Associations Between Breastfeeding, Maternal Emotional Availability, and Infant–Mother Attachment: The Role of Coparenting

Abstract

Background:

Breastfeeding is a parenting practice that combines close intimate contact with the opportunity to be sensitive and responsive to the infant, and may have direct and indirect relations with infant attachment. However, researchers have produced inconsistent findings, suggesting there may be other mechanisms involved. Coparenting may play a significant role, as it has been consistently associated with mother–infant relationships.

Research aims:

The aims of this study were to examine: (1) whether breastfeeding would be directly associated with infant–mother attachment; (2) whether this association was also indirect, through mothers’ quality of caregiving; and (3) whether partners’ coparenting support moderates breastfeeding’s indirect association with attachment

Methods:

This was a prospective, longitudinal study that drew data from a larger NIH-funded study on sleep and family relationships (R01HD052809). Mothers reported on their feeding practices and coparenting relationships. Independent observations were used to assess mothers’ emotional availability toward infants. A separate team of observers assessed infant–mother attachment.

Results:

Exclusive breastfeeding during the first 6 months, and longer duration of any breastfeeding across the 1st year, were directly associated with more secure infant–mother attachment. These associations were also indirect, through maternal emotional availability. Coparenting was a significant moderator, such that the influence of longer breastfeeding duration on improved emotional availability, and, in turn, on more secure attachment, was significant only for mothers who perceived coparenting quality to be low.

Conclusion:

Findings highlighted the importance of breastfeeding on both the quality of mothering and infant attachment, but also emphasized that coparenting support may be particularly important for mothers who are unable to breastfeed.

Key messages

• Previous mixed findings on the association between breastfeeding and infant attachment have suggested that there may be potential moderating influences, such as coparenting support.

• Breastfeeding was associated with infant–mother attachment, directly and indirectly through mothers’ emotional availability.

• The association between breastfeeding duration and mothers’ emotional availability, and, in turn, on attachment, was significant only for mothers who perceived quality of coparenting to be poor.

• The importance of breastfeeding on mother and infant outcomes was highlighted, but the role of partners’ coparenting support was also emphasized, especially for mothers who were unable or unwilling to breastfeed for any number of reasons.

Background

Breastfeeding is the act of feeding an infant milk from the breast (Labbok & Starling, 2012). The benefits of breastfeeding for both mothers and infants are well-established (World Health Organization, 2022). Breastfeeding’s associations with health and socio-emotional outcomes have been extensively studied and have been found to support a broad range of improved health outcomes in children and mothers (Pattison et al., 2019; Unar-Munguía et al., 2019). As such, the American Academy of Pediatrics (AAP, 2022) recommends exclusive breastfeeding for the first 6 months of infant life, and breastfeeding with the introduction of other foods as mutually desired for 2 years and beyond. In recent years, growing attention has been placed on examining the influence of breastfeeding on attachment (Linde et al., 2020). Based on attachment theory (Bowlby, 1969), breastfeeding may have direct and indirect relations with infants’ quality of attachment, with maternal caregiving as the mediator. However, empirical findings have been mixed, suggesting there may be moderators influencing this association. Coparenting, or the coordination of childrearing between two or more individuals (McHale & Jones, 2021), may be a significant influence, as coparenting support has been associated with increased quality of mother–infant relationship in previous studies ( C. Y.Kim et al., 2021). As such, we explore coparenting quality as a moderator in the associations between breastfeeding, maternal caregiving, and the quality of infant–mother attachment.

A central tenet of infant–caregiver attachment is that the quality of parenting, rather than behaviors or practices of parenting, predicts quality of attachment in children (Solomon & George, 2018). Indeed, in Harlow's (1958) groundbreaking study of rhesus monkeys, attachment was not accounted for by the provision of food but by the contact comfort experienced by the infant in the context of caregiving and feeding. Breastfeeding may represent a unique intersection of both parenting quality and practice, because it is a type of parenting practice that may lead the mother to be in close intimate contact with the infant and be more sensitive and responsive to the infant’s cues (Whitfield & Ventura, 2019). Many interactions between mother and infant in the early months of life take place in the context of feeding. Thus, the quality and method of feeding may play a significant role in both the quality of mother–infant interactions and the development of infant–caregiver attachment. Indeed, the association between breastfeeding and maternal sensitivity has been consistently supported empirically (Feldman et al., 2023; Weaver et al., 2018). Researchers have attributed this association to psychological and physiological explanations, in that breastfeeding may lead to positive mother–infant emotional bonds (Peñacoba & Catala, 2019), increased production of oxytocin and prolactin (Uvnäs-Moberg et al., 2020), lower levels of postpartum maternal stress and depressive symptoms (Xia et al., 2022), and greater maternal brain activation and response to her infant’s cries ( P.Kim et al., 2011).

Breastfeeding requires greater likelihood of skin-to-skin, ventral-to-ventral contact between mother and infant. Skin-to-skin contact theoretically promotes both the mother’s bond with her infant and the infant’s quality of attachment to the mother (Norholt, 2020). Indeed, the activities that often come with skin-to-skin contact, for example, maintaining eye contact and touching, facilitate parents and infants to form intimate emotional bonds. Despite the theoretical direct association between breastfeeding and attachment, however, there have been mixed findings (Linde et al., 2020). Interestingly, significant associations were found in studies that have assessed attachment using observer-rated Attachment Q-Set (AQS; Vaughn et al., 2021) or the Toddler Attachment Sort-45 (Gibbs et al., 2018; Jackson, 2016; Weaver et al., 2018), whereas inconsistent findings were seen in studies that used the Strange Situation (Beijers et al., 2013; Britton et al., 2006; Tharner et al., 2012). Using trained observers to rate the AQS provides an objective and continuous assessment of attachment that is not limited by the categorical scoring of the Strange Situation or the issue of maternal bias in a mother-reported AQS security score. More empirical evidence is still needed to explore this association and the underlying mechanisms.

One such mechanism may relate to the quality of partners’ coparenting support, which has consistently been shown to have a positive influence on the quality of mother–infant relationships ( C. Y.Kim et al., 2021). Indeed, according to family systems theory (Cox & Paley, 1997), family dynamics, for example, the partners’ support in childrearing, can have significant influence over individual development and close relationships. However, it is of interest whether coparenting will interact with breastfeeding to predict mother and infant outcomes. Partner support has been positively associated with higher initiation and continuation of breastfeeding in mothers (Ogbo et al., 2020; Yan et al., 2023), but no researchers have examined whether partners’ coparenting support plays a role beyond the mothers’ decision to breastfeed, on the longitudinal influence of breastfeeding on the mother–infant relationship. From coparenting theory (McHale & Jones, 2021), it is possible that supportive coparenting would promote a higher quality of mothering, and, in turn, more secure attachment, regardless of the presence or amount of breastfeeding.

In summary, the aims of this study were to examine: (1) whether breastfeeding would be directly associated with infant–mother attachment; (2) whether this association was also indirect, through mothers’ quality of caregiving; and (3) whether partners’ coparenting support moderates breastfeeding’s indirect association with attachment (Figure 1)

Figure 1.

Conceptual Model.

Method

Research Design

This was a prospective, longitudinal study that drew data from a larger NIH-funded study on sleep and family relationships (R01HD052809). This study design allowed us to understand the influences of breastfeeding and coparenting on the mother–infant relationship across the first 18 months of life. This study was approved by the Pennsylvania State University’s institutional review board.

Setting and Relevant Context

Mothers were recruited from the obstetric floors of two hospitals in the United States, in Central Pennsylvania: the Mount Nittany Medical Center, and the Milton S. Hershey Medical Center. Obstetric services at these two sites deliver over 2400 infants per year, with about 2000 infants born full-term. Demographic information obtained from the Central Pennsylvania region as a whole indicates a racial/ethnic distribution of approximately 80% White, 7% African American, 8% Latino, and 5% Asian/Pacific Islander/American Indian, and census information obtained from both hospital sites indicates a racial/ethnic distribution that reflects these percentages.

According to data from the US Centers for Disease Control and Prevention (2022), 24.6% of infants in Pennsylvania were exclusively breastfed for 6 months, and 35.2% were still being breastfed at 12 months. There are currently no policies around family leave in Pennsylvania. At both Mount Nittany and Milton S. Hershey Medical Centers, breastfeeding classes and lactation support providers were available. Additional support was available via phone at numerous breastfeeding help lines offered by Pennsylvania.

Sample

The target population for this study was newborns and their mothers who were 24 to 48 hours post-birth. Inclusion criteria were as follows: 18 years or older, fluent in English, and must live with their infant as an independent family unit. Exclusion criteria were: prematurely born infants (< 37 weeks genstational age), infants with low birth weight (< 2500 g), infants with chromosomal abnormalities, infants experiencing perinatal complications that delays hospital discharge for more than 1 week, and mothers with positive perinatal toxicology screens.

We recruited 167 mothers and their infants for the study. To reduce selection bias, we randomly selected and recruited 92 dyads from Mount Nittany Medical Center, and 75 dyads from the Milton S. Hershey Medical Center. Participants were compensated $100 for completing the in-home assessments at each timepoint, with an additional $50 for the in-home attachment assessment at 12 and 18 months. Our final sample included data from 158 dyads, after excluding nine dyads for incomplete breastfeeding data. According to the analysis of variance and chi-square tests, the final sample (N = 158) did not differ from the original (N = 167) on demographic variables (p > 0.05), suggesting that the final sample was an adequate representation of the population. We conducted an a priori power analysis, which indicated that a sample size of 120 provided very high power (95%) with as many as nine predictors, suggesting our final sample size (N = 158) was adequate for providing sufficient power to find associations.

Measurement

Demographic Characteristics

Participants filled out demographic information during the first in-home assessment. Variables used for analysis in the current study included age, parity, marital status, education, employment, race, income, and infant sex assigned at birth.

Breastfeeding

Breastfeeding status by 6 months

At 6 months, participants were asked about their feeding practices for the last 6 months (see the online Supplemental Materials). Using their responses, we created a variable that categorized them into three ordinal groups: (1) mothers who exclusively formula-fed for 6 months (no breastfeeding); (2) mothers who both breastfed and formula-fed for at least some time in the last 6 months (some breastfeeding); and (3) mothers who exclusively breastfed for 6 months.

Duration of any breastfeeding across the year

At 12 months, participants were asked whether they were still breastfeeding, and, if they had stopped, at what infant age (in months) they stopped. These items were recoded into a continuous numerical variable, to indicate the age (in months) when breastfeeding was ceased, with a 0 indicating infants who were exclusively formula-fed from birth, and a 12 indicating infants who were still being breastfed at 12 months. We did not differentiate the participants who were exclusively breastfeeding from those who were both breastfeeding and solids/formula-feeding; thus, both groups were considered as “breastfeeding.”

Coparenting Quality

Participants reported on the couple’s coparenting quality at 1, 3, 6, 9, and 12 months using the 47-item Your Parenting Team Scale (Feinberg et al., 2012; see the online Supplemental Materials). The overall scores were highly stable across timepoints (r = 0.74 to 0.87) and were combined to create a composite coparenting score across the year. Your Parenting Team was internally reliable across all timepoints (alpha values were 0.70 to 0.93).

Maternal Emotional Availability (EA)

We drew from emotional availability (EA) theory (Biringen et al., 2014) to evaluate the emotional quality of mother–infant interactions. EA is a unique parenting construct that derives directly from attachment theory, in that it emphasizes the emotional connection between the parent and the child, both during times of distress and non-distress. At 1, 3, 6, 9, and 12 months, trained coders watched videotaped interactions of daytime free-play (10–15 minutes) and naturalistic bedtime interactions (M = 71.11 minutes) and used the Emotional Availability Scales (EAS; Biringen et al., 2014) to score each participant on four dimensions: sensitivity, non-intrusiveness, non-hostility, and structuring. For both daytime and bedtime EA, the scores from each dimension were standardized into z-scores and aggregated into a composite EA score for each timepoint. Because the composite scores were correlated across the timepoints for both daytime (r = 0.22–0.53) and bedtime (r values = 0.25–0.61), we combined the scores into one mean daytime EA score and one bedtime EA score across the year for every participant. Then, the mean daytime and bedtime EA scores were averaged to create a final combined EA score for each participant (ranged −31.46 to 15.39). We obtained interrater reliability from 42 randomly selected videos across the five timepoints, which equaled to about 35% of the total number of families that had complete video data. Intraclass correlations ranged from 0.89 to 0.96 for daytime EA and 0.80 to 1.00 for bedtime EA.

Infant–Mother Attachment Security

At 12 and 18 months, researchers blind to other study data observed mothers and their infants in their home for about 2 hours and used the Attachment Q-Set (AQS; Waters, 1995) to sort 90 cards of infants’ attachment-exploration characteristics into nine categories. Then, infants’ scores on each item were correlated with a criterion sort, a security score of the hypothetically most secure child sorted by attachment experts. The resulting correlation coefficient for the child served as the security score. Intraclass correlations ranged from 0.87 to 0.96 across both time points. Scores were highly correlated across the two timepoints (r = 0.53) and were averaged into one final score.

Covariates

Socioeconomic risk was calculated based on the participant mothers’ and fathers’ education levels (1 = some high school, 2 = high school graduate; 3 = some college; 4 = Associate degree; 5 = Bachelor’s degree; 6 = Some graduate school; 7 = Master’s degree; 8 = PhD, law, or medical degree) and the family’s average income-to-needs ratio from 1 and 12 months (computed as yearly family income divided by the federal poverty threshold based on family size and year of data collection). Participants reported on the infants’ sleep arrangements at 1, 3, 6, 9, and 12 months. Because sleep arrangements are often fluid during infancy, a new binary variable was created, differentiating the predominantly solitary sleepers from the predominantly cosleepers to examine infant sleep arrangements across the entire 1st year.

Data Collection

Participants were recruited from March 2009 to December 2011. Follow-up data were collected for 18 months until June 2012. Families were visited in their homes at 1, 3, 6, 9, 12, and 18 months, on multiple occasions at each timepoint. During the first visit, participants gave informed consent and filled out demographic information, including reports of infant feeding and coparenting. Participants and their infants were also instructed to play as they normally would for about 10–15 minutes, and this interaction was videotaped. A few days later, researchers visited a second time to set up two to four pieces of video equipment around the home where the infants’ bedtime routines typically took place for recordings of mother–infant bedtime interactions. Participants were instructed to turn on the cameras one hour before the infant's bedtime and turn them off the next morning when the infant awoke. At 12 and 18 months, different researchers visited the family on a separate day to observe and score infant–mother attachment.

Participants’ identity were kept fully confidential and stored in a data file where all information were de-identified. Video recordings were stored in a password protected lab.

Data Analysis

To investigate the complex relations among the variables of interest, a series of analyses were conducted, including mediation and moderated mediation analyses using the lavaan package in R (Version 2023.06.2+561; Rosseel, 2012). Mediation analyses were conducted to examine whether the relations between breastfeeding status and duration and infant–mother attachment were mediated by EA (Aims 1 and 2). We then performed a moderated mediation analysis to investigate the potential moderated mediation associations of the participant’s perception of coparenting on the relation between breastfeeding status and duration and infant–mother attachment, mediated by EA (Aim 3). We used bootstrapped 95% confidence intervals with 1000 samples to test our models. Significance was determined if the confidence interval did not include zero. We centered the predictor, moderator, and mediator variables to mitigate issues related to collinearity. Covariates were included in all models but were removed from a path if there were no significant associations. Multiple imputation was used to handle any missing data. Descriptive statistics and correlations among main study variables were calculated using SPSS (Version 29) to ensure that the assumptions for mediation analysis were met (see Tables 1 and 2).

Table 1.

Demographics of Included Participant Mothers (N = 158).

Characteristic	n (%)
Participant’s Age in Years
18–24	28 (17.7)
25–34	101 (63.9)
35 +	28 (27.7)
Marital Status
Single, no partner	3 (1.9)
Single, living with partner	23 (14.6)
Married or in a relationship, living with partner	128 (81.0)
Married or in a relationship, not living with partner	4 (2.5)
Employment Status
Unemployed	62 (39.2)
Employed	96 (60.8)
Education
Some high school	2 (1.3)
High school graduate	20 (12.7)
Some college	29 (18.4)
Associate degree	14 (8.9)
Bachelor’s degree	36 (22.8)
Some graduate school	10 (6.3)
Master’s degree	35 (22.2)
PhD, law, or medical degree	12 (7.6)
Race
White	133 (84.2)
Black	5 (3.2)
Hispanic/Latina	9 (5.7)
Asian	5 (3.2)
Mixed	5 (3.2)
Yearly Family Income
$0–$30,000	32 (20.3)
$30,001–$60,000	47 (29.7)
$60,001–$90,000	41 (25.9)
$90,001–$140,000	26 (16.5)
$140,001 +	7 (4.4)
Parity
Primiparous	59 (37.3)
Multiparous	99 (62.7)

Note. Missing: Participant’s age in years = 1, race = 1, yearly family income = 5. Demographics were collected at 1 month postpartum.

Table 2.

Correlations Among Study Variables (N = 158).

Variable	Breastfeeding Status (6 m)	Breastfeeding Duration (0-12 m)	Coparenting (1-12 m)	EA (1–12 m)	AQS	Sleep Arrangements	SER
Breastfeeding Status (6 m)	1
Breastfeeding Duration (0–12 m)	0.61 (< 0.001)	1
Coparenting (1–12 m)	−0.08 (0.35)	0.04 (0.66)	1
EA (1–12 m)	0.21 (0.01)	0.25 (0.002)	0.26 (0.002)	1
AQS	0.19 (0.02)	0.33 (< 0.001)	0.21 (0.02)	0.39 (< 0.001)	1
Sleep Arrangements	−0.10 (0.23)	0.02 (0.83)	−0.24 (0.01)	−0.37 (< 0.001)	−0.21 (0.02)	1
SER	−0.11 (0.17)	−0.35 (< 0.001)	−0.23 (0.01)	−0.39 (< 0.001)	−0.41 (< 0.001)	0.19 (0.03)	1
M	2.03	6.89	126.25	−0.23	0.39	0.38 *	0.06
SD	0.58	4.81	22.97	10.69	0.21	0.49	2.33

Note. Correlations are r values (p values in parentheses); 6 m = 6 months, 0–12 m = birth through 12 months; 1–12 m = 1 month through 12 months; AQS = Attachment Q-Set; EA = mothers’ emotional availability; SER = socioeconomic risk.

^*Percentage of infants in predominantly cosleeping arrangements across the 1st year.

Results

Characteristics of the Sample

Sample demographics are presented in Table 1. At 1-month postpartum, participants were between 18 and 43 years (M = 29.32, SD = 5.30). We did not gather information on mothers’ gender identity. Eighty-eight percent of fathers were employed. The median family income was US $60,000 (M = $68,113, SD = $46,115). Fifty-two percent of infants (n = 82) were assigned female at birth. Descriptive statistics and correlations among the main study variables are presented in Table 2.

Mediation Analyses

First, we ran two simple mediation models. In the first model, the direct association from breastfeeding status by 6 months to infant–mother attachment by 12 and 18 months was significant (b = 0.06, SE = 0.03, 95% CI [0.004, 0.115], β = 0.15). Breastfeeding duration was also directly associated with attachment in the second model (b = 0.01, SE = 0.004, 95% CI [0.003, 0.018], β = 0.22).

In these same mediation models, both breastfeeding status and duration were also indirectly associated with attachment, through EA (b = 0.02, SE = 0.009, 95% CI [0.003, 0.039], β = 0.04; b = 0.002, SE = 0.001, 95% CI [0.0004, 0.005], β = 0.05, respectively).

Moderated Mediation Analyses

We then tested separate moderated mediation models for each predictor, with the hypothesized moderated path included (Figure 1). Results from both moderated mediation models are shown in Table 3a for breastfeeding status, and Table 3b for breastfeeding duration. In Table 3b, the interaction between breastfeeding duration and coparenting was significantly associated with EA. Specifically, as presented in Figure 2, the positive association between breastfeeding duration and EA was significant only when mothers’ perceptions of coparenting quality was low (1 standard deviation lower than sample mean). At higher levels of coparenting, breastfeeding duration was not significantly related to EA. As indicated by the significant index of moderated mediation, coparenting significantly moderated the indirect association between breastfeeding duration and attachment through EA. The significant conditional indirect association at 1 standard deviation below the mean of coparenting suggested that this indirect association was only significant at low levels of coparenting quality.

Table 3.

Results From Moderated Mediation Models: (a) Breastfeeding Status by 6 Months as Predictor; and (b) Duration of Any Breastfeeding Across the First Year as Predictor (N = 158).

(a) Breastfeeding Status by 6 months.

Paths	b (SE)	95% CI	β
EA ~ BF Status (X; Path A1)	0.59 (0.23)	[0.121, 1.051]	0.16
EA ~ SER (Control)	−0.63 (0.19)	[−1.060, −0.290]	−0.30
EA ~ Sleep Arrangement (Control)	−1.09 (0.35)	[−1.776, −0.389]	−0.25
EA ~ Coparenting (W; Path A2)	0.32 (0.19)	[−0.035, 0.709]	0.15
EA ~ BF Status*Coparenting (Interaction; Path A3)	−0.47 (0.29)	[−0.983, 0.149]	−0.13
AQS ~ EA (M; Path B)	0.03 (0.01)	[0.007, 0.048]	0.26
AQS ~ SER (Control)	−0.06 (0.02)	[−0.097, −0.022]	−0.26
AQS ~ BF Status (Direct; Path C)	0.06 (0.03)	[0.010, 0.120]	0.17
Conditional Indirect Effects
Low Coparenting (−1 SD)	0.03 (0.02)	[0.005, 0.065]	0.08
High Coparenting (+1 SD)	0.003 (0.01)	[−0.018, 0.026]	0.01
Index of Moderated Mediation	−0.01 (0.01)	[−0.037, 0.002]	−0.03

(b) Duration of Any Breastfeeding Across the 1st Year.

Paths	b (SE)	95% CI	β
EA ~ BF Duration (X; Path A1)	0.08 (0.04)	[0.022, 0.160]	0.19
EA ~ SER (Control)	−0.48 (0.19)	[−0.826, −0.101]	−0.22
EA ~ Sleep Arrangement (Control)	−1.27 (0.32)	[−1.925, −0.681]	−0.29
EA ~ Coparenting (W; Path A2)	0.35 (0.17)	[−0.005, 0.672]	0.16
EA ~ BF dur*Coparenting (Interaction; Path A3)	−0.08 (0.04)	[−0.154, −0.006]	−0.17
AQS ~ EA (M; Path B)	0.02 (0.01)	[0.004, 0.037]	0.21
AQS ~ SER (Control)	−0.05 (0.02)	[−0.088, −0.016]	−0.23
AQS ~ BF Duration (Direct; Path C1)	0.01 (0.003)	[0.002, 0.015]	0.19
Conditional Indirect Effects
Low Coparenting (−1 SD)	0.003 (0.002)	[0.001, 0.008]	0.07
High Coparenting (+1 SD)	0.000 (0.001)	[−0.002, 0.002]	0.003
Index of Moderated Mediation	−0.002 (0.001)	[−0.005, −0.0001]	−0.04

Note. BF = breastfeeding; EA = mothers’ emotional availability across the 1st year; SER = socioeconomic risk; sleep arrangement is a binary variable with 1 = predominantly cosleepers across the year and 0 = predominantly solitary sleepers across the year; AQS = Attachment Q-Set. Path is significant if the confidence interval does not include zero (significant confidence intervals bolded in the table).

Figure 2.

Mothers’ Perceived Coparenting as Moderator Between Breastfeeding Duration and Emotional Availability (Moderation on Path A in Figure 1).

Discussion

In the current study, we examined breastfeeding’s direct and indirect relations with maternal EA and quality of infant–mother attachment, exploring coparenting as a moderator. Our findings of direct, positive associations between breastfeeding status and duration and attachment replicated findings from previous studies that have used the Attachment Q-Set (AQS) to assess infant attachment (Gibbs et al., 2018; Jackson, 2016; Weaver et al., 2018).

In addition, our study was the first to find parenting quality to significantly mediate the association between exclusive breastfeeding and higher quality of attachment. To our knowledge, there are only two other studies that formally explored maternal parenting as a mediator of this association, and neither of the studies found significant mediation (Britton et al., 2006; Tharner et al., 2012). We attribute our significant findings to the use of the naturalistic, observer-rated AQS, and the use of a parenting measure that is rooted in attachment theory to measure the quality of parenting across multiple contexts, one of which we examined in an unstructured bedtime context. We believe that using measures that are naturalistic, theoretically driven, and observed across multiple contexts contributed to finding significant associations between breastfeeding, parenting quality, and attachment. Further, these findings were consistent with previous literature which suggests that breastfeeding offers increased opportunities for physical and emotional connections for mothers and infants, contributing directly to the quality of infant–mother attachment, and that breastfeeding provides psychological and physiological benefits to mothers’ interactions with her child (Feldman et al., 2023; Uvnäs-Moberg et al., 2020).

Interestingly, the influence of breastfeeding duration on mothers’ EA, and, in turn, infants’ quality of attachment, were significant only for those mothers who perceived their coparenting quality to be low, but not for infants and their mothers who perceived coparenting quality to be high. In other words, longer breastfeeding duration had expected positive influence on maternal EA and infant–mother attachment when mothers did not feel supported by their partner in their childrearing effort, but breastfeeding duration did not matter for infants and mothers who experienced high quality coparenting support from their partner. This suggested that although breastfeeding may have a clear positive influence on maternal parenting and infant attachment, the quality of support mothers receive from their partner in regards to childrearing may serve as a protective factor for those mothers who are unable to breastfeed for any number of reasons. This finding is especially important to consider in countries like the United States where there are no parental leave policies to support postpartum parents and many mothers are not able to continue breastfeeding after returning to work. As such, medical professionals and intervention programs should take a family systems perspective when discussing breastfeeding and childcare with new families and make every effort to involve partners in the discussion.

Moreover, although the results emphasize clear benefits of breastfeeding on maternal caregiving and infant attachment, questions remain. For example, we simply asked mothers whether they were “breastfeeding” and did not explore whether this included pumping or donor’s milk. Thus, it was up to each mother’s discretion to define what breastfeeding meant. Differentiating groups of mothers that are exclusively nursing versus mothers who also pump could provide more insight on what it is about breastfeeding that produces higher EA in mothers and more secure attachment in infants. Is it the increased skin-to-skin intimate contact? Is it the psychological influence of the intent to breastfeed? Are there underlying hormonal mediating factors (e.g. oxytocin production)? Future studies should aim to address these mechanisms.

Limitations

We note several limitations in our study. First, despite the broad range of income levels in our sample, our participants were largely homogeneous in terms of race and ethnicity. This could potentially affect the generalizability of our study to the broader population. In addition, our sample consisted only of heterosexual two-parent couples. Thus, we could only examine coparenting between mothers and fathers in the same household, although mothers could have received coparenting support from any other individual who would have been unaccounted for in the present study.

Conclusion

The current study highlights the behavioral and socio-emotional benefits of breastfeeding for both mothers and infants, and supports and extends the AAP’s (2022) guidelines. In addition, our findings indicate that, although exclusive breastfeeding across the first 6 months and continuous breastfeeding across the 1st year confer benefits to parenting and infant socio-emotional development, coparenting support may be particularly important for mothers who may be unable or unwilling to breastfeed.