Childhood Emotional Maltreatment and Couple Functioning Among Women Across Transition to Parenthood: A Process Model

Abstract

Research on the implications of childhood emotional maltreatment for later romantic relationship well-being has been accumulating. More efforts are needed to systematically delineate the etiological chains of the more proximal processes explaining why childhood emotional maltreatment, as a more distal risk factor, may ultimately result in later romantic relationship malfunctioning. Using multiple-wave data from a diverse community sample of primiparous women across the transition to parenthood (N = 159), we conducted a process model in which adult attachment, emotion regulation difficulties, and depressive symptoms were examined as potential mediators for the associations between childhood emotional maltreatment and various aspects of later couple functioning. Results demonstrate that adult attachment anxiety played a key role in accounting for the associations between childhood emotional maltreatment and the negative aspects of later couple functioning (i.e., conflict and ambivalence) through its associations with emotion regulation difficulties and depressive symptoms. In contrast, adult attachment avoidance more directly mediated the associations between childhood emotional maltreatment and the positive dimensions of later couple functioning (i.e., love and maintenance), without other intervening mediators. Overall, the currently identified risk chains contribute to a more nuanced understanding of the mechanisms via which childhood emotional maltreatment shapes later romantic relationship functioning and also help clinicians specify potential targets for effective interventions aimed at diminishing the detrimental consequences of childhood emotional maltreatment for later intimate relationship well-being.

Among various forms of childhood maltreatment, emotional maltreatment has been shown to be particularly influential in shaping adult adjustment (e.g., Infurna et al., 2016). It encompasses emotional abuse (i.e., verbal assaults and demeaning behaviors toward children by adults that are harmful for children’s self-worth) and emotional neglect (i.e., caregivers’ irresponsibility in satisfying children’s psychological needs for love and nurturance) (Bernstein et al., 1994). As compared to physical and sexual maltreatment, emotional maltreatment is more pervasive because it can occur in isolation and also constitutes a core component of other types of maltreatment, but its consequences are usually not immediate or easily observable and thus timely intervention is difficult (Stoltenborgh, Bakermans-Kranenburg, Alink, & van IJzendoorn, 2012). Recently, research on the implications of childhood emotional maltreatment for later romantic relationship well-being has been accumulating (Reyome, 2010). Findings of these studies suggest that individuals who had experienced early emotional maltreatment tend to have difficulties in later romantic relationships, because such early adversities likely contribute to vulnerabilities that may elevate the risk of close relationship distress, including insecure attachment (Riggs, Cusimano, & Benson, 2011), deficits in emotion regulation (Bradbury & Shaffer, 2012), and psychological distress (Perry, DiLillo, & Peugh, 2007).

In the present study, we seek to extend prior literature by testing how these mechanisms may operate together to explain why adulthood couple malfunctioning may track its roots back to childhood emotional maltreatment (Figure 1). Theoretically, we provide a process delineation of the etiological chains of the more proximal risk processes that account for why childhood emotional maltreatment, as a more distal risk factor, may ultimately result in later couple maladjustment. Testing such a process model also has applied implications because it may help identify intervention avenues to break the risk chains generated by early emotional maltreatment. Our study also contributes to the literature by replicating prior findings and testing the proposed model in a diverse community sample of primiparous women experiencing the transition to parenthood. This transition is a family life stage characterized by challenges that contribute to considerable changes in couple relationship well-being, mental health problems, and emotion regulation abilities (Nelson, Kushlev, & Lyubomirsky, 2014). Moreover, due to the increased cues for painful abuse memories (e.g., taking care of a dependent infant), mothers with a history of early maltreatment tend to have difficulties in adapting to this transition (Menke, Morelen, Simon, Rosenblum, & Muzik, 2018). Also given that maternal maladaptation to this transition has been linked to compromised parenting, poor offspring adjustment, and substantial social costs (Field, 2010), the practical significance of our study is further highlighted.

Figure 1. Childhood Emotional Maltreatment and Adulthood Couple Functioning: A Process Model Involving Adult Attachment, Emotion Regulation Difficulties, and Psychological Distress.

Note. Given the complexity of the proposed model and the present modest sample size, we examined different aspects of couple functioning separately (i.e., one aspect of couple functioning at a time; Panels A, B, C, and D). EMAL = Recalled Childhood Emotional Maltreatment, ANX = Adult Attachment Anxiety; AVD = Adult Attachment Avoidance, ERD = Emotion Regulation Difficulties, DEP = Depressive Symptoms, CFL = Couple Functioning – Love, CFM = Couple Functioning – Maintenance, CFC = Couple Functioning – Conflict, CFA = Couple Functioning – Ambivalence. Pre = Prenatal Wave, 6M = 6 Months Postpartum Wave, 1Y = Year Postpartum Wave. The following covariates were included in all analyses: Maternal Race (Pre), Maternal Education (Pre), Maternal Age (Pre), Maternal Neuroticism (Pre), Couple Relationship Status (Pre), Couple Relationship Length (Pre), Family Income-to-needs Ratio (Pre), and Child Negative Affect (6M). Covariates were included in the model as exogenous variables predicting couple functioning and also correlating with both predictors and mediators, but all the relevant correlation lines and predicting pathways are not depicted in the figure to increase clarity, and the relevant results are available from the authors upon request. All reported estimated parameters are standardized. Also to simplify presentation, measurement errors and residuals are not shown in the figure. Also for clarity, (a) pathways with p > .05 (two-tailed) are depicted in grey dash lines and the relevant parameter estimates are not reported; (b) pathways with p < .05 (two-tailed) are depicted in black solid lines; and (c) pathways for the direct effects from adult attachment and emotion regulation difficulties to various aspects of couple functioning were estimated in all analyses and only those significant ones are depicted in the figure. * p < .05, ** p < .01, *** p < .001 (two-tailed).

Theoretical Foundation

The current study is grounded in the model by Riggs (2010) on the processes via which early emotional abuse may consequently shape later romantic relationship functioning. This model proposes that, without intervention, childhood emotional abuse can disrupt the adaptive development of the attachment system, which, in turn, contributes to cognitive-affective deficits (e.g., emotional dysregulation and negative internal working models) that may elevate the risk of later social maladjustment and psychological distress and ultimately culminate in adulthood romantic relationship malfunctioning in different domains. Informed by Riggs’ model, we constructed a process model (Figure 1) in which childhood emotional maltreatment, as a distal risk factor, triggers a cascade of proximal risk processes before ultimately resulting in later couple malfunctioning. We hypothesized that childhood emotional maltreatment interferes with the normal development of attachment system. Attachment insecurity is then reflected in emotion regulation difficulties which increase vulnerability to depressive symptoms that eventually undermine multiple aspects of couple functioning.

Empirical Background

Childhood Emotional Maltreatment and Adult Attachment.

Early interactions with caregivers shape individuals’ attachment style and contribute to the development of internal working models composed of beliefs concerning selves, others, and interpersonal relations (Mikulincer & Shaver, 2016). Being raised by available and sensitive caregivers is critical for children to develop secure attachment. Securely attached children tend to perceive themselves as worthy of being loved, believe that others are responsive and trustworthy, and view interpersonal relations as safe and warm. In contrast, exposure to emotional maltreatment likely threatens individuals’ attachment security and results in a maladaptive internal working model characterized by negative representations with respect to selves, others, and social relations (Riggs, 2010).

The relational schemas implicated in the internal working model guide individuals’ expectations and interpretations of social information and shape their social behaviors (Dykas & Cassidy, 2011). Given that the early-formed attachment style and internal working model are relatively stable over time (while remaining open to revision in light of later life experiences) and can be carried over into adulthood (Waters, Merrick, Treboux, Crowell, & Albersheim, 2000), individuals who are insecurely attached due to childhood emotional maltreatment are likely to have trouble in later intimate relationships. Such a pattern of stability may be more likely in community samples, such as ours, than in high risk samples. Meta-analyses (Pinquart, Feußner, & Ahnert, 2013) indicate that attachment stability is lower in samples at high social risk than those at low social risk, in which the social risk can be defined by factors such as child maltreatment and parental psychopathology. In addition, there exists malleability of attachment security among children from maltreating families, which is evidenced by their higher susceptibility to beneficial effects of interventions as compared to those from nonmaltreated samples (Facompré, Bernard, & Waters, 2018).

Avoidance and anxiety are two central dimensions of adult attachment. As Mikulincer and Shaver (2019) stated, individuals scoring high on anxiety tend to exaggerate the seriousness of threats and over-emphasize their vulnerabilities (i.e., hyperactivation strategies), partly because they believe that the signs of helplessness may elicit attachment figures’ attention and thus increase the likelihood of receiving care. In contrast, individuals with high avoidance tend to defensively block or inhibit emotional states associated with a sense of weakness (i.e., deactivation strategies), partly because such strategies may protect them from experiencing attachment needs that would otherwise activate the attachment system and increase the chance of feeling vulnerable and getting rejected.

Due to their worries about being rejected and abandoned, adults with high attachment anxiety have a hypervigilant appraisal—monitoring system that is closely monitoring relationship-threatening cues (Campbell, Simpson, Boldry, & Kashy, 2005); in contrast, because of the importance that they assign to self-reliance and their needs to keep a safe distance, adults with high attachment avoidance are less likely to display affection when interacting with romantic partners (Collins, Cooper, Albino, & Allard, 2002). Ironically, the maladaptive actions of anxious or avoidant adults may contribute to what they fear the most (Mikulincer & Shaver, 2012). Anxious adults’ amplified perceptions of and overreactions to relationship-threatening cues may destabilize their relationships and eventually lead to relationship loss; and avoidant adults’ tendency to dismiss partner’s complaints and downplay the significance of conflicts to keep emotional distance, may incur more emotion-arousing interactions.

Accumulating research has corroborated that adult attachment, including both anxiety and avoidance, can serve as a mediating mechanism for the association between childhood emotional maltreatment and later couple relationship well-being (e.g., Riggs et al., 2011). Individuals who had experienced more emotional maltreatment in childhood tend to report higher adult attachment anxiety and avoidance, which, in turn, contribute to later couple maladjustment.

Childhood Emotional Maltreatment and Later Emotion Regulation.

Young children acquire their emotion regulation and interpersonal skills primarily via emotion socialization (Morris, Silk, Steinberg, Myers, & Robinson, 2007). Caregivers who commit emotional maltreatment may themselves have difficulties in regulating emotions and behaviors under stressful circumstances, and their dysfunctional behaviors provide children with maladaptive models of emotion regulation.

At the same time, positive emotion socialization processes occur less frequently in families with maltreatment (Shipman, Schneider, Fitzgerald, Sims, Swisher, & Edwards, 2007). The abusive family environment prevents children from expressing feelings, and maltreatment perpetrators are less likely to coach children about the identification of the causes and consequences of emotions, the ways to appropriately display emotions, and the strategies to effectively manage emotions; instead, in the face of child distress, abusive caregivers tend to demonstrate minimizing, punitive, and distress reactions (Edwards, Shipman, & Brown, 2005). Due to the chronic invalidation of their emotions, maltreated children likely internalize the beliefs that their emotions are unacceptable and thus engage in increased emotion inhibition and avoidance (Krause, Mendelson, & Lynch, 2003).

Both maladaptive behavioral models and negative emotion socialization processes impair the healthy development of children’s emotion regulation skills and interpersonal competence. After entering a close relationship, maltreated children’s deficits in emotion regulation are likely to be manifested in dysfunctional behaviors when interacting with romantic partners (e.g., aggression and avoidance), which foreshadows later relationship distress. Consistent with this notion, an emerging body of research (e.g., Bradbury & Shaffer, 2012) has indicated that the implications of childhood emotional maltreatment for later couple relationship well-being are at least partly accounted for by emotion regulation difficulties and associated problematic behaviors, such as intimate partner violence, deficits in the ability to read partners’ emotions, and maladaptive problem solving strategies.

Childhood Emotional Maltreatment and Later Psychological Distress.

The link between childhood maltreatment and later psychopathology has been extensively examined. Childhood emotional maltreatment has been identified as one particularly salient precursor for adult mental health issues, especially depressive symptoms (Infurna et al., 2016). Depressive symptoms as an intrapersonal stressor may spillover into the couple system and lead to relationship distress via undermining partners’ spousal role functioning (Heene, Buysse, & Van Oost, 2005). That is, depressive symptoms not only drain resources that otherwise partners might have devoted to promoting relationship (e.g., love expression) or solving relationship problems (e.g., constructive conflict resolution), but also may contribute to the formation of a gloomy context in which the couple functions on a daily basis. A handful of studies have identified depressive symptoms as a mediator for the link between childhood emotional maltreatment and later couple well-being (e.g., Liu et al., 2018).

Various Mechanisms Operate as an Etiological Chain.

Theoretically, childhood emotional maltreatment as a distal risk factor may trigger a cascade of more proximal risk processes before ultimately resulting in later couple distress. Each of the aforementioned factors constitutes one core part of this chain (Riggs, 2010), and the interrelations among them can be deduced from prior studies.

A review by Mikulincer and Shaver (2019) indicates that individual differences in attachment are reflected in their patterns of emotion regulation. When regulating negative emotions in stressful situations, individuals with high attachment avoidance tend to block or inhibit any emotional state and deny or suppress any emotion-related thoughts that are incongruent with their goals of keeping attachment needs deactivated, whereas individuals with high attachment anxiety tend to perceive negative emotions as congruent with their goals of intimacy-seeking and often exaggerate the seriousness of threats and their vulnerability to elicit attachment figures’ attention.

Furthermore, individuals who have difficulties in emotion regulation are susceptible to depressive symptoms. A substantial body of research has demonstrated that when individuals cannot effectively manage their emotional responses to everyday events, they tend to suffer from longer periods of and more severe levels of depressive symptoms (Aldao, Nolen-Hoeksema, & Schweizer, 2010). Depressed people differ from their non-depressed counterparts in the more frequent use of maladaptive emotion regulation strategies (Joormann & Stanton, 2016). In turn, there is a well-documented link between depressive symptoms and compromised couple functioning (Heene et al., 2005). Thus, a potential etiological chain of mediating processes might be: Childhood emotional maltreatment disrupts the adaptive development of the attachment system, attachment insecurity is reflected in difficulties in emotion regulation, emotion regulation difficulties increase vulnerability to depressive symptoms, and depressive symptoms ultimately undermine couple functioning.

The Current Study

Aims and Hypotheses.

Based on prospective data from a diverse sample of primiparous women across the transition to parenthood (N = 159), the central goal of the present study is to provide a delineation of the mediating mechanisms underlying the association between childhood emotional maltreatment and later couple relationship functioning by testing a process model that involves adult attachment, emotion regulation, and psychological distress (Figure 1). Guided by Riggs’ (2010) model, childhood emotional maltreatment is hypothesized to contribute to the development of adult attachment anxiety and avoidance; adult attachment anxiety and avoidance, in turn, predict difficulties in emotion regulation; emotion regulation difficulties further give rise to depressive symptoms; and depressive symptoms ultimately impair multiple aspects of couple functioning.

Analytic Approach.

Temporally ordered analyses were conducted using assessments across multiple waves and the baseline levels for some mediators were controlled for, helping more appropriately address the temporality of associations among variables (Preacher, 2015). Extensive demographic factors were controlled for, including maternal race, education, age, couple relationship status and length, and family income-to-needs ratio. Prior research has linked these variables with individuals’ reports of both childhood experiences and adulthood adjustment (e.g., Leerkes, Supple, Su, & Cavanaugh, 2015). Maternal neuroticism and child temperament were also considered. Neuroticism represents a maternal characteristic that may shape their self-reports of present and past experiences in general, because neurotic people tend to focus on the negative side of the world and experience distress regardless of the situation (Norris, Leaf, & Fenn, 2018). Further, neuroticism has also been demonstrated to be closely associated with emotion regulation difficulties, depressive symptoms, and marital malfunctioning (Paulus, Vanwoerden, Norton, & Sharp, 2016). We also included child temperament as a covariate because women in our sample were experiencing the transition to parenthood. Taking care of a difficult child may contribute to new mothers’ distress (Solmeyer & Feinberg, 2011) and elicit their memories of how they had been treated by their own parents, especially for those who had been maltreated (Menke et al., 2018). Considering the effects of these covariates helps rule out the possibility that the currently observed associations are a function of these variables rather than the study variables of interest. Last, it should be noted up front that given the complexity of our model and the modest size of the sample, we have limited statistical power to detect small indirect effects of childhood emotional maltreatment on couple functioning. Thus, our findings should be interpreted with caution and replicated with larger samples.

Method

Participants

Participants in the present sample were drawn from a larger study focusing on the origins of maternal sensitivity that included 259 primiparous mothers and their infants from the southeastern United States. Given that one of the key variables in the current study was couple functioning, women whose relationship status was single, divorced, separated, or widowed at the prenatal wave, the 6 months postpartum wave, or the 1 year postpartum wave were removed from the analytic sample. In addition, because mothers’ recalled childhood maltreatment experiences were assessed during the lab visit at the 2 year postpartum wave, attrition further reduced the sample size. Ultimately, this resulted in a sample of 159 mothers, of whom 94 (59.10%) self-identified as European American and 65 (40.90%) as African American. These mothers ranged in age from 18 to 44 years old (M = 26.23, SD = 5.66) at the prenatal assessment. Around 20.40% had a high school diploma or less, 24.20% had some college or a 2-year college degree, and 55.40% had a 4-year college degree or beyond. Eighty-six mothers (54.10%) were married and living together, 33.30% were non-married but living together, and 12.60% were non-married and not living together.” The length of couple relationship (assessed at the prenatal wave) ranged from .25 to 16 years, with an average of 4.22 years (SD = 3.27). Annual family income ranged from less than $2,000 to more than $100,000 (Median = $30,000–$39, 999).

To test potential selection effects, based on prenatal data we conducted a multivariate analysis of variance (MANOVA) to examine the differences in some key study variables (i.e., adult attachment anxiety and avoidance, emotion regulation difficulties, depressive symptoms, and different aspects of couple functioning) and covariates (i.e., maternal race, education, age, neuroticism, couple relationship status and length, and family income-to-needs ratio) between mothers who were the present analytic sample and those who were not. Results demonstrate that mothers included in the present analytic sample were older, had higher educational attainment, longer couple relationship duration, and were more likely to be married and living together with their partners. Yet, all the significant differences were small in their effect sizes (η² = .04-.06). Importantly, the two groups did not differ in race, family income-to-needs ratio, neuroticism, prenatal depressive symptoms, adult attachment, emotion regulation difficulties, or different aspects of couple functioning.

Procedures

Mothers were recruited from childbirth education and breast-feeding classes, local obstetric practices, and clinics, as well as referrals from other participants via flyers or presentations by members of the research team. Inclusion criteria were that women had to be 18 or older, African American or European American, fluent in English, and expecting their first child. Enrolled women were mailed consent forms and a packet of surveys. Women returned consent forms and surveys when visiting our lab for an interview 6–8 weeks prior to their due dates. Women were followed when infants were about 6 months (M = 6.39 months), 1 year (M = 13.90 months), and 2 years old (M = 27.32 months). At each wave, mothers and infants were observed in interactions and mothers also completed a series of surveys. Mothers reported their childhood maltreatment experiences prior to the 2-year visit. Mothers received $50, $ 100, and $ 120 at the conclusion of each lab visit for participation. A gift to mothers and a toy to infants were also given at each wave (i.e., canvas tote bag, water bottler, infant onesie, bib and sippy cup with study logo). The protocol was approved by the Institutional Review Board at the University of North Carolina at Greensboro (IRB 09 – 0035, Triad Child Study).

Measures

Couple Functioning.

The Romantic Relationship Questionnaire (RRQ; Braiker & Kelley, 1979) was used to assess couple functioning at the prenatal and the 1-year postpartum waves. It is a 25-item measure which assesses four aspects of romantic relationships, including love (10 items), maintenance (5 items), ambivalence (5 items), and conflict (5 items). Participants indicated the extent to which each item description applied to their relationship on a 9-point scale with differing responses for each aspect of functioning: 1 (not at all) to 9 (very much) for love; 1 (not at all) to 9 (a lot) for maintenance; 1 (not at all) to 9 (extremely) for ambivalence; and 1 (not serious at all) to 9 (very serious) for conflict. Example items include: “To what extent do you love your partner at this stage?” (love); “How much time do you and your partner spend discussing and trying to work out problems between you?” (maintenance); “How ambivalent or unsure are you about continuing in the relationship with your partner?” (ambivalence); and “When you and your partner argue, how serious are the problems or arguments?” (conflict). Item scores for each subscale were averaged to yield four composites, with higher scores indicating higher love, maintenance, ambivalence, and conflict. Cronbach’s alphas were .91, .81, .84, and .66 at the prenatal wave and .94, .84, .85,and.71 at 1-year postpartum wave for love, conflict, ambivalence, and maintenance subscales, respectively.

Childhood Emotional Maltreatment.

Childhood emotional maltreatment was assessed when the infants were around 2 years old using the emotional abuse (5 items) and neglect subscales (5 items) form the Childhood Trauma Questionnaire (CTQ) (Bernstein et al., 1994). The CTQ has been shown to be a reliable and valid retrospective assessment of individuals’ experience of early traumas in community samples (Scher, Stein, Asmundson, McCreary, & Forde, 2001). Mothers were asked to rate how well each statement described their experiences as a child on a 5-point scale ranging from 1 (never true) to 5 (very often true). A sample item (reversed) for emotional neglect is “There was someone in my family who helped me feel that I was important or special.” A sample item for emotional abuse is “People in my family called me things like stupid, lazy, or ugly.” Items for each subscale were summed, with higher scores indicating higher levels of emotional abuse or neglect. Cronbach’s αs were .89 for the emotional neglect subscale and .84 for the emotional abuse subscale.

Emotional neglect and abuse correlated with each other positively (r = .69, p < .001), and were thus averaged to represent the overall level of emotional maltreatment (α = .81).

Adult Attachment

Mothers completed the Experiences in Close Relationships Scale (ECR; Brennan, Clark, & Shaver, 1998) to assess adult attachment prenatally. It is a 36-item measure composed of two 18 item subscales, anxiety and avoidance. Participants indicate how strongly they agree with each item on a 7-point scale from 1 (disagree strongly) to 7 (agree strongly). Example items include: “I worry about being abandoned” and “I try to avoid getting too close to my partner.” Items for each subscale are averaged, with higher scores indicating higher avoidance and anxiety. Cronbach’s αs were .88 for the anxiety scale and .93 for the avoidance scale.

Emotion Regulation Difficulties.

Mothers completed the 36-item Difficulties in Emotion Regulation Scale (DERS; Gratz & Roemer, 2004) prenatally and at 6 months postpartum by rating how often specific items apply to them on a 5-point scale ranging from 1 (almost never) to 5 (almost always). The DERS has 6 dimensions: Non-acceptance (6 items; a tendency to have negative secondary emotional responses to one’s negative emotions, or non-accepting reactions to one’s distress); Goals (5 items; difficulties concentrating and accomplishing tasks when experiencing negative emotions); Impulse (6 items; difficulties remaining in control of one’s own behavior when experiencing negative emotions); Awareness (6 items; difficulties in the ability to acknowledge emotions); Strategies (8 items; the belief that there is little can be done to regulate emotions adaptively, once an individual is upset); and Clarity (5 items; the extent to which individuals are unclear about and unaware of the emotions they are experiencing). A total score was calculated as the average of subscales, with higher scores indicating higher emotion regulation difficulties. Cronbach’s alphas were .91 at the prenatal wave and .93 at 6 months postpartum.

Depressive symptoms.

The 20-item Center for Epidemiologic Studies—Depression Scale (Radloff, 1977) was used to measure depression at the prenatal wave and the 1-year postpartum wave. It consists of a checklist of depressive cognitions, moods, and feelings (e.g., “I felt that I could not shake off the blues even with help from my family or friends”). Mothers indicated how often they felt a particular way in the prior week on a 4-point scale ranging from 1 (rarely/never) to 4 (most of the time). Items were summed to derive a global score, with higher scores indicating higher depressive symptoms. Cronbach’s αs were .87 at the prenatal wave and .91 at 1 year postpartum.

Covariates.

Mothers completed the Infant Behavior Questionnaire-Revised Very Short Form at 6 months postpartum. We used mothers’ reports on negative affect (12 items; e.g., “When tired, how often did your baby show distress?’) on a 7-point scale from 1 (never) to 7 (always) as a measure of infant temperamental difficulty, with higher scores indicating higher difficulty (α = .74). Neuroticism was assessed at the prenatal wave using the neuroticism subscale (12 items) from the NEO Five-Factor Inventory (α = .81). Mothers reported their age, race, education family income, couple relationship status and duration at the prenatal phase. Family income-to-needs ratio was calculated using family income and the number of individuals in the household.

Analytic Strategies

Using SPSS 22.0, we performed analyses to examine descriptive statistics and zero-order bivariate intercorrelations among study variables and between study variables and covariates. Using Mplus 8.4, we conducted path analyses to test hypotheses. In Figure 1, mothers’ childhood emotional maltreatment, adult attachment at the prenatal wave, emotion regulation difficulties at 6 months postpartum, and depressive symptoms at 1 year postpartum were specified as exogenous variables predicting different aspects of couple functioning at 1 year postpartum. Given the model complexity and modest sample size, we examined different aspects of functioning separately (Panels A-D).

Moreover, adult attachment anxiety and avoidance at the prenatal wave, emotion regulation difficulties at 6 months postpartum, and depressive symptoms at 1 year postpartum were specified as mediators for the association between mothers’ childhood emotional maltreatment and different aspects of couple functioning at 1 year postpartum. Adult attachment anxiety and avoidance at the prenatal wave were specified as predictors to emotion regulation difficulties at 6 months postpartum and depressive symptoms at 1 year postpartum; and emotion regulation difficulties at 6 months postpartum was specified as a predictor to depressive symptoms at 1 year postpartum.

In addition, emotion regulation difficulties, depressive symptoms, and different aspects of couple functioning assessed at the prenatal wave as a baseline control were specified as exogenous variables predicting emotion regulation difficulties at 6 months postpartum, depressive symptoms at 1 year postpartum, and different aspects of couple functioning at 1 year postpartum, respectively. Covariates were included in the model as exogenous variables predicting different aspects of couple functioning at 1 year postpartum and also correlating with both predictors and mediators.

Indirect effects with different aspects of couple functioning as outcomes were assessed using bootstrapping (Preacher & Hayes, 2008). The standard errors (S.E) and confidence intervals (CIs) for were based on 2,000 bootstrap resamples. Conclusions regarding mediation are based on whether the indirect pathways are significant when examining 95% bias-corrected bootstrapped CIs around the unstandardized indirect associations. CIs that do not span 0 reflect significant indirect effects. Last, we evaluated model adequacy using various indices: Chi-Square statistic (χ²), comparative fit index (CFI), root-mean-square error of approximation (RMSEA), and standardized root-mean-square residual (SRMR). A non-significant χ², CFI > .90, RMSEA < .08, and SRMR < .08 indicate an acceptable model fit (Kline, 2011). Missing values were primarily due to unavailability of data from a specific wave and were handled via the full-information maximum likelihood method (FEML, Acock, 2005).

Results

In Table 1, we present: (a) the descriptive statistics for study variables and covariates; and (b) the bivariate intercorrelations among study variables and between study variables and covariates. Women’ reports of couple functioning at 1-year postpartum, depressive symptoms at 1-year postpartum, emotion regulation difficulties at 6 months postpartum, attachment anxiety and avoidance at the prenatal wave, and recalled childhood maltreatment correlated with each other in expected directions. Each covariate was correlated with at least two of the study variables, which further justifies the necessity of controlling for their effects in the primary analyses.

Table 1.

Descriptive Statistics for and Bivariate Intercorrelations Among Key Study Variables and Between Key Study Variables and Covariates.

		M	SD	1	2	3	4	5	6	7	8	9
Key Study Variables
1.	Recalled Childhood Emotional Maltreatment (Pre)	7.98	3.62	—
2.	Adult Attachment Anxiety (Pre)	3.14	1.01	.21	—
3.	Adult Attachment Avoidance (Pre)	2.16	.96	.37	.42	—
4.	Emotion Regulation Difficulties (6M)	1.71	.43	.25	.50	.42	—
5.	Depressive Symptoms (1Y)	10.46	9.16	.33	.32	.40	.45	—
6.	Couple Functioning-Love (1Y)	7.55	1.39	−.19	−.13	−.43	−.16	−.27	—
7.	Couple Functioning-Maintenance (1Y)	6.24	1.32	>−.04	−.001	−.20	−.06	−.07	.61	—
8.	Couple Functioning-Conflict (1Y)	4.20	1.63	.06	.33	.36	.12	.30	−.37	−.08	—
9.	Couple Functioning-Ambivalence (1Y)	2.58	1.77	.12	.20	.41	.17	.47	−.71	−.25	.57	—
Covariates
10.	Maternal Race (Pre) a	59.10	—	.01	−.001	−.14	.03	−.19	.29	−.07	−.14	−.34
11.	Maternal Education (Pre)	4.24	1.87	−.22	−.12	−.19	−.02	−.38	.19	−.06	−.15	−.31
12.	Maternal Age in Years (Pre)	26.23	5.66	−.12	−.02	−.15	−.06	−.20	.15	−.12	−.15	−.25
13.	Maternal Neuroticism (Pre)	2.48	.61	.28	.55	.34	.41	.48	−.19	.10	.33	.37
14.	Couple Relationship Status (Pre) b	54.10	—	.13	.18	.40	.06	.38	−.27	.02	.15	.37
15.	Couple Relationship Length in Years (Pre)	4.22	3.27	−.09	−.01	−.24	−.07	−.29	.22	−.03	−.08	−.32
16.	Family Income-to-Needs Ratio (Pre)	3.24	2.14	−.18	−.05	−.15	−.03	−.30	.28	.06	−.08	−.29
17.	Infant Negative Affect (6M)	3.41	.96	.02	.12	.10	.15	.24	−.12	.06	.10	.16
18.	Emotion Regulation Difficulties (Pre)	1.88	.41	.24	.54	.46	.56	.41	−.11	.003	.23	.22
19.	Depressive Symptoms (Pre)	12.92	7.96	.28	.41	.41	.32	.47	−.17	.06	.25	.29
20.	Couple Functioning-Love (Pre)	7.90	1.09	−.16	−.18	−.52	−.09	−.19	.43	.16	−.20	−.34
21.	Couple Functioning-Maintenance (Pre)	6.66	1.13	−.19	−.15	−.39	−.17	−.15	.22	.30	−.06	−.12
22.	Couple Functioning-Conflict (Pre)	4.17	1.48	.10	.29	.44	.15	.13	−.26	.04	.51	.29
23.	Couple Functioning-Ambivalence (Pre)	2.37	1.53	.15	.36	.58	.20	.35	−.44	−.05	.42	.60

Note. n ranges from 134 to 159. n represents the number of participants for each pair of correlation given missing data for some variables.

^{a, b}The means for maternal race and couple relationship status reflect the percentage of European American mothers and the percentage of mothers who were in a married relationship, respectively. When coding maternal race, “European American Mothers” were coded as “1” and “African American Mothers” were coded as “0.” When coding couple relationship status, “Married and Living Together Couples” were coded as “1”, “Non-Married but Living Together Couples” were coded as “2”, and “Non-Married and not Living Together Couples” were coded as “3.” Pre = Prenatal Wave, 6M = 6 Months Postpartum Wave, 1Y = 1 Year Postpartum Wave. The bolded correlations indicate significance with at leastp < .05 (2-tailed).

The model with “love” as the outcome showed an excellent fit: χ² = 28.839, df = 20, p = .091, RMSEA = .053 with 90% CI [.000, .093], CFI = .976, SRMR = ..065. As depicted in Panel A of Figure 1 and reported in Table 2, one indirect pathway emerged. Mothers’ childhood emotional maltreatment was negatively associated with their reports of love at 1 year postpartum through its positive association with attachment avoidance at the prenatal wave (B = −.039, S.E. = .019, 95% CI [−.091, −.012], β = −.101). Likewise, the model with “maintenance” as the outcome demonstrated an excellent fit: χ² = 32.754, df = 24, p =.109, RMSEA = .048 with 90% CI [.000, .085], CFI = .972, SRMR = .063. As depicted in Panel B of Figure 1 and reported in Table 2, one indirect pathway was identified. Mothers’ early emotional maltreatment was negatively associated with their reports of maintenance at the 1 year postpartum via its positive association with attachment avoidance at the prenatal wave (B = −.026, S.E. = .017, 95% CI [−.075, −.003], β = −.070).

Table 2.

The Significant Indirect Effects with Couple Functioning as the Ultimate Outcome.

Specific Significant Indirect Pathways in the Model	Bootstrapped Estimates for Indirect Effects
	B	S.E.	95% CI	β
Love Model
EMAL → AVD at Pre → CF-Love at 1Y	−.039	.019	[−.091, −.012]	−.101
Maintenance Model
EMAL → AVD at Pre → CF-Maintenance at 1Y	−.026	.017	[−.075, −.003]	−.070
Conflict Model
ANX at Pre -→ ERD at 6 M → DEP at 1Y -→ CF-Conflict at 1Y	.033	.018	[.008, .087]	.020
ERD at 6 M → DEP at 1Y → CF-Conflict at 1Y	.320	.165	[.083, .752]	.083
EMAL -→ ANX at Pre -→ ERD at 6 M → DEP at 1Y → CF-Conflict at 1Y	.002	.001	[.000, .007]	.004
Ambivalence Model
ANX at Pre → ERD at 6 M → DEP at 1Y → CF-Ambivalence at 1Y	.036	.020	[.010, .102]	.020
ERD at 6 M → DEP at 1Y → CF-Ambivalence at 1Y	.346	.171	[.103, .817]	.082
EMAL → ANX at Pre → ERD at 6 M → DEP at 1Y → CF-Ambivalence at 1Y	.002	.002	[.000, .011]	.004

Note. EMAL = Recalled Childhood Emotional Maltreatment, ANX = Adult Attachment Anxiety; AVD = Adult Attachment Avoidance, ERD = Emotion Regulation Difficulties, DEP = Depressive Symptoms, CF = Couple Functioning. Pre = Prenatal Wave, 6M = 6 Months Postpartum Wave, 1Y = 1 Year Postpartum Wave. The pathways with nonsignificant indirect effects based on the bias-corrected Bootstrapped 95% CI are not reported to simplify presentation.

The model with “conflict” as the outcome had an excellent fit to the data: χ² = 26.131, df = 21, p = .202, RMSEA = .986 with 90% CI [.000, .082], CFI = .986, SRMR = .067. As depicted in Panel C of Figure 1 and reported in Table 2, three indirect pathways emerged. First, mothers’ attachment anxiety at the prenatal wave was positively associated with their emotion regulation difficulties at 6 months postpartum, which, in turn, positively predicted their reports of couple conflicts at 1 year postpartum via contributing to their depressive symptoms at 1 year postpartum (B = .033, S.E. = .018, 95% CI [.008, .087], β = .020). Second, mothers’ emotion regulation difficulties at 6 months postpartum was positively associated with their reports of couple conflicts at 1 year postpartum via contributing to depressive symptoms at 1 year postpartum (B = .320, S.E. = .165, 95% CI [.083, .752], β = .083). Last, mothers’ early emotional maltreatment was positively associated with their attachment anxiety at the prenatal wave, which, in turn, was positively associated with their difficulties in regulating emotions at 6 months postpartum; and such difficulties ultimately predicted their reports of couple conflicts at 1 year postpartum via a positive association with depressive symptoms at 1 year postpartum (B = .002, S.E. = .001, 95% CI [.000, .007], β = .004).

The model with “ambivalence” as the outcome fit the data very well: χ² = 22.723, df = 17, p = . 159, RMSEA = .046 with 90% CI [.000, .091], CFI = .986, SRMR = .063. As depicted in Panel D of Figure 1 and reported in Table 2, three indirect pathways were identified. First, mothers’ attachment anxiety at the prenatal wave was positively associated with their emotion regulation difficulties at 6 months postpartum, which, in turn, positively predicted their reports of ambivalence at 1 year postpartum via contributing to their depressive symptoms at 1 year postpartum (B = .036, S.E.=.020, 95% CI [.010, .102], β = .020). Second, mothers’ emotion regulation difficulties at 6 months postpartum was positively associated with their ambivalence at 1 year postpartum via contributing to their depressive symptoms at 1 year postpartum (B = .346, S.E. = .171, 95% CI [.103, .817], β = .082). Last, mothers’ early emotional maltreatment was positively associated with attachment anxiety at the prenatal wave, which, in turn, was positively associated with their emotion regulation difficulties at 6 months postpartum; and such difficulties ultimately predicted their reports of ambivalence at 1 year postpartum via a positive association with depressive symptoms at 1 year postpartum (B = .002, S.E. = .002, 95% CI [.000, .011], β = .004). In terms of effect sizes (Kenny, 2012), standardized indirect effects .01 were interpreted as “small”, .09 as “medium”, and .25 as “large.”

Discussion

The goal of this study was to test a process model to delineate a risk chain of more proximal processes via which childhood emotional maltreatment as a more distal factor may predict later couple malfunctioning. In general, consistent with hypotheses, results reveal an etiological cascade by which childhood emotional maltreatment disrupts the adaptive development of attachment system. Attachment anxiety (but not avoidance) is reflected in difficulties in regulating emotions; emotion regulation difficulties further increase vulnerability to depressive symptoms; and depressive symptoms ultimately contribute to both conflict and ambivalence in later couple relationships. In addition, our results also indicate that childhood emotional maltreatment may undermine love and maintenance in later romantic relationships via its contribution to attachment avoidance (but not anxiety). Overall, this study provides some evidence supporting the model by Riggs (2010) on the processes through which early emotional abuse may shape later romantic relationship functioning.

A Risk Chain Composed of Various Mediating Mechanisms

Each of the mediators (i.e., adult attachment, emotion regulation difficulties, depressive symptoms) involved in the identified risk chain has been separately corroborated in prior research (Bradbury & Shaffer, 2012; Perry et al., 2007; Riggs et al., 2011). The uniqueness of our study lies in examining the interrelations among these mediators as guided by a theoretical model by Riggs (2010). Delineating how various mechanisms may operate as a long cascade from the more distal factors to the more proximal factors is critical for a complete and nuanced understanding of the developmental processes via which early emotional maltreatment exerts its far-reaching influences on later couple relationship well-being. Such an unifying model is also beneficial for the more rapid expansion of knowledge base in this field because such a model makes it easier to integrate findings across studies.

From a practical perspective, the identified risk mediating chain clearly indicates that different factors may work as a successive series of interlocking, shifting gears (Repetti et al., 2011). That is, alterations in a more distal gear would set in motion subsequent shifts in a more proximal gear, and the shifts in the most proximal gear would ultimately result in the focal outcomes. Thus, our findings specify potential targets for interventions aimed at diminishing the effects of childhood emotional maltreatment on later couple relationship well-being. For example, for women who suffered from early emotional maltreatment, attending training programs on emotion regulation may help them cope with the challenges of childhood emotional maltreatment for their romantic relationship adjustment.

The Differential Roles of Adult Attachment Anxiety and Avoidance

It is interesting that attachment anxiety plays a key role in the aforementioned risk chain accounting for the implications of childhood emotional maltreatment for the negative aspects of later couple functioning (i.e., conflict and ambivalence), whereas attachment avoidance mediates the associations between childhood emotional maltreatment and the positive aspects of later couple functioning (i.e., love and maintenance) but does not constitute a core part of the longer risk chain primarily due to the nonsignificant link between attachment avoidance and emotion regulation difficulties. Several lines of research may be helpful for understanding such findings.

First, different attachment dimensions may be associated with distinct emotion regulation strategies (Mikulincer & Shaver, 2019). Individuals with high anxiety are characterized by worry of abandonment and tend to use hyperactivation strategies to elicit attachment figures’ care; in contrast, individuals scoring high on avoidance are typically fearful of intimacy and uncomfortable with dependence and thus tend to suppress their emotions to reach the goals of keeping attachment needs deactivated. Consistent with this, Wei et al. (2005) found that attachment anxiety and avoidance contributed to negative mood and interpersonal problems via distinct affect regulation strategies. Specifically, the link between attachment anxiety, negative mood, and interpersonal problems was mediated only by emotional reactivity (i.e., emotionally hypersensitive to problems), whereas the link between attachment avoidance, negative mood, and interpersonal problems was mediated only by emotional cutoff (i.e., suppress negative feelings and maximize distance from others). Thus, future research is needed to utilize the established measures that are more appropriate for assessing hyperactivating/deactivating affect regulation strategies (e.g., the Differentiation of Self Inventory, DSI; Skowron & Friedlander, 1998) to test this possibility. Likewise, future research may also benefit from examining the roles of various subtypes of emotion regulation difficulties assessed with the DERS in explaining the distinct implications of attachment anxiety versus avoidance (for a detailed discussion, see Marganska, Gallagher, & Miranda, 2013). We did not address this question in the present study primarily due to the limited statistical power associated the modest sample size.

Second, as Shaver and Mikulincer (2002) suggested, attachment avoidance and anxiety involve the differential activation of the cognitive, affective, and behavioral subsystems of attachment. Accordingly, in addition to emotion regulation as an affective mechanism, the alternative cognitive processes (e.g., attribution; Collins, Ford, Guichard, & Allard, 2006) and behavioral processes (e.g., communication; Guerrero, Farinelli, & McEwan, 2009) that were not tested in the present study merit attention in future research to more adequately account for the distinct roles of attachment avoidance and anxiety in shaping romantic relationship outcomes. Relatedly, from a practical perspective, as attachment (in)security comprehensibly shapes the ways that people think, feel, and behave (Dykas & Cassidy, 2011; Mikulincer & Shaver, 2019), therapists should sensitively consider the particular characteristics of the cognitive, emotional, and behavioral subsystems that are differentially activated by attachment anxiety versus avoidance, when working with childhood maltreatment survivors to deal with attachment issues and their implications for romantic relationship well-being.

Third, as Fraley and Shaver (2000) proposed, the major interpersonal functions of each adult attachment dimension may be different. For anxious individuals who are highly worried about being abandoned, their appraisal-monitoring system tend to be set at a lower threshold and their hypervigilance may intensify close monitoring and distorted appraisal of relationship-threatening cues (e.g., conflict and ambivalence) more than the support-related cues (Campbell et al., 2005). For individuals scoring high on attachment avoidance scale, given that they value self-reliance and one of their interpersonal goals is to maintain a safe emotional distance and minimize dependence, they are less likely to engage in affectionate exchanges with romantic partners (Collins et al., 2002; Mikulincer & Shaver, 2012). Indeed, results of a meta-analytic review by Li and Chan (2012) indicated that attachment anxiety and avoidance were both negatively associated with various aspects of intimate relationship quality; yet, attachment avoidance was more negatively associated with satisfaction, connectedness, and support in relationships, whereas attachment anxiety was more positively associated with conflict in relationships. Such patterns were consistent with our findings.

Last, it is important to note that prior research (Mikulincer, Shaver, Cassidy, & Berant, 2009) suggests that individuals scoring high on attachment avoidance tend to display a defensive denial of vulnerabilities and minimize the subjective awareness of distress in order to keep the needs of attachment deactivated and maintain a “strong” self-reliant figure. Thus, when assessing problematic characteristics (e.g., emotion regulation difficulties and depressive symptoms) with self-report surveys among highly avoidant individuals, the obtained responses may be biased (i.e., underreporting); in contrast, research utilizing the more objective measures of emotion regulation and distress (e.g., physiological indices) should be less susceptible to the self-report biases (e.g., Maunder, Lancee, Nolan, Hunter, & Tannenbaum, 2006). Thus, future research may benefit from utilizing multiple methods when measuring emotion regulation and distress to see if the current results are replicable.

The Context of Transition to Parenthood

The present findings should be understood in the context of transition to parenthood. The transition to parenthood is a unique life stage characterized by both joys and stressors (associated with preparing for the arrival of the baby and raising the young child) and during which all the currently examined mediating and outcome factors are quite salient and open to changes (Nelson et al., 2014). Taking care of a highly dependent infant also may elicit memories of how new parents were treated by their parents, especially for those who had been maltreated (Menke et al., 2018). The present study reveals the interrelations among these factors and how they operate as a cascade from the more distal factors to the more proximal factors in shaping new mothers’ relational adaptation, highlighting several targets for interventions, especially those focusing on new mothers with early traumas. For example, emotionally focused marital therapy, an approach built on attachment theories and couple interaction research, may be particularly helpful for promoting relationship well-being among couples in which one or both partners had experienced early traumas (Dalton, Greenman, Classen, & Johnson, 2013), as our results suggested that adult attachment and emotion regulation difficulties serve as two core explanatory mechanisms for the effects of childhood emotional abuse on later couple functioning.

Another implication of our findings is that mothers’ adaptation to the transition to parenthood is multifaceted and their adaptation in different domains are interrelated. For example, as shown in our model, maternal intrapersonal (e.g., psychological distress, emotion regulation difficulties) and interpersonal functioning (e.g., couple relationship well-being) were closely related with each other. However, the majority of existing interventions focus on new mothers’ adaptation in only a single domain, with few targeting adaptations in different domains and their interrelations simultaneously. Also considering the debate on the effectiveness of existing intervention programs (Pinquart & Teubert, 2010), our findings suggest that developing interventions addressing adaptation in different domains at the same time may be one avenue to more successfully assist new mothers.

Limitations and Future Directions

Limitations of this study and additional avenues for future research should be noted. First, the present sample is a non-representative, community sample of primiparous women, and the sample size was modest. We excluded women whose relationship status was single, divorced, separated, or widowed at any wave because one of the central variables was couple functioning; thus, the present sample focuses on women in relatively stable relationships. It is likely that some women with high attachment avoidance were not included in the analytic sample due to not being in a close relationship (Pepping & MacDonald, 2019). Future studies should examine the present model in mothers with and without partners to identify potential differences. Notably, the range of couple relationship duration in our sample was relatively large. To some extent, relationship duration can be viewed as a proxy for critical characteristics of the couple relationship (e.g., closeness, commitment). In particular, the interdependence between partners is primarily derived from their shared living experiences over the course of their relationships. All things being equal, longer relationship duration may imply higher levels of interdependence between partners (Veroff, 1999). It seems necessary for future research to explore the potential differences of the currently tested models as a function of relationship duration.

As to the present sample, concerns also may arise about the incidence rate of maltreatment in this community sample. Defining the “moderate to severe” emotional maltreatment assessed with the CTQ using the cut-off score criteria by Walker et al. (1999), in the present sample, the rates were 10.10% with 95% CI [5.37%, 14.75%] and 20.1% with 95% CI [13.88%, 26.38%] for emotional neglect and emotional abuse, respectively. The current rate of emotional neglect is higher than that reported by Scher et al. (2004) in a community sample of women (5.3%, 95% CI [3.4%, 7.3%], t = 1.99, p = .048), whereas the current rate of emotional abuse is comparable to that reported by Scher et al. (2004) (14.3%, 95% CI [11.2%, 17.3%], t = 1.83, p = .07). Nevertheless, efforts should be made to see if similar results can be found with larger, more representative and higher risk samples.

Second, mothers recalled childhood maltreatment. The validity of retrospective measures has been questioned because memories may be subject to distortion and early events may be selectively recalled (Tajima, Herrenkohl, Huang, & Whitney, 2004). Nevertheless, research indicates that adults’ reports of childhood parenting experiences are stable over time and correlate significantly, albeit modestly, with parents’ reports, and more strongly with siblings’ reports (Brewin, Andrews, & Gotlib, 1993; Harlaar et al., 2008); and in particular, an examination of the self-reported prevalence and test-retest reliability of exposure to child maltreatment using the CTQ among women around pregnancy also indicates an adequate consistency in reporting (Cammack et al., 2016).

Third, shared method and informant variance may inflate the links identified in this study, given that all constructs were assessed with self-report surveys filled by women themselves. Research will benefit from using multi-informant and multi-method designs when assessing contiguous constructs. Fourth, we examined statistical power using the post hoc Monte Carlo simulation analysis with the obtained model estimates used as population values and 10,000 replications (Thoemmes, MacKinnon, & Reiser, 2010). Results indicated that we had limited power to detect small indirect effects of childhood maltreatment on couple functioning. Research with larger samples is needed.

Fifth, we did not collect data from fathers. Future research should collect data from both mothers and fathers and approach the current model from a dyadic perspective to adequately tackle the complexity within the interdependent enterprise of couple relationship (Liu et al., 2018). Fathers and mother also may experience the transition to parenthood differently (Mickelson & Biehle, 2017), which further highlights the necessity of obtaining data from both fathers and mothers.

Sixth, according to the family systems theory, the link between intrapersonal functioning and couple relationship functioning is likely bidirectional (Kung, 2000). We conducted cross-lagged models in which depressive symptoms and the specific dimensions of couple functioning assessed at the prenatal and 1 year postpartum waves were considered. Although within each wave, depressive symptoms and couple functioning correlated with each other in expected directions, they were not associated with each other over time. The relevant results are available from authors upon request. Nevertheless, research is needed to systematically test the potential reciprocity in the link between individual functioning and couple functioning across the transition to parenthood (Choi, 2016).

Last, we cannot provide a more rigorous estimation of the proposed model given the current data set. Following recommendations for testing mediational hypotheses (Preacher, 2015), a more rigorous approach may be conducting a full cross-lagged panel model. Using the current data set, the best we could do was temporally ordering most of the examined factors across different assessment waves, while controlling for the baseline levels for most of the mediators and all outcomes. Thus, our findings await replications with more “ideal” approaches for mediation models.

Conclusion

Our results demonstrate an etiological risk chain of proximal processes (i.e., adult attachment anxiety and avoidance, emotion regulation difficulties, and depressive symptoms) explaining why childhood emotional maltreatment, as a distal risk factor, has far-reaching influences on later couple functioning. Practical implications of our findings lie not only in specifying factors that may be intervened to diminish the link between early emotional maltreatment and later couple malfunctioning, but also in highlighting avenues to help first-time mothers whose childhood had been clouded with emotional maltreatment successfully navigate challenges during the transition to parenthood.