Satisfaction with parental responsibilities and disorganized attachment among infants of mothers at risk for depression

Abstract

Growing attention has been placed on examining the family environment as antecedent of attachment, including the coparenting relationship. Parents’ satisfaction with the coparenting relationship may be particularly of interest when parents are at heightened risk for depression, as depression has been consistently linked to negative coparenting, poor quality of parenting, and insecure infant attachment. However, no study has examined the effects of parents’ satisfaction with the coparenting relationship on attachment. The present study examined mothers’ satisfaction with division of childrearing responsibilities, a component of coparenting, and its longitudinal and cross-sectional links with infant disorganized attachment, examining the quality of mothering as a mediator, in a sample of infants and mothers at elevated risk for depression (N=234). We assessed maternal depressive symptoms at 3, 6, and 12 months of infant age, mothers’ satisfaction with the division of parental responsibilities at 3 and 12 months, the quality of mothering at 6 and 12 months, and infant disorganized attachment at 12 months. Mediation analyses revealed that at 12 months, mothers who were unsatisfied with fathers’ childrearing responsibilities had poorer quality of mothering, which in turn was linked to disorganized attachment in their infants. However, the longitudinal indirect association between satisfaction with childrearing responsibilities at 3 months and disorganization at 12 months mediated by maternal parenting at 6 months was not significant. Findings emphasize the importance of partner support in childrearing for mothers at risk for depression in shaping a healthy relationship between mothers and their infants, particularly as infants get older.

Attachment theory establishes the parents’ ability to be sensitive, prompt, appropriate, and responsive in their interactions with their infant to be the core predictor of infant attachment security (Ainsworth et al., 1974; Bowlby, 1969). However, decades of research since the formation of attachment theory have found the sensitivity-attachment link to be weaker than Bowlby had theorized (e.g., De Wolff & van IJzendoorn, 1997; Nievar & Becker, 2008), and as a result, attention has shifted to examining additional mechanisms that may account for the variability in attachment security. In recent years, increasing attention has been paid to taking a family systems perspective (Cox & Paley, 1997) that emphasizes the role of family contextual factors on infant attachment security, including coparenting.

Coparenting is an important family-level process during the transition to parenthood, as parents experience high levels of stress and strain that comes with new parenthood (Feinberg, 2002). As a result, parents’ satisfaction with the coparenting relationship during this period may be especially predictive of other areas of functioning, such as the parent-infant relationship. Of particular interest in the present study is mothers’ satisfaction with one component of coparenting, the division of daily parental responsibilities. We are interested to see whether mothers’ satisfaction with fathers’ support in parental roles and responsibilities will be salient in families with mothers at heightened risk for depression. Depressed mothers are at elevated risk for coparenting conflict (Tissot et al., 2017) and poor quality of parenting (Goodman et al., 2020; Lovejoy et al., 2000), and their infants are at increased risk for disorganized attachment (Bigelow et al., 2018; Hayes et al., 2013; Teti et al., 1995). Together, these findings suggest a link between coparenting and attachment in infants and mothers at risk for depression. In the present study, we aim to understand the links between mothers’ satisfaction with the division of parental responsibilities, quality of mothering, and disorganized infant-mother attachment in a sample of infants and their mothers at elevated risk for depression.

Coparenting satisfaction

Coparenting is defined as the degree to which two or more individuals support, share, and divide childrearing responsibilities (Feinberg, 2003; Talbot & McHale, 2003). Research on coparenting has focused on the quality of this coordinated parenting effort, and less on parents’ satisfaction with the coparenting arrangement (Riina & McHale, 2012). However, exploring parents’ feelings about the parenting team may provide a unique perspective and understanding of the interpersonal processes within the coparenting relationship (Bradbury et al., 2000; Van Egeren, 2004).

Coparenting is a complex, triadic relationship that is comprised of multiple subcomponents characterized by both emotional and physical manifestations of coparenting behaviors. One component of coparenting relates to the division of daily roles, tasks, and duties that goes into caring for the child (Feinberg, 2003). Parents’ satisfaction with the division of parental roles and responsibilities may be especially important in the transition to parenthood because parents’ abilities to support each other in their parenting roles play an integral role in organizing family processes during this early stage of parenthood (Feinberg et al., 2016; Kim et al., 2021). The birth of a new infant is a period when the infant becomes a priority and the central focus of the family, with growing labor and responsibilities related to childrearing. As a result, how satisfied parents are with their ability to work together as a team to rear the infant and divide parenting tasks would be a crucial indicator and determinant of individual and family functioning (Feinberg et al., 2016; Gordon et al., 2022). Thus, the couple’s satisfaction with the division of responsibilities in the postpartum can spill over into other areas of functioning, such as the quality of the parent’s individual interactions with the infant, which then in turn can affect infants’ quality of attachment (Kim et al., 2021).

Indirect relationship between satisfaction with parental responsibilities and attachment

From a family systems perspective (Cox & Paley, 1997), a parent may be more competent and sensitive in their individual interactions with their infant when they feel supported by the spouse in their parental roles and responsibilities (Caldera & Lindsey, 2006; Kim et al., 2021). Social support has been well-established to play a significant role on the quality of parenting and overall parental functioning (e.g., He et al., 2021; Radunovich et al., 2017; Taraban & Shaw, 2018). When this support comes from the spouse and is in direct relation to childrearing, it may play an even more salient and positive role in the parent’s individual interactions with their infant, relative to general social support. Indeed, coparenting is unique in that it combines the couple’s relationship quality with their individual relationships with the child, and research suggests coparenting to be more closely related to parenting than any other types of family processes (Feinberg, 2002). Because attachment theory establishes a sensitive parent-infant interaction to be the “final common pathway” to a secure attachment (Bowlby, 1969), we can hypothesize an indirect pathway between satisfaction with the spouses’ support in daily childrearing responsibilities and attachment, through the quality of parenting.

Coparenting and attachment in the literature

To our knowledge, no published empirical work has explored the indirect relationship between satisfaction with coparental roles and attachment. Studies examining the coparenting-attachment link have done so in the context of the quality of coparenting. For example, one study has found an indirect association between mothers’ perceptions of coparenting quality and infant-mother attachment, through parenting quality, using currently accepted tests of mediation (Kim et al., 2021). Specifically, they found an indirect association between mother-reported coparenting quality across the first year and the quality of infant-mother attachment at infants’ 12 months, mediated by maternal emotional availability across the first year. Post hoc analyses found longitudinal linkages as well, finding mother-reported coparenting quality at 1 month postpartum to be significantly associated with attachment at 12 months, again mediated by maternal emotional availability across the first year. The empirical evidence for the indirect link between coparenting quality and attachment supports the idea that a parent’s satisfaction with the spouses’ support in parenting duties may also be indirectly associated with attachment security, through the quality of parenting.

Why examine these associations in women at elevated risk for depression?

The relationship between coparenting and attachment has not yet been observed in a sample of women who are at risk for depression, although examining these associations among women at elevated risk for depression may be important for several reasons. First, infants of depressed mothers are already at risk for developing insecure attachment, particularly disorganized attachment (e.g., Erickson et al., 2019; Hayes et al., 2013). Disorganized infants differ from secure, avoidant, and resistant infants in that while the latter three exhibit “organized” patterns of strategies in responding and adapting to stressful situations, disorganized infants lack organized strategies in regulating their distress (Granqvist et al., 2017; Lyons-Ruth & Jacobvitz, 2018). For example, although they are insecurely attached, avoidant and resistant infants still have a consistent, predictable pattern of handling distress, whether it is by maximizing or minimizing their negative emotions, and being characterized by a heightened attachment versus exploratory system. However, disorganized infants do not have an organized strategy in responding to distress, due to the fact that they view their attachment figure (who should be the source of comfort) as a source of alarm. This leads to odd, contradictory, and conflicting behaviors toward the attachment figure during times of distress, such as freezing, stilling, and trying to move away from the attachment figure but simultaneously trying to approach them. Empirical and theoretical work have suggested that when parents struggle from mental disorders or have accumulative risks and stressors, they are more likely to behave in ways that are frightening to the infant (Granqvist et al., 2017; van IJzendoorn et al., 1999). Thus, because our sample is comprised of infants and their mothers with a heightened risk for depression, and therefore at an elevated risk for disorganized attachment, it may be especially important to examine precursors of disorganized attachment in particular. Moreover, studies have found infants with disorganized attachment to be at heightened risk for negative outcomes such as externalizing, even compared to infants with insecure-avoidant or insecure-resistant attachments (Fearon et al., 2010; van IJzendoorn et al., 1999). This emphasizes the need to explore mechanisms that lead to disorganization in infancy, in order to intervene as early as possible.

Second, depression in mothers can also have deleterious spillover effects on their interactions with the infant (Goodman et al., 2020; Lovejoy et al., 2000). Maternal depression has consistently been linked with lower levels of sensitivity in the literature. That is, during interactions with their children, greater depression in mothers is associated with lower warmth and responsiveness, higher harshness and intrusiveness, more frequent disengagement, and lower self-efficacy (e.g., Dix & Moed, 2019; Goodman, Simon, et al., 2022). Mothers are especially at risk for depression during the postpartum period (Reck et al., 2008), when they experience a major stress from the new role and responsibilities that follow the transition into parenthood (Cowan & Cowan, 2000). Indeed, the childbearing years for women is noted to be the highest period for onset of depressive disorders (Goodman, 2007), with at least one in five women experiencing depressive symptoms in the first year of postpartum (O’Hara & Wisner, 2014).

For all of these reasons, coparenting, specifically the spouse’s support in childrearing roles and responsibilities, may be especially salient for infants and their mothers at elevated risk for depression in the postpartum period. If depressed mothers are highly satisfied with the spouses’ parental role support during the stress and chaos of the postpartum period, that will likely have a positive effect on mothers’ individual parenting. In addition, infants are entirely dependent on their parents in the first year of life, and the parents’ ability to support and cooperate with the partners’ parenting efforts may shape the infant’s attachment system, more so than any other time in the infant’s life (Solomon & George, 2018).

Current study

In summary, the present study of mothers at risk for depression investigated the relationship between mothers’ coparenting satisfaction and attachment in infants. We examine a specific component of coparenting, the division of day-to-day parental roles and responsibilities, and how satisfied mothers are with the fathers’ support in these daily parenting responsibilities. Specifically, we explore the following questions based on the theoretical and empirical evidence presented: Is coparenting satisfaction in mothers at elevated risk for depression associated with disorganized attachment in their infants at 1 year? Is this association indirect, through the mothers’ quality of interactions with their infant?

Method

Procedures

Data were collected as part of a larger, longitudinal study examining pathways to infant vulnerability to the development of psychopathology. All women and a subset of their partners participated in an informed consent procedure. All procedures were approved by the Emory University Institutional Review Board (protocol number: IRB00004249). This study was not preregistered. At enrollment, prior to 16 weeks gestation, women participated in diagnostic interviews to capture lifetime depression diagnoses, to determine eligibility for the study. Mothers provided data on demographics. At infant age 3-, 6-, and 12- months of age, mothers completed a depression symptom rating scale, mothers (and fathers) completed a parenting responsibilities questionnaire and mothers and their infants participated in a face-to-face interaction. At 12-months of age, mothers and infants also participated in the Strange Situation. In the methods, we report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. Data and study materials are not publicly available.

Participants

For this report we selected 234 women and their infants from a larger project, Perinatal Stress and Gene Influences: Pathways to Infant Vulnerability, a prospective, longitudinal investigation of women at risk for perinatal depression due to past depression and/or anxiety. Pregnant women were recruited from several sources, including a women’s mental health program in a university psychiatry department, local obstetrical and mental health practitioners, and media announcements. Women were eligible if they were less than 16 weeks gestation based on their last menstrual period (LMP), between the ages of 18 and 45 at enrollment, and had a lifetime psychiatric history of a depressive or anxiety disorder as assessed using the Structured Clinical Interview for DSM-IV Axis 1 Disorders (SCID; First et al., 1995). Women with history of anxiety disorder, along with women with history of depressive disorders, were eligible to participate given how highly comorbid they are with each other and that they are both considered to belong within the broader category of internalizing disorders. Moreover, 46% of individuals with history of major depression also have a lifetime history of anxiety disorders, based on Kessler et al.’s (2015) psychiatric epidemiology work. Thus, expanding the inclusion criteria to those with history of anxiety disorder enhanced the likelihood of identifying women who will develop depression. Women were excluded for active suicidality or homicidality, having psychotic symptoms, meeting DSM-IV criteria for bipolar disorder, schizophrenia, an active eating disorder, an active substance use disorder within 6 months before the last menstrual period, a positive urine drug screen, and/or having an illness requiring treatment that could influence infant outcomes (e.g., epilepsy, asthma, autoimmune disorders, anemia, or abnormal thyroid stimulating hormone concentrations).

At delivery, included mothers (N = 234) ranged in age from 20.7 to 44.5, with a mean age of 33.78 (SD = 4.51). The majority of mothers (n= 207, 88.5%) were married, 7 (3.0%) were divorced, 12 (5.1%) were never married but cohabiting with partner, and 8 (3.4%) were never married and living alone. Mothers’ years spent in education ranged from 10 to 21, with mean years of 16.52 (SD = 2.01). Eighty-nine percent of mothers were White, 9.0% Black, 1.3% Asian, 0.9% Native American, and 0.4% were Mixed. Mean Hollingshead Four-Factor Index scores were 50.75 (SD = 8.85), indicating on average, mothers were broadly middle socioeconomic class.

For this report, we selected the subset of mothers who had participated in infant assessments; we also selected one infant randomly from each of 6 twin pairs, resulting in the final sample (N = 234). Eighty-three mothers (35.5%) in the final sample completed coparenting assessment at one timepoint, and 130 mothers (55.6%) completed both coparenting assessments. Included mothers (N = 234) versus those from the larger study sample who had not participated in any infant assessments (N = 84), respectively, were typically married (88.5% vs. 85.4%), white (89% vs. 85.4%), educated (79.1% vs. 73.1% had completed 16 years of education), of similar socioeconomic status (mean Hollingshead Four-Factor Index scores were 50.75 vs. 48.71, SD = 8.85 vs. 10.44, which reflects broadly middle socioeconomic status; Hollingshead, 1975), and of similar age (33.78 vs. 33.42 years at delivery, SD = 4.51 vs. 4.64). For marital status, race, and education, differences were near zero or small, respectively, φs = .03, .04 and .05; ps = .57, .52 and .40 (thresholds for small, medium, and large correlations are .10, .30, .50 absolute; Cohen, 1988). And for socioeconomic status and age, respectively, standardized mean differences (Cohen’s d) were .22 and .08, ps = .19 and .65 (thresholds for small, medium, and large ds are 0.20, 0.50, and 0.80; Cohen, 1988).

Measures

Structured Clinical Interview for DSM-IV Axis I Disorders (SCID; First et al., 1995).

The SCID is a semi-structured diagnostic interview designed to assess Axis I disorders of the Diagnostic and Statistical Manual- Fourth Edition (DSM-IV; American Psychiatric Association, 1994). We used the SCID to determine the lifetime presence/absence of major depressive episodes. Masters level clinicians administered this SCID, with reliability determined by a senior clinical psychologist who listened to the audiotapes of each interview and, after discussion with the interviewer, independently assigned diagnoses; additional details can be found in Davis et al. (2019). From these interviews, we determined whether the mother met diagnostic criteria or not for major depression in their lifetimes.

Mothers’ depression symptom severity.

We used the Beck Depression Inventory – Second Edition (BDI-II; Beck et al., 1996) to measure mothers’ depressive symptoms at 3-, 6-, and 12-months postpartum. The 21 items each correspond with a particular symptom of depression and are rated on a 4-point scale (0–3); respondents are asked to think about the past two weeks. The score is the sum; scores of 14 or higher indicate at least mild depression. In addition to evidence for reliability and consistency in clinical and nonclinical samples (e.g., Steer et al., 1997), the BDI-II also serves well as a screening test during pregnancy and postpartum (Ji et al., 2011; Steer & Scholl, 1990; Su et al., 2007). Coefficient alpha in the current study were .90, .92, and .88, for 3-, 6-, and 12- months postpartum, respectively. Descriptive statistics for these total scores are in Table 1.

Table 1.

Summary of means, standard deviations, and correlations among study variables

	1	2	3	4	5	6	7	8
1. Depressive symptoms (3m)	1.00
2. Depressive symptoms (6m)	.76***	1.00
3. Depressive symptoms (12m)	.71***	.63***	1.00
4. Satisfaction with parental responsibilities (3m)	.19*	.16*	.19*	1.00
5. Satisfaction with parental responsibilities (12m)	.15†	.21*	.10	.55***	1.00
6. Mothers’ parenting (6m)	.14†	.12	.16†	−.05	−.04	1.00
7. Mothers’ parenting (12m)	.02	−.13	−.10	−.02	−.33***	.35***	1.00
8. Disorganized attachment (12m)	.05	.01	.07	−.01	.01	−.25**	−.21*	1.00
M	8.70	8.68	7.71	12.52	14.25	3.59	3.52	.21a
SD	7.67	8.02	7.57	11.28	11.91	.39	.55	.41

Note. 3m = 3 months. 6m = 6 months. 12m = 12 months. Disorganized attachment was scored dichotomously (0, 1), so the correlations reported are point-biserial. For mothers’ satisfaction with parental responsibilities, higher number indicates less satisfaction.

^aIndicates percent positive for disorganized attachment

^†p < .10.

^*p < .05.

^**p < .01.

^***p < .001.

Measure of satisfaction with division of daily parental responsibilities.

At infant ages 3 and 12 months, mothers completed a measure of perceived and preferred parental responsibility. We combined two published measures with good psychometric properties, the Parental Responsibility Scale (PRS; McBride & Mills, 1993) and the Child Care Activity Questionnaire (CCAQ; Montague & Walker-Andrews, 2002). In addition, we divided one PRS item (“Determine appropriate time and putting child to bed at night”) into two questions (“Determine appropriate time for putting the child to bed at night” and “Put baby to bed at night”), resulting in a list of 24 common tasks involved in caring for infants. Mothers also had the option to state that an item described a task that was not applicable (N/A). For example, “Pick up baby at day care/sitter” would be N/A if the baby does not go to day care or to a childcare provider. No items were found to have high rates of N/A (all were less than 37%) and thus we retained all items and replaced missing data (responses indicated as N/A) with the mean of the participant’s score.

As in the original scales, mothers were instructed to indicate who has primary responsibility for each task on a five-point scale from 1 (mother almost always responsible) to 5 (father almost always responsible), with responsibility defined as remembering, planning, and scheduling the task. Then mothers were asked to indicate how they would like the division of responsibility to be for each task. The range of scores on these two scales was potentially from 24 to 120, with higher scores indicating a higher perceived or preferred father responsibility. For the present study, we relied on a calculated score indexing the difference between the mothers’ perceived and preferred division of responsibility (ideal minus perceived). The range of scores was −7 to +48, with higher positive scores indicating less satisfaction with the degree of father responsibility.

Mothers’ quality of parenting.

We observed mothers interacting with their infant at 6- and 12-months of infant age during a 5-minute unstructured free-play procedure, with the instructions to mothers to ‘please play with your baby in a way that is typical for you.’ Supporting the use of 5-minute observational samples is a large longitudinal study that supported construct validity of a latent parenting construct derived from observational ratings of a 5-minute teaching task (Nordahl et al., 2020) and a meta-analytic review of mothers’ depression in relation to parent-child interaction qualities, which found weak and inconsistent evidence that interaction duration was a moderator (Lovejoy et al., 2000).

We used a well-regarded, reliable and valid rating system of the parent items from the NICHD Study of Early Child Care and Youth Development (SECCYD) to analyze qualities of mothers’ interaction from video-records (see NICHD Early Child Care Research Network, 1999). In addition, we added two items from the Parent-Child Early Relational Assessment (Clark, 1985) to enhance salience of the rating system for comorbid mood problems common in interactions of women at risk of depression. We used 5-point versions of the rating scales, from 1 (not at all characteristic) to 5 (highly characteristic), which were expansions of the original 4-item version (Owen, 2006) and have been widely used (Mills-Koonce et al., 2007; Nordahl et al., 2020).

The score used in the present analyses is the mean of the Positive Engagement scale, factor-analytically derived from analyses of a multi-site dataset (N = 647) that used the same set of rating scales and found the factor to be well-defined (Goodman, Muzik, et al., 2022). The Positive Engagement scale consists of five items: positive regard, sensitivity, cognitive stimulation, detachment (negatively loaded) and depressed/withdrawn/apathetic mood (negatively loaded). We determined inter-observer reliability by having at least 20% of randomly selected video segments at each infant age rated by a second trained observer. Observers had adequate agreement as determined by being within 1 point on each rating scale, i.e., greater than .80 on all items. They had agreement between .90 and 1.00 on most items.

Strange Situation.

The Strange Situation (Ainsworth et al., 1978) is the most widely used procedure for measuring the quality of attachment on infants between one and two years of age. The Strange Situation is a 24-minute video recorded laboratory procedure that involves observing the infant in a comfortable but unfamiliar room, with the mother, with the mother and a stranger, with the stranger, and alone across multiple three-minute episodes. The infant’s attachment security is coded based on his or her behavior during reunions with the mother. We sent video recordings of the Strange Situation to expert coders at the Institute for Child Development at the University of Minnesota, blinded to other information on the mother or infant. Infants were assigned one of the following primary classifications: (A) Insecure-avoidant; (B) Secure; (C) Ambivalent-Resistant; and (D) Disorganized. From these classifications, we created a dichotomous variable that identified infants who were classified as disorganized as their primary attachment category versus infants classified as organized. In this sample, 20.1% of infants (N = 31 out of 154 who had participated in the Strange Situation) were classified as disorganized for their primary attachment category. This percentage is in line with rates of disorganized attachment found in groups of infants with depressed parents (21%) (van IJzendoorn et al., 1999).

Sociodemographic variables.

Mothers reported on their age, race and ethnicity, education, marital status, and living situation.

Analytic plan

We used IBM SPSS Statistics Version 29 to compute descriptive statistics and run bivariate correlations among study variables. Then, to test the indirect effects between maternal coparenting satisfaction and infant disorganized attachment mediated by mothers’ quality of interaction with their infant, we used the lavaan package in R (Rosseel, 2012). Specifically, we conducted mediation analyses using structural equation modeling, with mothers’ coparenting satisfaction (at either 3 or 12 months) as the predictor, mother-infant interaction quality (at either 6 or 12 months) as the mediator, and infant attachment (at 12 months) as the outcome. We followed the recommended guidelines for mediation and tested for indirect associations using the bias-corrected bootstrapped 95% confidence intervals based on 5000 samples (Preacher & Hayes, 2004; Shrout & Bolger, 2002). We reported unstandardized coefficients with standard errors, 95% confidence intervals, and the standardized coefficients.

Percentage of missing data ranged from 12.4% to 34.2% for depression variables, 18.8% to 31.6% for coparenting variables, 34.2% for attachment, and observations of mother-infant interactions had the highest missing rates, ranging between 33.8% to 54.3%. Of the 234 included mother-infant dyads, 155 (66.2%) had codable videotaped free-play interactions at infants’ 6 months, and 107 (45.7%) at infants’ 12 months. No significant differences in depressive symptoms, coparenting satisfaction, and attachment status were found between mothers and infants who participated in videotaped interactions at either 6 or 12 months, and mothers and infants who did not. Using SPSS 29, we ran Little (1988)’s MCAR analyses to test for missing data assumptions. Both the p-value and the relative chi-square value indicated that our data was missing at random (χ²(147) = 153.72, p = .34). Thus, we employed the multiple imputation strategy to account for missing data, to best account for the mix of continuous and categorical data. Previous studies have suggested that for data missing at random, multiple imputation has minimum bias and performs well even when missingness is as large as 50% (Madley-Dowd et al., 2019; Mishra & Khare, 2014).

Results

Preliminary analyses

Table 1 presents descriptive statistics and correlations for all study variables. Depressive symptom means at each age were below the cut score for clinically significant levels of depression, with the standard deviations indicating meaningful variability. A repeated measures ANOVA indicated that mothers’ mean depressive symptoms did not significantly differ across timepoints. Not shown in the table is that 19.5%, 21.5%, and 18.8% of women exceeded the cut score for clinically significant depression at infant ages 3-, 6-, and 12- months, respectively. Additionally, 18.8% of the mothers scored above the cut-off at one measurement period, 7.3% at two measurement periods, and 4.7% at all three measurement periods. Mothers’ satisfaction with the division of parental responsibilities at infant ages 3- and 12- months did not differ significantly, based on a within-subjects t-test. About one-fifth of the infants (20.5%) were classified as having a disorganized attachment style.

Results from the correlation analyses revealed stability to be high among mothers’ depressive symptoms across 3, 6, and 12 months (rs = .63 to .76; ps < .001). Similarly, mothers’ satisfaction with parental responsibilities and mothers’ parenting showed moderate to high stability across timepoints (rs= .55 and .35, respectively, ps < .001). There were significant but weak relations between mothers’ depressive symptoms across all timepoints and satisfaction with parental responsibilities at 3 months (rs = .16 to .19, ps < .05), and between depressive symptoms at 6 months and satisfaction with responsibilities at 12 months (r = .21, p < .05). Additionally, satisfaction with parental responsibilities at 12 months had moderately strong negative relations with the quality of mothering at 12 months (r = −.33, p < .001). Mothers’ quality of parenting at 6 and 12 months were negatively correlated with disorganized attachment (rs = −.21 to −.25, ps < .05).

We also compared the satisfaction of parental responsibilities in mothers living with their partners to mothers who were not. Although a small portion of the included mothers were single or divorced (11.5%), all mothers who completed the Parenting Responsibilities Scale had a partner to share parenting responsibilities with, regardless of their marital status or living situation. At 3 months, 13 (7%) out of 187 mothers who completed the Parenting Responsibilities Scale were not cohabiting with a partner. At 12 months, 11 (7%) out of 154 mothers were not cohabiting with a partner. We tested the differences in satisfaction scores between mothers who were cohabiting with their partner versus not and found significant differences at both 3 and 12 months (ps < .05), such that mothers who did not live with a partner were significantly less satisfied with the division of parenting responsibilities compared to cohabiting mothers at both 3 months (M = 23.85 vs M = 11.68, respectively; lower scores indicate higher satisfaction) and 12 months (M = 26.91 vs M = 13.27, respectively).

Tests of study hypotheses

Results from the main analyses are presented in Table 2 and Figures 1 and 2. We conducted a mediational model using 3 months satisfaction, 6 months parenting, and 12 months attachment (Table 2a and Figure 1). Mothers’ satisfaction with parental responsibilities at infants’ 3 months of age was not significantly associated with the quality of mothering at 6 months (path a). However, path b was significant, such that poorer quality of mothering at 6 months was significantly associated with infant disorganized attachment at 12 months (b = −.71, SE = .20, 95% CI [−1.098, −.324], β = −.30). The direct, indirect, and total effects were not significant, suggesting no linkages between mothers’ satisfaction with parental responsibilities at 3 months and infant disorganized attachment at 12 months. The final model explained 1.0% of the variance in mothers’ quality of parenting at 6 months, and 9.2% of the variance in disorganized attachment at 12 months.

Table 2.

Bootstrapping estimates of mediating effects in the links between mothers’ satisfaction with division of parental responsibilities and disorganized attachment through maternal parenting: (a) longitudinal mediation model using 3 months satisfaction, 6 months parenting, and 12 months disorganization; (b) cross-sectional mediation model using 12 months satisfaction, 12 months parenting, and 12 months disorganization

(a)
	Longitudinal mediation model
Paths	b (SE)	95% CI	β
Parenting ~ Satisfaction (path a)	−.001 (.003)	[−.008, .004]	−.03
Disorganization ~ Parenting (path b)	−.71 (.20)	[−1.098, −.324]	−.30
Disorganization ~ Satisfaction (direct)	.003 (.01)	[−.014, .023]	.02
Path a * path b (indirect)	.001 (.002)	[−.003, .007]	.01
Direct + (path a * path b) (total)	.003 (.01)	[−.015, .024]	.03
(b)
	Cross-sectional mediation model
Paths	b (SE)	95% CI	β
Parenting ~ Satisfaction (path a)	−.02 (.004)	[−.028, −.013]	−.40
Disorganization ~ Parenting (path b)	−.62 (.15)	[−.884, −.311]	−.35
Disorganization ~ Satisfaction (direct)	−.01 (.01)	[−.027, .004]	−.12
Path a * path b (indirect)	.01 (.004)	[.006, .022]	.14
Direct + (path a * path b) (total)	.002 (.01)	[−.016, .017]	.02

Note. Significant paths are bolded.

Figure 1. Longitudinal mediational model examining the indirect effect of mothers’ satisfaction with the division of parental responsibilities at 3 months on infant disorganized attachment at 12 months, through mothers’ quality of parenting at 6 months.

Note. DE = direct effect; IE = indirect effect (in parentheses).

^aStandardized coefficients were reported (bolded if significant).

Figure 2. Cross-sectional mediational model examining the indirect effect of mothers’ satisfaction with the division of parental responsibilities at 12 months on infant disorganized attachment at 12 months, through mothers’ quality of parenting at 12 months.

Note. DE = direct effect; IE = indirect effect (in parentheses).

^aStandardized coefficients were reported (bolded if significant).

Although we did not find longitudinal indirect links between 3 months satisfaction, 6 months parenting, and 12 months disorganized attachment, we hypothesized a significant indirect association using only data collected at 12 months, because of the significant zero-order correlations found between satisfaction with responsibilities, quality of mothers’ parenting, and attachment at 12 months. Thus, we conducted a cross-sectional mediation model using the same variables collected at infants’ 12 months (Table 2b and Figure 2). Path a was significant, such that mothers’ lower satisfaction with division of parental responsibilities at 12 months (larger difference between ideal and perceived division of responsibilities) was associated with poorer quality of mothering at 12 months (b = −.02, SE = .004, 95% CI [−.028, −.013], β = −.40). Lower quality of mothering at 12 months was also associated with disorganized attachment in infants at 12 months (b = −.62, SE = .15, 95% CI [−.884, −.311], β = −.35). There were no direct linkages, but the indirect link was significant, such that the quality of mothers’ parenting significantly mediated the link between maternal satisfaction with parental responsibilities and infant disorganized attachment (b = .01, SE = .004, 95% CI [.006, .022], β = .14). The total effect was not significant. The final model explained 16.3% of the variance in maternal parenting quality at 12 months, and 10.4% of the variance in disorganized attachment at 12 months.

Discussion

The current study of mothers at elevated risk for depression in the postpartum year investigated the indirect relations between mothers’ satisfaction with fathers’ support in daily parenting responsibilities and infant-mother disorganized attachment, examining the quality of mothering as a mediator. First, from our preliminary analyses, we found that mothers cohabiting with a partner are significantly more likely to be satisfied with the division of parenting responsibilities compared to mothers who do not live with their partner. Mothers who do not reside with their partners may be more dissatisfied with the division of parental tasks, as they would naturally have to take on more day-to-day responsibilities around childcare compared to mothers who have more consistent support at home. Moreover, if mothers are separated or divorced from their partner, they may hold resentment or anger from the ended romantic relationship or the separation process, which may spill over into how mothers perceive the current coparental relationship.

In the main analyses, we found that lower satisfaction with the division of parental responsibilities in mothers was cross-sectionally linked to infant disorganized attachment at 12 months through poorer quality of mothering. However, the longitudinal indirect link between satisfaction at 3 months and 12 months disorganization was not supported. Specifically, mothers’ satisfaction with fathers’ support in parental responsibilities at 3 months did not predict her quality of parenting at 6 months, although her parental role satisfaction and quality of parenting were linked at 12 months. This suggests that the degree to which mothers were satisfied with their spouses’ support and involvement in childcare at 3 months did not matter in how she interacted with her child at 6 months. These findings are not consistent with previous findings from a general population sample, that quality of caregiving is longitudinally predicted by coparenting, particularly coparenting in the early months of life (Kim et al., 2021). It may be that in regards to the day-to-day responsibilities of taking care of the child, mothers expect less from fathers in the earlier months of parenthood, compared to by 12 months, when infants’ range of daily activities expand. Indeed, previous research have suggested that both mothers and fathers believe that the role of mothers in childrearing are more important than the role of fathers in the early months of infancy, but both mothers and fathers believed that fathers’ role becomes increasingly more important as infants approach 12 months of age (Favez et al., 2016). In addition, mothers may also be able to put more time and effort into parenting in the earlier months of life, and be less reliant on fathers to take on a share of the childrearing duties, as mothers are typically on leave in the first few months and are able to care for their infants full-time, even in the United States where legally protected parental leave is limited. However, as mothers return to work and are juggling more responsibilities by 12 months, they may be more reliant on the spouses’ support in order to be better parents themselves. Thus, satisfaction with childrearing support in the early months of postpartum may not matter as much for mothers’ behavioral functioning (and in turn, infant attachment) as it does in later months when mothers expect fathers to step up in their roles as coparents.

Another explanation could be that the role of early coparenting relationships may be specific to general population samples and not generalize to samples of mothers at high risk for depression. Indeed, the role of social and family support for depressed parents has been unclear in the literature (e.g., Taraban et al., 2017), with several studies finding marital quality and partner support to have a negative effect on depressed parents (Engle & McElwain, 2013; Taraban et al., 2017, 2019). Perhaps for mothers at risk for depression, spouses’ support in daily childrearing tasks and responsibilities is not as salient as it is for mothers in the general population. It may be that in the earlier months of parenthood, mothers with a history of depression may benefit from more emotional forms of coparenting support, such as encouragement and acknowledgement of mothers’ competence and abilities as parents, verbal affirmations and expressions of support and agreement about mothers’ parenting decisions and authority, and warmth and respect towards mothers’ contributions and commitment to childcare. This begs the need for more research on which specific components of coparenting may be particularly salient for mothers at risk for depression.

In addition, our study assessed mothers’ satisfaction with a component of the coparenting relationship, whereas previous studies examining the coparenting-attachment link examined the quality of coparenting. It is possible that the quality of coparenting in the early postpartum predicts later parenting quality, but parents’ satisfaction with the coparenting quality does not. Future studies would benefit from measuring both quality and satisfaction with coparenting.

The indirect relations between mothers’ satisfaction with fathers’ childrearing support and infant disorganized attachment were significant cross-sectionally at 12 months, but not longitudinally. It is noteworthy, however, that mothers’ quality of interaction with the infant was significantly linked to disorganized attachment in infants both cross-sectionally and longitudinally. Findings are consistent with attachment theory, which suggests that the parents’ quality of caregiving is the “final common pathway” to infants’ quality of attachment (Bowlby, 1969). Moreover, our findings extend previous research that points to the quality of parenting in explaining the high rates of disorganization in infants of depressed parents (e.g., Hayes et al., 2013), emphasizing the importance of parent-infant interaction quality throughout the first year of life for later infant socioemotional outcomes. Nonetheless, mothers’ satisfaction with childrearing responsibilities by 12 months did indirectly influence infants’ disorganization, suggesting that the health of the family system may be increasingly important in organizing the mother-infant relationship as the infant grows older, especially among mothers at risk for depression.

Limitations

Our study is the first to establish indirect links between satisfaction with a coparenting component and disorganized attachment among infants and their mothers at risk for depression. However, we acknowledge several limitations. First, our sample was primarily white, middle-class, and well-educated, which could limit the generalizability of the findings to the broader population. It is possible that for mothers who face systemic racism and discrimination or who have fewer educational or financial resources than others, coparenting satisfaction may be more salient in predicting their behavior, even in the early months of postpartum. Because these mothers face additional risks from having more distress and fewer resources, they may especially benefit from their partner’s support in childcare. More research is needed to examine these associations in minoritized or low resourced families. Second, we were limited to examining mothers’ satisfaction of coparenting and its links with infant-mother disorganized attachment, because we only had data on infants’ attachment to mothers and not fathers. Additionally, our sample consisted of mothers in heterosexual and predominantly cohabiting/married relationships, and as a result, we only examined childrearing efforts between mothers and fathers, although coparenting can occur between any two or more individuals. It is possible that mothers received childrearing support from other family or friends that was unaccounted for in our study. Third, elevated depressive symptoms may be associated with a bias toward perceiving the partner negatively, which could be reflected in lower satisfaction scores. However, correlations between our depression measure and satisfaction scores ranged from .10 to .21 (reported in Table 1); thus, these relatively small correlations suggest that negative cognitive biases were not playing a large role. Moreover, although our sample comprised of women at high risk for depression, mean levels of depressive symptoms for the sample as a whole were still minimal at each timepoint, although nearly 20% of women were in the mild, moderate, or severe depression range across all timepoints. Thus, it is of interest whether our findings may be replicated in a sample of clinically depressed women. Fourth, as previously mentioned, we were only able to find the predicted mediated relations from satisfaction with parental responsibilities to maternal behavior to infant disorganization using cross-sectional measurement at 12 months, which do not meet the temporal requirement for mediation. As we have suggested in the discussion, future studies should examine other components of coparenting support, or the quality of coparenting, to examine whether there are longitudinal indirect links between early coparenting, maternal behavior, and infant disorganization among women at risk for depression. Lastly, our measure of parenting focused solely on maternal behavior, and we were unable to consider infants’ behavior in mothers’ interactions with the infant. Future work would benefit from including child behavior in the prediction of disorganized attachment.

Conclusion and future directions

The present study emphasizes the importance of the family system in the transition to parenthood as it impacts infants and mothers at risk for depression. Specifically, we underscore the crucial role of mothers’ satisfaction with fathers’ support in childrearing across the infants’ first year, as it plays an important role in organizing the quality of the mother-infant relationship, and in turn, the infants’ likelihood of having a disorganized attachment system. As such, interventions targeting maternal postpartum depression should continue to not only aim to reduce mothers’ depression symptoms but also to improve the quality of maternal caregiving throughout the first year of life and help women to achieve greater satisfaction in their relationship with their partner. These outcomes, combined, hold promise to mitigate the risk for disorganized attachment in infants, especially as infants get older, in order to best support a healthy relationship for mothers with depression histories and their infants.