Constructive and Destructive Marital Conflict, Parenting, and Children’s School and Social Adjustment

Abstract

This study addresses the links between destructive and constructive marital conflict and mothers’ and fathers’ parenting to understand associations with children’s social and school adjustment. Multi-method, longitudinal assessments of 235 mothers, fathers, and children (129 girls) were collected across kindergarten, first, and second grades (ages 5-7 at Time 1; ages 7-9 at Time 3). Whereas constructive marital conflict was related to both mothers’ and fathers’ warm parenting, destructive marital conflict was only linked to fathers’ use of inconsistent discipline. In turn, both mothers’ and fathers’ use of psychological control was related to children’s school adjustment, and mothers’ warmth was related to children’s social adjustment. Reciprocal links between constructs were also explored, supporting associations between destructive marital conflict and mothers’ and fathers’ inconsistent discipline. The merit of examining marital conflict and parenting as multidimensional constructs is discussed in relation to understanding the processes and pathways within families that affect children’s functioning.

Exposure to marital conflict places children at greater risk for developing adjustment problems including aggression, conduct disorders, anxiety, and depressive symptomatology (Buehler, Lange, & Franck, 2007; El-Sheikh, Buckhalt, Mize, & Acebo, 2006; Grych & Fincham, 1990). Traditionally, researchers have focused on examining the relationship between the frequency of marital conflict and child adjustment (Emery, 1982). Recently studies have documented that it is also critical to consider how conflict is handled when attempting to understand how conflict impacts families (Goeke-Morey, Cummings, Harold, Shelton, & 2003; Goodman, Barfoot, Frye, & Belli, 1999; Grych & Fincham, 1990).

Analogue and diary studies have demonstrated ways of categorizing conflict into two categories based on children’s emotional and behavioral reactions: constructive and destructive marital conflict tactics. The use of constructive tactics such as support, verbal and physical affection, problem solving, and resolution elicit positive emotional reactions (e.g., happiness) from children (e.g., Cummings, Goeke-Morey, & Papp, 2003; Goeke-Morey, et al., 2003). In contrast, the use of destructive tactics such as verbal hostility, physical aggression, non-verbal anger, and withdrawal elicit negative emotional and behavioral reactions (e.g., anger, sadness, behavior dysregulation) from children (e.g., Cummings et al., 2003; Goeke-Morey et al., 2003). Destructive marital conflict tactics may make children more vulnerable to developing adjustment problems (Buehler et al., 2007; Katz & Low, 2004; Lindsey, Cowell, Frabutt, & MacKinnon-Lewis, 2006). Constructive marital conflict, on the other hand, may beneficially impact children by teaching them problem solving skills, effective ways of communicating, and ultimately fostering more positive social relations (McCoy, Cummings, & Davies, 2009; Goodman, et. al, 1999). These findings support the importance of examining marital conflict as a multidimensional construct.

Marital Conflict and Parenting

Although a substantial body of research illustrates that both constructive and destructive conflict impact children, a consensus is that research must advance beyond examining the direct effects of conflict on children (Cox & Paley, 2003; Grych & Fincham, 2001). Family systems theory holds that understanding family influences on children’s development requires taking into account multiple family relationships and their interrelations, including both the marital and parent-child systems (Cox & Paley, 2003). This perspective highlights the importance of incorporating parent-child relations when attempting to understand the underlying pathways between constructive and destructive marital conflict and child functioning.

Stemming from family systems theory, the spillover hypothesis suggests that the positivity or negativity experienced in the interparental relationship may spill over or transfer into the parent-child relationship (Erel & Burman, 1995). Marital partners who are satisfied and receive support from their spouse tend to be more available and responsive to their children (Easterbrooks & Emde, 1988), whereas partners who are plagued by destructive marital conflict tend to lack emotional availability and are less responsive to their children’s needs (Sturge-Apple, Davies & Cummings, 2006). The notion of the coparenting relationship also helps to explain the link between distressed marriages and difficulties in the parent-child relationship. Research suggests that the coparenting relationship is disrupted in distressed marriages, contributing to less parental involvement and negative parenting behaviors (Margolin, Gordis, & John, 2001; McHale, 1995).

As research has primarily focused on destructive marital conflict, limited research has investigated the impact of constructive marital conflict on parenting. Specifically, few studies have addressed whether constructive marital conflict may foster “positive spillover,” resulting in more positive parent-child interactions. Consistent with this perspective, Rinaldi and Howe (2003) found that constructive conflict was associated with positive outcomes across multiple systems in the family.

Notably, parenting also consists of multiple dimensions. Parenting is often classified into three broad categories: warmth, psychological control (e.g., intrusiveness, instilling anxiety, controlling through guilt), and behavioral control/discipline (Barber, Maughan, & Olsen, 2005; Shaefer, 1965). Each of these dimensions of parenting has been related to children’s socioemotional development (Cummings, Davies, & Campbell, 2000). As studies have demonstrated that marital conflict is associated in varying ways with specific parenting behaviors (e.g., Buehler & Gerard, 2002; Krishnakumar & Buehler, 2000), it is important to examine multiple dimensions of parenting to understand how exactly marital conflict impacts parenting. For example, in a meta-analysis, Krishnakumar and Buehler (2000) found that marital conflict was most strongly associated with harsh parenting and acceptance, as opposed to other types of parenting. Moreover, specific domains of parenting are associated with different child outcomes (e.g., Buehler, Benson, & Gerard, 2006; Palmer & Hollin, 2001; Schoppe-Sullivan, Schermerhorn, & Cummings, 2007). Schoppe-Sullivan and colleagues, for example, found that parental warmth was negatively related to children’s externalizing symptoms, whereas parental behavioral control was positively associated with internalizing symptoms.

Furthermore, research is mixed regarding gender differences in relations between marital conflict and parenting (e.g., Belsky, Gilstrap, & Rovine, 1984; Buehler et al., 2006; Cox & Paley, ey, 2003; Kaczynski, Lindahl, & Laurenceau, 2006). Some research suggests that mothers and fathers are equally susceptible to the negative effects of marital conflict on parenting (Cummings, Merrilees, & George, 2010). Yet, there is also evidence suggesting that fathers’ parenting may be more vulnerable (i.e., the father vulnerability hypothesis). Fathers may have a harder time differentiating their roles as husband and father, resulting in emotions in one relationship transferring directly over into other familial relationships (Belsky, et al., 1984).

Child Adjustment: School and Social Functioning as Outcome Variables

School and social adjustment are particularly significant outcomes for children in early childhood as they are indicative of stage-salient developmental tasks (Bascoe, Davies, Sturge-Apple, & Cummings, 2009). During this particular developmental period, children are spending more time away from home; they are increasingly immersed in school and peer activities (Gross, Shaw, & Moilanen, 2008; Rimm-Kauffman & Pianta, 1999). Early school and social adjustment are also related to later trajectories of children’s socioemotional development, such as emotional, psychological and academic functioning (Meagher, Arnold, Doctoroff, Dobbs, & Fisher, 2009; Troop-Gordon & Ladd, 2005). Understanding factors that influence these developmental processes in early childhood may be valuable in preventing later socioemotional difficulties. Previous studies suggest that marital conflict is indirectly related to children’s social and school adjustment through parenting, including friendship quality, peer relations, and social competence (Bascoe et al., 2009; Katz & Low, 2004; Lindsey, Caldera, Ladd & Tankersley, 2009), as well as academic achievement and attention difficulties (Bascoe et al., 2009; Davies, Woitach, Winter, & Cummings, 2008). Additionally, parental warmth, psychological control and inconsistent discipline are all associated with children’s school and social adjustment (e.g., Casas, Weigel, Crick, Ostrov, Woods, et al., 2006; Doyle & Markiewicz, 2005; Wentzel, 1994).

Reciprocal Links Among Family Processes

Family systems theory stresses considering alternative, reciprocal pathways between variables, as family systems are interlinked (Cox & Paley, 2003). Although marital conflict is directly linked to the parent-child relationship, the parent-child relationship may also influence the marital relationship (Cox & Paley, 2003; Steinberg & Silverberg, 1987; Schoppe-Sullivan, Mangelsdorf, Forsch, & McHale, 2004). In addition, the child may impact family subsystems. Children’s behavioral problems, for example, have repeatedly been shown to impact parenting behaviors (Burke, Pardini, & Loeber, 2008; Hipwell, Keenan, Kasza, Loeber, Stouthamer-Loeber, & Bean, 2008; Larsson, Viding, Rijsdijk, & Plomin, 2008). Moreover, children’s responses to marital conflict may over time affect interparental conflict (Schermerhorn, Cummings, DeCarlo, & Davies, 2007). It is unclear if children’s school and social adjustment may impact family functioning.

Current Study

Longitudinal relations between constructive and destructive marital conflict, mothers’ and fathers’ parenting (i.e., psychological control, warmth, and inconsistent discipline) and children’s school and social adjustment in early to middle childhood were examined. This particular developmental stage is critical to study, as early and middle childhood school and social competence are developmental markers of children’s later adjustment (Meagher, et al., 2009; Troop-Gordon & Ladd, 2005). The following aims were addressed: (1) how specific dimensions of mothers’ and fathers’ parenting are impacted by constructive and destructive marital conflict behaviors, respectively, (2) how various family processes (i.e., marital conflict, parenting) influence children’s social and school adjustment, and (3) reciprocal links between constructs. To further strengthen the aims of this study, autoregressive controls of children’s previous adjustment were included.

For the first aim, we expected that, over time, destructive marital conflict would negatively impact parenting, whereas constructive conflict would positively impact parenting. Both mothers’ and fathers’ parenting were expected to be impacted by marital conflict. If in fact differences in vulnerabilities were found, it was expected that fathers’ parenting would be more vulnerable, due to previous literature related to the father vulnerability hypothesis (e.g., Belsky, et al., 1984). For the second aim, positive dimensions of parenting were expected to be associated with positive school adjustment and lower levels of child social problems. Moreover, for the third aim, reciprocal links were expected given the transactional nature of family relations. Children’s school and social adjustment were expected to be associated with more positive parenting and less negative parenting, whereas children’s difficulties in adjustment were expected to have the inverse effect. Mothers’ and fathers’ parenting was also expected to influence marital conflict, with positive parenting being associated with constructive marital conflict and negative parenting being associated with destructive marital conflict.

Method

Participants

Participants included 235 families recruited from a moderate-sized metropolitan area in the Northeast and a small city in the Midwest (see Davies, Sturge-Apple, Cicchetti, & Cummings, 2007; McCoy et al., 2009; for other studies that have used this sample). Families were recruited using flyers, postcards, and advertisements dispersed through local community events and agencies, such as daycare providers. Families were eligible to participate if couples had been living together for at least three years, had a child enrolled in kindergarten, and could complete questionnaires in English.

The majority of couples were married (89%; N = 209) and had been living together for an average of 11 years (SD = 4.9). At Time 1, mean age for mothers was 35 (SD = 5.6), 36.8 (SD = 6.2) for fathers, and 6 years old for children (107 boys; 129 girls; range from 5 – 7 years old; SD = .5). Participants were 77% European American, 16.2% African American, and 6.8% of participants reported other races. Mean family income range was between $40,000 and $54,999 a year. Ninety-five percent (N = 224) of female participants and 89% (n = 209) of male participants were the biological parents. The majority of families participated in all three waves, 95% (N = 224) retention at Time 2 and 91% (N = 214) at Time 3. Independent samples t-tests were conducted comparing families who dropped out versus those who were retained to assess group differences. Numerous demographic characteristics at Time 1 (e.g., family income, race/ethnicity), as well as study variables at Time 1 (e.g., marital conflict, parenting, child adjustment), were examined. Only three significant differences emerged: mothers’ use of psychological control, fathers’ use of psychological control and parents’ report of children’s exclusion by peers. Scores on all 3 of those variables were higher for those that dropped out. As a result, participation was controlled for in final analyses.

Procedure

Data for this study were collected across three time points (equally spaced one year apart) when children were in kindergarten, first, and second grade. This study was conducted under the approval and direction of the University’s Institutional Review Board. Both parents and children provided consent and assent, respectively, before participating. At each time point, data collection was divided into two visits. Mother, father, and child participated in the first visit, whereas only the mother and child participated in the second. Participation by fathers in laboratory visits is traditionally difficult to achieve; therefore, the study design only included fathers in the first visit to reduce the likelihood of attrition. Fathers, however, did complete all of the questionnaires and tasks directly related to the central aims of the study (i.e., marital conflict, parenting, child adjustment). Each visit lasted approximately 2.5 hours. During the first visit, mothers, fathers and child each independently completed interviews and questionnaires. At the first and third waves, mothers and fathers also engaged in two 10-minute marital problem solving interactions. During these marital interactions, mothers and fathers selected and discussed two topics that they identified as problematic in their relationship. They were instructed to work toward finding a resolution to the problem. These interactions were videotaped and later coded for specific marital conflict tactics. During the second visit at each of the three measurement occasions, mother and child completed additional questionnaires and interviews. At each wave, parents consented to the participation of their child’s primary teacher to complete assessments about the child’s school adjustment (e.g., school liking, behavior in school). One hundred percent of teachers at Time 1 and 87% (N = 182) of teachers at Time 3 participated in this study. Families were compensated for their participation ($130 = Time 1, $160 = Time 2, $190 = Time 3).

Constructive and destructive marital conflict definitions were derived from analogue and diary studies (Cummings et al., 2002, 2003, 2004; Goeke-Morey et al., 2003), as well as intervention studies conducted by Markman and colleagues (1993). Specifically, cooperation, resolution, problem solving, and support were identified as constructive marital conflict behaviors, as studies have demonstrated that children tend to have positive emotional reactions to these marital conflict tactics. By contrast, physical aggression, nonverbal anger, verbal anger, and withdrawal were identified as destructive conflict behaviors, as research has shown children have negative emotional reactions. Additionally, prevention studies in this area have also conceptualized marital conflict similarly (Markman et al., 1993).

Measures

Destructive marital conflict

At wave one and wave three, couples independently completed the Conflict Problem Solving Scales (CPS; Kerig, 1996), which is a measure designed to assess marital conflict tactics. For the purposes of this study, only the physical aggression subscale (seven items; e.g., “Slap partner”) was used as an indicator of destructive marital conflict. Couples reported on their own and their partner’s use of the conflict behaviors, resulting in both a self- and partner-report for each conflict tactics (e.g., wife report of self and husband report of wife) that was summed. These composite individual scores were then averaged across husband and wife to provide a conflict score for the couple (correlation between husbands and wives scores = .59). Reliability coefficients for both mothers’ and fathers’ reporting of physical aggression was .91.

Couples also completed the O’Leary Porter Scale (OPS; Porter, & O’Leary, 1980) at wave one and wave three to assess children’s exposure to marital hostility (10 items; e.g., “How often do you and/or your partner display verbal hostility (raised voices, etc.) in front of your child?”). Couples’ individual subscale scores were averaged to create a score for children’s total exposure to marital hostility (correlations between mothers’ and fathers’ scores = .54). Higher scores indicated greater exposure to hostile marital conflict. The reliability coefficient for mothers’ and fathers’ report was .86.

Observational assessments collected at waves one and three were used to create a multi-method, multi-dimensional assessment of destructive marital conflict. Specifically, based on definitions derived from the Marital Daily Records Protocol (Cummings et al., 2002), marital interactions completed in the lab were coded for both verbal anger (saying something in angry tone of voice or discussing angry content) and nonverbal anger (expressing frustration without saying words, such as rolling eyes or angrily sighing). Two research assistants were extensively trained on the coding system, achieving adequate reliabilities on all coding categories (i.e., intraclass coefficients were .88 for females’ nonverbal anger and 1.0 for males’ nonverbal anger; .97 for females’ verbal anger and 1.0 for males’ verbal anger) based on independently coded interactions. Coders rated destructive behaviors every 30 seconds throughout the 10-minute interactions, based on a scale from 0 = “did not occur” to 2 = “strongly occurred.” Scores were summed across both interactions to create a separate score for both mothers’ and fathers’ use of each of the destructive behaviors, and then mothers’ and fathers’ behaviors were averaged to create one couple score for both nonverbal and verbal anger. Correlations between mothers’ and fathers’ scores were .78 for verbal anger and .16 for nonverbal anger.

Constructive marital conflict

Cooperation (six items; e.g., “Talk it out with the other one”) and resolution (13 items; e.g., “We feel that we’ve resolved it, or come to an understanding”) subscales from the Conflict and Problem Solving Scales (CPS; Kerig, 1996), which was collected at waves 1 and 3, were used to assess elements of constructive marital conflict. Couples reported on both their own and their partner’s use of the conflict behaviors, resulting in both a self- and partner-report (e.g., wife report of self and husband report of wife) that was summed for each conflict tactic. These composite individual scores were then averaged across husband and wife to provide a conflict score for the couple. Correlations for mothers’ and fathers’ scores were the following: cooperation = .64; resolution = .53. Reliability coefficients for mothers’ and fathers’ reports were .89 for cooperation and .90 for resolution.

Observational assessments collected at waves one and three were also used to create a multi-method, multi-dimensional measure of constructive marital conflict. Specific codes utilized in this study were support (reassuring spouse that one is listening or supporting spouse’s statement) and problem solving (suggesting a possible solution or compromise) as indicators of constructive marital conflict (see description under destructive marital conflict for a more detailed description of the coding protocol). Correlations for mothers’ and fathers’ scores were the following: problem solving = .62; support = .26. Reliability coefficients were .83 for females’ problem solving; 93 for males’ problem solving; .94 for females’ use of support; .89 for males’ use of support.

Parental warmth

The Parental Acceptance-Rejection Questionnaire (PARQ; Rohner, 1990) collected at wave two, which consists of 20 items, assessed parents’ use of warm/accepting behaviors (higher scores representing more warmth; e.g., “You say nice things about your child”). Mothers and fathers reported on both their use of warmth, as well as their partner’s use of warmth. Self and partner reports of each individual were then averaged together to create two manifest variables: mothers’ parenting (i.e., mother report of self and father report of mother) and fathers’ parenting (i.e., father report of self and mother report of father). Correlations between self and partner report were the following: mothers’ warmth = .33; fathers’ warmth = .31. Reliability coefficients in this sample for mothers were .95 and fathers were .96.

Parental inconsistent discipline

The Alabama Parenting Questionnaire (APQ; Shelton, Frick, & Wootton, 1996) collected at wave 2, assessed parents’ use of inconsistent discipline (e.g., “You threaten to punish your child and then do not actually punish him/her”). Self- and partner-report was collected; six items were summed and higher scores represent greater use of inconsistent discipline practices. Reliability coefficients for mothers in this sample were .79 and .78 for fathers. Consistent with previous measures, self and partner reports were averaged to create separate manifest scores for mothers’ inconsistent discipline and fathers’ inconsistent discipline. Correlations between self and partner report were the following: mothers’ inconsistent discipline = .38; fathers’ inconsistent discipline = .28.

Parental psychological control

Mothers and fathers completed the Parent Version of the Child Report of Parenting Behavior Inventory (PV-CRPBI; Margolies & Weintraub, 1977) during wave 2, which measures parent’s use of psychological control (e.g., “You talk to your child again and again about anything s/he does”). All 15 items were summed and higher scores represented greater use of psychological control. Reliability coefficients for both mothers and fathers in this sample were .90. Again, self and partner reports were averaged to create separate manifest scores for mothers and fathers. Correlations between self and partner report were the following: mothers’ psychological control = .31; fathers’ psychological control = .30.

Children’s positive school adjustment

Children’s teachers completed the Teacher Rating of School Adjustment at waves one and three (TRSSA; Birch & Ladd, 1997), which assessed multiple aspects of children’s school adjustment, including cooperation (seven items; e.g., “Follows teacher’s directions”), self-directedness (four items; e.g., “Works independently”), and school liking (five items; e.g., “Likes to come to school”). Each of these subscales was used as separate indicators for children’s school adjustment. Reliability coefficients for subscales ranged from .65 to .85.

Children’s negative social adjustment

Children’s mothers and fathers separately completed the Child Behavior Scale at waves one and three (CBS; Ladd & Proflit, 1996). This assessed aspects of children’s social adjustment, including prosocial behavior (seven items; e.g., “Helps other children”), asocial behavior (six items; e.g., “Likes to be alone”), and exclusion by peers (six items; “Peers refuse to let this child play with them”). ). Items were summed initially to create two separate scores for mother and father report, which were then summed and averaged together to create one score for parental report of children’s school adjustment. Correlations between mothers’ and fathers’ report were as follows: prosocial behavior = .37; asocial behavior = .49; exclusion by peers = .46. Reliability coefficients for mothers’ and fathers’ report on this measure were .79 for prosocial behavior, .79 for asocial behavior, and .83 for exclusion by peers.

Plan of Analyses

Structural equation modeling analyses were conducted using AMOS 16.0 (Arbuckle & Wothke, 1999), which uses full information maximum likelihood estimation in the presence of missing data. Models included a combination of latent variables and manifest variables; latent variables were used when at least three manifest assessments were available (i.e., constructive and destructive marital conflict, school adjustment, and social adjustment). Notably, for both marital conflict latent variables, there was significant variability in the factor loadings, in that questionnaire data loaded significantly higher than observational data. Varying factor loadings are commonly seen in multi-method studies (e.g., Bank, Burraston, & Snyder, 2004; Davies, et al., 2007; Majdandzic & van den Boom, 2007). Although the observational data had much lower factor loadings, they were all significant. Therefore, the decision was made to include both the questionnaire and the observational data as a means of capturing a multi-method approach, which allows for a more valid assessment of marital conflict. All parenting behaviors were manifest variables, as only two indicators of behaviors were observed in the study. Although traditionally, the χ² statistic is used to examine the model fit for the data, this statistic has been criticized for being biased in large samples. Thus, additional fit indices were used to examine model fit, including the χ²/df, comparative fit index (CFI), and root mean square error of approximation n (RMSEA; Arbuckle & Wothke, 1999). Notably, as the focus of the study was on the impact of marital conflict on family processes, parenting was only explored at Time 2.

Results

All manifest variables comprising latent constructs were significantly correlated supporting the use of the latent constructs (see Table 1). Also, correlations among the parenting variables were moderately associated, providing justification for examining parenting behaviors as separate constructs.

Table 1.

Intercorrelations and Descriptive Statistics for Study Variables (N = 236)

	Kindergarten								First Grade						Second Grade
Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
Time 1-Conflict
Child Age: Kindergarten
1. Cooperation	1
2. Resolution	.69*	1
3. Support	.15*	.11	1
4. Problem Solving	.15*	.16*	.22*	1
5. Hostility	-.40*	-.56*	-.07	-.09	1
6. Physical Aggression	-.38*	-.51*	-.09	-.17*	.51*	1
7. Verbal Anger	-.30*	-.42*	.13	-.16*	.28*	.25*	1
8. Nonverbal Anger	-.09	-.07	-.22*	-.25*	.18*	.17*	.17*	1
Time 2 -Parenting
Child Age: First Grade
9. Mom Warmth	.26*	.27*	-.02	.08	-.09	-.12	-.02	-.15	1
10. Dad Warmth	.31*	.27*	.09	.15*	-.15*	-.15*	-.07	-.15*	.72*	1
11. Mom Discipline	.23*	-.23*	-.10	-.11	.31*	.16*	-.01	.13	-.22*	-.22*	1
12. Dad Discipline	-.07	-.06	-.06	-.04	.20*	.18*	-.09	.10	-.09	-.10	.63*	1
13. Mom Psych Control	-.11	-.03	-.10	-.05	.13	.14*	-.00	.06	.13	.08	.27*	.28*	1
14. Dad Psych Control	.03	.07	-.11	-.04	.03	.09	-.01	.04	.25*	.29*	.18*	.28*	.83*	1
Time 3 -Adjustment
Child Age: Second Grade
15. Peer Exclusion-PR	-.08	-.10	-.01	-.12	.06	.12	.01	-.08	-.13	-.13	.06	.18*	.09	.12	1
16. Prosocial-PR	.11	.08	.05	.06	-.04	.01	.05	-.02	.37*	.34*	-.20*	-.16*	-.08	.03	-.27*	1
17. Asocial-PR	.02	-.05	-.05	-.06	.00	.06	.03	-.10	-.02	-.08	.05	.16*	-.07	-.06	.53*	-.18*	1
18. School Liking-TR	-.15*	-.06	-.01	-.06	.09	.06	.02	.05	.09	.06	-.20*	-.17*	-.13	-.12	-.26*	.15	-.13	1
19. School Coop.TR	-.04	.00	.03	-.01	.01	.04	-.01	.01	.09	.03	-.08	-.13	-.21*	-.24*	-.27*	.21*	-.04	.60*	1
20. Self-Direct-TR	-.05	-.01	.04	.05	-.00	-.02	-.00	.00	.13	.09	-.06	-.12	-.20*	-.24*	-.30*	.21*	-.02	.50*	.64*	1
Means	14.58	4.18	.77	5.16	13.61	1.73	1.43	.88	176.08	166.77	27.53	26.05	28.40	25.97	2.32	5.31	2.66	9.14	10.24	7.29
Standard Deviations	2.22	10.62	1.13	4.37	4.35	2.01	5.02	1.13	13.95	16.91	5.33	5.45	4.40	4.64	.46	.56	.57	1.69	2.48	2.56

Note: PR = Parent Report; TR = Teacher Report

Marital Conflict and Parenting: Examining Different Pathways for Mothers and Fathers

To address aim one, a model was constructed to include constructive and destructive marital conflict simultaneously at Time 1, as well as both mothers’ and fathers’ parenting behaviors (warmth, psychological control, inconsistent discipline) simultaneously at Time 2 (see Figure 1). This model allowed for constructive and destructive marital conflict, as well as mothers’ and fathers’ parenting to be accounted for, resulting in a stringent test of the relationship among these variables. Variables measured at the same time points were allowed to covary with each other (correlations ranged from .08 to .83). Model fit indices suggested that this model adequately represented the data (χ² (55) = 108.19, p < .05; χ²/df = 1.97, CFI = .95, RMSEA = .06). Consistent with hypotheses, constructive marital conflict was associated positively with warm parenting for both mothers (β = .57, SE = 1.71, p < .01) and fathers (β = .44, SE = 1.94, p < .05), whereas destructive marital conflict positively related to inconsistent discipline for fathers only (β = .54, SE = .39, p < .05).

Figure 1.

Relationship between constructive and destructive marital conflict and maternal and paternal parenting behaviors (N = 236).

Note. f denotes variables with factor loadings fixed to equal 1. * p < .05. ** p < .01. Specific correlations between parenting variables can be found in Table 1(correlation table). R² = amount of variance explained

Marital Conflict, Parenting and Children’s School Adjustment and Social adjustment

Next, to address aim two, longitudinal models examined the impact of marital conflict and parenting on children’s social and school adjustment. Model complexity was limited due to the size of the sample. Therefore, two models were fit to examine relations for r mothers and fathers separately, as it is suggested that studies should have ten participants per indicator. This would ultimately necessitate a sample size of at least 260 (Nunally, 1967) in order to estimate both mothers and fathers parenting simultaneously. Each model included pathways between constructive and destructive marital conflict at Time 1, specific parenting behaviors at Time 2, and children’s positive school adjustment and negative social adjustment at Time 3. Time 1 social and school adjustment were included in the models for more stringent longitudinal examination. Constructs measured at the same time were allowed to covary with one another; specific correlations between the latent constructs can be seen in the figures. Correlations between parenting variables were not reported in the figures for clarity; those correlations are in the correlation table (see Table 1). Reporters of child adjustment from Time 1 to Time 3 were also allowed to covary to reduce monoreporter bias; these specific correlations can be seen in the correlation table as well (see Table 1).

As shown in Figure 2, for mothers, constructive marital conflict at Time 1 was related to warm parenting at Time 2 (β = .58, SE = 1.67, p < .01). Controlling for prior levels of children’s school and social adjustment at Time 1, warm parenting was negatively related to children’s poor social adjustment at Time 3 (β = -.13, SE = .01, p < .05), whereas use of psychological control was negatively related to children’s positive school adjustment at Time 3 (β = -.21, SE = .02, p <.01). As a follow-up, the significance of the pathway from constructive marital conflict to children’s social adjustment via mothers’ warm parenting was assessed using the Sobel method. Findings suggest that the pathway is only marginally significant (z = -1.66, p = .098). This model adequately fit the data (χ² (217) = 344.52, p < .05; χ²/df = 1.59, CFI = .91, RMSEA = .05).

Figure 2.

Relationship between constructive and destructive marital conflict, maternal parenting behaviors and children’s school and social adjustment accounting for children’s prior adjustment levels (N = 236).

Note. f denotes variables with factor loadings fixed to equal 1. * p < .05. ** p < .01. Specific correlations between parenting variables can be found in Table 1(correlation table).

For fathers, as s shown in Figure 3, constructive marital conflict at Time 1 was also linked with warm parenting at Time 2 (β = .46, SE = 2.01, p < .05). Additionally destructive marital conflict at Time 1 was associated with the use of inconsistent discipline at Time 2 (β = .52, SE = .40, p < .05). Also, as with mothers, psychological control at Time 2 was negatively linked to children’s positive school adjustment at Time 3 (β = -.21, SE = .02, p < .05). This model adequately fit the data (χ² (217) = 335.14, p < .05; χ²/df = 1.54, CFI = .91, RMSEA = .05).

Figure 3.

Relationship between constructive and destructive marital conflict, paternal parenting behaviors and children’s school and social adjustment accounting for children’s prior adjustment levels (N = 236).

Note. f denotes variables with factor loadings fixed to equal 1. * p < .05. ** p < .01. Specific correlations between parenting variables can be found in Table 1(correlation table).

Reciprocal model: Child Adjustment Impacting Parenting and Marital Conflict

To address the third aim, an alternative reciprocal model was tested, in which children’s school and social adjustment were included at Time 1 as predictors of parenting at Time 2, as well as constructive and destructive marital conflict at Time 3 (see Table 2). Thus, models examined the impact of children’s school and social adjustment on parenting, and the influence of these on constructive and destructive marital conflict over time. Autoregressive controls of marital conflict at Time 1 were also included. Notably, fathers, psychological control at Time 2 was positively associated with destructive marital conflict at Time 3 (β = .16, SE = .04, p < .05), and fathers’ use of inconsistent discipline at Time 2 was positively associated with destructive marital conflict at Time 3 (β = .16, SE = .03, p < .05). The model adequately fit the data (χ² (242) = 390.38, p < .05; χ²/df = 1.61, CFI = .90, RMSEA = .05). For mothers, only a significant relationship existed between inconsistent discipline at Time 2 and destructive marital conflict at Time 3 (β = .15, SE = .03, p < .05). This model adequately fit the data (χ² (242) = 388.54, p < .05; χ²/df = 1.61, CFI = .90, RMSEA = .05).

Table 2.

Bidirectional Model Assessing Children’s School and Social Adjustment as Predictors of Parenting and Marital Conflict (N = 236)

Fathers’ Parenting Model	β	SE
T1 Social Adjustment → T2 Father’s Warmth	-.17	3.76
T1 Social Adjustment → T2 Father’s Inconsistent Discipline	.03	1.18
T1 Social Adjustment → T2 Father’s Psychological Control	.04	.99
T1 School Adjustment → T2 Father’s Warmth	.09	1.67
T1 School Adjustment → T2 Father’s Inconsistent Discipline	-.12	.54
T1 School Adjustment → T2 Father’s Psychological Control	-.14	.46
T1 Constructive Marital Conflict → T3 Constructive Marital Conflict	.63**	.08
T1 Destructive Marital Conflict → T3 Destructive Marital Conflict	.83**	.07
T1 Social Adjustment → T3 Constructive Marital Conflict	.09	.30
T1 Social Adjustment → T3 Destructive Marital Conflict	-.05	.13
T1 School Adjustment → T3 Constructive Marital Conflict	-.02	.57
T1 School Adjustment → T3 Destructive Marital Conflict	-.04	.26
T2 Father’s Warmth → T3 Destructive Marital Conflict	.03	.01
T2 Father’s Warmth → T3 Constructive Marital Conflict	.12	.01
T2 Father’s Inconsistent Discipline → T3 Destructive Marital Conflict	.16*	.03
T2 Father’s Inconsistent Discipline → T3 Constructive Marital Conflict	-.10	.02
T2 Father’s Psychological Control → T3 Destructive Marital Conflict	.16*	.04
T2 Father’s Psychological Control → T3 Constructive Marital Conflict	-.03	.02
Mothers’ Parenting Model
T1 Social Adjustment → T2 Mother’s Warmth	-.07	2.87
T1 Social Adjustment → T2 Mother’s Inconsistent Discipline	.02	1.09
T1 Social Adjustment → T2 Mother’s Psychological Control	.08	.91
T1 School Adjustment → T2 Mother’s Warmth	.12	1.38
T1 School Adjustment → T2 Mother’s Inconsistent Discipline	-.13	.53
T1 School Adjustment → T2 Mother’s Psychological Control	-.08	.43
T1 Constructive Marital Conflict → T3 Constructive Marital Conflict	.65**	.08
T1 Destructive Marital Conflict → T3 Destructive Marital Conflict	.82**	.07
T1 Social Adjustment → T3 Constructive Marital Conflict	.07	.28
T1 Social Adjustment → T3 Destructive Marital Conflict	-.05	.54
T1 School Adjustment → T3 Constructive Marital Conflict	-.01	.13
T1 School Adjustment → T3 Destructive Marital Conflict	-.06	.26
T2 Mother’s Warmth → T3 Destructive Marital Conflict	.10	.01
T2 Mother’s Warmth → T3 Constructive Marital Conflict	.02	.01
T2 Mother’s Inconsistent Discipline → T3 Destructive Marital Conflict	.20**	.04
T2 Mother’s Inconsistent Discipline → T3 Constructive Marital Conflict	-.07	.02
T2 Mother’s Psychological Control → T3 Destructive Marital Conflict	.09	.04
T2 Mother’s Psychological Control → T3 Constructive Marital Conflict	.00	.02

Note. T = Time;

^*p < .05.

^**p < .01.

Discussion

Although many studies have explored associations between marital conflict and parenting (e.g., Erel & Burman, 1995; Shoppe-Sullivan et al., 2007), the study of how constructive and destructive marital conflict, respectively, differentially impact multiple ple aspects of parenting and child adjustment is a new direction. This study incorporated both constructive and destructive marital conflict, as well as multiple dimensions of mothers’ and fathers’ parenting behaviors, in an attempt to capture a broader family systems perspective. Exploring all parenting behaviors simultaneously allowed for alternative parenting behaviors to be accounted for, creating less bias in analyses and more precision in identifying the relations among the constructs. Findings supported links among dimensions of marital conflict, parenting, and child adjustment over time.

Results provided some support for a positive spillover effect occurring, as constructive marital conflict was related to higher levels in warm parenting for both mothers and fathers. Similar to a negative spillover effect and consistent with family systems theory, positive emotions from one family relationship may transfer across family relationships (Cox & Paley, 2003; Rinaldi & Howe, 2003). Findings also parallel recent prevention research. Increases in parent’s use of constructive marital behavior are in turn related to increases in positive parenting practices overtime (Cummings, Faircloth, Mitchell, Cummings, & Schermerhorn, 2008).). Notably, the precise underlying mechanisms accounting for this positive transference needs further exploration. As previous research has shown that co-parenting helps to account for the associations between destructive marital conflict and less positive parenting behaviors (Katz & Gottman, 1996), co-parenting may be one of the underlying mechanisms. Constructive marital conflict may contribute to parents’ ability to effectively coparent, which in turn may lead to more positive parenting behaviors.

Findings also provided some evidence that the association between marital conflict and specific parenting behaviors may in fact differ based on parent gender. Whereas positive relations between constructive marital conflict and mothers’ positive parenting were found, no support emerged for destructive conflict contributing ing to vulnerabilities in mothers’ parenting. Previous research has suggested that mothers may be better able to compartmentalize their roles as parent versus spouse (Thompson & Walker, 1989). When faced with destructive marital conflict, mothers may still able to differentiate their roles in the family, resulting in mothers still being available caregivers to their children.

In contrast, for fathers, destructive marital conflict was linked to an increase in inconsistent discipline, which parallels findings in other studies (Belsky, Youngblade, Rovine, & Volling, 1991; Parke & Beitel, 1988). Early childrearing years may be a particularly vulnerable time for fathers, potentially making them more susceptible to the negative consequences of marital conflict. During this developmental period, fathers may have difficulty separating their roles as parent and spouse when there are challenges in the marital relationship (Belsky et al., 1984). Stemming from gender role theory (Thompson & Walker, 1989), the role of parent and childcare taker may be more clearly defined for mothers, as opposed to fathers. Ultimately this may contribute to fathers’ difficulties compartmentalizing their various family roles. Additionally, disruptions in the marital relationship may relate to problems in the coparenting relationship, which may foster inconsistent discipline practices for fathers (Sturge-Apple et al., 2006). Notably, this study only found qualified support for the fathering vulnerability hypothesis, in that only one parenting dimension appeared to be more susceptible to problems in the marital relationship, as compared to mothers. Additionally, since no specific analyses compared the coefficients from destructive marital conflict to inconsistent discipline for mothers versus fathers; caution should be used regarding interpreting the mother-father gender difference. Replicating and further examining this finding may provide further clarification.

Consistent with previous findings, psychological cal control was associated with children’s school adjustment (e.g., Casas et al., 2006) for both mothers and fathers. Psychological control may be a particularly damaging parental behavior for children’s development as a result of the manipulation and lack of boundaries between parent and child that result (Barber & Harmon, 2002). Additionally, in the context of psychological control, children tend to have difficulty creating a sense of identity and self-reliance, which are related to early childhood achievement (NICHD Early Child Care Research Network, 2008). Mothers’ and fathers’ parenting also appeared to differentially relate to children’s social adjustment. In the current study, children’s social adjustment was only influenced by maternal warmth. Even though gender roles in families have shifted over the years, women may still be primarily responsible for many aspects of day to day care giving that occurs in the early childhood years (Wood & Repetti, 2004). This day to day care giving contact may help to explain why only maternal warmth was related to children’s social adjustment.

After examining the relations of marital conflict on parenting and child adjustment, alternative analyses were conducted to explore whether children acted as active agents in the family. Although previous studies have linked children’s characteristics to difficulties in the parent-child and marital relationship (Burke, et al, 2008; Hipwell, et al., 2008; Larsson, et al., 2008), children’s school and social adjustment were not related to parenting nor marital conflict over time. Extreme behavior problems, such as externalizing or antisocial behavior, for example, may be more closely related to parenting. These behaviors may be more pronounced for parents, resulting in parents having to directly intervene and discipline children as a response to those actions. Children’s school and social adjustment may be more subtle, as parents may not have to directly discipline children as a result of their social behavior. Parenting, on the other hand, was related to marital conflict, which is consistent with family systems theory that family subsystems dynamically influence one another (Cox & Paley, 2003; Minuchin, 1985). Specifically, parents’ inconsistent discipline was related to marital conflict one year later. Disruptions in the discipline process may represent a relatively significant breakdown of the co-parenting relationship, which has been linked to marital adjustment (Schoppe-Sullivan, et al., 2004). Inconsistent discipline may also be viewed as one spouse undermining the other’s discipline practices, and parents may feel as though they do not have a sense of solidarity and common purpose (Feinberg, 2002). This may ultimately increase conflict in the marital relationship.

Notably, findings are also consistent with literature related to the father vulnerability hypothesis (e.g., Belsky, et al., 1984). Specifically, father’s use of psychological control was positively related to destructive marital conflict, whereas mothers’ use of psychological control was not. Previous studies have documented a link between distance in the father-son relationship and marital satisfaction (Steinberg & Silverberg, 1987). As father involvement is closely tied to other constructs, such as maternal stress and marital satisfaction, it may be particularly influential on marital conflict (Feldman, 2000; Levy-Shiff, 1994) more so then mothers’ parenting.

Finally, constructive and destructive conflict had qualitatively different effects on family functioning and child adjustment. Constructive marital conflict was linked to positive parenting practices, and destructive marital conflict was associated with more negative parenting practices. Thus, consistent with family systems theory, it is important to examine family relationships from a multidimensional standpoint, as differences emerged based on how family relationships were conceptualized. Findings suggest that couples who engage in positive, constructive interactions may in turn have more positive relations in multiple domains (e.g., parenting, child adjustment). Additionally, exploring marital conflict from a multidimensional perspective helps to create better specificity in linking family constructs together, as well as provides specific directions for future prevention work.

Limitations and Conclusion

Several limitations are important to consider. First, these findings should be replicated in other age groups to see how the influence of parenting behaviors on adjustment may change over time. For example, although research has documented a link between autonomy-granting and school adjustment in adolescence (e.g., Gray & Steinberg, 1999); less is known about the impact of psychological control on school adjustment in adolescence. Second, no child report was included. The inclusion of child data would have been ideal; it would have allowed for children’s perception of marital conflict to be incorporated into this study. Child report, unfortunately, was not available. Additionally, future studies should explore dimensions of destructive marital conflict in more depth (e.g., high and low expressions of negative affect during marital conflict). Also, there may be a fundamental difference between couples’ expression of anger versus contempt and the impact that each of those conflict behaviors have on the family system (Gottman, Coan, Carrere, & Swanson, 1998; Katz & Gottman, 1996).

Moreover, alternative analytical models should be considered, as using growth mixture modeling would be advantageous in allowing for couples to be grouped by level of marital conflict (e.g., mostly destructive, mostly constructive). Finally, this sample was homogenous in race/ethnicity and socioeconomic status limiting the generalizability of these findings. Although this sample is representative of both the areas where the sample was drawn from, future studies should attempt to have a more racially and socioeconomically diverse sample in order to understand the role that culture may play in these relations. ns. Previous research has documented that the impact of marital conflict on parenting and child adjustment may differ by ethnicity (Lindahl & Malik, 1999). Additionally, specific marital conflict tactics, such as physical aggression, may vary as a function of differences in cultures (McLoyd, Cauce, Takeuchi, & Wilson, 2000).

Conclusion

Despite these limitations, this study highlights the need to explore the interrelation of multiple family factors simultaneously, as guided by a family systems perspective, to further our understanding of children’s adjustment. Additionally, consistent with a developmental psychopathology perspective, this study emphasizes the need to explore both positive and negative aspects of family relationships. Emphasis is also placed on the importance of examining both mothers’ and fathers’ parenting behaviors to elucidate relations between marital conflict and children’s social and school adjustment. These findings thus advance the understanding of marital conflict and parenting utilizing a family systems explanation for children’s development.