Explaining Heterogeneity of Daily Conflict Spillover in the Family: The Role of Dyadic Marital Conflict Patterns.

In this multi-informant, longitudinal, daily diary study, we investigated whether long-term dyadic patterns of marital conflict resolution explain the heterogeneity in short-term day-to-day cross-lagged associations between marital conflict intensity and mother-adolescent conflict intensity. The sample consisted of 419 adolescents (44.6% girls, Mage  = 13.02, SD = 0.44, at T1; Mage  = 17.02, SD = 0.44, at T5), their mothers (N = 419, Mage  = 44.48, SD = 4.17, at T1), and their fathers (N = 419, Mage  = 46.76, SD = 4.99, at T1). Mothers and fathers reported on their marital conflict resolution strategies annually across 5 years. Mother-father daily conflict intensity (mother-reported) and mother-adolescent daily conflict intensity (mother- and adolescent-reported) were assessed for 75 days across 5 years. We hypothesized that long-term marital conflict resolution patterns would moderate the short-term daily dynamics of conflict between the marital and the mother-adolescent dyads. Latent Class Growth Analysis revealed four types of families based on long-term dyadic marital conflict resolution, including families where mostly constructive or mostly destructive conflict resolution was used. Dynamic Structural Equation Modeling was used to investigate the daily levels and short-term daily dynamics of conflict, revealing that for most families there were no day-to-day lagged associations between marital conflict and mother-adolescent conflict. Results showed that long-term conflict resolution patterns did not moderate the short-term dynamics of daily conflict. However, differences among long-term marital conflict resolution patterns were found in the levels of daily conflict, such that in families with long-term destructive conflict resolution patterns, daily conflict intensity was higher.

Conflict is an inevitable aspect of family relationships. Conflict among a family dyad, like the parental dyad, can spread to another dyad in the family, like the mother–adolescent dyad (Sears, Repetti, Reynolds, Robles, & Krull, 2016). Indeed, the spillover hypothesis predicts a positive association between marital conflict and parent–child conflict, because the negative mood and the friction stemming from marital conflict render parents more prone to conflict with their children (Erel & Burman, 1995; Sears et al., 2016). Alternatively, the compensatory hypothesis predicts a negative association between marital conflict and parent–child conflict, because parents try to compensate for the dissatisfaction in their conflictual marriage by investing in a more positive relationship with their child (Cox, Paley, & Harter, 2001; Kouros, Papp, Goeke‐Morey, & Cummings, 2014). Moreover, the compartmentalization hypothesis postulates that there is no association between marital conflict and parent–child conflict, because the boundaries of the marital dyad are strong enough to withhold marital conflict from inflicting the parent–child dyad. Spillover, compensation, and compartmentalization (Erel & Burman, 1995; Krishnakumar & Buehler, 2000) are three hypotheses about how conflict in one family dyad might affect conflict in other dyads of the family. In other words, these three hypotheses apply to within‐family effects of conflict.

Many studies have investigated these hypotheses and each of these hypotheses has received support (spillover: Nelson, Boyer, Villarreal, & Smith, 2017; compensation: Kouros et al., 2014; compartmentalization: Mastrotheodoros, Van Lissa, et al., 2020). Moreover, these studies have shown that there is significant heterogeneity in associations, indicating that not all families function similarly (Sturge‐Apple, Davies, Cicchetti, & Fittoria, 2014). This implies that these three within‐family hypotheses might all bear some truth, yet each of the hypotheses might apply to different groups of families. Some families might function according to the spillover hypothesis, in which marital conflict leads to mother–adolescent conflict. Other families might show a compensatory pattern, in which marital conflict deters mother–adolescent conflict. Yet other families align more with the compartmentalization hypothesis, in which marital conflict is unrelated to mother–adolescent conflict. Such between‐family differences in within‐family dynamics have important implications for interventions (Kerig, 2016). Therefore, it is important to understand why in different families, mother–adolescent conflict is affected differently by marital conflict. Or, in other words, which between‐family differences might determine the heterogeneity in within‐family dynamics.

A factor that potentially explains between‐family differences in how marital conflict affects mother–adolescent conflict might be how marital couples manage marital conflicts. Some couples handle marital conflict with conflict resolution styles consisting of behaviors like yelling, blaming, and withdrawing, that have been found to be less constructive and relate to lower wellbeing and marital satisfaction (Gottman, 1993). Other couples use conflict resolution styles consisting of attempts to bridge differences and find compromises, which have been found to relate to better adjustment outcomes (Whitton, James‐Kangal, Rhoades, & Markman, 2018).

The purpose of this study was twofold. First, we investigated between‐family differences in dyadic marital conflict patterns based on constructive and destructive resolution tactics. Second, we aimed to examine whether these between‐family differences in marital conflict resolution can explain the heterogeneity in the daily dynamics of family conflict, focusing on the daily interplay of marital conflict and mother–adolescent conflict.

Dyadic Marital Conflict Resolution Patterns

Marital conflict resolution strategies can be constructive, if they elicit more positive than negative reactions, or destructive, if the opposite is true (Goeke‐Morey, Cummings, Harold, & Shelton, 2003). Constructive conflict resolution strategies, by means of renegotiating the needs, roles, and boundaries in the marital relationship, may help build bridges, or strengthen existing ones. On the contrary, destructive marital conflict resolution strategies, by means of behaviors that alienate and demean the partners, can lead to depression and marital dissolution (Markman, Rhoades, Stanley, Ragan, & Whitton, 2010; Whitton et al., 2018).

Constructive conflict resolution consists of behaviors such as positive problem solving, support and trying to understand each other’s position, and finding compromises, whereas destructive conflict resolution consists of behaviors such as attacking each other, withdrawing, or giving in without standing for one’s opinion or desires (Mastrotheodoros, Van der Graaff, Deković, Meeus, & Branje, 2019a; McCoy, Cummings, & Davies, 2009). Whether marital conflict is resolved in a constructive or a destructive manner has important implications for family and child functioning (Gao, Du, Davies, & Cummings, 2019; McCoy et al., 2009). For example, constructive conflict has been associated with more supportive and warmer parenting (McCoy, George, Cummings, & Davies, 2013), whereas destructive conflict has been associated with less supportive parenting (Gao et al., 2019; McCoy et al., 2013).

Even though several past studies applied variable‐focused approaches to studying conflict resolution strategies, there are various reasons to endorse a person‐centered, dyadic approach; such an approach focuses on conflict resolution patterns (Bogat, von Eye, & Bergman, 2016). First, individuals may use both constructive and destructive tactics. A parent that almost exclusively and consistently employs destructive strategies can be expected to contribute to their family’s emotional climate differently compared to a parent that employs some destructive, but mostly constructive strategies. Second, the dyadic partners in a couple might not use the same constellation of conflict resolution strategies. The emotional climate in a family where both parents employ mostly constructive marital conflict strategies might be different compared to a family where one parent employs destructive marital conflict strategies and the other parent uses constructive marital conflict strategies (Gottman, 1993). A person‐centered dyadic approach can simultaneously take into account (a) different aspects of conflict resolution (constructive and destructive behaviors); and (b) different reporters (mothers and fathers). Thereby, such an approach might be better suited to tackle population heterogeneity and provide a better insight in the family dynamics than segregated conflict resolution strategies. Moreover, a person‐centered approach is inductive and allows the unobserved subpopulations to emerge from patterns in the data, rather than imposing patterns on the data by the researcher.

Indeed, the usefulness of applying person‐centered typologies on studying marital conflict has been empirically supported (Bogat et al., 2016; Wickrama, Klopack, O’Neal, & Neppl, 2020). For example, taking into account the interactive effect of both partners’ conflict resolution behaviors predicted later divorce over and above the segregated behaviors of each partner (Birditt, Brown, Orbuch, & McIlvane, 2010), supporting a dyadic approach. Furthermore, based on both constructive and destructive conflict resolution behaviors of both mothers, and fathers, four groups of couples have been found, that is, Concordant Constructive, Mother Destructive, Father Destructive, and Concordant Destructive (Kopystynska, Paschall, Barnett, & Curran, 2017). Mothers in the Mother Destructive group were harsher toward their children compared to mothers in the Concordant Constructive group, a finding that supports the usefulness of dyadic patterns of marital conflict resolution behaviors. Therefore, person‐centered typologies are a promising way to investigate how marital conflict resolution patterns can have implications for family functioning.

In this study, we used a person‐centered approach in families with an adolescent child, in order to derive patterns of dyadic marital conflict resolution of mothers and fathers, across adolescence.

Daily Within‐Family Mother–Adolescent Conflict Dynamics

It is well‐established that conflict among family members, like marital conflict (Repetti, Reynolds, & Sears, 2015) and parent–child conflict (e.g., Laursen, Coy, & Collins, 1998; Repetti et al., 2015) occurs in most families. Investigating parent–child conflict is particularly interesting during adolescence, as conflicts become more intense during this period, especially for the mother–adolescent dyad (Mastrotheodoros, Van der Graaff, Van der Graaff, Deković, Meeus, & Branje, 2020). However, there is less agreement regarding how marital conflict affects mother–adolescent conflict. Three hypotheses offer contradicting propositions (Krishnakumar & Buehler, 2000), assuming positive (spillover hypothesis; Erel & Burman, 1995), negative (compensatory hypothesis; Krishnakumar & Buehler, 2000), or null (compartmentalization hypothesis; Krishnakumar & Buehler, 2000) associations between marital conflict and mother–adolescent conflict. Even though it is often assumed that such associations imply marital effects on children, family systems are dynamic (Cox et al., 2001) and children might affect parents as well. Based on the circular causality principle (Lobraico, Brinberg, Ram, & Fosco, 2020; Minuchin, 1974), we can expect bidirectional effects of daily conflict between the marital and the parent–child dyads. Indeed, empirical research has found that the parent–child relationship may also have an effect on the marital relationship (Mastrotheodoros et al., 2019a; Sears et al., 2016; Whiteman, McHale, & Crouter, 2007).

Although the theories do not specify the time scale in which the postulated effects of marital conflict on parent–adolescent conflict would emerge, most studies have applied daily diary methods to investigate the three hypotheses. Some of these studies supported the spillover hypothesis: in days that mothers reported more conflict with their spouses or partners, they also reported more conflict (Nelson et al., 2017) and lower relationship quality (Kouros et al., 2014) with their adolescents. Similarly, daily positive and negative experiences of parents in the marital relationship were positively associated with positive and negative adolescent emotions, respectively, supporting the spillover hypothesis (Merrilees, McCormick, Hsueh, Chou, & Cummings, 2018). One of these studies also found support for compensation: mothers reported that relationship quality with their adolescent children improved following days with marital conflict (Kouros et al., 2014). Finally, other studies found evidence for compartmentalization: marital conflict was not found to significantly affect adolescent–mother conflict the next day (Mastrotheodoros, Van Lissa, et al., 2020). In short, daily diary studies have offered support for all three hypotheses regarding effects of marital conflict on parent–adolescent conflict or relationship quality.

In several daily diary studies that examined conflict processes, significant variance was found around the average daily‐level effects (Mastrotheodoros, Van Lissa, et al., 2020). This variance indicates that the average effect does not necessarily apply to all families and that different effects (spillover, compensation, and compartmentalization) might hold for different groups of families within the same study. That is, marital conflict might be associated with mother–adolescent conflict positively in some families, negatively in other families, and not at all (null effects) in other families, indicating between‐family heterogeneity in within‐family dynamics. The question that arises then is which factors may explain such between‐family differences.

In this study, we focused on the heterogeneity in bidirectional day‐to‐day effects between marital conflict intensity and mother–adolescent conflict intensity. Specifically, we tested the hypothesis that the dispositional marital conflict resolution patterns might explain heterogeneity in within‐family dynamics.

Marital Conflict Patterns and Heterogeneity in Within‐Family Conflict Dynamics

Between‐family dispositional characteristics have recently attracted empirical attention as candidate features of family life that may explain between‐family heterogeneity in within‐family dynamics (Fosco & Lydon‐Staley, 2019; Jarvis, McClure, & Bolger, 2019). The dyadic patterns of marital conflict that characterize a family across years (Gottman, 1993; Whitton et al., 2018), may be one reason for between‐family differences in within‐family associations between marital conflict and mother–adolescent conflict. According to family systems theory, the marital subsystem and the parent–child subsystem are interdependent (Minuchin, 1974). Couples that employ destructive conflict patterns might have difficulty maintaining the boundaries of the marital subsystem (Fosco, Lippold, & Feinberg, 2014). Therefore, there are higher chances that the marital conflict will transcend the boundaries of the parental subsystem and will affect the parent–child relationship.

Empirical research has supported the hypothesis that family dispositional characteristics, like patterns of marital conflict resolution, play an important role in explaining daily fluctuations in family life. For example, in couples with higher marital aggression, dyadic boundaries are more permeable (Timmons, Arbel, & Margolin, 2017), which might contribute to spillover of conflict within the same family. Also, lower dispositional family cohesion moderated the negative effect of daily fluctuations in cohesion on adolescent adjustment, implying, again, more “permeable” family boundaries in those suboptimally functioning families (Fosco & Lydon‐Staley, 2019). Therefore, the dyadic patterns parents use to resolve conflict might constitute a dispositional aspect of the marital subsystem which might make daily conflict spillover more or less probable. Despite extant evidence, whether dyadic patterns of marital conflict affect the within‐family, day‐to‐day associations among marital conflict intensity and mother–adolescent conflict intensity remain unexplored.

The Present Study

In sum, previous theoretical accounts and empirical evidence support the idea that differences between families exist in whether marital conflict spills over to mother–adolescent conflict. Furthermore, theory and research point to the possibility that marital subsystem characteristics, like the dyadic patterns of marital conflict resolution, might influence the daily dynamics among marital conflict intensity and mother–adolescent conflict intensity. In other words, between‐family differences in marital conflict resolution patterns might help explain within‐family processes of daily conflict spillover in the family. This possibility has not been empirically tested. Therefore, the current multi‐informant, longitudinal daily diary study aimed at answering the following questions: (1) What are the dyadic patterns of marital conflict resolution behaviors that differentiate families at the between‐family level? (RQ1) We expected that at least four dyadic patterns of marital conflict resolution would emerge (Kopystynska et al., 2017; Whitton et al., 2018), indicating: (a) families where both spouses use mainly constructive behaviors, (b) families where both spouses use mainly destructive behaviors, (c) families where the mother, but not the father, uses mainly constructive behaviors, and (d) families where the father, but not the mother, uses mainly constructive behaviors. (2) Do these between‐family differences in dyadic marital conflict resolution patterns moderate the degree to which daily conflict within one family dyad affects daily conflict in another family dyad? (RQ2) We expected that marital conflict patterns would moderate the within‐family dynamics of daily conflict, such that a stronger spillover effect of daily marital conflict on daily mother–adolescent conflict would be observed in families in which parents employ primarily destructive conflict (Fosco et al., 2014).

Method

Participants

The sample consisted of 419 adolescents (44.6% girls, M age = 13.02, SD = 0.44, at T1; M age = 17.02, SD = 0.44, at T5), their mothers (N = 419, M age = 44.48, SD = 4.17, at T1), and their fathers (N = 419, M age = 46.76, SD = 4.99, at T1) who took part in five annual assessments of an ongoing longitudinal study (RADAR, Branje & Meeus, 2018) in the Netherlands, from 2006 to 2010. From the initial number of families participating in the study (N = 497), only the families that stayed intact across the five years were chosen for the current study. Adolescents were recruited from randomly selected elementary schools from the province of Utrecht as well as from three other big cities in the Netherlands. During the first measurement wave, adolescents were in 7th Grade. Most adolescents were native Dutch (94.8%) and came from households with medium or high socioeconomic status (91.4%). More information on the sample selection procedures can be found in Branje and Meeus (2018).

Procedure

The study was approved by the medical ethics committee of Utrecht University. Before the start of the study, parents were required to provide informed consent, and adolescents to provide assent. Data were collected via annual home visits during which participants filled‐in self‐report questionnaires, and procedures were the same for all five waves.

To assess annual marital conflict resolution, mothers and fathers filled out questionnaires during annual home visits that, for the most part, took place between February and March each year. Trained research assistants provided verbal instructions in addition to written instructions that accompanied the questionnaires. To assess daily marital conflict and daily mother–adolescent conflict, mothers and adolescents were asked to fill in an online daily diary for five consecutive days, in June, September, and December each year, for five consecutive years, resulting in 75 daily diaries in total (5 days*3 weeks*5 years). Confidentiality was guaranteed, and the data were processed anonymously. Each wave families received 100 euros for their participation.

Measures

Interparental conflict management

We used the Conflict Resolution Styles Inventory (CRSI, Kurdek, 1994) to assess marital conflict resolution annually. The CRSI consists of 20 items and measures four conflict resolution strategies with five items each: Conflict Engagement, Compliance, Withdrawal, and Problem Solving. The CRSI has been translated and used in Dutch and has shown good psychometric properties (Van Doorn, Branje, & Meeus, 2007). The items are addressed using a Likert scale from 1 (never) to 5 (always), and all items were preceded by the quote “During an argument or conflict with my husband/wife I do the following things.” Example items for each subscale are “Personally attack him/her,” and “Exploding and getting out of control.” (Conflict Engagement); “Not being willing to stick up for myself,” and “Giving in with little attempt to present my side of the issue.” (Compliance); “Focusing on the problem at hand,” and “Sitting down and discussing differences constructively.” (Problem Solving); “Remaining silent for long periods of time,” and “Withdrawing, acting distant and not interested.” (Withdrawal). Mothers reported about their own conflict resolution behavior in arguments with their husbands, and fathers reported about their own conflict resolution behavior in arguments with their wives. To obtain a more robust measure of marital conflict resolution, we used repeated measures of the mother‐ and father‐reported four subscales across five years (see Analytic Plan). In each of the five waves internal consistency coefficients were good across scales, reporters, and waves, ranging between α = .75 and α = .90.

Daily interparental and mother–adolescent conflict

We used two adapted items from the Negative Interactions subscale of the Network of Relationships Inventory—short form (De Goede, Branje, & Meeus, 2009; Furman & Buhrmester, 1985) to assess marital conflict and mother–adolescent conflict on the daily level. The two items mothers and adolescents addressed were: “How much did you and your partner/child/mother get upset with or mad at each other today?”, and “How much did you and your partner/child/mother argue with each other today?”. Mothers reported about conflict with their partners (Mother‐Father report, MF), and mothers and adolescents reported about mother–adolescent conflict (Mother–Adolescent, MA; Adolescent–Mother, AM). Answers were given on a 7‐point Likert scale, ranging from 1 (Not at all) to 7 (Very much). Internal reliability was good across days, across years, and for all informants: α’s .81 to .88 (MF report); .80 to .86 (MA report); .83 to .85 (AM report). The MA and the AM reports correlated strongly (from .46 to .48 across waves), therefore we combined them into one score of mother–adolescent conflict.

Attrition and Missing Values

The majority of adolescents (84.7%), mothers (84.7%), and fathers (76.1%) were still involved in the study at Wave 5, and the average participation rate across the five waves was 91.3%, 91.2%, and 90% for adolescents, mothers, and fathers, respectively. Little’s MCAR test (Little, 1988) was not significant [χ2 (767) = 812, p = .125], indicating that the assumption of missingness being completely at random was tenable. Therefore, data from all 419 families could be included in the analyses using Full Information Maximum Likelihood.

Analytic Plan

As preliminary analyses, to test the appropriateness of the CRSI for dyadic longitudinal modeling with our data, we applied the following steps. First, we examined the factor structure of the CRSI using Confirmatory Factor Analyses separately for the five annual assessments, and for mothers and fathers. Second, we tested the longitudinal measurement invariance of the CRSI across the five years (van de Schoot, Lugtig, & Hox, 2012), separately for mothers and fathers, using the measEq.syntax function of the semTools package in R (Jorgensen et al., 2018). In these increasingly restricted models, changes in model fit were deemed acceptable if deltas (Δs) were within the cutoffs recommended by Chen (2007): ΔCFI ≤ −.010, ΔRMSEA ≤ .015, and ΔSRMR ≤ .030. Given the non‐normality of the scales, we chose a maximum likelihood estimator with robust standard errors.

To answer the first research question on the dyadic patterns of mother‐reported and father‐reported conflict resolution strategies, we applied the following steps. First, we estimated univariate Latent Growth Curve Models (LGCM, (Wang & Wang, 2012) in lavaan (Rosseel, 2012), for each of the four subscales of the CRSI (Problem Solving, Conflict Engagement, Compliance, Withdrawal), and separately for mother and father reports (eight LGCMs in total). Conducting univariate LGCMs is a recommended first step before applying Latent Class Growth Analysis (Jung & Wickrama, 2008). For each LGCM, we examined whether linear, quadratic, or cubic slopes fit the data best, based on model fit indices. Model fit was deemed good based on the following thresholds: CFI > .95, TLI > .95, RMSEA < .08, SRMR < .08, as well as based on a comparatively lower BIC (Little, 2013). Second, we applied Latent Class Growth Analyses (Nagin, 2005) in Mplus 8.2 (Muthén & Muthén, 2018) to explore the dyadic patterns of mother‐ and father‐reported conflict resolution strategies, across time.

To answer the second research question regarding whether the short‐term daily dynamics of within‐family conflict intensity vary by between‐family differences in the long‐term dyadic patterns of marital conflict resolution strategies we proceeded in the following three steps. First, as a preliminary step, we tested for measurement invariance of the daily diaries (a) across days within years separately for each year, and (b) across weeks for all five years simultaneously. Specifically, using a series of CFAs, we applied increasingly stricter models to test for configural, metric, and scalar invariance; these levels correspond to equality of factor structure, item factor loadings, and item intercepts (van de Schoot et al., 2012). In the first set of these analyses, we treated weekdays as items of a construct; in the second set we treated weeks as items of a construct. For example, we tested whether (a) the factor loadings and item intercepts of Monday, Tuesday, […] and Friday in week 1, in year 1, were the same as the factor loadings and item intercepts of, respectively, Monday, Tuesday, […] and Friday in weeks 2 and 3 of the same year; and (b) the factor loadings and item intercepts of weeks 1–3, in year 1 were the same as the factor loadings and item intercepts of, respectively, weeks 1–3 in years 2–5.

In the second step, we applied Dynamic Structural Equation Modeling (Asparouhov, Hamaker, & Muthén, 2018), to examine the day‐to‐day bidirectional effects between marital conflict intensity and mother–adolescent conflict intensity. DSEM is an innovative newly developed analytic technique that combines the advantages of three analytical frameworks: time‐series, multilevel, and Structural Equation Models (McNeish & Hamaker, 2020). It offers the possibility of simultaneously modeling intensive longitudinal data (like in time‐series) of multiple individuals (like in multilevel modeling) and of several possible predictor, mediator, and outcome (latent) variables (like in Structural Equation Modeling). In this study, the time‐series part was specified as an AR(1) process. However, we manually entered a 10‐day gap between weeks, forcing the autoregressive parameter to reach zero, thereby accounting for the fact that not all days were consecutive. The multilevel part of DSEM consisted of estimating family‐specific (random) within‐family effects. The SEM part consisted of the possibility to simultaneously model two variables and their bidirectional effects, while taking measurement error into account. The factor scores of this analysis represented the family‐specific bidirectional effects among daily mother–father and mother–adolescent conflict and could be positive, negative, or null. These factor scores were saved to be used in the final step.

In the third step, we used the saved scores from DSEM as outcomes in a final run of LCGA, to examine the effect of the conflict resolution patterns on the factor scores. We reran the best fitting LCGA (from research question 1), but this time applying the Bolck–Croon–Hagenaars (BCH) procedure (Bakk & Vermunt, 2016) in Mplus 8.2. This procedure allows the estimation of LCGA with distal continuous outcomes of the latent classes. In other words, this technique estimates the latent classes as predictors of distal outcomes (in this case, daily conflict and day‐to‐day cross‐lagged effects). In addition, we used the option TYPE = COMPLEX in Mplus in this step to account for the interdependence among dyads in the same family.

Results

Descriptive Statistics

Table S1 provides the means, standard deviations, and bivariate correlations of the mother‐reported and father‐reported CRSI scales across the five waves.

Factor Structure and Longitudinal Measurement Invariance of the CRSI

Table S2 presents the results of the CFAs and the longitudinal measurement invariance analyses of the CRSI. The factor structure of the CRSI was acceptable based on both mother and father reports, across time points (CFI: .916 to .954; TLI: .901 to .946; RMSEA: .049 to .066; SRMR: .047 to .058; see Table S2 for more details). Furthermore, the longitudinal measurement invariance analyses supported the configural, metric, and scalar models for both the mother‐ and the father‐reported CRSI (ΔCFI: .000 to −.002; ΔTLI: .000 to −.001; ΔRMSEA: .000; ΔSRMR: .000 to .003, for pairwise model comparisons; see Table S2 for more details). These three forms of invariance indicate equality of factor structure, factor loadings, and item intercepts, respectively, for the mother‐ and father‐reported forms of the CRSI. Therefore, the use of the mother‐ and father‐reported CRSI in the next analytic steps is warranted.

Dyadic Patterns of Marital Conflict Resolution Strategies

Table S3 presents the results of the LGCMs for the mother‐ and the father‐reported CRSI scales. Based on fit indices (CFI, RMSEA, SRMR), linear slopes fit most of the subscales best, with two exceptions. For the mother reports, Conflict Engagement and Compliance were best represented by cubic slopes, compared to linear and quadratic slopes.

Table S4 presents the fit indices of the LCGAs with different number of classes. Figures S1–S6 present the means of the four CRSI scales for mothers and fathers in the different classes. Based on the fit indices (Table S4) and the interpretability of the classes (Figures S1–S6), we chose a 4‐class solution for the configurations of the CRSI, for mothers and fathers. In the solutions beyond the 4‐class, the new classes were only variations of already existing classes in the 4‐class solution.

Table 1 presents the means and confidence intervals of the growth parameters (intercepts and slopes) of the four CRSI scales across the four classes, for mothers and fathers separately, whereas Figure S7 graphically presents the scores of mothers and fathers in each group, along with the sample grand mean, to aid interpretation. The groups did not differ significantly on the rates of change of the conflict management strategies; only intercept differences emerged (Table 1). Based on these means, we labeled the classes “Couple Constructive” (class #1, n = 132), “Couple Destructive” (class #2, n = 53), “Father Constructive—Mother Average” (class #3, n = 141), and “Mother Constructive—Father Submissive” (class #4, n = 91). Mothers and fathers in the “Couple Constructive” class generally showed the highest scores on Problem Solving, and the lowest scores on Conflict Engagement, Compliance, and Withdrawal. In contrast, mothers and fathers in the “Couple Destructive” class had low Problem Solving, and high destructive conflict. In the “Mother Constructive—Father Submissive” class, mothers showed more constructive and less destructive conflict (except for Conflict Engagement) than the sample grand mean, while fathers showed more Withdrawal and Compliance than the sample grand mean, and average Conflict Engagement. Finally, in the “Father Constructive—Mother Average” class, fathers had the same profile as fathers in the “Couple Constructive” class, that is, high Problem Solving, and low destructive conflict, whereas mothers scored average on all scales, with lower constructive and higher destructive conflict scores compared to fathers.

Table 1

Means and 95% Confidence Intervals for Growth Parameters of the Mother‐, and Father‐reported Conflict Resolution Styles Inventory for the Four Classes

Conflict Resolution Patterns	Couple Constructive (n = 132)		Couple Destructive (n = 53)		Mother Constructive—Father Submissive (n = 91)		Father Constructive—Mother Average (n = 141)
	M (95% CI)		M (95% CI)		M (95% CI)		M (95% CI)
	Mothers	Fathers	Mothers	Fathers	Mothers	Fathers	Mothers	Fathers
Intercepts
PS	4.17a,1 (4.05–4.28)	3.90b,1 (3.79–4.01)	3.243 (2.91–3.57)	3.342 (3.09–3.60)	3.83a,2 (3.70–3.95)	3.36b,2 (3.23–3.49)	3.70b,2 (3.58–3.81)	3.93a,1 (3.81–4.06)
CE	1.60a,3 (1.48–1.72)	1.43b,2 (1.35–1.51)	2.43a,1 (2.09–2.77)	2.01b,1 (1.58–2.44)	2.082 (1.84–2.32)	1.961 (1.73–2.19)	2.05a,2 (1.91–2.19)	1.51b,2 (1.40–1.62)
CO	1.70b,4 (1.59–1.80)	2.16a,2 (2.06–2.26)	2.671 (2.37–2.97)	2.641 (2.41–2.88)	1.95b,3 (1.80–2.10)	2.71a,1 (2.57–2.85)	2.202 (2.08–2.33)	2.152 (2.06–2.25)
WI	1.44b,4 (1.35–1.52)	1.76a,2 (1.64–1.88)	2.981 (2.76–3.20)	2.681 (2.46–2.90)	1.89b,3 (1.75–2.03)	2.77a,1 (2.61–2.93)	2.27a,2 (2.10–2.44)	1.76b,2 (1.66–1.87)
Linear slopes
PS	0.02 (0.00–0.04)	0.02 (−0.01–0.04)	−0.04 (−0.09–0.00)	−0.01 (−0.06–0.04)	0.01 (−0.02–0.04)	0.00 (−0.03–0.04)	0.02 (0.00–0.04)	0.00 (−0.04–0.03)
CE	−0.19 (−0.33–−0.05)	−0.04 (−0.05–−0.02)	−0.16 (−0.54–0.23)	−0.08 (−0.11–−0.04)	−0.03 (−0.22–0.16)	−0.04 (−0.08–−0.01)	−0.16 (−0.31–−0.01)	−0.04 (−0.06–−0.02)
CO	−0.10 (−0.27–0.08)	−0.01 (−0.04–0.01)	−0.38 (−0.64–−0.11)	−0.01 (−0.06–0.04)	0.02 (−0.20–0.25)	−0.02 (−0.05–0.01)	−0.05 (−0.20–0.10)	−0.02 (−0.04–0.01)
WI	−0.03 (−0.05–−0.01)	−0.03 (−0.06–−0.01)	0.00 (−0.06 to 0.05)	−0.01 (−0.08–0.06)	−0.02 (−0.06–0.02)	−0.02 (−0.05–0.01)	−0.02 (−0.04–0.00)	−0.02 (−0.05–0.00)
Quadratic slopes
CE	0.07 (−0.01–0.15)	—	0.06 (−0.19–0.31)	—	−0.01 (−0.13–0.10)	—	0.06 (−0.04–0.16)	—
CO	0.05 (−0.06–0.15)	—	0.22* (0.06–0.39)	—	0.00 (−0.13–0.14)	—	0.03 (−0.06–0.12)	—
Cubic slopes
CE	−0.01 (−0.02–0.00)	—	−0.01 (−0.05–0.03)	—	0.00 (−0.02 to 0.02)	—	−0.01 (−0.02–0.01)	—
CO	−0.01 (−0.03–0.01)	—	−0.04 (−0.06–−0.01)	—	0.00 (−0.02–0.02)	—	−0.01 (−0.02–0.01)	—

CE = conflict engagement; CI = confidence intervals; CO = compliance; PS = problem solving; WI = withdrawal.

Values with different letter indicators indicate statistically different means across sex within each class, where “a” denotes a higher mean than “b”. Values with different number indicators indicate statistically different means across classes within each sex, where “1” indicates higher mean than “2,” “2” indicates higher mean than “3,” and “3” indicates a higher mean than “4.”

Day‐to‐Day Dynamics of Marital Conflict and Mother–adolescent Conflict

Tables S5 and S6 present the results of the measurement invariance of the daily diaries of conflict. Table S7 presents the results of the DSEM analyses, while Figure S8 shows a graphical representation of DSEM as applied in this study. The between‐family correlation of marital conflict with mother–adolescent conflict was positive, significant, and strong (r = .68), indicating that families with higher marital conflict, compared to other families, also had higher mother–adolescent conflicts. The within‐family within‐day correlation of marital conflict and mother–adolescent conflict was positive and significant, but weak (r = .18), indicating that days with higher marital conflict intensity tended to also be days with higher mother–adolescent conflict intensity.

Furthermore, the day‐to‐day autoregressive stability of marital conflict was positive, significant, but weak (r = .17), indicating a weak carry‐over effect within the dyad. That is, on average, days with high marital conflict relative to the family’s own mean tended to also be followed by days with relatively high marital conflict; a similar carry‐over effect was also found for mother–adolescent conflict (r = .16).

On average, no significant day‐to‐day cross‐lagged effects were found between marital conflict and mother–adolescent conflict. That is, the occurrence of mother–adolescent conflict on a certain day was unrelated to the intensity of marital conflict that occurred the day before, and vice versa. However, there were significant variances of the two cross‐lagged effects. These variances indicate significant differences between families in the day‐to‐day cross‐lagged effects between marital conflict and mother–adolescent conflict, implying that some families might show a spillover pattern, others a compensatory pattern, and yet others might show a compartmentalization pattern. Figures 1 and S9 present the histograms of the two cross‐lagged effects—from marital to mother–adolescent, and vice versa. Most families were characterized by null effects, supporting the compartmentalization hypothesis, but sizeable subgroups emerged with negative (compensatory) and positive (spillover) effects (Figures 1 and S9). To make these histograms, we used the 95% credible intervals of the cross‐lagged effects to group families in those with significant negative cross‐lagged effects (in which the 97.5% credible interval was negative), null effects (in which the credible intervals included zero), and significant positive effects (in which the 2.5% credible interval was positive).

Figure 1

Histogram of the Day‐to‐Day Standardized Cross‐lagged Effect of Marital Conflict on Mother–adolescent Conflict, Grouped by Valence of Statistical Significance, Based Only on Two‐parent Families that Stayed Intact Across the Study Duration.

Differences Between Dyadic Marital Conflict Patterns in the Day‐to‐Day Dynamics of Conflict

Finally, Table 2 presents the mean comparisons among the four marital conflict resolution patterns in the average levels of daily marital conflict, daily mother–adolescent conflict, as well as the average day‐to‐day cross‐lagged effects of marital conflict on next‐day mother–adolescent conflict, and vice versa. In addition, Table 2 presents the global, between‐family, bivariate correlations between marital conflict intensity and mother–adolescent conflict intensity, separately for each dyadic marital conflict pattern. Significant mean differences emerged in the levels of both daily marital and daily mother–adolescent conflict intensity, but almost no significant differences among classes were found in the day‐to‐day cross‐lagged effects, and in the between‐family bivariate correlations. Daily marital conflict was the highest in the Couple Destructive group, and the lowest in the Couple Constructive group, with the Mother Constructive‐Father Submissive and the Father Constructive‐Mother Average groups falling in‐between. Similarly, mother–adolescent dyads in families of the Couple Constructive class reported the lowest level of daily mother–adolescent conflict intensity, with mother–adolescent dyads in families of the other three classes not differing significantly among each other. Mean differences in the day‐to‐day cross‐lagged effects across groups were nonsignificant. Even though the mean of the day‐to‐day effect of mother–adolescent conflict on next‐day marital conflict (“reverse spillover”) was higher in the Couple Constructive class compared to the Father Constructive‐Mother Average class, the effect was null in both classes, therefore this mean difference is rather spurious.

Table 2

Means, and Confidence Intervals of Levels of Daily Conflict, and Day‐to‐Day Cross‐Lagged Effects, as well as Bivariate Between‐Person Correlations Across Conflict Resolution Patterns

Daily Conflict	Conflict Resolution Patterns
	Couple Constructive (n = 132)		Couple Destructive (n = 53)		Father Constructive—Mother Average (n = 141)		Mother Constructive—Father Submissive (n = 91)
	M	95% CI	M	95% CI	M	95% CI	M	95% CI
Level
Marital conflict	2.603	2.30–2.90	4.841	3.97–5.70	2.962	2.67–3.26	3.182	2.81–3.54
Mother–adolescent conflict	3.133	2.82–3.45	4.851	4.06–5.64	4.242	3.80–4.68	4.162	3.62–4.70
Autoregressive stabilities
Marital conflict	0.14	0.11–0.17	0.18	0.12–0.25	0.18	0.15–0.20	0.16	0.13–0.20
Mother–adolescent conflict	0.15	0.14–0.17	0.17	0.13–0.21	0.16	0.14–0.17	0.15	0.13–0.17
Cross‐lagged effects
Marital → Mother‐Adol.	0.00	−0.03–0.03	0.00	−0.05–0.05	−0.01	−0.03–0.02	0.01	−0.02–0.04
Mother‐Adol. → Marital	0.021	0.00–0.03	0.011,2	−0.02–0.03	−0.012	−0.02–0.00	0.011,2	−0.01–0.03
Bivariate between‐person correlations a	r		r		r		r
Marital with Mother‐Adol.	.77***		.75***		.72***		.76***

Conflict Resolution Patterns are the result of the four‐class solution of the Latent Class Growth Analysis on the four scales of the Conflict Resolution Strategies Inventory as reported by mothers and fathers. Levels represent the between‐person estimates of the average daily conflict. Cross‐Lagged Effects represent the within‐person average day‐to‐day cross‐lagged effects. Estimates with different numeric indicators differ significantly at p = .05, where “1” indicates a higher mean.

Fisher’s Z was nonsignificant for all Pearson’s r comparisons.

p < .001.

Discussion

Families differ in whether and how strongly daily marital conflict intensity influences mother–adolescent conflict intensity. Given the leading role of the marital dyad in the family (Fosco et al., 2014), such between‐family differences in the daily dynamics of conflict transmission might be attributed to the dyadic patterns of marital conflict resolution. In the present study, we examined whether the dyadic patterns of marital conflict resolution moderate the day‐to‐day cross‐lagged effects from marital conflict intensity to mother–adolescent conflict intensity. Moreover, because of the interdependency among the family dyads, and given that parent–adolescent conflict can bear a toll on the marital relationship (Mastrotheodoros et al., 2019a), causing marital stress (van Eldik, Prinzie, Deković, & de Haan, 2017), and precipitating marital conflict, we also tested the opposite direction of effects, from daily mother–adolescent conflict intensity to marital conflict intenstiy. Overall, we found significant heterogeneity across families in the within‐family dynamics of daily conflict. For most families daily interparental conflict was unrelated to next‐day mother–adolescent conflict (compartmentalization), yet for sizable subgroups of families, daily interparental conflict had enduring positive (spillover) or negative (compensation) effects on next‐day mother–adolescent conflict. The opposite direction of effects (mother–child conflict affecting interparental conflict) was also mostly nonsignificant and the spillover and compensation patterns were much more rare. Finally, we found little or no evidence for moderation of the within‐family daily dynamics of conflict intensity by the dyadic patterns of marital conflict resolution.

In accordance with our expectations (Birditt et al., 2010; Kopystynska et al., 2017; Whitton et al., 2018) we found four different dyadic patterns of marital conflict resolution, including three patterns in which both or one of the parents used constructive conflict management, and one pattern in which both parents used mostly destructive conflict management strategies. Contrary to our expectations, however, these marital conflict patterns did not significantly explain the variance in the day‐to‐day cross‐lagged effects between marital conflict and mother–adolescent conflict. This implies that how strongly daily conflict in one dyad in the family affects daily conflict in another dyad in the same family does not depend on long‐term marital conflict resolution patterns.

Dyadic Patterns of Marital Conflict Strategies

By using self‐reports for both constructive and destructive conflict strategies from mothers and fathers across five years, we were able to examine the dyadic patterns describing how married couples manage conflict. The results are in accordance with extant research using both self‐report data (Birditt et al., 2010; Kopystynska et al., 2017) and observations (Markman et al., 2010). Many couples (31.6%) were classified as “Couple Constructive”, because they depicted the highest levels of problem solving across years, and low levels of destructive conflict strategies, like conflict engagement, compliance, and withdrawal. Furthermore, in these couples no significant differences were found between mothers and fathers. Two further groups (Mother Constructive‐Father Submissive, 21.8%, and Father Constructive‐Mother Average, 33.8%) described couples in which partner differences were apparent, and one partner typically applied a more destructive conflict management pattern, while the other partner applied a less destructive, or even constructive conflict management pattern. Finally, a fourth pattern (Couple Destructive, 12.7%), the least common, described couples in which partners did not significantly differ from each other, and they both typically employed a destructive conflict management pattern.

This is the first study to assess both constructive and destructive marital conflict strategies as perceived by mothers and fathers, across their children’s adolescence. Conflict configurations found in previous studies (Birditt et al., 2010; Kopystynska et al., 2017; Whitton et al., 2018) were not only confirmed, but were also shown to describe long‐term patterns of how marital dyads resolve conflict. Given the negative implications that destructive patterns have been shown to have for marital satisfaction (e.g., Birditt et al., 2010; Markman et al., 2010; Whitton et al., 2018) and adolescent development (e.g., Fosco et al., 2014), our findings underline the need to focus on those couples who employ mainly destructive marital conflict management strategies.

Dynamics of Daily Conflict and the Role of Dyadic Patterns of Marital Conflict

This study confirmed the heterogeneity among families in the day‐to‐day cross‐lagged effects of marital conflict intensity on next‐day mother–adolescent conflict intensity, and vice versa. For most families, the two within‐family day‐to‐day cross‐lagged effects examined in this study were not significant (n = 331, 79.8%, for the within‐family effect from marital conflict to next‐day mother–adolescent conflict, and n = 369, 89.1%, for the reverse effect). These results are in accordance with the compartmentalization hypothesis, which postulates that the marital dyad is “insulated”, and can withhold marital conflict from spilling over to the mother–adolescent dyad (Krishnakumar & Buehler, 2000). Of note, the daily timeframe we applied in this study might be too long for spillover to be apparent. For example, experimental evidence showed that spillover was apparent but dissipated within few hours (Jouriles & Farris, 1992). Also, we chose to examine daily mother–adolescent conflict during adolescence because this is a formative period for the parent–adolescent relationships: parenting changes (Mastrotheodoros, Van der Graaff, Deković, Meeus, & Branje, 2019b), and parent–adolescent conflict intensity increases (Mastrotheodoros, Van der Graaff, et al., 2020). However, different patterns of associations between marital conflict and parent–adolescent conflict could be observed at different developmental stages (Gao & Cummings, 2019).

Despite the null effect applying to most families, for significant proportions of the families there was a significant within‐family day‐to‐day cross‐lagged effect, especially from marital conflict to next‐day mother–adolescent conflict. Thus, a more appropriate question than asking whether spillover, compensation, or compartmentalization take place is to investigate the between‐family heterogeneity in within‐family dynamics (Keijsers et al., 2016; Papp, 2004) and to elucidate which families are characterized by which pattern. What might be the reasons for these between‐family differences in within‐family daily dynamics?

Contrary to our expectations, the four dyadic patterns of marital conflict resolution could not explain this heterogeneity in the day‐to‐day within‐family cross‐lagged effects among marital conflict and mother–adolescent conflict. Even though sizable proportions of the sample depicted spillover‐like, or compensatory‐like day‐to‐day within‐family cross‐lagged effects, whether a family was more characterized by one pattern or the other was unrelated to the dyadic marital conflict patterns. In addition, no group differences were found on the bivariate correlations between marital and mother–adolescent conflict. These findings go against our hypothesis that destructive marital conflict patterns would increase the chances of spillover from marital conflict to mother–adolescent conflict on the daily level.

The heterogeneity in day‐to‐day within‐family cross‐lagged effects coupled with our finding of strong between‐family correlations among marital and mother–adolescent conflict indicate that other between‐family characteristics explain these differences across families. However, instead of focusing on global, annually assessed family characteristics, the future search for factors that might explain between‐family heterogeneity in daily conflict spillover might benefit by focusing on factors measured on a daily timescale. For example, even though the global tendency to experience negative affect (i.e., Neuroticism) does not explain within‐day spillover‐like associations, daily negative mood partly does (Sears et al., 2016). This indicates the need to match the time‐scales of the phenomena under study (Keijsers & van Roekel, 2018). That is, day‐to‐day effects between marital and mother–adolescent conflict intensity might be affected by the pattern of resolution of the specific conflict, rather than the general pattern of conflict resolution across years. Future studies, thus, may offer a better understanding of the within‐family processes by investigating daily patterns of family functioning with intensive repeated measures designs (Repetti et al., 2015).

Even though the dyadic ways with which parents resolve conflict did not explain the between‐family differences in the day‐to‐day cross‐lagged effects, marital conflict resolution groups differed significantly and meaningfully in the levels of daily marital conflict and mother–adolescent conflict. This is an important finding because it confirms a widely held, but rarely empirically tested notion that the daily life experiences are linked to the global, stable patterns of marital functioning (Granic & Hollenstein, 2015; Repetti et al., 2015). Specifically, the marital conflict patterns as conceptualized in this study were not only significantly associated with marital conflict on the daily level, but also with the mother–adolescent conflict. Therefore, families with different patterns of interparental conflict resolution do differ in how intensely they experience daily conflict, but they do not differ significantly in how daily conflict spreads within the family. To conclude, in families where parents employ more constructive and less destructive conflict resolution, the daily experiences of family members are less conflict‐ridden, compared to families with more destructive resolution patterns.

Strengths and Limitations

This multi‐informant, longitudinal, and daily diary study is the first to examine the heterogeneity in day‐to‐day cross‐lagged effects between marital conflict intensity and mother–adolescent conflict intensity. We applied state‐of‐the‐art analytic techniques, with robust measures, and a large sample, to answer our research questions. This paper has several methodological strengths.

A particular strength of this study is the use of robust, multi‐informant, and longitudinally assessed measures that spanned over a relatively long period of time. Most daily diary studies on family dynamics span few days or weeks, whereas this study combined a panel approach with a measurement burst daily diary approach. Also, the global predictor in this study—marital conflict resolution strategies—was a different aspect of conflict than the mere aggregation of the daily‐level conflict. By focusing on conflict resolution strategy patterns, we were able to examine whether the way parents handle conflict in the marital relationship affects spillover processes of conflict intensity in the family.

The explicit focus on heterogeneity is another strength of this study. The combination of rich data and advanced techniques helped us better acknowledge the differences between families, both on the global, annual level, and on the daily level.

Compared with past daily diary research, the use of DSEM is another particular strength of this study, because this technique allows a more stringent test of the theoretically proposed within‐family mechanisms. Specifically, compared to multilevel modeling, DSEM allows to test a lagged effect from, for example, T1 interparental conflict to T2 mother–adolescent conflict while simultaneously controlling for: (a) the effect of T1 mother–adolescent conflict on T2 mother–adolescent conflict; (b) the T1 correlation between interparental conflict and mother–adolescent conflict; (c) the T2 correlation between interparental conflict and mother–adolescent conflict; and (d) the effect that T1 mother–adolescent conflict has on T2 interparental conflict. In addition, DSEM applies latent person‐mean centering, which is not based on observed scores, and takes measurement error into account. Even though this approach is technically possible in MLM, centering is most often based on observed scores in this type of modeling.

Yet, some limitations should be noticed. The sample consisted of families with a middle/upper‐middle socioeconomic background, belonging to the majority ethnic group, in a relatively affluent western country. Therefore, results might not generalize to families from different socioeconomic strata, ethnic backgrounds, and/or countries. Second, all measures used in this study were self‐reports. Observed behaviors might differ from what people report for themselves, thus different results might have emerged if we used observations of conflict, instead of self‐report. In addition, this study investigated daily conflict only between adolescents and mothers. As such, we were unable to test effects on daily father–adolescent conflict. Another limitation is that the daily timescale is not necessarily the ideal timeframe to investigate spillover. Although the relevant theories do not specify the timescale on which the effects would take place, evidence suggests that spillover can be short‐lived, dissipating within few hours (Jouriles & Farris, 1992). The DSEM currently cannot estimate random within‐day associations, therefore we were unable to test whether dyadic marital conflict patterns moderate the within‐day associations between daily marital and mother–adolescent conflict intensity. Finally, we were not able to investigate within‐family random effects across time in this study. Therefore, the within‐ and across‐days associations in this study are assumed to be stable across the duration of the study. This is an assumption that might not hold true. Future studies should investigate both random effects across families, and across time.

Conclusion

Families differ in how marital conflict during one day might affect mother–adolescent conflict the next day. For some families, days with marital conflict might be followed by days with mother–adolescent conflict (spillover). For other families, such days may be followed by days with lower than average mother–adolescent conflict (compensation). Yet in other families instances of mother–adolescent conflict might be unrelated to instances of marital conflict (compartmentalization; e.g., Krishnakumar & Buehler, 2000). In addition, couples differ in the ways they manage marital conflict (e.g., Birditt et al., 2010). This study investigated whether the dyadic patterns of marital conflict management can explain the heterogeneity in the daily dynamics between marital conflict and mother–adolescent conflict. The results showed that, even though meaningful differences between families exist in the ways couples resolve conflict, these dyadic patterns do not explain the heterogeneity in the within‐family daily dynamics of conflict.

Table S1. Means, Standard Deviations, and Bivariate Correlations Among the Mother‐, and Father‐Reported CRSI Scales Across the Five Years.Table S2. Fit Indices of the Confirmatory Factor Analyses and the Longitudinal Measurement Invariance Analyses for the Conflict Resolution Strategies Inventory, for Mother‐, and Father‐reports, Across Waves 1–5.Table S3. Fit Indices for the Latent Growth Curve Models of the Conflict Resolution Styles Inventory with Linear, Quadratic, and Cubic Specification, for Mother‐, and Father‐RePOrts.Table S4. Fit Indices for the Latent Class Growth Analyses with Different Numbers of Classes.Table S5. Tests of Measurement Invariance of Conflict Daily Diaries Across Three Weeks within Each Year.Table S6. Tests of Measurement Invariance of Conflict Daily Diaries Across Years.Table S7. Within‐family Coefficients (Correlations, Autoregressive Stabilities, and Cross‐Lagged Effects), and Between‐Family Coefficients (Correlations, Mean Levels) for the Daily Dynamics Among Marital Conflict and Mother‐Adolescent Conflict, Controlling for Adolescent Sex.Figure S1. Two‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S2. Three‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S3. Four‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S4. Five‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S5. Six‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S6. Seven‐class Solution of the LCGA on the Father‐ and Mother‐reported CRSI.Figure S7. Four‐class Final Solution of the LCGA on the Father‐ and Mother‐Reported CRSI, with the Sample Grand Mean, Based on Two‐Parent Families Staying Intact Throughout the Duration of Study.Figure S8. Standardized Coefficients from the Dynamic Structural Equation Modeling for Interparental Conflict and Mother‐Adolescent Conflict.Figure S9. Histogram of the Day‐to‐Day Standardized Cross‐Lagged Effect of Mother‐Adolescent Conflict on Marital Conflict, Grouped by Valence of Statistical Significance, Based Only on Two‐parent Families that Stayed Intact Across the Study Duration.

Click here for additional data file.