The Association Between Breastfeeding Duration and Attachment: A Genetically Informed Analysis

Abstract

Introduction:

A growing body of research explores whether breastfeeding during infancy influences the development of attachment security in offspring. Studies to date have generally yielded inconsistent results, with some studies detecting an association between breastfeeding and attachment security, and others failing to do so. The purpose of this study is to empirically consider whether (1) any association between breastfeeding and offspring attachment security is robust to both familial and genetic confounding and (2) whether the breastfeeding-attachment relationship is significant for both male and female offspring.

Materials and Methods:

The present study uses data from the Early Childhood Longitudinal Study: Birth Cohort (ECLS-B) to examine whether twins who differ in their extent of exposure to breastfeeding exhibit different attachment patterns by the time they reach toddlerhood.

Results:

The results suggest that, independent of genetic and shared environmental influences, breastfeeding duration increases the security of attachment in offspring, but only among females.

Conclusions:

A longer duration of breastfeeding may help to facilitate a secure attachment among female offspring.

Introduction

Alarge body of research has suggested that numerous factors may help or hinder a child's ability to form a secure attachment, from physical and sexual abuse¹ to genetics.² One of the factors that scholars have begun to examine as a predictor of attachment security is breastfeeding.^3–5 Although breastfeeding clearly entails the transmission of key nutrients and growth factors to the infant,^6,7 it is nonetheless an inherently relational process that is characterized by skin-on-skin contact, cradling, and (in many cases) eye contact and vocalization,⁸ all of which may serve as key facilitators of attachment security. Attachment security is reflected in the child's ability to use their parent as “a secure base from which to explore the environment.”³ This essential component of socio-emotional development is facilitated by a process, in which the child's safety and survival are maintained through reciprocal parent-infant interactions that are characterized by closeness, sensitivity, and responsiveness.

As might be expected, some studies have detected a link between breastfeeding practices, mother-infant bonding, and attachment security during infancy,^3,5,9 childhood,⁴ and even into adulthood.¹⁰ Even so, research on the topic thus far is somewhat equivocal, with other studies suggesting little to no evidence of a relationship between breastfeeding and attachment.^11–13 Still, in the face of inconsistent results, the rationale behind a breastfeeding-attachment link remains quite sound. As Gribble so eloquently stated, “Breastfeeding may assist attachment development via the provision of regular intimate interaction between mother and child; the calming, relaxing, and analgesic impact of breastfeeding on children; and the stress relieving and maternal sensitivity promoting influence of breastfeeding on mothers.”⁴ Breastfeeding, moreover, has been found to increase the maternal oxytocin response^14,15 and maternal sensitivity to the cues of offspring,^5,16 both of which have been linked to attachment security.^17–20 Still, of the studies conducted to date on the link between breastfeeding and attachment, virtually none (1) account for familial and genetic confounding and (2) examine the breastfeeding-attachment link separately for males and females.

There is growing concern among scholars that, even after including statistical controls for important confounders in the family environment, residual confounding may render the relationship between breastfeeding and various child outcomes spurious, due to omitted variable bias and/or poor measurement.^21–24 Thus, while a number of studies using between-family comparisons support the link between breastfeeding and attachment (i.e., samples with one child per household), research using within-family designs is sorely lacking.²¹ Studies using such designs are essential, as a host of familial factors (e.g., socioeconomic status) and maternal traits (e.g., intelligence, intent to breastfeed) may confound the link between breastfeeding and attachment detected in many prior studies.²⁵

It is also possible, for example, that shared genetic factors between the mother and her child underpin both the mother's propensity to initiate and continue breastfeeding, as well as the child's ability to form a secure attachment, as mothers and their biological children share 50% of their distinguishing DNA. Thus, genetically influenced traits might (1) be shared by mother and child due to shared genetic material, (2) reduce the likelihood of successful prolonged breastfeeding, and (3) decrease the likelihood of a secure attachment. In this example, the mother would transmit both genetic risk and environmental risk to the child, and these risks are likely correlated, making it difficult to tease out the independent effects of genes and the environment on attachment security.²⁶ The present study seeks to address these gaps in the literature by examining whether male and female same-sex twins who differ in their duration of breastfeeding also differ in their attachment patterns during toddlerhood.

Materials and Methods

Data from the Early Childhood Longitudinal Study: Birth Cohort (ECLS-B) were used in the current study. The ECLS-B is a nationally representative study of children born in the United States in 2001. Using a stratified sampling approach, ECLS-B researchers sampled birth certificates registered with the National Center for Health Statistics in the year 2001, which covers ∼99% of U.S. births that occur in a given year. Children were deemed ineligible if (1) they died before the age of 9 months, (2) they were adopted before the age of 9 months, or (3) their mothers were younger than 15 at the time of birth. The study collects data from various sources, including parents, teachers, school administrators, and birth certificates, following subjects through five waves of data collection until their kindergarten school year. Importantly, while over 10,000 children were included in the study at wave 1, a large subsample of over 1,600 twins was also obtained at wave 1 of data collection. To conduct a genetically informative analysis, the final sample is limited to the same-sex twin pair sample of nearly 1,000 twins (n = 976; 238 monozygotic and 738 dizygotic). Ethical approval was granted for the current study from the Institutional Review Board at the University of Texas at San Antonio.

Attachment measures

At the second wave of data collection, when the focal children were ∼2 years of age, the development of attachment relationships between parents and each participating offspring was assessed using the Toddler Attachment Sort-45 Item (TAS-45), which is a modified version of the Attachment Q-Sort (AQS).^27,28 The TAS-45 is an assessment tool designed to detect the extent to which children are able to use their parent (typically the mother) as a secure base, from which they feel free to explore the world, yet confident to return to in times of distress.^29,30 Trained interviewers, assisted by a computer, were charged with the task of sorting 45 cards, first into two piles (i.e., “applies” or “not applies”), but ultimately into four piles (from, “almost always applies” to “rarely or hardly ever applies”). These observations were recorded over the course of ∼2 hours, with sorting lasting roughly 10 minutes. Listed on the cards were child behaviors, and corresponding parent-child interactions, which are reflective of the degree of attachment security (e.g., “Child seeks and enjoys being hugged by mother”). These behaviors were categorized into the following nine dimensions, or “hot spots”: is comfortably cuddly, is cooperative, enjoys company, is independent, seeks attention, is upset by separation, avoids others/not sociable, is demanding/angry, and is moody/unsure/unusual. In general, children who are securely attached to their caregivers engage in more warm, cooperative, and/or sociable behaviors and fewer clingy/anxious, demanding, angry, and/or moody behaviors.^31,32 Importantly, each of the attachment hotspots was also examined as outcome variables in the present study (For a complete list of items that comprise the hot spots, see Appendix A).

In addition, examined as outcome variables were two continuous measures designed to assess the overall level of attachment security as follows: security and dependency. Security taps the ability to use the caregiver as a secure base and a resource to return to in times of distress, whereas dependency suggests clinginess and fearfulness that does not typify secure attachment. These measures are provided in the ECLS-B data and were obtained by calculating correlations between the resulting score on the attachment sort (from all nine dimensions) and a criterion sort representing a hypothetical child with ideal levels of security.²⁷ Ultimately, children who score lower on dependency and higher on the security item are considered to be more securely attached. Alternatively, children who score low on security and high on dependency suggest an ambivalent attachment style, whereas children who score low on both security and dependency suggest an avoidant attachment style. Disorganized attachment is often considered to be a blend of avoidant and ambivalent styles and is most epitomized in the hotspot “Moody/unsure/unusual.” Importantly, a number of prior studies using the ECLS-B have used these same comprehensive measures as indicators of attachment security^31,32 and a comprehensive meta-analysis has buttressed the validity of these measures.³³

Finally, dichotomous measures of the four attachment styles proposed by Bowlby were also included as outcome variables.^29,30 These styles are secure (i.e., explores the environment freely, but returns to parent for comfort when distressed), ambivalent (i.e., extreme distress when separated from parent, inability to be comforted upon reunion), disorganized (i.e., confused/disoriented in the presence of the parent—tries to please yet simultaneously rejects the parent), and avoidant (i.e., avoids others and denies own needs).

Breastfeeding duration

A continuous measure of breastfeeding (in months) was derived from questions asked of the mother during the first and second waves of data collection. These questions included “Did you ever breastfeed the child?,” “Are you still breastfeeding the child?,” and “For how many months did you breastfeed the child?” These same items, or slight variations of these items, were asked again at wave 2, to capture the variation in the duration of breastfeeding for those children who were still breastfeeding at the first wave of data collection, but had stopped by the second wave. Data from waves 1 and 2 were used to construct a variable that measured the duration of breastfeeding (in months) for each child. There was a slight truncation of the measure at 23 months of breastfeeding, as some subjects (<1% of the sample) were still breastfeeding at wave 2, but had not yet aged past 23 months. Thus, these subjects, along with the few subjects who breastfed for 23 months or more, were assigned a value of 23. Otherwise, subjects were assigned a value on this variable according to the number of months of breastfeeding.

DeFries–Fulker (DF) analysis was used in the present study. DF analysis is an effective way to control for residual confounding that can be attributed to shared environmental factors (i.e., household economic disadvantage, maternal education, etc.) and genetic influences, which strengthen causal inferences about the influence of specific nonshared environmental factors that are examined simultaneously (e.g., sibling differences in breastfeeding duration). Genetically informed tests that use twin differences as a way to tap the influence of breastfeeding on attachment are virtually nonexistent. This oversight in the literature is one of the impetuses for this study.

The DF equation used is as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}K_1 = b_0 + b_1 ( K_2 - K_{ \rm m} ) + b_2 [ R * ( K_2 - K_{ \rm m} ) ] + b_3 { \rm ENVDIF} + e ,\end{align*} \end{document}

K₁ in the above equation represents the attachment score (i.e., the outcome variable) for one of the twins being analyzed, K₂ represents their co-twin's attachment score, R is an indicator of the genetic similarity between the kinship pair (1 for MZ twin pairs and 0.5 for DZ twin pairs), and R * K₂ is an interaction term that multiplies the co-twin's attachment score by their degree of genetic similarity with their twin. K_m represents the mean value of K₂ (or, in this study, the mean attachment score of the co-twins). Therefore, the parenthetical statement K₂−K_m signifies that K₂ is mean centered in this equation. Finally, ENVDIF represents the breastfeeding duration difference score between twins from the same twin pair. Approximately, 12% of twins were divergent on the measure of breastfeeding duration, providing sufficient variation for a genetically informed analysis.

Moreover, b₀ represents the constant, b₁ represents the proportion of the variance in attachment that is explained by shared environmental influences, and b₂ is the proportion of the variance in attachment that is explained by genetic influences. Importantly, b₃ in the equation does not represent a latent factor, but instead represents a regression coefficient, and needs to be interpreted as such (e.g., using critical t-values, p-values, etc.). The error term (e) encompasses the effects of the nonshared environment on the attachment outcome and error. To maximize the information available on twin pairs in the ECLS-B, and in line with prior research, twins were double entered and Huber–White standard errors were used to account of the clustering of observations when estimating the statistical significance of the results.^34,35 In some models (e.g., models in Table 2), baseline equations were used, which specify no nonshared environment of interest. These equations are identical, except they do not include the term ENVDIF. Furthermore, in the four attachment style models, a slight variation of DF analysis known as logistic DF analysis was used to account for binary outcomes.³⁶

Results

The results of the descriptive analyses for the final sample are presented in Table 1 (n = 902). Generally speaking, the descriptive statistics reveals a slightly stronger tendency toward secure attachment in females relative to males across attachment hotpots and classifications. To be precise, while 66% of female subjects were classified as securely attached, only 54% of male subjects were classified as securely attached. On average, females were also breastfed slightly longer than males, suggesting, at least at the bivariate level, that individuals who breastfeed for longer durations may develop a more secure attachment to their caregiver. Of course, the analyses depicted in Table 1 are merely descriptive and not explanatory, nor do they control for the influence of shared environmental or genetic factors.

Table 1.

Descriptive Statistics

	Males			Females
Variable	Mean	Standard deviation	Range	Mean	Standard deviation	Range
Attachment factor scores
Security	0.36	0.39	−0.73 to 0.98	0.42	0.34	−0.86 to 0.98
Dependency	−0.39	0.37	−0.98 to 0.86	−0.03	0.34	−0.78 to 0.83
Attachment hotspots
Comfortably cuddly	0.28	0.30	−0.61 to 0.82	0.36	0.28	−0.77 to 0.77
Cooperative	0.29	0.38	−0.68 to 0.87	0.40	0.32	−0.55 to 0.89
Enjoys company	0.10	0.41	−0.82 to 0.83	0.13	0.40	−0.76 to 0.81
Independent	0.11	0.33	−0.69 to 0.79	0.04	0.32	−0.72 to 0.72
Seeks attention	−0.04	0.27	−0.63 to 0.70	−0.04	0.26	−0.61 to 0.65
Upset by separation	−0.07	0.28	−0.63 to 0.70	−0.09	0.27	−0.67 to 0.77
Avoids others/not sociable	0.02	0.29	−0.65 to 0.71	−0.03	0.27	−0.60 to 0.73
Demanding/angry	−0.07	0.29	−0.57 to 0.64	−0.11	0.28	−0.64 to 0.73
Moody/unusual	−0.51	0.56	−1.60 to 1.01	−0.54	0.51	−1.45 to 1.22
Attachment styles
Secure	0.54	0.5	0 to 1	0.66	0.47	0 to 1
Ambivalent	0.17	0.38	0 to 1	0.13	0.34	0 to 1
Disorganized	0.17	0.37	0 to 1	0.10	0.31	0 to 1
Avoidant	0.12	0.32	0 to 1	0.10	0.30	0 to 1
Breastfeeding variables
Breastfeeding (in months)	2.77	4.56	0 to 23	3.21	4.61	0 to 23
Breastfeeding (difference score)	0	1.08	−12 to 12	0	1.64	−20 to 20

Table 2 displays the results of a set of DF analyses examining the influence of the shared environment and heritability on attachment factor scores and hotspots (*p < 0.05; **p < 0.01). This step in the analyses is an important one, as it reveals whether attachment factor scores and hotspots are influenced by genetic and shared environmental factors and, thus, whether a genetically informed analysis of the link between breastfeeding and attachment is truly warranted. As expected, the results do indeed suggest that both the shared environment and genetic factors influence attachment patterns, but that the degree of influence is contingent on the exact dimension of attachment being examined. The results broadly indicate that a significant portion of the variance in hotspots associated with greater attachment security (e.g., comfortably cuddly, cooperative, enjoys company, etc.) is explained by the shared environment, whereas a significant portion of the variance in hotspots associated with lower attachment security (e.g., upset by separation, demanding/angry, etc.) is explained by genetic factors. Thus, both genetics and the shared environment appear to play a role in the different facets of attachment security that influence attachment style. Ultimately, a significant portion of dependency (clinginess/fearfulness) was explained by genetic factors (i.e., 69%), whereas a significant portion of security was explained by shared environmental factors (i.e., 52%). In either case, however, a sizeable portion of the variance in attachment still remains to be explained by nonshared environmental factors, and one such factor may be differences in the duration of breastfeeding.

Table 2.

DeFries–Fulker Analysis of the Shared Environment and Heritability as Predictors of Attachment Factor Scores and Hotspots

	Heritability (h²)	Shared environment (c²)
Attachment factor scores
Security	0.19 (0.13)	0.52^** (0.09)
Dependency	0.69^** (0.14)	0.00 (0.11)
Attachment hotpots
Comfortably cuddly	0.16 (0.15)	0.48^** (0.10)
Cooperative	0.23 (0.13)	0.48^** (0.09)
Enjoys company	0.47^** (0.11)	0.20^* (0.08)
Independent	0.56^** (0.15)	0.03 (0.11)
Seeks attention	0.58^** (0.13)	0.11 (0.10)
Upset by separation	0.70^** (0.14)	0.00 (0.11)
Avoids others/not sociable	0.05 (0.12)	0.58^** (0.08)
Demanding/angry	0.45^** (0.13)	0.22^** (0.09)
Moody/unusual	0.64^** (0.08)	0.04 (0.12)
N	902

p < 0.05, two-tailed; ^**p < 0.01, two-tailed.

Table 3 explores the key question of this study: Do twin differences in breastfeeding duration predict differences in attachment patterns and styles? The results of the DF analyses indicate that, for the full sample, longer durations of breastfeeding correspond to greater attachment security and lower dependency scores, even after accounting for shared environmental and genetic influences. Thus, relative to their co-twins, twins who breastfeed for longer durations are (1) significantly more likely to exhibit secure attachment patterns and (2) significantly less likely to exhibit dependency. Longer durations of breastfeeding were also associated with higher scores on hotspots indicative of attachment security and lower scores on hotspots associated with ambivalence (i.e., attention seeking, upset by separation, and demanding/angry). Interestingly, breastfeeding did not appear to influence avoidance (lack of sociability) or unusual moodiness, which suggests it may be less relevant (or irrelevant) to avoidant and disorganized attachment styles, compared to its more clear relevance for secure and ambivalent styles in the full sample.

Table 3.

In Twins, Do Differences in Breastfeeding Duration Predict Differences in Attachment Factor Scores, Hotspots, and/or Styles?

	Breastfeeding duration (in months)
Attachment factor scores	Full sample (b/SE)	Males (b/SE)	Females (b/SE)
Security	0.03^** (0.01)	−0.05^** (0.02)	0.04^** (0.01)
Dependency	−0.02^** (0.01)	0.00 (0.02)	−0.03^** (0.01)
Attachment hotpots
Comfortably cuddly	0.02^* (0.01)	−0.02 (0.02)	0.02^** (0.01)
Cooperative	0.03^** (0.01)	−0.01 (0.01)	0.03^** (0.00)
Enjoys company	0.02^** (0.01)	−0.03 (0.02)	0.03^** (0.01)
Independent	0.02^** (0.01)	−0.01 (0.02)	0.02^** (0.01)
Seeks attention	−0.02^** (0.01)	−0.01 (0.02)	−0.02^** (0.00)
Upset by separation	−0.02^* (0.01)	−0.01 (0.01)	−0.02^* (0.01)
Avoids others/not sociable	0.01 (0.00)	−0.02 (0.01)	0.01^* (0.00)
Demanding/angry	−0.03^** (0.00)	−0.01 (0.01)	−0.03^** (0.00)
Moody/unusual	−0.03 (0.02)	0.11^** (0.02)	−0.05^** (0.02)
Attachment styles	Full sample (OR/SE)	Males (OR/SE)	Females (OR/SE)
Secure	1.12^** (0.05)	0.85 (0.12)	1.15^** (0.06)
Ambivalent	0.85^** (0.04)	0.93 (0.15)	0.85^** (0.04)
Disorganized	0.92 (0.05)	1.04 (0.17)	0.89 (0.06)
Avoidant	1.10 (0.06)	1.37^* (0.22)	1.07 (0.06)
n	902	452	450

p < 0.05, two-tailed; ^**p < 0.01, two-tailed.

OR, odds ratio; SE, standard error.

Separate analyses for males and females were then conducted to explore whether the link between breastfeeding duration and attachment security persists for offspring of both sexes. The findings suggest that the results supporting a positive association between breastfeeding and secure attachment are driven by the female twins in the sample. Specifically, while a longer duration of breastfeeding in males failed to increase the odds of a secure attachment classification, it significantly increased the odds of a secure attachment classification in females. For example, each additional month of breastfeeding among female infants (1) increased the odds of obtaining a secure attachment classification by 15% and (2) decreased the odds of obtaining an ambivalent attachment classification by 15%. No such benefits were derived from breastfeeding for male offspring once genetic and the shared environmental factors were taken into account. Moreover, every attachment hotspot was associated with breastfeeding in females (largely in the expected direction), whereas only one attachment hotspot was associated with breastfeeding duration among male offspring (i.e., moody/unusual), and it was not in the expected direction. Breastfeeding actually appeared to correspond to lower security scores in males, as well as a higher risk of an avoidant attachment classification, despite its clear association with more favorable attachment patterns in females.

Discussion

The present study sought to explore whether (1) breastfeeding influences attachment patterns in offspring, independent of genetic and shared environmental factors and (2) whether these effects exists for both male and female offspring. The results revealed that, in the full sample, longer durations of breastfeeding are associated with a greater likelihood of a secure attachment style and a lower likelihood of an ambivalent attachment style, independent of shared environmental and genetic factors. Put differently, even when twins living in the same household were examined, the twin who breastfed for longer had a significantly greater likelihood of developing a secure attachment style and a lower likelihood of developing an ambivalent attachment style, than their co-twin who breastfed for a shorter duration. An important caveat to this finding, however, is that this effect appears to be restricted to the females in the sample.

Interestingly, the findings of the present study are not the first to suggest that the benefits of breastfeeding for child developmental outcomes may be more pronounced and/or limited to female offspring. Prior research has revealed, for instance, that the influence of breastfeeding on offspring verbal ability during childhood is stronger for females than for males.³⁷ More recent research has also found that breastfeeding experiences predict better psychological well-being among females during adulthood³⁸ and improve cognition and memory in females through adolescence and adulthood.³⁹ In these studies, however, such benefits were not detected for male offspring. Although prior research examining breastfeeding-gender interactions did not examine attachment specifically, they nonetheless suggest that there may be some differences in male and female sensitivity to breastfeeding. The specific mechanisms underlying enhanced female sensitivity to breastfeeding in these studies were not identified by the authors. Similarly, the reason behind the divergent results across the genders also remains unknown in the present study, which is the first to explore the link between breastfeeding and attachment separately for male and female offspring. It is possible that breastfeeding generally failed to contribute to male attachment patterns in the present analysis due to the genetically informed nature of the study. To be precise, prior research has revealed that male attachment patterns are more heritable (i.e., genetically influenced) than female attachment patterns.⁴⁰ If the same genetic processes that influence attachment patterns in male offspring also influence breastfeeding duration among male offspring, then a longer duration of breastfeeding may not independently predict attachment patterns in males in a genetically informed analysis (despite predicting a secure attachment in females). Future studies may benefit from exploring the biological and/or social underpinnings that may elucidate the protective effect of breastfeeding in the case of female attachment. More research examining whether biological sex of offspring may interact with breastfeeding to influence developmental outcomes is also needed, as prior null findings of within-family studies may be a function of distinct effects across male and female subjects.

Despite the many strengths of the current study, including its genetically informative approach, it has a number of limitations. First, the data provide no means of determining whether (and how frequently) infants were bottle-fed with breastmilk v. fed directly from the breast, nor do they provide the motivations for divergent feeding practices across a sizeable portion of twin pairs. These details would have been preferable to tease apart the effect of breastfeeding as a unique form of nutrition, as well as a bonding/social experience. Furthermore, the effect of breastfeeding exclusivity could not be determined due to insufficient variation in exclusivity across twins within the same twin pair. Second, the data do not include measures of attachment at the earliest stages of infancy. It is likely that the association between breastfeeding and attachment in the current study reflects a bidirectional relationship, wherein breastfeeding and attachment mutually reinforce one another. Even so, ancillary analyses reveal that the supportive results in the female sample are robust to temperamental, health, and affective differences between twins during infancy, despite revealing a significant association between infant temperament and breastfeeding duration. Consequently, the findings suggest that breastfeeding independently influences attachment security among females during toddlerhood. Finally, the extent to which the differences in attachment patterns between female twins who differed in their duration of breastfeeding persisted past toddlerhood and influenced other developmental and psychosocial outcomes could not be ascertained.

Conclusion

Ultimately, the findings of the current study suggest that breastfeeding can facilitate a secure attachment, at least in the case of female toddlers. Although the current study contradicts a recent within-family examination of the breastfeeding-attachment relationship,²¹ it also extends it by (1) specifically modeling and accounting for the proportion of the variance in attachment due to genetic factors (in addition to shared environmental factors) and (2) examining this relationship separately for males and females. Additional genetically informed research is required to effectively tease apart the intricate overlap between breastfeeding and infant attachment and, hopefully, inform policies aimed at supporting mothers' efforts to breastfeed their offspring.

Footnotes

Disclosure Statement

No competing financial interests exist.

Appendix A: Items for the Low Attachment Security Measure

Hot Spot 1: Is Comfortably Cuddly (Wave 2)

Hot Spot 2: Is Cooperative (Wave 2)

Hot Spot 3: Enjoys Company (Wave 2)

Hot Spot 4: Is Independent (Wave 2)

Hot Spot 5: Seeks Attention (Wave 2)

Hot Spot 6: Is Upset by Separation (Wave 2)

Hot Spot 7: Avoids Others/Not Sociable (Wave 2)

Hot Spot 8: Is Demanding/Angry (Wave 2)

Hot Spot 9: Is Moody/Unsure/“Unusual” (Wave 2)

References

Shapiro

, Alytia

. Adolescent survivors of childhood sexual abuse: The mediating role of attachment style and coping in psychological and interpersonal functioning. Child Abuse Neglect, 1999; 23:1175–1191.

Raby

, Cicchettii

, Carlson

, et al. Genetic and caregiving-based contributions to infant attachment unique associations with distress reactivity and attachment security. Psychol Sci, 2012; 23:1016–1023.

Britton

, Britton

, Gronwaldt

. Breastfeeding, sensitivity, and attachment. Pediatrics, 2006; 118:e1436–e1443.

Gribble

. Mental health, attachment and breastfeeding: Implications for adopted children and their mothers. Int Breastfeed J, 2006; 1:1.

Tharner

, Luijk

, Raat

, et al. Breastfeeding and its relation to maternal sensitivity and infant attachment. J Dev Behav Pediatr, 2012; 33:396–404.

Ballard

, Ardythe

. Human milk composition: Nutrients and bioactive factors. Pediatr Clin N Am, 2013; 60:49–74.

Lönnerdal

. Bioactive proteins in breast milk. J Paediatr Child H, 2013; 49:1–7.

Liu

, Leung

, Yang

. Breastfeeding and active bonding protects against children's internalizing behavior problems. Nutrients, 2013; 6:76–89.

Lai

, Hung

, Stocker

, et al. Postpartum fatigue, baby-care activities, and maternal–infant attachment of vaginal and cesarean births following rooming-in. Appl Nurs Res, 2015; 28:116–120.

10.

Schwarze

, Hellhammer

, Stroehle

, et al. Lack of breastfeeding: A potential risk factor in the multifactorial genesis of borderline personality disorder and impaired maternal bonding. J Pers Disord, 2015; 29:610–626.

11.

Else-Quest

, Janet

, Clark

. Breastfeeding, bonding, and the mother-infant relationship. Merrill Palmer Q, 2003; 4:495–517.

12.

Jansen

, de Weerth

, Riksen-Walraven

. Breastfeeding and the mother–infant relationship—A review. Dev Rev, 2008; 28:503–521.

13.

Schulze

, Carlisle

. What research does and doesn't say about breastfeeding: A critical review. Early Child Dev Care, 2010; 180:703–718.

14.

Feldman

, Gordon

, Zagoory-Sharon

. Maternal and paternal plasma, salivary, and urinary oxytocin and parent–infant synchrony: Considering stress and affiliation components of human bonding. Dev Sci, 2011; 14:752–761.

15.

Strathearn

, Lyengar

, Fonagy

, et al. Maternal oxytocin response during mother–infant interaction: Associations with adult temperament. Horm Behav, 2012; 61:429–435.

16.

Kim

, Feldman

, Mayes

, et al. Breastfeeding, brain activation to own infant cry, and maternal sensitivity. J Child Psychol Psychiatry, 2011; 52:907–915.

17.

Leerkes

. Maternal sensitivity during distressing tasks: A unique predictor of attachment security. Infant Behav Dev, 2011; 34:443–446.

18.

Nelson

, Panksepp

. Brain substrates of infant–mother attachment: Contributions of opioids, oxytocin, and norepinephrine. Neurosci Biobehav R, 1998; 22:437–452.

19.

Pederson

, Moran

, Stiko

, et al. Maternal sensitivity and the security of infant-mother attachment: A Q-sort study. Child Dev, 1990; 61:1974–1983.

20.

Strathearn

, Fonagy

, Amico

, et al. Adult attachment predicts maternal brain and oxytocin response to infant cues. Neuropsychopharmacology, 2009; 34:2655–2666.

21.

Colen

, Ramey

. Is breast truly best? Estimating the effects of breastfeeding on long-term child health and wellbeing in the United States using sibling comparisons. Soc Sci Med, 2014; 109:55–65.

22.

Evenhouse

, Reilly

. Improved estimates of the benefits of breastfeeding using sibling comparisons to reduce selection bias. Health Serv Res, 2005; 40:1781–1802.

23.

Kramer

, Fombonne

, Igumnov

, et al. Effects of prolonged and exclusive breastfeeding on child behavior and maternal adjustment: Evidence from a large, randomized trial. Pediatrics, 2008; 121:e435–e440.

24.

Kramer

, Fombonne

, Matush

, et al. Long-term behavioural consequences of infant feeding: The limits of observational studies. Paediatr Perinat Epidemiol, 2011; 25:500–506.

25.

Mathews

, Leerkes

, Lovelady

, et al. Psychosocial predictors of primiparous breastfeeding initiation and duration. J Hum Lact, 2014; 30:480–487.

26.

Shelton

, Collishaw

, Rice

, et al. Using a genetically informative design to examine the relationship between breastfeeding and childhood conduct problems. Eur Child Adolesc Psychiatry, 2011; 20:571–579.

27.

Andreassen

, Fletcher

. Early Childhood Longitudinal Study, Birth Cohort (ECLS-B): Psychometric report for the 2-year data collection. Methodology report. National Center for Education Statistics, 2007, NCES 2007–2084.

28.

Waters

, Deane

. Defining and assessing individual differences in attachment relationships: Q-methodology and the organization of behavior in infancy and early childhood. Monogr Soc Res Child, 1985; 50:41–65.

29.

Bowlby

. Attachment and Loss: Attachment. V. Basic Books, New York, NY, 1969.

30.

Bowlby

. Attachment and loss: Retrospect and prospect. Am J Orthopsychiatry, 1982; 52:664–678.

31.

Bronte-Tinkew

, Scott

, Horowitz

. Male pregnancy intendedness and children's mental proficiency and attachment security during toddlerhood. J Marriage Fam, 2009; 71:1001–1025.

32.

Roisman

, Fraley

. A behavior-genetic study of parenting quality, infant attachment security, and their covariation in a nationally representative sample. Dev Psychol, 2008; 44:831–839.

33.

Van IJzendoorn

, Vereijken

, Bakermans-Kranenburg

, et al. Assessing attachment security with the attachment Q sort: Meta-analytic evidence for the validity of the observer AQS. Child Dev, 2004; 75:1188–1213.

34.

Kohler

, Rodgers

. DF-analyses of heritability with double-entry twin data: Asymptotic standard errors and efficient estimation. Behav Genet, 2001; 31:179–191.

35.

Rodgers

, Buster

, Rowe

. Genetic and environmental influences on delinquency: DF analysis of NLSY kinship data. J Quant Criminol, 2001; 17:145–168.

36.

Schwartz

, Beaver

. Revisiting the association between television viewing in adolescence and contact with the criminal justice system in adulthood. J Interpers Violence, 2015 [Epub ahead of print]; DOI:0886260515576970.

37.

Quinn

, O'Callaghan

, Williams

, et al. The effect of breastfeeding on child development at 5 years: A cohort study. J Paediatr Child H, 2001; 37:465–469.

38.

Cable

, Bartley

, McMunn

, et al. Gender differences in the effect of breastfeeding on adult psychological well-being. Eur J Public Health, 2012; 22:653–658.

39.

Foroushani

, Mohammad

, Mahmoodi

, et al. Effect of breastfeeding on cognitive performance in a British birth cohort/Effets de l'allaitement sur les performances cognitives dans une cohorte de naissance britannique. East Mediterr Health J, 2010; 16:202–209.

40.

Minnis

, Reekie

, Young

, et al. Genetic, environmental and gender influences on attachment disorder behaviours. Br J Psychiatry, 2007; 190:490–495.