University of Southern California Dissertations and Theses

The old ball and linkage: couples’ prenatal conflict behavior, cortisol linkage, and postpartum depression risk

(USC Thesis Other)

The old ball and linkage: couples’ prenatal conflict behavior, cortisol linkage, and postpartum depression risk

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

The Old Ball and Linkage: Couples’ Prenatal Conflict Behavior, Cortisol Linkage, and
Postpartum Depression Risk

by

Mona Khaled

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)

August 2021

Copyright 2020 Mona Khaled
ii

Acknowledgements
I want to thank my dissertation committee, Dr. Darby Saxbe, Dr. Gayla Margolin, Dr.
Mark Lai, Dr. Dorian Traube, and Dr. Brain Baucom, for their feedback during and since
inception of this project, for challenging me to reflect and evolve, and for inspiring me to grow
as a clinical researcher. I am very appreciative of your time and commitment to reviewing this
project.
I want to extend an enormous thank you to Dr. Mark Lai, who has provided invaluable
consultation and guidance for many steps and phases of this project. I’ve been very grateful for
his support.
Additionally, I am so fortunate for my lab mate, co-therapist, friend, and on-call
statistician, Geoff Corner, who has been extremely generous with offering me his time and mind.
He has been monumental to not only this project, but my first first-author publication. I am
forever indebted to him.
I hold a deep sense of gratitude for my irreplaceable lab mates, who have not only
provided emotional support and comfort over the years, but also academic and intellectual
guidance: the graduate students and lab managers of the NEST Lab, Hannah Khoddam, Geoff
Corner, Alyssa Morris, Sarah Stoycos, Narcis Marshall, Sofi Cardenas, Katie Horton, Bryna Tsai
and Nia Barbee.
A very special thank you to my dissertation chair and research advisor, Dr. Darby Saxbe.
I think back to the very first time we met in 2014 in Chris Dunkel Schetter’s lab meeting, after
she gave a Brown Bag talk at UCLA, and reflect on how far I’ve come since then. I owe so much
of my academic, professional, and personal growth to her. I am extremely thankful for her one-
iii

of-a-kind mentorship and unconditional support during the most challenging moments of
graduate school.
I am so fortunate to have experienced graduate school with such brilliant, hardworking,
genuine, loving, compassionate colleagues and friends. I am a better researcher, clinician, lab
mate, and friend as a result of growing alongside them: my cohort mates, Anna Blanken,
Vanessa Calderon, Hannah Rasmussen, Miriam Rubenson, and Elissa McIntosh. Also, to others
who have been vital in supporting me through this journey, within USC, Kelly Durbin and
Sylvanna Vargas, and outside, my high school friends; Rachel and Patrice; and Jeanette.
I want to express my sincere gratitude to my roommate and best friend, Tiffany Tran,
who has provided me with immense comfort and unwavering support throughout every step of
graduate school. I could not have made it through this process without her.
I want to thank my siblings, and most importantly, my wonderful mother, Ebtesam, for
standing behind me and having the utmost faith in me. For many reasons, I would not be at this
place in my life or able to pursue my dream if it weren’t for her. I dedicate this accomplishment
to her.

iv

TABLE OF CONTENTS
Acknowledgements ......................................................................................................................... ii
List of Tables ...................................................................................................................................v
List of Figures ................................................................................................................................ vi
Abstract ......................................................................................................................................... vii
General Introduction ........................................................................................................................1
Current Studies.....................................................................................................................2
References ............................................................................................................................4
Chapter 1: Dyadic risk factors for postpartum depression: Does prenatal relationship conflict
behavior pose a risk? ........................................................................................................................6
Current Study .....................................................................................................................11
Methods..............................................................................................................................13
Results ................................................................................................................................22
Discussion ..........................................................................................................................32
References ..........................................................................................................................40

Chapter 2: Physiological Linkage in Pregnancy: Couples’ Cortisol, Negative Conflict Behavior,
and Postpartum Depression............................................................................................................52
Current Study .....................................................................................................................56
Methods..............................................................................................................................57
Results ................................................................................................................................67
Discussion ..........................................................................................................................75
References ..........................................................................................................................81

General Discussion ........................................................................................................................87
Conclusion .........................................................................................................................89

v

List of Tables
1.1 Participant and couple characteristics ......................................................................................23
1.2 Descriptives for key study variables ........................................................................................23
1.3 Zero-order correlations for all study and control variables .....................................................25
1.4 APIM results from models predicting maternal and paternal PPD ..........................................26
1.5 Sensitivity analyses for APIM results with positive conflict behavior models .......................27
1.6 Intercepts and variances for negative behavior segments and reciprocity effects ...................28
1.7 Associations between negative escalation and PPD ................................................................31
1.8 Between-level effects for negative escalation predicting PPD ................................................31
1.9 Between-level effects for negative escalation predicting PPD with covariates .......................32
2.1 Participant and couple characteristics ......................................................................................67
2.2 Descriptives for key study variables ........................................................................................68
2.3 Zero-order correlations for study variables ..............................................................................69
2.4 Cortisol linkage: Partner cortisol predicting own cortisol level ..............................................70
2.5 Associations between cortisol linkage and negative conflict behavior ...................................70
2.6 Associations between cortisol linkage and PPD ......................................................................72
2.7 Associations between cortisol linkage, negative conflict behavior, and PPD .........................73
2.8 Interaction effects of own and partner negative conflict behavior as a moderator between
cortisol linkage and maternal PPD ...........................................................................................74
2.9 Interaction effects of own and partner negative conflict behavior as a moderator between
cortisol linkage and paternal PPD ............................................................................................74

vi

List of Figures
1.1 APIM results from models predicting maternal and paternal PPD ..........................................27
2.1 Overview of prenatal visit procedures and saliva collection time-points ................................60
2.2 Association between cortisol linkage and mothers’ negative conflict behavior ......................71
2.3 Association between cortisol linkage and fathers’ negative conflict behavior ........................71
2.4 Association between cortisol linkage and fathers’ PPD ..........................................................73

vii

Abstract
Given that poor relationship functioning is one of the strongest predictors of maternal and
paternal postpartum depression (PPD), couple conflict behavior is a valuable dyadic process to
explore. The current dissertation examined whether couple’s observed behaviors during prenatal
conflict interactions predict postpartum depressive symptoms for mothers and fathers.
Additionally, couples’ correlated cortisol patterns can indicate a mutual influence of
physiological and psychological states and may reflect adaptive or maladaptive relationship
functioning depending on the context of linkage. Given that cortisol has also been implicated in
postpartum depression, it is valuable to investigate these dyadic interpersonal processes during
the transition to parenthood. Thus, this study also examined whether couple cortisol linkage
during pregnancy is associated with mother and fathers’ negative relationship conflict behavior
and their subsequent postpartum depressive symptoms. A total of 82 opposite-sex couples
expecting their first child engaged in a conflict discussion and completed measures of
relationship satisfaction and depressive symptoms during pregnancy. At approximately six-
months postpartum, couples completed a measure of postpartum depressive symptoms. The
findings suggest that for some couples during prenatal conflict, couple’s negative behaviors may
be benign, and yet, fathers’ positive behavior may facilitate poor postpartum adjustment for their
partners. Furthermore, these dissertation findings also present a context of couple cortisol linkage
in which stronger physiological associations between partners may indicate adaptive processes.
Directions for future research and potential mechanisms are discussed.
1

General Introduction
The transition to parenthood is arguably one of the most disruptive and transformative
processes that one can experience. Mothers undergo drastic anatomical, hormonal, cardiac,
respiratory, and musculoskeletal changes (Soma-Pillay, Nelson-Piercy, Tolppanen, & Mebazaa,
2016). While fathers do not first-hand experience the physical effects of pregnancy, some fathers
do mimic some of their partners’ physical changes, and this synchrony of physical alterations
may relate to more paternal investment in the relationship and parenting (Brennan, Ayers,
Ahmed, & Marshall-Lucette, 2007; Saxbe et al., 2017). Expectant parents experience changes in
nearly all other facets of life: psychological, identity, role, sleep, environmental, and financial
changes. For a subset of parents, they will also experience significant clinical mood changes,
with between 8% and 35% of mothers developing postpartum depression (Gavin, Lindhorst, &
Lohr, 2011; Shorey et al., 2018) and between 6% to 10% of fathers (Epifanio, Genna, De Luca,
Roccella, & La Grutta, 2015; Paulson & Bazemore, 2010).
Experiencing postpartum depression (PPD) after the birth of a child is like adding insult
to injury; during an already physically demanding, emotionally taxing, and sleep-deprived period
of time, some parents are enduring despondent moods, apathy or disinterest, hopelessness,
feelings of worthlessness, and even thoughts about dying. Interestingly, some symptoms of
depression just happen to overlap with the postpartum period, making it harder to disentangle
what is attributable to clinical mood dysregulation versus normative postpartum adjustment:
significant changes in sleep, loss of energy/fatigue, difficulty concentrating, and changes in
appetite.
Given the added challenges that postpartum depression can contribute to a period in
which psychological resources are already low, it is undoubtedly clear how valuable preventative
2

research on PPD can be for some parents. Identifying which individuals are at risk for
developing PPD and applying that knowledge to prenatal interventions can make a world of a
difference for couples making the transition to parents. Seeing as low relationship satisfaction
has been found to be the most reliable and robust risk factor for PPD in mothers and fathers
(Gawlik et al., 2014; Wee, Skouteris, Pier, Richardson, & Milgrom, 2011; Yim, Tanner
Stapleton, Guardino, Hahn-Holbrook, & Dunkel Schetter, 2015), the couple relationship has the
power to contribute to or buffer from the development of postpartum depressive symptoms.
Therefore, investigating which features of the couple relationship are linked to PPD is a fruitful
area to explore.
Current Studies
The two studies of this dissertation explore whether postpartum depressive symptoms are
associated with 1) couples’ prenatal conflict interactions, and 2) couples’ physiological linkage
during pregnancy. Chapters 1 and 2 both investigated the role of behaviorally coded couple
conflict interactions during pregnancy among the same sample of couples. Expectant couples
engaged in a 15-minute prenatal conflict discussion in which topics included areas of their
relationship that they would like to change. These discussions were coded for mothers and
fathers’ positive and negative conflict behaviors.
In Chapter 1, mothers and fathers’ positive and negative conflict behaviors were
conceptualized as potential risk factors of PPD and were tested to examine if display of these
conflict behaviors increased risk of PPD for their partners. Additionally, we examined whether
reciprocal negative behavior, in which individuals respond to their partners’ displays of negative
behavior with more negative behavior, was associated with PPD symptoms. By focusing on
couples’ observed behaviors during conflict as possible links to PPD, we hoped to provide some
3

useful targets for future interventions. Identifying specific, concrete behavior that puts an
individual or their partner at higher risk for developing PPD can contribute to prenatal couple
treatment as a preventative measure.
In Chapter 2, the focus was on couples’ physiological linkage as it relates to conflict
behavior and PPD. Specifically, are couples who show correlated levels of cortisol more likely to
exhibit negative behavior during conflict and at an increased risk for PPD? Moreover, among
expectant couples with stronger cortisol synchrony, does an individual’s negative conflict
behavior influence their partner’s PPD symptoms? Chapter 2 is designed to assess whether
physiological linkage among pregnant couples makes partners more susceptible to negative
emotions and mood dysregulation.
Both Chapter 1 and Chapter 2 offer unique investigations to the study of couples and
PPD. Chapter 1 is the first examination in which couples’ prenatal conflict behavior is tested as a
predictor of PPD. Additionally, we are not aware of any other studies that explore expectant
couples’ reciprocal negative conflict behavior as a potential risk factor for maternal and paternal
PPD. Chapter 2 is one of just six studies to date that has assessed cortisol linkage in pregnant
couples. Both of these studies contribute to the literature on couples’ adjustment during the
transition to parenthood, and specifically, ways in which the couple relationship may confer risk
for or buffer development of PPD symptoms through couple conflict behavior and physiological
linkage.

4

References
Brennan, A., Ayers, S., Ahmed, H., & Marshall-Lucette, S. (2007, August). A critical review of
the Couvade syndrome: The pregnant male. Journal of Reproductive and Infant Psychology.
Routledge . https://doi.org/10.1080/02646830701467207
Epifanio, M. S., Genna, V., De Luca, C., Roccella, M., & La Grutta, S. (2015). Paternal and
maternal transition to parenthood: The risk of postpartum depression and parenting stress.
Pediatric Reports, 7(2), 38–44. https://doi.org/10.4081/pr.2015.5872
Gavin, A. R., Lindhorst, T., & Lohr, M. J. (2011). The prevalence and correlates of depressive
symptoms among adolescent mothers: Results from a 17-year longitudinal study. Women &
Health, 51(6), 525–545.
https://doi.org///dx.doi.org.libproxy1.usc.edu/10.1080/03630242.2011.606355
Gawlik, S., Müller, M., Hoffmann, L., Dienes, A., Wallwiener, M., Sohn, C., … Reck, C. (2014).
Prevalence of paternal perinatal depressiveness and its link to partnership satisfaction and
birth concerns. Archives of Women’s Mental Health, 17(1), 49–56.
https://doi.org/10.1007/s00737-013-0377-4
Paulson, J. F., & Bazemore, S. D. (2010). Prenatal and postpartum depression in fathers and its
association with maternal depression: A meta-analysis. JAMA, 303(19), 1961–1969.
https://doi.org/10.1001/jama.2010.605
Saxbe, D. E., Edelstein, R. S., Lyden, H. M., Wardecker, B. M., Chopik, W. J., & Moors, A. C.
(2017). Fathers’ decline in testosterone and synchrony with partner testosterone during
pregnancy predicts greater postpartum relationship investment. Hormones and Behavior,
90, 39–47. https://doi.org/10.1016/j.yhbeh.2016.07.005

5

Shorey, S., Chee, C. Y. I., Ng, E. D., Chan, Y. H., Tam, W. W. S., & Chong, Y. S. (2018,
September 1). Prevalence and incidence of postpartum depression among healthy mothers:
A systematic review and meta-analysis. Journal of Psychiatric Research. Elsevier Ltd.
https://doi.org/10.1016/j.jpsychires.2018.08.001
Soma-Pillay, P., Nelson-Piercy, C., Tolppanen, H., & Mebazaa, A. (2016). Physiological
changes in pregnancy. Cardiovascular Journal of Africa, 27(2), 89–94.
https://doi.org/10.5830/CVJA-2016-021
Wee, K. Y., Skouteris, H., Pier, C., Richardson, B., & Milgrom, J. (2011, May 1). Correlates of
ante- and postnatal depression in fathers: A systematic review. Journal of Affective
Disorders. Elsevier B.V. https://doi.org/10.1016/j.jad.2010.06.019
Yim, I., Tanner Stapleton, L., Guardino, C., Hahn-Holbrook, J., & Dunkel Schetter, C. (2015).
Biological and psychosocial predictors of postpartum depression: Systematic review and
call for integration. Annual Review of Clinical Psychology, 11, 99–137.
https://doi.org/10.1146/annurev-clinpsy-101414-020426

6

Chapter 1: Dyadic risk factors for postpartum depression: Does prenatal relationship
conflict behavior pose a risk?
Although the birth of a baby is often thought to be a joyous event, the postpartum period
is also a time of increased health risks for parents, including increased prevalence of depression
(Condon, Boyce, & Corkindale, 2003; Perren, Von Wyl, Bürgin, Simoni, & Von Klitzing, 2005;
Stowe & Nemeroff, 1995), decreased relationship satisfaction (Doss, Rhoades, Stanley, &
Markman, 2009), and heightened physiological stress (Groer, Davis, & Hemphill, 2002;
Mastorakos & Ilias, 2003). Many new parents experience their first episode of major depression
during the postpartum period (Stowe & Nemeroff, 1995), increasing the risk for future
depressive episodes (Monroe & Harkness, 2005). Postpartum depression (PPD) can have
devastating effects on both parent and child well-being (Field, 2010; Murray, Fiori-Cowley,
Hooper, & Cooper, 1996; Stein et al., 1991), compromising a parent’s ability to be sensitive and
empathic to infant cues (Murray, Fiori-cowley, & Hooper, 1996; Zajicek-Farber, 2009),
increasing the risk for family conflict (Burke, 2003), and even potentially causing harm to the
infant (Cadzow, Armstrong, & Fraser, 1999; Zajicek-Farber, 2009).
Although both fathers and mothers experience PPD and there is evidence that PPD risk is
shared within couples (Paulson & Bazemore, 2010), few studies have focused on prenatal
predictors of fathers’ PPD or examined PPD within a dyadic context. This is surprising, given
that low social support and poor relationship quality have been consistently linked with PPD
(Beck, 2001; Collins, Dunkel-Schetter, Lobel, & Scrimshaw, 1993). Yet, studies have generally
overlooked the role of dynamic prenatal couple relationship processes in the etiology of PPD.
PPD Prevalence and Consequences
7

The prevalence of postpartum depression is estimated to be between 8% and 35% for
mothers (Gavin, Lindhorst, & Lohr, 2011; Shorey et al., 2018) and 6% to 10% for fathers
(Epifanio, Genna, De Luca, Roccella, & La Grutta, 2015; Paulson & Bazemore, 2010), with
potentially higher rates of PPD depending on racial, geographical, or partner context (Goodman,
2004). In the short-term, postpartum depression can impair parents’ emotional responsiveness to
infants (Arteche et al., 2011; Murray, Fiori-Cowley, et al., 1996; Zajicek-Farber, 2009) and
reduce the likelihood of safe parenting practices (Field, 2010; Zajicek-Farber, 2009), increasing
risk to infants’ emotional and physical health. Postpartum depression also has far-reaching
implications and long-term impacts on child development (Brand & Brennan, 2009; Ertel &
Gillman, 2010; Gress-Smith, Luecken, Lemery-Chalfant, & Howe, 2012; Murray, 1992) and
overall family well-being (Burke, 2003), yet prevention efforts and treatments have been
modestly effective (Goldstein, Rosen, Howlett, Anderson, & Herman, 2020; O’Hara & McCabe,
2013). With this cascade of enduring and widespread consequences, it is necessary to understand
what factors may contribute to the development of PPD.
Risk Factors for PPD
Established risk factors for maternal PPD include history of depression, symptoms of
depression and anxiety during pregnancy, neuroticism, low self-esteem, stressful life events,
chronic stress, low partner and social support, and poor relationship quality (Iles, Slade, & Spiby,
2011; Milgrom et al., 2008; O’Hara & McCabe, 2013; Robertson, Grace, Wallington, & Stewart,
2004; Yim, Tanner Stapleton, Guardino, Hahn-Holbrook, & Dunkel Schetter, 2015). Until
recently, most PPD research focused on mothers’ experience of pregnancy, the transition to
parenthood, and adjustment during the postpartum period. There is now a growing body of
literature examining the risk factors and implications of PPD in fathers. Similar to mothers’ risk
8

factors for PPD, a consistent finding is that relationship quality has been shown to be predictive
of paternal PPD, with low relationship satisfaction reliably increasing risk of paternal PPD
symptoms (Bielawska ‐Batorowicz & Kossakowska ‐Petrycka, 2006; Don & Mickelson, 2012;
Morse, Buist, & Durkin, 2000; Gawlik et al., 2014; Wee, Skouteris, Pier, Richardson,
&Milgrom, 2011). In fact, paternal depression is correlated with maternal depression (Paulson &
Bazemore, 2010), with the strongest and most common predictor of paternal PPD being partners’
depressive symptoms (Goodman, 2004).
Psychosocial assessments or screenings during pregnancy can increase awareness of risk
for PPD, but do not reduce the likelihood of developing depressive symptoms in the postpartum
period (Austin, Priest, & Sullivan, 2008). So, although these risk factors can be readily detected,
treating them to prevent the incidence of PPD is more complex. Thus, identifying risk factors of
PPD that can be amenable to treatment and increase preventive efforts is necessary and an
important public health concern.
Characteristics of the Couple Relationship and PPD
Over the past two to three decades, research on the transition to parenthood has slowly
shifted to include fathers and dyadic approaches to examine partner effects on PPD. As
mentioned, poor couple relationship functioning is the most consistent risk factor of PPD for
both mothers and fathers (Beck, 2001; Bielawska ‐Batorowicz & Kossakowska ‐Petrycka, 2006;
Condon et al., 2003; deMontigny, Girard, Lacharité, Dubeau, & Devault, 2013b; O’Hara &
Swain, 1996; Ripley et al., 2016). Other more specific aspects of relationship quality and
functioning have also been found to contribute to maternal and paternal PPD. For example, a
recent study found that dyadic consensus on family matters and lifestyle (i.e., friendships, free-
time, religion, money) and affectional expression (i.e., displays of affection and sexual intimacy)
9

were strong predictors of PPD for both mothers and fathers, with low dyadic consensus being the
most robust risk factor.
While poor relationship quality and low partner agreement may influence the
development of PPD, conversely, having a supportive partner may reduce or prevent depression
during the postpartum period (Tanner Stapleton et al., 2012). Research findings on prenatal
relationship factors that buffer PPD symptoms, such as high partner support, can help inform
preventative couple interventions. However, partner support can be intangible and subjective
since actual, received support is less of an indication of well-being than whether an individual
perceives their partner as available to offer support (Reinhardt, Boerner, & Horowitz, 2006).
Additionally, preference for type of support (e.g., instrumental support, emotional support, etc.)
may vary across individuals and situations (Reinhardt et al., 2006), causing it to be a complicated
target of intervention. A couple’s observed interaction, on the other hand, can be dissected for
precise and explicit behaviors, emotions, and communication patterns which are more
discernable, quantifiable, and concrete targets for intervention. Therefore, this study aimed to
identify patterns of observed couple conflict interactions during pregnancy that are potential risk
factors for postpartum depression to help inform dyadic, preventative interventions for couples
during the transition to parenthood.
In addition to relationship quality, partners’ negative moods during pregnancy may also
be a risk factor for later PPD symptoms. For instance, one study that examined the effect of
partner’s mood on maternal and paternal PPD found that self and partner reports of negative
affect states during pregnancy predicted higher PPD for both partners (Morse et al., 2000). A
contagion of distress may occur within some couples, in which partners’ negative mood states
influence each other and create a reciprocal effect. Indeed, the negative affect reciprocity
10

hypothesis suggests that among distressed couples, partners tend to react to each other’s stress or
negative affect states with reciprocal tension and negative affect, further escalating levels of
distress (Gottman, Coan, Carrère, & Swanson, 1998; Levenson & Gottman, 1983). Given that
negative affect reciprocity in couples may be a marker of relationship distress, it would be
valuable to identify whether partners’ reciprocal negative exchanges are predictive of PPD.
Additionally, partners’ negative behavior reciprocity is more concrete than low relationship
quality, making them a more tangible risk factor of PPD to target for intervention. However, no
studies to date have directly tested whether negative affect reciprocity among expecting couples
increases risk for PPD.
Couple Interactions During Pregnancy and PPD
The majority of studies assessing the couple relationship and PPD symptoms are cross-
sectional in nature (Bielawska ‐Batorowicz & Kossakowska ‐Petrycka, 2006; Boyce & Hickey,
2005; Da Costa et al., 2017; Logsdon & Usui, 2001; Wee et al., 2011), and those of which that
use a longitudinal design tend to focus on mothers only (Parade, Blankson, Leerkes,
Crockenberg, & Faldowski, 2014) or fathers only (Gawlik et al., 2014; Pinto, Samorinha,
Tendais, & Figueiredo, 2019) rather than examining the couple as a unit. The few studies that do
examine both mothers and fathers longitudinally, from pregnancy to postpartum, use self-report
measures of relationship features (Bower, Jia, Schoppe-Sullivan, Mangelsdorf, & Brown, 2013;
Canário & Figueiredo, 2016; Figueiredo et al., 2018; Morse et al., 2000).
One of the few studies that assessed the association between couples’ prenatal
interactions and mothers’ and fathers’ PPD found no effect of self-reported positive or negative
couple interactions during pregnancy on postpartum depression (Figueiredo et al., 2018). Yet,
findings indicated that negative interactions reported over multiple time-points from pregnancy
11

to postpartum predicted sharper increases in depressive symptoms from three to 30 months
postpartum. No significant results were reported for positive couple interactions and maternal or
paternal depressive symptoms. However, it is unclear whether these patterns of self-reported
couple interactions and PPD would apply to observed couple conflict interactions.
To our knowledge, only one study has investigated the effect of observed prenatal marital
conflict interactions on depression during the postpartum period and found that couples who
consistently displayed destructive problem-solving behaviors from pregnancy to 24 months
postpartum reported more concurrent depressive symptoms (Houts, Barnett-Walker, Paley, &
Cox, 2008). While this study examined change in conflict patterns over time, assessment of
couples’ observed conflict behavior at a single time-point during pregnancy distinguishes
dynamic patterns that may contribute to subsequent PPD symptoms. This approach allows for
more precise identification of problematic patterns and can better inform interventions designed
to target postpartum psychological functioning. To our knowledge, no study has used a
prospective design to assess observed couple conflict interactions during pregnancy as a
predictor of subsequent maternal and paternal PPD symptoms. Additionally, researchers have not
explored the impact of prenatal negative affect reciprocity as a risk factor for PPD.
Current Study
The current study takes a dyadic approach to identifying risk factors for postpartum
depression (PPD). Couples’ observed conflict communication during pregnancy was assessed as
a risk factor for PPD using a sample of couples recruited in pregnancy and studied through the
first year postpartum. Because of evidence that the couple relationship exerts a particularly
strong influence on fathers’ adjustment to parenthood (Morris, Khaled, Corner, & Saxbe, Under
12

review), we used analyses that explore sex-specific actor and partner pathways for the effect of
prenatal couple conflict behavior on PPD.
Since relationship satisfaction is strongly related to couple’s behavioral exchanges (Karney
& Bradbury, 1995) and observed interactions (Woodin, 2011), we also assessed relationship
satisfaction and included it as a covariate in analyses that examined observed couple conflict
behavior and PPD to highlight unique behavioral contributions to PPD (Baucom, Leo, Adamo,
Georgiou, & Baucom, 2017). Prenatal depressive symptoms are known to be a consistent predictor
of PPD for both mothers and fathers (Leung, Letourneau, Giesbrecht, Ntanda, & Hart, 2017), and
therefore, we also controlled for participants’ reported depressive symptoms during pregnancy in
sensitivity analyses.
This study is one of the first to longitudinally investigate observed prenatal couple
conflict interactions as a risk factor for PPD, making a novel contribution to research on the early
detection, treatment, and prevention of PPD. Couple conflict communication may be more
amenable to intervention than other relationship factors (i.e., overall relationship satisfaction,
partner support) because it is more behaviorally concrete. This study is also the first to
longitudinally measure both mother and father PPD and assess how joint couple behavior
contributes to risk in each parent.
First, we examined associations between conflict behavior and PPD, testing for both
within-person and cross-partner effects. We hypothesized that an individual’s overall levels of
positive behavior across the conflict discussion will predict fewer postpartum depressive
symptoms for both that individual and their partner, whereas overall negative conflict behavior
will predict higher depressive symptoms for both that individual and their partner (H1).
13

Second, we examined the reciprocity or escalation of negative conflict behaviors by
testing whether an individual’s level of negative conflict behavior exhibited during a 30-second
time-segment predicts their partner’s subsequent negative behavior. We hypothesized that
individuals’ negative behavior in one segment would predict their own and their partners’
subsequent negative behavior during the following segment, indicating negative behavior
reciprocity (H2). Finally, we expected that within-person and cross-partner negative behavior
reciprocity would predict maternal and paternal postpartum depressive symptoms (H3).
Methods
Participants
The current study draws from the larger HATCH (Hormones Across the Transition to
Childrearing) study, which investigates couple-level adjustment across the transition to
parenthood. The study follows couples from mid-to-late pregnancy across the first year
postpartum and includes in-person assessments that occur during pregnancy and during the
postpartum period. Participants are cohabitating, opposite-sex couples in Los Angeles expecting
their first child and they comprise a non-clinical, community sample. Participants were recruited
through posting flyers, social media (e.g., advertising to parenting-related online communities),
and word of mouth. The current study uses data from two in-lab visits, one during late pregnancy
and one approximately six months postpartum. During the prenatal visit, informed consent was
obtained from participants and all study procedures were approved by the university’s
Institutional Review Board.
Of the original sample of 95 couples who provided prenatal data, nine couples did not
return to complete the postpartum visit for a variety of reasons (e.g., relocated, no longer available,
didn’t want to participate). Observational coding for the conflict discussion was not possible for
14

two couples (due to missing audio and a corrupted video file transfer), so these two couples were
dropped from analyses. Additionally, data on prenatal relationship satisfaction was missing for
one mother and data on postpartum depression was missing for another mother. These two mothers
and their partners were excluded from analyses due to the dyadic nature of our study testing partner
effects. Therefore, the total sample size is 82 couples.
Procedure
Prenatal visit. The prenatal visit occurred in mid-to-late pregnancy (M = 28.4 weeks of
pregnancy, SD = 4.2 weeks, range = 20 to 39 weeks). During the prenatal visit, participants
engaged in three video-recorded discussions, including the conflict discussion, and completed
questionnaires about their current depressive symptoms and relationship satisfaction.
Approximately one hour into the study visit, participants were given a list of relationship-related
topics (e.g., sex and intimacy, household chores, time spent together, time spent on electronics,
etc.), and instructed to rate the extent they would like to change each area in regard to their
relationship. Following the protocol that has been used in many other observational couple conflict
studies (Gordis, Margolin, Spies, Susman, & Granger, 2010), the experimenter selected the three
topics that appeared to be the greatest sources of conflict (e.g., topics that both partners rated
highly, and/or topics where partners’ ratings were particularly divergent). Couples then discussed
these topics for fifteen minutes and were instructed to start with the topic of greatest concern and
progress to the other topics if time remains. Couples were told:
“Now, we’re going to have you sit together for 15 minutes and talk about areas of
concern in your relationship…These conversations sometimes get intense, and that is
OK. Try to discuss this topic as you would at home, and make sure you each get your
point across.”
15

Following the conflict discussion and the two other discussion tasks (a discussion about hopes
and expectations for the pregnancy, and a discussion about the anticipated division of baby care
responsibilities), participants completed questionnaire measures, including measures of
depressive symptoms and relationship satisfaction.
Postpartum visit. The postpartum visit takes place approximately six to eight months
after the birth with the infant present (M = 7.2 months; SD = 0.8 months; range = 6.0 to 9.8
months postpartum). The visit parallels the prenatal visit with similar tasks and procedures to
facilitate comparison of prenatal and postpartum data. Couples again complete questionnaires,
including measures of postpartum depressive symptoms.
Coding of Couple Conflict Behavior
As stated above, 93 discussions were available for coding. Prenatal couple conflict
behavior for all 93 discussions was assessed using the well-established and widely-used
observational coding scheme, the Specific Affect Coding System (SPAFF; Gottman, McCoy,
Coan, & Collier, 1996). The SPAFF identifies specific affective behavior based on verbal
content, vocal tone, facial expression, gestures, and body movement. The SPAFF includes five
positive codes (affection, enthusiasm, humor, interest, validation), 12 negative codes (anger,
belligerence, contempt, criticism, defensiveness, disgust, domineering, fear/tension, sadness,
stonewalling, threats, whining), and one neutral code. The SPAFF has shown excellent construct
and criterion validity and has been employed in many studies of couple relationship behavior
(Heyman, 2001).
Behavioral coding of the conflict discussion was completed in two parts: the first 49
conflict discussion videos were coded halfway through data collection using the Noldus XT
Observer system (Noldus, 1991) and the next 44 discussions were coded near the completion of
16

data collection using Datavyu (Datavyu Team, 2014). The first set of 49 discussions were coded
by a group of four research assistants. Two raters coded the entire first sample of 49 videos. Due
to attrition of raters, a third rater completed only the first 19 videos, and a fourth completed the
first 12. The second set of 44 discussions were coded by a group of three research assistants; all
three raters coded the entirety of the 44 videos.
Although behavioral coding of the 93 conflict discussions were completed at two
different time periods, all research assistants were trained using the SPAFF manual and were
trained by the same person, the first author. Additionally, although the two sets of research
assistants utilized different software to complete the behavioral coding, observational codes were
entered identically by all raters. Raters entered observational codes continuously (i.e., each time
an emotional behavior was expressed, a code was assigned until a different behavior was
displayed). Thus, for each behavioral code, this yielded a total time in seconds. For each
participant, we summed the duration of all positive behavioral codes and, separately, all negative
behavior codes to reflect the amount of time spent expressing each valanced behavior. Due to
coder disagreement, coders had varying durations for each valanced behavior. Thus, we averaged
the duration of positive behavioral codes across all coders and used this value for analyses. We
conducted the same calculation for negative and neutral behavioral code durations. We then
calculated the proportion of time that each participant spent exhibiting positive behavior by
dividing the duration of positive behavior for a given individual by the total discussion time. We
repeated these calculations for negative behavior. The average proportion of positive conflict
behavior was 0.29 (SD = 0.13, Range = 0.05 – 0.66) for mothers and 0.30 (SD = 0.13, Range =
0.07 – 0.75) for fathers. The average proportion of negative conflict behavior was 0.14 (SD =
0.12, Range = 0.00 – 0.52) for mothers and 0.13 (SD = 0.12, Range = 0.00 – 0.58) for fathers. In
17

other words, this suggests that, on average, mothers spent about 29% of the discussion engaged
in positive behavior and 14% of the discussion engaged in negative behavior, with the rest of the
discussion coded as neutral. Fathers showed similar proportions of behavior. Fathers, on average,
spent about 30% of the discussion displaying positive behavior and 13% of the discussion
displaying negative behavior.
Given that coding was completed by two groups, we assessed whether there were
differences in proportion of positive and negative conflict behavior between these two coding
groups. We separated the conflict behavior variables by group, therefore creating variables for
mothers’ positive and negative conflict behavior and fathers’ positive and negative behavior for
the each of the coding groups. We then conducted t-tests to compare the conflict behavior means
across the two groups. For mothers’ positive conflict behavior, there were no significant
differences between coding group one (M = .27 , SD = .15 ) and group two (M = .29 , SD = .11 ;
t (42) = -.62, p = .54). Likewise, for fathers’ positive conflict behavior, there were no significant
differences between group one (M = .29 , SD = .15 ) and group two (M = .31 , SD = .10 ; t (42)
= -.57, p = .57). However, for negative conflict behavior, group differences were trending
towards significance. For mothers’ negative conflict behavior, group one coded less negative
behavior (M = .12 , SD = .12) than group two (M = .17 , SD = .12 ; t (42) = -.2.05, p = .046.
Similarly, for fathers’ negative conflict behavior, group one coded less negative behavior (M =
.10 , SD = .09) than group two (M = .15 , SD = .13 ; t (42) = -2.00, p = .05). These group
differences could reflect actual differences in negative behavior exhibited by the latter half of
couples in the sample or could indicate distinct approaches to coding negative behavior by the
two groups of coders.
18

For the time-lagged analyses that examined prenatal negative conflict behavior exchanges
(i.e., examining an individual’s behavior in one segment predicting their own behavior and their
partner’s behavior in the next segment), the 15-minute conflict discussion was divided into 30-
second intervals, therefore creating 30 total segments of coded behavior for each partner. For
each segment, the proportion of time spent exhibiting negative behavior during the 30-seconds
was calculated. Therefore, each segment value indicates the extent to which an individual
expressed negative behavior during that interval of time. These time segment values were used
as the unit of analyses for predicting negative escalation across the discussion.
To assess inter-rater reliability among the raters, four intra-class correlations (ICCs;
Bartko, 1966) were calculated for total duration of positive behavior and total duration of
negative behavior. In other words, raters’ agreement of the amount of positive and negative
behavior expressed by participants was compared. ICCs were calculated separately for mothers
and fathers, yielding four ICC values: ICCs for mother positive, mother negative, father positive
and father negative. Then, the four ICCs were averaged to provide one ICC value for each group
of coders. ICC is widely used to assess inter-rater reliability in studies with two or more coders
(Hallgren, 2012). Across discussions, ICCs ranged from .70 - .83, depending on groups of coders
and number of discussions, reflecting moderate to good reliability (Koo & Li, 2016).
Measures
Relationship Satisfaction. Prenatal relationship satisfaction was assessed at the prenatal
visit using the Marital Adjustment Test (MAT), a 15-item self-report scale (Locke & Wallace,
1959) designed to distinguish distressed couples from satisfied couples. The scores for all 15
items were summed and higher scores indicate higher levels of relationship satisfaction. The
MAT is a well-validated measure of relationship satisfaction (Kimmel & Van Der Veen, 1974).
19

Typically, scores above 100 are interpreted to mean that the partner is satisfied with their
relationship (Abramowitz & Sewell, 1980), whereas scores ranging from 60 to 99 suggest
moderate relationship distress, and scores of 59 or lower indicate severe distress (Crane,
Allgood, Larson, & Griffin, 1990). Among the current sample, couples’ mean score was 122.8
(SD = 18.0, Range = 61 – 158), suggesting generally high levels of satisfaction.
Prenatal Depression. Prenatal depressive symptoms were measured at the prenatal visit
using the Beck Depression Inventory (BDI-II; Beck, Steer, & Carbin, 1988). The BDI is a 21-
item self-report questionnaire assessing depressive symptoms experienced in the past two weeks.
Items assess for symptoms including sadness, loss of interest and pleasure, pessimism about the
future, self-blame and self-dislike. Items are rated on a 4-point scale ranging from 0 to 3 (with
the exception of two items assessing increase or decrease of sleep and appetite rated on a 7-point
scale), with higher ratings corresponding to greater symptoms. Total scores range from 0 to 63.
The BDI is a well-established and well-validated measure of depression and has demonstrated
good internal consistency (Beck, Steer, & Carbin, 1988).
Postpartum Depression. Postpartum depressive symptoms were measured at the
postpartum visit using the Edinburgh Postnatal Depression Scale (EPDS; Cox & Holden, 2003).
The EPDS is a 10-item scale assessing postpartum depressive symptoms experienced in the past
7 days. Items assess for symptoms including feeling sad or miserable, self-blame, and looking
forward with enjoyment to activities. Items are rated on a 4-point scale ranging from 0 to 3, with
higher ratings corresponding to greater symptoms. The EPDS is widely-used, has been shown to
have good criterion validity, internal consistency and reliability (Adouard, Glangeaud-
Freudenthal, & Golse, 2005) and has been validated for use with fathers (Matthey, Barnett,
Kavanagh, & Howie, 2001). For the current study, we divided participants’ total EPDS score by
20

number of items, to yield an average item score. Therefore, maximum possible EPDS score for
this sample was 3.
Data Analysis
Analyses were conducted using SPSS Version 26 and Mplus Version 8.4 (Muthén &
Muthén, 2018). Descriptive statistics and bivariate correlations were calculated for key study
variables. T-tests were conducted to compare mothers and fathers on demographic characteristics
and key variables of interest (e.g., conflict behavior, postpartum depression).
Hypothesis 1: Conflict Behavior Predicting PPD. Analyses for hypothesis 1 (i.e.,
examining whether an individual’s prenatal conflict behavior predicts their own and their partner’s
PPD symptoms) were tested using structural equation modeling (SEM), specifically the Actor-
Partner Interdependence Model (APIM) with distinguishable dyads (Kenny, Kashy, Cook, &
Simpson, 2006). APIM addresses interdependence of dyadic data and, as a result, is widely used
when examining couples’ characteristics or behavior during a shared interaction. APIM
simultaneously estimates the effect of an individual’s behavior on their own outcome, or the actor
effect, as well as the effect of an individual’s behavior on their partner’s outcome, or the partner
effect. Positive conflict behavior and negative conflict behavior were tested in separate models
predicting maternal and paternal postpartum depression. All predictor variables and covariates
were grand mean centered. Conflict behavior was calculated to yield the proportion of time that
each participant spent exhibiting positive and negative behavior during the discussion. Therefore,
this proportion was between 0 and 1. We rescaled conflict behavior by multiplying these values
by 10 to aid in the interpretation of model intercepts.
Testing and selection of APIMs followed guidelines described by Fitzpatrick et al. (2016)
and Kenny and Cook (1999). A series of Actor-Partner Interdependence Models (APIM) structural
21

models were run to assess the relationship between positive and negative conflict behavior and
maternal and paternal postpartum depression. First, two fully saturated, unconstrained APIMs
were tested, in which positive and negative conflict behavior were run as predictors in separate
models and included covariates. After identifying the models with at least one significant path, we
then tested those models using four dyadic patterns: actor-only, partner-only, couple-oriented, and
a contrast pattern. For the actor-only pattern, the partner paths are constrained to be zero, whereas
for the partner-only pattern, the actor paths are constrained to be zero. A couple-oriented pattern
allows the actor and partner paths to be constrained to be equal. Finally, the contrast pattern sets
the actor and partner paths to be equal values, but opposite directions. Next, model fit was assessed
by comparing SABIC values across the four models and the original unconstrained APIM. The
pattern with the best fit was then used as the final model in the analyses.
All models controlled for maternal and paternal prenatal relationship satisfaction, the
number of weeks pregnant at the prenatal visit, and the infant’s age at the postpartum visit. Prenatal
relationship satisfaction was included as a covariate to account for the potential influence of
relationship distress on prenatal couple conflict behavior and PPD.
Finally, sensitivity analyses were conducted to assess whether the inclusion or removal of
covariates affected results. Because inclusion of prenatal depressive symptoms as a covariate in
models predicting postpartum depression would examine change in depression across the prenatal
to the postpartum period, altering the primary study question, prenatal depressive symptoms were
instead added as a covariate during sensitivity analyses. Final models were also tested without
controlling for relationship satisfaction.
Hypothesis 2 & 3: Negative Behavior Reciprocity Predicting PPD. Analyses for
hypothesis 2 (determining whether patterns of prenatal negative conflict behavior are related to
22

postpartum depression) were tested using multilevel cross-lagged time series analysis through
dynamic structural equation modeling (DSEM) to investigate the influence of an individual’s
behavior at one timepoint on both their behavior and their partners’ behavior at the next timepoint.
In other words, we tested whether maternal negative conflict behavior at time t – 1 predicted both
maternal and paternal negative behavior at time t, as well as whether paternal negative behavior at
time t – 1 predicted both paternal and maternal negative behavior at time t. Thus, DSEM
simultaneously estimates the time-lagged effects of an individual’s behavior on their own outcome,
or autoregressive effects, as well as the effect of an individual’s behavior on their partner’s
outcome, or cross-lagged effects. We utilized models that allowed for all within-person random
effects to be correlated. Next, we examined associations between each of these paths with PPD.
Finally, we used the autoregressive and cross-lagged effects to predict PPD (i.e., whether negative
escalation exchanges predict higher symptoms of PPD). Maternal and paternal PPD were tested in
separate models.
Similar to data analysis for hypothesis 1, all models controlled for maternal and paternal
prenatal relationship satisfaction, the number of weeks pregnant at the prenatal visit, and the
infant’s age at the postpartum visit. All predictor variables and covariates were grand mean
centered. Also, for this time-series hypothesis, negative conflict behavior was calculated to yield
the proportion of time that mothers and fathers spent exhibiting negative behavior during each 30-
second segment across the discussion. Therefore, this proportion was between 0 and 1. We rescaled
negative conflict behavior by multiplying these values by 10 to aid in the interpretation of model
intercepts.
Results
23

Sample demographics including race and education are displayed in Table 1. Descriptive
statistics for key study variables and t-tests comparing mothers and fathers are reported in Table
2. T-tests indicated that within our sample, mothers were younger than fathers and reported
higher prenatal relationship satisfaction, prenatal depressive symptoms, and postpartum
depressive symptoms than fathers. There were no significant differences in expression of
positive or negative conflict behavior between mothers and fathers.
Chapter1: Table 1
Participant and couple characteristics

Female
Participants
(n = 82)

Male
Participants
(n = 82)

Participant Characteristics n (%)
Race
Asian / Pacific Islander 15 (18.3%) 15 (18.3%)
Black / African American 5 (6.1%) 6 (7.3%)
Hispanic / Latinx 19 (23.2%) 16 (19.5%)
White / Caucasian 37 (45.1%) 41 (50.0%)
Other 6 (7.3%) 4 (4.9%)
Educational Attainment
High School / GED 1 (1.2%) 2 (2.4%)
Some College 12 (14.7%) 13 (15.9%)
College Degree 31 (37.8%) 40 (48.8%)
Master’s Degree 30 (36.6%) 16 (19.5%)
Professional or Doctoral Degree 8 (9.8%) 11 (13.4%)
Couple Characteristics
Married 68 (82.9%)
Dating/Cohabiting 14 (17.1%)

Chapter 1: Table 2
Descriptives for Key Study Variables

Female Participants Male Participants

M (SD)
Range
M (SD)
Range
t-test
a

Study Variables
Age
31.3 (4.31)
21 - 39
33.3 (5.64)
22 - 57
-4.23***
Positive Conflict Behavior
.29 (.13)
.05 - .66
.30 (.13)
.07 - .75
-.91
24

Negative Conflict Behavior
.14 (.12)
0 - .52
.13 (.12)
0 - .58
1.34
Prenatal Relationship Satisfaction
125.0 (16.65)
68 - 158
120.5 (19.36)
61 - 158
2.24*
Prenatal Depression Score (BDI)
10.0 (5.35)
0 - 30
7.9 (6.26)
0 - 37
2.88**
Postpartum Depression Score (EPDS)
.69 (.50)
0 – 2.0
.47 (.43)
0 – 1.78
3.29**
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10
a
Critical values are from paired samples t-tests evaluating gender differences on key study variables.
Abbreviations: BDI = Beck Depression Inventory; EPDS = Edinburgh Postnatal Depression Scale;

Zero-order correlations for all variables are shown in Table 3. Positive and negative
conflict behaviors were not correlated with PPD for mothers or fathers, but both partners’
negative conflict behavior was positively associated with paternal prenatal depression. Conflict
behaviors were positively associated within couples, and negative and positive conflict behaviors
were inversely correlated for both mothers and fathers. Prenatal relationship satisfaction was
inversely associated with negative conflict behavior for both partners, positively associated with
positive conflict behavior for mothers, and inversely associated with prenatal depression for both
partners and postpartum depression for fathers. Mothers’ and fathers’ prenatal depression were
positively associated with each other, yet mothers’ and fathers’ postpartum depression were not
significantly associated. However, for both mothers and fathers, prenatal depression was
positively associated with their own postpartum depression. Moreover, fathers’ postpartum
depression was negatively associated with both mothers’ and fathers’ prenatal relationship
satisfaction and prenatal depression.
Hypothesis 1: Conflict Behavior Predicting PPD
For the model with positive conflict behavior predicting postpartum depression, at least
one path was significant. Therefore, we tested this model with the four dyadic patterns described
above and compared model fit. However, no paths were significant in the model with negative
25

conflict behavior predicting postpartum depression. Thus, individual or partner expression of
negative conflict behavior was not related to mothers’ or fathers’ postpartum depression. Since
there was not at least one significant path in the initial negative conflict behavior APIM, we did
not test negative behavior with the four dyadic patterns.
Chapter 1: Table 3
Zero-order correlations for all study and control variables
1 2 3 4 5 6 7 8 9 10 11
1. Maternal Positive Conflict Behavior -
2. Paternal Positive Conflict Behavior .62** -
3. Maternal Negative Conflict Behavior
-.33** -.19
⊥
-
4. Paternal Negative Conflict Behavior -.19
⊥
-.29** .73** -
5. Maternal Prenatal Relationship
Satisfaction
.22* .13 -.26* -.26* -
6. Paternal Prenatal Relationship
Satisfaction
.23* .16 -.36** -.30** .50** -
7. Maternal Prenatal Depression (BDI) -.11 -.05 .12 .15 -.60** -.31** -
8. Paternal Prenatal Depression (BDI) -.10 -.18
⊥
.32** .27** -.32** -.47** .36** -
9. Maternal Postpartum Depression
(EPDS)
-.06 .16 .01 -.05 -.17 -.02 .32** -.13 -
10. Paternal Postpartum Depression
(EPDS)
-.02 -.03 .10 .12 -.25* -.44** .31** .54** .16 -
11. Weeks Pregnant at Prenatal Visit .01 -.04 -.13 -.10 -.02 .16 -.09 -.14 -.15 -.20 -
12. Infant Age at Postpartum Visit .08 .06 -.09 -.17 .03 .21
⊥
.07 .04 .13 .30** -.07
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10
Abbreviations: BDI = Beck Depression Inventory; EPDS = Edinburgh Postnatal Depression Scale;
When comparing model fit among the unconstrained and dyadic pattern models for
positive conflict behavior, a contrast model emerged as the best fit for both mothers’ and fathers’
postpartum depression (i.e., produced the lowest SABIC values). Table 4 presents the
unstandardized effects for the initial fully-saturated, unconstrained models for both positive and
negative conflict behavior, as well as the best fitting model for positive conflict behavior. Figure
1 also displays the fully-saturated unconstrained model for negative conflict behavior and the
best fitting model for positive conflict behavior. Consistent with hypothesis 1, mothers’ prenatal
positive conflict behavior predicted lower maternal postpartum depression. However,
26

unexpectedly, fathers’ positive conflict behavior predicted higher maternal postpartum
depression. Neither mothers’ nor fathers’ positive conflict behaviors were associated with
paternal postpartum depression.
Chapter 1: Table 4
APIM results from models predicting maternal and paternal PPD

Unconstrained Model

Best Fitting Model
(Contrast Pattern)
Models with Positive Conflict Behavior β (SE) β (SE)
Intercept Mothers .69*** (.05) .69*** (.05)
Intercept Fathers .45*** (.04) .45*** (.04)
Residual Variance Mothers .20*** (.03) .20*** (.03)
Residual Variance Fathers .15*** (.02) .15*** (.02)
Effects of Positive Conflict Behavior
Mothers’ Positive → Mothers’ PPD (Actor) -.10* (.05) -.12** (.05)
Fathers’ Positive → Mothers’ PPD (Partner) .14** (.05) .12** (.05)
Fathers’ Positive → Fathers’ PPD (Actor) -.03 (.04) -.04 (.04)
Mothers’ Positive → Fathers’ PPD (Partner) .05 (.04) .04 (.04)
Predictor & Residual Covariance
Mothers’ Positive → Fathers’ Positive 1.18*** (.24) 1.18*** (.24)
Mothers’ PPD → Fathers’ PPD .03 (.02) .03 (.02)
Models with Negative Conflict Behavior
Intercept Mothers .70*** (.05) -
Intercept Fathers .46*** (.04) -
Residual Variance Mothers .21*** (.03) -
Residual Variance Fathers .14*** (.02) -
Effects of Negative Conflict Behavior
Mothers’ Negative → Mothers’ PPD (Actor) -.01 (.06) -
Fathers’ Negative → Mothers’ PPD (Partner) -.06 (.06) -
Fathers’ Negative → Fathers’ PPD (Actor) .05 (.05) -
Mothers’ Negative → Fathers’ PPD (Partner) -.07 (.05) -
Predictor & Residual Covariance
Mothers’ Negative → Fathers’ Negative 1.03*** (.20) -
Mothers’ PPD → Fathers’ PPD .02 (.02) -
Significance: * = p < .05, ** = p < .01, *** = p < .001
All models include n = 82 couples and control for both partners’ reports of prenatal relationship satisfaction,
weeks pregnant at the prenatal visit, and infant age at the postpartum visit. All variables are grand mean centered
to aid in the interpretation of model intercepts. Unstandardized effects are reported.
Abbreviations: PPD = Postpartum depression;
Sensitivity Analyses. Sensitivity analyses were tested for positive conflict behavior
models. After adding prenatal depression as a covariate to the final positive behavior model,
results remained the same in regards to significance, direction, and magnitude of effects.
Likewise, when prenatal relationship satisfaction was removed from the final model, results
27

remained unchanged in significance, direction and magnitude. Sensitivity analyses are reported
in Table 5.
Chapter 1: Figure 1. APIM results from models predicting maternal and paternal PPD.

The top panel depicts results for the best fitting positive conflict behavior model predicting mothers’ and
fathers’ postpartum depression. The bottom panel depicts results for the fully-saturated, unconstrained
negative conflict behavior model predicting mothers’ and fathers’ postpartum depression. Both models
include n = 82 couples and control for both partners’ reports of prenatal relationship satisfaction, weeks
pregnant at the prenatal visit, and infant age at the postpartum visit.

Chapter 1: Table 5
Sensitivity analyses for APIM results with positive conflict behavior models

Best Fitting
Model
(Contrast pattern)
Model with
Prenatal
Depression Added

Model with
Relationship
Satisfaction
Removed
Models with Positive Conflict Behavior β (SE) β (SE) β (SE)
Intercept Mothers .69*** (.05) .65*** (.05) .69*** (.05)
Intercept Fathers .45*** (.04) .49*** (.04) .47*** (.05)
Residual Variance Mothers .20*** (.03) .19*** (.03) .22*** (.03)
28

Residual Variance Fathers .15*** (.02) .12*** (.02) .17*** (.03)
Effects of Positive Conflict Behavior
Mothers’ Positive → Mothers’ PPD (Actor) -.12** (.05) -.12** (.04) -.12* (.05)
Fathers’ Positive → Mothers’ PPD (Partner) .12** (.05) 12** (.04) .12* (.05)
Fathers’ Positive → Fathers’ PPD (Actor) -.04 (.04) -.02 (.04) -.02 (.04)
Mothers’ Positive → Fathers’ PPD (Partner) .04 (.04) .02 (.04) .02 (.04)
Predictor & Residual Covariance
Mothers’ Positive → Fathers’ Positive 1.18*** (.24) 1.18*** (.24) 1.16*** (.24)
Mothers’ PPD → Fathers’ PPD .03 (.02) .04* (.02) .03 (.02)
Significance: * = p < .05, ** = p < .01, *** = p < .001
All models include n = 82 couples and control for both partners’ reports of prenatal relationship satisfaction,
weeks pregnant at the prenatal visit, and infant age at the postpartum visit. All variables are grand mean centered
to aid in the interpretation of model intercepts. Unstandardized effects are reported.
Abbreviations: PPD = Postpartum depression;

Hypothesis 2: Negative Escalation Exchanges Predicting PPD
Descriptives for Negative Behavior Within Segments. Table 6 presents unstandardized
effects for intercepts and variances for negative behavior during time-segments as well as
negative escalation exchanges. For each individual, we calculated the within-person mean of
negative behavior across time-segments (i.e., across the entire discussion, the person’s average
level of negative conflict behavior). Next, we computed the average of all individuals’ mean
levels of negative behavior.
Chapter 1: Table 6

Intercepts and Variances for Negative Behavior Segments and Reciprocity Effects
Intercepts and variances
Effect Posterior Median 95% Credible Interval
Grand mean (Mother negative seg) -.03 [-.24, .18]
Grand mean (Father negative seg) .00 [-.51, -.19]
Within-Person Residual Variance (Mother negative seg) 1.92 [1.81, 2.04]
Within-Person Residual Variance (Father negative seg) 1.73 [1.64, 1.84]
Between-Person Variance (Mother negative seg) .64 [.39, 1.01]
Between-Person Variance (Father negative seg) .58 [.34, .99]
Variance (Mother t-1 to Mother t) .03 [.01, .05]
Variance (Father t-1 to Father t) .04 [.02, .08]
Variance (Mother negative t-1 to Father negative t ) .05 [.02, .09]
Variance (Father negative t-1 to Mother negative t) .04 [.02, .08]
Regression path intercepts
Predictor at time t - 1 Outcome at time t Posterior Median 95% credible interval
29

Mother negative Mother negative .36 [.29, .42]
Father negative Father negative .44 [.36, .51]
Mother negative Father negative .11 [.04, .19]
Father negative Mother negative .14 [.07, .22]
Note. Unstandardized effects are reported. Bolded values indicate significant effects in which the credible interval
does not contain 0. Negative seg refers to proportion of negative behavior exhibited in a 30-second time segment.
Across participants’ within-person means of negative conflict behavior, the average of all
mothers’ mean negative behavior was -.03 and the average of all fathers’ mean negative behavior
was zero. These within-person means were non-significant, indicating that on average,
individuals’ level of negative behavior did not differ from zero. Yet, the within-person residual
variance for mothers’ and fathers’ negative conflict behavior indicated that there are large
within-person fluctuations in negative behavior across the discussion. The average residual
variance of mothers’ negative behavior was 1.92 and, comparably, the average residual variance
of fathers’ negative behavior was 1.73. This suggests that the amount of negative behavior
expressed at any given point in the discussion may vary widely for a given person.
Additionally, we found large between-mother variability of .64 (SD = sqrt(.64) = .80)),
indicating that some mothers had higher or lower within-person means than the average for all
mothers. Across fathers’ mean levels of negative conflict behavior, the average of all fathers’
mean negative behavior was .01 and similarly, there was large between-father variability of .58
(SD = sqrt(.58) = .76)), indicating that some fathers have higher or lower within-person means
than the average for all fathers. In other words, there are significant within-person fluctuations as
well as between-person differences for average negative behavior expressed during the
discussion for both mothers and fathers. The ICC in the context of dynamic multilevel modeling
is calculated as the ratio of between-person variance to total variance, indicating that the ICC is
the proportion of total variance that is accounted for by more stable between-person differences.
The ICCs for negative escalation is .28 for mothers and .22 for fathers. This suggests that 28% of
30

the total variance of mothers’ negative escalation is at the between-person level and 22% of the
total variance of fathers’ negative escalation is at the between-person level.
Autoregressive and Cross-Lagged Effects for Negative Escalation. The results
indicate significant inertia (i.e., autoregressive effects) in mothers’ and fathers’ negative conflict
behavior, such that an individuals’ higher negative conflict behavior at time t – 1 predicts their
higher negative behavior at time t. More specifically, every one-unit increase in mothers’
negative behavior at time t – 1 corresponds to a .36-point increase in mothers’ own negative
behavior at time t and every one-unit increase in fathers’ negative behavior at time t – 1
corresponds to a .44-point increase in fathers’ own negative behavior at time t. For both mothers’
and fathers’ autoregressive effects, results show little between-person variability.
Additionally, as expected, mothers’ and fathers’ cross-lagged escalation, or partner
effects, in negative behavior was also significant. Mothers’ negative conflict behavior at time t –
1 predicts fathers’ negative behavior at time t, such that a one-unit increase in mothers’ negative
behavior at time t – 1 predicts a .11-point increase in fathers’ negative behavior at time t.
Similarly, a one-unit increase in fathers’ negative behavior at time t – 1 predicts a .14-point
increase in mothers’ negative behavior at time t. Results show that there is some between-couple
variability among these paths of .05 (SD = sqrt(.05) = .22)) for mothers’ influencing fathers’
subsequent negative affect and .04 (SD = sqrt(.04) = .20)) for fathers’ influencing mothers’
subsequent negative affect.
Hypothesis 3: Negative Escalation Exchanges Predicting PPD
Table 7 displays standardized effects for associations between negative escalation
variables and maternal and paternal PPD. None of the autoregressive and cross-lagged negative
escalation effects were associated with maternal or paternal PPD. Next, we tested whether
31

negative escalation exchanges predict PPD. Results for between-level unstandardized effects are
presented in Table 8. All paths testing negative behavior predicting PPD were non-significant.
Therefore, autoregressive and cross-lagged negative escalation exchanges were not predictive of
maternal or paternal PPD.
Chapter 1: Table 7

Associations between Negative Escalation and PPD
Negative Escalation Paths Posterior Median 95% credible interval
Maternal PPD
Mother t-1 to Mother t (Autoregressive) .26 [-.01, .06]
Father t-1 to Father t (Autoregressive) -.04 [-.05, .04]
Mother negative t-1 to Father negative t (Cross-Partner) .11 [-.02, .06]
Father negative t-1 to Mother negative t (Cross-Partner) -.01 [-.02, .06]
Paternal PPD
Mother t-1 to Mother t (Autoregressive) -.03 [-.05, .01]
Father t-1 to Father t (Autoregressive) .01 [-.04, .04]
Mother negative t-1 to Father negative t (Cross-Partner) .04 [-.03, .04]
Father negative t-1 to Mother negative t (Cross-Partner) .09 [-.03, .05]
Note. Standardized effects for correlations are reported. Bolded values indicate significant effects in which the
credible interval does not contain 0.

Chapter 1: Table 8

Between-Level Effects for Negative Escalation Predicting PPD

Negative Escalation Paths Posterior Median 95% Credible Interval
Outcome: Maternal Postpartum Depression
Mother t-1 to Mother t (Autoregressive) 1.22 [-.45, 3.91]
Father t-1 to Father t (Autoregressive) -.22 [-1.22, .76]
Mother negative t-1 to Father negative t (Cross-Partner) .43 [-.50, 1.42]
Father negative t-1 to Mother negative t (Cross-Partner) -.16 [-1.39, 1.02]
Outcome: Paternal Postpartum Depression
Mother t-1 to Mother t (Autoregressive) -.82 [-2.65, .54]
Father t-1 to Father t (Autoregressive) .03 [-.83, .93]
Mother negative t-1 to Father negative t (Cross-Partner) .15 [-.65, 1.05]
Father negative t-1 to Mother negative t (Cross-Partner) .35 [-.64, 1.61]
Note. Unstandardized effects are reported. Bolded values indicate significant effects in which the credible interval
does not contain 0.
Finally, covariates were added to the model, including maternal and paternal relationship
satisfaction, number of weeks pregnant at the prenatal visit, and infant age at the postpartum
visit. Results for between-level covariate effects are presented in Table 9. All main paths testing
negative behavior predicting PPD remained non-significant.
32

Chapter 1: Table 9

Between-Level Effects for Negative Escalation Predicting PPD With Covariates
Predictors and Covariates Posterior Median 95% Credible Interval
Outcome: Maternal Postpartum Depression
Predictors
Mother t-1 to Mother t (Autoregressive) 1.33 [-.31, 3.77]
Father t-1 to Father t (Autoregressive) -.21 [-1.20, .74]
Mother negative t-1 to Father negative t (Cross-Partner) .50 [-.45, 1.45]
Father negative t-1 to Mother negative t (Cross-Partner) .02 [-1.27, 1.25]
Covariates
Mothers’ relationship satisfaction -.01 [-.02, -.01]
Fathers’ relationship satisfaction -.01 [-.01, .01]
Weeks pregnant .00 [-.01, .00]
Infant age .02 [-.01, .06]
Outcome: Paternal Postpartum Depression
Predictors
Mother t-1 to Mother t (Autoregressive) -.65 [-2.33, .72]
Father t-1 to Father t (Autoregressive) .21 [-.61, 1.02]
Mother negative t-1 to Father negative t (Cross-Partner) .06 [-.69, .81]
Father negative t-1 to Mother negative t (Cross-Partner) .21 [-.74, 1.37]
Covariates
Mothers’ relationship satisfaction .00 [-.01, .00]
Fathers’ relationship satisfaction -.01 [-.02, -.01]
Weeks pregnant .00 [.00, .00]
Infant age .03 [-.01, .06]
Note. Unstandardized effects are reported. Bolded values indicate significant effects in which the credible interval
does not contain 0.
Discussion
This study examined couple’s prenatal conflict behavior as a predictor of PPD and tested
whether reciprocal patterns of negative conflict behavior predicted PPD for mothers and fathers.
As expected, mothers’ prenatal positive conflict behavior predicted fewer maternal postpartum
depressive symptoms. However, contrary to expectations, fathers’ positive conflict behavior
predicted higher maternal postpartum depressive symptoms. These results remained significant
after adding prenatal depression as a covariate or removing relationship satisfaction from the
final positive behavior model. Negative conflict behavior did not predict either maternal or
paternal PPD. We also found that evidence for cross-partner escalation of negative conflict
behaviors, meaning when individuals exhibit negative behavior during one segment, their
partners are more likely to respond with negative behavior during the following segment. Thus,
33

these results are consistent with our expectation that negative behavior reciprocity would appear
within this sample. However, contrary to hypotheses, we did not find that patterns of negative
reciprocity during prenatal conflict predicted either maternal or paternal PPD symptoms.
As hypothesized, mothers who expressed more positive behavior during the conflict
discussion were less likely to develop PPD symptoms. We controlled for mothers’ prenatal
relationship satisfaction, suggesting that these findings are not explained only by mothers’
positive feelings towards their partners. This finding may reflect trait-level effects, such that
mothers with more positive dispositions overall tend to exhibit higher levels of positive behavior
during the discussion and also to feel more positively during the postpartum period. It is also
possible that mothers who behaved more positively with their partners were able to elicit more
support from those partners over the transition to parenthood. However, surprisingly, neither
positive nor negative conflict behavior predicted PPD for fathers. Future studies should continue
to explore maternal prenatal conflict behavior as a potential buffer of postpartum depression.
Surprisingly, and inconsistent with hypothesis 1, fathers’ positive conflict behavior
predicted higher maternal PPD symptoms. It is possible that when fathers exhibit positive
behavior during challenging situations, such as relationship conflict, they may inadvertent
dismiss mothers’ relational concerns. This result may also be a sign of fathers’ sense of denial or
avoidance of acknowledging difficult emotions during the conflict discussion, with a
continuation of avoidance of challenging emotions or topics during the postpartum period.
Indeed, other work has found that mothers who reported that their partners engaged in more
avoidance strategies during conflict were more likely to develop PPD symptoms six months
postpartum (Parade et al., 2014). Given that low perceived partner support is a direct risk factor
contributing to the development of maternal PPD (Dennis & Ross, 2006), partners’ avoidance
34

may translate into mothers’ perception of low partner support once the baby arrives, facilitating
maternal PPD symptoms. Additionally, if mothers’ attempt at discussing difficult issues with
their partners are met with positive, yet patronizing platitudes or appeasing behavior, it may
discourage mothers from continuing to express negative emotions to their partner, leading to
negative emotional suppression. In fact, individuals diagnosed with major depressive disorder
(MDD) show an increased suppression of negative emotions (Beblo et al., 2012). Moreover, the
conflict discussion prompt of the current study is for partners to discuss topics that they would
like to change in their relationship. These results may suggest that expressing high levels of
positive behavior to a partner while they discuss areas of the relationship they would like to
change could be counterproductive. More research is needed to elucidate these unexpected yet
intriguing results. In addition to examining observed behavior, future studies investigating
couple conflict discussions could assess for perceptions of partners’ responses to determine if
behavior coded as ‘positive’ is received as such by their partner.
Results remained significant after adding prenatal depression as a covariate to the final
positive behavior model. Thus, higher maternal prenatal positive conflict behavior was
associated with greater decreases in maternal depressive symptoms from prenatal to postpartum,
whereas higher paternal prenatal conflict behavior was associated with greater increases in
maternal depressive symptoms across the transition to parenthood. Likewise, when prenatal
relationship satisfaction was removed from the final model, results remained unchanged in
significance, direction and magnitude.
Whereas mothers’ and fathers’ positive conflict behavior was associated with maternal PPD
symptoms, positive conflict behavior was not associated with paternal PPD. In fact, positive and
negative conflict behavior were unrelated to paternal PDD. These findings are unexpected given
35

other evidence that the couple relationship shapes fathers’ adjustment to parenthood (Morris,
Khaled, Corner, & Saxbe, Under review)
While relationship satisfaction and quality have been shown to be a predictor of paternal PPD
(Gawlik et al., 2014; Wee et al., 2011), the fathers in this sample generally reported high levels
of satisfaction in their relationships, suggesting that these fathers may not be at great risk of
developing PPD. Additionally, we controlled for relationship satisfaction to underscore the
distinctive behavioral contributions to PPD, therefore perhaps removing the foremost influence
of the couple relationship on paternal PPD. However, when relationship satisfaction was
removed from the final positive conflict behavior model, effects of positive behavior on paternal
PPD remained non-significant. More recent research studies have found that among the most
robust risk factors for paternal PPD, aspects of parenting may largely contribute to the
development of PPD symptoms for fathers. In addition to relationship quality, parenting distress
and low perceived parenting efficacy (deMontigny, Girard, Lacharité, Dubeau, & Devault,
2013a) as well as lacking coparenting support (Bronte-Tinkew, Moore, Matthews, & Carrano,
2007) put fathers at a higher likelihood of reporting depressive symptoms postpartum. Perhaps
for fathers who report high levels of relationship satisfaction, adjustment to and challenges with
parenting and infant care may contribute to their PPD to a greater extent than couple interactions
during pregnancy. Moreover, another robust risk factor putting fathers’ at risk of developing
PPD is their partners’ PPD (Gawlik et al., 2014; Goodman, 2008; Wee et al., 2011). For fathers
whose partners (i.e., coparents) are depressed, this may lead to greater parenting and infant care
difficulty for fathers, compounding fathers’ risk for PPD. Future work can explore the extent to
which some factors pose a greater risk for fathers who report high relationship quality.
Inconsistent with our hypotheses, negative conflict behavior did not predict PPD
36

symptoms within either mothers and fathers. To a certain extent, negative behavior is to be
expected during a conflict discussion and indeed, satisfied romantic relationships are not void of
negative emotion (Richards, Butler, & Gross, 2003). Perhaps for the couples in our sample, who
report generally high levels of satisfaction with their partners, negative behavior during a conflict
discussion is normative and not indicative of poor adjustment during the postpartum period.
Additionally, for these couples, the average amount of negative behavior observed across these
discussions was relatively low. On average, couples in our study spent less than one-fifth of the
conflict discussion exhibiting negative behavior. In other words, out of the 15 minutes spent
engaging in the conflict discussion, on average, couples were observed exhibiting negative
behavior for approximately two of those minutes. Thus, other features of the interactions may
carry more weight, such as the ratio of positive to negative emotion or the specific types of
negative behavior expressed, such as contempt for the partner (Carrère & Gottman, 1999;
Gottman, 1999; Gottman, Coan, Carrère, & Swanson, 1998). Another study that investigated the
effect of observed relationship conflict and depressive symptoms in couples over the transition to
parenthood also found no relationship between negative marital interactions and depressive
symptoms during the prenatal period (Cox, Paley, Payne, & Burchinal, 1999). Although
observed conflict interactions and depressive symptoms were measured at the same time-points
in that study, it may also suggest that some negative interactions in the context of conflict may
serve a valuable purpose, for partners to express their needs and encourage desired changes to
enhance closeness and intimacy. Moreover, although we coded conflict behaviors, we did not
measure how partners perceived each other’s behavior. It may be that partner perceptions of
conflict behavior are more important for future adjustment than the behaviors themselves.
This study also found evidence of negative conflict behavior reciprocity in this sample. In
37

other words, when individuals exhibited negative behavior during one segment of time, they, as
well as their partners, were more likely to display negative behavior in the following segment.
This cyclical pattern of negative behavior is analogous to the negative affect reciprocity
hypothesis (Gottman et al., 1998; Levenson & Gottman, 1983), in which couples show a
continual pattern of negative affect exchanges during conflict. Distressed couples are more likely
to display these patterns of negative affect reciprocity, indicating that these patterns may be
indicative of or lead to poor relationship quality. However, at times, this feedback loop of
negativity during conflict may be inevitable, even for highly satisfied couples. In those moments,
attempts to repair the conflict by reducing negative affect and increasing positive affect may
distinguish couples who fare well from those who don’t (Gottman, Driver, & Tabares, 2015).
That is to say that negative affect or behavior reciprocity is not a guaranteed signal of adverse
outcomes. This is consistent with the results of this study in which within-person and within-
partner negative behavior reciprocity did not predict later PPD symptoms for mothers or fathers.
Perhaps among highly satisfied couples, negative behavior reciprocity may not have the same
deleterious implications for the emotional health of the relationship or partners as couples who
report relationship distress.
To our knowledge, this is the first study to examine couples’ observed conflict behavior
during pregnancy as a dyadic, longitudinal risk factor for maternal and paternal PPD symptoms
as well as test the effect of negative behavior reciprocity on PPD. While the majority of studies
investigating relationship quality, conflict, and PPD use self-report measures to assess for
relationship satisfaction and characteristics of conflict interactions, the current study utilized
observed conflict interactions which provide a more nuanced and richer snapshot of the couple’s
dynamics. Additionally, we utilized advanced statistical analyses to account for dyadic as well as
38

time-lagged intensive longitudinal data. This is the first study to assess couples’ interdependence
and reciprocal influence of observed conflict behavior on PPD symptoms. The longitudinal
design of this study also allows for examining the lasting influence of prenatal couple conflict
behavior on postpartum adjustment. However, even with longitudinal models, any potential
causal relationship needs to be interpreted with caution.
Limitations of this study include our reliance on a non-clinical, community sample
comprised of highly-educated, opposite-sex, highly-satisfied couples. Future research utilizing a
more diverse sample of couples is warranted. Comparing these effects among participants of
various backgrounds, same-sex couples, or within a clinical sample would contribute to the
understanding of how PPD risk factors vary across groups of individuals. Additionally, the
observational coding of the conflict discussions in this study were conducted by two groups of
research assistants. Ideally, the same raters would code all videos to ensure a reliable approach to
coding. Finally, across the sample, the average proportion of combined positive and negative
conflict behavior was 43%. In other words, on average, couples expressed valanced affective
behavior for less than half of the discussion, and exhibited neutral behavior for a majority of the
conflict discussion. It may be reasonable to assume that these proportions of valanced and
neutral behavior are not representative of these couples’ at-home conflict discussions. Therefore,
social desirability effects may have impacted partners’ expression of affective behavior in the
lab, producing dampened emotionality.
Despite these limitations, this study provides an important first step in understanding the
link between couples’ conflict interactions during pregnancy and postpartum depression for first-
time parents. Mothers’ own expression of positivity during prenatal conflict may be an indicator
of advantageous postpartum mood adjustment. However, fathers’ displays of positivity toward
39

their partners during conflict may not be as supportive as expected and instead, contribute to
mothers’ emotional difficulties during the postpartum period. Future studies can continue to
elucidate this relationship between positive conflict behavior and postpartum depression to help
inform interventions designed to prevent PPD. Additionally, research can investigate for which
couples and in which contexts positive and negative conflict behavior may be maladaptive for
postpartum mental health adjustment. Researchers can apply these questions to more naturalistic
settings and collect conflict interaction data through ecological methods to increase external
validity and generalizability. These intriguing findings open the door to further exploration and
set the scene to better understand how couples interactions can influence their mental health
adjustment over the transition to parenthood.

40

References
Abramowitz, S. I., & Sewell, H. H. (1980). Marital adjustment and sex therapy outcome. Journal
of Sex Research, 16(4), 325–337. https://doi.org/10.1080/00224498009551089
Adouard, F., Glangeaud-Freudenthal, N. M. C., & Golse, B. (2005). Validation of the Edinburgh
postnatal depression scale (EPDS) in a sample of women with high-risk pregnancies in
France. Archives of Women’s Mental Health, 8(2), 89–95.
Arteche, A., Joormann, J., Harvey, A., Craske, M., Gotlib, I. H., Lehtonen, A., … Stein, A.
(2011). The effects of postnatal maternal depression and anxiety on the processing of infant
faces. Journal of Affective Disorders, 133(1–2), 197–203.
Austin, M., Priest, S. R., & Sullivan, E. A. (2008). Antenatal psychosocial assessment for
reducing perinatal mental health morbidity. The Cochrane Database of Systematic Reviews,
(4). https://doi.org/10.1002/14651858.CD005124.pub2.Antenatal
Bartko, J. J. (1966). The intraclass correlation coefficient as a measure of reliability.
Psychological Reports, 19(1), 3–11.
Baucom, B. R. W., Leo, K., Adamo, C., Georgiou, P., & Baucom, K. J. W. (2017). Conceptual
and Statistical Issues in Couples Observational Research : Rationale and Methods for
Design Decisions, 31(8), 972–982.
Beblo, T., Fernando, S., Klocke, S., Griepenstroh, J., Aschenbrenner, S., & Driessen, M. (2012).
Increased suppression of negative and positive emotions in major depression. Journal of
Affective Disorders, 141(2–3), 474–479. https://doi.org/10.1016/j.jad.2012.03.019
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988a). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8(1),
77–100.
41

Beck, A. T., Steer, R. A., & Carbin, M. G. (1988b). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8(1),
77–100. https://doi.org/10.1016/0272-7358(88)90050-5
Beck, C. T. (2001). Predictors of Postpartum Depression. Nursing Research, 50(5), 275–285.
https://doi.org/10.1097/GRF.0000000000000368
Bielawska ‐Batorowicz, E., & Kossakowska ‐Petrycka, K. (2006). Depressive mood in men after
the birth of their offspring in relation to a partner’s depression, social support, fathers’
personality and prenatal expectations. Journal of Reproductive and Infant Psychology,
24(1), 21–29. https://doi.org/10.1080/02646830500475179
Bower, D., Jia, R., Schoppe-Sullivan, S. J., Mangelsdorf, S. C., & Brown, G. L. (2013).
Trajectories of couple relationship satisfaction in families with infants. Journal of Social
and Personal Relationships, 30(4), 389–409. https://doi.org/10.1177/0265407512457656
Boyce, P., & Hickey, A. (2005). Psychosocial risk factors to major depression after childbirth.
Social Psychiatry and Psychiatric Epidemiology, 40(8), 605–612.
https://doi.org/10.1007/s00127-005-0931-0
Brand, S. R., & Brennan, P. A. (2009). Impact of antenatal and postpartum maternal mental
illness: How are the children? Clinical Obstetrics and Gynecology, 52(3), 441–455.
https://doi.org/10.1097/GRF.0b013e3181b52930
Bronte-Tinkew, J., Moore, K. A., Matthews, G., & Carrano, J. (2007). Symptoms of major
depression in a sample of fathers of infants. Journal of Family Issues, 28(1), 61–99.
https://doi.org/10.1177/0192513X06293609
Burke, L. (2003). The impact of maternal depression on familial relationships. International
Review of Psychiatry, 15(3), 243–255.
42

Cadzow, S. P., Armstrong, K. L., & Fraser, J. A. (1999). Stressed Parents with Infants:
Reassessing Physical Abuse Risk Factors. Child Abuse & Neglect, 23(9), 845–853.
https://doi.org/10.1016/S0145-2134(99)00063-0
Canário, C., & Figueiredo, B. (2016). Partner relationship from early pregnancy to 30 months
postpartum: Gender and parity effects. Couple and Family Psychology: Research and
Practice, 5(4), 226–239. https://doi.org/10.1037/cfp0000066
Carrère, S., & Gottman, J. M. (1999). Predicting divorce among newlyweds from the first three
minutes of a marital conflict discussion. Family Process, 38(3), 293–301.
https://doi.org/10.1111/j.1545-5300.1999.00293.x
Collins, N. L., Dunkel-Schetter, C., Lobel, M., & Scrimshaw, S. C. M. (1993). Social Support in
Pregnancy: Psychosocial Correlates of Birth Outcomes and Postpartum Depression. Journal
of Personality and Social Psychology, 65(6), 1243–1258. https://doi.org/10.1037/0022-
3514.65.6.1243
Condon, J. T., Boyce, P., & Corkindale, C. (2003). Predictors of Postnatal Adjustment in First
Time Fathers. Australian and New Zealand Journal of Psychiatry, 37, 33–43.
Cox, J., & Holden, J. (2003). Perinatal mental health: A guide to the Edinburgh Postnatal
Depression Scale (EPDS). Retrieved from https://psycnet.apa.org/record/2004-14522-000
Cox, M. J., Paley, B., Payne, C. C., & Burchinal, M. (1999). The transition to parenthood:
Marital conflict and withdrawal and parent-infant interactions. In Martha J. Cox & J.
Brooks-Gunn (Eds.), Conflict and cohesion in families: Causes and consequences (pp. 87–
104). Routledge. Retrieved from https://www.routledge.com/Conflict-and-Cohesion-in-
Families-Causes-and-Consequences/Cox-Brooks-Gunn/p/book/9781410602039

43

Crane, D., Allgood, S., Larson, J., & Griffin, W. (1990). Assessing marital quality with
distressed and nondistressed couples: A comparison and equivalency table for three
frequently used measures. Journal of Marriage and the Family, 87–93.
Da Costa, D., Zelkowitz, P., Dasgupta, K., Sewitch, M., Lowensteyn, I., Cruz, R., … Khalifé, S.
(2017). Dads Get Sad Too: Depressive Symptoms and Associated Factors in Expectant
First-Time Fathers. American Journal of Men’s Health, 11(5), 1376–1384.
https://doi.org/10.1177/1557988315606963
Datavyu Team. (2014). Datavyu: A Video Coding Tool. Databrary Project: New York
University. Retrieved from http://datavyu.org.
deMontigny, F., Girard, M.-E., Lacharité, C., Dubeau, D., & Devault, A. (2013a). Psychosocial
factors associated with paternal postnatal depression. Journal of Affective Disorders,
150(1), 44–49. https://doi.org///dx.doi.org.libproxy1.usc.edu/10.1016/j.jad.2013.01.048
deMontigny, F., Girard, M. E., Lacharité, C., Dubeau, D., & Devault, A. (2013b). Psychosocial
factors associated with paternal postnatal depression. Journal of Affective Disorders,
150(1), 44–49. https://doi.org/10.1016/j.jad.2013.01.048
Dennis, C.-L., & Ross, L. (2006). Women’s perceptions of partner support and conflict in the
development of postpartum depressive symptoms. Journal of Advanced Nursing, 56(6),
588–599. https://doi.org/10.1111/j.1365-2648.2006.04059.x
Don, B. P., & Mickelson, K. D. (2012). Paternal postpartum depression: The role of maternal
postpartum depression, spousal support, and relationship satisfaction. Couple and Family
Psychology: Research and Practice, 1(4), 323–334. https://doi.org/10.1037/a0029148

44

Doss, B. D., Rhoades, G. K., Stanley, S. M., & Markman, H. J. (2009). The effect of the
transition to parenthood on relationship quality: An 8-year prospective study. Journal of
Personality and Social Psychology, 96(3), 601–619.
Epifanio, M. S., Genna, V., De Luca, C., Roccella, M., & La Grutta, S. (2015). Paternal and
maternal transition to parenthood: The risk of postpartum depression and parenting stress.
Pediatric Reports, 7(2), 38–44. https://doi.org/10.4081/pr.2015.5872
Ertel, K. A., & Gillman, M. W. (2010). Antenatal and postpartum depressive symptoms are
differentially associated with early childhood weight and adiposity, 24(2), 179–189.
https://doi.org/10.1111/j.1365-3016.2010.01098.x.Antenatal
Field, T. (2010). Postpartum depression effects on early interactions, parenting, and safety
practices: A review. Infant Behavior and Development, 33(1), 1–6.
https://doi.org/10.1016/j.infbeh.2009.10.005
Figueiredo, B., Canário, C., Tendais, I., Pinto, T. M., Kenny, D. A., & Field, T. (2018). Couples’
relationship affects mothers’ and fathers’ anxiety and depression trajectories over the
transition to parenthood. Journal of Affective Disorders, 238.
https://doi.org/10.1016/j.jad.2018.05.064
Fitzpatrick, J., Gareau, A., Lafontaine, M.-F., & Gaudreau, P. (2016). How to Use the Actor-
Partner Interdependence Model (APIM) To Estimate Different Dyadic Patterns in MPLUS:
A Step-by-Step Tutorial. The Quantitative Methods for Psychology, 12(1), 74–86.
https://doi.org/10.20982/tqmp.12.1.p074
Gavin, A. R., Lindhorst, T., & Lohr, M. J. (2011). The prevalence and correlates of depressive
symptoms among adolescent mothers: Results from a 17-year longitudinal study. Women &
Health, 51(6), 525–545.
45

Gawlik, S., Müller, M., Hoffmann, L., Dienes, A., Wallwiener, M., Sohn, C., … Reck, C. (2014).
Prevalence of paternal perinatal depressiveness and its link to partnership satisfaction and
birth concerns. Archives of Women’s Mental Health, 17(1), 49–56.
https://doi.org/10.1007/s00737-013-0377-4
Goldstein, Z., Rosen, B., Howlett, A., Anderson, M., & Herman, D. (2020, March 15).
Interventions for paternal perinatal depression: A systematic review. Journal of Affective
Disorders. Elsevier B.V. https://doi.org/10.1016/j.jad.2019.12.029
Goodman, J. H. (2004). Paternal postpartum depression, its relationship to maternal postpartum
depression, and implications for family health. Journal of Advanced Nursing, 45(1), 26–35.
https://doi.org/10.1046/j.1365-2648.2003.02857.x
Goodman, J. H. (2008). Influences of maternal postpartum depression on fathers and on father-
infant interaction. Infant Mental Health Journal, 29(6), 624–643.
https://doi.org/10.1002/imhj.20199
Gordis, E. B., Margolin, G., Spies, L. A., Susman, E. J., & Granger, D. A. (2010). Interparental
aggression and parent – adolescent salivary alpha amylase symmetry. Physiology &
Behavior, 100(3), 225–233. https://doi.org/10.1016/j.physbeh.2010.01.006
Gottman, J. M. (1999). The marriage clinic: A scientifically-based marital therapy. New York,
NY: WW Norton & Company. https://doi.org/10.1891/0889-8391.15.3.279
Gottman, J. M., McCoy, K., Coan, J., & Collier, H. (1996). The Specific Affect Coding System
(SPAFF) for observing emotional communicationin martial and family interaction.
Mahwah, NJ: Erlbaum.
Gottman, John M, Coan, J., Carrère, S., & Swanson, C. (1998). Predicting marital happiness and
stability from newlywed interactions. Journal of Marriage and Family, 60(1), 5–22.
46

Gottman, John M, Driver, J., & Tabares, A. (2015). Repair during marital conflict in newlyweds:
How couples move from attack-defend to collaboration. Journal of Family Psychotherapy ,
26(2), 85–108. https://doi.org/10.1080/08975353.2015.1038962
Gress-Smith, J. L., Luecken, L. J., Lemery-Chalfant, K., & Howe, R. (2012). Postpartum
depression prevalence and impact on infant health, weight, and sleep in low-income and
ethnic minority women and infants. Maternal and Child Health Journal, 16(4), 887–893.
https://doi.org/10.1007/s10995-011-0812-y
Groer, M. W., Davis, M. W., & Hemphill, J. (2002). Postpartum stress: Current concepts and the
possible protective role of breastfeeding. Journal of Obstetric, Gynecologic & Neonatal
Nursing, 31(4), 411–417. https://doi.org/10.1111/j.1552-6909.2002.tb00063.x
Hallgren, K. A. (2012). Computing inter-rater reliability for observational data: An overview and
tutorial. Tutorials in Quantitative Methods for Psychology, 8(1), 23–34.
Heyman, R. E. (2001). Observation of couple conflicts: Clinical assessment applications,
stubborn truths, and shaky foundations. Psychological Assessment, 13(1), 5–35.
https://doi.org/10.1037/1040-3590.13.1.5
Houts, R. M., Barnett-Walker, K. C., Paley, B., & Cox, M. J. (2008). Patterns of couple
interaction during the transition to parenthood. Personal Relationships, 15(1), 103–122.
https://doi.org///dx.doi.org.libproxy1.usc.edu/10.1111/j.1475-6811.2007.00187.x
Iles, J., Slade, P., & Spiby, H. (2011). Posttraumatic stress symptoms and postpartum depression
in couples after childbirth: The role of partner support and attachment. Journal of Anxiety
Disorders, 25(4), 520–530. https://doi.org/10.1016/j.janxdis.2010.12.006
Karney, B. R., & Bradbury, T. N. (1995). The longitudinal course of marital quality and stability:
A review of theory, methods, and research. Psychological Bulletin, 118(1), 3–34.
47

Kenny, D. A., & Cook, W. (1999). Partner effects in relationship research: Conceptual issues,
analytic difficulties, and illustrations. Personal Relationships, 6(4), 433–448.
https://doi.org/10.1111/j.1475-6811.1999.tb00202.x
Kenny, D. A., Kashy, D. A., Cook, W. L., & Simpson, J. A. (2006). Dyadic data analysis
(Methodology in the social sciences). New York, NY: Guilford.
Kimmel, D., & Van Der Veen, F. (1974). Factors of marital adjustment in Locke’s Marital
Adjustment Test. Journal of Marriage and Family, 36(1), 57–63.
https://doi.org/10.1007/sl0551-008-9945-3
Koo, T. K., & Li, M. Y. (2016). A Guideline of Selecting and Reporting Intraclass Correlation
Coefficients for Reliability Research. Journal of Chiropractic Medicine, 15(2), 155–163.
https://doi.org/10.1016/j.jcm.2016.02.012
Leung, B. M. Y., Letourneau, N. L., Giesbrecht, G. F., Ntanda, H., & Hart, M. (2017). Predictors
of Postpartum Depression in Partnered Mothers and Fathers from a Longitudinal Cohort.
Community Mental Health Journal, 53(4), 420–431. https://doi.org/10.1007/s10597-016-
0060-0
Levenson, R. W., & Gottman, J. M. (1983). Marital interaction: Physiological linkage and
affective exchange. Journal of Personality and Social Psychology, 45(3), 587–597.
https://doi.org/10.1037/0022-3514.45.3.587
Locke, H. J., & Wallace, K. M. (1959). Short marital-adjustment and prediction tests: Their
reliability and validity. Marriage and Family Living, 21(3), 251–255.
https://doi.org/10.2307/348022
Logsdon, M. C., & Usui, W. (2001). Psychosocial predictors of postpartum depression in diverse
groups of women. Western Journal of Nursing Research, 23(6), 563–574.
48

Mastorakos, G., & Ilias, I. (2003). Maternal and fetal hypothalamic-pituitary-adrenal axes during
pregnancy and postpartum. Annals of the New York Academy of Sciences, 997(1), 136–149.
Matthey, S., Barnett, B., Kavanagh, D. J., & Howie, P. (2001). Validation of the Edinburgh
Postnatal Depression Scale for men, and comparison of item endorsement with their
partners. Journal of Affective Disorders, 64(2–3), 175–184.
Milgrom, J., Gemmill, A. W., Bilszta, J. L., Hayes, B., Barnett, B., Brooks, J., … Buist, A.
(2008). Antenatal risk factors for postnatal depression: A large prospective study. Journal of
Affective Disorders, 108(1–2), 147–157. https://doi.org/10.1016/j.jad.2007.10.014
Monroe, S. M., & Harkness, K. L. (2005). Life stress, the “kindling” hypothesis, and the
recurrence of depression: Considerations from a life stress perspective. Psychological
Review, 112, 417–445.
Morris, A. R., Khaled, M., Corner, G., & Saxbe, D. E. (n.d.). Prenatal romantic relationship
satisfaction predicts postpartum parent-infant bonding for fathers, but not mothers.
Submitted to Parenting: Science and Practic.
Morse, C. A., Buist, A., & Durkin, S. (2000). First-time parenthood: Influences on pre-and
postnatal adjustment in fathers and mothers. Journal of Psychosomatic Obstetrics &
Gynecology, 21(2), 109–120. https://doi.org/10.3109/01674820009075616
Murray, L. (1992). The Impact of Postnatal Depression on Infant Development. Journal of Child
Psychology and Psychiatry, 33(3), 543–561.
Murray, L., Fiori-cowley, A., & Hooper, R. (1996). The Impact of Postnatal Depression and
Associated Adversity on Early Mother-Infant Interactions and Later Infant Outcome. Child
Development, 67(5), 2512–2526.

49

Murray, L., Fiori-Cowley, A., Hooper, R., & Cooper, P. (1996). The Impact of Postnatal
Depression and Associated Adversity on Early Mother-Infant Interactions and Later Infant
Outcome. Child Development, 67(5), 2512–2526.
Muthén, L. K., & Muthén, B. (2018). Mplus. The Comprehensive Modelling Program for
Applied Researchers: User’s Guide, 5. Retrieved from
http://www.statmodel.com/virg_nov_course.shtml
Noldus, L. P. (1991). The Observer: A software system for collection and analysis of
observational data. Behavior Research Methods, Instruments, & Computers, 23(3), 415–
429. https://doi.org/10.3758/BF03203406
O’Hara, M., & McCabe, J. (2013). Postpartum Depression: Current Status and Future Directions.
Annual Review of Clinical Psychology, 9, 379–407. https://doi.org/10.1146/annurev-
clinpsy-050212-185612
O’hara, M. W., & Swain, A. M. (1996). Rates and risk of postpartum depression—a meta-
analysis. International Review of Psychiatry, 8(1), 37–54.
https://doi.org/10.3109/09540269609037816
Parade, S. H., Blankson, A. N., Leerkes, E. M., Crockenberg, S. C., & Faldowski, R. (2014).
Close relationships predict curvilinear trajectories of maternal depressive symptoms over
the transition to parenthood. Family Relations: An Interdisciplinary Journal of Applied
Family Studies, 63(2), 206–218.
https://doi.org///dx.doi.org.libproxy1.usc.edu/10.1111/fare.12065
Paulson, J. F., & Bazemore, S. D. (2010). Prenatal and postpartum depression in fathers and its
association with maternal depression: A meta-analysis. JAMA, 303(19), 1961–1969.
https://doi.org/10.1001/jama.2010.605
50

Perren, S., Von Wyl, A., Bürgin, D., Simoni, H., & Von Klitzing, K. (2005). Depressive
symptoms and psychosocial stress across the transition to parenthood: Associations with
parental psychopathology and child difficulty. Journal of Psychosomatic Obstetrics &
Gynecology, 26(3), 173–183.
Pinto, T. M., Samorinha, C., Tendais, I., & Figueiredo, B. (2019). Depression and paternal
adjustment and attitudes during the transition to parenthood. Journal of Reproductive and
Infant Psychology. https://doi.org/10.1080/02646838.2019.1652256
Reinhardt, J. P., Boerner, K., & Horowitz, A. (2006). Good to have but not to use: Differential
impact of perceived and received support on well-being. Journal of Social and Personal
Relationships, 23(1), 117–129. https://doi.org/10.1177/0265407506060182
Richards, J. M., Butler, E. A., & Gross, J. J. (2003). Emotion regulation in romantic
relationships: The cognitive consequences of concealing feelings. Journal of Social and
Personal Relationships, 20(5), 599–620. https://doi.org/10.1177/02654075030205002
Ripley, J. S., Garthe, R. C., Perkins, A., Worthington Jr., E. L., Davis, D. E., Hook, J. N., …
Eaves, D. (2016). Perceived partner humility predicts subjective stress during transition to
parenthood. Couple and Family Psychology: Research and Practice, 5(3), 157–167.
https://doi.org///dx.doi.org.libproxy1.usc.edu/10.1037/cfp0000063
Robertson, E., Grace, S., Wallington, T., & Stewart, D. E. (2004). Antenatal risk factors for
postpartum depression: a synthesis of recent literature . Gen.Hosp.Psychiatry , 26(0163-
8343 (Print)), 289–295. https://doi.org/10.1016/j.genhosppsych.2004.02.006
Shorey, S., Chee, C. Y. I., Ng, E. D., Chan, Y. H., Tam, W. W. S., & Chong, Y. S. (2018,
September 1). Prevalence and incidence of postpartum depression among healthy mothers:
A systematic review and meta-analysis. Journal of Psychiatric Research. Elsevier Ltd.
51

Stein, A., Gath, D. H., Bucher, J., Bond, A., Day, A. N. N., & Cooper, P. J. (1991). The
Relationship between Post-natal Depression and Mother-Child Interaction. British Journal
of Psychiatry, 158, 46–52.
Stowe, Z. N., & Nemeroff, C. B. (1995). Women at risk for postpartum-onset major depression.
American Journal of Obstetrics and Gynecology, 173(2), 639–645.
Tanner Stapleton, L. R., Schetter, C. D., Westling, E., Rini, C., Glynn, L. M., Hobel, C. J., &
Sandman, C. A. (2012). Perceived partner support in pregnancy predicts lower maternal and
infant distress. Journal of Family Psychology, 26(3), 453–463.
https://doi.org/10.1037/a0028332
Wee, K. Y., Skouteris, H., Pier, C., Richardson, B., & Milgrom, J. (2011, May 1). Correlates of
ante- and postnatal depression in fathers: A systematic review. Journal of Affective
Disorders. Elsevier B.V. https://doi.org/10.1016/j.jad.2010.06.019
Woodin, E. M. (2011). A Two-Dimensional Approach to Relationship Conflict: Meta-Analytic
Findings. Journal of Family Psychology, 25(3), 325–335. https://doi.org/10.1037/a0023791
Yim, I., Tanner Stapleton, L., Guardino, C., Hahn-Holbrook, J., & Dunkel Schetter, C. (2015).
Biological and psychosocial predictors of postpartum depression: Systematic review and
call for integration. Annual Review of Clinical Psychology, 11, 99–137.
https://doi.org/10.1146/annurev-clinpsy-101414-020426
Zajicek-Farber, M. L. (2009). Postnatal depression and infant health practices among high-risk
women. Journal of Child and Family Studies, 18(2), 236–245.
https://doi.org/10.1007/s10826-008-9224-z

52

Chapter 2: Physiological Linkage in Pregnancy: Couples’ Cortisol, Negative Conflict
Behavior, and Postpartum Depression
Cortisol Linkage
Couples’ physiological states have been found to be correlated in a number of studies,
and a growing literature has specifically examined within-couple correlations in momentary
levels of the stress hormone cortisol (Timmons, Margolin, & Saxbe, 2015). Although correlated
levels of cortisol, or cortisol linkage, may be normative within cohabiting couples, there is also
evidence that linkage is stronger among couples with poorer relationship quality, more marital
strain, more hostility, and lower empathy (Laws, Sayer, Pietromonaco, & Powers, 2015; Liu,
Rovine, Cousino Klein, & Almeida, 2013; Saxbe & Repetti, 2010; Schneiderman, Kanat-
Maymon, Zagoory-Sharon, & Feldman, 2014). However, physiological linkage has shown
nuanced effects with relationship functioning, depending on individual and couple characteristics
and the context in which linkage occurs. For instance, some studies have also reported stronger
couple cortisol linkage associated with partners’ physical proximity, length of cohabitation, and
time spent together, suggesting that linkage may be a product of mutual environment and shared
experiences (Laws et al., 2015; Papp, Pendry, Simon, & Adam, 2013; Saxbe & Repetti, 2010).
In general, greater couple physiological linkage can be thought of as an attunement or
sensitivity to partners’ physiological or mood states (Timmons et al., 2015). As it stands, there is
nothing inherently harmful or maladaptive about partners’ linked physiological patterns. Rather,
physiological linkage is a dyadic process that demonstrates the extent to which partners’ physical
and psychological states mutually influence each other. It is highly likely that the relational
implications of physiological linkage depend on the context in which linkage occurs (Timmons
53

et al., 2015). The current study specifically focuses on cortisol linkage within couples
transitioning to parenthood.
Cortisol Linkage and Couple Conflict
Couple conflict represents one context that may influence partners’ cortisol linkage.
Physiological linkage may be a marker of relationship distress during conflict through the
process of negative affect reciprocity, in which partners respond to one another with continued
anger and aggression (John M Gottman, Coan, Carrère, & Swanson, 1998; Levenson & Gottman,
1983). This conflict response cycle may occur less often in satisfied relationships, where one
partner might try to disrupt the negative feedback loop by attempting to de-escalate the conflict.
Additionally, a theory of stress contagion posits that partners with closely linked physiological
levels or patterns are more susceptible to their partners’ psychological and physical stress
responses, since coordination of physiological arousal may signify a high sensitivity to partners’
stress or negative mood states.
One study that provides evidence for stress contagion found that linkage of acute cortisol
levels during an in-lab conflict discussion was associated with poor relationship functioning
(Laws et al., 2015). Among couples with lower relationship satisfaction, stronger cortisol linkage
was associated with higher anger and hostility during the conflict discussion (Laws et al., 2015).
Another study found that higher couple cortisol linkage was associated with less empathy during
a couple conflict interaction. Decreases in empathy were shown to mediate the relationship
between couple cortisol linkage and later relationship dissolution (Schneiderman et al., 2014).
In the context of conflict, partners who exhibit heightened levels of negative conflict
behavior may show mutually increasing stress responses, becoming more physiologically
reactive to their partner, and leading to more patterns of emotional and physiological escalation.
54

This is the process through which partners experience stress contagion or negative affect
reciprocity.
Martial conflict research has demonstrated that conflict interactions are associated with
higher levels of physiological linkage, greater negative affect, and more reciprocal negative
affect exchanges among dissatisfied couples (Levenson & Gottman, 1983). Type of discussion
moderated the effect of physiological linkage, such that distressed couples showed greater
physiological connectedness when engaging in a conflict discussion rather than a neutral
discussion task (i.e., conversation about the day’s events). These study results highlight how the
connection between physiological linkage and relationship distress depends on context, since
among distressed couples, physiological levels were most strongly linked when discussing
relationship problems. Interestingly, higher couple physiological linkage during the conflict
discussions were associated with decreases in relationship satisfaction three years later
(Levenson & Gottman, 1985). This suggests that relationship conflict is one context in which
higher couple physiological linkage may be a marker of relationship distress. There are likely
some contexts in which physiological linkage signifies closeness or connectedness and others in
which linkage confers risk (i.e., conflict, negative reciprocity). Whereas previous evidence has
shown associations between physiological linkage and affective reactivity in couple conflict, it is
also important to investigate these interpersonal psychophysiological dynamic processes among
other forms of mood dysregulation.
Cortisol Linkage and Postpartum Depression (PPD)
Symptoms of depression, including sad mood, lack of interest or pleasure in activities,
and cognitive difficulties, are theorized as possible consequences of dysregulated affective and
biological stress responses (Ehlert, Gaab, & Heinrichs, 2001). Dysregulation of stress response
55

systems, specifically the hypothalamic-pituitary-adrenal (HPA) axis, and imbalanced cortisol
activity have been implicated in major depression (Holsboer, 2000), and have been proposed as
mechanisms linking stress and depression (Gold, Machado-Vieira, & Pavlatou, 2015).
HPA axis dysregulation has also been proposed as a possible pathway for the
development of PPD (Glynn, Davis, & Sandman, 2013) During pregnancy, expectant mothers
experience large shifts in both stress and reproductive hormones, with cortisol increasing
substantially over the course of pregnancy, and decreasing suddenly after birth. Since women
experience heightened levels of HPA axis activity and cortisol levels during pregnancy, and
these hormonal processes have been implicated in depression, investigating prenatal cortisol
activity may help identify physiological profiles that can increase risk for PPD.
The limited number of studies that have explored cortisol and PPD demonstrate
inconsistent findings. While most studies examining cortisol and PPD report null findings
(Jolley, Elmore, Barnard, & Carr, 2007; see Yim, Tanner Stapleton, Guardino, Hahn-Holbrook,
& Dunkel Schetter, 2015 for a review), others have found links between PPD and higher prenatal
cortisol reactivity (Nierop, Bratsikas, Zimmermann, & Ehlert, 2006), lower postpartum salivary
cortisol postpartum (Groer & Morgan, 2007), and an impaired postpartum cortisol awakening
response (Taylor, Glover, Marks, & Kammerer, 2009). Additionally, while assessments of
cortisol activity during pregnancy can serve as potential risk factors for PPD, the majority of
studies examined cortisol during the postpartum period. Given that new parents may experience
increased stress and physiological disruptions (Maureen Wimberly Groer, Davis, & Hemphill,
2002), HPA axis functioning is a plausible, but understudied, link to PPD.
Cortisol Linkage in Pregnancy
56

Very few studies have investigated couple cortisol linkage during pregnancy. Among the
studies that have, all five have shown that expectant couples demonstrate cortisol linkage during
pregnancy (Berg & Wynne-Edwards, 2002; Braren, Brandes-Aitken, Ribner, Perry, & Blair,
2020; Edelstein et al., 2015; Saxbe et al., 2015; Storey, Walsh, Quinton, & Wynne-Edwards,
2000). Only two of these studies assessed for moderators of cortisol linkage to evaluate whether
cortisol synchrony is adaptive in the context of pregnancy. Braren et al. (2020) examined cortisol
linkage in relation to maternal self-reported psychological stress, finding stronger cortisol
linkage when maternal stress was high. In a study that examined both pregnant and non-pregnant
couples, Saxbe et al. (2015) found that couples with stronger cortisol linkage subsequently
reported higher levels of intimate partner aggression, and that these associations were not
moderated by pregnancy status. These findings provide support for stress contagion in
pregnancy, in which couples who are more physiologically connected may influence or feed off
of one another’s stress states. However, little is known about how cortisol linkage during
pregnancy relates to dyadic interactions and postpartum adjustment. This is surprising, given the
growing body of literature on physiological and mood linkage among couples and its
significance to conflict and depression. Moreover, it is valuable to explore these
psychophysiological processes during the transition to parenthood due to the complex and
multilayered implications for couples and their families.
Current Study
The current study applied a biopsychosocial approach to identifying dyadic, behavioral,
and mood correlates of physiological linkage in expectant couples. To our knowledge, this is the
first study to examine whether prenatal couple cortisol linkage is associated with mother and
fathers’ negative relationship conflict behavior and their subsequent PPD. Pregnancy may be a
57

context in which partners are more physiologically attuned to one another’s stress and mood
states. Drawing on the stress contagion literature and negative affect reciprocity evidence, we
expected that couples with stronger cortisol linkage would exhibit higher levels of negative
conflict behavior and report more symptoms of PPD.
First, we examined couple cortisol linkage by investigating associations between
mothers’ and fathers’ cortisol levels over the course of the prenatal visit. We hypothesized that
mothers’ and fathers’ cortisol levels would be positively associated for the six samples across the
prenatal visit (H1).
Second, we examined associations between couple cortisol linkage and mothers’ and
fathers’ prenatal negative conflict behavior. We hypothesized that higher cortisol linkage will be
associated with greater levels of negative conflict behavior for mothers and fathers (H2).
Third, we tested associations between couple cortisol linkage and mothers’ and fathers’
PPD. We expected that higher cortisol linkage will be associated with greater symptoms of PPD
(H3).
Lastly, we investigated whether negative conflict behavior moderated the relationship
between couple cortisol linkage and PPD. We expected that among couples with higher cortisol
linkage, negative conflict behavior will be associated with more symptoms of PPD (H4).
Methods
Participants
The current study draws from the larger HATCH (Hormones Across the Transition to
Childrearing) study, which investigates changes in hormones and behaviors across the transition
to parenthood. The study follows couples from mid-to-late pregnancy across the first year
postpartum and includes in-person assessments that occur during pregnancy and during the
58

postpartum period. Participants are cohabitating, opposite-sex couples in Los Angeles expecting
their first child and they comprise a non-clinical, community sample. Participants were recruited
through posting flyers, social media (e.g., advertising to parenting-related online communities),
and word of mouth. The current study uses data from two in-lab visits, one that occurred during
mid-late pregnancy and one that occurred at approximately six months postpartum. During the
prenatal visit, informed consent was obtained from participants and all study procedures were
approved by the university’s Institutional Review Board.
Of the original sample of 95 couples, cortisol observations were missing for one father,
due to insufficient saliva for assays. Nine couples did not complete the postpartum visit for a
variety of reasons (e.g., relocated, no longer available, didn’t want to participate). Additionally,
one mother failed to complete any items of postpartum depression, so she and her partner were
excluded from analyses due to the dyadic nature of our study testing partner effects.
Observational coding for the conflict discussion was not possible for two couples (due to missing
audio and a corrupted video file transfer), so these two couples were dropped from analyses.
Therefore, the total sample size is 82 couples.
Procedure
Prenatal visit. The prenatal visit occurred in mid-to-late pregnancy (M = 28.4 weeks of
pregnancy, SD = 4.2 weeks, range = 20 to 39 weeks). During the prenatal visit, participants
engaged in three video-recorded discussions, including the conflict discussion, and completed
demographic and psychosocial questionnaires, and provided six saliva samples over the course
of the visit. Approximately one hour into the study visit, participants were given a list of
relationship-related topics (e.g., sex and intimacy, household chores, time spent together, time
spent on electronics, etc.), and instructed to rate the extent they would like to change each area in
59

regard to their relationship. Following the protocol that has been used in many other
observational couple conflict studies (Gordis, Margolin, Spies, Susman, & Granger, 2010), the
experimenter selected the three topics that appeared to be the greatest sources of conflict (e.g.,
topics that both partners rated highly, and/or topics where partners’ ratings were particularly
divergent). Couples then discussed these topics for fifteen minutes and were instructed to start
with the topic of greatest concern and progress to the other topics if time remains. Couples were
told:
“Now, we’re going to have you sit together for 15 minutes and talk about areas of
concern in your relationship…These conversations sometimes get intense, and that is OK. Try to
discuss this topic as you would at home, and make sure you each get your point across.”
Following the conflict discussion and the two other discussion tasks (a discussion about
hopes and expectations for the pregnancy, and a discussion about the anticipated division of baby
care responsibilities), participants completed questionnaire measures.
Cortisol Procedure. The prenatal visit is scheduled to begin at 2pm to standardize
cortisol sampling time of day and avoid morning cortisol peak and steeper slopes of cortisol in
the first half of the day. Upon arriving to the visit, participants are guided to the research lab
kitchenette to rinse out their mouths with water to remove any food or other possible
contaminants before saliva sampling begins. Participants are instructed to avoid drinking water
or other beverages for the remainder of the visit. Next, participants complete a saliva
questionnaire including questions about time of last consumption of food, beverage, and
caffeine, any cuts or sores in the mouth, medications that can affect cortisol levels (e.g., steroid-
based anti-inflammatory medications), smoking, etc. Experimenters check the saliva
questionnaire for any potential sources of contamination (e.g., bleeding gums) prior to conducing
60

the first saliva sample collection. The first saliva sample is collected at least 10 minutes after
mouth rinsing and 25 minutes after participants arrive to the visit. If participants have drank any
beverage besides water in the last hour, the first saliva sample is postponed until one hour has
elapsed since their last beverage consumption. Experimenters document any relevant notes
regarding sampling error, discoloration, or other sample collection compliance issues.
Six saliva samples of 1mL are collected via passive drool over the course of the visit: 1)
at baseline, 2) baseline + 20 minutes, 3) immediately post conflict discussion; baseline + 35
minutes, 4) baseline + 50 minutes, 5) baseline + 70 minutes, 6) baseline + 90 minutes. Figure 1
displays an overview of the visit and saliva collection times.
Chapter 2: Figure 1. Overview of prenatal visit procedures and saliva collection time-points.

Immediately following the visit, saliva samples were frozen in a -80-degree freezer visit
and then shipped on dry ice to Dr. Clemens Kirschbaum’s research laboratory (Dresden Lab
Service; Dresden, Germany), specializing in salivary cortisol assays. The Kirschbaum lab uses
an enzyme immunoassay to extract concentrations of salivary cortisol in nmol/l. Each saliva
sample was assayed twice, and analyses were repeated if any pair of results differed by >7%.”
Postpartum visit. The postpartum visit takes place approximately six to eight months
after the birth with the infant present (M = 7.2 months; SD = 0.8 months; range = 6.0 to 9.8
61

months postpartum). The visit parallels the prenatal visit with similar tasks and procedures to
facilitate comparison of prenatal and postpartum data. Couples again complete questionnaires,
including measures of postpartum depressive symptoms.
Coding of Couple Conflict Behavior
As stated above, 93 discussions were available for coding. Prenatal couple conflict
behavior for all 93 discussions was assessed using the well-established and widely-used
observational coding scheme, The Specific Affect Coding System (SPAFF; Gottman, McCoy,
Coan, & Collier, 1996). The SPAFF identifies specific affective behavior based on verbal
content, vocal tone, facial expression, gestures, and body movement. The SPAFF includes five
positive codes (affection, enthusiasm, humor, interest, validation), 12 negative codes (anger,
belligerence, contempt, criticism, defensiveness, disgust, domineering, fear/tension, sadness,
stonewalling, threats, whining), and one neutral code. The SPAFF has shown excellent construct
and criterion validity and has been employed in many studies of couple relationship behavior
(Heyman, 2001).
Behavioral coding of the conflict discussion was completed in two parts: the first 49
conflict discussion videos were coded half-way through data collection using the Noldus XT
Observer system (Noldus, 1991) and the next 44 discussions were coded near data collection
completion using Datavyu (Datavyu Team, 2014). The first set of 49 discussions were coded by
a group of four research assistants. Two raters coded the entire first sample of 49 videos. Due to
attrition of raters, a third rater completed only the first 19 videos, and a fourth completed the first
12. The second set of 44 discussions were coded by a group of three research assistants; all three
raters coded the entirety of the 44 videos.
62

Although behavioral coding of the 93 conflict discussions were completed at two
different time periods, all research assistants were trained using the SPAFF manual and were
trained by the same person, the first author. Additionally, although the two sets of research
assistants utilized different software to complete the behavioral coding, observational codes were
entered identically by all raters. Raters entered observational codes continuously (i.e., each time
an emotional behavior was expressed, a code was assigned until a different behavior was
displayed). Thus, for each behavioral code, this yielded a total time in seconds. For each
participant, we summed the duration of all positive behavioral codes and, separately, all negative
behavior codes to reflect the amount of time spent expressing each valanced behavior. Due to
coder disagreement, coders had varying durations for each valanced behavior. Thus, we averaged
the duration of negative behavioral codes across all coders and used this value for analyses. We
then calculated the proportion of time that each participant spent exhibiting negative behavior by
dividing the duration of negative behavior for a given individual by the total discussion time. The
average proportion of negative conflict behavior was 0.14 (SD = 0.12, Range = 0.00 – 0.52) for
mothers and 0.13 (SD = 0.12, Range = 0.00 – 0.58) for fathers.
To assess inter-rater reliability among the raters, four intra-class correlations (ICCs;
Bartko, 1966) were calculated for total duration of positive behavior and total duration of
negative behavior. In other words, raters’ agreement of the amount of positive and negative
behavior expressed by participants was compared. ICCs were calculated separately for mothers
and fathers, yielding four ICC values: ICCs for mother positive, mother negative, father positive
and father negative. Then, the four ICCs were averaged to provide one ICC value for each group
of coders. ICC is widely used to assess inter-rater reliability in studies with two or more coders
63

(Hallgren, 2012). Across discussions, ICCs ranged from .70 - .83, depending on groups of coders
and number of discussions, reflecting moderate to good reliability (Koo & Li, 2016).
Measures
Postpartum Depression. Postpartum depressive symptoms were measured at the
postpartum visit using the Edinburgh Postnatal Depression Scale (EPDS; Cox & Holden, 2003).
The EPDS is a 10-item scale assessing postpartum depressive symptoms experienced in the past
7 days. Items assess for symptoms including feeling sad or miserable, self-blame, and looking
forward with enjoyment to activities. Items are rated on a 4-point scale ranging from 0 to 3, with
higher ratings corresponding to greater symptoms. The EPDS is widely-used, has been shown to
have good criterion validity, internal consistency and reliability (Adouard, Glangeaud-
Freudenthal, & Golse, 2005) and has been validated for use with fathers (Matthey, Barnett,
Kavanagh, & Howie, 2001). For the current study, we divided participants’ total EPDS score by
number of items, to yield an average item score. Therefore, maximum possible EPDS score for
this sample was 3.
Data Analysis
Analyses were conducted using SPSS Version 26 (IBM Corp, 2019) and Hierarchical
Linear Modeling Version 8.0 (HLM; Raudenbush, Bryk, & Congdon, 2019). Descriptive
statistics and bivariate correlations were calculated for key study variables. T-tests were
conducted to compare mothers and fathers on demographic characteristics and key variables of
interest (e.g., cortisol, conflict behavior, postpartum depression).
Cortisol data analyses. For descriptives and zero-order correlations, we calculated
aggregate cortisol output (area under the curve with respect to ground (AUCg)) using the
standard trapezoidal formula provided by Pruessner and colleagues (Pruessner, Kirschbaum,
64

Meinlschmid, & Hellhammer, 2003). All other cortisol analyses were conducted in HLM. When
examining raw cortisol data, we truncated extreme values or outliers by removing observations
that were >3 standard deviations above the sample mean. Truncating outliers resulted in <4% of
samples being dropped from analyses. Additionally, after truncating extreme values, we log-
transformed cortisol values in order to account for positive skew. Using information from
participant saliva questionnaires completed at the beginning of the prenatal visit as well as
experimenter notes about saliva sampling contamination, we created a dummy variable to
indicate possible sources of saliva sampling errors.
We added the following covariates to all models: within-person Level 1 covariates
including time of day for each saliva collection and saliva sampling errors (i.e., dummy-coded
variable for sources of error such as bleeding gums, sample discoloration, or other problems with
sampling procedures) and the between-person Level 2 covariate of mothers’ number of weeks
pregnant at the prenatal visit. Time of each saliva collection is included as a covariate to account
for the diurnal variability of cortisol levels over the course of the day. Number of weeks pregnant
is added as a covariate to control for mothers’ increasing levels of cortisol in later stages of
pregnancy (Yim et al., 2015).
Negative conflict behavior, EPDS scores, cortisol sampling times and weeks pregnant
were grand-centered for analyses to represent deviations from the sample or couples’ means.
Given the dyadic nature of our data, we decided to grand-center our variables rather than group-
center to be able to compare individuals to the entire sample as a whole.
Analyses for study hypotheses were tested using HLM to estimate couples’ cortisol
linkage and associations with negative conflict behavior and PPD. HLM is ideal for modeling
repeated measures of cortisol across time and examining variations at the within-person (Level
65

1) and between-person (Level 2) levels (Hruschka, Kohrt, & Worthman, 2005). HLM can also
successfully model data in instances when missing values result in varying numbers of
observations per individual and varying times between observations (Huta, 2016).
Hypothesis 1: Couple cortisol linkage. Analyses for hypothesis 1 were conducted to
determine whether mothers’ and fathers’ cortisol levels were correlated across the six samples of
the prenatal visit. In order to model couple cortisol linkage in HLM, separate models were used
to estimate (1) the influence of fathers’ cortisol levels on mothers’ cortisol and (2) the influence
of mothers’ cortisol levels on fathers’ cortisol. Thus, for the first series of models in HLM,
mothers’ log-transformed cortisol was added as the outcome variable to an unconditional model
(i.e., no predictors or other variables in the model). Next, fathers’ log-transformed cortisol was
added as a Level 1 predictor. Then, we added covariates including time of day and saliva
sampling errors at Level 1 and number of weeks pregnant at Level 2. The same procedure was
used to construct models with fathers’ log-transformed cortisol as the outcome variable, mothers’
log-transformed cortisol as a Level 1 predictor, and time of day, sampling errors, and weeks
pregnant added as covariates. Therefore, separate cortisol linkage intercepts and slopes are
estimated depending on whether mothers’ or fathers’ cortisol is the outcome or predictor.
Hypothesis 2: Couple cortisol linkage and negative conflict behavior. To examine
whether couple cortisol linkage is associated with negative conflict behavior, negative conflict
behavior was added as a Level 2 predictor to the previous two models. Thus, for the first model,
mothers’ log-transformed cortisol was the outcome, fathers’ log-transformed cortisol was added
as a level 1 predictor, time of day, sampling errors, and weeks pregnant were added as
covariates, and then mothers’ negative conflict behavior was added as a Level 2 predictor. This
allowed us to estimate the effect of mothers’ negative conflict behavior on mothers’ cortisol
66

intercept and cortisol linkage. For the second model, fathers’ log-transformed cortisol was the
outcome, mothers’ log-transformed cortisol was added as a Level 1 predictor, time of day,
sampling errors, and weeks pregnant were added as covariates, and then fathers’ negative
conflict behavior was added as a Level 2 predictor. This allowed us to estimate the effect of
fathers’ negative conflict behavior on fathers’ cortisol intercept and cortisol linkage.
Hypothesis 3: Couple cortisol linkage and PPD. Analyses for hypothesis 3, examining
the association between cortisol linkage and PPD, were conducted using the same steps as
hypothesis 2. The first model included mothers’ log-transformed cortisol as the outcome, fathers’
log-transformed cortisol as the Level 1 predictor, time of day, sampling errors, and weeks
pregnant as covariates, and then mothers’ PPD was added as a Level 2 predictor. This allowed us
to estimate the effect of mothers’ PPD on mothers’ cortisol intercept and cortisol linkage. The
second model was comprised of fathers’ log-transformed cortisol as the outcome, mothers’ log-
transformed cortisol as the Level 1 predictor, time of day, sampling errors, and weeks pregnant
as covariates, and then fathers’ PPD was added as a Level 2 predictor. This allowed us to
estimate the effect of fathers’ PPD on fathers’ cortisol intercept and cortisol linkage.
Hypothesis 4: Negative conflict behavior moderating cortisol linkage and PPD. To
test whether negative conflict behavior moderates the relationship between cortisol linkage and
PPD, we extracted empirical Bayes coefficients for cortisol linkage from HLM and then tested
interaction effects in SPSS. In order to extract empirical Bayes coefficients for cortisol linkage,
we created two models in HLM for (1) the influence of fathers’ cortisol levels on mothers’
cortisol and (2) the influence of mothers’ cortisol levels on fathers’ cortisol. Time of day was
added as a covariate to both models. In other words, for one model, we added mothers’ log-
transformed cortisol as the outcome variable to an unconditional model, then added fathers’ log-
67

transformed cortisol as the predictor and time of day as a covariate. For the second model,
fathers’ log-transformed cortisol was outcome variable in an unconditional model, then mothers’
log-transformed cortisol was added as a predictor and time of day as a covariate. Next, we saved
the Level 2 residual files which included the between-level empirical Bayes coefficients for
cortisol linkage.
In SPSS, we merged data files including negative conflict behavior, PPD, and the cortisol
linkage coefficients. Cortisol linkage coefficients were group-centered to create interaction
terms. Four interaction effects were tested using linear regression to determine if, among couples
with stronger cortisol linkage, own or partner’s negative behavior predicted PPD symptoms: (1)
mothers’ negative conflict behavior moderating the effect of cortisol linkage on maternal PPD,
(2) fathers’ negative conflict behavior moderating the effect of cortisol linkage on maternal PPD,
(3) mothers’ negative conflict behavior moderating the effect of cortisol linkage on paternal
PPD, (4) fathers’ negative conflict behavior moderating the effect of cortisol linkage on paternal
PPD.
Results
Sample demographics including race/ethnicity and education are displayed in Table 1.
Chapter 2: Table 1.
Participant and couple characteristics.

Female
Participants
(n = 82)

Male
Participants
(n = 82)

Participant Characteristics n (%)
Race
Asian / Pacific Islander 15 (18.3%) 15 (18.3%)
Black / African American 5 (6.1%) 6 (7.3%)
Hispanic / Latinx 19 (23.2%) 16 (19.5%)
White / Caucasian 37 (45.1%) 41 (50.0%)
Other 6 (7.3%) 4 (4.9%)
Educational Attainment
68

High School / GED 1 (1.2%) 2 (2.4%)
Some College 12 (14.7%) 13 (15.9%)
College Degree 31 (37.8%) 40 (48.8%)
Master’s Degree 30 (36.6%) 16 (19.5%)
Professional or Doctoral Degree 8 (9.8%) 11 (13.4%)
Couple Characteristics
Married 68 (82.9%)
Dating/Cohabiting 14 (17.1%)
Descriptive statistics for key study variables and t-tests comparing mothers and fathers
are reported in Table 2. T-tests indicated that among our sample, fathers were older than
mothers. Additionally, mothers had significantly higher levels of cortisol than fathers, which is
not surprising, given that cortisol levels are known to increase during pregnancy. On average,
mothers reported higher postpartum depressive symptoms than fathers. There was no significant
difference in expression of negative conflict behavior between mothers and fathers.
Chapter 2: Table 2.
Descriptives for Key Study Variables.
Female Participants Male Participants
M (SD)
Range
M (SD)
Range
t-test
a

Study Variables
Age
31.3 (4.31)
21 - 39
33.3 (5.64)
22 - 57
-4.23***
Cortisol (log-transformed)
2.25 (.49)
.70 – 3.17
1.60 (.62)
.06 – 2.97
21.50***
Postpartum Depression Score (EPDS)
.69 (.50)
0 – 2.0
.47 (.43)
0 – 1.78
3.29**
Negative Conflict Behavior
.14 (.12)
0 - .52
.13 (.12)
0 - .58
1.34
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10
a
Critical values are from paired samples t-tests evaluating gender differences on key study variables.
Abbreviations: BDI = Beck Depression Inventory; EPDS = Edinburgh Postnatal Depression Scale;

Zero-order correlations for all variables are shown in Table 3. There were no within-
couple associations between mothers’ and fathers’ aggregate cortisol (AUCg) or mothers’ and
fathers’ PPD. Fathers’ prenatal cortisol was positively associated with maternal PPD.
69

Additionally, there was a positive within-couple association between mothers’ and fathers’
negative conflict behavior.
Chapter 2: Table 3.
Zero-order correlations for study variables.

1 2 3 4 5 6
1. Maternal Cortisol -
2. Paternal Cortisol .14 -
3. Maternal Postpartum Depression
(EPDS)
-.14 .32** -
4. Paternal Postpartum Depression (EPDS) -.13 .14 .16 -
5. Maternal Negative Conflict Behavior -.06 .06 .02 .13 -
6. Paternal Negative Conflict Behavior .11 -.05 -.04 .13 .73*** -
7. Weeks Pregnant at Prenatal Visit .21
⊥
.07 -.09 -.18 -.13 -.10
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10
Abbreviations: EPDS = Edinburgh Postnatal Depression Scale;
Hypothesis 1: Couple cortisol linkage. Table 4 depicts results for hypothesis 1, with
separate cortisol linkage models for mothers’ cortisol as the outcome and fathers’ cortisol as the
outcome. Consistent with hypothesis 1, couples’ cortisol levels are positively associated,
meaning when one partner has higher cortisol levels than usual, their partner is likely to have
higher-than-usual cortisol levels as well, suggesting that couple cortisol linkage appeared within
this sample. For both mothers and fathers, partner cortisol was positively associated with own
cortisol levels after controlling for time of sampling, sampling errors, and number of weeks
pregnant (mothers,  = .17, p = .00; fathers,  = .22, p = .00).
Additionally, time of saliva samples was inversely associated with both partners’ cortisol
intercept (mothers,  = -.09, p = .00; fathers,  = -.11, p = .00), meaning that participants’
cortisol levels tended to decrease over the course of the visit. This is to be expected given the
natural decline in cortisol over the course of the day. For mothers, weeks pregnant was positively
associated with cortisol (  = .01, p = .00), indicating that, as expected, mothers’ cortisol levels
were increasing as they get farther along in their pregnancy.
70

Chapter 2: Table 4

Cortisol Linkage (Partner Cortisol Predicting Own Cortisol Level)
Fixed effect Coefficient (SE) t-ratio
Model Predicting Mothers’ Cortisol
Mothers’ intercept ( 0 iMOTH) 3.41 (.36) 9.60***
Level 2 weeks pregnant ( 01) .01 (.00) 4.43***
Slope of mothers’ time ( 1M) -.09 (.02) -3.98***
Slope of mothers’ saliva sampling error ( 2M) -.05 (.11) -.41
Partner (father) cortisol predicting mothers’ cortisol ( 3FM) .17 (.04) 4.00***
Model Predicting Fathers’ Cortisol
Fathers’ intercept( 0 iFATH) 2.85 (.56) 5.08***
Level 2 weeks pregnant ( 01) .00 (.00) .18
Slope of fathers’ time ( 1F) -.11 (.03) -3.33**
Slope of fathers’ saliva sampling error ( 2F) -.01 (.10) -.05
Partner (mother) cortisol predicting fathers’ cortisol( 3MF) .22 (.06) 3.50***
Note. Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10

Hypothesis 2: Couple cortisol linkage and negative conflict behavior. Table 5
displays results for associations between couple cortisol linkage and negative conflict behavior.
For mothers, negative conflict behavior was positively associated with their intercept of cortisol
(  = .13, p = .04), meaning that for mothers who exhibited more negative behavior during the
conflict discussion, they tended to have higher cortisol levels on average across the prenatal visit.
The association between fathers’ negative conflict behavior and cortisol intercept was also
positive and trending towards significance (  = .19, p = .06). Inconsistent with hypothesis 2,
couple cortisol linkage was inversely associated with negative conflict behavior, such that
stronger cortisol linkage was associated with less negative conflict behavior for both mothers ( 
= -.10, p = .00) and fathers (  = -.10, p = .02), after controlling for time of sampling, sampling
errors, and number of weeks pregnant. Figures 2 and 3 present line graphs depicting cortisol
linkage and negative conflict behavior for mothers and fathers.
Chapter 2: Table 5

Associations Between Cortisol Linkage and Negative Conflict Behavior
Fixed effect Coefficient (SE) t-ratio
71

Model Predicting Mothers’ Cortisol
Mothers’ intercept ( 0 iMOTH) 3.30 (.36) 9.26***
Level 2 weeks pregnant ( 01) .01 (.00) 4.74***
Level 2 maternal negative conflict behavior ( 02) .13 (.06) 2.04*
Slope of mothers’ saliva sampling error ( 1M) -.02 (.11) -.17
Slope of mothers’ time ( 2M) -.08 (.02) -3.69***
Partner (father) cortisol predicting mothers’ cortisol ( 3FM) .18 (.04) 4.55***
Level 2 maternal negative conflict behavior ( 31) -.10 (.03) -2.93**
Model Predicting Fathers’ Cortisol
Fathers’ intercept( 0 iFATH) 2.74 (.57) 4.81***
Level 2 weeks pregnant ( 01) .00 (.00) -.04
Level 2 paternal negative conflict behavior ( 02) .19 (.10)
1.90
⊥

Slope of fathers’ saliva sampling error ( 1F) .09 (.07) 1.16
Slope of fathers’ time ( 2F) -.10 (.03) -3.01**
Partner (mother) cortisol predicting fathers’ cortisol ( 3MF) .21 (.06) 3.57***
Level 2 paternal negative conflict behavior ( 31) -.10 (.04) -2.72*
Note. Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10

Chapter 2: Figure 2. Association between cortisol linkage and mothers’ negative conflict behavior.

This figure, generated in HLM 8.0, depicts the relationship between cortisol linkage (fathers’ cortisol
predicting mothers’ cortisol) and mothers’ negative conflict behavior. The mean and +/- 1 standard deviation of
mothers’ negative conflict behavior is displayed. For mothers, stronger cortisol linkage is associated with less
negative conflict behavior (i.e., -1 standard deviation), whereas weaker cortisol linkage is associated with more
negative conflict behavior (i.e., +1 standard deviation).

Chapter 2: Figure 3. Association between cortisol linkage and fathers’ negative conflict behavior.

72

This figure depicts the relationship between cortisol linkage (mothers’ cortisol predicting fathers’ cortisol) and
fathers’ negative conflict behavior. The mean and +/- 1 standard deviation of fathers’ negative conflict behavior is
displayed. For fathers, stronger cortisol linkage is associated with less negative conflict behavior (i.e., -1 standard
deviation), whereas weaker cortisol linkage is associated with more negative conflict behavior (i.e., +1 standard
deviation).

Hypothesis 3: Couple cortisol linkage and PPD. Table 6 displays results for
associations between cortisol linkage and PPD. For mothers, prenatal cortisol linkage was not
associated with PPD (  = .06, p = .46). Inconsistent with hypothesis 3, prenatal couple cortisol
linkage was inversely associated with fathers’ PPD symptoms, such that stronger cortisol linkage
was associated with fewer PPD symptoms for fathers (  = -.24, p = .03). Figure 4 depicts the
relationship between cortisol linkage and fathers’ PPD.
Chapter 2: Table 6

Associations Between Cortisol Linkage and PPD
Fixed effect Coefficient (SE) t-ratio
Model Predicting Mothers’ Cortisol
Mothers’ intercept ( 0 iMOTH) 3.30 (.41) 8.01***
Level 2 weeks pregnant ( 01) .00 (.00) 3.35**
Level 2 maternal postpartum depression ( 02) -.24 (.62) -.39
Slope of mothers’ saliva sampling error ( 1M) -.05 (.11) -.48
Slope of mothers’ time ( 2M) -.08 (.03) -3.08**
Level 2 maternal postpartum depression ( 21) .00 (.04) .12
Partner (father) cortisol predicting mothers’ cortisol ( 3FM) .17 (.04) 3.89***
Level 2 maternal postpartum depression ( 31) .06 (.08) .74
Model Predicting Fathers’ Cortisol
Fathers’ intercept( 0 iFATH) 2.82 (.62) 4.51***
Level 2 weeks pregnant ( 01) .00 (.00) .33
Level 2 paternal postpartum depression ( 02) -1.77 (1.37) -1.29
Slope of fathers’ saliva sampling error ( 1F) -.01 (.09) -.08
Slope of fathers’ time ( 2F) -.11 (.04) -3.07**
Level 2 paternal postpartum depression ( 21) .16 (.09)
1.73
⊥

Partner (mother) cortisol predicting fathers’ cortisol ( 3MF) .23 (.07) 3.52***
Level 2 paternal postpartum depression ( 31) -.24 (.11) -2.16*
Note. Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10
Next, we combined models for hypothesis 2 and 3, adding both negative conflict
behavior and PPD to the same model testing for cortisol linkage. Table 7 shows results for
negative conflict behavior and PPD added to the same model. All results remained the same,
73

with stronger cortisol linkage associated with less negative conflict behavior for mothers (  = -
.09, p = .01) and fathers (  = -.08, p = .01) and fewer paternal PPD symptoms (  = -.23, p = .02).
Chapter 2: Figure 4. Association between cortisol linkage and fathers’ PPD.

This figure depicts the relationship between cortisol linkage (mothers’ cortisol predicting fathers’ cortisol) and
fathers’ PPD symptoms. The mean and +/- 1 standard deviation of fathers’ PPD is displayed. For fathers, stronger
cortisol linkage is associated with fewer PPD symptoms behavior (i.e., -1 standard deviation), whereas weaker
cortisol linkage is associated with greater PPD symptoms (i.e., +1 standard deviation).

Hypothesis 4: Negative conflict behavior moderating cortisol linkage and PPD.
Table 8 displays results for the interaction effects of negative conflict behavior and cortisol
linkage on maternal PPD and table 9 shows interaction results for paternal PPD. All interaction
effects were non-significant.
Chapter 2: Table 7

Associations Between Cortisol Linkage, Negative Conflict Behavior and PPD
Fixed effect Coefficient (SE) t-ratio
Model Predicting Mothers’ Cortisol
Mothers’ intercept ( 0 iMOTH) 3.23 (.41) 7.91***
Level 2 weeks pregnant ( 01) .00 (.00) 3.85***
Level 2 maternal postpartum depression ( 02) -.15 (.14) -1.12
Level 2 maternal negative conflict behavior ( 03) .12 (.06)
1.86
⊥

Slope of mothers’ saliva sampling error ( 1M) -.03 (.10) -.26
Slope of mothers’ time ( 2M) -.08 (.03) -3.02**
Partner (father) cortisol predicting mothers’ cortisol ( 3FM) .17 (.04) 3.97***
Level 2 maternal postpartum depression ( 31) .07 (.07) .95
Level 2 maternal negative conflict behavior ( 32) -.09 (.03) -2.761**
Model Predicting Fathers’ Cortisol
Fathers’ intercept( 0 iFATH) 2.57 (.64) 4.00***
Level 2 weeks pregnant ( 01) .00 (.00) .18
74

Level 2 paternal postpartum depression ( 02) .68 (.23) 2.92**
Level 2 paternal negative conflict behavior ( 03) .14 (.08)
1.81
⊥

Slope of fathers’ saliva sampling error ( 1F) .10 (.08) 1.29
Slope of fathers’ time ( 2F) -.09 (.04) 2.58*
Partner (mother) cortisol predicting fathers’ cortisol ( 3MF) .24 (.06) 3.77***
Level 2 paternal postpartum depression ( 31) -.23 (.09) -2.63**
Level 2 paternal negative conflict behavior ( 32) -.08 (.03) -2.42*
Note. Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10

Chapter 2: Table 8.
Interaction effects of own and partner negative conflict behavior as a moderator between
cortisol linkage and maternal PPD.

Interaction of own negative
behavior and cortisol with
maternal PPD
Interaction of partners’
negative behavior and
cortisol with maternal PPD

F (78,3) = .36, p = .78 F (78,3) = .56, p = .64
Beta t Beta t
(Constant) - 5.21*** - 12.04***
Cortisol linkage .11 .99 .09 .83
Negative conflict behavior .07 .39 -.03 -.28
Interaction of cortisol linkage and
negative behavior
-.02 -.09 -.11 -.93
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10

Chapter 2: Table 9.
Interaction effects of own and partner negative conflict behavior as a moderator between
cortisol linkage and paternal PPD.

Interaction of own negative
behavior and cortisol with
paternal PPD
Interaction of partners’
negative behavior and
cortisol with paternal PPD

F (79,3) = 1.87, p = .14 F (79,3) = 1.61, p = .19
Beta t Beta t
(Constant) - 5.57*** - 5.34***
Cortisol linkage -.18 -1.61 -.18 -1.56
Negative conflict behavior .28 1.25 .21 .92
Interaction of cortisol linkage and
negative behavior
-.20 -.89 -.14 -.62
Significance: *** p < .001, ** p < .01, * p < .05,
⊥
p < .10

75

Discussion
This study evaluated couples’ cortisol levels during pregnancy, and examined effects
between cortisol linkage, prenatal negative conflict behavior, and PPD symptoms among first-
time mothers and fathers. As hypothesized, couples’ cortisol levels were positively associated,
indicating prenatal couple cortisol linkage within this sample. These results held after controlling
for time of sampling, sampling errors, and number of weeks pregnant. Mothers who exhibited
more negative behavior during the conflict discussion tended to have higher overall cortisol
levels across the visit, with a similar but non-significant (trending) association for fathers’
negative conflict behavior and fathers’ cortisol. Although we expected that stronger cortisol
linkage would be emerge within more distressed couples, we actually found the opposite:
stronger cortisol linkage was associated with lower levels of negative conflict behavior for both
mothers and fathers, and predicted fewer PPD symptoms for fathers (but not mothers). When
negative conflict behavior and PPD were added to the same model, results remained significant,
with stronger cortisol linkage associated with less negative conflict behavior for mothers and
fathers and fewer paternal PPD symptoms. Lastly, we tested moderation effects, but did not find
support for our hypothesis that negative conflict behavior would moderate the association
between cortisol linkage and subsequent PPD.
Our results provide evidence for couple cortisol linkage during pregnancy, consistent
with other studies finding that expectant couples show correlated levels of cortisol (Berg &
Wynne-Edwards, 2002; Braren et al., 2020; Edelstein et al., 2015; Storey et al., 2000). Contrary
to our hypotheses, we found that stronger couple cortisol linkage was associated with more
positive functioning for couples: specifically, lower levels of negative conflict behavior for both
parents and fewer paternal PPD symptoms. Said another way, weaker cortisol linkage among
76

couples was associated with higher negative conflict behavior and greater paternal PPD
symptoms. While we hypothesized that greater cortisol linkage would increase the likelihood
that partners would engage in more negative conflict and develop PPD symptoms, these results
suggest that couple cortisol linkage in and of itself is not a marker of heightened conflict,
relationship distress or stress contagion.
Research suggests that cortisol linkage is not inherently adaptive or maladaptive, but
rather is a dyadic process that demonstrates the extent to which partners mutually influence each
other. It is highly likely that the relational implications of physiological linkage depend on the
context in which linkage occurs (Timmons et al., 2015). Cortisol linkage may be a marker of
unhealthy dynamics in some contexts, and may signify more adaptive processes in other
contexts. For instance, while some studies have found cortisol linkage to be associated with
relationship strain and dissatisfaction (Laws et al., 2015; Liu et al., 2013; Saxbe & Repetti,
2010), studies have also reported stronger couple cortisol linkage associated with partners’
physical proximity, length of cohabitation, and time spent together, suggesting that linkage may
be a product of mutual environment and shared experiences (Laws et al., 2015; Papp et al., 2013;
Saxbe & Repetti, 2010).
We did not find evidence that prenatal cortisol linkage was related to maternal PPD.
These findings suggest that couple cortisol linkage during pregnancy may have differential
effects on mothers and fathers’ postpartum well-being. This is consistent with other work
suggesting that the couple relationship exerts a particularly strong influence on fathers’
adjustment to parenthood more so than for mothers (Belsky, Youngblade, Rovine, & Volling,
1991; Cox, Owen, Lewis, & Henderson, 1989).
77

Finally, we tested moderation effects of negative conflict behavior and cortisol linkage on
PPD symptoms to examine whether among couples with strong cortisol linkage, individuals
exhibiting more negative behavior during the conflict discussion would increase the likelihood of
their partner developing PPD. We hypothesized that couples with highly linked patterns of
cortisol may be more susceptible to transmission of their partners’ stress reactivity or negative
mood states. However, we found no significant interaction effects or evidence of stress contagion
among this sample.
In the context of pregnancy, in which partners are preparing for a significant joint
transition and mothers are undergoing substantial physiological changes, physiological linkage
may represent connection, closeness, and partner attunement. Indeed, this is a common
interpretation of Couvade syndrome, in which male partners of pregnant women experience a
variety of physical symptoms similar to that of their pregnant partner, yet with no physiological
basis or etiology, with the onset of symptoms occurring during their partner’s first trimester of
pregnancy and subsiding after the birth (Brennan, Ayers, Ahmed, & Marshall-Lucette, 2007).
Empirical studies have found that male partner’s affective involvement in his partner’s
pregnancy and the extent of his preparation for parenthood is associated with the frequency,
duration, and severity of his pregnancy symptoms (Clinton, 1987; Longobucco & Freston, 1989).
This provides support for couples’ physiological linkage during pregnancy representing partner
attunement. This dovetails with previous findings showing that higher levels of cortisol linkage
were related to shared environment and joint experiences (Laws et al., 2015; Papp et al., 2013;
Saxbe & Repetti, 2010).
Therefore, in some contexts, physiological linkage may signify positive relationship
functioning, bonding, and connectedness. Particularly during pregnancy, weaker physiological
78

linkage may indicate disengagement or emotional distance. Our results revealed that weaker
cortisol linkage among expectant couples was associated with higher negative conflict behavior
and greater paternal PPD symptoms. In our sample, pregnant couples’ weaker cortisol linkage
over the visit may convey emotional detachment and disharmony, resulting in partners’ elevated
levels of negative conflict behavior and subsequent paternal depressive symptoms during the
postpartum period. Therefore, our results suggest that pregnancy might be one context in which
weaker cortisol linkage may reflect less adaptive dyadic processes.
To date, very little is known about the implications of couple cortisol linkage in the
context of pregnancy. The current study results point to stronger couple cortisol linkage
facilitating partner attunement and closeness in expectant parents. This finding complements the
results of the only two other studies that assessed prenatal couple cortisol linkage with a
psychosocial moderator (Braren et al., 2020; Saxbe et al., 2015). The three other studies that
examined cortisol linkage among expectant parents did not assess for moderators of linkage to
evaluate whether cortisol synchrony is adaptive in the context of pregnancy (Berg & Wynne-
Edwards, 2002; Edelstein et al., 2015; Storey et al., 2000). Braren et al. (2020) also found
evidence for cortisol linkage among expectant couples. Additionally, cortisol linkage was
stronger among pregnant couples with higher maternal psychological stress. For couples with
lower maternal psychological stress, couple cortisol linkage was non-significant. The authors
also found effects of a two-way interaction, in which among couples with higher maternal
psychological stress, yet lower paternal cortisol, mothers also had lower levels of cortisol,
meaning that even when mothers are experiencing high levels of stress, partners’ low cortisol
levels may buffer mothers’ cortisol. Although their results support the notion that prenatal
cortisol linkage may heighten stress transmission, they conclude that physiological linkage
79

during pregnancy may indicate partner attunement and adaptive interpersonal processes, given
that in the case of high maternal stress, it may be adaptive for mothers to signal to their low-
cortisol partners to help dampen her physiological stress response. While our study did not
examine prenatal couple cortisol linkage in the context of psychosocial stress, more studies are
needed to identify psychosocial contexts for which physiological linkage during pregnancy
facilitates or impairs positive dyadic adjustment.
This study has a number of notable strengths. It is one of the only five studies to date that
has assessed couple cortisol linkage in pregnancy and is one of the only longitudinal
investigations. To our knowledge, it is the first study to examine prenatal couple cortisol linkage
as a potential biopsychosocial predictor of postpartum depressive symptoms. Our study
contributes to the understanding of physiological linkage and couple interactions during
pregnancy and couple mental health during the postpartum period. Given that pregnancy and the
transition to parenthood involves drastic biological and endocrinological changes for the
pregnant mother and possibly her partner, it is worthwhile to explore how physiology impacts
the couple relationship during this transitional period to better understand the interplay of
biological processes, interpersonal dynamics and individual and couple well-being. Additionally,
our use of observational coding to quantify couple conflict interactions presents a richer, more
nuanced view of the couple’s relationship than self-report measures of relationship quality or
dynamics.
The limitations of this study include a highly-educated, opposite-sex sample of
cohabitating couples who are not entirely representative of first-time parents. Additionally, as we
only collected six saliva samples from participants over the course of the prenatal visit, we were
not able to model diurnal cortisol rhythms and get a broader snapshot of cortisol linkage
80

variations across days of sampling. Additionally, the cortisol collection and the conflict
discussion both took place during the prenatal visit, making the temporal direction of effects
more difficult to discern. Moreover, we did not include a moderator of psychological stress or
relationship satisfaction, limiting the conclusions we can draw regarding cortisol linkage and
contexts of stress during pregnancy. Future research should build on the couple of existent
studies on cortisol linkage during pregnancy and couple dynamics, including moderators of
psychosocial stress.
Couple’s cortisol linkage is a dynamic and complex process and implications of
physiological linkage may depend on the dyadic and physiological context. In some contexts,
linkage may signify a sense of closeness or connectedness, while in others, it may indicate poor
relationship functioning or contagion of partners’ stress levels (Timmons et al., 2015). In the
context of strong cortisol linkage among couples with high relationship strain (Liu et al., 2013)
or low relationship satisfaction (Saxbe & Repetti, 2010), physiological linkage may represent
maladaptive processes or poor relationship functioning. However, in other contexts, strong
physiological linkage may be representative of partner attunement, shared environment and
experiences, or time spent together (Laws et al., 2015; Papp et al., 2013; D. Saxbe & Repetti,
2010). Future research should aim to identify specific contexts for which couple cortisol linkage
is more and less adaptive for relationship processes. Our findings extend previous work and
support evidence that cortisol linkage can be detected among expectant parents. This study
contributes to the field’s understanding of physiological linkage, couple dynamics, and
postpartum health and adjustment.

81

References
Adouard, F., Glangeaud-Freudenthal, N. M. C., & Golse, B. (2005). Validation of the Edinburgh
postnatal depression scale (EPDS) in a sample of women with high-risk pregnancies in
France. Archives of Women’s Mental Health, 8(2), 89–95.
Bartko, J. J. (1966). The intraclass correlation coefficient as a measure of reliability.
Psychological Reports, 19(1), 3–11.
Belsky, J., Youngblade, L., Rovine, M., & Volling, B. (1991). Patterns of marital change and
parent–child interaction. Journal of Marriage and the Family, 53(2), 487–498.
https://doi.org///dx.doi.org.libproxy1.usc.edu/10.2307/352914
Berg, S. J., & Wynne-Edwards, K. E. (2002). Salivary hormone concentrations in mothers and
fathers becoming parents are not correlated. Hormones and Behavior, 42(4), 424–436.
https://doi.org/10.1006/hbeh.2002.1841
Braren, S. H., Brandes-Aitken, A., Ribner, A., Perry, R. E., & Blair, C. (2020). Maternal
psychological stress moderates diurnal cortisol linkage in expectant fathers and mothers
during late pregnancy. Psychoneuroendocrinology, 111.
https://doi.org/10.1016/j.psyneuen.2019.104474
Brennan, A., Ayers, S., Ahmed, H., & Marshall-Lucette, S. (2007, August). A critical review of
the Couvade syndrome: The pregnant male. Journal of Reproductive and Infant Psychology.
Routledge . https://doi.org/10.1080/02646830701467207
Cox, J., & Holden, J. (2003). Perinatal mental health: A guide to the Edinburgh Postnatal
Depression Scale (EPDS). Retrieved from https://psycnet.apa.org/record/2004-14522-000
Cox, M. J., Owen, M. T., Lewis, J. M., & Henderson, V. K. (1989). Marriage, Adult Adjustment,
and Early Parenting, 60(5), 1015–1024.
82

Datavyu Team. (2014). Datavyu: A Video Coding Tool. Databrary Project: New York
University. Retrieved from http://datavyu.org.
Edelstein, R. S., Wardecker, B. M., Chopik, W. J., Moors, A. C., Shipman, E. L., & Lin, N. J.
(2015). Prenatal hormones in first-time expectant parents: Longitudinal changes and within-
couple correlations. American Journal of Human Biology, 27(3), 317–325.
Ehlert, U., Gaab, J., & Heinrichs, M. (2001). Psychoneuroendocrinological contributions to the
etiology of depression, posttraumatic stress disorder, and stress-related bodily disorders:
The role of the hypothalamus-pituitary-adrenal axis. Biological Psychology, 57(1–3), 141–
152. https://doi.org/10.1016/S0301-0511(01)00092-8
F. Clinton, J. (1987). Physical and emotional responses of expectant fathers throughout
pregnancy and the early postpartum period. International Journal of Nursing Studies, 24(1),
59–68. https://doi.org/10.1016/0020-7489(87)90039-3
Glynn, L. M., Davis, E. P., & Sandman, C. A. (2013). New insights into the role of perinatal
HPA-axis dysregulation in postpartum depression. Neuropeptides, 47(6), 363–370.
Gold, P. W., Machado-Vieira, R., & Pavlatou, M. G. (2015). Clinical and Biochemical
Manifestations of Depression: Relation to the Neurobiology of Stress. Neural Plasticity,
2015, 1–11. https://doi.org/10.1155/2015/581976
Gordis, E. B., Margolin, G., Spies, L. A., Susman, E. J., & Granger, D. A. (2010). Interparental
aggression and parent – adolescent salivary alpha amylase symmetry. Physiology &
Behavior, 100(3), 225–233. https://doi.org/10.1016/j.physbeh.2010.01.006
Gottman, J. M., McCoy, K., Coan, J., & Collier, H. (1996). The Specific Affect Coding System
(SPAFF) for observing emotional communicationin martial and family interaction.
Mahwah, NJ: Erlbaum.
83

Gottman, John M, Coan, J., Carrère, S., & Swanson, C. (1998). Predicting marital happiness and
stability from newlywed interactions. Journal of Marriage and Family, 60(1), 5–22.
Groer, Maureen W., & Morgan, K. (2007). Immune, health and endocrine characteristics of
depressed postpartum mothers. Psychoneuroendocrinology, 32(2), 133–139.
https://doi.org/10.1016/j.psyneuen.2006.11.007
Groer, Maureen Wimberly, Davis, M. W., & Hemphill, J. (2002). Postpartum stress: Current
concepts and the possible protective role of breastfeeding. Journal of Obstetric,
Gynecologic & Neonatal Nursing, 31(4), 411–417. https://doi.org/10.1111/j.1552-
6909.2002.tb00063.x
Hallgren, K. A. (2012). Computing inter-rater reliability for observational data: An overview and
tutorial. Tutorials in Quantitative Methods for Psychology, 8(1), 23–34.
Heyman, R. E. (2001). Observation of couple conflicts: Clinical assessment applications,
stubborn truths, and shaky foundations. Psychological Assessment, 13(1), 5–35.
https://doi.org/10.1037/1040-3590.13.1.5
Holsboer, F. (2000). The Corticosteroid Receptor Hypothesis of Depression.
Neuropsychopharmacology, 23(5), 477–501. https://doi.org/10.1016/S0893-
133X(00)00159-7
Hruschka, D. J., Kohrt, B. A., & Worthman, C. M. (2005). Estimating between- and within-
individual variation in cortisol levels using multilevel models. Psychoneuroendocrinology,
30(7), 698–714. https://doi.org/10.1016/j.psyneuen.2005.03.002
Huta, V. (2016). When to Use Hierarchical Linear Modeling. Cahiers Victoriens and
Edouardiens, 2016(84), 13–28. https://doi.org/10.4000/cve.2973
IBM Corp. (n.d.). IBM SPSS Statistics for Windows. Armonk, NY: IBM Corp.
84

Jolley, S. N., Elmore, S., Barnard, K. E., & Carr, D. B. (2007). Dysregulation of the
hypothalamic-pituitary-adrenal axis in postpartum depression. Biological Research for
Nursing, 8(3), 210–222. https://doi.org/10.1177/1099800406294598
Koo, T. K., & Li, M. Y. (2016). A Guideline of Selecting and Reporting Intraclass Correlation
Coefficients for Reliability Research. Journal of Chiropractic Medicine, 15(2), 155–163.
Laws, H., Sayer, A. G., Pietromonaco, P. R., & Powers, S. I. (2015). Longitudinal changes in
spouses’ HPA responses: Convergence in cortisol patterns during the early years of
marriage. Health Psychology, 34(11), 1076–1089. https://doi.org/10.1016/S2215-
0366(16)30284-X.Epidemiology
Levenson, R. W., & Gottman, J. M. (1983). Marital interaction: Physiological linkage and
affective exchange. Journal of Personality and Social Psychology, 45(3), 587–597.
Levenson, R. W., & Gottman, J. M. (1985). Physiological and Affective Predictors of Change in
Relationship Satisfaction. J oumal of Personality and Social Psychology (Vol. 49).
Liu, S., Rovine, M. J., Cousino Klein, L., & Almeida, D. M. (2013). Synchrony of diurnal
cortisol pattern in couples. Journal of Family Psychology, 27(4), 579–588.
Longobucco, D. C., & Freston, M. S. (1989). Relation of Somatic Symptoms to Degree of
Paternal-Role Preparation of First-Time Expectant Fathers. Journal of Obstetric,
Gynecologic & Neonatal Nursing, 18(6), 482–488. https://doi.org/10.1111/j.1552-
6909.1989.tb00500.x
Matthey, S., Barnett, B., Kavanagh, D. J., & Howie, P. (2001). Validation of the Edinburgh
Postnatal Depression Scale for men, and comparison of item endorsement with their
partners. Journal of Affective Disorders, 64(2–3), 175–184. https://doi.org/10.1016/S0165-
0327(00)00236-6
85

Nierop, A., Bratsikas, A., Zimmermann, R., & Ehlert, U. (2006). Are stress-induced cortisol
changes during pregnancy associated with postpartum depressive symptoms?
Psychosomatic Medicine, 68(6), 931–937.
https://doi.org/10.1097/01.psy.0000244385.93141.3b
Noldus, L. P. (1991). The Observer: A software system for collection and analysis of
observational data. Behavior Research Methods, Instruments, & Computers, 23(3), 415–
429. https://doi.org/10.3758/BF03203406
Papp, L. M., Pendry, P., Simon, C., & Adam, E. K. (2013). Spouses’ cortisol associations and
moderators: Testing physiological synchrony and connectedness in everyday life. Family
Processes, 52(2), 284–298. https://doi.org/10.1146/annurev.neuro.31.060407.125627.Brain
Pruessner, J. C., Kirschbaum, C., Meinlschmid, G., & Hellhammer, D. H. (2003). Two formulas
for computation of the area under the curve represent measures of total hormone
concentration versus time-dependent change. Psychoneuroendocrinology, 28(7), 916–931.
https://doi.org/10.1016/S0306-4530(02)00108-7
Raudenbush, S., Bryk, A., & Congdon, R. (2019). HLM 8 for Windows. Lincolnwood, IL:
Scientific Software International, Inc.
Saxbe, D. E., Adam, E. K., Schetter, C. D., Guardino, C. M., Simon, C., Mckinney, C. O., …
Kennedy, E. (2015). Cortisol Covariation Within Parents of Young Children: Moderation
by Relationship Aggression Darby. Psychoneuroendocrinology, 62, 121–128.
https://doi.org/10.1016/j.psyneuen.2015.08.006.Cortisol
Saxbe, D., & Repetti, R. L. (2010). For Better or Worse? Coregulation of Couples’ Cortisol
Levels and Mood States. Journal of Personality and Social Psychology, 98(1), 92–103.
https://doi.org/10.1037/a0016959
86

Schneiderman, I., Kanat-Maymon, Y., Zagoory-Sharon, O., & Feldman, R. (2014). Mutual
influences between partners’ hormones shape conflict dialog and relationship duration at the
initiation of romantic love. Social Neuroscience, 9(4), 337–351.
https://doi.org/10.1080/17470919.2014.893925
Storey, A. E., Walsh, C. J., Quinton, R. L., & Wynne-Edwards, K. E. (2000). Hormonal
correlates of paternal responsiveness in new and expectant fathers. Evolution and Human
Behavior, 21(2), 79–95. https://doi.org/10.1016/S1090-5138(99)00042-2
Taylor, A., Glover, V., Marks, M., & Kammerer, M. (2009). Diurnal pattern of cortisol output in
postnatal depression. Psychoneuroendocrinology, 34(8), 1184–1188.
https://doi.org/10.1016/j.psyneuen.2009.03.004
Timmons, A. C., Margolin, G., & Saxbe, D. E. (2015). Physiological Linkage in Couples and its
Implications for Individual and Interpersonal Functioning: A Literature Review. Journal of
Family Psychology, 29(5), 720–731. https://doi.org/10.1037/fam0000115.Physiological
Yim, I., Tanner Stapleton, L., Guardino, C., Hahn-Holbrook, J., & Dunkel Schetter, C. (2015).
Biological and psychosocial predictors of postpartum depression: Systematic review and
call for integration. Annual Review of Clinical Psychology, 11, 99–137.
https://doi.org/10.1146/annurev-clinpsy-101414-020426

87

General Discussion
This dissertation investigated couple relationships during pregnancy to identify dyadic
behaviors and physiological markers as risk factors for postpartum depression. In Chapter 1, we
identified associations between positive conflict behavior and PPD symptoms. Mothers’ positive
conflict behavior predicted fewer maternal PPD symptoms, while fathers’ positive conflict
behavior predicted more maternal PPD symptoms. We found that negative conflict behavior was
unrelated to maternal PPD, and overall conflict behavior was not associated with paternal PPD.
While individuals’ displays of negative behavior increased the likelihood of their partners’
responding with more negative behavior, this negative behavior reciprocity during conflict was
not predictive of PPD.
In Chapter 2, we discovered effects of couples’ physiological linkage on negative conflict
behavior and paternal PPD. Stronger cortisol linkage was associated with less negative conflict
behavior for both partners and fewer symptoms of paternal PPD. We found that negative conflict
behavior did not moderate the effect of cortisol linkage and PPD, meaning that among couples
with higher cortisol linkage, an individual’s negative conflict behavior did not predict their
partner’s PPD.
While these two studies presented intriguing and illuminating effects, nearly all results
were inconsistent with our hypotheses. This may speak to the need to consider specific contexts
when examining couple dynamics. For instance, in Chapter 1, we hypothesized that negative
conflict behavior would predict more PPD symptoms for both mothers and fathers. Yet, findings
indicated that negative behavior was not related to PPD. Negative behaviors expressed during
conflict interactions may be normative and not problematic. However, while this interpretation
may explain our findings, it may not apply to some samples or within some contexts. Our sample
88

was highly-educated and reported generally high levels of relationship satisfaction. Additionally,
these couples enrolled in a research study knowing they would have to spend hours together and
engage in discussions about their relationship. While our results may apply to similar
populations, they may not generalize to couples with low relationship quality, high levels of
aggression, or clinical samples. This may also be the case for our finding that fathers’ positive
conflict behavior was associated with more maternal PPD symptoms. Perhaps among highly
conflictual couples, expressions of positive behavior on the part of the father would be associated
with better adjustment for mothers. Among our sample, fathers’ high levels of positive behavior
in the context of a conflict discussion may be indicative of less engagement or supportive
processes. Future research should continue to explore contextual factors that aid in understanding
of the nuances of couple conflict behavior and partner postpartum adjustment.
Additionally, as mentioned, context is important. This is particularly relevant when
investigating couple physiological linkage. In Chapter 2, we also found unexpected results.
Couples’ cortisol linkage was associated with less negative behavior and fewer paternal PPD
symptoms. Dynamic interpersonal and physiological processes such as couple conflict
interactions and cortisol linkage are features of couple relationships that vary depending on
contextual factors. Thus, although we examined cortisol linkage in the context of couple conflict,
conflict may not in and of itself produce increased stress responses, particularly among highly
satisfied couples. Instead, other factors such as high psychological stress or low relationship
quality may provide more insight into adaptive and maladaptive interpersonal dynamics.
Nevertheless, Chapter 2 provides evidence that couple physiological linkage may be beneficial
for satisfied expectant couples and can confer positive postpartum adjustment. While it’s
necessary to understand contexts for which couple synchrony can enact stress responses and
89

emotional dysregulation, it’s equally as important to examine situations for which couples’ levels
of emotional, behavioral, and physiological attunement may signify positive relationship
functioning.
Conclusion
Taken together, these two studies contribute to knowledge on couples’ dynamic
behavioral and physiological processes during the transition to parenthood. Couples’ conflict
interactions and mutual physiological influence are not one-size-fits-all. Instead, our findings
show that these dynamics vary across couples and contexts. Undoubtedly, these processes can
also vary for the same couple day-to-day. Individuals’ mood, behavior, and physiological states
are ever-changing; while there is still much to know about the implications of partners’ mutually
influencing and shared states, these studies provide further insight into adaptive couple linkage.

Abstract (if available)

Abstract Given that poor relationship functioning is one of the strongest predictors of maternal and paternal postpartum depression (PPD), couple conflict behavior is a valuable dyadic process to explore. The current dissertation examined whether couple’s observed behaviors during prenatal conflict interactions predict postpartum depressive symptoms for mothers and fathers. Additionally, couples’ correlated cortisol patterns can indicate a mutual influence of physiological and psychological states and may reflect adaptive or maladaptive relationship functioning depending on the context of linkage. Given that cortisol has also been implicated in postpartum depression, it is valuable to investigate these dyadic interpersonal processes during the transition to parenthood. Thus, this study also examined whether couple cortisol linkage during pregnancy is associated with mother and fathers’ negative relationship conflict behavior and their subsequent postpartum depressive symptoms. A total of 82 opposite-sex couples expecting their first child engaged in a conflict discussion and completed measures of relationship satisfaction and depressive symptoms during pregnancy. At approximately six-months postpartum, couples completed a measure of postpartum depressive symptoms. The findings suggest that for some couples during prenatal conflict, couple’s negative behaviors may be benign, and yet, fathers’ positive behavior may facilitate poor postpartum adjustment for their partners. Furthermore, these dissertation findings also present a context of couple cortisol linkage in which stronger physiological associations between partners may indicate adaptive processes. Directions for future research and potential mechanisms are discussed.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Couple conflict during pregnancy: Do early family adversity and oxytocin play a role?

PDF

Pregnancy in the time of COVID-19: effects on perinatal mental health, birth, and infant development

PDF

Relationships and meaning: examining the roles of personal meaning and meaning-making processes in couples dealing with important life events

PDF

Theory of mind processing in expectant fathers: associations with prenatal oxytocin

PDF

You always say that: physical aggression perpetration, linguistic behavior, and conflict intensity in young adult dating couples

PDF

Using observed peer discussions to understand adolescent depressive symptoms and interpersonal interactions

PDF

Prenatal predictors of parental sensitivity in first-time mothers and fathers

PDF

Predictors and outcomes across the transition to fatherhood

PDF

Prenatal sleep health, cortisol, and gestational weight gain

PDF

Inhibitory control in first-time fathers: Neural correlates and associations with paternal mental health

PDF

Neural and behavioral correlates of fear processing in first-time fathers

PDF

Fathers’ social-cognitive network connectivity: associations with prenatal father-infant attachment and postpartum parenting behavior

PDF

The physiology of compassion in couples’ discussions about loss

PDF

Young adult dating couple interactions in daily life: links to family aggression and physiological processes

PDF

Investigating discrimination and depression within a couple context

PDF

Adolescent life stress and the cortisol awakening response: the moderating roles of emotion regulation, attachment, and gender

PDF

Effectiveness and implementation of a maternal depression treatment in a Head Start setting: a mixed method study

PDF

Couples’ neuroendocrine activity in response to family conflict discussions: the role of self-reported anger and previous marital aggression

PDF

Biological and behavioral correlates of emotional flexibility and associations with exposure to family aggression

PDF

Spouse aggression, depression, and physical health: a multivariate longitudinal study of midlife couples

Asset Metadata

Creator Khaled, Mona (author)

Core Title The old ball and linkage: couples’ prenatal conflict behavior, cortisol linkage, and postpartum depression risk

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Degree Conferral Date 2021-08

Publication Date 07/31/2021

Defense Date 07/21/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag conflict,cortisol linkage,Couples,OAI-PMH Harvest,postpartum depression,Pregnancy

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Saxbe, Darby (committee chair), Baucom, Brian (committee member), Lai, Mark (committee member), Margolin, Gayla (committee member), Traube, Dorian (committee member)

Creator Email Mkhaled@usc.edu,Mona.mkhaled@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC15672113

Unique identifier UC15672113

Legacy Identifier etd-KhaledMona-9959

Document Type Dissertation

Format application/pdf (imt)

Rights Khaled, Mona

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu