Understanding How and Why Effortful Control Moderates Children’s Vulnerability to Interparental Conflict

Abstract

This study examined the moderating role of effortful control in the association between interparental conflict and externalizing problems in a diverse sample of preschool children (N = 243; M age = 4.60 years). Using a multi-method, multi-informant, prospective design, findings indicated that the relation between interparental conflict and externalizing problems was only significant among children with poor effortful control. Children with high effortful control appeared to be protected against the negative effects of interparental conflict exhibiting low levels of externalizing problems despite increasing levels of interparental conflict. Towards identifying the mechanisms underlying the protective effects of effortful control, mediated moderation analyses indicated that children’s effortful control protects children against interparental conflict by reducing their angry reactivity to interparental conflict.

Qualitative and quantitative reviews demonstrate that children’s externalizing problems are one of the most consistent sequelae of destructive (i.e., hostile, aggressive) interparental conflict (Buehler et al., 1997; Grych & Fincham, 1990). Despite consistency in the findings, the relatively modest effect sizes underscore the substantial heterogeneity in the functioning of children exposed to high levels of interparental conflict (Buehler et al., 1997). Thus, a pressing task of a second generation of research is to identify sources of variability in the relationship between interparental conflict and externalizing problems (Cummings & Davies, 2002). Children’s effortful control is one promising mechanism that may elucidate the pathway between interparental conflict and externalizing problems. Effortful control is a temperamental self-regulatory capacity that involves inhibiting a reflexive dominant response (i.e., inhibitory control) and voluntarily enacting a subdominant response (i.e., activation control; Eisenberg, Smith, Sadovsky, & Spinrad, 2004; Rothbart & Bates, 2006). However, there is currently a paucity of research exploring the role of effortful control in understanding individual differences in children’s vulnerability to interparental conflict. To address this significant gap, the primary aim of this study was to test the relative viability of effortful control in its role as a possible moderator of the prospective association between interparental conflict and children’s externalizing problems. Consistent with protective models (Aiken & West, 1991; Luthar, Cicchetti, & Becker, 2000), we tested the hypothesis that effortful control would buffer children from the risks of destructive interparental conflict. To understand why effortful control acts as a moderator in contexts of interparental conflict, we next examined whether children’s angry reactivity to interparental conflict mediated the moderating role of effortful control in prospective associations between interparental conflict and children’s externalizing problems.

Effortful Control as a Moderator of Interparental Conflict

As illustrated in Path 1 of the conceptual model in Figure 1, our first aim was to examine whether children’s effortful control served as a protective factor in the association between their exposure to interparental conflict and their externalizing symptoms. Consistent with this premise, conceptual models have repeatedly cast children’s temperamental attributes as moderators of their vulnerability to interparental conflict based on their core features as temporally and situationally enduring behavioral dispositions (e.g., Cummings & Davies, 2002; Grych & Fincham, 1990; Harold & Sellers, 2018). As a central feature of temperament, effortful control is characterized as an early emerging and relatively stable behavioral attribute with constitutional roots (Rothbart, 1989; Rothbart & Derryberry, 1981). Thus, although effortful control may change in modest ways as a function of children’s socialization experiences (Eisenberg, Taylor, Widaman, & Spinrad, 2015; Kochanska, Murray, & Harlan, 2000), the enduring nature of individual differences in effortful control may protect children from repeated exposure to family adversity. For example, individual differences in effortful control are moderately stable during early and middle childhood periods and are associated with better anger regulation in stressful contexts (Eisenberg et al., 2015; Kochanska et al., 2000; Murphy, Eisenberg, Fabes, Shepard, & Guthrie, 1999). Moreover, conceptual and empirical work has delineated effortful control as a buffer for children who experienced high levels of parenting difficulties (e.g., Kiff, Lengua, & Zalewski, 2011; Morris et al., 2002). Thus, effortful control may prove particularly beneficial in protecting children in the context of interparental conflict by promoting strategies to manage reflexive, intense negative emotions (e.g., emotion-focused coping) and reduce emotional arousal and reactivity to family conflict (Repetti, Taylor, & Seeman, 2002).

Figure 1.

Conceptualization of children’s angry reactivity to interparental conflict as a mediator of the interaction between interparental conflict and effortful control in predicting children’s externalizing problems.

Despite the conceptual bases for hypothesizing that children’s vulnerability to family adversity may vary as a function of their temperamental regulation capacities, previous research examining children’s temperament as a moderator of interparental conflict has primarily examined temperamental reactivity (e.g., negative emotionality; Hentges, Davies, & Cicchetti, 2015; Pauli-Pott & Bechman, 2007). Although research has yet to systematically test effortful control as a source of heterogeneity in the risk experienced by children exposed to high levels of interparental conflict, two studies provide some preliminary support for the hypothesis. Davies and Windle (2001) demonstrated that among adolescents exposed to high levels of marital discord, those exhibiting good task orientation (i.e., attentional control) had significantly reduced rates of delinquency in comparison to adolescents with poor task orientation. However, attentional control is only one property of effortful control. Thus, as the only study to directly test effortful control as a moderator, David and Murphy (2007) showed that interparental conflict was associated with problematic peer relationships only for children who exhibited lower levels of effortful control. However, these findings were limited by reliance on a cross-sectional design, a single method (i.e., questionnaire) for assessing interparental conflict and effortful control, and a small (n = 62) and a predominantly White sample. Thus, to address existing gaps, the first aim of this study was to replicate and extend previous tests of effortful control as a moderator of interparental conflict by providing a more rigorous test of the moderating role of effortful control in a longitudinal model of interparental conflict and externalizing symptoms among a large sample of children from diverse demographic backgrounds.

Children’s Angry Reactivity to Interparental Conflict as a Mediating Mechanism

Examining the moderating role of effortful control is only the first step to understanding the heterogeneity of outcomes of children exposed to interparental conflict. Progress in characterizing the moderating role of effortful control can be facilitated by further examining why associations between interparental conflict and children’s externalizing problems vary across different levels of effortful control. Therefore, building off previous research examining the moderating role of effortful control in contexts of interparental conflict, the next step is to highlight mechanisms that are more immediate products of the interaction between interparental conflict and effortful control and, in turn, act as antecedents of children’s externalizing problems. As denoted by the mediated moderation model in Paths 2a and 2b of Figure 1, our second objective was to examine whether children’s angry reactivity to interparental conflict mediated the interaction between interparental conflict and effortful control in a longitudinal model predicting children’s externalizing problems.

According to risky family process models (Repetti, Robles, & Reynolds, 2011; Repetti et al., 2002), high conflict homes are proffered to increase children’s behavioral problems by sensitizing them to respond to interpersonal stressors in hostile ways. Exacerbating children’s angry hyperresponsivity, high conflict homes are further posited to constrain templates for understanding interpersonal problems and enacting constructive coping responses. In delving deeper into possible underlying processes in this cascade, social learning theory models propose that observational and enactive learning in high conflict homes promotes the development of children’s angry reactivity to emotionally arousing situations (Davies & Sturge-Apple, 2014; Eron, Huesmann, & Zelli, 1991; Margolin, Oliver, & Medina, 2001). In the observational or vicarious learning process component of the theory, exposure to hostile conflicts between parents increases children’s tendencies to respond in hostile ways to parental discord by modeling their destructive behaviors and reducing their inhibitions about aggressing. According to the enactive learning component of social learning theory (Emery, 1989), children’s hostile reactivity in subsequent contexts of conflict functions to distract parents from engagement in ongoing conflicts. Thus, by reducing exposure to the aversive nature of parental conflicts, children’s angry, coercive reactivity to conflicts is proposed to further intensify through negative reinforcement contingencies.

In highlighting the latter parts of this intergenerational cascade of hostility, risky family process and social learning theories share the premise that children’s tendencies to respond to interparental conflict in angry and hostile ways intensify and broaden into externalizing symptoms (Eron et al., 1991; Margolin et al., 2001; Repetti et al., 2002; 2011). In supporting this hypothesis, research has shown that children’s angry reactivity to conflict in the family is associated with their greater behavior problems. For instance, aggressive behavior and affect in response to family conflict was associated with increases in aggressive functioning a year later (Davis, Hops, Alpert, & Sheeber, 1998). Similarly, preschool children’s coercive patterns of responding to interparental conflict predicted increases in their externalizing problems one year later (Davies, Martin, Sturge-Apple, Ripple, & Cicchetti, 2016). Thus, children’s angry reactivity to interparental conflict has been repeatedly identified as a precursor of their externalizing symptoms.

Given the empirical and theoretical bases for expecting that children’s angry reactivity to interparental conflict may underpin the association between interparental conflict and externalizing problems, it is possible that effortful control may protect children exposed to interparental conflict from developing behavior problems by promoting effective coping strategies, such as anger regulation. In support of this hypothesis, previous research has reported associations between children’s self-regulation and adaptive responding to challenging interpersonal situations. For instance, Denham and colleagues (2014) reported that preschool children’s self-regulation was associated with subsequently lower levels of angry emotional and aggressive behavioral responses. Therefore, children with high levels of effortful control who are more skilled at managing intense negative emotions may be better able to suppress the reflexive, dominant aggressive behaviors and angry emotions stemming from exposure to family conflict and enact more deliberate and effective ways of managing their emotional arousal (Denham et al., 2014; Repetti et al., 2002). Coupled with prior documentation of children’s angry reactivity to interparental conflict as a precursor of externalizing problems, these findings highlight the plausibility that effortful control may serve as a protective factor in associations between interparental conflict and externalizing problems by enhancing their ability to regulate their angry and aggressive responding to family conflict.

Developmental Considerations

Based on several developmental considerations, our mediated moderation model was tested in a sample of children transitioning from preschool to the early school years. First, young children exposed to interparental conflict may be especially likely to engage in externalizing problems. In contrast to older children, young children are more likely to express their distress in the form of externalizing problems rather than internalizing behaviors (Glasberg & Aboud, 1981; 1982). This is particularly concerning because heightened behavioral problems during this developmental period have lasting implications on psychological adjustment observed throughout adolescence and young adulthood (Bongers, Koot, van der Ende, & Verhulst, 2003; Kouros, Cummings, & Davies, 2010). Second, as children transition into the early school years, they exhibit increases in sensitivity and negative affect in response to adult anger (Cummings, 1987; Cummings & Cummings, 1988). Third, complementary developmental processes operating during the preschool period may bolster the protective functions of effortful control in adverse socialization contexts. Significant normative improvements in children’s effortful control capacities across the preschool years indicate that it is an emerging developmental skill available for children to utilize in regulating their emotions in stressful family contexts (Jones, Rothbart, & Posner, 2003; Kochanaska et al., 2000). However, against this backdrop of normative change, there is substantial stability in individual differences in children’s effortful control abilities during preschool (Kochanska & Knaack, 2003; Kochanska, Murray, & Coy, 1997; Murphy et al., 1999). Thus, the role of effortful control as a moderator of interparental conflict may be particularly pronounced by virtue of the developmental salience (i.e., absolute change) and the enduring nature of individual differences (i.e., differential stability) in effortful control capacities during preschool.

Present Study

This study employs a multi-method (i.e., observation, interview, survey) and multi-informant (i.e., observer, mother, partner, teacher) design to reduce the operation of common method and informant variance. As delineated in Path 1 of Figure 1, our first aim of the study was to provide the first prospective test of effortful control as a moderator of children’s vulnerability to interparental conflict. Addressing limitations of previous cross-sectional work, this study utilizes a longitudinal, autoregressive design across three waves of data each spaced one year apart. This approach allows for the determination of whether the interaction between interparental conflict and effortful control during preschool predicts autoregressive change in children’s externalizing problems over a two-year period from preschool to first grade (Aiken & West, 1991; Selig & Little, 2012). Therefore, in drawing from the existing theoretical and empirical literature addressing the protective role of effortful control, we specifically tested the hypothesis that effortful control serves as a protective factor in the prospective association between interparental conflict and externalizing problems.

As denoted by Paths 2a and 2b in Figure 1, our second aim was to build on previous research on the moderating role of effortful control in the context of interparental conflict by testing a process model of why effortful control moderates associations between interparental conflict and children’s externalizing problems. Guided by risky family process and social learning theory models, we specifically examined whether children’s angry reactivity to interparental conflict mediates the interaction between interparental conflict and effortful control in the prediction of externalizing problems. To overcome the serious limitations (i.e., substantially biased parameter estimates) of cross-sectional analyses of mediation (Cole & Maxwell, 2003; Maxwell & Cole, 2007), our prospective analysis estimated successive change at each step of the mediational chain to delineate the prospective cascade of children’s externalizing problems over a two-year period. Thus, we specifically examined whether the interaction between interparental conflict and effortful control during preschool predicted change in externalizing symptoms in first grade through the mediational role of angry reactivity to interparental conflict during kindergarten in a fully lagged autoregressive design that controlled for angry reactivity and externalizing symptoms in preschool.

Method

Participants

Two hundred and forty-three families (i.e., mother, partner, and child) from a moderate-sized metropolitan area in the Northeast participated in our study. Families were recruited through multiple agencies including local preschools, Head Start programs, and public and private child care centers to obtain a diverse demographic sample. The longitudinal design included three annual measurement occasions, with a retention rate of 97 and 94% across contiguous waves of data collection. At Wave 1, children were on average 4.60 years old (SD = .44, range = 4 to 5 years old), and 56% of children were girls. Approximately half of the families identified as Black (48%), and the remainder identified as White (43%), multiracial (6%), or another race (3%). Sixteen percent of family members identified as Latino. Most families received public assistance (69%) and the median household income was $36,000 per year (range = $2,000 –$121,000). Parents’ median education was a GED or high school diploma. Almost half of the parents were married (47%) and, on average, parents had lived together for 3.36 years. The majority of parents (99% of mothers and 74% of fathers) were the child’s biological parents.

Procedures and Measures

Families visited our research laboratory at three waves of data collection each spaced one year apart. Experimenters provided rides to and from the research center for families without feasible transportation. All research procedures were approved by the Institutional Review Board at the University of Rochester prior to conducting the study (Title: “Children’s Development in the Family”; Approval #: 00030261). Remuneration for participation consisted of monetary compensation for families (i.e., $200, $240, and $280 at Waves 1, 2, and 3, respectively), small gifts for children (value ≤ $10), and $20 gift certificates for teachers who completed the questionnaire.

Interparental Conflict.

At Wave 1, parents participated in a 10-minute interparental problem solving task in which they were instructed to discuss common, problematic disagreements in their relationship (Gordis, Margolin, & John, 2001; Grych, 2002). Prior to children entering the room, parents selected multiple problematic issues to discuss allowing them to move on to another topic if they finished discussing a previous topic during the task. Parents could address any problematic issues that they were comfortable discussing in front of their children. Parents’ conflictual interactions began once the experimenter brought the child into the room, introduced them to a set of toys, and departed the room. Interaction were video-recorded for later coding.

Consistent with previous research (e.g., Coe, Davies, & Sturge-Apple, 2015), trained raters assessed each the mother and partner for specific behaviors during the conflict task along a 9-point continuous scale ranging from 1 (Not at all characteristic) to 9 (Highly characteristic). Scales included Anger, Aggression, Behavior Disorganization, and Controlling Behavior. Anger assesses the extent to which each partner displays signs of tension, frustration, irritation, fury, or anger. Aggression assesses the extent to which each partner uses harmful verbalizations of behavioral displays that are intended to harm the partner either psychologically or physically. Behavior Disorganization assesses affect expressions and conflict tactics that are volatile, unpredictable, and chaotic. Controlling Behavior assesses the extent to which each partner complains, protests, or dogmatically, selfishly, and rigidly asserts their viewpoint in a way that is intended to be coercive, controlling and/or aversive. Partners each received an overall score for each of the four scales. Intraclass correlation coefficients (ICC) were calculated to index interrater reliability based on two coders’ independent ratings of 20% of the videos. ICCs were as follows for mothers and partners, respectively: anger = .84 and .83; aggression = .85 and .83; behavior disorganization = .77 and .61; controlling behavior = .71 and .86. Scores on the eight scales were standardized and averaged together into a single indicator of interparental conflict for the analyses (α = .81). Although one ICC fell below .70, it is not uncommon for the higher-order composite to evidence good reliability when ICCs for ratings fall in the .50 and .60 range (e.g., Sohr-Preston et al., 2013).

Effortful control tasks.

We obtained observational assessments of children’s effortful control in three video-recorded tasks at Wave 1. First, in the Peg Tapping task, children were instructed to enact the rule of tapping a peg once on the table when the experimenter tapped it twice and vice versa over 16 trials (Bierman et al., 2008; Diamond & Taylor, 1996). The number of correct responses to the Peg Tapping task over 16 trials was designed to assess children’s ability to suppress the prepotent tendency to mimic the experimenter while following instructions to purposefully employ the opposite response. Thus, higher scores indicated higher effortful control. The Peg Tapping task has previously been utilized in preschool samples (Davies, Hentges, & Sturge-Apple, 2016; Rhoades, Greenberg, & Domitrovich, 2009).

Second, in the Black Boxes task, children were instructed to guess what objects were inside the three black boxes by touching, but not looking, at the objects (Van Brakel et al., 2004). Not only did the game consist of having children listen to the instructions quietly while standing against the wall, but it also required children to follow multiple directions during the game including approaching the boxes in a specific order, inhibiting impulses to peek in the boxes, and refraining from pulling the objects out of the box. To obtain a measure of effortful control, trained coders’ rated children’s impulsivity along a 9-point continuous scale ranging from 1 (No impulsivity) to 9 (Intense impulsivity). Scores were reverse scored so that higher scores reflected lower impulsivity and greater effortful control (Masked Author Citation; Olson, Shilling, & Bates, 1999). Thus, high effortful control was characterized by the ability to inhibit reflexive impulses (e.g., approaching the boxes during the instructions, looking at the objects in the boxes, attempting to pull the objects out of the boxes) while enacting contextually appropriate responses of complying with the rules of the game (e.g., following the ordering of the boxes, identifying the objects based on touch). The Black Boxes task has previously been utilized in preschool samples (Davies et al., 2016; Vreeke et al., 2012). Two coders independently overlapped in rating 30% of the videos (ICC = .94).

Third, in the Gift Delay task, children were instructed to refrain from enacting an impulse to open a gift until after the experimenter retrieved a bow from another room (Aksan & Kochanska, 2004). Consistent with previous approaches, we obtained molar ratings of children’s effortful control (e.g., Spinrad, Eisenberg, & Gaertner, 2007). Trained coders rated children’s effortful control along a 9-point dimensional scale ranging from 1 (No effortful control) to 9 (Intense effortful control). Behaviors reflecting no effortful control included standing and peeking into the bag or touching the gift in the bag within one minute of being left alone, or pulling the toy out of the bag during the delay period. Behaviors reflecting intense effortful control included the ability to inhibit impulses to move toward the bag, touch the bag, peak at the gift, or pull the gift out of the bag. Thus, high scores reflected the ability of children to successfully resist the temptation to open the gift (i.e., rewarding stimulus) while following the instructions to sit and wait until the experimenter returned before pulling the gift out the bag. The Gift Delay task has previously been utilized in preschool samples (Allan & Lonigan, 2011; Lengua, Honorado, & Bush, 2007; Murray & Kochanska, 2002). Two coders independently overlapped in rating 100% of the videos (ICC = .94).

In addition to the three observational assessments, the experimenter responsible for the child’s activities during the visit completed the California Child Q-set (CCQ; Block, 2008). Experimenters categorized 100 CCQ descriptors of children’s psychological adjustment into nine piles ranging from “extremely uncharacteristic” to “extremely characteristic.” Raters recorded children’s functioning based on an average of five cumulative hours of contact with children spanning visit tasks, transition periods, and transportation between homes and the research center (SD = 1 hour; Range = 3 to 8 hours). Consistent with previous research (e.g., Davies et al., 2016), effortful control was assessed using the nine-item CCQ Conscientiousness scale measuring children’s planful and organized behaviors (e.g., “Is attentive and able to concentrate,” “Is planful, thinks ahead”; α = .81). The CCQ conscientiousness scale has previously been utilized in preschool samples (Abe, 2005; Zastrow, Martel, & Widiger, 2018) and as a measure of effortful control in preschool samples (Davies et al., 2016). In support of the empirical overlap of the effortful control measures, a principal components analysis of the four measures with a varimax rotation produced a one factor solution based on parallel analysis and accounted for 53.25% of the variance. Therefore, consistent with previous approaches to assessing effortful control (e.g., Kochanska & Knaack, 2003), the Peg Tapping task, Black Boxes task, Gift Delay task, and experimenter Q-sort ratings were standardized and averaged together to create a composite measure of effortful control (α = .70).

Children’s angry reactivity to interparental conflict.

To obtain a measure of children’s angry reactivity to interparental conflict at Waves 1 and 2, trained experimenters administered the Interparental Disagreement Interview (IDI) to mothers. The IDI is a semi-structured, narrative interview designed to assess children’s behavioral and emotional reactivity during and following common, intense conflicts between parents (Davies, Cicchetti, & Martin 2012). Mothers were presented with a series of open-ended questions regarding children’s behavioral (e.g., “During these disagreements that [child] sees or hears, how does s/he respond?”) and emotional (e.g., “How do you think [child] feels during these disagreements?”) reactivity to interparental conflict. Interviews were video-recorded for later coding.

Indices of children’s angry and coercive reactivity were quantified from maternal narratives using three dimensions. As the first indicator of children’s angry reactivity to interparental conflict, trained raters coded for the presence (1) or absence (0) of four specific forms of angry and controlling behaviors during interparental conflict: (1) Anger, descriptions of facial expressions, behaviors (e.g., clenching fists, stomping feet), and verbalizations (e.g., yelling) reflecting frustration, irritation, and ire; (2) Aggression, indices of verbal or physical hostility directed to the self or others (e.g., hitting, yelling, belittling); (3) Involvement, reflected children’s bold intervention efforts to control the course of the conflict (e.g., getting physically in the middle of the disagreement, telling parents to stop); and (4) Triangulation, descriptions reflecting children’s attempts to form alliances or coalitions with one parent against the other (e.g., taking sides with one parent). Ratings of each dimension were summed together to create a behavioral count of angry reactivity to interparental conflict at Waves 1 (α = .64) and 2 (α = .63). Tallies ranged between 0 and 4. Intraclass correlation coefficients, based on coders who overlapped on at least 20% of the videos at each wave, were .91 at Wave 1 and .93 at Wave 2.

As the second indicator of children’s angry reactivity to interparental conflict, coders rated children’s hostile reactivity to interparental conflict along a 7-point dimensional scale based on maternal descriptions of multiplicity, intensity, and patterning of their coercive and angry behaviors. Low scores (i.e., 0 = “None”) on the angry reactivity code were reserved for interviews in which there was no indication that the child experienced anger, frustration, irritation, or aggression). In contrast, high levels of angry reactivity (i.e., 6 = “High”) reflected maternal portrayals of their children who exhibit highly intense anger expressions that are highly coercive, prolonged, and dysregulated. To obtain interrater reliability, multiple coders provided independent ratings of at least 20% of the videos at each wave. The resulting ICC values were .79 at Wave 1 and .92 at Wave 2.

For the third indicator, trained coders provided molar ratings of children’s dominant patterns of conflict reactivity based on maternal descriptions of their behavior during and immediately following interparental conflicts. Dominant patterns of responding were defined as strong tendencies to control parental interactions through angry, domineering, and coercive means. Coders rated children’s dominant response patterns along a 5-point continuous scale ranging from 1 (Not at all characteristic) to 5 (Highly characteristic). Intraclass correlation coefficients, based on independent coders overlapping on at least 20% of the videos at each wave, were .80 and .90 at Waves 1 and 2, respectively. The three resulting measures were specified as manifest indicators of children’s angry reactivity at each wave. Previous research has documented that coder ratings of maternal narratives from the IDI provide psychometrically sound measures (e.g., predictive validity) of children’s emotional (e.g., angry) reactivity in other samples of young children (e.g., Hentges et al., 2015; Manning, Davies, & Cicchetti, 2014).

Children’s externalizing problems.

We utilized a multi-informant approach to assess children’s externalizing problems at Waves 1 and 3. First, both mothers and partners completed five externalizing subscales from the parent version of the McArthur Health and Behavior Questionnaire (HBQ; Ablow et al., 1999). Each HBQ item is rated along a 3-point scale ranging from 0 (Never or not true) to 2 (Often or very true). The subscales consisted of overt hostility (four items; e.g., “Gets in many fights”), conduct problems (12 items; e.g., “Lies or cheats”), ODD symptoms (nine items; e.g., “Argues a lot with adults”), inattention (six items; e.g., “Distractible, has trouble sticking to any activity”), and impulsivity (nine items; e.g., “Has difficulty awaiting turn in games or groups”). Consistent with HBQ guidelines (Ablow et al., 1999; Armstrong, Goldstein, & The MacArthur Working Group on Outcome Assessment, 2003), items from five subscales consisting of overt hostility, oppositional defiant disorder (ODD) symptoms, inattention, and impulsivity were averaged together to obtain maternal and partner reports of children’s externalizing problems at Waves 1 and 3 (α range from .87 and .91). In addition, teachers completed the same five externalizing subscales from the teacher version of the HBQ using the same response alternatives as the parent version. The five teacher subscales were averaged together to create the third indicator of a latent construct of children’s externalizing problems at Waves 1 and 3 (α = .91 at each wave).

Covariates.

Due to documented associations with child adjustment, planned covariates include: (a) child sex (1 = girls; 2 = boys); (b) household income per capita, calculated as the ratio of total annual family income relative to the number of individuals living in the home; and (c) children’s internalizing problems. We utilized a multi-informant approach to assess internalizing problems at Wave 3. Due to low correlations with teacher report of internalizing problems, we created a composite utilizing mothers’ and partners’ reports of the depression (seven items; e.g., “Feels worthless or inferior”; maternal α = .66; partner α = .72), overanxious (12 items; “Worries about things in the future”; maternal α = .79; partner α = .77), and withdrawal (nine items; “Likes to be alone”; maternal α = .77; partner α =.70) subscales from the parent version of the McArthur Health and Behavior Questionnaire (HBQ; Ablow et al., 1999). Each HBQ item is rated along a 3-point scale ranging from 0 (Never or not true) to 2 (Often or very true). Items from the six subscales were averaged together to obtain a composite of maternal and partner report of children’s internalizing problems at Wave 3 (α = .76).

Results

Table 1 provides the means, standard deviations, and correlations among the study’s primary variables. Structural equation modeling (SEM; McArdle, 2009) was used to test the moderating role of effortful control as a moderator of prospective associations between interparental conflict and children’s externalizing symptoms and the role of children’s angry reactivity to interparental conflict as a mediator of the moderating effects. Maximum likelihood approaches to estimating data produce accurate parameter estimates for all types of missing data when the percentage data missing is below 20% (Schlom er, Bauman, & Card, 2010). Given that missing data in our sample met this criterion (i.e., M = 8.9%), we used full information likelihood (FIML) in AMOS 25.0 (Arbuckle, 2017) to estimate missing data and retain the full sample in our analyses.

Table 1.

Means, standard deviations, and correlations among the variables in the primary analyses of the study.

	Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
Covariates
1. Child Sex	--	--	--
2. Socioeconomic Status	9.60	6.22	.03	--
3. Internalizing Problems	.33	.20 −	.01	.13	--
Wave 1 Primary Predictors
4. Interparental Conflict	0.00	1.56 −	.04	−.03	−.14*	--
5. Effortful Control	0.00	0.77 −	.19*	.04	.24*	−.03	--
6. Interparental Conflict × Effortful Control	−0.03	1.16 −	.13	−.13	−.01	.08	−.04	--
Wave 1 Angry Reactivity to Interparental Conflict
7. Angry Behavioral Count	1.34	1.16	.14*	−.05	.01	.14*	−.01	.05	--
8. Hostile Reactivity	2.25	1.09	.14*	.00	−.12	.13*	−.09	.03	.78*	--
9. Dominant Pattern	1.70	1.82	.11	−.01	−.04	.12	−.04	.02	.76*	.82*	--
Wave 2 Angry Reactivity to Interparental Conflict
10. Angry Behavioral Count	1.57	1.07	.07	−.07	−.02	.12	−.08	−.10	.42*	.38*	.42*	--
11. Hostile Reactivity	2.02	1.48	.05	−.03	−.12	.20*	−.10	−.12	.31*	.35*	.39*	.76*	--
12. Dominant Pattern	1.72	0.99	−.02	−.03	−.10	.14*	−.09	−.12	.30*	.30*	.33*	.77*	.86*	--
Wave 1 Externalizing Problems
13. Mother Report	0.41	0.27	.12	.35*	−.04	.10	−.15*	−.10	.09	.09	.06	.04	.14*	.12	--
14. Teacher Report	0.31	0.37	.10	−.03	−.11	.00	−.34*	−.04	.08	.02	.01	−.00	.05	.06	.30*	--
15. Partner Report	0.41	0.27	.05	.26*	.00	−.04	−.14*	−.07	.08	.08	.09	−.03	.00	−.00	.37*	.28* --
Wave 3 Externalizing Problems
16. Mother Report	0.40	0.32	.11	.44*	−.04	.14*	−.22*	−.16*	.18*	.19*	.14*	.17*	.20*	.20*	.73*	.24*	.30*	--
17. Teacher Report	0.35	0.40	.04	.09	−.27*	.12	−.45*	−.00	.06	.13	.14	.10	.13	.19*	.29*	.43*	.17*	.39*	--
18. Partner Report	0.41	0.29	.06	.55*	−.02	.11	−.10	−.12	−.05	−.01	.01	−.01	−.00	−.00	.45*	.15	.57*	.41*	.22*

Note.

^*p < .05.

Primary Analysis I: Effortful Control as a Moderator of Interparental Conflict

First, we examined whether the interplay of children’s effortful control and interparental conflict predicted children’s Wave 3 externalizing problems prior to the inclusion of children’s angry reactivity to interparental conflict as a mediator of the moderating effects of effortful control. We mean centered composites of interparental conflict and effortful control to reduce problems with multicollinearity in the moderator analyses (Marsh et al., Wen, & Hau, 2004). We specified the three predictors (i.e., interparental conflict, children’s effortful control, and their multiplicative interaction) as predictors of children’s Wave 3 externalizing problems. To increase methodological rigor, we estimated autoregressive paths between children’s Wave 1 and Wave 3 externalizing problems constraining factor loadings of each indicator to be equal over time. Preliminary correlational analyses revealed significant associations between children’s sex, internalizing problems, and socioeconomic status and variables in the primary analyses, thus we included children’s sex, internalizing problems, and socioeconomic status as covariates. We specified correlations: (a) between all exogenous variables and (b) between the residual errors of the same manifest indicators corresponding to the same reporter of externalizing problems across time. However, to reduce complexity, only significant correlations are depicted in Figure 2.

Figure 2.

Structural equation model examining the moderating effect of effortful control in the pathway between interparental conflict and children’s externalizing problems. Significant paths are bolded whereas nonsignificant paths are nonbolded. Parameter estimates for the structural paths are standardized path coefficients. Correlations were estimated among all predictors; however, only significant correlations are modeled. *p < .05. **p < .01.

The model for the moderating role of children’s effortful control, which is shown in Figure 2, provided a satisfactory fit with the data, χ² (31, N = 243) = 68.10, p <.01, root mean square error of approximation (RMSEA) = .07, comparative fit index (CFI) = .93, and χ²/df ratio = 2.197. Analysis of the main effects of the predictors revealed that both interparental conflict, β = .18, p <.01, and effortful control, β = −.17, p =.01 predicted children’s Wave 3 externalizing problems. Furthermore, the interparental conflict x effortful control interaction term predicted children’s Wave 3 externalizing problems even after including structural paths involving the other predictors and covariates, β = −.11, p =.03.

To further understand the form of the moderation, we calculated the graphical plot and simple slope analyses of the interaction from the parameter estimates of the full SEM model in Figure 2. As depicted in Figure 3, the graphical plot is consistent with conceptualizations of protective-stabilizing effects (Luthar et al., 2000). Whereas children low in effortful control demonstrated increases in externalizing problems with greater exposure to interparental conflict, children high in effortful control exhibited relatively stable and low levels of externalizing problems as interparental conflict increased. We probed the interaction at high and low (±1 SD of the M) levels of interparental conflict for children high and low (±1 SD of the M) in effortful control (Liu, Zhou, Wang, Liang, & Shi, 2018; Preacher, Curran, & Bauer, 2006). Confirming visual interpretations of the graphical plot, findings from simple slope analyses indicated that interparental conflict was significantly associated with children’s Wave 3 externalizing problems for children low in effortful control, b =.04, p <.01, but not for children high in effortful control, b = .003, p =.79.

Figure 3.

A graphical plot depicting the simple slopes of the interaction between interparental conflict and effortful control in predicting children’s Wave 3 externalizing symptoms, at ±1 SD of interparental conflict.

Primary Analysis II: Children’s Angry Reactivity to Interparental Conflict as a Mediator of the Moderating Effects of Effortful Control

Given the significant moderation findings in our first primary analysis, the second analytic step was designed to address the question of why the magnitude of interparental conflict as a predictor of children’s externalizing problems varied as a function of their effortful control. Next, we estimated a mediated moderation model to examine whether children’s angry reactivity to interparental conflict mediated the moderating role of effortful control in the pathway between interparental conflict and children’s externalizing symptoms. We utilized the mean centered composites of interparental conflict and effortful control and specified the same set of structural paths modeled in Figure 2. Additionally, we estimated paths between all predictors, covariates (except Wave 3 internalizing problems), and children’s Wave 1 externalizing problems and children’s Wave 2 angry reactivity to interparental conflict. In turn, children’s angry reactivity to interparental conflict was specified as a predictor of children’s Wave 3 externalizing problems. To increase methodological rigor, we estimated autoregressive paths between children’s Wave 1 and Wave 2 angry reactivity to interparental conflict constraining factor loadings for each manifest indicator to be equal over time. We also estimated correlations: (a) between all exogenous variables and (b) between the residual errors of the same manifest indicators of angry reactivity and externalizing symptoms across time. To reduce complexity, only significant correlations are shown in Figure 4.

Figure 4.

Structural equation model examining the mediating effect of children’s angry reactivity to interparental conflict in the pathway between the interaction between interparental conflict and effortful control in predicting children’s externalizing problems. Significant paths are bolded whereas nonsignificant paths are nonbolded. Parameter estimates for the structural paths are standardized path coefficients. Correlations were estimated among all predictors; however, only significant correlations are modeled. *p < .05. **p < .01.

The resulting model provided a good fit with the data, χ² (99, N = 243) = 154.70, p <.01, RMSEA =.05, comparative fit index (CFI) = .97, and χ²/df ratio = 1.58. Analysis of the main effects of the predictors revealed that interparental conflict predicted children’s angry reactivity to interparental conflict at Wave 2, β =.13, p =.04. Furthermore, the interparental conflict x effortful control interaction term predicted children’s Wave 2 angry reactivity to interparental conflict, β = −.17, p =.01. Children’s Wave 2 angry reactivity to interparental conflict, in turn, predicted children’s Wave 3 externalizing problems, β =.13, p =.02. Further supporting mediated moderation, bootstrapping tests using the RMediation software program revealed a significant indirect path involving the interparental conflict x effortful control interaction term, children’s angry reactivity to interparental conflict, and children’s Wave 3 externalizing problems, 95% CI [−0.0094, −0.0003] (Tofighi & MacKinnon, 2011).

Demonstrating that the interaction between interparental conflict and effortful control is comparable across models of children’s angry reactivity to interparental conflict and externalizing problems would provide additional evidence for the mediational role of angry reactivity to interparental conflict. Thus, we next characterized the moderating role of effortful control in the association between interparental conflict and children’s Wave 2 angry reactivity to interparental conflict using the parameter estimates from the full SEM model depicted in Figure 4. We probed the interaction at high and low (±1 SD of the M) levels of interparental conflict for children high and low (±1 SD of the M) in effortful control (Liu et al., 2018; Preacher et al., 2006). As depicted in Figure 5, the graphical plot is consistent with a protective-stabilizing form of moderation (Luthar et al., 2000). Whereas children low in effortful control demonstrated increases in angry reactivity to interparental conflict as their exposure to interparental conflict increased, angry reactivity to interparental conflict among children high in effortful control remained low as interparental conflict increased. Thus, the form of the interparental conflict x effortful control interaction was similar in the prediction of children’s angry reactivity to interparental conflict and externalizing problems. Second, consistent with the findings for externalizing problems, simple slope analyses indicated that interparental conflict was significantly associated with greater angry reactivity to interparental conflict at Wave 2 for children low in effortful control, b =.21, p <.01, but not for children high in effortful control, b = −.06, p =.33. Further support for angry reactivity to interparental conflict as a mediator of the moderating effects of effortful control is provided by the comparability of the graphical plots and simple slope analyses across externalizing problems and angry reactivity to interparental conflict.

Figure 5.

A graphical plot depicting the simple slopes of the interaction between interparental conflict and effortful control in predicting children’s Wave 2 angry reactivity to interparental conflict, at ±1 SD of interparental conflict.

Discussion

Although effortful control has previously been identified as a protective factor in contexts of interparental conflict, these findings were limited to a single cross-sectional study of a relatively small and predominantly White sample (David & Murphy, 2007). Therefore, our first aim was to provide a more authoritative test of the moderating effects of effortful control in the prospective link between interparental conflict and children’s externalizing problems among a diverse sample of families. The findings from the multi-informant, multi-method approach revealed that interparental conflict predicted greater externalizing problems over a two-year period for children with poor effortful control, but not for children with high effortful control. Our second aim was to address the process-oriented question of why high effortful control buffers children against the negative effects of interparental conflict. As hypothesized by multiple theories, interparental conflict uniquely predicted subsequent increases in children’s angry reactivity to interparental conflict which, in turn, was associated with their subsequent externalizing problems one year later. Disrupting this cascade, our mediated moderation analyses revealed that children with high effortful control were at a reduced risk for developing externalizing problems in the context of interparental conflict due to their decreased tendency to respond to interparental conflict in angry and aggressive ways.

In our first analytic step, moderation analyses revealed a heightened vulnerability to the effects of interparental conflict among children with poor effortful control. More specifically, the relation between interparental conflict and externalizing problems was only significant among children displaying poor effortful control with high levels of interparental conflict predicting greater externalizing problems. In building on these findings, effortful control may promote children’s ability to manage reflexive, intense negative emotions that help to reduce their reactivity to family conflict (Repetti et al., 2002). Accordingly, the null association between interparental conflict and externalizing problems among children with high effortful control suggests that effortful control is one potential factor that protects children against interparental conflict. Consistent with conceptualizations of protective-stabilizing effects (Luthar et al., 2000), high levels of effortful control appear to confer stability in externalizing problems despite increasing levels of interparental conflict.

To understand why effortful control buffers children against the negative effects of interparental conflict, we also aimed to identify more proximal mechanisms that are disrupted by effortful control in the cascade of risk leading to externalizing problems. Guided by models identifying children’s emotional and behavioral reactivity to interparental conflict as one possible mechanism underlying associations between interparental conflict and externalizing problems (Davies & Sturge-Apple, 2014; Eron et al., 1991; Margolin et al., 2001; Repetti et al., 2002; 2011), we examined children’s diminished levels of angry reactivity to interparental conflict as a mediator of the moderating role of effortful control. Consistent with conceptualizations of family risk and social learning theory models, we proposed that the protective role of effortful control in reducing children’s vulnerability to interparental conflict would first disrupt their angry reactivity to interparental conflict. Reductions in children’s tendency to respond to interparental conflict in angry and aggressive ways, in turn, were posited to reduce children’s vulnerability to externalizing problems. In supporting our hypothesis, the results indicated that children’s angry reactivity to interparental conflict accounted for the interplay between interparental conflict and effortful control in predicting externalizing problems over a two-year period.

In accord with the first proposed link in the mediational chain, the interaction between interparental conflict and effortful control significantly predicted children’s angry reactivity to interparental conflict one year later. More specifically, interparental conflict was only a significant predictor of angry reactivity to interparental conflict for children displaying poor effortful control. Thus, consistent with the protective-stabilizing form of moderation (Luthar et al., 2000), children with high effortful control evidenced low levels of angry reactivity regardless of whether they were previously exposed to high or low levels of interparental conflict. Interpreted within social learning theory (Eron et al., 1991; Margolin et al., 2001), it is possible that high effortful control may reduce the operation of two key classes of pathogenic processes in high conflict homes. First, in reflecting the vicarious learning process within the theory (Bandura, 1976; Cox, Paley, & Harter, 2001), repeatedly observing parents engage in angry conflicts is proposed to engender children’s reflexive tendencies to model hostile and coercive behaviors and increase their negative arousal in ways that lead to the disinhibition of aggression (Cummings & Zahn-Waxler, 1992). In this context, children with high effortful control may be better able to inhibit the automatic mimicry of hostile behaviors and regulate emotional arousal in ways that prevent disinhibition and dysregulation of behavior. Second, in highlighting negative reinforcement processes, social learning theory formulations have posited that propensities to respond to parental conflicts with progressively greater anger may be further intensified because it effectively detracts parents from their conflicts, thereby resulting in reduced exposure to the stressor (Cox et al., 2001; Emery, 1989; Patterson, 1982). Thus, through their effective suppression of disruptive behavioral responses to interparental conflict, children with high effortful control may halt the operation of negative reinforcement contingencies that function to intensify angry and aggressive behaviors.

Emotion contagion models may provide another plausible explanation for our mediated moderation findings (e.g., Morris, Silk, Steinberg, Myers, & Robinson, 2007). Emotion contagion is characterized by the reflexive mimicry of another leading to the matching of expressions, behaviors, and emotions (Hatfield, Cacioppo, Rapson, 1993; Saarni, Mumme, & Campos, 1998). Thus, children may “catch” or “mimic” parental displays of intense, negative emotions during conflict, thereby increasing their tendency to react in a similar manner. Consistent with this hypothesis, previous research has documented that children’s angry behavioral displays increase following repeated exposure to angry adult interactions (Cummings, Iannotti, & Zahn-Waxler, 1985). In further extending the implications of this work, our findings may reflect the possibility that effortful control disrupts this cascade of hostility. More specifically, children with high effortful control may be able to suppress the automatic tendency to match parental displays of anger expressed in interparental conflict.

In addition to its documentation as a moderator in pathways between interparental conflict, children’s angry reactivity to interparental conflict, and externalizing problems, our findings indicated that effortful control was a distinctive predictor of children’s externalizing problems two years later. Thus, these results are consistent with previous empirical documentation that children’s self-regulation is associated with diminished angry and aggressive behavior (Denham et al., 2014). Taken together, these studies indicate the important role that self-regulatory abilities may play in suppressing children’s reflexive angry reactivity in ways that may prevent damage to interpersonal relationships. Given the support for the dual role of effortful control as a protective factor and predictor, targeting children’s effortful control and their more general self-regulatory deficits may be a key area for intervention (Bierman et al., 2008; Liew, 2012). Importantly, these skills are beginning to develop as children transition into the early school years and have been demonstrated to not only have implications for children’s social competence, but also school adjustment and readiness, and behavioral adjustment (e.g., David & Murphy, 2007; Eisenberg et al., 2004; Murphy et al., 1999). Therefore, self-regulatory interventions may be a key factor in reducing children’s early adjustment problems that may be early outcomes in a cascading process leading to psychopathology.

Although we found support for the role of angry reactivity as a mediator of the interaction between interparental conflict and effortful control in a longitudinal model predicting externalizing problems, the relatively modest effect sizes for mediated moderation findings underscores the importance of expanding the search for mechanisms that underpin the protective effects of effortful control. For example, children’s processing and appraisals of conflict may be a complementary explanatory mechanism for the significant interaction between interparental conflict and effortful control (Grych & Fincham, 1990). Greater control capacities to enact voluntary, contextually nuanced responses to environmental stimuli may be manifested in the ability to shift attention away from the saliency of threatening stimuli in ways that limit the development of hostile encoding and appraisals of interpersonal relationships. As a result, effortful control may confer some of its protective effects by limiting the development of negative social representations that are precursors to the development of psychological problems. As an alternative plausible pathway, evocative processes may be operating whereby the cumulative experience of couples coping with high conflict in their relationship and the challenges of raising a child with poor effortful control may synergistically increase parenting difficulties (e.g., unresponsiveness, harshness, detachment) and, in turn, children’s externalizing problems. In support of this possibility, children’s poor effortful control and interparental conflict have each been identified as viable antecedents of parenting difficulties (e.g., Kiff et al., 2011; Krishnakumar & Buehler, 2008). Thus, examining whether parenting difficulties help to account for the disproportionate vulnerability of children who experience poor effortful control and elevated interparental conflict is an important future direction.

As another qualification to our findings, understanding how parental personality characteristics may be operating in the context of interparental conflict is an important future direction for research. For example, parents in highly conflictual relationships may be more likely to experience high levels of anger and poor regulation abilities. Thus, it is possible that the mediated moderation findings may be explained by the intergenerational transmission of a temperamental pattern of high anger and poor effortful control that, over time, lays the foundation for children’s externalizing problems (Harold, Elam, Lewis, Rice, & Thapar, 2012). However, genetic processes do not appear to fully explain the risk experienced by children from high conflict homes (e.g., Harold, Leve, & Sellers, 2017). Consistent with this research, the earlier emergence of anger as an inherited temperamental trait relative to effortful control would suggest that the personality characteristics of highly conflictual couples would first be evident in children’s angry reactivity to conflict. In contrast, our findings indicated that interparental conflict was associated with subsequent increases in angry reactivity only for children who previously experienced poor effortful control.

Discussion of other limitations and qualifications are also warranted in fully interpreting the findings of the study. First, despite a relatively diverse sample in terms of sociodemographic characteristics, the results of this study may not generalize to other samples (e.g., affluent families, clinical samples of children). Second, the analyses revealed the moderating role of effortful control in predicting children’s externalizing problems was modest in magnitude. Given that much of the heterogeneity in children’s functioning is not explained by effortful control, future research may benefit from examining other sources of variability (e.g., cognitive appraisals, sibling relationships, maternal sensitivity). Third, although our tests of effortful control as a protective factor for externalizing sequelae were guided by previous work, effortful control may also moderate other forms of vulnerability and functioning experienced by children from high conflict homes (e.g., internalizing symptoms, academic skills). Finally, increasing diversity of measures and methods of assessment is an important task for future research. For example, although we utilized a psychometrically sound assessment of children’s angry reactivity to interparental conflict within a larger multi-method measurement battery, our approach of rating maternal descriptions of their children’s responses to conflict does not capture all components of their emotional experiences. Therefore, it would be beneficial to supplement this methodology with measures that can capture their subjective anger in response to the conflicts.

In conclusion, our study was the first longitudinal test of the moderating role of effortful control in the prospective association between interparental conflict and externalizing problems. Extending previous cross-sectional findings (David & Murphy, 2007), our results indicated that effortful control served as a protective-stabilizing factor that offset the risk posed by interparental conflict in the subsequent development of children’s externalizing problems two years later. In furthering a process-oriented model, mediated moderation analyses demonstrated that reductions in children’s angry reactivity to interparental conflict accounted for the protective role of effortful control in decreasing children’s vulnerability to externalizing problems in the context of interparental conflict. Our findings suggest that the proposed protective properties of effortful control not only disrupt the cascading process of early risk, but also independently predict more proximal processes in this cascade. Thus, targeting children’s deficits in effortful control may be a key area for intervention in reducing children’s early adjustment problems as they transition to the early school years.