Primiparous Mothers’ Parenting Self-Efficacy in Managing Toddler Distress: Childhood Non-Supportive Emotion Socialization, Adult Attachment Style and Toddler Temperament as Antecedents

Abstract

Early maternal sensitivity to child distress is predictive of child subsequent social-emotional adjustment. A mother’s global parenting self-efficacy shapes her adaptive responses to child challenging behaviors (e.g., negative emotions). However, little is known about the antecedents of maternal self-efficacy in managing child distress. Using longitudinal data from a diverse sample of 259 primiparous mothers and their toddlers, we tested a model predicting maternal self-efficacy in managing toddler distress. Mothers’ remembered childhood experiences of maternal non-supportive emotional socialization were positively associated with their self-reports of attachment avoidance and anxiety. Further, a negative link between mothers’ self-reports of attachment avoidance and their self-efficacy in managing toddler distress was identified when toddlers displayed higher negative affect. Most importantly, a conditional indirect pathway was found. That is, only when toddlers displayed higher negative affect, mothers’ remembered early experiences of maternal non-supportive emotional socialization were negatively associated with their self-efficacy in managing toddler distress through a positive association with their self-reports of attachment avoidance. Although maternal attachment assessed with Adult Attachment Interview was also included in analyses as a parallel mediator to self-reported attachment, no relevant effects emerged. These results suggest that by disrupting attachment development, parents’ early emotion socialization experiences hold long-standing implications for their subsequent confidence in managing child distress. Child distress as a threatening, attachment-related stimulus may contextualize such effects via interfering with the operation of caregiving system, especially for avoidant parents. Such findings may inform the designs of more targeted interventions to assist first-time mothers in navigating emotionally-evocative challenges during toddlerhood.

A mother’s parenting self-efficacy, defined as the degree to which she is confident in her abilities to perform parenting tasks competently, has long been shown to play a salient role in shaping maternal well-being, parenting quality, as well as child adaptation (Schuengel & Oosterman, 2019). Maternal parenting self-efficacy also has long been an important intervention target due to its modifiable property (e.g., Gilkerson et al., 2020). Notably, to more effectively promote predictive validity and also guide the designs of more targeted interventions, researchers have increasingly advocated for studies investigating domain-specific (versus the global) parenting self-efficacy (Cao et al., 2022; Junttila et al., 2015). Coping with child distress represents one of the most emotionally evocative tasks during the first few years of life, especially for primiparous mothers (Kurth et al., 2011). There is substantial evidence supporting developmental cascades in which early emotion-related parenting predicts children’s subsequent social-emotional adjustment through shaping children’s acquisition of social-emotional competence (Leerkes & Bailes, 2019; Perry et al., 2020). In particular, as compared to maternal sensitivity to child non-distress, maternal adaptive responses to child distress cues have been found to be uniquely more predictive of child subsequent social-emotional adaptation (Conradt & Ablow, 2010; Leerkes, 2011; Leerkes et al., 2009; McElwain & Booth-LaForce, 2006). Thus, it seems important to understand the origins of maternal self-efficacy in managing child distress, which refers to a mother’s confidence in her abilities to accurately detect, understand, and appropriately respond to child negative emotions (Cao et al., 2022; Leerkes, 2010; Leerkes, Crokenburg, & Burrous, 2004). However, research in this area still remains extremely scarce.

The present study seeks to address this gap by testing a model of the antecedents of maternal self-efficacy in managing child negative emotions. First, this model represents a novel extension of the emotion socialization model proposed by Eisenberg et al. (Eisenberg, Cumberland et al., 1998; Eisenberg, Spinrad et al., 1998). In our model, mothers’ self-efficacy in managing child distress was considered as one social competence outcome that is possibly related to their emotion socialization experiences in childhood. Also drawing on Bowlby’s attachment theory (Bowlby, 1969/1982), we tested mothers’ adult attachment as one intrapersonal database mechanism that might be underlying the link between their remembered childhood parental emotion socialization and the extent to which they feel confident in their abilities to manage child negative emotions. In testing so, we noted that previous research has identified a “trivial to small” overlap between self-reported and interview-based adult attachment security, and yet both are predictive of key outcomes (Jones et al., 2015; Roisman et al., 2007). Thus, each of the two approaches appears to tap distinct aspects of adult attachment security. Accordingly, scholars have called for research administrating adult attachment interview (AAI) and self-reports of adult attachment in tandem to examine their relative, unique links with various facets of parenting, including parenting self-efficacy (Cao et al., 2022; Jones et al., 2015; Roisman et al., 2007 for discussions). Thus, both self-reported and interview-based adult attachment was considered in the model. Last, our examination is also informed by Dix’s (1991) affective process model of parenting and guided by recent propositions by attachment theorists with regards to the implications of parental attachment for parenting behaviors, emotions, and cognitions (Jones et al., 2015; Mikulincer & Shaver, 2019a). Accordingly, child distress as a threatening attachment-related stimulus likely contextualizes parenting in the moment by shaping the potential implications of a mother’s attachment behavioral system for the operation of her caregiving behavioral system. Thus, we also tested the possible moderating role of child negative affect in the association between maternal attachment and self-efficacy in managing child negative emotions.

The Role of Mothers’ Childhood Experience of Non-Supportive Emotion Socialization

Parental emotion socialization can be conceptualized as the process through which parents teach their children to understand the causes and consequences of their own and others’ emotions and to appropriately express and regulate their own emotions (Eisenberg, Cumberland et al., 1998). Although a wide array of parental emotion socialization practices exists, we focus on maternal non-supportive reactions to child distress, including ignoring, dismissing or punishing children for showing negative emotions and getting upset when children display negative emotions (Fabes et al., 2002). We focus on mothers’ recollection of their mothers’ parenting because: (1) prior research has demonstrated that mothers’ reports of their own mothers’ parenting are predictive of maternal parenting self-efficacy whereas their reports about fathers are not (e.g., Leerkes & Crockenberg, 2002); and (2) prior analyses using the current sample indicate that a high proportion of reports about fathers is missing non-randomly, which introduces selection biases (e.g., Cao et al., 2018). We focus on the recalled non-supportive emotion socialization given prior evidence showing that it plays a more salient role in shaping adults’ subsequent emotion-related parenting cognitions and practices than the recalled supportive emotion socialization (e.g., Leerkes, Bailes, & Augustine, 2020).

As Bandura (1977) proposed, “People do not rely on experienced mastery as the sole source of information concerning their level of self-efficacy. Many expectations are derived from vicarious experience.” (p. 197). When mothers were children, they might benefit from having an early effective model of emotion-related parenting by experiencing how their parents responded to their (i.e., the mothers as young children) distress and also by observing their parents’ competently coping with infant or toddler siblings’ negative emotions if they had siblings. Such behaviors can be readily retrieved and reproduced in their later interactions with their own children. Such modeling processes may also convey messages to children that they are capable of satisfying their own children’s needs. As a result, the modeling and vicarious experience may facilitate mothers’ emotion-related parenting self-efficacy via enhancing their performance attainment in managing child distress.

In addition to the modeling and observational learning processes, other mechanisms may be also applicable to understanding the implications of early experiences of parental emotion socialization for mothers’ subsequent emotion-related parenting self-efficacy. In line with the tenets of Eisenberg, Spinrad, et al.’s revised model, early experiences of parental non-supportive emotion socialization have been extensively shown to have salient roles in hindering children’s acquisition of critical social-emotional competence, such as understanding of emotions, regulation of emotions, and affective stance toward emotions, and thus pose threats to child subsequent social-emotional adaptation (Engle & McElwain, 2011; Fabes et al., 2001; Perry et al., 2020). An accumulating body of research has indicated that mothers’ early experiences of non-supportive emotion socialization undermines their emotion-related competence and social-emotional functioning during early parenthood (Cao et al., 2018; Leerkes, Supple, Su et al., 2015). Further, such experiences disrupt how they make meaning of and regulate their emotional arousal in the face of children’s distress cues (e.g., crying; Leerkes & Crockenberg, 2006; Leerkes & Siepak, 2006), which in turn would interfere with their abilities to respond to child negative emotions in an adaptive manner (i.e., the intergenerational transmission of non-supportive emotion socialization; Lee et al., 2021; Leerkes et al., 2020). According to the cognition–affect–behavior process framework outlined by Cao et al. (2022) for the emergence and evolution of maternal self-efficacy in emotion-related parenting, mothers’ deficits in cognitive and affective abilities in responding to child distress sensitively are likely to elevate the risk of their failures in soothing child distress, which contribute to mothers’ lack of confidence in managing child distress. Thus, we expect a negative link between mothers’ early experiences of maternal non-supportive emotion socialization and their later confidence in managing child distress.

Mothers’ Adult Attachment as a Mediator: AAI-Based Versus Self-Report Measures

Drawing from Bowlby’s view (1969/1982) that the most fundamental function of attachment is to promote infants’ survival, attachment theorists (e.g., Thompson, 1997) have argued that parental responses to infant distress cues or safety bids should be particularly relevant to the development of attachment security. Accordingly, the formation of children’s insecure working models is thought to mainly stem from early experiences of harsh, withdrawn or inconsistent caregiving. In line with such propositions, early experiences of non-supportive emotion socialization either within or outside of the normal range (e.g., emotional maltreatment) have also been linked with insecure adult attachment style (Cao et al., 2020; Magai et al., 2004). Notably, primiparous mothers’ remembered care from their mothers has been found to be predictive of their global parenting self-efficacy indirectly through shaping their global self-esteem (Leerkes & Crockenberg, 2002). As a positive view of self is a key aspect of a secure internal working model, this study provides some support to the potential mediating role of adult attachment security in the link between early experiences of non-supportive emotion socialization and subsequent parenting-related self-efficacy.

Avoidance and anxiety are two central dimensions of adult attachment. Anxiously attached individuals and avoidantly attached individuals demonstrate distinct characteristics from each other in many domains (see Mikulincer & Shaver, 2012, 2016, 2019b; Mikulincer et al., 2009; Shaver et al., 2019). Given the unpredictable caregiving that they received in the early years of life, individuals scoring high on attachment anxiety have a hypervigilant appraisal-monitoring system dominated by worries about being abandoned. In the face of challenges, they tend to exaggerate the seriousness of threats and overemphasize their vulnerabilities, because they believe that such signs of helplessness may elicit attachment figures’ attention and increase the likelihood of receiving care. In contrast, due to the attachment figure’s unresponsiveness in early childhood, individuals with high attachment avoidance have doubts about others’ trustworthiness. As a result, they value self-reliance, are fearful of intimacy, feel uncomfortable with others’ dependence, and prefer to maintain a safe emotional distance with others. In stressful situations, they tend to block their inner feelings that are associated with a sense of weakness by minimizing or dismissing their own and others’ needs and downplaying the significance of the situation. Such strategies may protect them from experiencing attachment needs that would otherwise activate the attachment system and increase the chance of getting rejected.

Before laying out a rationale for the link between adult attachment and maternal self-efficacy in managing child distress, we noted that in spite of their common roots in Bowlby’s theory, self-reported adult attachment style and interview-based adult attachment representations appear to tap distinct aspects of adult attachment security. An insightful view about such differences has been offered by Roisman et al. (2007, pp. 679–680). Self-report adult attachment measures centers on how individuals perceive and understand their experiences in the present or recent close relationships, whereas the AAI coding focuses on whether adults are able to produce coherent narratives about their early experiences with primary caregivers. The coherence of such narratives represents the degree to which they would use defensive mechanisms when recalling and describing early attachment experiences, irrespective of how they view such experiences now. The two measurement traditions also differ in their assumptions about the circumstances under which individual differences in attachment are most readily revealed. Researchers utilizing the self-report measures hold the belief that the attachment working models shape attachment-related behaviors under the conditions of stress or threat. In contrast, researchers using the AAI do not consider such danger or threat as a key condition for the identification of the links between coherence of mind and interpersonal behaviors. Instead, AAI-based adult attachment security appears to more reflect a general interpersonal asset that subserves effective coping and interpersonal cooperation. Given that a distressed child seeking comfort can be viewed as an attachment-related threat stimulus for parents, it seems warranted to expect that the self-reported adult attachment style may be more relevant for our proposed model.

Despite the modest correlation between self-reported and AAI-based adult attachment, they are similarly related to a wide range of attachment-relevant constructs in theoretically expected ways (e.g., emotion regulation, defensive processes, social information processing, relationship functioning) and both also appear to be reliably associated with various facets of parenting (see Jones et al., 2015; Shaver et al., 2000). As to responses to child distress in particular, parents with greater attachment insecurity on self-report measures tend to view an upset infant seeking comfort as more threatening, be less tolerant of infant distress (River et al., 2019), make more negative attributions about child distress cues (Leerkes & Siepak, 2006), have more difficulties in adaptively responding to child needs (Mills-Koonce et al., 2011), especially in highly emotion-charged tasks (Edelstein et al., 2004; Sher-Censor et al., 2020). Likewise, mothers with an insecure attachment representation assessed with AAI tend to be less responsive to child distress cues and interpret such cues more negatively (Leerkes et al., 2015; Lowell et al., 2021; Spangler et al., 2010) and struggle to regulate their emotions in the face of child distress (Ablow et al., 2013; Leerkes et al., 2015; Xu & Groh, 2021).

As to parenting self-efficacy specifically, extensive studies demonstrate that adults who are insecurely attached based on self-report assessments tend to hold more pessimistic attitude toward childrearing and expect childcare to be more challenging. They also tend to report lower competence in parental roles and lower confidence in the ability to relate to children (see a summary in Cao et al., 2022). In contrast, only two somewhat indirectly relevant studies were identified for the implications of AAI attachment for parenting self-efficacy. A study by Shlafer et al. (2015) on low-income first-time mothers found that mothers’ AAI coherence of mind measured at age 26 years predicted their supportive parenting six years later. One indicator for parenting quality in this study was parental confidence (i.e., sense of efficacy in parental role), but the specific link between AAI coherence of mind and parental confidence was not tested. In a study of first-time pregnant women who were abused in childhood, Kunseler et al. (2016) found that the unresolved/not unresolved status of abuse experiences assessed with AAI did not explain variance in parenting self-efficacy, but AAI coherence of mind was not examined in this study. As such, due to the sparseness of relevant research, the direct links still remain unclear between AAI-based attachment representations and parenting self-efficacy; and there is a lack of research on such links in low-risk community samples. Yet, given the widely-documented implications of AAI attachment representations for the affective and cognitive processes that are variously implicated in parental responses to child distress (Ablow et al., 2013; Leerkes et al., 2015; Lowell et al., 2021; Spangler et al., 2010; Xu & Groh, 2021), the links between AAI attachment representations and parenting self-efficacy merit additional examination. Take altogether, we expect that maternal adult attachment, especially attachment style assessed with self-report measures, may be an explanatory mechanism by which mothers’ early experiences of non-supportive emotion socialization are linked with their subsequent self-efficacy in dealing with child distress.

The Contextualizing Role of Child Negative Affect in Mother-Child Interaction

Parenting is an emotional endeavor (Kerr et al., 2021). Parent and child affect both serve to organize parenting (Dix, 1991). From a dynamic systems perspective, children’s distress perturbs parents’ internal system away from equilibrium, and parents facing children’s more challenging behaviors tend to experience greater perturbation in the moment (Zhang et al., 2022). Disruptive child behaviors in challenging parenting situations can shape parental momentary thoughts of self-efficacy and feelings of stress (Schulz et al., 2019). Mothers whose children get upset frequently and intensely tend to feel more anxious about performing emotion-related parenting tasks and also may attribute such child fussiness to themselves and thus experience diminished confidence in their emotion-related parenting competence (Leerkes & Crockenberg, 2006; Lipscomb et al., 2011). As Hajal and Paley (2020) stated, “When children are highly dysregulated and seemingly unresponsive to their parents’ efforts to help them recover, parents may hear a message that they are failing at their most central task, which is to take care of and protect their child.” (p. 414). Thus, the varied levels of child distress constitute the specific contexts in which parents’ attachment and caregiving behavioral systems operate. Research has highlighted the importance of quantifying the context of emotion-related parenting by measuring children’s distress levels or identifying discrete moments when they display intense distress (Augustine & Leerkes, 2019; Ravindran et al., 2022).

A dependent young child’s distress represents an attachment-related stimulus for mothers (see Jones et al., 2015). The activation of mothers’ attachment and caregiving behavioral systems is shaped by their perceptions of contextual threats. How mothers appraise and respond to contextual threats are affected by their individual differences in attachment security. An upset infant seeking parents’ comfort should be perceived as especially threatening for mothers with attachment insecurity. This is partly because these mothers’ early caregiving experiences, particularly those in distress contexts, have contributed to the development of their defensive mechanisms to cope with threats. In turn, such increased activation of a mother’s own attachment needs may significantly undermine her ability to appropriately taking care of an upset child’s needs. As Jones et al. (2015) stated, “It may be best to think of child behavior … as increasing activation of a caregiving system operating within the context of threat assessment that is influenced by the parent’s attachment system.” (p. 46).

Notably, extant research (Edelstein et al., 2004; Feeney & Collins, 2001; Simpson et al., 1992) have indicated that individuals’ ability to provide support to either a romantic partner or a young child under the threat of separation or in anxiety-producing situations is influenced by both the provider’s self-reported attachment style and the recipient’s affective state that define the contexts under which the interpersonal interactions unfold. In brief, results of these studies suggest that emotional support and caregiving to a romantic partner or a young child are influenced by an interaction between the provider’s self-reports of attachment avoidance (instead of anxiety) and the recipient’s intensity of distress. Specifically, when threat or stress is low, attachment avoidance is often non-significantly associated with the emotional support/caregiving provision; in contrast, when threat or stress is high, attachment avoidance is negatively associated with the emotional support/caregiving provision.

Edelstein et al. (2004, pp. 45–46) proposed several mechanisms that may account for why avoidantly attached parents tend to be less supportive toward their highly dependent and upset children. Avoidant parents tend to minimize and neglect children’s attachment needs, prefer to keep a safe emotional distance from others, be uncomfortable with disclosure of vulnerabilities and emphasize self-reliance, be reluctant to allow others to depend on them, and be fearful of intimacy. These tendencies reflect coping strategies in the face of threats that these parents have developed out of early experiences of parental rejection or neglect about their bids for comfort in distress situations. Further, such maladaptive attachment-related characteristics likely interfere with the operation of avoidant parents’ caregiving behavioral system in response to children’s negative emotions and result in low confidence in managing child distress. Indeed, avoidant parents tend to experience greater stress after child birth, reap less joys of caregiving, and perceive parenting as less satisfying and meaningful (Nelson-Coffey et al., 2017; Rholes et al., 2006). Avoidantly attached adults tend to perceive themselves as less competent in preforming caregiving tasks. Their parenting-related anxiety may lead them to withdraw from upset children when parental adaptive emotional responses are most urgently needed (Feeney & Collins, 2003; Rholes et al., 1997). Taken collectively, we expect to identify an interaction effect between maternal self-reported attachment avoidance and child negative affect, such that avoidance will only be associated with lower self-efficacy in managing child distress among mothers whose toddlers display higher negative affect. Further, to determine if the moderating effect is indeed unique to maternal self-reported attachment avoidance, at the meantime we also test the interaction between maternal self-reported attachment anxiety and child distress as well as the interaction between maternal interviewed-based attachment representation and child distress.

The Present Study

The central goal of this study is to identify the origins of maternal self-efficacy in managing child negative emotions and to elucidate the mechanisms by which such effects occur. Our analyses were based on longitudinal data from a diverse sample of 259 first-time mothers and their toddlers. We hypothesize that mothers’ childhood experiences of maternal non-supportive emotion socialization will negatively but indirectly predict their subsequent self-efficacy in coping with child distress through positive links with their insecure adult attachment style/representations. Moreover, we also expect that child negative affect will moderate the associations between mothers’ insecure adult attachment style/representations and their self-efficacy in managing child distress. That is, the negative associations between mothers’ insecure adult attachment style/representations and their self-efficacy in managing child distress will be stronger when the child demonstrates higher negative affect than when the child displays lower negative affect. Notably, mothers’ self-reported attachment avoidance may be especially relevant for this interaction hypothesis, given the higher theoretical relevance of self-reported attachment style for our proposed model as compared to AAI-based attachment representations (Roisman et al., 2007) and prior empirical evidence showing the interaction between self-reported attachment avoidance (instead of anxiety) and contextual threats in predicting caregiving provision (Edelstein et al., 2004; Feeney & Collins, 2001). To our knowledge, this is the first study focusing on the origins of maternal self-efficacy specifically in managing child negative emotions. Findings of this study may inform the designs of more targeted interventions aimed at assisting first-time mothers in navigating emotionally-evocative challenges during a difficult period of parenthood (i.e., toddlerhood) that is characterized by more frequent and intense child negative affect in the context of normative autonomy struggles. That is, toddlerhood is such a critical period for the emergence of autonomous and self-regulated behavior, during which many of the interactions between parent and child are marked by efforts of the parent to exert control over the child as the child seeks autonomy and independence from the parent (Brownell & Kopp, 2007). Last, the present study may also help identify avenues for practices aimed at improving the quality of early emotion socialization in the family setting, which likely yield long-term benefits by promoting child social-emotional functioning (e.g., emotion regulation).

Methods

Participants and Procedures

Participants were derived from a larger project focusing on the origins of maternal sensitivity. Two hundred and fifty-nine primiparous mothers, including 128 European American, 123 African American, and 8 multiracial, and their infants from the southeastern United States participated in the larger project. Notably, this sample is non-representative of adult women living in the data collection region, because the original project exclusively focused on pregnant primiparous mothers and also particularly oversampled African American mothers to increase diversity. Mothers’ age ranged from 18 to 44 years, with a Mean of 25.1. As to education attainment, 27% had a high school diploma or less, 27% had attended but not completed college, and 46% had a 4-year college degree. A majority (57%) were married or living with the child’s father, 24% were in a relationship but not living with the child’s father, and 19% were single. Annual family income ranged from less than $2,000 to over $100,000, with a median of $35,000. All infants (51% female) were full term and healthy. Of the initial 259 participants, 230 participated in the 6-month questionnaire assessment, 227 participated in the 1-year questionnaire assessment, and 198 completed the 2-year observation as well as the interview after such observation. The reasons for not participating across waves included: (a) too busy to continue, (b) the infant died, (c) did not respond to multiple attempts of contact, and (d) voluntarily withdrew from the study. Results of attrition analyses indicated minimal attrition bias (see the Supplementary Material for details). Expectant mothers were recruited from various sources, including childbirth classes in local hospitals (n = 95) and the public health department (n = 28), breastfeeding classes through the Special Supplemental Nutrition Program for Women, Infants and Children (n = 100), obstetric practices (n = 12), and word of mouth (n = 24). To be eligible, a woman should be expecting a singleton, be African American or European American, be 18 or older, and be fluent in English. Additional details for data collection procedures were described in the Supplementary Material. The protocol was approved by the institutional review board at the University of North Carolina at Greensboro (IRB No. 09–0035, Triad Child Study).

Measures

Mothers’ Childhood Experience of Non-Supportive Emotion Socialization.

At the prenatal wave, women recalled mothers’ reactions to their negative emotions in specific ways across nine situations (e.g., being scared of injections) during their first 16 years of life. A modified version of the Coping with Children’s Negative Emotions Scale (CCNES; Fabes et al., 2002) was used. For each situation, on a 7-point scale ranging from 1 (very unlikely) to 7 (very likely), participants rated how likely their mothers were to react in each of the three ways of nonsupportive reactions: minimizing reactions, punitive reactions, and distress reactions. Sample items for each reaction can be found in the Supplementary Material. By averaging the ratings on items across the various scenarios in a scale, subscale scores were generated. Cronbach’s alphas were .67, .78, and .77 for distress reactions, punitive reactions, and minimizing reactions subscales, respectively. Then, a latent construct was created using three manifest reaction indicators.

Mothers’ Adult Attachment.

At the prenatal visit, mothers were administered the Adult Attachment Interview (AAI; George et al., 1984–1996). The AAI is a semistructured interview in which participants are asked to describe their early relationships with primary caregivers and the implications of those experiences for them. Verbatim de-identified transcripts were coded by coders who were trained through and reliable with Dr. Mary Main’s lab using the AAI Scoring and Classification System (Main, Goldwyn, & Hesse, 2003–2008). Interrater agreement for the 2-way autonomous (secure)/insecure classification system was 82%, k = .63, p < .001, based on 50 double-coded cases. In addition, we also used the coherence of mind rating (1 = not at all coherent to 9 = very coherent) as a criterion measure for adult attachment security. It is a summary measure of participants’ ability to describe early attachment experiences as well as the influences of such experiences on their current functioning in an organized manner. Prior research has indicated that the coherence of mind rating is the AAI rating that best distinguishes between secure and insecure adults and it is significantly associated with observed parenting (Bosquet & Egeland, 2001; Cowan et al., 1996). Interrater reliability for this rating was good, intraclass correlation was .75, p < .001.

Mothers also completed the 36-item Experiences in Close Relationships scale (ECR; Brennan, Clark, & Shaver, 1998) to assesses adult attachment style at the prenatal, 6-month, 1-year, and 2-year postpartum waves. The ECR is composed of two 18-item subscales, anxiety and avoidance. Participants indicated the extent to which they agree with each item on a 7-point scale ranging from 1 (disagree strongly) to 7 (agree strongly). Example items include: “I worry about being abandoned” and “I try to avoid getting too close to my partner.” Cronbach’s alphas ranged from .93 to .94 for the avoidance subscale and from .88 to .90 for the anxiety subscale across waves. The scores at 6-month and 1-year waves were averaged to yield a more reliable estimate of attachment styles during infancy, which was used in the primary analyses as potential mediators, whereas the mean scores at prenatal and 2-year waves were included as key covariates in primary analyses.

Toddlers’ Negative Affect.

Toddler affect during the 2-year interactive tasks (free play, clean up task, locked box task, and spider task) was continuously rated from digital media files during the emotion-eliciting tasks using INTERACT V.9 software (Mangold International GmbH, Arnstorf, Germany). Toddler affect was rated on a 7-point scale ranging from (1) high positive affect (e.g., intense smile) to (7) high negative affect (e.g., sobs intensely). This scale was adapted from Braungart-Rieker and Stifter’s study (1996) based on infants’ vocalizations, facial expressions, and body tension. Event-based coding was used (i.e., once a code was activated, it remained active until another code was selected). Coders were blind to other data, reliability cases were selected at random, and disagreements were resolved via consensus. Inter-rater reliability was Kappa = .73 (weighted kappa = .81) based on 33 double coded videos. The average ratings of toddler affect across tasks was calculated, with higher scores indicating toddlers displayed more frequent and intense negative affect. Notably, 97% of toddlers became distressed (i.e., defined as a negative affect level ≥ 5) at some point across the 3 tasks, indicating the effectiveness of these tasks at eliciting child negative emotions, but the distress duration was relatively brief (Mean = 1.96 minutes, SD = 1.85, Range = 0–8.07 minutes).

Mothers’ Self-Efficacy in Managing Toddler Distress.

Mothers viewed the recorded segments of their interaction with toddlers during the distress-eliciting tasks (i.e., clean up task, locked box task, and spider task) while their toddler was cared for by a member of the research team in an adjacent room. After viewing each clip, mothers rated how good they thought they were at the following four items during the recorded task: knowing what your child wanted or needed, understanding what your child was feeling, knowing how to handle the situation and your child, and how good they felt they handled the situation. The response scale ranged from not good at all (1) to very good (4). This questionnaire was modeled after Teti and Gelfand’s (1991) Maternal Self-Efficacy Scale, which has been shown to be predictive of sensitive maternal behaviors in many studies (e.g., Hess et al., 2004; Teti & Gelfand, 1991). Items were averaged within each task to get task specific measures of efficacy and then across the 3 tasks to create a composite measure of efficacy. Cronbach’s alphas were .88, .86, .89 for clean-up task, locked box task, and spider task, respectively, and .90 for the total composite.

Covariates.

Mothers reported their age, race, child sex, prior experience with infants, and family income at the prenatal phase. They were asked to report their prior experience with infants on a 4-point scale ranging from 0 (none) to 4 (a lot). Although all mothers were primiparous, they varied substantially in prior experience caring for infants (M = 2.15, SD = .874), with 22% reported no or little, 37% reported some, and 41% reported a lot. Family income and the number of individuals living in the household were used to calculate family income-to-needs ratio (M = 2.644, SD = 1.993).

Analytic Strategies

Our hypotheses were tested using structural equation modeling (Byrne, 2013) in Mplus Version 8.4 (Muthén & Muthén, 1998–2017). As depicted in Figure 1, we conducted a model in which mothers’ childhood experiences of their own mothers’ non-supportive reactions to child distress (i.e., minimizing, distress, and punitive reactions) were specified as a latent construct (assessed at the prenatal wave) to predict their self-efficacy in managing toddler distress (assessed at 2-year wave), with AAI based maternal attachment (assessed at prenatal wave) and self-reported maternal attachment (assessed at 6-month and 1-year waves) tested as a potential mediator and child negative affect (assessed at 2-year wave) tested as a potential moderator. In terms of the interview-based maternal attachment, given that there was a high bivariable correlation between AAI coherence of mind and AAI dichotomous classification (r = .839, p < .001) and also considering the complexity of our model as well as the greater statistical power afforded by the utilization of continuous (as compared to dichotomous) variables, we used the continuous score in primary analyses. Potential indirect effects were assessed using the bootstrapping technique (Preacher & Hayes, 2008) with 5,000 resamples. The 95% confidence intervals (CIs) for unstandardized indirect effects that do not include zero indicate statistically significant effects.

Figure 1.

The model analyses results.

Notes. All coefficients are standardized and identified with all covariates included. For clarity, (a) pathways with p > .05 are depicted in gray color and pathways with p < .05 are depicted in black color; (b) direct and mediating effects paths are presented in solid lines, whereas moderating paths are shown in dash lines; and (c) the correlation lines and predicting pathways involving covariates are not depicted, but the relevant coefficients are available in the Supplementary Material. As an additional set of sensitivity analyses, AAI-based attachment also was separately examined as a dichotomous classification score (secure versus insecure) and the patterns of model results were the same as those using the AAI continuous score (coherence of mind).

To test the moderating role of toddler negative affect in mother-child interactions, product terms were created by multiplying mean-centered maternal adult attachment indicators and mean-centered toddler negative affect. Then, product terms were specified as exogenous variables predicting maternal self-efficacy. Further, to illustrate the patterns for the identified interactive effects, simple slope analyses were conducted following the recommendations by Aiken and West (1991). Specifically, simple slopes for the association between maternal attachment avoidance and maternal self-efficacy in managing child distress were examined at +1 SD and −1 SD from the mean of the moderator (i.e., child negative affect). Last, to test the conditional indirect effects, the mediating model was conducted again using the bootstrapping technique with 5,000 resamples at the higher and lower levels of child negative affect (i.e., 1SD above and 1SD below the mean).

Notably, in all analyses a series of covariates assessed at the prenatal wave were included as exogenous variables predicting outcomes while also correlating with predictors, including child sex, maternal age, maternal race, maternal education, maternal prior experience with infant, and family income-to-needs ratio. These covariates were selected based on both prior studies (e.g., Hess et al., 2004) and the results of preliminary zero-order bivariate correlation analyses. In addition, self-reported adult attachment style at both prenatal and 2-year waves were also included as key covariates in model analyses. Controlling for adult attachment style at the prenatal wave was to rule out possibility that the longitudinal association between mothers’ early experiences of non-supportive emotion socialization and their subsequent parenting self-efficacy was not solely a function of attachment-based differences in how mothers reported on their childhood experiences. Covarying out the effect of attachment style at the 2-year assessment would control for any potential confounders related to mothers’ changes in attachment over the transition to parenthood, given the relative malleability of adult attachment during this life period (e.g., Rholes et al., 2021; Simpson et al., 2003). This also helped rule out the possibility that the identified mediating effects might be an artifact of a concurrent link between attachment styles and parenting self-efficacy.

Model adequacy was evaluated using a series of indices, including Chi-Square statistic (χ²), the comparative fit index (CFI), the root-mean-square error of approximation (RMSEA), and the standardized root-mean-square residual (SRMR). A non-significant χ² , CFI > .90, RMSEA < .08, and SRMR < .08 indicate an acceptable model fit (Kline, 2015). For the present data, Little’s Missing Completely at Random (MCAR) test was nonsignificant (χ² = 224.552, df = 219, and p = .384), indicating a missing value pattern of being completely at random. The full-information maximum likelihood method (FIML, Acock, 2005) was used to handle missing data in primary analyses.

Transparency and Openness

This study’s design and its analyses were not pre-registered. However, the relevant analytic data file, supporting document, and Mplus syntax have been uploaded to the UNC dataverse hosted by the Odum Institute, which can be retrieved via this link: https://doi.org/10.15139/S3/CCYYQP. In previous sections, we have reported how we recruited participants, how we determined our sample size, all data exclusion criteria (if any), all data collection procedures, and all used measures in the study. Any additional materials and details are available from the authors upon reasonable request.

Results

Descriptive statistics for and zero-order bivariable correlations among study variables were reported in Table 1. The model as presented in Figure 1 demonstrated a good fit to the data: χ² = 70.233, df = 55, p = .081, RMSEA = .033 with a 90% CI [.000, .054], CFI = .982, SRMR = .033, N = 259. Mothers’ recalled childhood experiences of maternal non-supportive reactions to child distress was positively associated with both maternal attachment avoidance (β = .237, p < .001) and attachment anxiety (β = .252, p < .001).

Table 1.

Descriptive statistics for and zero-order bivariate intercorrelations among key study variables.

	1	2	3	4	5	6	7	8
Key Study Variables
1. CENES-Minimizing Reaction (Mother Recalled at Prenatal Wave)	‐‐‐‐
2. CENES-Punitive Reaction (Mother Recalled at Prenatal Wave)	.672	‐‐‐‐
3. CENES-Distress Reaction (Mother Recalled at Prenatal Wave)	.664	.717	‐‐‐‐
4. ECR-Avoidance (Mother Self-Report at 6-Month & 1-Year Waves)	.245	.164	.193	‐‐‐‐
5. ECR-Anxiety (Mother Self-Report at 6-Month & 1-Year Waves)	.194	.240	.211	.416	‐‐‐‐
6. AAI-Continuous Score (Coherence of Mind) (Interviewed at Prenatal Wave)	−.021	−.044	−.034	−.092	.030	‐‐‐‐
7. Toddler Negative Affect when Interacting with Mother (Observed at 2-Year Wave)	.146	.116	.111	.019	−.033	−.129	‐‐‐‐
8. Maternal Self-Efficacy in Managing Toddler Distress (Video-Based Self-Rating at 2-Year Wave)	−.060	.026	.080	−.236	−.173	.118	−.083	‐‐‐‐
Key Covariates
Child Sex (Mother Self-Report at Prenatal Wave)a	−.003	−.031	−.025	.085	−.015	−.039	.128	.025
Maternal Age in Years (Mother Self-Report at Prenatal Wave)	.017	.045	.010	−.209	−.036	.218	−.109	.174
Maternal Race (Mother Self-Report at Prenatal Wave)b	.097	.001	.095	.282	.040	−.254	.119	−.005
Maternal Education (Mother Self-Report at Prenatal Wave)c	−.072	−.053	−.040	−.248	−.082	.398	−.198	.185
Maternal Prior Experience with Infant (Mother Self-Report at Prenatal Wave)	−.046	.013	−.081	−.040	−.132	−.128	.018	.103
Family Income-to-Needs Ratio (Mother Self-Report at Prenatal Wave)	−.135	−.038	−.115	−.241	−.036	.232	−.148	.128
ECR-Avoidance (Mother Self-Report at Prenatal Wave)	.234	.085	.221	.689	.288	−.103	.071	−.205
ECR-Anxiety (Mother Self-Report at Prenatal Wave)	.239	.264	.249	.323	.663	.046	.072	−.125
ECR-Avoidance (Mother Self-Report at 2-Year Wave)	.197	.112	.153	.634	.356	−.186	.006	−.241
ECR-Anxiety (Mother Self-Report at 2-Year Wave)	.147	.188	.120	.319	.624	−.060	.035	−.121
Mean	3.147	2.929	2.631	2.541	3.026	5.313	4.188	3.577
Standard Deviation	1.217	1.194	1.274	1.134	1.033	1.460	.262	.396

Notes. n ranges from 186 to 259 given missing data for some variables. n represents the number of participants for each pair of zero-order correlation. CENES = Childhood Experience of Non-Supportive Emotion Socialization, ECR = Experiences in Close Relationships scale, AAI = Adult Attachment Interview. The italic correlation values indicate significance with p < .05 (2-tailed), whereas the bolded correlation values indicate significance with p < .01 (2-tailed). The reported correlations between binary variables (i.e., child sex and maternal race) and continuous variables were the Point-Biserial correlations and correlations among binary variables were the tetrachoric correlations.

^a.The codes for child sex: 1 = male, 2 = female.

^b.The codes for maternal race: 1 = European American, 2 = African American.

^c.The codes for maternal education: 1 = Some high school, 2 = high school or GED, 3 = some college, 4 = 2-year college degree, 5 = 4-year college degree, 6 = postgraduate work, and 7 = graduate degree.

Although no main associations were identified between mothers’ self-reports of attachment avoidance or anxiety and their self-efficacy in managing toddler distress (β= −.116, p = .300 for avoidance; β= −.147, p = .144 for anxiety), toddler negative affect was found to significantly moderate the link between maternal attachment avoidance and maternal self-efficacy in managing toddler distress (β= −.158, p = .032). The relevant simple slopes are depicted in Figure 2. There was a negative association between maternal attachment avoidance and self-efficacy in managing toddler distress when toddlers displayed higher levels of negative affect (b = −.104, S.E. = .050, β = −.274, p = .027), but there was no association when toddlers demonstrated lower levels of negative affect in mother-child interactions (b = .017, S.E. = .048, β = .049, p = .724).

Figure 2.

Illustration of the moderating role of toddler negative affect in the association between maternal attachment avoidance and self-efficacy in managing toddler distress.

Further conditional indirect effects analyses indicated that when toddlers displayed higher levels of negative affect in mother-child interactions, there existed an indirect effect from mothers’ childhood experiences of maternal remembered non-supportive emotion socialization to maternal self-efficacy in managing child distress through maternal attachment avoidance (b = −.029, S.E. = .019, 95% CI [−.079, −.002], β = −.072). This indirect effect was nearly “medium” in terms of magnitude. As to effect size criteria, standardized indirect effects around .01 can be interpreted as small, around .09 as medium, and around .25 as large (Kenny, 2012). In contrast, such an indirect effect was nonsignificant when toddlers demonstrated lower levels of negative affect in mother-child interactions (b = .005, S.E. = .016, 95% CI [−.029, .035], β = .012).

No significant associations were identified for coherence of mind assessed via the AAI. Mothers’ recalled early experiences of maternal non-supportive reactions to child distress was not associated with their AAI coherence of mind (β= −.055, p = .426). Likewise, mothers’ AAI coherence of mind was not associated with their self-efficacy in managing toddler distress (β= .059, p = .423). Last, no significant interaction emerged between mothers’ AAI coherence of mind and toddlers’ negative affect in predicting mothers’ self-efficacy in managing toddler distress (β= −.094, p = .165). Notably, such patterns of results did not change when conducting a separate, parallel set of model analyses using dichotomous classification for AAI-based maternal attachment (i.e., secure versus insecure). This is not unexpected, considering a high zero-order bivariable correlation between AAI coherence of mind and AAI dichotomous classification (r = .839, p < .001).

Discussion

In line with one of our primary hypotheses, we identified a moderating role of child negative affect in the association between mothers’ attachment avoidance and their self-efficacy in managing child distress, such that a negative association between attachment avoidance and self-efficacy emerged only when the child displayed higher negative affect. Furthermore, and most importantly, because of this moderation, a conditional indirect pathway emerged. That is, only when toddlers displayed higher negative affect, mothers’ remembered early experiences of maternal non-supportive emotional socialization were negatively associated with their self-efficacy in managing toddler distress through a positive association with their self-reports of attachment avoidance. The moderation finding adds to an accumulating body of evidence showing that romantic partners’ or caregivers’ attachment avoidance is negatively associated with their ability to provide emotional support or caregiving when the recipients’ (i.e., romantic partners or young children) distress is high (Edelstein et al., 2004; Feeney & Collins, 2001; Simpson et al., 1992). From an attachment theory perspective (Jones et al., 2015; Mikulincer & Shaver, 2019a), the activation of mothers’ attachment behavioral systems and caregiving behavioral systems is shaped by their perceived contextual threat. The levels of child distress reflect the intensity of attachment-related threat that mothers face when interacting with children. Although managing a highly distressed child seeking comfort is stressful for most mothers, it is particularly threatening for avoidant mothers because of their attachment-related vulnerabilities (e.g., being fearful of intimacy, uncomfortable with others’ dependence, and reluctant to seek help, valuing self-reliance, dismissing others’ needs, deactivating affiliation desires) (Feeney & Collins, 2003; Mikulincer & Shaver, 2019b; Rholes et al., 2006; Rholes et al., 1997; Shaver et al., 2019). Moreover, avoidant mothers’ various attachment-related vulnerabilities should be most revealed in the face of intense child distress (i.e., an attachment-related threat) and work in conjunction with each other as defensive mechanisms that interfere with the adaptive operation of their caregiving behavioral system, including cognitive processing of child distress cues, affective arousal and regulation in response to child distress, and actions to soothe child distress. Mothers’ disrupted caregiving system elevates their risk of failures in managing child distress, which should contribute to their lack of confidence in managing child distress (Cao et al., 2022).

Our findings also suggest that mothers’ childhood experiences of non-supportive emotion socialization contribute to the development of their insecure attachment styles as adults (Leerkes, 2011; Magai et al., 2004; McElwain & Booth-LaForce, 2006). Such attachment insecurity may further diminish their confidence in coping with their own children’s negative emotions after becoming mothers, especially under attachment-threatening contexts that are defined by high levels of child distress (Cao et al., 2022; Jones et al., 2015). Such a risk chain or cascade is implicated in the more extended series of “interlocking and shifting gears” that constitute the vicious intergenerational transmission cycles of maladaptive parenting (Kerr & Capaldi, 2019) and attachment insecurity (van Ijzendoorn & Bakermans-Kranenburg, 2019). Practically, this process-oriented delineation helps identify modifiable targets for interventions that are aimed at promoting auspicious dynamics across generations, given the increased malleability of parenting self-efficacy, attachment style, and emotion-related capacities during the transitional life periods such as the transition to parenthood (Arriaga et al., 2021; Gilkerson et al., 2020; Rutherford et al., 2015; Simpson et al., 2003). Theoretically, our findings provide support to a core proposition in Bowlby’s attachment theory that children develop their internal working models primarily out of early interactions with caregivers, especially parental responses to child distress cues or safety bids. In addition, our findings extend the Eisenberg et al.’s model of emotion socialization (Eisenberg, Cumberland et al., 1998) by suggesting that parents’ emotion socialization experiences in childhood may also shape their self-efficacy in managing child distress. This link may serve as a key mechanism that helps bridge the Eisenberg et al.’s model and the voluminous research focusing on the intergenerational transmission mechanisms of parenting and attachment (Kerr & Capaldi, 2019; van Ijzendoorn & Bakermans-Kranenburg, 2019).

Notably, a unique characteristic of the current study is considering AAI-based and self-reports of adult attachment in tandem in a single model, which echoes the long-standing calls from prominent attachment theorists (e.g., Jones et al., 2015; Shaver et al., 2000). In line with prior research (e.g., Roisman et al., 2007), we found “trivial to null” correlations between the AAI-based attachment representations (either categorical or continuous) and self-reports of attachment styles, suggesting a limited overlap between adult attachment security assessed with the two approaches. Interestingly, it is the self-reported attachment style (rather than the AAI-based attachment) that was found to mediate the link between mothers’ remembered childhood experiences of non-supportive emotion socialization and their self-efficacy in managing child distress, but only under attachment-threatening contexts that are defined by high levels of child distress. As Roisman et al. (2007) noted, “evidence that self-reported attachment style (but not the AAI) is associated with interpersonal behavior principally under conditions of attachment-related threat speaks to the fidelity of the social psychological measurement tradition with Bowlby’s account of the conditions under which insecurity would be expected to be revealed in interpersonal behavior.” (p. 694). A distressed child seeking comfort can be viewed as an attachment-related threat stimulus for parents. The assumption underlying the self-report measures of adult attachment is that the influences of attachment working models on attachment-related behaviors are most significant under the conditions of stress or threat. Thus, it is not surprising to find that the self-reported attachment style is more relevant for our model.

Nevertheless, some alternative explanations and/or future research directions merit attention. First, the shared-method variance between self-reported attachment and self-reported parenting self-efficacy may inflate their association. Second, prior studies have indicated high continuity (e.g., the overall mean test–retest correlation was .72; Stern et al., 2018) and relative malleability (e.g., decline in the levels of self-rated attachment insecurity due to spousal support; Rholes et al., 2021; Simpson et al., 2003; Spieker et al., 2011) of maternal adult attachment style/representation over the transition to parenthood. Future research may benefit from testing if the changes in self-reported and interviewed-based adult attachment security may play distinct roles in predicting parenting self-efficacy and other parenting outcomes. Relatedly, a study by Arriaga et al. (2021) suggests that first-time parents’ self-reported attachment tendencies might change as they derive personal competence and self-efficacy from their new parenting roles, highlighting the necessity of addressing the temporal dynamics of the link between new parents’ adult attachment style and parenting self-efficacy over time.

Last, secure base script knowledge (SBSK) as an alternative approach to using the AAI in assessing adult attachment representations should be acknowledged. Specifically, “those who receive consistent sensitive and competent care abstract the central features of those experiences to form a secure base script summarizing the temporal-causal sequence of effective secure base use and support.” (Waters & Roisman, 2019; p. 162). As Mikulincer and Shaver (2022) stated, once activated a secure base script serves as a guide for a person to adaptively regulate cognitive and affective processes when coping with stress and distress. SBSK has been shown to hold important implications for maternal ability to effectively manage emotional responding to infant distress (e.g., Groh et al., 2015). Unfortunately, data on SBSK were currently unavailable in the used dataset. Future research likely benefits from scoring SBSK from AAI transcripts and testing its potential roles in our model.

Applied Implications

Findings of this study provide some implications for interventions seeking to ease the stressful transition to parenthood for primiparous mothers. First, it is important to pay attention to new mothers who experienced higher non-supportive parental emotion socialization in childhood, because these mothers are more likely to be insecurely attached as adults and likely lack confidence in managing their own children’s distress. Accordingly, to better inform early parenthood prevention and intervention, assessments of mothers’ childhood (emotion-related) parenting experiences should be systematically conducted before the arrival of the infant to more timely identify individuals who are more susceptible to low self-efficacy in emotion-related parenting.

Practitioners should attend to the needs of mother-child dyads in which mothers carry the burdens of attachment insecurity, especially high avoidance, and children have high negative emotionality. Mothers in such dyads likely face difficulties in managing child negative emotions. For avoidant mothers, high child negative affectivity may be perceived as particularly aversive. As noted earlier, these mothers typically have some unique attachment-related vulnerabilities and tend to engage in various defensive mechanisms in the face of threats that they might develop out of early experiences of parental rejection or neglect about their bids for comfort when distressed. Increased activation of an avoidant mother’s attachment needs in the face of child distress may in turn interfere with the adaptive operation of their caregiving system (Jones et al., 2015; Mikulincer & Shaver, 2019a). Such disruptions likely foster mother’s sensitization to infant distress, escalate their negative emotions, and increase their perceived powerless about control over parenting performance.

In relevant counseling and intervention practice, educating new mothers to become aware of these otherwise “unconscious” processes may help them better understand their cognitions, emotions and behaviors when coping with child distress. Further, on the basis of such education, further training mothers to appropriately regulate their emotions and behaviors in the face of child negative emotions should be more effective in improving the quality of early emotion socialization, which may facilitate child social-emotional adjustment in the long run. Notably, families headed by insecurely attached parents are likely to benefit most from practice of this type as it helps diminish the vicious intergenerational transmission cycle of non-supportive emotion socialization and attachment insecurity (Leerkes et al., 2020; van Ijzendoorn & Bakermans-Kranenburg, 2019).

Limitations and Future Directions

Limitations of the present study should be noted. First, mothers provided retrospective reports of their childhood experiences. Research using retrospective self-report assessments of adults’ childhood experiences with their parents has yielded mixed results. On one hand, there appears to be notable evidence showing that adults’ recalled experiences in the early years of life are well predictive of their subsequent adaptions in various life domains (e.g., Chopik & Edelstein, 2019; Szepsenwol et al., 2015). Research has also indicated that adults’ reports of childhood parenting tend to be relatively stable over time and correlate significantly (yet modestly) with their parents’ reports, and more strongly with siblings’ reports (Brewin et al., 1993; Harlaar et al., 2008).

On the other hand, “memories may be subject to distortion, and events may be selectively recalled; early experiences may be forgotten, and perceptions of childhood events may be shaped by subsequent experiences” (Tajima et al., 2004, p. 424). In comparison to prospective measures, retrospective early experiences tend to show stronger associations with outcomes that were assessed with self-reports and weaker associations with outcomes that were more objectively measured (Reuben et al., 2016). Meta-analytic evidence indicated a low agreement between prospective and retrospective measures of childhood experiences (Baldwin et al., 2019). Notably, a recent study by Nivison et al. (2021) found that adults’ retrospective reports of childhood caregiving experiences were weakly associated with the actual caregiving measured in childhood, highlighting limitations in the utilization of retrospective self-reports of childhood parenting as a proxy for prospective data. Future efforts are pressing to address the sparseness of longitudinal data spanning from early childhood to adulthood. Our findings should be cautiously interpreted and await to be replicated with such data.

Relatedly, although our findings may, to some extent, have implications for understanding the intergenerational transmission mechanisms of emotion-related patenting, we could not provide direct evidence for such processes, given the retrospective assessment of mother’s childhood experiences. Future research may benefit from gaining access to datasets with direct assessments of early childhood experiences and subsequent parenting to test the replicability of the current findings, such as the Minnesota Longitudinal Study of Risk and Adaptation (MLSRA, see https://innovation.umn.edu/parent-child; Sroufe et al., 2005).

Second, fathers’ data for many participants were unavailable because fathers were absent in childhood. In addition, data for participants’ spouses were not collected in the larger project from which our sample were derived. However, according to the family systems perspective, there should be interdependence between fathers’ and mothers’ emotion-related parenting practices and self-efficacy (e.g., Blandon, 2015). The systemic complexities in the current model await to be addressed by research collecting data from both parents in the current families and in the families of origin.

Third, we acknowledge that a key limitation or alternative possibility for the current model is that mothers’ current states of mind with respect to attachment may influence how they recalled and reported on early experiences of caregiving and personal vulnerabilities. In particular, insecure adults may have particular biases, such that preoccupied/anxious adults tend to over-report negative experiences and dismissing/avoidant adults tend to underreport such experiences (Dykas et al., 2014; Mikulincer et al., 2009). When interpreting our findings, such influences should be considered.

Last, mother’s global (parenting) self-efficacy or self-efficacy in other parenting domains were not assessed in the larger project. Thus, we cannot control for the relevant effects or address related questions. In particular, to increase predictive validity and guide the designs of more targeted and effective interventions, scholars have increasingly advocated for research investigating the origins of domain-specific versus domain-general parenting self-efficacy (e.g., Junttila et al., 2015). Thus, future research is needed to make a more nuanced differentiation in parenting self-efficacy in different domains (e.g., emotion socialization, cognitive stimulation, and routine care) and identify uniquely salient antecedents of parenting self-efficacy in specific domains.

Conclusion

Parenthood is emotionally demanding. Coping with young children’s negative emotions is one of the most challenging tasks, especially in toddlerhood characterized by more frequent and intense child negative affect in the context of normative autonomy struggles. This study represents an initial effort trying to identify critical predictors of maternal self-efficacy specifically in managing toddler distress. Our findings suggest that by disrupting the normal development of attachment, parents’ childhood experiences of non-supportive emotion socialization hold long-standing consequences for their subsequent confidence in managing negative emotions of their own children. Further, child distress as a threatening stimulus may contextualize such effects through interfering with the operation of caregiving system, especially for avoidant parents. Built on the current study, more research is needed to obtain a thorough understanding of the origins of parental self-efficacy in managing child negative emotions. Such work would inform the designs of more targeted interventions aimed at improving the quality of early emotion socialization in the family setting, which may yield long-term development benefits by promoting child social-emotional functioning.