for example, increasing funding for education and allowing young adults to remain on parental

4 insurance until mid -20s may serve to prepare young adult children for a smoother launch and improve parents’ wellbeing in the meantime.

References

Eickmeyer, Kasey J. 2016. Over 25 Years of Change in Men's Entry into Fatherhood, 1987-

2013. Family Profiles, FP-16-10. Bowling Green, OH: National Center for Family &

Marriage Research. http://www.bgsu.edu/ncfmr/resources/data/family-

profiles/eickmeyer-fatherhood-25years-change-age-entry-fp-16-10.html

Fingerman, Karen L. 2017. "Millennials and Their Parents: Implications of the New Young

Adulthood for Midlife Adults." Innovation in Aging 1(3):1-16.Fingerman, Karen L., Yen-

Pi Cheng, Lauren Tighe, Kira S. Birditt and Steven Zarit. 2012. "Relationships between

young adults and their parents." Pp. 59-85 in Early adulthood in a family context, edited

by A. Booth, S.L. Brown, N.S. Landale, W.D. Manning and S.M. McHale. New York:

Springer Publishers.

Palloni, Alberto, Carolina Milesi, Robert G. White and Alyn Turner. 2009. "Early childhood

health, reproduction of economic inequalities and the persistence of health and mortality

differentials." Social Science & Medicine 68(9):1574-1582.

Read, Jen’nan G. and Bridget K. Gorman. 2011. "Gender and health revisited." Pp. 411-429 in

Handbook of the Sociology of Health, Illness, and Healing, edited by B. Pescosolido, J.

Martin, J. McLeod and A. Rogers. Springer, New York, NY.

Umberson, Debra, Tetyana Pudrovska and Corinne Reczek. 2010. "Parenthood, childlessness,

and well‐being: A life course perspective." Journal of Marriage and Family 72(3):612-

629.

Chapter 2. Midlife Parental Context and Mental Health

Parenthood at midlife (ages 40-60) is an important social context at a crucial life course stage for individual wellbeing, including mental health. Yet, extant research tends to focus on either the health consequences of parenting minor children during young adulthood (e.g., Musick, Meier,

& Flood, 2016) or elderly care and the parent-child relationship at parents’ later life (e.g.,

Sechrist, Suitor, Howard, & Pillemer, 2014). Midlife parenthood may be especially influential for mental wellbeing considering the unique heterogeneity of today’s midlife parents (i.e., the late baby boomers, born around late 1950s and early 1960s). Two recent demographic trends drive this midlife parenthood heterogeneity. First, childbearing postponement was increasingly prevalent among the late baby boomers (Eickmeyer et al., 2018), and thus individuals from this cohort are more likely to have younger children in their early midlife (40-50) than their own parents and grandparents (Pew Research Center, 2010; Eickmeyer et al., 2018). Second, for those midlife boomers who have young adult children, children’s delayed first departure from parental household means that these parents today are also more likely to house young adult children

(Pew, 2013, 2016). The heterogeneity of midlife parenthood is important for health and wellbeing, as research has shown that parenting minor children and adult children have different ramifications for parental wellbeing including mental health (Umberson, Pudrovska, & Reczek,

2010). Notably, the life course stage between age 40 and age 50 – the period from early midlife to mid midlife, is a time when both the family structure and health profiles undergo significant

6 changes (Fingerman, 2017; Lachman, 2004). However, prior research has seldom considered how midlife parenthood contexts and mental health differ for age 40 and age 50.

Notably, the experiences of mothers and fathers are starkly different, with women providing more coresidential care for both young and adult children across the life course

(Umberson et al., 2010). A gendered life course perspective suggests that parenthood imposes different constraints and demands on men and women; factors in turn shown to be linked with varying health outcomes including mental health (Umberson, Liu, Mirowsky, & Reczek, 2011).

In particular, women are more likely to report higher levels of psychological distress than men

(Bird, 2009); women’s mental health is also more sensitive to relationship with children because motherhood is more central to women’s identity (Carr, 2004; Milkie, Bierman, & Schieman,

2008). These gender differences suggest distinctive association between midlife parenthood and mental wellbeing for women and men. On the other hand, recent literature suggests a convergence in men’s and women’s social roles (Altintas & Sullivan, 2016; Pew Research

Center, 2013), and thus the relationship between midlife parenthood and wellbeing may not drastically differ by gender.

To address these gaps, this study uses nationally representative data of the American cohort of midlife late baby boomers to answer the following questions: (1) How does children’s life course stage and parent-child coresidence at midlife matter for individuals’ psychological wellbeing and how does this relationship differ between age 40 and age 50? (2) How do these patterns vary both across and within gender? This study contributes to existing literature by considering the heterogeneity of midlife parenthood among the late baby boomers. Further, this study also casts attention on the scenario of parenting both minor and adult children in the

7 household, one emerging parenthood context for late boomers at age 40. Using data from

National Longitudinal Survey of Youth-1979, I employ OLS regression models to examine the research questions. This study is innovative and can contribute to the family and health literature depicting the landscape as well as the mental health effects of midlife parenthood among a contemporary cohort. These endeavors will provide evidence to policy makers and help direct resources to midlife individuals who need additional support in face of the various challenges associated with parenthood, improving the family wellbeing as a whole.

Background

Parenthood context and psychological wellbeing at midlife

Unraveling the relationships between parenthood context and health at midlife is crucial because midlife is when chronic health problems begin to surface and population health disparities further diverge (Read & Gorman, 2011; Palloni et al. 2009). A life course perspective suggests the heterogeneous parenthood context—particularly in terms of children’s life course stage and parent-child coresidence— is very intricately linked with individuals’ midlife mental health (Umberson et al., 2010). Furthermore, women’s mental health is shown to be more responsive to the relationship with children than men (Milkie et al., 2008; Reczek & Zhang,

2016), and thus the effects of children’s life course stage and parent-child coresidence on midlife adults’ mental health likely differ for men and women (Umberson et al., 2010). In the following sections, I will review and analyze current literature on how five parenthood contexts, including

(1) coresident minor children, (2) coresident adult children, (3) both minor and adult children in the household, (4) the childless, and (5) parents with no coresident children matter for midlife parents’ mental health, and how these relationships differ by gender.

(1) Minor children and midlife parental mental health. The average age at first birth increased significantly over the past four decades, wherein late baby boomers are more likely to have younger children at their 40s than their parents’ generation. Within this cohort, men are more likely to have younger children than women at age 40, as men tend to have children at old ages than women (Eickmeyer, 2016). Parenting young children is strongly associated with individual’s mental health. Parents with minor children generally report lower levels of psychological wellbeing than both the childless and those with adult children (Evenson & Simon,

2005). Higher demands of parenting responsibilities for minor children accompanied by greater levels of strain, stress, and physical fatigue are among the main contributing factors of this finding (Kluwer & Johnson, 2007; Pollmann-Schult, 2014). However, recent evidence complicates the association between parenting minor children and mental health. For instance, some studies show that parenting minor children is associated with more positive subjective wellbeing including higher levels of parental satisfaction and happiness, lower levels of stress, and higher likelihood to perceive life as meaningful and joyful, compared to the childless, or compared to times when parents are not engaging with young children (Blair-Loy, 2003; Musick et al., 2016; Nomaguchi, 2012), especially when demographic covariates such as marital status are accounted for (Umberson, Thomeer, & Williams, 2012). Regardless of the association between parenting minor children and mental wellbeing, research consistently shows that mothers shoulder more childrearing responsibilities and experience more stress and greater fatigue while caring for young children than fathers (Umberson et al., 2010; Musick et al., 2016).

Notably, the majority of research on minor children and parental mental wellbeing, including those cited above, situates in the context of parent-minor child coresidence. In fact,

American parents, especially mothers, care for young children in the household until they leave the nest for higher education, job, and family formation (Furstenburg, 2010). Parental coresidence with minor children is not only the dominant childrearing living arrangement in the

U.S., it is directly linked with everyday hassles, stress, and rewards that are connected with individuals’ mental health (Umberson et al., 2010; Musick et al., 2016). Taken together, coresidence with minor children can be linked with lower levels of mental wellbeing among parents, especially mothers. This association may be attenuated once the sociodemographic covariates are considered.

(2) Young adult children and midlife parental mental health. Midlife parents maintain frequent contact with and offer continued emotional, financial, and instrumental support to adult children (Fingerman et al., 2012), and parents’ mental health is positively associated with a supportive relationship with adult children, and negatively associated with a strained relationship with adult children (Fingerman, 2017; Umberson et al., 2010). Additionally, an integration of the life course perspective and stress process theory (Milkie et al., 2008; Umberson & Reczek, 2007) suggests that midlife parents’ mental wellbeing is linked with young adult children through a process of intergenerational stress proliferation, wherein young adults’ stressors spread through the family network and increase parental distress. For instance, parents’ psychological distress increases when parents perceive young adults need too much help during the transition to adulthood (Fingerman et al., 2012). Notably, because mothers have a closer relationship with adult children, maintain more frequent contact, offer more emotional and instrumental support, and report greater closeness as well as conflict with/to adult children than fathers, midlife

10 mothers’ mental wellbeing are also more sensitive to the relationship with adult children and adult children’s stressors (e.g., Arnett & Schwab, 2012; Fingerman et al., 2016).

Young adult children increasingly postpone their first departure from home, wherein the proportion of young adults living in parental household has increased for the past four decades

(Pew Research Center, 2015a). Consequently, living with parents has become the most common living arrangement among the 18-to-34-year-old (Pew Research Center, 2016). Because women tend to have earlier first birth and shoulder more parenting responsibilities than men, more young adult children live with mothers than fathers. Research using data from earlier cohorts generally finds that coresidence with emerging young adult children (e.g., post-college-age adult children) is associated with higher psychological distress among parents (Pudrovska, 2009). However, midlife parents report increased acceptance of young adult children’s extended stay after the mid-2010s. For example, a vignette study conducted prior to the great recession (late 2000s to early 2010s) finds that adults of all ages do not appear to prefer grown children’s extensive coresidence with parents unless the younger generation encounters severe economy deficiency or was single or childless (Seltzer, Lau, & Bianchi, 2012). Similarly, recent research shows that young adults’ coresidence with parents is only associated with worse marital quality for parents only if the coresidence happened before the Great Recession (Davis et al., 2016). Because the great recession happened when the late baby boomers were in their early 40s, coresidence with adult children is likely to be associated with decreased mental wellbeing among parents, especially mothers.

(3) Both minor children and adult children and parental mental health. Virtually no research has examined how parenting both coresidential minors and adult children, an emerging

11 parenthood context at age 40 for the late boomers, is associated with mental health.

Psychological theories of adult development describe midlife as a life course stage of self- assessment and evaluation (DeVries et al., 2007). This perspective posits that midlife wellbeing largely hinges on individuals’ ability to handle multiple roles (Ann & Cooney, 2006). In the case of parenthood, midlife individuals who have both coresidential adult children and minor children are dealing with different parenting roles on a daily basis (e.g., more physical care, transportation care for minor children, more emotional care for adult children maybe). While balancing the parenting roles of both adult and young children, midlife parents, especially mothers, are thus more prone to experience more psychological distress than parents in parenthood contexts.

On the other hand, coresidence with both minor and adult children may be no more stressful than coresidence with minor children only. As an aphorism suggests – “A mother is only as happy as her least happy child,” the strain of parenting both minor and adult children may be more similar to that of the strain of parenting coresident minor children only, which generally involves more work and causes more psychological distress than parenting adult children (Evenson & Simon, 2005). Additionally, a parent that could have been bothered by adult children’s delayed departure may direct their primary attention to the minor children and thus not become overly stressed by the stalled progress in adult children’s transition. Prior work has shown that factors including parents’ work-family stress and being an older child are positively associated with children’s housework time (Gager, Sanchez, & Demaris, 2009). In line with this research, adult children in the household may contribute more to the household at the presence of younger siblings, relieving some parental stress. In this case, parents to both

12 coresident minor and adult children may have similar mental health as those with coresidential minor children only.

(4) Childlessness and mental health at midlife. To better understand the effects of parenting coresidential children, I consider individuals with no coresidential children in the analysis for comparison, including the childless and parents with no coresidential children while maintaining the differentiation between these two categories. These two groups are used to examine how the lack of parenting stress for minor children and adult children, described above, can relate to midlife adults’ mental wellbeing. The inclusion of these two groups along with other groups of parents can serve as the first step to provide a comprehensive assessment of the relationship between parenthood and mental health, with an emphasis on children’s coresidence and life course stages. Even though the childless and parents with non-residential children may not experience the daily stress of caring for coresident children, they may have other stressors in life that affect their mental wellbeing.

The rate of childlessness has increased for the past three decades, with men being more likely to remain childless than women at age 40 (Eickmeyer, 2016). In general, recent research using national American data of older cohorts (e.g., Health and Retirement Survey) finds that compared to parents, childless people in mid- to later-life displayed lower levels of depression

(Bures, Koropeckyj-Cox, & Loree, 2009). This finding is corroborated by various studies, especially among women at midlife, because women, especially the never married, tend to be highly educated and more socially active (Wenger, Dykstra, Melkas, & Knipscheer, 2007;

Koropeckyj-Cox & Call, 2007). On the other hand, childless men tend to have lower SES and thus they are more likely to have worse mental health than fathers (Dykstra & Hagestad, 2007).

Additionally, fatherhood is far less demanding, and often perceived as more fun than motherhood because fathers spend more time doing interactive activities with children while mothers do more physical chores and onerous tasks (Musick et al., 2016). As a result, the positive effect of fatherhood on wellbeing is evident even among the most disadvantaged group of men (Edin & Nelson, 2013). A negative association between childlessness and mental health is thus likely present among men, but not women.

(5) Parents with no coresidential children and mental health at midlife. The share of children growing up in two-parent families has been declining since the 1960s (Pew Research

Center, 2015b). Meanwhile, the proportion of non-residential parents has been increasing, especially among fathers. For example, a pew research report based on national data collected in

2006-2008 shows that about 27% of men with children 18 or younger were living apart from at least one of their children, and 16% of them were residing with none of their children (Pew

Research Center, 2011). Notably, the majority of parents with no residential children during early midlife (age 40-50) may not be empty nest parents, based on the demographic statistics of age at first birth and parity (Eickmeyer et al., 2018). Individuals with lower SES are more likely to be non-residential parents and those with no coresidential children at all are particularly disadvantaged (Pew Research Center, 2011; Anderson, Kohler, & Letiecq, 2005). As a result, parents with no coresidential children may experience more economic stress and thus have higher mental distress than others. Additionally, coparenting stress including conflicts with the custodial parent and logistic strains may set this group of individuals up for higher levels of psychological distress (Anderson, et al., 2005; Jayakody & Seefeldt, 2005). Further, the negative

14 association between non-residential parenthood and mental wellbeing may be more prominent among women due to the salience of mother’s identity to women (Carr, 2004).

Taken together, the existing literature overlooks the complexity of midlife parenthood among the late baby boomers and can mask important distinctions in mental health by specific parenthood contexts including parenting coresidential minor children only, caring for both minor and adult coresidential children, parenting only adult children in the household, the childless, and parents with no coresidential children. This study aims to amend this research gap by addressing two questions: (1) How does children’s life course stage and parent-child coresidence at midlife matter for individuals’ psychological wellbeing? (2) How does parental gender matter in these relationships? This study further considers how these relationships differ between age 40 and age

50. In addition, the role of gender is examined in analysis using between-gender comparisons as well as within-gender comparisons.

Methods

Data

Data are from the National Longitudinal Survey of Youth-1979, which includes a nationally representative sample of 12,686 men and women aged between 14 and 22 in 1979.

Respondents have been interviewed annually since 1979 and biennially since 1994. The data collection is ongoing and the most recent wave available is 2014. Beginning in 1998, NLSY79 launched a 40-and over health module to collect a wealth of information on health, including psychological wellbeing. The health module was administered to individuals once, after they turned age 40. The survey years of the 40’s health module comprise 1998 through 2006 with a two-year interval. Beginning in 2008, another health module was administered to individuals after they turned age 50. Each respondent would also participate in the 50’s health module once, 15 the survey years of which spanned between 2008 and 2014. This study employed data through

2008, using information from both the 40’s and 50’s health modules to address the research questions.

I first limited the analytical sample to 8,510 men and women who either remained childless or had valid information on both the age and residential status of their children. Men are more likely than women to report missing fertility information than women. 39 respondents were further excluded due to missing information on mental health. The final analytical sample is thus composed of 8,471 men and women who either remained childless or had complete information on the residential status and life course stage of their children, all of these individuals also had valid mental health responses. More specifically, this study contains two analytical samples, corresponding to the two analysis described further in the section of analytical strategy. The first sample includes men and women at age 40 (n=8,131). The second sample comprises men and women at age 50 (n=7,324).

Measures

Psychological Distress captures symptoms of anxiety and depression (Mirowsky & Ross,

1989). In this study, psychological distress was measured using a 7-item version of Center for

Epidemiological Studies Depression (CES-D) scale. The CES-D scale was based on answers from the following questions: “how often do you felt that way during the past week”: (1) I did not feel like eating; my appetite was poor; (2) I had trouble keeping my mind on what I was doing; (3) I felt depressed; (4) I felt that everything I did was an effort; (5) My sleep was restless;(6) I felt sad; (7) I could not get “going”. For each question, respondents selected answers from rarely or none of the time, some or little of the time, occasional or a moderate

16 amount of the time, and most of all of the time. NLSY researchers combined answers from the seven questions into a single scale, which I log transformed to address the positive skew. The transformed scores were then multiplied by 100 for the convenience of interpretation (see Clarke et al., 2011 for a similar approach). A higher CESD score denotes higher levels of psychological distress. Alpha reliabilities were 0.83 at both age 40 and age 50.

Composition of coresidential children at midlife, which captures the family structure in terms of children’s life course stage and parent-child coresidence, is the main independent variable. It comprises five categories: presence of coresidential minor children only, presence of both coresidential minor and adult children, presence of adult children only, the childless, and parents who have no coresidential children. Children’s age and coresidential status were identified from respondents’ answers on the birth year and usual residence of each of their children. Other parenthood-related variables are also included so as to provide a comprehensive picture of parenthood context at midlife. Number of children in the household was derived from respondents’ self-reported number of children in the household; it was top-coded at four due to much fewer responses beyond four children. Two dummy variables, presence of non-residential minor child and presence of non-residential adult child were also based on answers on the birth year and usual residence of each of the children.

Covariates that are shown to be associated with both midlife parenthood context and mental wellbeing are included in multivariate analysis. The covariates include race/ethnicity

(Non-Hispanic White, Non-Hispanic Black, Hispanic), highest grade completed, marital status

(Never married, married, other), employment status (employed, other), family income (adjusted for inflation and reported in 2014 dollars, and log transformed to account for positive skew of the

17 original distribution), family size (top coded at six), health insurance coverage (private, public, self-pay), and self-rated health (fair/poor, good, very good, excellent). Because early transition to parenthood is linked with both parenthood contexts at midlife and mental wellbeing (Aitken et al., 2016), I also generated a dummy variable to account for the experience of early childbearing

(i.e., whether one had a first birth prior to age 20).

Analytic Strategy

I ran three sets of models to test the association between midlife parenthood context and individuals’ psychological distress. The first set of tests employed OLS models and examined how children’s life course stage and parent-child coresidence matters for individuals’ mental health at age 40. The second set of tests used OLS models among those who reached age 50.

These two sets of models aim to shed light on the between-individual differences in the association between midlife parenthood and mental wellbeing at age 40 and age 50 separately.

Both of these models were run on the age-specific subsamples first and then on men and women separately.

To maximize the data usage, I employed multiple imputation methods with chained equations to deal with missing values on covariates. Most of the covariates had missing values at a suggested cut point of 5% (Schafer, 1999), except family income, for which about 15% of responses were invalid. Sensitivity analysis that excluded family income but contained other highly-correlated SES variables such as health insurance type and highest achieved education showed qualitatively similar results as those presented. Conditional distribution for missing values on all variables was generated by Gibbs sampling techniques (Royston 2005, Van Buuren

2012). Five distinct data sets were produced because 3 to 4 datasets are deemed appropriate for models with up to 20% missing (Royston 2005, Van Buuren 2012). Dependent variables with 18 missing values were included in the imputation model and later deleted for the analysis (Van

Hippel, 2007). All the descriptive and multivariate analyses are conducted using the mi command in Stata/MP 15.0 (College Station, TX: StataCorp LP).

Results

Descriptive Results

Table 1 presents gender-specific descriptive information of all the analyzed variables for samples of the age 40, age 50. All of the descriptive results are stratified by gender. Significance tests conducted to examine gender differences are also shown in Table 1. According to Table 1, at both age 40 and 50, men had higher levels of psychological distress than women (Age 40:

164.32 VS 192.95, p < 0.001; Age 50: 174.44 VS 209.15, p < 0.001). At age 40, 24% of men remained childless, the majority of men – about 43% had only minor children in the household, and 23% of men had children but had no coresidential children at the time. Only 7% of men had both minor and adult children in the household and a very small proportion of them, 3% lived with coresidential adult children only. Women had different parenthood contexts in terms of children’s age and coresidence at age 40. Compared to men at the same age, the 40-year-old women were less likely to remain childless (16%) and more likely to live with children as only

9% of women did not have coresidential children at home. The proportion of women living with minor and adult children, and women living with adult children only, were significantly greater than that of their male counterparts (18% VS 7%, p < 0.001; 10% VS 3%, p < 0.001).

Compared to age 40, parenthood contexts at age 50 changed for both men and women: the percentage of coresidence with minor children only decreased, whereas the proportion of coresidence with adult children and the proportion of parents with no coresidential children

19 increased. Substantial gender differences persisted, as women at age 50 continued to have higher likelihood of living with adult children than men at the same age (29% VS 16%, p < 0.001).

Multivariate Results

Table 2 presents results from the OLS regression models examining the association between midlife parenthood contexts and mental health at age 40. Results from Model 1 with both men and women show that compared to parents with both residential minor and adult children, the childless (b = 12.77, p < 0.05), parents with residential minor children only (b =

8.66, p < 0.05), and parents with adult children only (b = 15.69, p < 0.01) all had more psychological distress at age 40, adjusting for the covariates. Additional analysis that used different reference groups did not reveal differences among other comparison groups. At age 40, women had substantially more psychological distress than men (b = 22.97, p < 0.001). Yet, analysis including the interaction term between gender and parenthood context (see Table 9 in

Appendix) finds that the general association between parenthood context and mental health did not vary by gender. Predicted values generated by Model 1 were plotted in Figure 1. According to Figure 1, individuals with both minor and adult children in the household had the lowest level of psychological distress, which was significantly lower than all the other parenthood contexts except parents with no residential children, likely because the latter had a wider confidence interval; women’s levels of psychological distress were higher than men’s in every parenthood context and in similar magnitude.

Model 2 and 3 in Table 2 shows OLS regression results estimating the relationship between midlife parenthood contexts and mental health at age 40 for men and women respectively. Model 2 shows that men living with minor children only had more psychological

20 distress than men with both coresidential minor and adult children (b = 13.69, p < 0.05), accounting for other sociodemographic covariates. Additional differences among other comparison groups were not found. Among women, those coresiding with adult children only reported more psychological distress than those living with both minor and adult children (b =

17.90, p < 0.05), adjusting for other covariates. Supplementary analysis shows that mothers living with adult children only also had higher levels of psychological distress than those with coresidential minor children (b = 14.36, p < 0.05).

Table 3 displays OLS regression results that examined the association between parenthood contexts and mental health at age 50. Midlife parenthood contexts do not appear to explain the variation in CESD scores at age 50 among the full sample, men, and women, respectively. In sum, the OLS results indicate between-individual differences in mental health by parental contexts at age 40 but not at age 50.

Discussion

Parenthood is an important social context that is intricately linked with mental health across the life course. At midlife, children’s life course stage and coresidence have significantly changed among the late baby boomers compared to the past two generations. Yet, previous research fails to consider the heterogeneity of midlife parenthood contexts in relation to mental wellbeing among the late baby boomers. Guided by a gender life course perspective and using nationally representative data of NLSY79, this study fills this research gap by examining the association between parenthood context in terms of parent-child coresidence and children’s life course stage and midlife adults’ mental health at midlife. I further study how this association differs for men and women. This study has three main contributions to the literature, described in detail below. 21

First, among all the parenthood contexts in terms of children’s life course stage and parent-child coreidence, results from the OLS regression models suggest that coresidence with adult children only is linked with elevated psychological distress. Young adults’ extensive coresidence is shown to violate the normative expectation of children leaving the nest after they become adults (Settersten, 1998). Coresidence with adult children, a behavior that likely violates age norm is subsequently linked with increased stress, which can proliferate through family network and influence parental wellbeing (Millkie et al., 2008). Furthermore, the connection between specific age norm and other family members’ wellbeing are situated within a particular historical context. This particular group of parents experienced their early midlife (early 40s) prior to the great recession, a time when young adults’ delayed departure from home was frowned upon (Davis et al., 2016). As a result, findings of this study aligned with results gleaned from older cohorts (e.g., Pudrovska, 2009) and show that coresidence with adult children can increase parents’ psychological distress.

Second, this study demonstrates that coresidence with both minor and children appear to have protective effects on mental health among middle-aged individuals relative to other coresidential patterns. For example, results from the OLS models for the 40-year-olds show that the scenario of living with both minor and adult children is associated with the lowest level of psychological distress among all the parenthood contexts including the childless at age 40. This finding corresponds to recent research, which finds that parents seek meaning and happiness from caring for young children and reap benefits for their mental wellbeing as a result (Blair-

Loy, 2003; Musick et al., 2016). Yet, what needs further explanation is why coresidence with minor children as well as adult children also yields positive mental health. It is possible that

22 residential minor children capture parents’ primary attention; the mental health of those with both minor and adult children in the household was thus more similar to those with coresidential minor children only. Additionally, adult children in the household might be able to help with some household chores or care for their younger siblings (Gager et al., 2009); in this case, parents may not be negatively affected by adult children’s late departure as is found when only adult children remain in the home. This finding suggests that scholars need to study a broader range family dynamics in order to develop a better understanding of the determinants of parental wellbeing, or wellbeing of other family members. This is because the relationship between two family members and the subsequent health effects may be modified by the presence of another family member.

Lastly, although women had higher levels of psychological distress than men across all parenthood contexts, the relationship between coresdiential parenthood contexts and mental health did not vary by gender. Nevertheless, many significant differences found in the main model with both men and women included were largely driven by results in the gender-specific models. This is highly related to the fact that women are more likely to experience some parenthood contexts, whereas other types of parenthood are largely experienced by men. For instance, descriptive results show that women were much more likely than men to live with adult children, as women transition to parenthood at a younger age than men. Women also do provide more residential support to adult children (Evenson & Simon, 2005). Consequently, the negative effects of parenting coresidential adult children on mental health found in the full sample, manifested in both the OLS results at age 40 and the fixed effects models, are largely driven by significant results produced by the sample of women. Overall, this study advances current

23 research, showing that it is primarily women who care for adult children who need more support in the changing economics and take the accompanied emotional toll.

This study is not without limitation. First, selection bias may be at play. Individuals from lower socioeconomic status, who have worse health outcomes than those with more socioeconomic resources across the life course (Link & Phelan, 2010), tend to have earlier transition to parenthood and thus are more likely to have adult children at age 40. This study addresses this potential bias by including a comprehensive list of socioeconomic variables andcontrolling for early first birth,. However, other selection factors may still influence the results. Second, the drastic differences in the relationship between parenthood and mental health between age 40 and age 50 are under-investigated. Future research should direct more attention to the changing family dynamics and health consequences during this specific life course stage of early midlife.

Despite the limitations, this study addresses the heterogeneity of midlife parenthood among the late baby boomers and identifies specific groups of middle-aged individuals – mothers to the coresidential adult children, might be at risk of e psychological distress. The analysis also suggests potential mental health benefits of raising minor children in midlife and reveals the complex effects of parenting both minor and adult children under the same roof. This study advances current literature in understanding the relationship between parenthood and wellbeing, paying particular attention to the structural positions of gender, parent-child coresidence, children’s life course stages, and presence of multiple children. The findings of the study provide potentially important policy recommendations to support parents and families in need.

References

Allison, Paul D. 1994. "Using panel data to estimate the effects of events." Sociological Methods

& Research 23(2):174-199.

Altintas, Evrim and Oriel Sullivan. 2016. "Fifty years of change updated: Cross-national gender

convergence in housework." Demographic Research 35:455-470.

Anderson, Elaine A., Julie K. Kohler and Bethany L. Letiecq. 2005. "Predictors of depression

among low-income, nonresidential fathers." Journal of Family Issues 26(5):547-567.

Arnett, Jeffrey J. and Joseph Schwab. 2012. The Clark University poll of emerging adults:

Thriving, struggling, and hopeful. Worcester, MA: Clark University.

Blair-Loy, Mary. 2009. Competing devotions: Career and family among women

executives. Harvard University Press.

Bures, Regina M., Tanya Koropeckyj-Cox and Michael Loree. 2009. "Childlessness, parenthood,

and depressive symptoms among middle-aged and older adults." Journal of Family

Issues30(5):670-687.

Carr, Deborah. 2004. "“My Daughter Has a Career; I Just Raised Babies”: The Psychological

Consequences of Women's Intergenerational Social Comparisons." Social Psychology

Quarterly 67(2):132-154.

DeVries, Helen M., Sarah Kerrick and Megan Oetinger. 2007. "Satisfactions and regrets of

midlife parents: A qualitative analysis." Journal of Adult Development 14(1-2):6-15.

Dykstra, Pearl A. and Gunhild O. Hagestad. 2007. "Roads less taken: Developing a nuanced

view of older adults without children." Journal of Family Issues 28(10):1275-1310.

Edin, Kathryn and Timothy J. Nelson. 2013. Doing the best I can: Fatherhood in the inner

city. University of California Press. 25

Eickmeyer, Kasey J. 2016. Over 25 Years of Change in Men's Entry into Fatherhood, 1987-

2013. Family Profiles, FP-16-10. Bowling Green, OH: National Center for Family &

Marriage Research. http://www.bgsu.edu/ncfmr/resources/data/family-profiles/eickmeyer-

fatherhood-25years-change-age-entry-fp-16-10.html

Eickmeyer, Kasey J., Payne, Krista K., Brown, Susan L., and Manning, Wendy D. (2018).

Crossover in the median age at first marriage and first birth: Thirty-five years of change.

Family Profiles, FP-17-22. Bowling Green, OH: National Center for Family & Marriage

Research. https://doi.org/10.25035/ncfmr/fp-17-22

Evenson, Ranae J. and Robin W. Simon. 2005. "Clarifying the relationship between parenthood

and depression." Journal of Health and Social Behavior 46(4):341-358.

Fingerman, Karen L., Kyungmin Kim, Kira S. Birditt and Steven H. Zarit. 2016. "The Ties That

Bind: Midlife Parents' Daily Experiences With Grown Children." Journal of Marriage and

Family 78:431-450.

Fingerman, Karen L. 2017. "Millennials and Their Parents: Implications of the New Young

Adulthood for Midlife Adults." Innovation in Aging 1(3):1-16

Fingerman, Karen L., Yen-Pi Cheng, Lauren Tighe, Kira S. Birditt and Steven Zarit. 2012.

"Relationships between young adults and their parents." Pp. 59-85 in Early adulthood in a

family context, edited by A. Booth, S.L. Brown, N.S. Landale, W.D. Manning and S.M.

McHale. New York: Springer Publishers.

Furstenberg Jr, Frank F. 2010. "On a new schedule: Transitions to adulthood and family

change." The Future of Children 20(1):67-87.

Gager, Constance T., Laura A. Sanchez and Alfred Demaris. 2009. "Whose time is it? The effect

of employment and work/family stress on children’s housework." Journal of Family

Issues30(11):1459-1485.

Jayakody, Rukmalie and Kristin S. Seefeldt. 2005. "Families across households: Who gets

counted and the implications for family functioning.".

Johnson, David. 2005. "Two‐wave panel analysis: Comparing statistical methods for studying

the effects of transitions." Journal of Marriage and Family 67(4):1061-1075.

Kluwer, Esther S. and Matthew D. Johnson. 2007. "Conflict frequency and relationship quality

across the transition to parenthood." Journal of Marriage and Family 69(5):1089-1106.

Koropeckyj-Cox, Tanya and Vaughn R. Call. 2007. "Characteristics of older childless persons

and parents: Cross-national comparisons." Journal of Family Issues 28(10):1362-1414.

Lachman, Margie E. 2004. "Development in midlife." Annual Review of Psychology. 55:305-

331.

Milkie, Melissa A., Alex Bierman and Scott Schieman. 2008. "How adult children influence

older parents' mental health: Integrating stress-process and life-course perspectives." Social

Psychology Quarterly 71(1):86-105.

Mirowsky, John and Catherine E. Ross. 1989. "Psychiatric diagnosis as reified

measurement." Journal of Health and Social Behavior:11-25.

Musick, Kelly, Ann Meier and Sarah Flood. 2016. "How parents fare: Mothers’ and fathers’

subjective well-being in time with children." American Sociological Review 81(5):1069-

1095.

Nomaguchi, Kei M. 2012. "Parenthood and psychological well-being: Clarifying the role of child

age and parent–child relationship quality." Social Science Research 41(2):489-498.

Palloni, Alberto, Carolina Milesi, Robert G. White and Alyn Turner. 2009. "Early childhood

health, reproduction of economic inequalities and the persistence of health and mortality

differentials." Social Science & Medicine 68(9):1574-1582.

Pew Research Center. 2010. “The New Demography of American Motherhood.” Retrieved from

http://assets.pewresearch.org/wp-content/uploads/sites/3/2010/10/754-new-demography-of-

motherhood.pdf

Pew Research Center. 2011. “A Tale of Two Fathers More Are Active, but More Are Absent.”

Retrieved from http://assets.pewresearch.org/wp-content/uploads/sites/3/2011/06/fathers-

FINAL-report.pdf

Pew Research Center. 2013. “Modern Parenthood Roles of Moms and Dads Converge as They

Balance Work and Family.” Retrieved from

http://www.pewsocialtrends.org/files/2013/03/FINAL_modern_parenthood_03-2013.pdf

Pew Research Center. 2015a. More millennials living with family despite improved job market.

Washington, DC: Pew Research Center. Retrieved from

http://www.pewsocialtrends.org/2015/07/29/more-millennials-living-with-family-despite-

improved-job-market/

Pew Research Center, 2015b. “Parenting in America: Outlook, worries, aspirations are strongly

linked to financial situation.” Retrieved from

http://www.pewsocialtrends.org/files/2015/12/2015-12-17_parenting-in-

america_FINAL.pdf

Pew Research Center. 2016. For the first time in modern era, living with parents edges out other

living arrangements for 18- to 34-year-olds. Washington, DC: Pew Research Center.

Retrieved from http://www.pewsocialtrends.org/2016/05/24/for-first-time-in-modernera-

living-with-parents-edges-out-other-living-arrangementsfor-18-to-34-year-olds/

Phelan, Jo C., Bruce G. Link and Parisa Tehranifar. 2010. "Social conditions as fundamental

causes of health inequalities: theory, evidence, and policy implications." Journal of Health

and Social Behavior 51(1_suppl):S28-S40.

Pollmann‐Schult, Matthias. 2014. "Parenthood and Life Satisfaction: Why Don't Children Make

People Happy?" Journal of Marriage and Family 76(2):319-336.

Pudrovska, Tetyana. 2009. "Parenthood, stress, and mental health in late midlife and early old

age." The International Journal of Aging and Human Development 68(2):127-147.

Read, Jen’nan G. and Bridget K. Gorman. 2011. "Gender and health revisited." Pp. 411-429

in Handbook of the sociology of health, illness, and healing" Gender and health

revisited."Springer.

Reczek, Corinne and Zhe Zhang. 2016. "Parent–Child Relationships and Parent Psychological

Distress: How Do Social Support, Strain, Dissatisfaction, and Equity Matter?" Research on

Aging 38(7):742-766.

Royston, Patrick. 2005. "Multiple imputation of missing values: update of ice." Stata

Journal 5(4):527.

Schafer, Joseph L. 1999. "Multiple imputation: a primer." Statistical Methods in Medical

Research 8(1):3-15.

Sechrist, Jori, J. Jill Suitor, Abigail R. Howard and Karl Pillemer. 2014. "Perceptions of equity,

balance of support exchange, and mother–adult child relations." Journal of Marriage and

Family 76(2):285-299.

Seltzer, Judith A., Charles Q. Lau and Suzanne M. Bianchi. 2012. "Doubling up when times are

tough: A study of obligations to share a home in response to economic hardship." Social

Science Research 41(5):1307-1319.

Settersten Jr, Richard A. 1998. "A time to leave home and a time never to return? Age

constraints on the living arrangements of young adults." Social Forces 76(4):1373-1400.

Umberson, Debra, Hui Liu, John Mirowsky and Corinne Reczek. 2011. "Parenthood and

trajectories of change in body weight over the life course." Social Science &

Medicine 73(9):1323-1331.

Umberson, Debra, Tetyana Pudrovska and Corinne Reczek. 2010. "Parenthood, childlessness,

and well‐being: A life course perspective." Journal of Marriage and Family 72(3):612-629.

Umberson, Debra and Corinne Reczek. 2007. "Interactive stress and coping around parenting:

Explaining trajectories of change in intimate relationships over the life course." Advances in

Life Course Research 12:87-121.

Umberson, Debra, Mieke B. Thomeer and Kristi Williams. 2013. "Family status and mental

health: Recent advances and future directions." Pp. 405-431 in Handbook of the sociology of

mental health"Family status and mental health: Recent advances and future

directions."Springer.

Van Buuren, Stef. 2012. Flexible imputation of missing data. CRC press.

Wenger, G. C., Pearl A. Dykstra, Tuula Melkas and Kees C. Knipscheer. 2007. "Social

embeddedness and late-life parenthood: Community activity, close ties, and support

networks."Journal of Family Issues 28(11):1419-1456.

Table 1. Sociodemographic and Health Characteristics of the Sample (Unweighted, Imputed) Age 40 Men Age 40 Women Age 50 Men Age 50 Women Mean or Mean or Mean or Mean or Percent SD Percent SD Percent SD Percent SD CESD score 164.32 109.32 192.95*** 112.27 174.44 111.28 209.15*** 110.41 Presence of coresidential children/by age Childless 24.44 16.20*** 23.07 15.87+ Only minor kids (ref) 42.76 46.97 17.17 13.59 Minor kids + Adult kids 6.99 17.88*** 8.02 9.35*** Only adult kids 2.74 10.10*** 16.04 29.30***

No children at home 23.06 8.85*** 35.70 31.89***

Non-residential minor children 24.19 4.94*** 8.77 1.09***

Non-residential YA child(ren) 20.29 27.96*** 42.81 61.29***

# children in the household (top-coded at 4) 1.18 1.25 1.56*** 1.18 0.82 1.05 0.91*** 1.01

Race/Ethnicity NonHispanic White (ref) 50.94 50.02 51.30 19.38 NonHispanic Black 29.31 30.61 29.59 31.27+ 19.75 Hispanic 19.37 19.11 49.35

Marital Status 21.28 Never Married 15.95*** 18.59 14.03** 59.06 Married or Cohabiting (ref) 58.23 58.04 53.98 19.66 Previously married 25.83*** 23.37 31.99*** Highest Grade Completed 13.21 2.57 13.54*** 2.60 13.27 2.57 13.56*** 2.57 Employment Status Employed (ref-not employed) 88.70 82.28*** 82.74 76.82***

11.23 Logged family income-(inflation adjusted) 11.17*** 11.24 11.13*** Family size 2.04 1.56 3.34*** 1.42 2.61 1.38 2.70** 1.33 Health Insurance 32

72.31 Private (ref) 74.81 66.60 69.68 Public 5.51 8.51*** 16.06 15.81 Self-pay 22.18 16.69*** 17.34 14.51** Self-rated health Fair/Poor (ref) 11.02 14.50 17.70 21.76 25.97 29.55+ Good 30.05 32.18* 40.36 36.01*** Very good 36.46 33.42*** 22.65 19.94*** Excellent 15.78 12.64*** 14.28 34.11*** First birth at or before age 20 14.08 34.21*** N (Person) 3,903 4,228 3,454 3,870 Source: NLSY79 and 79 Young Adults Survey (1994-2014). Note: The bivariate differences bewteen men and women are included in the table. + p<.10 * p<.05 ** p<.01 *** p<.001

Table 2. Results from OLS Regression Models Predicting Psychological Distress at Age 40 Model 1 - Full Sample Model 2 - Men Model 3 - Women b SE b SE b SE Presence of coresidential children/by age Childless 12.77* 6.12 16.98 + 9.42 6.12 8.36 Only minor kids 8.66* 4.07 13.69* 6.95 3.55 5.19 Minor kids + Adult kids Ref Ref Ref Only adult kids 15.69** 5.88 2.37 11.88 17.90* 6.98 No children at home 9.77 6.46 15.37 9.77 6.04 9.04

Non-residential minor children 4.28 4.20 0.1 5 5.20 15.30 + 7.9 1 Non-residential YA child(ren) -0.78 3.96 -6.59 5.73 2.49 5.50 # children in the household -3.01 2.32 -1.23 3.17 -3.00 3.46

Gender (ref: men) 22.97*** 2.54 Race/Ethnicity (ref: NonHispanic White) NonHispanic Black -7.35 ** 2.85 -5.10 4.06 -9.19 * 4.0 7 Hispanic -16.84*** 3.11 -12.07** 4.40 -21.25*** 4.41 Marital Status (ref: Married) Never Married 4.4 6 3.9 8 11.1 2* 5.6 5 -2.88 5.7 6 Previously married 13.74*** 3.58 17.38** 5.50 10.25* 4.93 Highest Grade Completed -1.14* 0.52 -2.35** 0.73 -0.10 0.74 Employment Status (ref-not employed) Employed -7.58 * 3.70 -20.7 1** 6.3 6 -1.8 6 4.6 6 Logged family income -13.63*** 2.40 -8.76* 3.44 -16.57*** 3.55 Family size 3.65* 1.72 4.06+ 2.21 2.03 2.71 Health Insurance (ref: Private) Public 34.72 *** 5.4 5 24.24 * 9.4 6 39.44 *** 7.0 7 Self-pay 13.45*** 3.36 20.59*** 4.68 4.13 4.94 Self-rated health (ref: Fair/Poor) Good -50.83 *** 3.96 -58.2 1*** 6.0 5 -46.2 6*** 5.2 9 Very good -78.62*** 3.91 -85.60*** 5.88 -73.48*** 5.29 Excellent -101.51*** 4.29 -107.21*** 6.35 -97.85*** 5.90 First birth at or before age 20 6.65 4.08 7.73 6.19 5.80 5.58 Intercept 385.18*** 26.81 353.90*** 38.34 430.65*** 39.30 N (NLSY79) 8,131 3,903 4,228 + p<.10 * p<.05 ** p<.01 *** p<.001 Source: NLSY79 and 79 Young Adults Survey.

Table 3. Results from OLS Regression Models Predicting Psychological Distress at Age 50 Model 1-Full Sample Model 2-Men Model 3-Women b SE b SE b SE Presence of coresidential children/by age Childless -7.09 7.00 -10.42 10.19 -3.85 9.71 Only minor kids 0.03 5.35 2.34 7.83 -0.77 7.43 Minor kids + Adult kids Ref Ref Ref Only adult kids 0.33 5.19 3.07 8.1 8 0.35 6.79 No children at home 0.27 6.63 -1.44 9.97 4.15 8.96

Non-residential minor children 4.4 4 5.96 2.4 6 6.7 7 5.27 15.69 Non-residential YA child(ren) -3.74 3.47 -9.29+ 5.22 0.04 4.67 # children in the household -3.27 2.75 -6.06 4.12 0.20 3.73

Gender (ref: men) 27.09*** 2.58 Race/Ethnicity (ref: NonHispanic White) NonHispanic Black -17.8 3*** 2.97 -11.8 9** 4.3 8 -23.18 *** 4.10 Hispanic -22.61*** 3.28 -17.21*** 4.84 -27.79*** 4.47 Marital Status (ref: Married) Never Married 6.3 7 4.27 11.0 6+ 6.29 2.70 6.00 Previously married 6.07+ 3.54 11.39* 5.63 3.23 4.71 Highest Grade Completed 0.69 0.55 0.61 0.82 0.76 0.75 Employment Status (ref-not employed) Employed -20.2 2*** 3.58 -25.6 7*** 5.9 4 -15.78 *** 4.53 Logged family income -10.70*** 2.49 -11.14** 3.82 -9.64* 3.76 Family size Health Insurance (ref: Private) Public 22.68 *** 4.11 21.52 *** 5.9 7 23.63 *** 5.74 Self-pay 16.74*** 3.79 11.18* 5.54 21.30*** 5.33 Self-rated health (ref: Fair/Poor) Good -52.80 *** 3.6 1 -54.10 *** 5.54 -51.58 *** 4.76 Very good -74.22*** 3.69 -68.19*** 5.58 -80.20*** 4.94 Excellent -99.42*** 4.46 -91.18*** 6.54 -109.04*** 6.19 First birth at or before age 20 8.66** 3.27 13.03* 5.51 6.48 4.18 Intercept 362.96*** 27.86 368.02*** 41.70 376.94*** 41.60 N (NLSY79) 7,324 3,454 3,870 + p<.10 * p<.05 ** p<.01 *** p<.001 Source: NLSY79 and 79 Young Adults Survey.

Figure 1. Predicted Psychological Distress by the Composition of Coresident Children

205

195

185 Predicted Psych Distress by Child 175 Composition and Gender Men 165 Predicted Psych Distress by Child 155 Composition and Gender Women 145 Only Minor Only adult No Childless minor kids kids + kids children Adult kids at home

Chapter 3. Intergenerational Coresidence and Mothers’ Body Weight at Midlife

A moderate midlife body weight is associated with favorable health outcomes including lower morbidity and extended longevity (Breeze et al. 2006). Consequently, the dramatic increase in the overweight and obese population since 1960s has been a major public health concern in the

United States (Wang and Beydoun 2007). In fact, more than two thirds of US adults are overweight (BMI>=25) and about 38% are obese (BMI>=30) (Flegal et al. 2016). Notably, recent American cohorts are obese for a longer duration of their lives, a trend that is particularly pronounced among women (Lee et al. 2010). Additionally, recent reports suggest that obesity rate is highest among middle-aged Americans (early 40s to mid 60s) relative to individuals in all the other age groups (Mendes 2010). Yet, the social determinants that are associated with body weight among midlife American women remain understudied.

One important social context for midlife women is manifested in the coresidential arrangements with emerging adult children. Nest-leaving of adult children is a historically key experience for midlife parents. Over the past two decades, however, demographers have noted drastic changes in home departure behavior, including high rates of what are known as

“boomerang” children who return to the maternal home after departing, as well as the “never left” who have remained in the parental home longer than expected. Boomerang and never left coresidential patterns appear to buffer young adults from economic shocks (e.g., economic recession, unemployment) and do not negatively affect young adults’ emotional and psychological well-being (Copp et al. 2015; Fingerman et al. 2016). Yet, what remains unknown 37 is how these changing coresidential patterns affect the well-being —including body weight—of another major family member in the coresidential home: mothers. A life course perspective argues intergeneraional coresidential patterns are very likely to affect parent’s body weight, especially mothers, due to the generations’ linked lives (Carr 2004; Milkie, Bierman and

Schieman 2008). Due to changing patterns of motherhood, midlife mothers now experience a more diverse set of coresidential patterns with their adult children that may affect well-being, such as body weight. Yet, virtually no empirical research examines this potential relationship.

In order to fill this research gap, this study asks: How are intergenerational coresidential biographies during mother’s midlife associated with one key health outcome: mothers’ body weight trajectories? We further test whether the effects of intergenerational coresidential patterns on mothers’ body weight trajectories are explained by two potential mechanisms: emerging adults’ life events and mothers’ socioeconomic characteristics. Body weight index (BMI) is a key indicator of health at midlife and is sensitive to life course changes including the parent- child relationship, and thus provides the ideal lens to examine the proposed effects of coresidence (Umberson et al. 2011).

Background

Intergenerational Coresidence and Mother’s Body Weight: A Life Course Perspective

Body weight is a health indicator that changes over the life course. Given that adult weight trajectories follows a slight upward trend across the life course (Adams and Schoenborn 2006), a relative moderate body weight (non-overweight) with a stable (or slightly increasing) trajectory suggests good and stable health whereas a dramatically fluctuating body weight trajectory suggests significant health change in response to various negative life events including those encountered by young adults through linked lives (Baltrus et al. 2005). Previous research has 38 established that motherhood is strongly associated with body weight changes across the life course, with this research focusing on the effects of parenting young children on parental body weight. For example, Laroche and colleagues (2013) find that living with young children (under five years old) is associated with accelerated weight gain among mothers relative to fathers and childless women across their young adulthood (18 to 30). This is likely because parenting young children is associated with a higher-calories diet (e.g., higher levels of fat, sugar, and meat) and less physical activity (Berge et al. 2011; Elstgeest, Mishra and Dobson 2012; Nomaguchi and

Bianchi 2004). In one of the few studies that examined the effects of parenting and mothers’ body weight as children age into adulthood, Umberson and colleagues (2011) found coresidence with adult children was associated with higher body weight at the baseline, further showing that the departure of adult children to be associated with greater weight gain for both fathers and mothers. Yet, Umberson and colleagues did not examine how young adults’ emerging coresidential patterns are related to mothers’ body weight change for the current cohort of midlife women, because their analyzed data was from an older cohort of parents (born between

1890 and 1962). The present study fills this important gap by studying intergenerational coresidential biographies on the current cohort of midlife mothers’ body weight.

Intergenerational coresidence trajectories are established as efficient indicators of children’s transitioning pathways to young adulthood (Houle and Warner 2017; Sassler,

Ciambrone and Benway 2008), and thus should be considered in terms of its association with maternal body weight changes at midlife (Grossman 2005; Newman 2012). Indeed, intergenerational coresidential patterns of midlife mothers (40-60) and their emerging adult children (18-30 year old) have undergone significant changes, in part because emerging adults

39 have experienced a decline in employment, increasing importance in higher education, the detachment of marriage and childbearing, and the postponement of first marriage (Cherlin 2010;

Cherlin 2014; Furstenberg 2010). According to a Pew Research report, 36% of emerging adults

(18-31) who come of age in the 2000s were living in parents’ household in 2012; in comparison,

32% of the emerging adults of the same age range were coresiding with parents in 2007 (Fry

2013). As a result, living with mothers has become the most common residential arrangement for the 18-to-34-year old in 2014, more so than living with cohabiting partner or spouse or living alone (Fry 2016).

Scholarly research on intergenerational coresidential biographies have mostly focused on the sociodemographic predictors of coresidence (e.g., socioeconomic status (SES)) (Sandberg‐

Thoma, Snyder, and Jang 2015; Seiffe-Krenke 2006; Ward and Spitze 2007) or the effects of coresidence on adult children’s economic and psychological well-being (Copp et al. 2015;

Kaplan 2012). However, how these coresidential biographies matter for the other player in these coresidential relationships: mothers, especially in terms of major health indictors such as body weight, remains unknown. In this study, we take a first step to amend this research gap to examine how intergenerational coresidential patterns shape midlife mothers’ body weight over time.

Three primary types of intergenerational coresidential patterns have gained prevalence in the past two decades in the U.S.: (1) children who left the parental household and never returned between ages 18-30 (i.e., the “gone for good”), (2) children who exited and came back at least once between the ages of 18-30 (i.e., “boomerangers”), and (3) children who did not exit parental household between the typical years of exit, 18-30 (i.e., prolonged coresidence, or the “never

40 left”) (Burn and Szoeke 2016; Copp et al. 2015; Newman 2012). Below, we further explicate each intergenerational coresidential pattern in the U.S. in order to build the case for why these patterns may differentially shape midlife mothers’ BMI.

The gone-for-good. Despite the increasing delay in home departure (Furstenberg,

Rumbaut, and Settersten 2005; Newman 2012), almost all young adults agree that they should leave parental household by age 25 (Settersten 1998). In a report that uses data from NLSY97—

Millennial cohort of emerging adults in the U.S., 47% of men and 46% of women leave parental household after 18 and do not return by age 24 (Payne 2011). The majority of these young adults move in with a spouse or cohabiting partner, with the remainder either live alone or enter semi- autonomous living such as dormitories and military barracks, or move in with other family members (Goldscheider and Goldscheider 1999). Many young adults also relocate geographically away from their parental home because the educational or employment opportunities are in another location (Goldscheider and Goldscheider 1999). In general, the independently-living young adults are more likely to be employed and earn more than those who live with parents (Payne and Copp 2013). Among the three coresidential patterns, a gone-for- good coresidential pattern may be most favorable for midlife mothers’ waistline, as intergenerational coresidence involves daily stressors, which are important biosocial mechanisms that influence body weight (Pearlin et al. 2005). For example, the transition to an empty nest is associated with reduced daily hassles (Dennerstein et al. 2002), reduced parent-child conflicts

(Ward and Spitze 2007), and decreased work-family conflicts (Erickson, Martinengo, and Hill

2010), all of which are important stressors that can link with body weight. In other words, a

41 gone-for-good coresidential pattern means fewest stressors; as a result, mothers of the gone-for- good may have the most moderate midlife body weight trajectory among all three groups.

The boomerangers. In this study, boomerang children are defined as those who have left the parental home after 18 but return before the age of 30. According to a nationwide survey of

2,048 adults conducted in 2011 by the Pew Research Center, 40 % of 18-to-24-year-old boomerang after the initial departure (Parker 2012) — this does not include boomeranging during college enrollment, but does include children who return home after dropping out of or completing college. Postponement in young adult role transitions (e.g., transition to marriage and home ownership) and deficiency in young adult’s economic resources—such as the loss of income, the lack of employment opportunities, or the greater cost of sustaining independent living— are associated with higher likelihood of returning home (Houle and Warner 2017;

Sassler et al. 2008; Smits, Van Gaalen, and Mulder 2010). Compared to mothers of the gone-for- good, mothers of the boomerangers likely encounter more coresidential stressors. Moreover, boomerangers’ mothers may need to tackle additional stressors while adjusting to children’s transition back to the nest, such as the negotiation of housework division and financial responsibilities. Consequently, midlife mothers of the boomerangers may have higher body weight than mothers of the gone-for-good.

The never-left. Recent research on intergenerational continuous coresidence into children’s young adulthood highlights a third less common, but increasingly important subgroup of adult children: adult children that have not left the parental household prior to age 30. For instance, using a sample of 1,113 older mothers aged 58-to-97, Wiemers and colleagues (2017) find that 6.85% of women had never-left adult children in their household at the time of the

42 survey. Although there is little known about this group of young adults and their parents, young adults who are male, unmarried, unemployed, economically needy, and those who are working on their post-secondary education are more likely to be never- or later-leavers (Burn and Szoeke

2016; Goldscheider and Goldscheider 1999; Sassler et al. 2008). Parents embedded in a stable family structure (stably married or consistently single) are more likely to live with young adult children (Smits, Van Gaalen, and Mulder 2010). Compared to mothers of the gone-for-good, mothers of the never-left are exposed to the daily stressors for a longer duration of time, which likely takes a toll on their waistline. While mothers of the boomerangers need to handle additional stressors associated with children’s transitioning back home, mothers of the never-left may need to manage a situation wherein children lack progress in transition to adulthood. As a result, mothers of the boomerangers and mothers of the never-left likely have distinct body weight trajectories reacting to different stressors, although the exact differences are hard to theorize due to the lack of relevant research.

Taken together, we hypothesize that:

H1: Mothers of the gone-for-good will experience the lowest BMI score at baseline and

the most moderate increase in their BMI trajectory among the three groups.

Coresidential Patterns and Mother’s Body Weight: Life Course Mechanisms

Guided by the life course perspective, we further analyze two mechanisms of the connection between mother-child coresidential patterns and mothers’ body weight: (1) adult children’s life events and (2) mothers’ SES. Mother-child coresidential patterns are linked with mothers’ body weight trajectories likely through the stress related to young adults’ life events/transitions.

Because body weight is very sensitive to perceived stress and psychological distress (Greeno and

Wing 1994), a life course perspective further suggests mothers’ body weight trajectory is connected with intergenerational coresidential patterns via a process of intergenerational stress proliferation, wherein adult children’s stressors grow and spread through the family network and become mothers’ stressors (Milkie et al. 2008; Pearlin 1999). As evidence of this perspective, current research shows that mothers’ mental wellbeing decreases when adult children encounter problems or negative life events (e.g., divorce, unemployment, drinking/drug problem, health problems, and problems with law) (Greenfield and Marks 2006; Kalmijn and De Graaf 2012). In fact, adult children’s negative life events are found to be highly associated with increased parental concerns and poorer intergenerational relationship (Bierman and Milkie 2008, Birditt,

Fingerman and Zarit 2010; Milkie et al. 2008), which are the proliferated stressors for parents that do not bold well for multiple aspects of parental wellbeing including body weight. Among the current cohort of young adults, the gone-for-good experience a relatively smoother transition to adulthood, whereas boomerang young adults encounter more detours and setbacks (Payne and

Copp, 2013; Houle and Warner 2017). Similar to a boomeranger, a never-left young adult may also experience a bumpier transition to adulthood than the gone-for-good. In fact, the never- or late-leavers are found to achieve fewer adulthood milestones (e.g., stable employment, marriage) than those who have left (Burn and Szoeke 2016).

Taken together, we hypothesize that:

H2: Considering young adults’ life events will at least partially explain the differences in

maternal BMI trajectories between mothers of the gone-for-good and mothers in the two

other groups.

Mother’s SES can serve as a second mechanism that explains the effect of intergenerational coresidence-related stress on mother’s BMI. A robust literature has shown that

SES is inversely associated with women’s body weight across the life course (Ailshire and

House 2011; Holowko et al. 2016). This is because individuals of different SES have fundamentally different knowledge of, access to, and use of medical care to maintain good health

(De Lew and Weinick 2000; Link and Phelan 2000). For example, individuals of high SES are found to have healthier dietary practices, more regular physical exercise (Wardle and Griffith

2001), and less depression—all contributors to body weight (Lorant et al. 2003). Coresidential patterns are also likely to be associated with maternal SES. One study using a small regional sample finds that parents of lower education were more likely to house young adult children than parents (Fingerman et al. 2015), but they did not differentiate between the boomerangers and the never-left. In addition to the potential selection effects, if young adults experience a coresidential trajectory that causes proliferated stressors for mothers, mothers’ of higher SES should be able to counteract some of these negative effects. Today, gone-for-good young adults are more likely to support themselves with their own earnings and/or pooling resources shared with their partners

(Fingerman et al. 2015; Furstenberg 2010); parents may use more of their income to enhance their own wellbeing. On the other hand, young adults boomerang to parental household mostly because they have encountered economic setbacks (Smits, Van Gaalen, and Mulder 2010).

Never-left postpone their departure likely because they are not economically ready (Burn and

Szoeke 2016). Recent research has also shown that extensive intergenerational coresidence substantially reduces parental financial assets and savings (Maroto 2017). In this case, high SES parents may be able to support their adult children while buffering any effect on their own

45 health, whereas parents of lower SES may have to assist their children at the expense of their own wellbeing. We thus hypothesize that:

H3: Considering maternal SES will at least in part explain the differences in maternal

BMI trajectories between mothers of the gone-for-good and mothers in one or both of the

two other groups.

The Present Study

This study examines the role of intergenerational coresidential patterns on midlife mothers’ BMI trajectories. This study compares how mother’s body weight is impacted by these three types of mother-child coresidential biography between midlife mothers and young adult children: (1) young adult left and never returned (the “gone-for-good”), (2) young adult never left the household of mother (the “never left”), (3) young adult made transition(s) back home after the initial departure (“boomerang children”). We further test two important mechanisms – young adults’ life events and mothers’ socioeconomic characteristics, in explaining the relationship between intergenerational coresidence and maternal weight change across midlife.

Data and Methods

Description of the Data

This study uses the National Longitudinal Survey of Youth 1979 to investigate the impact of intergenerational coresidential biographies on mother’s BMI during midlife. In 1979, 12,686 respondents, aged 14-22, were randomly selected and followed over time. Respondents were interviewed annually between 1979 and 1993 and biennially since 1994. The most recent wave of data available is for 2014. Despite the initial intention of studying the baby boomers’ labor market experiences as they transitioned from adolescence into adulthood, the NLSY79 remains one of the best datasets with which to examine health and its determinants in a longitudinal 46 fashion. In 1986, data collection efforts began on the NLSY79-Children (NLSY79-CH), which is comprised of all children born to NLSY79 mothers, to gather information regarding a wide range of social, economic, and psychological characteristics. It is these children who form the basis of the NLSY79 – Young Adult Survey, which officially began in 1994 when the oldest of the

NLSY79 Children reached their 14th birthday. For the current study, we combine data from female members of the original NLSY79 cohort (NLSY79) with information from their young adult children (NLSY79-YA) who were 18 years of age or older from 1994 to 2014.

There are 3,301 NLSY79 mothers who had at least one adult child (18 years or older) by

2014 and for whom we had valid information concerning the latter’s living arrangements. We excluded about 0.07% of the observations for which the NLSY79 young adults did not reside with their mothers prior to turning 18. We also excluded year-specific data for NLSY79-YA respondents who were under the age of 18 or institutionalized during a given survey year. An additional 3% of cases were not included in the subsequent analyses due to missing values on the dependent variable of interest (i.e. mother’s BMI). Furthermore, less than 1% of values on weight and/or height reported by NLSY79 mothers were extreme (i.e. greater than 3 standard deviations from the mean) and likely attributable to respondent error or interviewer miscoding.

We addressed this issue of outlying observations by substituting these extreme values with person-specific mean BMIs. We also conducted sensitivity analyses to determine if our findings depended on the inclusion of these outlying BMI values and found regression results remained qualitatively unchanged. These inclusion criteria generated an analytic sample of 7,220 mother- young adult child pairs that were followed from 1994 through 2014. Thus, the unit of analysis in the paper is person-years (N=28,010).

Description of the Measures

The outcome measure is body weight index (BMI), which is calculated by dividing self-reported maternal weight (in kilograms) by height (in meters) squared. Self-reported weight and height have established validity in nationally representative surveys, largely due to the ease and accuracy with which respondents can recall them (Gorber et al. 2007). Although in recent years there has been debate concerning the utility of BMI as an indicator of overall health and subsequent morbidity and mortality (Ahima and Lazar 2013), the effects of stress on this particular health outcome are well defined and provide a biologically plausible pathway through which young adult coresidential trajectories are likely to impact mother’s wellbeing. We model

BMI as a continuous variable in growth curve analysis, which allows us to examine the trajectory and degrees of change in BMI as it varies incrementally over time (Umberson et al. 2011). For ease of understanding, one BMI unit is comparable to about 3.13 kg (almost 7 pounds) for a woman of average height. Because change in BMI for the middle-age women is primarily driven by change in weight, we thus use “body weight” or “weight” to refer to the dependent variable in subsequent sections.

To construct our primary independent variable, coresidence trajectories, we relied on information gleaned from NLSY79-YA from age 18 onward. This study examines coresidence trajectories starting at age 18— the legal age of adulthood. Age 30 is used as the end point for two reasons: (1) theoretically by age 30, most of the young adults have finished all the tertiary education and they either have made the first transition out of parental household or have made the first return by then (Parker 2012); (2) for the majority of the young adults in NLSY79 by the most recent wave (2014), they are about to turn 30; a small number of those who are older than

30 represent a very selective group (e.g., born to the teen mothers, lower SES). In each survey 48 year, the young adult respondent is asked about their current residential situation. They can choose from a variety of options including but not limited to living with their mother, their father, both parents, another relative, or their own dwelling unit. Based upon this information, we categorized NLSY79-YA young adults into three categories. The “never left” group was comprised of those who remained in their childhood home from age 18 onward. Those who left their childhood home either before or by age 18 and never returned were classified as the “gone for good” category. Young adult respondents who exited from their childhood home before or by age 18 and transitioned back, either once or multiple times, after age 18 were labeled

“boomerang” young adults. For example, at age 40, a mother had an 18-year-old son, who stayed in maternal home in the childhood, left at age 21, and returned to maternal home at age 26; in the most recent wave, this mother was 50 years old and this son of hers was 28 years old. In this case, the coresidential pattern for this mother-child pair is categorized as “boomeranger” and this mother’s body weight between age 40 and 50 would be analyzed in relation to the coresidential patterns.

Although most of the never-left adult children may ultimately become gone-for-good or boomerangers as they age, we argue that it is important to keep them as a separate category at this life course stage for a number of reasons. First, this category captures a delayed transition out of parental household (Burn and Szoeke 2016). Second, the experience of housing an adult child who has never exited prior to 30 may have a unique impact on mothers’ body weight changes. In subsequent sensitivity analyses (available upon request), we divided the

“boomeranger” category into two additional groups—those who remained within their mother’s residence upon returning home and those who, at some point in the future, left their childhood

49 home. Sensitivity analyses conducted using either approach to classifying the “boomerangers” remained remarkably consistent across models. Thus, for ease of interpretation and to preserve statistical power, we combined all the “boomerangers” into a single group.

We account for a number of potential confounders in multivariate regression analyses that have been previously shown to be associated with both young adult living arrangements

(Burn and Szoeke 2016) and mother’s BMI (Baltrus et al. 2005). The first set of control variables were chosen to capture key sociodemographic conditions that characterize the young adult’s major life events and transitions (H2). They include sex (male/female), marital status (never married, married, cohabiting, other), parental status (yes/no), educational attainment (below high school, high school, college or above), and employment status (neither in school or working, in school, working, in school and at work). Sociodemographic and SES-related control variables for

NLSY79 mothers were incorporated into subsequent models (H3). They are as follows: race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic), highest grade completed, marital status (never married, married or cohabiting, other), family income (adjusted for inflation and reported in 2014 dollars, and transformed to address positive skewness), number of persons in the family which was top coded at seven plus, and region of residence (northeast, north- central, south, and west) to address the influences of housing prices on co-residential patterns.

Finally, in the most comprehensive models, we add covariates to control for important maternal health behaviors such as cigarette smoking (non-smoker, ex-smoker, current smoker), frequency of physical exercise (daily, weekly, little or none), and health insurance coverage (private, public, self-pay).

Analytical Approach

We employed multilevel regression models (i.e., growth curve models with random intercepts and random coefficients) to examine how young adult’s coresidential biographies influence their mother’s BMI during midlife. Growth curve analysis is appropriate for this study because of its flexibility in recognizing the variability of baseline weight as well as the heterogeneity in weight trajectories across individuals over time. As a result, growth curve models are well-positioned to distinguish between within-individual variability and between-individual change. To address the correlations within individuals and between observations, all models estimate a random intercept and slope for each mother; all the other covariates are modeled as fixed effects.

The analytical approach is further elaborated by the following equation:

2 푦푖푗 = 훽1 + 훽2 훿푗 + 훽3푡푖푗 + 훽4푡푖푗 + 훽5푥5푖푗 … + 훽푛푥푛푖푗 + 휍1푗 + 휍2푗푡푖푗 + 휖푖푗 where 푦푖푗 represents the BMI of mother i at time j, 훽1 is the population-averaged intercept (or the initial BMI score across all mothers), 훿푗 is the indicator variable for mother-child coresidential patterns, 푡푖푗 is the age of mother j at time i, x5 - xn stand for each additional covariate, 휍1푗 is the random intercept, 휍2푗 is the random slope, and 휖푖푗 is the individual- and time-specific error term.

Regression models are estimated using the mixed command in Stata 15.0. Robust standard errors were calculated using the Huber/White correction estimate and clustered at the highest level of aggregation (the NLSY79 mother).

We employed multiple imputation methods with chained equations to deal with missing values on covariates. The proportion of missingness for all variables around the suggested cut point of 5% (Schafer 1999) except for mother’s family income, for which 13% of person-years did not contain valid responses. To address potential bias, we included other variables that are

51 highly correlated with family income such as insurance type and maternal education. We performed sensitivity analysis excluding the variable of family income (available upon request); the results remained qualitatively similar to those presented in the study. Gibbs sampling techniques generated conditional distribution for missing values on all variables (Royston 2005,

Van Buuren 2012). According to the suggestion in extant literature that 3 to 5 datasets is appropriate for models with up to 20% missing, we produced five distinct datasets (van Buuren,

2012; Royston, 2005). Guided by previous research (Von Hippel 2007), we first ran the imputation model including the observations with missing values on BMI; we then deleted these observations. To accommodate the imputed data, all descriptive statistics and regression analyses are executed using the mi command in Stata/MP 15.0 (College Station, TX: StataCorp LP).

Results

Descriptive Findings

Descriptive results for the sample as a whole and stratified by parent-child coresidential biographies are presented in Table 4. Significance tests are conducted to document differences by intergenerational coresidential patterns. Among the 7,220 mother-young adult child pairs in the sample, over half of the young adults (61%) followed a coresidential biography of gone-for- good (GFG), 17% fell in the category of the never-left (NL), and 23% of young adults were boomerangers (BMG). Mean body mass index for the entire sample of mothers was 29.31 kg/m2. Among the three groups, mothers of the BMG and the NL had similar BMI scores at

29.94 kg/m2 and 29.74 kg/m2, respectively, which were significantly higher than BMI scores among mothers of the gone for good (28.96 kg/m2) (p <0.001).

Mothers who had distinct intergenerational coresdiential biographies appear to differ from one another on key sociodemographic indicators. Compared to mothers of the GFG young 52 adults, boomerangers’ mothers had lower SES (e.g., lower family income, fewer years of completed education, and higher rate of public insurance), whereas mothers of the never-left had similar or even higher SES. Young adults’ demographic covariates also differed from one another by intergenerational coresidence. Compared to the gone-for-good, boomerang young adults were less likely to be married and completed fewer years of education. Young adults of the never-left achieved fewer milestones of adulthood (e.g., transition to marriage and parenthood, completion of post-secondary education) than gone-for-good young adults likely because the former group was younger in age. For additional details of the descriptive results, please refer to Table 4.

Multilevel Regression Results

Unconditional growth curve models with no covariates (available upon request) show that the BMIs of NLSY79 mothers increased by 0.12 points on the BMI scale per year (b= 0.12, p< .001). This translates to about 3.5 pounds of weight gain over five years (or) for a female with a height of 5 feet 3 inches, the average height of American women (CDC, 2012). We also found substantial variation in the random intercept (var(훽1)= 212.27 , p< .001) and slope

(var(훽3)= 0.086 , p< .001), which suggests considerable between-person and within-person differences in the body weight trajectories.

Results from multilevel regression analyses predicting BMI among mothers by young adult children’s coresidential biographies are presented in Table 5. We centered the time variable

(mother’s age) so that the coefficients for different coresidential groups represent mothers’ mean

BMI score at age 35, which is the starting age of early midlife. Model 1 is the baseline model and only contains the main independent variables – coresidential biographies, age, and

53 coresidence*age. According to Model 1, mothers of the NL were, on average, 1.14 kg/m2 (p

<0.01) heavier than mothers of GFG at baseline. In this model, the BMI score of the GFG mothers increased at a rate of 0.130 kg/m2 per year (b = 0.130, p < 0.001). The significant negative coefficient of never-left*age indicated that the BMI score of the NL mothers increased at a rate that is 0.077 unit slower than that of the GFG mothers (b = -0.077, p < 0.001). As a result, sensitivity analysis shows that by age 55 the BMI scores of GFG and NL mothers are no longer different from each other. In terms of the BMI difference between BMG mothers and

GFG mothers, Model 1 indicates that the former group had an average BMI score that is 1.059 kg/m2 higher (p <0.001) than the latter; the lack of significance in the coefficient of boomerangers*age suggests that BMG mothers and GFG mothers had similar body weight trajectory over the studied period. In other words, BMG mothers were, on average, 1.059 kg/m2 heavier than the GFG mothers across midlife.

Predicted values from Model 1 were plotted in Figure 2, which shows that NL mothers had higher BMI score than GFG mothers in early midlife, but such difference gradually decreased and ultimately converged at around age 50. On the other hand, BMG mothers had higher body weight than GFG mothers at baseline and the weight difference was consistent across midlife. Taken together, Hypothesis 1 is partially supported in that GFG mothers had a lower initial BMI and more moderate increase in the BMI trajectory across midlife (e.g., the lowest BMI at midlife among the three groups), while NL mothers had a higher baseline BMI but had the slowest rate of increase in BMI among the three groups.

In order to test Hypothesis 2, we controlled for young adults’ life events in Model 2.

Compared to results in Model 1, the coefficient of the baseline difference (at age 35) in BMI

54 between mothers to the NL and the GFG increases slightly from 1.144 to 1.539 and remains statistically significant (p < 0.001); the difference in the rate of change decreases slightly from

0.077 to 0.056 kg/m2 per year and the significance level decreases (from p < 0.001 to p < 0.01).

In other words, compared to GFG mothers, NL mothers still had higher body weight at early midlife but weight increased less over time. Sensitivity analysis shows that the BMI scores of

GFG and NL mothers converged at early 50s (and thus no statistically significant difference at this time point). In turn, the baseline difference in BMI score between BMG mothers and GFG mothers slightly decreases from 1.059 to 1.018 but remains statistically significant (p <0.001)

The null effect of boomeranger*age suggest that the BMI scores of these two groups were increasing at the same rate. In other words, the BMI differences between these two groups remain consistent across midlife. In sum, these results offer minimal support to Hypothesis 2 because the inclusion of young adults’ life events does little to narrow the differences in maternal

BMI trajectories by mother-child coresidential patterns.

In order to test hypothesis 3, we further control for mothers’ SES covariates in Model 3.

Compared to results in Model 2, the coefficient of the baseline difference in BMI between NL mothers and GFG mothers changes little (from 1.539 to 1.180) and remains statistically significant (p < 0.001); the coefficient of the rate of change (b = -0.062) is also similar to that in

Model 2 (b = -0.056) and remains statistically significant (p < 0.01). This is likely because maternal SES does not differ much between these two groups (see Table 4 for details). On the other hand, the introduction of maternal covariates in Model 3 substantially reduces both the coefficient magnitude and significance level of the BMI difference across midlife between BMG mothers and GFG mothers, from 1.018 (p <0.001, Model 2) to 0.674 (p <0.05), suggesting that a

55 considerable amount of difference in maternal body weight differences in midlife between these two groups are due to differences in maternal SES. Taken together, Hypothesis 3 is supported because the introduction of maternal SES covariates reduces body weight differences by mother- child coresidential patterns, especially those between BMG mothers and GFG mothers.

Figure 3, which includes the predicted body weight values generated from Model 3, demonstrates that mothers of the NL started with higher baseline BMI than mothers of GFG; controlling for both young adults’ life events and maternal SES covariates, the body weight of

GFG mothers increased at a moderate speed while NL mothers appeared to have a decreasing weight trajectory over time, compared to mothers of GFG. Despite the convergence of BMI among GFG mothers and NL mothers in early 50s, BMG mothers continued to have higher body weight than GFG mothers across midlife.

Discussion

The rise of overweight and obese population in the U.S. has drawn much research attention, yet the underlying social determinants of body weight at midlife remain under-addressed. Research is especially needed on the body weight of midlife women because women tend to be obese for a prolonged period of time than men and midlife is the life course stage with the highest obesity rate. This study examines whether the social context of motherhood explains midlife women’s body weight changes. Midlife mothers today are parenting an increasingly diverse set of young adult children who take multiple pathways to independence, with ramifications for both young adult children and their parents. This study is the first to examine how mother-adult child coresidential patterns are associated with mother’s body weight trajectories at midlife. Further, this study tests the role of two explanatory mechanisms to explain any potential significant relationship: young adults’ life events and maternal SES. Results from growth curve models 56 show that coresidential patterns are associated with both baseline BMI differences as well as different BMI trajectories. Below, we discuss our three research hypotheses and highlight our main contributions to the literature.

First, our findings strongly suggest that coresidential biographies matter for midlife mothers’ well-being both at baseline and over time, even when all covariates are included. Our hypothesis 1, that midlife mothers who have children that are gone-for-good will experience the lowest BMI score at baseline and the most moderate increase in their BMI trajectory than other mother-child profiles, is mostly, but not fully, supported. We find that mothers with gone-for- good children experienced the lowest baseline BMI; at 40, having a child who has left the nest and not returned is associated with lower BMI than midlife mothers whose children boomeranged back into the home or whose adult children had never left the home. BMI trajectories into the late 50’s show that mothers of the boomerangers continued to have higher body weight than mothers to the gone-for-good until late 50s. However, mothers of the never-left and mothers of the gone-for good converged in their early fifties.

In order to explain these differences, we hypothesized two potential explanations for mothers’ body weight trajectories across coresidential patterns. First, previous literature has suggested that maternal wellbeing is linked with young adults’ ongoing life events through a process of intergenerational stress proliferation (Milkie et al. 2008; Pearlin 1999). The life events examined in this study, including young adults’ transitions in romantic relationships, parental status, and economic status, represent the most occurred life course transitions for young adults at this stage. Thus, we tested the role of children’s life events on the effect between coresidential biographies and mothers’ BMI, hypothesizing that young adults’ life events would narrow the

57 differences in maternal BMI trajectories between mothers of the gone-for-good and mothers in the two other groups (Hypothesis 2). Hypothesis 2 was not supported, as the inclusion of young adults’ life events does little to narrow the differences in maternal BMI trajectories by intergenerational coresidential patterns. In other words, as different combinations of these major life events experienced by mothers’ young adult children are suggestive of different coresidential pathways, they do not contribute to the significant relationship between coresidential patterns and mothers’ midlife BMI. This may be because midlife mothers today are more adjusted to children’s heterogeneous life course patterns including delayed parenthood and marriage, delayed employment, and delayed educational attainment and they thus do not experience stress by providing further assistance to young adult children (Fingerman et al. 2012b; Furstenberg

2010; Settersten and Ray 2010). In fact, current research has corroborated that parental assessment of emerging young adults’ success does not overlap with the objective indicators of adult roles (i.e., employment, education, marital status) (Fingerman et al. 2012a). As a result of the increasing acceptance of children’s divergent pathways to adulthood, young adults’ life events thus fail to explain the relationship between the coresidential patterns and mothers’ midlife body weight.

Additionally, we hypothesized that maternal SES would explain the relationship between intergenerational coresidential patterns and mothers’ body weight at midlife (Hypothesis 3).

Hypothesis 3 is partially supported, as the introduction of maternal SES covariates explains some, but not all of the body weight differences between mothers of the boomerangers and mothers to the gone-for-good at baseline. Our analysis suggests that a boomerang coresidential trajectory is in fact more prevalent among mothers of lower SES. Accounting for maternal SES

58 thus explains a sizable amount of the difference in the midlife body weight between boomerangers’ mothers and mothers whose children are gone for good from the late 30s to late

50s. Current research finds that young adults are more likely to boomerang when they experience economic or educational setbacks (Coate, Krolikowski, and Zabek 2017; Houle and Warner

2017), while coresidence with young adult children does significantly reduce maternal saving

(Maroto, 2017). Accordingly, mothers of the boomerangers may use their own economic resources to counteract the economic setbacks experienced by their children, reducing the resources available to improve their own health. The economic stress in combination with less resource for health-promoting behaviors (e.g., healthier diets) can be further linked with an increase in maternal body weight. Maternal SES, as potential indicators of a selection effect as well as economic stress, thus explains part of the association between intergenerational coresidence and maternal body weight at midlife.

It is worth noting that differences in the midlife body weight trajectories of the boomerangers’ and gone-for-good’s mothers are not completely explained by maternal SES. As the rate of change in body weight between late 30s and late 50s did not vary between these two groups, it is possible that the driving force of such difference is a pre-existing selection factor

(e.g., other uncontrolled SES characteristics). Because current literature has established that adolescent mothers are particularly disadvantaged as they have lower educational attainment, lower family income and poorer midlife health (Hoffman 2008), we performed sensitivity analysis that excludes adolescent mothers (i.e., mothers having their first child before age 20) in the analysis and then compared it to the main results. The descriptive analysis shows that adolescent mothers were indeed more likely to have boomerang adult children. Multivariate

59 analysis excluding the adolescent mothers reports no BMI difference between boomerangers’ mothers and mothers to the gone-for-good when controlling for maternal SES. Taken together, mothers whose children are gone-for-good may receive potential health benefits as their children achieve a normative coming-of-age pathway for emerging young adulthood. On the other hand, mothers whose children never-left or boomeranged may, over time, cope with additional stress caused by these non-normative patterns in ways that is linked with higher body weight once they reach their 50’s. This finding extends prior literature, most of which finds that the normative behaviors of adult children is positively associated with parents’ psychological wellbeing (e.g.,

Fingerman et al. 2011; Pudrovska 2009), in that study reveals that the normative leaving-home behavior of young adult children, has substantial impact on mothers’ physical health as well.

Limitation and Conclusion

This study contains limitations that warrant future research. First, while all women in the sample are midlife, due to heterogeneity in childrearing patterns not all children born to these women have completed their coresidential pathways between age 18 and 30 because not all of them have reached age 30 by the most recent wave. However, the variation in young adults’ age does not interfere in our research endeavors, because the primary focus of this study is on the coresidential patterns of midlife mothers, not on young adults. The main independent variable is mother-child coresidential patterns during mother’s midlife, and we aim to describe how these patterns are linked with mothers’ body weight changes at this exact life course stage.

Nevertheless, we conducted sensitivity analysis to address this potential bias. For example, we ran sensitivity analysis including only mothers whose adult children reached at least age 25 by the most recent wave, the cutpoint age of emerging young adulthood (Arnett 2007); the results

60 are comparable to those presented in the text. We also ran sensitivity analysis on mothers aged

40-to-55, another widely-used cutpoint of midlife; results are qualitatively similar to those presented in the text.

Second, this study cannot completely exclude reverse causality. Young adults may stay longer in maternal household or boomerang because mothers have high or dramatically changing

BMI, although mother-child coresidence is mostly to attend to young adults’ needs instead of vice versa at this life course stage (Fingerman et al. 2012b). We ran sensitivity analysis that includes both mothers’ and young adults self-rated health. Similar results are found. Third, measurement error can bias the findings. Because NLSY79 young adults are interviewed every other year and some even have missing waves (e.g., non-interviewed), it is possible that some transition episodes are not captured in the data. We address this by controlling for the missing waves in the supplementary analysis, which reports similar results as those presented in the text.

Despite potential measurement error, current research finds individuals’ own report of residence to be more accurate than living arrangement inferred from the household roster because it can eliminate situations such as temporary living situation (e.g. residential spells in college) (Houle

& Warner 2017). Finally, one mother can have multiple children who experience different coresidential pathways. Findings from additional analysis that randomly selects one child for one mother resemble those presented in the text.

Despite the limitations, this study expands prior research on the links between mothering young adults and women’s midlife health. Intergenerational coresidential patterns are found to have sustained effects on maternal body weight trajectories across midlife. This study thus contributes to the literature by unraveling how mothering experiences beyond the core parenting

61 years continue to shape mothers’ physical health as both generations’ linked lives unfold. The findings also provide updates to the changing demographics of family life for midlife mothers and the related health consequences. By doing so, this study offers insights into other alarming social issues, such as the stark increase in morbidity and all-cause mortality among many middle-aged Americans since the 2000s (Case and Deaton 2015). The increasing demands and diverse experiences of parenting young adult children may be another prominent angle to address this health emergency. Taken together, situated in the current-day economy when young adults need more parental resources to launch, more public assistance and resources should be directed to young adults and their family members such as midlife mothers to improve their wellbeing.

References

Adams, Patricia F and Charlotte A Schoenborn. 2006. "Health Behaviors of Adults: United

States, 2002-04." Vital and Health Statistics. Series 10, Data from the National Health

Survey (230):1-140.

Ahima, Rexford S and Mitchell A Lazar. 2013. "The Health Risk of Obesity—Better Metrics

Imperative." Science 341(6148):856-58. doi: 10.1126/science.1241244.

Ailshire, Jennifer A and James S House. 2011. "The Unequal Burden of Weight Gain: An

Intersectional Approach to Understanding Social Disparities in Bmi Trajectories from

1986 to 2001/2002." Social forces; a scientific medium of social study and interpretation

90(2):397-434. doi: https://doi.org/10.1093/sf/sor001.

Arnett, Jeffrey Jensen. 2007. "Emerging Adulthood: What Is It, and What Is It Good For?".

Child Development Perspectives 1(2):68-73.

Baltrus, Peter T, John W Lynch, Susan Everson-Rose, Trivellore E Raghunathan and George A

Kaplan. 2005. "Race/Ethnicity, Life-Course Socioeconomic Position, and Body Weight 62

Trajectories over 34 Years: The Alameda County Study." American Journal of Public

Health 95(9):1595-601.

Berge, Jerica M, Nicole Larson, Katherine W Bauer and Dianne Neumark-Sztainer. 2011. "Are

Parents of Young Children Practicing Healthy Nutrition and Physical Activity

Behaviors?". Pediatrics 127(5):881-87.

Bierman, Alex and Melissa A Milkie. 2008. "Intergenerational Stress Proliferation between

Adult Children and Parents: Contingencies by Functional Timing and Parent's Gender."

Advances in Life Course Research 13:343-67. doi: 10.1016/S1040-2608(08)00013-0.

Birditt, Kira S, Karen L Fingerman and Steven H Zarit. 2010. "Adult Children’s Problems and

Successes: Implications for Intergenerational Ambivalence." Journals of Gerontology

Series B: Psychological Sciences and Social Sciences 65(2):145-53. doi:

https://doi.org/10.1093/geronb/gbp125.

Breeze, E., R. Clarke, M. J. Shipley, M. G. Marmot and A. E. Fletcher. 2006. "Cause-specific

mortality in old age in relation to body mass index in middle age and in old age: follow-

up of the Whitehall cohort of male civil servants." International Journal of Epidemiology

35(1):169-178.

Burn, Katherine and Cassandra Szoeke. 2016. "Boomerang Families and Failure-to-Launch:

Commentary on Adult Children Living at Home." Maturitas 83:9-12. doi:

http://dx.doi.org/10.1016/j.maturitas.2015.09.004.

Carr, Deborah. 2004. "My Daughter Has a Career; I Just Raised Babies”: The Psychological

Consequences of Women's Intergenerational Social Comparisons." Social Psychology

Quarterly 67(2):132-54. doi: 10.1177/019027250406700202.

Case, Anne and Angus Deaton. 2015. "Rising Morbidity and Mortality in Midlife among White

Non-Hispanic Americans in the 21st Century." Proceedings of the National Academy of

Sciences of the United States of America 112(49):15078-83. doi:

10.1073/pnas.1518393112 [doi].

Cherlin, Andrew J. 2010. "Demographic Trends in the United States: A Review of Research in

the 2000s." Journal of Marriage and Family 72(3):403-19. doi: 10.1111/j.1741-

3737.2010.00710.x.

Cherlin, Andrew J. 2014. Labor's Love Lost: The Rise and Fall of the Working-Class Family in

America. New York: Russell Sage Foundation.

Coate, Patrick, Pawel Krolikowski and Mike Zabek. 2017. "Parental Proximity and the Earnings

Consequences of Job Loss." Economic Commentary 3(February).

Copp, Jennifer E, Peggy C Giordano, Monica A Longmore and Wendy D Manning. 2015.

"Living with Parents and Emerging Adults’ Depressive Symptoms." Journal of Family

Issues 38(16):2254-76. doi: 10.1177/0192513X15617797.

Dennerstein, Lorraine, Emma Dudley, and Janet Guthrie. 2002. "Empty nest or revolving door?

A prospective study of women's quality of life in midlife during the phase of children

leaving and re-entering the home." Psychological Medicine 32(3): 545-550.

De Lew, Nancy and Robin M Weinick. 2000. "An Overview: Eliminating Racial, Ethnic, and

Ses Disparities in Health Care." Health care financing review 21(4):1-7.

Elstgeest, Liset EM, Gita D Mishra and Annette J Dobson. 2012. "Transitions in Living

Arrangements Are Associated with Changes in Dietary Patterns in Young Women." The

Journal of nutrition 142(8):1561-67. doi: 10.3945/jn.112.158188.

Hill, E. Jeffrey, Jenet J. Erickson, Kaylene J. Fellows, Giuseppe Martinengo, and Sarah M.

Allen. 2014. "Work and family over the life course: Do older workers differ?." Journal of

Family and Economic Issues 35(1): 1-13.

Fingerman, Karen L, Yen-Pi Cheng, Lauren Tighe, Kira S Birditt and Steven Zarit. 2012a.

"Relationships between Young Adults and Their Parents." Pp. 59-85 in Early Adulthood

in a Family Context, edited by A. Booth, S. L. Brown, N. S. Landale, W. D. Manning and

S. M. McHale. New York, NY: Springer.

Fingerman, Karen L, Yen‐Pi Cheng, Eric D Wesselmann, Steven Zarit, Frank Furstenberg and

Kira S Birditt. 2012b. "Helicopter Parents and Landing Pad Kids: Intense Parental

Support of Grown Children." Journal of Marriage and Family 74(4):880-96. doi:

10.1111/j.1741-3737.2012.00987.x.

Fingerman, Karen L, Meng Huo, Kyungmin Kim and Kira S Birditt. 2016. "Coresident and

Noncoresident Emerging Adults’ Daily Experiences with Parents." Emerging Adulthood

5(5):337-50. doi: https://doi.org/10.1177/2167696816676583.

Fingerman, Karen L., Kyungmin Kim, Eden M. Davis, Frank F. Furstenberg, Kira S. Birditt and

Steven H. Zarit. 2015. "“I'll Give You the World”: Socioeconomic Differences in

Parental Support of Adult Children." Journal of Marriage and Family 77(4):844-65. doi:

10.1111/jomf.12204.

Flegal, Katherine M, Deanna Kruszon-Moran, Margaret D Carroll, Cheryl D Fryar and Cynthia

L Ogden. 2016. "Trends in Obesity among Adults in the United States, 2005 to 2014."

JAMA 315(21):2284-91.

Fry, R. Pew Research Center. 2013. "A Record 21.6 Million In 2012: A Rising Share of Young

Adults Live in Their Parents’ Home." Washington, DC: Pew Research Center: August 1.

http://www.pewsocialtrends.org/2013/08/01/a-rising-share-of-young-adults-live-in-their-

parents-home/

Fry, R. Pew Research Center. 2016. "For First Time in Modern Era, Living With Parents Edges

Out Other Living Arrangements for 18- to 34-Year-Olds." Washington, DC: Pew

Research Center: May 24. http://www.pewsocialtrends.org/2016/05/24/for-first-time-in-

modern-era-living-with-parents-edges-out-other-living-arrangements-for-18-to-34-year-

olds/

Furstenberg, Frank F, Rubén G Rumbaut and Richard A Settersten. 2005. "On the Frontier of

Adulthood: Emerging Themes and New Directions." On the frontier of adulthood:

Theory, research, and public policy:3-25.

Furstenberg, Frank F. 2010. "On a New Schedule: Transitions to Adulthood and Family

Change." The Future of Children 20(1):67-87. doi: 10.1353/foc.0.0038.

Goldscheider, Francis and Calvin Goldscheider. 1999. The Changing Transition to Adulthood,

Vol. 17. Thousand Oaks, California: Sage.

Gorber, S Connor, M Tremblay, David Moher and B Gorber. 2007. "A Comparison of Direct Vs.

Self‐Report Measures for Assessing Height, Weight and Body Mass Index: A

Systematic Review." Obesity reviews 8(4):307-26. doi: 10.1111/j.1467-

789X.2007.00347.x.

Greenfield, Emily A and Nadine F Marks. 2006. "Linked Lives: Adult Children’s Problems

and Their Parents’ Psychological and Relational Well‐Being." Journal of Marriage

and Family 68(2):442-54. doi: 10.1111/j.1741-3737.2006.00263.x.

Greeno, Catherine G and Rena R Wing. 1994. "Stress-Induced Eating." Psychological Bulletin

115(3):444-64.

Grossman, Lev. 2005. "Grow Up? Not So Fast." Time Magazine.

Hoffman, Saul D. 2008. "Kids having kids: Economic costs & social consequences of teen

pregnancy." The Urban Insitute.

Holowko, Natalie, Mark Jones, Ilona Koupil, Leigh Tooth and Gita Mishra. 2016. "Combined

Effect of Education and Reproductive History on Weight Trajectories of Young

Australian Women: A Longitudinal Study." Obesity 24(10):2224-31. doi:

10.1002/oby.21610.

Houle, Jason N and Cody Warner. 2017. "Into the Red and Back to the Nest? Student Debt,

College Completion, and Returning to the Parental Home among Young Adults."

Sociology of Education 90(1):89-108. doi: https://doi.org/10.1177/0038040716685873.

Kalmijn, Matthijs and Paul M De Graaf. 2012. "Life Course Changes of Children and Well‐

Being of Parents." Journal of Marriage and Family 74(2):269-80. doi: 10.1111/j.1741-

3737.2012.00961.x.

Kaplan, Greg. 2012. "Moving Back Home: Insurance against Labor Market Risk." Journal of

Political Economy 120(3):446-512.

Lee, Joyce M, Subrahmanyam Pilli, Achamyeleh Gebremariam, Carla C Keirns, Matthew M

Davis, Sandeep Vijan, Gary L Freed, William H Herman and James G Gurney. 2010.

"Getting Heavier, Younger: Trajectories of Obesity over the Life Course." International

Journal of Obesity 34(4):614-23.

Link, Bruce G and Joanne C Phelan. 2000. "Evaluating the Fundamental Cause Explanation for

Social Disparities in Health." Handbook of medical sociology 5:33-46.

Lorant, Vincent, Denise Deliège, William Eaton, Annie Robert, Pierre Philippot and Marc

Ansseau. 2003. "Socioeconomic Inequalities in Depression: A Meta-Analysis." American

Journal of Epidemiology 157(2):98-112.

Maroto, Michelle. 2017. "When the Kids Live at Home: Coresidence, Parental Assets, and

Economic Insecurity." Journal of Marriage and Family. doi: 10.1111/jomf.12407.

Milkie, Melissa A, Alex Bierman and Scott Schieman. 2008. "How Adult Children Influence

Older Parents' Mental Health: Integrating Stress-Process and Life-Course Perspectives."

Social Psychology Quarterly 71(1):86-105. doi: 10.1177/019027250807100109.

Newman, Katherine S. 2012. The Accordion Family: Boomerang Kids, Anxious Parents, and the

Private Toll of Global Competition. Boston: Beacon Press.

Nomaguchi, Kei M and Suzanne M Bianchi. 2004. "Exercise Time: Gender Differences in the

Effects of Marriage, Parenthood, and Employment." Journal of Marriage and Family

66(2):413-30. doi: 10.1111/j.1741-3737.2004.00029.x.

Parker, Kim. 2012. "The Boomerang Generation: Feeling Okay about Living with Mom and

Dad." Washington, DC: Pew Research Center Social and Demographic Trends project,

March 15. http://www.pewsocialtrends.org/2012/03/15/the-boomerang-generation/

Payne, Krista K. 2011. "On the Road to Adulthood: Leaving the Parental Home." National

Center for Family & Marriage Research: FP-11-02.

http://ncfmr.bgsu.edu/pdf/family_profies/file98800.pdf.

Payne, Krista K. and Jennifer Copp. 2013. "Young Adults in the Parental Home and the Great

Recession." National Center for Family and Marriage Research: FP-13-07.

http://ncfmr.bgsu.edu/pdf/family_profiles/file126564.pdf.

Pearlin, Leonard I. 1999. "The Stress Process Revisited." Pp. 395-415 in Handbook of the

Sociology of Mental Health: Springer.

Pearlin, Leonard I., Scott Schieman, Elena M. Fazio, and Stephen C. Meersman. 2005. "Stress,

health, and the life course: Some conceptual perspectives." Journal of health and Social

Behavior 46(2): 205-219.

Royston, Patrick. 2005. "Multiple Imputation of Missing Values: Update of Ice." Stata Journal

5(4):527-36.

Sandberg‐Thoma, Sara E, Anastasia R Snyder and Bohyun Joy Jang. 2015. "Exiting and

Returning to the Parental Home for Boomerang Kids." Journal of Marriage and Family

77(3):806-18. doi: 10.1111/jomf.12183.

Sassler, Sharon, Desiree Ciambrone and Gaelan Benway. 2008. "Are They Really Mama’s

Boys/Daddy’s Girls? The Negotiation of Adulthood Upon Returning to the Parental

Home." Pp. 670-98 in Sociological Forum, Vol. 23: Wiley Online Library.

Schafer, Joseph L. 1999. "Multiple Imputation: A Primer." Statistical methods in medical

research 8(1):3-15.

Seiffe-Krenke, Inge. 2006. "Leaving Home or Still in the Nest? Parent-Child Relationships and

Psychological Health as Predictors of Different Leaving Home Patterns." Developmental

psychology 42(5):864-76. doi: http://dx.doi.org/10.1037/0012-1649.42.5.864.

Settersten Jr, Richard A and Barbara Ray. 2010. "What's Going on with Young People Today?

The Long and Twisting Path to Adulthood." The Future of Children 20(1):19-41. doi:

10.1353/foc.0.0044.

Settersten, Richard A. 1998. "A Time to Leave Home and a Time Never to Return? Age

Constraints on the Living Arrangements of Young Adults." Social Forces 76(4):1373-

400. doi: https://doi.org/10.1093/sf/76.4.1373.

Smits, Annika, Ruben I Van Gaalen and Clara H Mulder. 2010. "Parent–Child Coresidence:

Who Moves in with Whom and for Whose Needs?". Journal of Marriage and Family

72(4):1022-33. doi: 10.1111/j.1741-3737.2010.00746.x.

Umberson, Debra, Hui Liu, John Mirowsky and Corinne Reczek. 2011. "Parenthood and

Trajectories of Change in Body Weight over the Life Course." Social science & medicine

73(9):1323-31. doi: 10.1016/j.socscimed.2011.08.014.

Van Buuren, Stef. 2012. Flexible Imputation of Missing Data. Boca Raton, Florida: CRC press.

Von Hippel, Paul T. 2007. "Regression with Missing Ys: An Improved Strategy for Analyzing

Multiply Imputed Data." Sociological Methodology 37(1):83-117.

Wang, Youfa and May A Beydoun. 2007. "The Obesity Epidemic in the United States—Gender,

Age, Socioeconomic, Racial/Ethnic, and Geographic Characteristics: A Systematic

Review and Meta-Regression Analysis." Epidemiologic reviews 29(1):6-28.

Ward, Russell A and Glenna D Spitze. 2007. "Nestleaving and Coresidence by Young Adult

Children: The Role of Family Relations." Research on Aging 29(3):257-77. doi:

https://doi.org/10.1177/0164027506298225.

Wardle, Jane and J Griffith. 2001. "Socioeconomic Status and Weight Control Practices in

British Adults." Journal of Epidemiology and Community Health 55(3):185-90.

Table 4. Sociodemographic and Health Characteristics of the Sample (Imputed) Whole sample Gone for good Boomerang kids Never Left Mean / % SD Mean / % SD Mean / % SD Mean / % SD Mother’s BMI score 29.31 6.90 28.96 6.84 29.94*** 6.94 29.74*** 6.97 Children's Residential Pattern 18-30 the Gone for good 60.86 / / / Boomerang kids 22.37 / / / the Never left 16.77 / / /

Children's Characteristics Age 22.96 3.55 23.18 3.55 23.41*** 3.58 21.54*** 3.11 Sex 51.62 Female 54.68 48.74*** 44.36***

Marital Status 66.87 Never Married 58.35 67.76 96.57 Married or cohabiting 30.10 38.53 27.8*** 2.58*** Previously married 3.03 3.12 4.44+ 0.85***

Parental status 37.55 Yes (is a parent) 41.36 43.69+ 15.52***

Educational Attainment (time-varying)

Below High School 17.75 16.28 21.23 18.43

High School 67.67 66.87 66.47*** 72.14 14.59 College or above 16.85 12.3*** 9.43*** Economic status

Neither in school or working 1.34 1.39 1.51 0.95 In school 5.57 4.58 4.61 10.42*** Working 65.14 66.68 72.06 50.30 In school and working 27.95 27.35 21.82* 38.33*** Mother's Characteristics

Age 46.73 4.82 46.43 4.78 46.15* 4.67 48.62*** 4.70 Race/Ethnicity NonHispanic White 42.24 46.26 33.96 38.70 NonHispanic Black 35.49 33.42 41.33*** 35.21** 22.27 Hispanic 20.32 24.7*** 26.1***

Region of Residence

Northeast 12.57 10.91 12.04 19.30

North Central 25.11 27.50 21.95** 20.63*** 43.09 South 42.60 43.15 40.75*** 19.24 West 18.99 19.86 19.32***

Marital Status 11.07 Never Married 10.09 13.71 11.11 54.68 Married or Cohabiting 55.12 48.92*** 60.77 Previously married (Separated, 34.25 Divorced, Widowed) 34.78 37.38* 28.12** 10.95 Logged family income-(inflation adjusted) 10.96 10.84*** 11.07*** 47998.64 Median income (inflation adjusted) 48855.19 40281.10 57177.18 Family size 3.25 1.53 3.07 1.51 3.28*** 1.58 3.9*** 1.36 Highest Grade Completed 12.87 2.50 12.92 2.49 12.53*** 2.35 13.13* 2.69 Health Insurance 64.70 Private 64.82 60.83 69.46 Public 15.48 15.07 17.17** 14.73 Self-pay 19.82 20.11 22.01* 15.82*** Smoke Status Non-smoker 46.76 45.14 46.51 52.94 24.17 24.68 Ex-smoker 22.14+ 25.04* Current smoker 29.08 30.18 31.36 22.02*** N (Person-Years) 28,010 17,046 6,266 4,698 73

N (NLSY79- child-mother pairs) 7,220 4,108 1,263 1,849 Source: NLSY79 and 79 Young Adults Survey (1994-2014). Note: The bivariate differences between gone-for-good and boomerangers or never-left are included in the table. + p<.10 * p<.05 ** p<.01 *** p<.001

Table 5. Results from Multilevel (Growth Curve) Regression Models with Random Coefficients Predicting Mom's BMI by Young Adult's Residential Biography, 1994-2014 Model 1 Model 2 Model 3 b SE b SE b SE Fixed effects Intercept 27.403*** 0.174 25.407*** 0.357 27.231*** 1.050 Children's Residential Pattern since Age 18 the Gone for good ref ref ref the Never left 1.144** 0.352 1.539*** 0.359 1.180*** 0.354 Boomerangers 1.059*** 0.277 1.018*** 0.277 0.674* 0.274 Age (centered at 35) 0.130*** 0.009 -0.010 0.019 0.052** 0.019 Children's Residential Pattern*age

the Gone for good*age ref ref ref the Never left*age -0.077*** 0.021 -0.056** 0.021 -0.062** 0.021 Boomerangers*age 0.002 0.015 -0.004 0.015 -0.009 0.015 Children's Characteristics

Age 0.151*** 0.020 0.092*** 0.019 Sex (ref: Male)

Female 0.332* 0.155 0.250+ 0.150 Marital Status (ref: Never Married)

Married or cohabiting -0.027 0.054 0.034 0.053

Previously married -0.007 0.137 0.033 0.136

Parental status (ref: Not a parent)

Yes (is a parent) 0.101 0.072 0.031 0.072 Educational Attainment (ref: Below High School)

High School -0.079 0.078 -0.028 0.078

College or above -0.198+ 0.116 -0.085 0.116 Economic Status (ref: Neither in school nor working) In school -0.047 0.202 0.051 0.205 Working 0.081 0.198 0.206 0.201 In school and working 0.060 0.196 0.200 0.199 Mother's Characteristics Race/Ethnicity (ref: NonHispanic White) NonHispanic Black 3.437*** 0.288

Hispanic 1.620*** 0.345 Logged Family income (inflation adjusted) -0.125* 0.058

Family size 0.081* 0.035 Marital Status (ref: Married or Cohabiting)

Never Married 0.119 0.285

Previously married -0.380** 0.123

Highest Grade Completed -0.113* 0.052

Health Insurance (ref: Private)

Public 0.382** 0.121

Self-pay -0.078 0.084 Region of Residence (ref: Northeast)

North Central -0.091 0.282 South 0.285 0.254 West 0.120 0.321 Smoke Status (ref: Non-smoker) Ex-smoker -0.058 0.145 Current smoker -0.734*** 0.170

Random effects Level 1 residual 4.837*** 4.842*** 4.830*** Level 2 age 0.085*** 0.085*** 0.082*** Level 2 intercept 48.659*** 48.626*** 45.718*** N (Person-Years) 28,010 28,010 28,010 N (NLSY79-Young adult child-mother pairs) 7,220 7,220 7,220 Source: NLSY79 and 79 Young Adults Survey (1994-2014). + p<.10 * p<.05 ** p<.01 *** p<.001

Figure 2. Predicted BMI by Coresidential Pattern (Table 5- Model 1)

Figure 3. Predicted BMI by Coresidential Pattern (Table 5- Model 3)

Chapter 4. Mother-child relationship quality and mental health at midlife

Intergenerational relationships between parents and children are highly associated with both generations’ wellbeing across the life course (Umberson, Pudrovska, & Reczek, 2010). Existing scholarship has established that positive qualities in parent-child ties are associated with higher parental wellbeing, whereas negative emotions in parent-child ties are associated with lower parental wellbeing (Merz, Schulze, & Schuengel, 2010; Milkie, Bierman, & Schieman, 2008).

Increasing evidence suggests that intergenerational ambivalence, the simultaneous presence of positive and negative sentiments in parent-child relationships, is also highly related to parental wellbeing, especially mental health (Fingerman, Pitzer, Lefkowitz, Birditt, & Mroczek, 2008;

Gilligan, Suitor, & Pillemer, 2015; Kiecolt, Blieszner, & Savla, 2011; Suitor, Gilligan, &

Pillemer, 2011; Tighe, Birditt, & Antonucci, 2016). Yet, three significant gaps exist in the literature that prevents a full understanding of intergenerational ties and well-being.

First, the majority of current research focuses on correlates of relationship quality with adult children for older parents (above 60) (e.g., Gilligan et al., 2015; Lee & Szinovacz, 2016;

Lendon, Silverstein, & Giarrusso, 2014; Pillemer & Suitor, 2002; Thomas et al., 2017). Much fewer studies focus on midlife parents exclusively (for exceptions, see Kiecolt, Blieszner, &

Salvla, 2011). Notably, parenthood at midlife (40-60) is particularly demanding as adolescent and emerging young adult children are undergoing various life transitions and redefining their social relationships, which can be connected with heightened conflict and ambivalence toward

parents and further influence parents’ mental health. Second, intergenerational ties are distinct from many other one-to-one family relationships (e.g., marital relationship), as most parents in the U.S. have more than one child (Pew Research Center, 2018). Parental wellbeing is likely connected with all parent-child relationships in a family (Fingerman et al., 2012; Tosi & Grundy,

2018), however, the predominant approach is to consider how individual parent-child dyad matters for parent’s mental wellbeing, not how individual parent-child ties in constellation with one another shape parental well-being (for exceptions, see Ward, 2008). Third, most of the current studies, especially those on parent-child ambivalence, tend to use cross-sectional data collected from a regional sample to examine the relationship between midlife parents’ mental health and intergenerational quality (e.g., Fingerman et al., 2008; Gilligan et al., 2015). Reverse causality needs to be addressed as poor mental wellbeing can lead to strains in social ties.

In order to address these gaps, this study draws on the family and health literature, especially ambivalence theory, and asks: How do intergenerational relationship quality matter for mothers’ mental health at midlife, considering mother’s relationship quality with multiple children and paying particular attention to mother-child ambivalence? More specifically, this study analyzes how multiple adolescent and young adult children’s reports of relationship quality with their mother, categorized by uniformly close, collective ambivalent, and uniformly unclose, are associated with mother’s mental health at age 50? Further, how are changes in intergenerational relationship quality associated with changes in mother’s mental health over time? Using a multigenerational, national sample of mothers and their adolescent and young adult children, this research adds nuances to understanding complex intergenerational

relationships at mothers’ mid-life, with further implications for improving both mother’s and family’s wellbeing.

Theoretical and Empirical Background

Midlife mothers’ and intergenerational relationship quality mental wellbeing

Family and health scholars have long recognized that positive sentiments in intergenerational relationships are connected with parents’ better mental wellbeing whereas negative components in parent-child ties are associated with worse mental health among parents (Koropeckyj-Cox,

2002; Merz et al., 2010; Milkie et al., 2008; Ward, 2008). This is typically theorized to operate via social support and social strain. Social support serves to buffer stress from diminishing mental wellbeing (Thoits, 2011), and social support from children is found to be associated with increased of life satisfaction and decreased loneliness, which all contributes to better mental health among parents (Chen & Silverstein, 2000; Lang & Schutze, 2002; Long & Martin, 2000).

In contrast, strained intergenerational relationships are linked with parents’ sense of loneliness, lower levels of happiness, and even negative treatment toward parents, which all lead to deteriorating mental health (Koropeckyj-Cox, 2002; Milkie et al., 2008). Compared to social support, strain in parent-child relationships tends to have more prominent and lasting effects on parental wellbeing (Umberson, 1992; Baumeister et al., 2001; Charles, 2010).

In addition to focusing on either intergenerational support or strain, over the past two decades, a burgeoning literature has identified the fundamental interplay between positive and negative elements in parent-child ties—what is known as ambivalence (Connidis & McMullin,

2002; Pillemer & Lüscher, 2004; Pillemer & Suitor, 2002). Current work suggests that ambivalent ties have detrimental effects on mental wellbeing (Suitor et al., 2011; Lee &

Szinovacz, 2016), in part because the presence of both positive and negative sentiments is unpredictable and inherently stressful (Uchino, Holt-Lunstad, Smith, & Bloor, 2004). More recent evidence finds higher parent-child ambivalence scores to be linked with higher levels of psychological distress among older adults and their grown children (Fingerman et al., 2008;

Gilligan et al., 2015; Lee & Szinovacz, 2016).

Existing literature casts most attention on the effects of support, strain, and intergenerational ambivalence on parents of older parents’ wellbeing, yet much fewer studies focus on parents at midlife (Fingerman, 2017). Such omission needs to be addressed as increasing evidence suggests that parents at midlife are exposed to intensified intergenerational support, strain, and ambivalence as children are entering adolescence and transitioning into emerging young adulthood (Kiecolt et al., 2011; Tighe et al., 2016). At this life course stage, adolescent and emerging young adult children establish independence from parents and experience important life event transitions such as enrollment in another educational institution

(e.g., high school or college), departure from parental home, participation in romantic relationships, and transition to employment. During this time, parent-child conflicts increases but intergenerational support, especially those from parents to children, persist and may even increase (Aquilino, 2006; Noack & Buhl, 2004; Fingerman et al., 2012). Meanwhile, intergenerational ambivalence likely intensifies between emerging young adults and their midlife parents (Fingerman et al., 2016). Furthermore, midlife mothers’ health is more susceptible to intergenerational conflict and ambivalence as women experience much more parenting stress than men (Umberson et al., 2010). This study amends the research gap by studying the association between midlife mothers’ mental health and intergenerational relationship,

considering both positive and negative ties and with particular focus on intergenerational ambivalence.

Rethinking intergenerational relationship quality: One mother and multiple children

Intergenerational relationship is distinct in that it can account for multiple relationship dynamics between mothers and children. Most American parents have more than one child. In 2014, about two thirds of U.S. women aged 40-44 had two or more children (Pew Research Center, 2015).

Research consistently shows that mother and her offspring’s lives are linked (Elder, Johnson, &

Crosnoe, 2003), wherein wellbeing is linked with her interactions with multiple children and life events of multiple children. For example, Fingerman and colleagues find that parental wellbeing is associated with both the successes and failures of all grown children (Fingerman et al., 2012).

Specifically, parents who had two children encountering problems experienced poorer mental health than parents with one child suffering problems. More stressors from multiple children thus take a heavier toll on mother’s wellbeing. Another study suggests that mothers with no frequent contact with all children had more depressive symptoms than mothers with frequent contact with at least one child (Tosi & Grundy, 2018). In other words, frequent contact with at least one child may greatly mitigate the negative effects of an estranged relationship with another child on mother’s mental health. These two studies suggest that considering mother’s relationship with only one child likely leads to biased conclusions about mother’s mental wellbeing.

In terms of intergenerational relationship quality and mental health: when one mother receives health benefits from her supportive and affectionate relationship with one child, the positive effects of this relationship on wellbeing may be reduced due to the strained relationship with another child. Ward (2008) finds that the presence of both negative and positive feelings

with different grown children in the family network, or “collective ambivalence,” to be associated with lower levels of parental happiness, especially among mothers. Yet, how mother’s mental wellbeing is influenced by the simultaneous presence of positive and negative sentiments from multiple children in the family, remains largely unknown. Additionally, prior work has not sufficiently addressed how a mother’s positive ties with all children or a mother’s negative ties with all children differently explain maternal wellbeing. Contributing to a burgeoning line of research efforts that considers multiple members in a family relationship, this study conceptualizes intergenerational, support, strain, and ambivalence between one midlife mother and her multiple children in one scenario.

Following interdependence theory, which posits that one party’s wellbeing can be influenced by the other’s appraisals of the relationship (Fingerman et al., 2008; Rusbult &

VanLange, 2003), this study examines multiple children’s assessment of parent-child ties in relation to mothers’ psychological wellbeing. In the case of closeness, all children would report high levels of support are characterized as uniformly close; all children would report high levels of strain are characterized as uniformly unclose; in the case of collective ambivalence, one mother has at least one child reporting overall positive feelings toward mother whereas the other children reported less positive ties with mother. Previous work, especially those that measured individual ambivalence from Griffin’s formula, finds that higher ambivalence scores are associated with worse mental health (Fingerman et al., 2008; Gilligan et al, 2015; Lee &

Szinovacz, 2016). This previous research suggests that individuals in an ambivalent relationship likely have more distress than those in either an overall supportive or an overall strained relationship. In line with this work, I hypothesize that mothers with collective ambivalent

relationship with multiple children experience more psychological distress than mothers whose adolescent and young adult children reporting either uniformly close or uniformly unclose ties with mother. Further, guided by research that demonstrates positive effects of supportive ties and detrimental consequences of negative ties on wellbeing (Merz et al., 2010; Milkie et al., 2008;

Umberson et al., 2010), I expect mothers with uniformly close relationship with multiple children experience less psychological distress than mothers with uniformly unclose ties with children.

Intergenerational Relationship Quality and Mother’s Mental Health: a longitudinal study design

In addition to examining the relationship between mother-multiple children ties and maternal mental health at midlife, this study further improves prior literature by using a longitudinal research design. Most of the prior research relies on cross-sectional data (e.g., Fingerman et al.,

2008; Gilligan et al., 2015), which invites bias pertaining to reverse causality. Poor mental health such as symptoms of anxiety is found to be associated with declining relationship satisfaction over time (Rehman et al., 2015). Children’s concerns about parental health likely increase intergenerational ambivalence (Willson, Shuey, & Elder, 2003). Notably, children’s reported worries about parental health are common when parents are at midlife and are still in good health

(Hay, Fingerman, & Lefkowitz, 2008). In line with this research, this study examines how parent-child relationship quality is associated with mother’s midlife mental health over time.

The dimension of changes in social ties and mental health is particularly relevant for adults at midlife, as midlife individuals are exposed to various transitions in their own lives as well their children’s, which likely set them up for wellbeing fluctuations (Gilligan et al., 2015).

Prior work posits that midlife adults’ emotions are more sensitive to negative sentiments such as family conflicts compared to older adults (Blanchard-Fields et al. 2004; Charles & Carstensen,

2007). This study thus compares transitions into highly strained parent-child relationships with positive transitions or no change in parent-child ties, in association with mother’s mental health changes over time. The highly strained relationships include uniformly unclose and ambivalent ties, because both contain negative emotions. Further, in the relationship between one mom and multiple children, an exposure model suggests that parenting stress from one child can lead to worse mental health among parents; a cumulative stress and reward model argues that parenting stress from multiple children will lead to incrementally higher mental distress (Fingerman et al.,

2012). Taken together, this study analyzes how negative changes in intergenerational ties, including changes into ambivalence relationship and declines in mother-child relationship, from both one child and all children combined, are associated with mothers’ mental health.

Other Factors Associated with Intergenerational Ties and Mental Health

This study considers other characteristics that are associated with intergenerational ties and psychological wellbeing. These characteristics include race, marital status, employment status, educational level, family income, family size, and insurance type, and self-rated health.

Specifically, I account for race because African American mothers’ wellbeing is more responsive to qualities of intergenerational relationship (Milkie et al., 2008). Mother’s marital status matters because unmarried parents tend to value their ties with children more than other social ties

(Fingerman, Hays, & Birditt, 2004). Socioeconomic characters captured by educational level, family income, and health insurance type are among the most effective predictors of wellbeing

(Clarke et al., 2011; Farmer & Ferraro, 2005). Poorer health and unemployment are negatively

associated with mental wellbeing (Clarke et al., 2011; Shieman & Glavein, 2011). Family size is not only related to the number of coresident children a mother currently has, but also helps to account for the dilution of family resources when included in a model along with measures of family income. In addition, mother’s mental wellbeing likely predicts her relationship with children. In order to reduce the effects of reverse causation, this study employs a longitudinal study design and control for mother’s prior psychological wellbeing. All of these related variables are considered throughout the analysis.

The present study

Drawing on the family and health literature, this study examines the association between intergenerational relationship quality and mothers’ mental health at midlife, considering one mother’s relationship with multiple adolescent and young adult children. The research endeavor is guided by two main research questions: first, how is mother’s mental health at age 50 predicted by mother’s relationship with multiple adolescent and young adult children, measured by uniformly close, collective ambivalent (some children report close relationship with mom whereas others do not), and uniformly unclose? Second, how are the negative transitions in mother-child ties, including increases in ambivalent feelings and declines in mother-child closeness, associated with changes in maternal mental health? This paper addresses multiple gaps in the literature on intergenerational ties and wellbeing by elucidating the complex intergenerational ties at a particularly stressful yet scarcely researched life course stage. This study also contributes to the literature by using data from a national sample and employing a longitudinal design to account for potential selection effects, details of which are depicted in the following section.

Data and Methods

Data

Data are from National Longitudinal Survey of Youth-1979, an ongoing panel study. In 1979, a nationally representative sample of 12,686 men and women aged between 14 and 22 were interviewed. Follow-up surveys were conducted annually until 1994 and biannually until 2014, the most recent wave available. Beginning in 2008, a 50-and over health module was launched to collect detailed information on individual wellbeing including mental health. Each respondent participates in the 50’s health module once, usually right after they turned 50. The survey years of this health module currently spanned between 2008 and 2014. In 1986, a separate survey,

NLSY79-Children (NLSY79-CH) was administered to collect a wide range of social, economic, developmental, and health information on children born to NLSY79 women. These children form the basis of the NLSY79 – Young Adult Survey (NLSY79-YA), which officially began in

1994. The NLSY79 children, once turned 14, would participate in NLSY79-YA survey, conducted biannually until 2014.

This study employed data from NLSY79 women and data from their adolescent and young adult children in NLSY79-YA. By 2014, 3,162 NLSY79 mothers, who had at least one child aged 14 or older, had participated in the 50’s health module. 24 mothers were dropped due to missing values on the dependent variables. The analysis used two waves of data – one that contains the 50’s health module (used as Wave 2) and the prior wave (Wave 1). For instance, if a respondent participated in the 50’s health module in 2010 (Wave 2), her answers from year 2008

(Wave 1), if available, would also be used in the analysis. This restriction criteria result in a final sample of 2,534 midlife women.

Measures

Dependent Variables: Psychological Distress

Psychological Distress is measured with a 7-item version of the Center for Epidemiological

Studies Depression Scale (CES-D), from the 50’s health module. Respondents reported frequencies in the past week that they experienced the following symptoms: (1) I did not feel like eating; my appetite was poor; (2) I had trouble keeping my mind on what I was doing; (3) I felt depressed; (4) I felt that everything I did was an effort; (5) My sleep was restless;(6) I felt sad;

(7) I could not get “going”. For each item, respondents selected answers from rarely or none of the time, some or little of the time, occasional or a moderate amount of the time, and most of all of the time. NLSY researchers combined the seven answers into a single scale, which I log transformed to address the positive skew. The transformed scores were further multiplied by 100 so as to increase interpretability of the results (see Clarke et al., 2011 for a similar approach).

Higher CESD scores indicate greater psychological distress (Cronbach’s alpha = 0.83).

Independent Variables: Intergenerational Relationship at Wave 1

Children’s responses from NLSY79-YA were used to construct the independent variables.

Starting in 2000, respondents in NLSY79-YA reported how close they felt toward mother, the answer options of which were extremely close, quite close, fairly close, and not very close. More than 50% of the answers concentrated in “extremely close”, therefore, I combined the three less close options into one single category of “not so close”. Taking answers from multiple children to one mom from wave 1, I generated a variable of intergenerational relationship with three categories: uniformly close, uniformly unclose, and ambivalent wherein at least one child described relationship with mom as not so close but others reported close ties with mom.

Independent Variables: Negative Changes in Intergenerational Ties between Wave 1 & Wave 2

Using the closeness measures reported by multiple children, I generated four dummy variables to capture the negative changes in intergenerational ties between Wave 1 and Wave 2. The first two measures emphasized on intergenerational ambivalence: (1) mothers’ relationship with all adolescent and young adult children changed from non-ambivalent (uniformly close or uniformly unclose) to ambivalent; (2) intergenerational ties changed from uniformly close to ambivalent.

The other two dummies took a broader focus on the decline in relationship quality: (3) at least one child felt less close with mother at Wave 2; (4) the overall intergenerational relationship changed from uniformly close to uniformly unclose or ambivalent at Wave 2.

Control Variables

The covariates include race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic), marital status (Never married, married, other), employment status (employed, other), highest grade completed, family income (adjusted for inflation and reported in 2014 dollars, and log transformed to account for positive skew of the original distribution), family size (top coded at six), health insurance coverage (private, public, self-pay), and self-rated health (fair/poor, good, very good, excellent). These variables are all measured at Wave 1. Mother’s prior psychological distress are measured by the 7-item of CES-D scales from the 40’s health module, administered once to NLSY79 respondents when they turned 40 (1998-2006).

Analytical Strategy

This study used Ordinary Least Squares regression models (OLS) with lagged dependent variables to examine the association between intergenerational relationship quality and mother’s mental health at age 50. Specifically, I studied the role of mother-children ties at Wave 1 in

predicting mother’s mental health at Wave 2. I also examined how the negative changes in intergenerational ties between Wave 1 & Wave 2 explain variation in mother’s mental wellbeing at Wave 2. Mother’s psychological distress reported at around age 40 was used as a proxy for the mental wellbeing at Wave 1 and included in all models. The inclusion of previous health conditions in the model is shown to reduce effects of reverse causality (Johnson, 2005).

Coefficients in lagged dependent variables can thus be taken as the average change in the dependent variable between Wave 1 & Wave 2 in relation to the estimated effect from the independent variable (Finkel, 1995; Williams, Sassler, & Nocholson, 2008). Additionally, Wald tests were used to examine differences between coefficients within the same model. Likelihood- ratio tests were conducted to compare model fits across nested models. Multiple imputation with chained equation was employed to maximize data usage (Royston, 2005; Van Buuren, 2012;

White, Royston & Wood, 2011). All the descriptive and multivariate analyses are estimated using the mi command in Stata/MP 15.0 (StataCorp, 2017).

Results

Descriptive Results

Table 6 displays descriptive information for the variables used in the analysis. On average, mothers in the sample had a distress score of 130.38 at Wave 2 (around age 50). At Wave 1

(around age 48), about 38% of individuals had uniformly close relationship with their adolescent and young adult children, 30% of mothers had ambivalent relationship with her children, and

32% of mothers had uniformly unclose relationship with her children. Respondents had an average of 1.87 adolescent or young adult children at Wave 1 and 60% of mothers had two or more children.

Results from regression with no covariates (not displayed in Table 6) show that at age 50, mothers with uniformly close intergenerational relationship reported an average distress score of

125.56, which is lower than that of mothers with ambivalent intergenerational ties (b= -9.40, p <

0.05) and that of mothers with uniformly unclose parent-children ties (b= -6.08, p = 0.17). The breakdown of mother’s overall relationship with multiple children at Wave 2 remains similar to that of Wave 1 (37% uniformly close, 31% ambivalent, 32% uniformly unclose; not shown in

Table 6), yet some changes were present. Between Wave 1 and Wave 2, 5% of mothers transitioned into an ambivalent relationship with children from a uniformly unclose relationship

(n = 124), and another 5% of mothers transitioned into an ambivalent intergenerational relationship from a uniformly close relationship (n = 139). In terms of all types of relationship decline, statistics from Table 6 shows that 17% of mothers had at least one child reporting declining relationship in the two-year span and 31% of mothers experienced an overall decline in intergenerational relationship with multiple children, transitioning out of a uniformly close relationship.

Multivariate Results

Regression models predicting change in psychological distress at Wave 2 are shown in Table 7 and Table 8. Model 1 From Table 7 estimates the association between Wave 1 intergenerational ties and change in mental health. Model 1 through 4 from Table 8 estimate effects of transitions in intergenerational ties. Model 1 indicates that compared to mothers with uniformly close relationship with children, mothers with ambivalent relationship with children reported higher levels of psychological distress (b= 9.14, p < 0.05), controlling for all the sociodemographic covariates. Distress score of mothers with uniformly unclose relationship with children was not

statistically different from mothers in the two other groups. Sensitivity analysis (see Table 10 in

Appendix) shows that results in Model 1 were primarily driven by mothers with fewer children

(two or less). About 40% of mothers in the sample had only one adolescent or young adult child.

Regression results without this group of mothers did not reveal differences in mental wellbeing by mother-child relationship quality. Model 1 & 2 from Table 8 suggest that transitions into an ambivalent intergenerational relationship, regardless of its relationship status at Wave 1, were not predictive of change in psychological distress. Model 3 shows that when at least one child reported declining relationship with mother, mother’s psychological distress would increase by about 7.66 unit (p < 0.05). The last model, Model 4 finds no association between an over decline in mother-children relationship and mother’s mental health.

This study tested other intergenerational dynamics in relation to midlife mother’s mental health (see Table 11 in Appendix). Increases in mother-children closeness, considering the relationship with both one child and multiple children, were not associated with mother’s psychological distress. Children’s reported closeness with mother was not predictive of mother’s psychological distress at age 40, in part due to a smaller sample size of adolescent and young adult children. Lastly, a longer time span of change in mother-children closeness could be theoretically relevant for women’s midlife health. I used group based trajectory models (Jones and Nagin, 2007) and found three main mother-child closeness trajectories across mother’s early midlife (40-50). These three trajectories (extremely close ties; close ties, and unclose ties) displayed little change over time and could not explain variations in mother’s midlife mental health.

Discussion

Intergenerational relationship quality is highly influential for parents’ wellbeing across the life course. However, substantial gaps remain, especially in assessing the role that relationships with multiple children play in midlife parents’ wellbeing. Moreover, collective intergenerational ambivalence, the presence of both positive and negative emotions from multiple children, has received insufficient research attention given its theorized impacts on maternal well-being.

Additionally, the majority of existing literature uses a cross-sectional design and is thus limited in demonstrating the effects of intergenerational relationship quality on mental health. This study addresses these research gaps by answering two questions: First, how is mother’s mental health at age 50 associated with her relationship with multiple children? Second, how do negative transitions in mother-child ties matter for changes in mother’s mental health? I used data from a national sample and constructed the main independent variables with multiple children’s responses about mother-child relationships. Models from OLS regression with lagged dependent variables find that midlife mothers in an ambivalent relationship with adolescent and young adult children, as well as mothers with at least one child reporting declining ties with mother over time, were at higher risk of increasing psychological distress at age 50.

This study has three main contributions to the literature. First, I find that collective ambivalence, wherein at least one child reported unclose relationship with mother but other children described close feelings toward mother, was associated with significantly higher psychological distress among mothers, compared to mothers who had uniformly close relationship with multiple children. The mental health effects of mothers with uniformly unclose intergenerational ties were in between that of the uniformly close and collective ambivalence, but

the differences were not statistically significant. These results resonate with prior studies that reported detrimental effects of ambivalent parent-child relationships on parental mental health

(Fingerman et al., 2008; Gilligan et al., 2015; Lee & Scze, 2016; Ward, 2008), with unique contributions. An emerging line of research has extended the concept of an ambivalent relationship between one child and one parent to collective ambivalence, which recognizes both positive ties and negative ties in multiple relationships rather than individual dyads (Ward, 2008;

Ward et al., 2008; Reczek, 2016). Among this work, Ward (2008) examines mother’s mixed feelings toward multiple children – a mother holding positive feelings toward on child and negative feelings toward another simultaneously, and finds this to be associated with less happiness. Reczek (2016) approaches collective ambivalence from a holist perspective by studying the perceived collective ambivalence in a broader family unit, wherein gay and lesbian adults’ experience parent, sibling, “in-law”, and extended kin as having both positive and negative beliefs and behaviors. This study extends the current research by constructing collective ambivalence with accounts from multiple children, and further connects these accounts with mother’s mental wellbeing. In other words, collective ambivalence, defined as multiple family members – multiple children’s perceptions toward mother, is verified to carry significant weight in predicting mother’s wellbeing. This finding complements prior work in demonstrating the relational nature of the ambivalent construct, wherein accounts from multiple family members should be considered to improve our standing in the complex family dynamics (Pillemer &

Suitor, 2008; Reczek, 2016).

Second, findings in the study suggest that mothers’ mental health is highly responsive to intergenerational relationship quality at midlife. A wealth of existing scholarship reveals that

older parents’ wellbeing is influenced by a complexity of intergenerational ties, including intergenerational support and strain (Thomas et al., 2017), mother’s favoritism toward specific children (Suitor et al., 2017), parental dissatisfaction (Reczek & Zhang, 2016), and relationship equity (De Jong Gierveld & Dykstra, 2008). Much less research attention has been cast on midlife parents (Fingerman, 2017). In view of the increasing challenges of parenting adolescent and young adult children at midlife as well as the changing parent-child relationship qualities at this life course stage (Fingerman et al., 2012), this study examines the intergenerational relationship qualities and maternal mental health at midlife. Findings suggest that collective ambivalence, rather than a uniformly close or a uniformly unclose relationship, is associated with more psychological stress for midlife mothers. The collective ambivalence between midlife mothers and adolescent and young adult children may set the stage for family dynamics of parents’ favoritism, unbalanced parent-child exchanges, or parental dissatisfaction in later life, with significant implications for the wellbeing of both generations.

Third, the study further explicates the relationship between intergenerational relationship quality and mother’s midlife mental health by shedding light on the effects of negative transitions in mother-child ties. Results suggest that declining relationship reported by at least one child was associated with mother’s elevated distress, but transitions into collective ambivalence from either a uniformly close or a uniformly unclose relationship were not associated with changes in maternal wellbeing. The findings correspond to the literature that suggests stronger effects of negative relationship on wellbeing over time (Larsen, 2009). These results also resonate with an exposure model for the connection between family relationships and health, which posits that parents’ stress from the relationship with one child can result in

decreased wellbeing (Fingerman et al., 2012). The lack of association between transitions into ambivalence and maternal wellbeing may be in part due to the fact that at midlife, relatively fewer mothers experienced increased collective ambivalence between the two points in the study

(see the descriptive statistics at Table 6). Existing research finds that children’s self-report ambivalence peaked at the life course stage of adolescence and emerging young adulthood and then declined at least until later young adulthood (Tighe et al., 2016). In turn, collective ambivalence may change little or slightly declines during mother’s midlife, thus contributing to the null association with maternal wellbeing at this specific life course stage. Taken together, the findings about negative changes and mother’s psychological distress complement the literature on intergenerational ties and midlife parents’ wellbeing. These results add nuances by examining changes in intergenerational ties over time and differentiating between multiple types of negative transitions.

Despite the contributions, limitations of this study warrant future study. First, the main independent variable was derived from one question administered to adolescent and young adult children. Commonly used questions on intergenerational support (e.g., how much support one received, how often one felt loved and cared for) and strain (e.g., how often one felt being critical by other family members or the relationship is too demanding) were not included in either NLSY79 or NLSY79-YA (Ward, 2008; Reczek & Zhang, 2016). Future research should make use of data with more comprehensive information on intergenerational relationships from multiple family members to deepen our understanding on mother-multiple children’s ties and wellbeing. Second, this study joins an emerging line of research by using longitudinal data to reduce biases from reverse causality (Kiecolt et al., 2011; Tighe et al., 2016). Yet, the research

design can be improved. For example, the measure of prior psychological distress was not exactly measured at Wave 1; the measure used in the analysis was from mother’s self-rated

CESD scores at age 40. Future research should tackle these potential biases using detailed and timelier measures of health. Third, the main dependent variable, based on the 7-item CESD scores, represented only one dimension of psychological wellbeing. Future work should consider other measures of wellbeing in relationship to mother’s ties with multiple children. Fourth, gender differences in family relationships and wellbeing have been well-documented in prior literature (Gilligan et al., 2017; Tosi & Grundy, 2018), yet how fathers are affected by their ties with multiple children could not be tested due to data limitation. Future work should address this.

Despite limitations, this study extends prior research on intergenerational relationship and maternal wellbeing at midlife. The findings establish that one midlife mother’s mental wellbeing is linked with her relationship with all of her adolescent and young adult children, collectively.

Using ongoing panel data on a contemporary cohort, the results deepen our understanding about midlife mothers who experience increasing parenting demands and responsibilities to adolescent and coming of age children compared to their parents’ generation. Notably, the findings also highlight the importance of considering multiple children’s accounts of relationship with mother.

Future studies should continue to advance the knowledge on family dynamics at midlife, how these dynamics are linked with both generations’ wellbeing across the life course, and how these relationships are conditioned by social contexts such as race, class, and gender.

References

Aquilino, William S. 2006. Family relationships and support systems in emerging adulthood. In

J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st

century (pp. 193–217). Washington, DC: American Psychological Association. Baumeister,

Roy F., Ellen Bratslavsky, Catrin Finkenauer, and Kathleen D. Vohs. 2001. Bad is stronger

than good. Review of General Psychology, 5, 323–370. doi:10.1037//1089-2680.5.4.323.

Blanchard-Fields, Fredda, Renee Stein and Tonya L. Watson. 2004. "Age differences in

emotion-regulation strategies in handling everyday problems." The Journals of Gerontology

Series B: Psychological Sciences and Social Sciences 59(6):P261-P269.

Charles, Susan T. and Laura L. Carstensen. 2007. "Emotion regulation and aging." Handbook of

Emotion Regulation 6:307-327.

Charles, Susan T. 2010. "Strength and vulnerability integration: A model of emotional well-

being across adulthood." Psychological Bulletin 136(6):1068.

Cichy, Kelly E., Lefkowitz, Eva S. and Karen L. Fingerman. 2012. "Conflict engagement and

conflict disengagement during interactions between adults and their parents." Journals of

Gerontology Series B: Psychological Sciences and Social Sciences 68(1):31-40.

Clarke, Philippa, Victor Marshall, James House and Paula Lantz. 2011. "The social structuring

of mental health over the adult life course: advancing theory in the sociology of

aging."Social Forces 89(4):1287-1313.

Connidis, Ingrid A. and Julie A. McMullin. 2002. "Ambivalence, family ties, and doing

sociology." Journal of Marriage and Family 64(3):594-601.

De Jong Gierveld, Jenny and Pearl A. Dykstra. 2008. "Virtue is its own reward? Support-giving

in the family and loneliness in middle and old age." Ageing & Society 28(2):271-287.

Dykstra, PA and JJ Mandemakers. 2007. "Parent–child consensus and parental well-being in mid

and late life.".

Elder Jr, Glen H., Monica K. Johnson and Robert Crosnoe. 2003. "The emergence and

development of life course theory." Pp. 3-19 in Handbook of the life course" The emergence

and development of life course theory." Springer.

Elliott, Sinikka, Rachel Powell and Joslyn Brenton. 2015. "Being a good mom: Low-income,

Black single mothers negotiate intensive mothering." Journal of Family Issues 36(3):351-

370.

Farmer, Melissa M. and Kenneth F. Ferraro. 2005. "Are racial disparities in health conditional on

socioeconomic status?" Social Science & Medicine 60(1):191-204.

Fingerman, Karen L., Pei-Chun Chen, Elizabeth Hay, Kelly E. Cichy and Eva S. Lefkowitz.

2006. "Ambivalent reactions in the parent and offspring relationship." The Journals of

Gerontology Series B: Psychological Sciences and Social Sciences 61(3):P152-P160.

Fingerman, Karen L., Yen-Pi Cheng, Kira Birditt and Steven Zarit. 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 Sciences and Social

Sciences 67(2):184-193.

Fingerman, Karen L., Yen-Pi Cheng, Lauren Tighe, Kira S. Birditt and Steven Zarit. 2012.

"Relationships between young adults and their parents." Pp. 59-85 in Early adulthood in a

family context, edited by A. Booth, S.L. Brown, N.S. Landale, W.D. Manning and S.M.

McHale. New York: Springer Publishers. Fingerman, Karen L., Elizabeth L. Hay and Kira

S. Birditt. 2004. "The best of ties, the worst of ties: Close, problematic, and ambivalent

social relationships." Journal of Marriage and Family 66(3):792-808.

100

Fingerman, Karen L., Lindsay Pitzer, Eva S. Lefkowitz, Kira S. Birditt and Daniel Mroczek.

2008. "Ambivalent relationship qualities between adults and their parents: Implications for

the well-being of both parties." The Journals of Gerontology Series B: Psychological

Sciences and Social Sciences 63(6):P362-P371.

Fingerman, Karen L., Meng Huo, Kyungmin Kim and Kira S. Birditt. 2017. "Coresident and

noncoresident emerging adults’ daily experiences with parents." Emerging Adulthood

5(5):337-350.

Fingerman, Karen L. 2017. "Millennials and Their Parents: Implications of the New Young

Adulthood for Midlife Adults." Innovation in Aging 1(3):1-16.

Finkel, Steven E. 1995. Causal analysis with panel data. Thousand Oaks, California: Sage

Publications.

Gilligan, Megan, J. J. Suitor, Scott Feld and Karl Pillemer. 2015. "Do positive feelings hurt?

Disaggregating positive and negative components of intergenerational

ambivalence." Journal of Marriage and Family 77(1):261-276.

Hay, Elizabeth L., Karen L. Fingerman and Eva S. Lefkowitz. 2008. "The worries adult children

and their parents experience for one another." The International Journal of Aging and

Human Development 67(2):101-127.

Johnson, David. 2005. "Two‐wave panel analysis: Comparing statistical methods for studying

the effects of transitions." Journal of Marriage and Family 67(4):1061-1075.

Jones, Bobby L. and Daniel S. Nagin. 2007. "Advances in group-based trajectory modeling and

an SAS procedure for estimating them." Sociological Methods & Research 35(4):542-571.

101

Kiecolt, K. J., Rosemary Blieszner and Jyoti Savla. 2011. "Long‐term influences of

intergenerational ambivalence on midlife parents' psychological well‐being." Journal of

Marriage and Family 73(2):369-382.

Koropeckyj‐Cox, Tanya. 2002. "Beyond parental status: Psychological well‐being in middle and

old age." Journal of Marriage and Family 64(4):957-971.

Kowal, Amanda K., Jennifer L. Krull and Laurie Kramer. 2006. "Shared understanding of

parental differential treatment in families." Social Development 15(2):276-295.

Larsen, Randy. Psihologijske teme. "The contributions of positive and negative affect to

emotional well-being ." 2009 18(2):247-266.

Lee, Hyo J. and Maximiliane E. Szinovacz. 2016. "Positive, Negative, and Ambivalent

Interactions With Family and Friends: Associations With Well‐being." Journal of Marriage

and Family78(3):660-679.

Lendon, Jessica P., Merril Silverstein and Roseann Giarrusso. 2014. "Ambivalence in older

parent–adult child relationships: Mixed feelings, mixed measures." Journal of Marriage and

Family 76(2):272-284.

Livingston, G. (2015). Childlessness falls, family size grows among highly educated women.

Retrieved from: http://assets.pewresearch.org/wp-content/uploads/sites/3/2015/05/2015-05-

07_children-ever-born_FINAL.pdf

Merz, Eva-Maria, Hans-Joachim Schulze and Carlo Schuengel. 2010. "Consequences of filial

support for two generations: A narrative and quantitative review." Journal of Family

Issues31(11):1530-1554.

102

Milkie, Melissa A., Alex Bierman and Scott Schieman. 2008. "How adult children influence

older parents' mental health: Integrating stress-process and life-course perspectives." Social

Psychology Quarterly 71(1):86-105.

Noack, Peter and Heike M. Buhl. 2004. "Child-parent relationships." Growing Together:

Personal Relationships Across the Lifespan:45-75.

Pew Research Center. 2010. “The New Demography of American Motherhood.” Retrieved from

http://assets.pewresearch.org/wp-content/uploads/sites/3/2010/10/754-new-demography-of-

motherhood.pdf

Pew Research Center. 2015. “Childlessness Falls, Family Size Grows Among Highly Educated

Women.” Retrieved from http://assets.pewresearch.org/wp-

content/uploads/sites/3/2015/05/2015-05-07_children-ever-born_FINAL.pdf

Pew Research Center. 2018. 7 facts about U.S. moms. Retrieved from

http://www.pewresearch.org/fact-tank/2018/05/10/facts-about-u-s-mothers/

Pillemer, Karl and Kurt Lüscher. 2004. "Intergenerational ambivalences." New Perspectives on

Parent-Child Relations in Later Life.

Pillemer, Karl and J. J. Suitor. 2002. "Explaining mothers' ambivalence toward their adult

children." Journal of Marriage and Family 64(3):602-613.

Reczek, Corinne. 2016. Ambivalence in Gay and Lesbian Family Relationships. Journal of

Marriage and Family, 78(3), 644-659.

Reczek, Corinne and Zhe Zhang. 2016. "Parent–Child Relationships and Parent Psychological

Distress: How Do Social Support, Strain, Dissatisfaction, and Equity Matter?" Research on

Aging 38(7):742-766.

103

Royston, Patrick. 2005. "Multiple imputation of missing values: update of ice." Stata

Journal 5(4):527.

Rusbult, Caryl E. and Paul A. Van Lange. 2003. "Interdependence, interaction, and

relationships." Annual Review of Psychology 54(1):351-375.

Schieman, Scott and Paul Glavin. 2011. "Education and work-family conflict: Explanations,

contingencies and mental health consequences." Social Forces 89(4):1341-1362.

StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.

Suitor, J. J., Megan Gilligan and Karl Pillemer. 2011. "Conceptualizing and measuring

intergenerational ambivalence in later life." Journals of Gerontology Series B:

Psychological Sciences and Social Sciences 66(6):769-781.

Thoits, Peggy A. 2011. "Mechanisms linking social ties and support to physical and mental

health." Journal of Health and Social Behavior 52(2):145-161.

Thomas, Patricia A., Hui Liu, Debra Umberson and J. J. Suitor. 2017. "Family relationships and

well-being." Innovation in Aging 1(3):1-11.

Tighe, Lauren A., Kira S. Birditt and Toni C. Antonucci. 2016. "Intergenerational ambivalence

in adolescence and early adulthood: Implications for depressive symptoms over

time."Developmental Psychology 52(5):824.

Tosi, Marco and Emily Grundy. 2018. "Intergenerational contacts and depressive symptoms

among older parents in Eastern Europe." Aging & Mental Health:1-7.

Uchino, Bert N., Julianne Holt-Lunstad, Timothy W. Smith and Lindsey Bloor. 2004.

"Heterogeneity in social networks: A comparison of different models linking relationships

to psychological outcomes." Journal of Social and Clinical Psychology 23(2):123-139.

104

Umberson, Debra, Tetyana Pudrovska and Corinne Reczek. 2010. "Parenthood, childlessness,

and well‐being: A life course perspective." Journal of Marriage and Family 72(3):612-629.

Van Buuren, Stef. 2012. Flexible imputation of missing data. CRC press.

Ward, Russell 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

Sciences63(4):S239-S247.

White, Ian R., Patrick Royston and Angela M. Wood. 2011. "Multiple imputation using chained

equations: issues and guidance for practice." Statistics in Medicine 30(4):377-399.

Williams, Kristi, Sharon Sassler and Lisa M. Nicholson. 2008. "For better or for worse? The

consequences of marriage and cohabitation for single mothers." Social Forces 86(4):1481-

1511.

Willson, Andrea E., Kim M. Shuey and Glen H. Elder Jr. 2003. "Ambivalence in the relationship

of adult children to aging parents and in‐laws." Journal of Marriage and Family 65(4):1055-

1072.

105

Table 6. Sociodemographic and Health Characteristics of the Sample (Unweighted, Imputed) NLSY79 (1998-2014) Mean or Percent SD CESD score at W2 130.38 91.89 Mother-children relationship quality at W1 Uniformly close 37.69 Ambivalent 30.43 Uniformly unclose 31.89 Changes in intergen ties between W1 and W2

From non-ambiv to ambiv 10.38

From uniformly close to ambiv 5.49

At least one child felt less close at W2 16.50

From uniformly close to unclose or ambiv 30.66 Number of YA children 1.87 0.88 Control variables at W1

Race/Ethnicity NonHispanic White 48.11 NonHispanic Black 31.73 Hispanic 20.17

Marital Status

Never Married 9.98

Married or Cohabiting 57.14

Previously married 32.87

Highest Grade Completed 13.39 1.90 Employment Status Employed 78.81

Logged family income-(inflation adjusted) 11.12

Family size 3.11 1.35 Health Insurance Private 69.40

Public 13.93 Self-pay 16.67 Self-rated health Fair/Poor 14.89 Good 30.50

Very good 35.37

Excellent 19.24

CESD score at W1 118.27

N (Person) 2,534 106

Table 7. Results from OLS Regression Models Predicting Psychological Distress at Wave 2 Model 1 b SE Mother-children relationship quality at W1 Uniformly close (ref) Ambivalent 9.137* 4.039 Uniformly unclose 5.200 3.951 Control variables at W1 Race/Ethnicity (ref: NonHispanic White) NonHispanic Black -14.287*** 4.121 Hispanic -21.021*** 4.472 Marital Status (ref: Married) Never Married 4.262 6.574 Previously married -1.834 4.470 Highest Grade Completed 0.537 0.967 Employed (ref: Not employed) -13.396** 4.593 Logged family income-(inflation adjusted) -10.685** 3.543 Family size 0.951 1.354 Health Insurance (ref: Private) Public 16.760** 5.887 Self-pay 7.723 5.258 Self-rated health (ref: Fair/Poor) Good -10.251+ 5.388 Very good -24.483*** 5.511 Excellent -28.500*** 6.218 CESD score at W1 0.343*** 0.020

Intercept 227.178*** 38.619 R-Squared 0.21 N (NLSY79) 2,534 + p<.10 * p<.05 ** p<.01 *** p<.001 Source: NLSY79 and 79 Young Adults Survey.

107

Table 8. Results from OLS Regression Models Predicting Psychological Distress at Wave 2 Model 1 Model 2 Model 3 Model 4 b SE b SE b SE b SE Changes in intergen ties between W1 and W2 From non-ambiv to ambiv 1.00 5.34 From uniformly close to ambiv 4.49 7.146 At least one child felt less close at W2 7.6 6* 3.5 4 From uniformly close to unclose or ambiv 3.72 4.39 Control variables at W1 Race/Ethnicity (ref: NonHispanic White) NonHispanic Black -14.6 1*** 4.10 -14.6 6*** 4.10 -14.1 9*** 4.10 -14.4 7*** 4.10 Hispanic -21.44*** 4.47 -21.51*** 4.47 -21.45*** 4.46 -21.41*** 4.47 Marital Status (ref: Married) Never Married 3.87 6.58 3.85 6.58 4.35 6.58 3.96 6.58 Previously married -1.23 4.47 -1.21 4.47 -1.37 4.46 -1.24 4.47 Highest Grade Completed 0.62 0.97 0.64 0.97 0.64 0.97 0.61 0.97 Employed (ref: Not employed) -13.36** 4.59 -13.37** 4.59 -13.11** 4.59 -13.32** 4.59 Logged family income-(inflation adjusted) -10.72** 3.54 -10.75** 3.54 -10.78** 3.54 -10.76** 3.54 Family size 1.39 1.33 1.39 1.33 1.26 1.33 1.43 1.33 Health Insurance (ref: Private) Public 17.32** 5.89 17.3 8** 5.89 17.11** 5.88 17.28** 5.89 Self-pay 8.10 5.25 8.09 5.25 7.71 5.25 8.00 5.25 Self-rated health (ref: Fair/Poor) Good -10.0 7+ 5.39 -10.04+ 5.39 -10.2 3+ 5.39 -10.02+ 5.39 Very good -23.94*** 5.51 -23.84*** 5.52 -23.91*** 5.50 -23.89*** 5.51 Excellent -28.24*** 6.22 -28.23*** 6.22 -28.29*** 6.21 -28.23*** 6.22 CESD score at W1 0.34*** 0.02 0.34*** 0.02 0.34*** 0.02 0.34*** 0.02 108

Intercept 228.95*** 38.55 228.74*** 38.53 227.16*** 38.54 228.75*** 38.55 R-Squared 0.21 0.21 0.21 0.21 N (NLSY79) 2,534 2,534 2,534 2,534 + p<.10 * p<.05 ** p<.01 *** p<.001 Source: NLSY79 and 79 Young Adults Survey.

109

Chapter 5. Conclusion

As one of the most influential and lasting social roles, parenthood is intricately linked with both psychological and physical wellbeing across the life course (Umberson, Pudrovska, & Reczek,

2010). Yet, one major gap exists in this literature – despite the theoretical linkage, how the varying contexts of raising young children into adulthood matter for individual wellbeing across midlife receives little research attention (Umberson et al., 2010; Umberson, Crosnoe, & Reczek,

2010; Fingerman, 2017). This research gap requires immediate attention considering the increasingly diverse parenthood landscape as midlife parents today continue to provide substantial economic, emotional, and residential support to children during the transition to adulthood (Fingerman et al., 2012). Furthermore, midlife is a life course stage when population health differences further escalate (Read & Gorman, 2011). Recent research suggests that the health declines are particularly relevant to specific race-ethnicity, gender, and socioeconomic groups among the midlife adults in the contemporary cohort (Case & Deaton, 2015).

Building on this gap, the three empirical chapters of my dissertation ask: how do various parenthood contexts matter for midlife adults’ mental and physical health? Guided by a life course perspective and prior research on family and health, this dissertation examines effects of two main aspects of parent-child ties – structure (e.g., coresidence) and content (e.g., emotional support) on midlife adults’ mental and physical wellbeing. All three empirical chapters used data

110

from NLSY79, and the sample sizes and analytical methods used in each chapter were based on specific research questions.

Chapter 2, the first empirical chapter, examines how children’s life course stage and parent-child coresidence at midlife matters for individuals’ psychological wellbeing, and further articulates how this relationship differs for men and women. Results from both OLS regression models suggest that mothers with only adult children in the household experienced higher levels of psychological distress than other midlife women, whereas coresidence with both minor and adult children might be conducive to mental wellbeing for both men and women at midlife. Next,

Chapter 3 tests how three emerging mother-child coresidential biographies: the gone-for-good, the boomerangers, and the never-left, mattered for mothers’ BMI trajectories. Results from growth curve modeling suggest that boomerangers’ mothers were at higher risk of gaining additional weight across midlife relative to mothers to the gone-for-good. In turn, mothers whose children never left the home experienced declining body weight over time relative to mothers whose children left “on time.” The last empirical chapter, Chapter 4, shifts the focus from parent- child coresidence to relationship quality. Specifically, this chapter examines the association between intergenerational relationship quality and midlife mother’s mental wellbeing. According to the results from OLS regression models with lagged dependent variables, midlife mothers in a collective ambivalent relationship with multiple adolescent and young adult children had more psychological distress than mothers in a uniformly close or a uniformly unclose relationship with all children. Mothers with at least one child reporting declining feelings toward mother also appeared to have worse mental health over time.

111

The three empirical chapters are interconnected and they all shed light on the complex relationship between midlife parenthood contexts and adults’ wellbeing. Below, I explicate the findings in more detail, organizing the implications in themes and highlighting the contributions to existing literature, with implications for future research.

First, findings from all three chapters address the main research question – parenthood contexts of raising young children into adulthood are indeed linked with both mental and physical wellbeing at midlife. These results highlight the increasingly demanding experiences of parenting emerging young adult children (Fingerman, 2017; Furstenberg, 2010). As Americans experience more twists and turns during the transition to adulthood and become less likely to achieve the traditional markers of adulthood such as completion of education, leaving parental home, launch of a stable job with livable income, and marriage (Vespa, 2017), U.S. parents and their wellbeing are likely impacted by lives of their coming of age children simultaneously. This study focuses on one emerging change in young adults’ lives – coresidence with parents, and examines how it relates parents’ psychological wellbeing and body weight trajectory. In contrast to studies that reveal little or no effects of coresidence on young adults mental and economic wellbeing (Copp et al., 2015; Kaplan, 2012), findings in this dissertation suggest that coresidence with parents, especially with mothers, likely pose a greater risk to parents’ health at midlife.

Future studies should continue to draw on and extend the research that connects stressful life events of adult children to parental wellbeing (Milkie et al., 2008; Barr et al., 2018). In my future endeavors, I plan to examine how children’s life event during the transition to adulthood, such as having student loans, job loss, or substance use problems is associated with parental mental, physical, and financial wellbeing. This work will add nuance to the influential parent-child ties

112

and unveil the challenges that different families face in launching their adult children against the current economic backdrop.

Second, findings from my dissertation suggest that when young adult children encounter more problems in their coming-of-age journey, midlife mothers are the ones who take more responsibility in providing care to children. Midlife mothers also experience costs on wellbeing as a result of countering the stress in life with their young adult children. This is consistent with prior work, which suggests that midlife mothers engage in more frequent intergenerational contact, provide more support, and report closer ties as well as more conflicts with grown children than fathers, across socioeconomic status (Arnett & Schwab, 2012; Fingerman et al.,

2016). Maternal health has also been established to be more susceptible to childrearing stress

(Umberson et al., 2010). Findings from my dissertation extend this literature by showing the motherhood penalty in wellbeing as midlife mothers help children navigate a more uncertain pathway to adulthood (Barr et al., 2018), which has clear implications for the scholarship on gender stratification. Specifically, future work should cast more attention to midlife motherhood in relation to the stalled and uneven gender revolution (England, 2010) – how does midlife mothers’ support for emerging adult children contribute to the family division of labor and alter mother’s employment decisions and trajectories? Additionally, how are women in the sandwich generation – who provides care to multiple generations fare in view of a heavier burden of supporting the younger generation? Furthermore, how are these relationships conditioned on other social contexts such as race, class, sexuality, and marital status?

Third, this dissertation advances prior research by demonstrating the importance of considering multiple family members in relation to individual wellbeing. A life course

113

perspective argues that the family members’ lives are linked (Elder, Johnson, Crosnoe, 2003). To better conceptualize the dynamics and effects of linked lives in a family network, family and health scholars argue that multiple family members should be considered and family level data should be used (Umberson et al., 2010; Reczek, 2016). My dissertation, Chapter 2 and Chapter 4 more specifically, respond to this call and account for multiple children in mother’s ties with children at midlife. Results from Chapter 2 suggest that the coresidence with multiple children at different life course stages likely changes the family dynamics (e.g., older children helping young siblings and relieving mother’s caregiving burden) and further influences parents’ mental health. It is unclear whether some older children postpone their departure from home to help midlife parents with young siblings. Future research should delve into this possibility. Results from Chapter 4 complements prior work in demonstrating the relational nature of intergenerational relationship quality, wherein accounts from multiple children should be considered (Pillemer & Suitor, 2008). Future work should continue on this endeavor to improve our understanding of the complex family dynamics, with particular attention on how the structural factors (e.g., gender, sexuality race, class) change relationship quality and lead to mixed emotions in the family network.

Taken together, my dissertation is among the first to thoroughly assess how midlife parental health is shaped by various parenthood contexts, especially those pertaining to young adults’ coming-of-age experiences. The findings add nuance to the influential parent-child ties and unveil the challenges that different families face in launching their adult children against the current economic backdrop. This dissertation contributes to a more comprehensive endeavor in tackling the specific political issues while caring for a broader population of American families.

114

References

Arnett, Jeffrey J. and Joseph Schwab. 2012. The Clark University poll of emerging adults:

Thriving, struggling, and hopeful. Worcester, MA: Clark University.

Barr, Ashley B., Leslie G. Simons, Ronald L. Simons, Steven R. Beach and Robert A. Philibert.

2018. "Sharing the Burden of the Transition to Adulthood: African American Young

Adults’ Transition Challenges and Their Mothers’ Health Risk." American Sociological

Review: 0003122417751442.

Case, Anne and Angus Deaton. 2015. "Rising morbidity and mortality in midlife among white

non-Hispanic Americans in the 21st century." Proceedings of the National Academy of

Sciences of the United States of America112(49):15078-15083.

Copp, Jennifer E., Peggy C. Giordano, Monica A. Longmore and Wendy D. Manning. 2015.

"Living with parents and emerging adults’ depressive symptoms." Journal of Family

Issues(02/01/2017 pages). google.

Elder Jr, Glen H., Monica K. Johnson and Robert Crosnoe. 2003. "The emergence and

development of life course theory." Pp. 3-19 in Handbook of the life course"The emergence

and development of life course theory."Springer.

England, Paula. 2010. "The gender revolution: Uneven and stalled." Gender Soc.Gender and

Society 24(2):149-166.

Fingerman, Karen L., Yen-Pi Cheng, Lauren Tighe, Kira S. Birditt and Steven Zarit. 2012.

"Relationships between young adults and their parents." Pp. 59-85 in Early adulthood in a

family context, edited by A. Booth, S.L. Brown, N.S. Landale, W.D. Manning and S.M.

McHale. New York: Springer Publishers. Fingerman, Karen L., Elizabeth L. Hay and Kira

115

S. Birditt. 2004. "The best of ties, the worst of ties: Close, problematic, and ambivalent

social relationships." Journal of Marriage and Family 66(3):792-808.

Fingerman, Karen L., Kyungmin Kim, Kira S. Birditt and Steven H. Zarit. 2016. "The Ties That

Bind: Midlife Parents' Daily Experiences With Grown Children." Journal of Marriage and

Family 78:431-450.

Fingerman, Karen L. 2017. "Millennials and Their Parents: Implications of the New Young

Adulthood for Midlife Adults." Innovation in Aging 1(3):1-16.

Furstenberg Jr, Frank F. 2010. "On a new schedule: Transitions to adulthood and family

change." The Future of Children 20(1):67-87.

Kaplan, Greg. 2012. "Moving back home: Insurance against labor market risk." Journal of

Political Economy 120(3):446-512.

Milkie, Melissa A., Alex Bierman and Scott Schieman. 2008. "How adult children influence

older parents' mental health: Integrating stress-process and life-course perspectives." Social

Psychology Quarterly 71(1):86-105.

Pillemer, Karl and J. J. Suitor. 2008. "Collective ambivalence: Considering new approaches to

the complexity of intergenerational relations." The Journals of Gerontology Series B:

Psychological Sciences and Social Sciences63(6):S394-S396.

Read, Jen’nan G. and Bridget K. Gorman. 2011. "Gender and health revisited." Pp. 411-429

in Handbook of the Sociology of Health, Illness, and Healing, edited by B. Pescosolido, J.

Martin, J. McLeod and A. Rogers.Springer, New York, NY.

Reczek, Corinne. 2016. "Ambivalence in gay and lesbian family relationships." Journal of

Marriage and Family 78(3):644-659.

116

Umberson, Debra, Robert Crosnoe and Corinne Reczek. 2010. "Social relationships and health

behavior across the life course." Annual Review of Sociology 36:139-157.

Umberson, Debra, Tetyana Pudrovska and Corinne Reczek. 2010. "Parenthood, childlessness,

and well‐being: A life course perspective." Journal of Marriage and Family 72(3):612-629.

Vespa, Jonathan. (2017). The Changing Economics and Demographics of Young Adulthood:

1975–2016. Current Population Reports, P20-579, U.S. Census Bureau, Washington,

DC.

117

Complete References

Aquilino, William S. 2006. Family relationships and support systems in emerging adulthood. In

J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st

century (pp. 193–217). Washington, DC: American Psychological Association.

Baumeister, Roy F., Ellen Bratslavsky, Catrin Finkenauer, and Kathleen D. Vohs. 2001.

Bad is stronger than good. Review of General Psychology, 5, 323–370. doi:10.1037//1089-

2680.5.4.323.

Adams, Patricia F and Charlotte A Schoenborn. 2006. "Health Behaviors of Adults: United

States, 2002-04." Vital and Health Statistics. Series 10, Data from the National Health

Survey (230):1-140.

Ahima, Rexford S and Mitchell A Lazar. 2013. "The Health Risk of Obesity—Better Metrics

Imperative." Science 341(6148):856-58. doi: 10.1126/science.1241244.

Ailshire, Jennifer A and James S House. 2011. "The Unequal Burden of Weight Gain: An

Intersectional Approach to Understanding Social Disparities in Bmi Trajectories from

1986 to 2001/2002." Social forces; a scientific medium of social study and interpretation

90(2):397-434. doi: https://doi.org/10.1093/sf/sor001.

Allison, Paul D. 1994. "Using panel data to estimate the effects of events." Sociological Methods

& Research 23(2):174-199.

118

Altintas, Evrim and Oriel Sullivan. 2016. "Fifty years of change updated: Cross-national gender

convergence in housework." Demographic Research 35:455-470.

Anderson, Elaine A., Julie K. Kohler and Bethany L. Letiecq. 2005. "Predictors of depression

among low-income, nonresidential fathers." Journal of Family Issues 26(5):547-567.

Arnett, Jeffrey J. and Joseph Schwab. 2012. The Clark University poll of emerging adults:

Thriving, struggling, and hopeful. Worcester, MA: Clark University.

Baltrus, Peter T, John W Lynch, Susan Everson-Rose, Trivellore E Raghunathan and George A

Kaplan. 2005. "Race/Ethnicity, Life-Course Socioeconomic Position, and Body Weight

Trajectories over 34 Years: The Alameda County Study." American Journal of Public

Health 95(9):1595-601.

Barr, Ashley B., Leslie G. Simons, Ronald L. Simons, Steven R. Beach and Robert A. Philibert.

2018. "Sharing the Burden of the Transition to Adulthood: African American Young

Adults’ Transition Challenges and Their Mothers’ Health Risk." American Sociological

Review: 0003122417751442.

Blanchard-Fields, Fredda, Renee Stein and Tonya L. Watson. 2004. "Age differences in

emotion-regulation strategies in handling everyday problems." The Journals of

Gerontology Series B: Psychological Sciences and Social Sciences 59(6):P261-P269.

Berge, Jerica M, Nicole Larson, Katherine W Bauer and Dianne Neumark-Sztainer. 2011. "Are

Parents of Young Children Practicing Healthy Nutrition and Physical Activity

Behaviors?". Pediatrics 127(5):881-87.

119

Bierman, Alex and Melissa A Milkie. 2008. "Intergenerational Stress Proliferation between

Adult Children and Parents: Contingencies by Functional Timing and Parent's Gender."

Advances in Life Course Research 13:343-67. doi: 10.1016/S1040-2608(08)00013-0.

Birditt, Kira S, Karen L Fingerman and Steven H Zarit. 2010. "Adult Children’s Problems and

Successes: Implications for Intergenerational Ambivalence." Journals of Gerontology

Series B: Psychological Sciences and Social Sciences 65(2):145-53. doi:

https://doi.org/10.1093/geronb/gbp125.

Blair-Loy, Mary. 2009. Competing devotions: Career and family among women

executives. Harvard University Press.

Breeze, E., R. Clarke, M. J. Shipley, M. G. Marmot and A. E. Fletcher. 2006. "Cause-specific

mortality in old age in relation to body mass index in middle age and in old age: follow-

up of the Whitehall cohort of male civil servants." International Journal of Epidemiology

35(1):169-178.

Bures, Regina M., Tanya Koropeckyj-Cox and Michael Loree. 2009. "Childlessness, parenthood,

and depressive symptoms among middle-aged and older adults." Journal of Family

Issues30(5):670-687.

Burn, Katherine and Cassandra Szoeke. 2016. "Boomerang Families and Failure-to-Launch:

Commentary on Adult Children Living at Home." Maturitas 83:9-12. doi:

http://dx.doi.org/10.1016/j.maturitas.2015.09.004.

Carr, Deborah. 2004. "“My Daughter Has a Career; I Just Raised Babies”: The Psychological

Consequences of Women's Intergenerational Social Comparisons." Social Psychology

Quarterly 67(2):132-154.

120

Case, Anne and Angus Deaton. 2015. "Rising Morbidity and Mortality in Midlife among White

Non-Hispanic Americans in the 21st Century." Proceedings of the National Academy of

Sciences of the United States of America 112(49):15078-83. doi:

10.1073/pnas.1518393112 [doi].

Charles, Susan T. and Laura L. Carstensen. 2007. "Emotion regulation and aging." Handbook of

Emotion Regulation 6:307-327.

Charles, Susan T. 2010. "Strength and vulnerability integration: A model of emotional well-

being across adulthood." Psychological Bulletin 136(6):1068.

Cherlin, Andrew J. 2010. "Demographic Trends in the United States: A Review of Research in

the 2000s." Journal of Marriage and Family 72(3):403-19. doi: 10.1111/j.1741-

3737.2010.00710.x.

Cherlin, Andrew J. 2014. Labor's Love Lost: The Rise and Fall of the Working-Class Family in

America. New York: Russell Sage Foundation.

Cichy, Kelly E., Lefkowitz, Eva S. and Karen L. Fingerman. 2012. "Conflict engagement and

conflict disengagement during interactions between adults and their parents." Journals of

Gerontology Series B: Psychological Sciences and Social Sciences 68(1):31-40.

Clarke, Philippa, Victor Marshall, James House and Paula Lantz. 2011. "The social structuring

of mental health over the adult life course: advancing theory in the sociology of

aging."Social Forces 89(4):1287-1313.

Coate, Patrick, Pawel Krolikowski and Mike Zabek. 2017. "Parental Proximity and the Earnings

Consequences of Job Loss." Economic Commentary 3(February).

121