Defense Mechanisms of Pregnant Mothers Predict Attachment Security, Social-Emotional Competence, and Behavior Problems in Their Toddlers

Pregnancy can be a significant stressor for mothers (1), especially for low-income women (2). The ability to adaptively cope with stress before, during, and after pregnancy—both consciously and unconsciously—may affect a mother’s ability to optimally support her child’s psychological development. Coping mechanisms fall into three broad categories: seeking social support, cognitive strategies, and involuntary defense mechanisms (3). This study focused on defense mechanisms—automatic psychological processes that mediate reactions to emotional conflicts and internal or external stressors. Special attention was given to the adaptive defenses (e.g., altruism) because they promote optimal adaptation by keeping individuals aware of their thoughts and feelings and the consequences of their actions. This differs from pathological defenses, in which both thought and feeling are ignored (4). Thus, healthy defenses should allow mothers to cope with stressors and remain attuned to their child’s thoughts, feelings, behaviors, and needs.

Infants are born into a complex social world with an innate capacity for social relatedness. A major developmental task in the first year of life is to develop at least one attachment relationship, which is characterized by a preference for a specific caregiver (or more than one), who is expected to care for the infant and with whom the infant shares a close, emotional bond (5, 6). According to attachment theory, the capacity for social relatedness ensures the child’s survival. Attachment behaviors, such as crying and smiling, are present at birth to facilitate this process; these behaviors increase the likelihood that a caregiver will respond to, care for, and protect the child. Caregivers also assist considerably with the infant’s emotional and behavioral regulation through their responses to infant cues (7, 8). By the end of the first year, infants’ cognitions about the self, others, and the world (i.e., internal working models) are also beginning to develop as the result of experiences with the people most important to them (5), and these cognitions subsequently guide the child’s behaviors and future interactions. Thus, the infant’s emotional, behavioral, and social capacities are organized and firmly rooted in the relational context with primary attachment figures during the first several years of life (6, 9).

When the attachment relationship goes well, caregivers consistently respond to the child’s needs in an accurate and attuned way, and the parent-child relationship is said to be “secure” because the child can count on and expect the attachment figure to provide appropriate care in times of need (5). Secure attachment relationships are associated with better social-emotional and behavioral functioning for the child both concurrently and in their future (10). In contrast, when parents are not able to provide consistent, sensitive care (for a variety of reasons, such as their own mental health difficulties and problematic defensive processes), the parent-child relationship suffers and is considered “insecure” because the young child is forced to develop a less adaptive relationship strategy in order to maintain proximity with the parent. Attachment insecurity is in turn associated with less adaptive developmental outcomes and greater social-emotional problems for the child (9).

We hypothesized that mothers who frequently utilize mature defenses are more attuned to their child’s needs and respond more appropriately to meet those needs, helping the child to develop adequate behavioral and emotional regulatory capacities early in life. Mature defense mechanisms may be especially important during times of stress and major relational changes, as is the case during the prenatal and postpartum periods. Therefore, a greater capacity to cope with the physical and emotional demands of preparing for and parenting a new infant, particularly in the context of other psychosocial adversity, is likely to facilitate a healthy parent-child relationship and optimal child development. Thus, we hypothesized that the maturity of defense mechanisms in at-risk pregnant mothers would predict their toddlers’ attachment security, social-emotional competence, and behavior problems.

Method

Participants

Participants were 84 pregnant women recruited for a larger longitudinal study (11) of the role of psychosocial risk in women and their young children. Data were collected during the third trimester of pregnancy, at 3 months postpartum, and when children were 1 and 2 years old. For the present study, data from the prenatal and 2-year waves were chosen, since we believed that the strengths and challenges of the children would be more clearly observable at age 2, especially social-emotional competence and potential behavior problems. Participants were recruited through fliers posted at public locations and agencies serving mostly low-income families.

Procedures

Pregnant women interested in participating in the study were screened by telephone. Inclusion criteria were age ≥18 years and English fluency. The majority of assessments were conducted in the family home by two or three research assistants who had received intensive training from the second author. At each wave of data collection, self-report questionnaires were read aloud to enhance participant understanding of each question and limit random responding. Each protocol was recorded in order for the research assistant to assess for participant literacy difficulties. Financial compensation was provided in the form of gift cards.

Interviews were conducted in the third trimester and lasted approximately 2.5–3 hours. The assessments, which were administered in a standardized order, included a demographic measure, a semistructured clinical interview assessing mothers’ representations of their infant in utero, and self-report measures. Mothers were contacted approximately 2 weeks after the baby’s due date to gather birth information, and a brief telephone interview was administered when infants were approximately 3 months old. Postnatal assessments were conducted around the child’s first and second birthdays, each lasting approximately 3–3.5 hours. Demographic and self-report data were collected, and observations of parent-child interactions were made.

Because of factors associated with poverty status and other psychosocial risks, participants were relatively transient with their living arrangements. To reduce attrition, participants were contacted every 3 months to update contact information as well as that of friends and relatives who would know their location if they could not be reached. Study staff conducted home visits in person to update information if participants could not be reached by telephone or mail. At the 2-year point, 84 of 120 mother-child pairs had completed all study measures.

The study was approved by the institutional review board of Eastern Michigan University. Informed consent was obtained from each participant at every wave of data collection.

Measures

Maternal defense mechanisms.

The DSM-IV Defensive Functioning Scale was developed as a proposed axis for DSM to provide researchers and clinicians with a standard list of defense mechanisms and definitions for diagnostic and research purposes (12, 13). The scale includes 31 defenses organized into seven hierarchical levels of functioning: high adaptive, mental inhibitions, minor image-distorting, disavowal, major image-distorting, action, and defensive dysregulation. Defenses are rated each time they emerge during a clinical interview. Total scores for each level of defense and an overall defensive functioning score (a measure of maturity of defense) were calculated. The overall defensive functioning score is calculated by multiplying the score for each level (ranging from 7, “high adaptive,” to 1, “defensive dysregulation”) by the total number of defenses at each level and dividing that score by the total number of defenses. Defensive dysregulation (i.e., psychotic) defenses were not observed in this study. Interrater agreement of the scales has ranged from fair to excellent, and their validity has been supported through studies of defenses and symptoms or disorders, personality disorder, social or occupational functioning, immature defenses, childhood or adult abuse, relatedness, improvement in depressive episodes, and therapists’ use of psychodynamic-interpersonal interventions (14–23). Interviews of mothers were coded by a psychologist-psychoanalyst (J.H.P.) with 24 years of experience rating defense mechanisms. Interrater agreement for the defenses was calculated from 25 randomly selected interviews rated by a doctoral candidate in clinical psychology (L.R.) who underwent extensive training. Raters were blind to all participant information. Interrater agreement (intraclass correlation) (24) for this study was as follows: overall defensive functioning, 0.81 (excellent); high adaptive, 0.76 (excellent); mental inhibition, 0.66 (good); minor image-distorting, 0.70 (good); disavowal, 0.79 (excellent); major image-distorting, 0.59 (fair); action level, 0.49 (fair).

The Defensive Functioning Scale was used to code maternal narratives resulting from the Working Model of the Child Interview (25), which was administered during the pregnancy interview. The Working Model of the Child Interview is a 1-hour (audio-recorded) semistructured interview that assesses a caregiver’s representations of the infant, the self as a caregiver, and the mother-infant relationship. The original coding system (26) includes qualitative scales (e.g., acceptance, coherence, and involvement) and categorical classifications (balanced, disengaged, and distorted); the present study utilized the narratives to evaluate mothers’ level of defenses.

Child attachment security.

The Attachment Q-Set (27) was used to measure toddler attachment security at the 2-year visit. It includes 90 items of infant behavior indicative of attachment security and insecurity, along a continuum; these items are later sorted by the researchers into nine equal piles of 10 statements ranging from “least characteristic of the child” to “most characteristic of the child” and are based on naturalistic observations of infant behavior toward the caregiver/mother during the 3-hour home visit. Research assistants were extensively trained for 6 months prior to attending home visits and sorting. Training included extensive readings and discussions about attachment theory, infant attachment behaviors, and attachment assessments. Two to three research assistants from the team (advanced undergraduates and clinical psychology graduate students) attended every home visit. After each visit, all research assistants who attended the home visit completed the Q-sort together, using consensus methodology, usually within 24–48 hours. Because home observations were made while research assistants were attending to all that was occurring in the home (e.g., the lead interviewer was focused on appropriate administration of measures, others were focused on monitoring the baby’s activities, etc.), the decision was made early on to have all assistants complete the Q-sort together, rather than independently, so that all observations were taken into account in the final Q-sort. Research assistants discussed every item together, consulted their interview notes, shared their observations, and decided on card placement. Every home visit was also discussed with the study’s principal investigator (A.H.B.) during weekly lab meetings with all research assistants present. Disagreements (which were rare) were conferenced. Each participant’s sort was then correlated with a criterion sort developed by experts in the field that represents the “ideal secure” infant. The resulting correlations were then used as the variable in analyses, with higher scores representing greater attachment security.

Child social-emotional competence and behavior problems.

The Brief Infant-Toddler Social and Emotional Assessment (28) was administered to evaluate the toddler’s social, emotional, and behavioral problems, as well as competencies, according to maternal report. The 42-item scale assesses child difficulties and competencies within the past month on a 3-point scale from 0 (not true/rarely) to 2 (very true/often). Items are summed to form the problem scale (31 items) and the competence scale (11 items). The measure has been demonstrated to have strong psychometric properties (28). In the present study, Cronbach’s alpha was 0.83 for the problem scale and 0.85 for the competence scale.

Demographic characteristics.

A questionnaire was administered to mothers to gather data on characteristics such as age, education, income, marital status, and ethnicity (Table 1). The Conflict Tactics Scale (29) was used to assess partner abuse during pregnancy with items covering minor and severe forms of abuse. Participants rated items on the subscales for physical abuse (12 items), sexual coercion (seven items), and psychological abuse (eight items) in reference to their partner’s behavior during pregnancy. Physical and sexual abuse items were combined into a total score, and the final physical and psychological abuse scores were dichotomized (absent or present). In this study, Cronbach’s alpha was 0.63 for the physical abuse scale and 0.79 for the psychological abuse scale.

Data Analysis

Means, standard deviations, and ranges were computed for each variable (Table 2). Pearson correlations (two-tailed) are reported in Table 3. Significant correlations between predictor (mothers’ defenses) and criterion variables (toddler variables) determined inclusion in regression analyses. Hierarchical multiple regression analyses were performed with demographic and partner abuse variables entered into the first block and defense levels entered into the second block in order to determine whether defenses predict attachment security, social-emotional competence, and behavior problems beyond what is accounted for by the demographic and partner abuse variables. To avoid the issue of collinearity, the overall defensive functioning scores were run in separate analyses from defense levels.

Results

Correlations Between Mothers’ Defenses, Demographic Variables, and Abuse During Pregnancy

Correlations between demographic or partner abuse variables and defense mechanisms are reported in Table 3. Only marital status failed to correlate significantly with any defense or toddler outcome variables, and it was not included in any subsequent analyses.

Correlations Between Maternal Defenses and Toddler Outcomes

Mothers’ overall defensive functioning and high adaptive level defense scores were significantly and positively correlated with child attachment security and social-emotional competence (Table 4). Mothers’ overall defensive functioning and high adaptive level defense scores were also significantly and negatively correlated with child behavior problems. The correlation between disavowal and behavior problems approached but fell short of significance (p=0.07). Disavowal was the only other defense level that significantly correlated with toddler outcomes. Therefore, high adaptive and disavowal level defenses were included in the regression analyses predicting attachment security and social-emotional competence. Separate regressions were run with overall defensive functioning predicting toddler outcomes.

Predicting Toddler Outcomes With Overall Defensive Functioning and Defense Levels

For regression analyses, demographic and partner abuse variables were entered into the first block and overall defensive functioning, high adaptive, and disavowal defense variables were entered as the second block. Results of the six analyses are reported in Table 5.

Mothers’ Defense Mechanisms Predict Attachment Security, Social-Emotional Competence, and Behavior Problems in Toddlers

Demographic and partner abuse variables significantly predicted attachment security (p=0.001), and together they explained 25% of the variance in attachment security. Overall defensive functioning significantly predicted attachment security after variance from the first block was accounted for (p=0.015), explaining an additional 6% of the variance. In a separate analysis, high adaptive and disavowal level defenses, entered as the second block, significantly predicted attachment security (p=0.001), explaining an additional 13% of the variance.

Demographic and partner abuse variables in the first block failed to predict toddler social-emotional competence. Overall defensive functioning, entered as the second block, approached significance (p=0.06) in predicting social-emotional competence after variance from the first block was accounted for (p=0.015), explaining an additional 5% of the variance. In a separate analysis, high adaptive and disavowal level defenses significantly predicted social-emotional competence after variance from the first block was accounted for (p=0.007), explaining an additional 12% of the variance.

Finally, demographic and partner abuse variables in the first block failed to predict toddler behavior problems. Overall defensive functioning, entered as the second block, also failed to predict behavior problems. In a separate analysis, high adaptive level defenses, entered as the second block, significantly predicted behavior problems, explaining an additional 6% of the variance.

Discussion

To our knowledge, this is the first prospective study evaluating maternal defense mechanisms as predictors of toddler attachment security, social-emotional competence, and behavior problems. Our findings demonstrate that greater use of healthy, adaptive defenses by pregnant mothers, assessed through their representations of their child in utero, predicted greater attachment security, better social-emotional competence, and fewer behavior problems in their toddlers after accounting for the significant relationships between demographic or partner abuse variables and child outcomes. Similarly, the findings demonstrate that less frequent use of maladaptive defenses (disavowal level defenses) predicted greater attachment security, better social-emotional competence, and fewer behavior problems. Thus, it appears that, as predicted, a mother’s ability to use healthy defense mechanisms during stressful times allows her to support the child’s psychological development and regulatory capacities in the earliest years of life.

These findings raise questions regarding the mechanisms by which certain maternal defenses contribute to toddlers’ social-emotional adjustment. We propose several such mechanisms, including maternal caregiving behavior and the mother’s capacity to mentalize about the child during pregnancy and after birth. Mentalization is the process by which an individual is able to consider and anticipate what is in the mind of another and use mental states to understand behavior (30–32). For a parent, mentalization, also known as parental reflective functioning (33, 34), involves entering into the inner world of the child and helping the child make sense of and organize his or her experiences.

In our sample, the most frequently occurring high adaptive defense mechanisms were humor, anticipation, altruism, suppression, and self-observation, all of which may promote more sensitive caregiving and greater parental mentalizing. For instance, the appropriate use of humor likely increases shared positive affect and attentive playfulness with the child and may help diffuse tension or negative affect under stress. Anticipation promotes the mother’s ability to store up and/or measure her own resources, considering ahead of time when the child is likely to need her support; planning and anticipation may also reduce the occurrence of stressful experiences in general. Altruism allows the mother to set her own needs aside to care for the child, placing the child’s attachment needs at the forefront, while also continuing to gain gratification from doing so. The use of suppression in response to daily stress, anger, and fatigue while parenting a toddler diminishes the chances of any direct aggression or hostility toward the child and instead allows for a “cooling down” period before responding. Finally, self-observation refers to the mother examining her feelings, wishes, fears, and maladaptive ways of potentially responding to the child. Through the use of self-observation, the mother can choose not to follow her immediate impulses in responding to a difficult situation with the toddler (e.g., blaming [projection or devaluation] or hitting the child [acting out]). Self-observation most overlaps with mentalizing, although all of the healthy defenses probably overlap somewhat with a mother’s capacity for reflective functioning. Therefore, overall, the findings in our study are consistent with past research demonstrating that higher levels of maternal reflective functioning are associated with more sensitive caregiving behavior, as well as greater infant attachment security (35, 36).

In contrast, the use of disavowal level defenses (denial, projection, rationalization) may distort perceptions of a child’s mental states and behavior. For example, denying a toddler’s need for emotional refueling after separations may interfere with the child’s sense of security because the mother may not offer support or soothing at such times. Anecdotally, we have also heard mothers in our sample describing the toddler as “not needing me” because of the child’s (pseudo) independence or because of the child serving in some role-reversed way in order to meet the needs of the mother (another form of denial of the child’s needs). Additionally, projecting malevolent intentions to a child’s behaviors (e.g., “He won’t go to bed because he wants me to be miserable,” or as seen in the first of the two vignettes presented in the shaded box) is likely to result in negative maternal responses that do not match the child’s internal experience. As a result, repeated experiences with a mother who displays such projections likely results in the toddler developing an incoherent sense of self, having difficulty understanding and organizing interpersonal experiences, and exhibiting poor emotion regulation and insecure, if not disorganized, attachment behavior. Indeed, a few reports (37, 38) indicate that maternal negative projections toward their infant or toddler are associated with problematic caregiving behavior, as well as child internalizing and externalizing problems and disorganized attachment. It is important to note, however, that no existing studies to our knowledge have measured projection per se by trained coders. Furthermore, the importance of negative maternal projections is highlighted by the fact that they are the focus of some forms of infant-parent psychotherapy (39).

This study had several strengths. One was its prospective, multimethod design, following women and their children from pregnancy through age 2. Another strength was that the ratings of the mothers’ defenses were based on a semistructured interview that elicited representations of their unborn child; to our knowledge, defense mechanisms have not previously been studied in pregnant parents to predict later child outcomes. Furthermore, ratings of defense mechanisms were not based on mothers’ defense use in general. The Defensive Functioning Scale allows for a comprehensive assessment of defenses ranging from normal to pathological functioning. A third strength was the success in maintaining contact with 84 at-risk mothers over a 2-year period. As for limitations, mentalization capacities and observed maternal behavior among participants were not directly assessed. Consequently, statements about causality cannot be made with certainty, and interpretations of the findings need to be examined in future studies. Whether or not mothers’ defense mechanisms directly influence attachment security and other aspects of child development, or are mediated in part by maternal mentalization and maternal behavior, remains an empirical question.

In summary, findings from this study suggest that maternal defense mechanisms are important in understanding the nature of the earliest parent-child relationship and subsequent child adjustment in the early years of life. Examining maternal defenses and other forms of coping may be particularly important during the prenatal and postnatal periods because this is typically a physically and emotionally demanding time, especially for women experiencing adversity and risk such as single parenthood, economic disadvantage, and difficulties with partners and social support. The findings underscore the importance of assessing defense mechanisms as part of a comprehensive evaluation of pregnant mothers and mothers with young children.