Child and Parent Factors Predictive of Mothers’ and Fathers’ Perceived Family Functioning

Abstract

Understanding family functioning, particularly the risk and protective factors that may contribute to adaptive versus maladaptive family processes, is critical to promote child and family well-being and resilience. Guided by family systems theory and the family stress model, this study longitudinally investigated parents’ and toddlers’ individual-level challenges and parental internal resources as potential predictors of subsequent perceived family functioning (i.e., general family functioning and family conflict), while accounting for earlier couple functioning. First-time mothers’ and fathers’ reports of psychological distress, perceived child behavior problems, intrapersonal resources (i.e., coping, emotion regulation, and empathy), and couple functioning (i.e., relationship satisfaction and intimate partner violence victimization) were assessed when their child was 18 months and 4 years of age. Using autoregressive path models, results suggested that greater maternal distress and paternal perceptions of toddler behavior problems predicted later reports of perceived family dysfunction, whereas parental intrapersonal resources predicted more adaptive perceived family functioning. Overall, findings indicate mutual or bidirectional influence within the family, consistent with the family systems principle of reciprocal causality. Clinical implications are discussed, including the need to bolster parents’ existing strengths as well as providing prevention and early screening of depressive symptoms among parents.

Family therapy and other family-based psychotherapeutic approaches aim to strengthen family functioning, or the transactional patterns and processes by which the family operates and achieves its goals (Carr, 2019; Walsh, 2016). These therapies are grounded in family systems theory, which conceptualizes the family unit as an organized social system in which nested subsystems are interdependent and mutually influential (Emery, 2014). Although family functioning is associated with the theory of attachment (i.e., attachment styles between dyads such as romantic partners or parent–child), family functioning in the context of family systems theory considers system-level factors outside of individuals or dyads, such as the overall family environment (O’Gorman, 2012). However, despite the increasing interest in and use of family therapy and similar therapeutic modalities (Carr, 2019), a clear consensus on the key elements contributing to adaptive family functioning has yet to be reached (Lebow & Stroud, 2012).

Understanding and bolstering family functioning is critical to efforts that promote family and child wellness and resilience in the face of hardship (Walsh, 2016). Poor family functioning, or family dysfunction, is important to detect due to its detrimental effects on individual family members’ adjustment and dyadic relationships (Letourneau et al., 2013; Newland, 2015; Shigeto et al., 2013). Yet identifying positive family functioning and related protective factors from a strength-based approach is also meaningful (Newland, 2015; Pollock et al., 2015). Strong family functioning contributes to better child adjustment, including prosocial and self-regulatory behaviors (Renzaho & Karantzas, 2010; Shigeto et al., 2013). These outcomes highlight the need for further empirical study of family functioning, particularly factors that may contribute to or predict adaptive versus maladaptive processes.

Theoretical Issues

In addition to family systems theory, the Family Stress Model (FSM) provides a useful framework in considering factors within the family system that may impact overall family functioning. Although the FSM was originally developed to focus on the stresses of economic hardship, the FSM can be used to examine other stressors and pathways that may place parents and children at risk of negative outcomes (Masarik & Conger, 2017). The FSM postulates a series of explanatory pathways through which stress disturbs the family: A stressor evokes parent psychological distress, which may lead to couple relationship problems and disrupted parenting; disrupted parenting in turn contributes to child adjustment problems (Masarik & Conger, 2017). This chain of influence may occur through the spillover effect, in which disrupted processes in one family subsystem transfer or “spill over” to other family relationships or individuals, producing further difficulties (Pu & Rodriguez, 2020, 2021), with some indications of stronger spillover effects for mothers than fathers (Pu & Rodriguez, 2021). Throughout these processes, individual, family, and community influences may exacerbate or mitigate family stress (Masarik & Conger, 2017). For example, mothers’ and fathers’ effective parental coping strategies predict their decreased depressive symptoms over time, which in turn predict reduced child internalizing and externalizing symptoms (Wadsworth et al., 2013). The FSM therefore suggests that parents’ mental health and internal resources, couple functioning, and children’s behaviors are all relevant to consider when examining family functioning.

Contributors to Family Functioning

According to family systems theory, the couple or interparental dyad comprises the executive subsystem that sets the tone for all other family interactions and individual outcomes (Hindman et al., 2013). Thus, problems within the couple dyad, such as intimate partner violence (IPV), are likely to worsen perceptions of overall family functioning. Physical and psychological victimization by an intimate partner is associated with poorer perceived family functioning for both women and men (Heru et al., 2007). Although perceptions of family functioning do not differ between women and men of nonviolent couples, female victims of IPV report poorer family functioning than male victims of IPV, suggesting that male-perpetrated IPV may have a larger perceived impact on the family than female-perpetrated IPV (Chrysos et al., 2005). Mothers’ reports of poorer couple functioning also predict their subsequent at-risk parenting (Pu & Rodriguez, 2020). Conversely, positive couple relationship qualities such as high partner satisfaction may be protective for the family. Higher partner satisfaction is associated with greater parental warmth during family interactions (Stroud et al., 2015), with warmth or expressiveness as a common theme across family functioning models (Lebow & Stroud, 2012). Higher partner satisfaction also relates to lower parenting stress, higher use of positive dyadic coping strategies, and lower use of negative dyadic coping strategies (Sim et al., 2017), reflecting an interplay between intrapersonal and interpersonal factors.

Couple and family dysfunction may be influenced by individual family members’ personal struggles, consistent with the spillover hypothesis. For parents, endorsing mental health concerns, particularly depressive symptoms and disorders, is one of the strongest predictors of family dysfunction (Hughes & Gullone, 2008). Postpartum depression affects up to 10%–15% of mothers (Letourneau et al., 2013) and 5%–10% of new fathers (Eddy et al., 2019) and presents serious mental health challenges to affected individuals, which may reverberate to their family functioning. Mothers’ and fathers’ depressive symptoms place children at risk of adverse socioemotional and behavioral outcomes (Breaux et al., 2014; Letourneau et al., 2010; Rominov et al., 2016), perhaps due to their modeling of ineffective emotion regulation (Kerns et al., 2017). In addition to being linked to deficits in emotion regulation (Berking et al., 2014), adults’ depressive symptoms are also highly associated with impaired empathic abilities (Cusi et al., 2011) and poorer marital quality, including increased couple conflict (Cummings et al., 2005), women’s increased risk of experiencing IPV (Whisman, 2019), and hostile couple behaviors of both themselves and their partners (i.e., actor and partner effects; Sutton et al., 2017). Thus, parental personal distress and related couple relationship difficulties could threaten overall family functioning.

Like their parents, children may also experience personal difficulties that alter family and couple dynamics. Consistent with the principles of child evocative effects and reciprocal causality in family systems theory (Choe et al., 2013; Combs-Ronto et al., 2009), perceived child traits such as difficult temperamental qualities can elicit harsher or more inconsistent parental responses (Lee et al., 2013; Lengua & Kovacs, 2005; Pu & Rodriguez, 2020). Perceived child internalizing symptoms (Crawford et al., 2011; Hughes & Gullone, 2008) and externalizing behaviors (Leeman et al., 2016; Letourneau et al., 2013) both relate to increased family dysfunction; however, the cross-sectional designs used by most studies, or the omission of possible child-driven or reciprocal effects in existing longitudinal analyses (e.g., Letourneau et al., 2013), makes it difficult to parse out the directionality of these associations. For parents, the stress of raising children they perceive as difficult may also negatively impact their partner relationship, imperiling their communication or couple satisfaction (Zemp et al., 2017). Mothers’ perceptions of more toddler behavior problems predict their later report of IPV victimization (Pu & Rodriguez, 2021), indicating that mothers who perceive their children as more difficult are more likely to experience later couple discord and violence victimization. Thus, children with behavioral difficulties may be at higher risk of experiencing dysfunctional family environments.

Lastly, parents may have access to internal or intrapersonal resources that buffer against family dysfunction. Although protective factors have been studied less extensively than risk factors (Newland, 2015), evidence points to positive problem-focused coping skills as a key resource for parents (Altiere & von Kluge, 2009; Masarik et al., 2016). In a cross-sectional study, mothers and fathers who reported greater use of effective coping strategies also reported high family cohesion, a common element of adaptive family functioning (Altiere & von Kluge, 2009); however, given the cross-sectional design, the directionality of this association is unclear. In a longitudinal design, parents’ use of effective problem-solving protected against later hostile couple behaviors in times of stress, whereas adults who were less effective at problem-solving experienced the largest increases over time in couple hostility in response to stress (Masarik et al., 2016). Given this preliminary evidence, continued longitudinal investigation of parental coping skills in relation to overall family functioning is warranted. Of note, mothers and fathers report using divergent coping strategies at different frequencies (Pinelli, 2000), suggesting possible gender differences in parental coping strategy utilization that merits further inquiry.

Other intrapersonal resources of parents that may protect against family dysfunction are their emotion regulation and empathic abilities. Adults’ successful emotion regulation use predicts less severe depressive symptoms overtime, signifying a protective effect (Berking et al., 2014). In addition, children often emulate their parents’ emotion regulation strategies, suggesting that parents’ effective self-regulation not only benefits themselves, but also facilitates adaptive child emotion socialization (Hajal & Paley, 2020). Similarly, parents’ empathy is essential to maintaining family resilience (Geiger et al., 2016). Maternal empathy is associated with positive child outcomes such as prosocial behaviors (Farrant et al., 2012) and may protect against toddler behavior problems when mothers experience high stress (Walker & Cheng, 2007). Although no prior studies appear to have examined parental emotion regulation or empathy as predictors of family functioning over time, these findings suggest that such parental qualities could benefit the overall family system as well. Further investigation of such risk and protective factors will be crucial to elaborate existing theoretical models on reciprocal influence in the family, particularly in studies that can improve upon current methodology by using longitudinal designs, multiple family informants, and dyadic data analysis (Conger et al., 2010; Whisman, 2019).

The Present Study

Guided by family systems theory and the family stress model, this study investigated potential individual-level risk versus protective predictors of perceived family functioning in a diverse sample of parents and young children assessed from toddlerhood into the preschool years. Contributors to family functioning (i.e., general family functioning, family conflict, partner satisfaction, IPV victimization) included parent and child challenges separately (i.e., parent psychological distress and child behavior problems) and parent intrapersonal resources (i.e., coping, emotion regulation, and empathic abilities), while accounting for earlier couple functioning (i.e., partner satisfaction, IPV victimization). Given the historical lack of inclusion of fathers in child and family research (Barker et al., 2017), pathways were analyzed for mothers and fathers separately as well as dyadically (i.e., with parents nested as a couple) so that gender differences in perceived family functioning could be explored. Less parent psychological distress, fewer perceived child behavior problems, stronger parental intrapersonal resources, and better couple functioning when children were 18 months were hypothesized to predict better perceived family functioning at 4 years (see Figure 1 for full hypothesized model).

Figure 1. Theoretical Model of Bidirectional Relations.

Note. Bidirectional relations between parent psychological distress, perceived child behavior problems, parents’ intrapersonal resources, and couple/family functioning from child age 18 months to 4 years.

Method

Participants

Participants were enrolled in the prospective longitudinal “Following First Families” (Triple-F) Study, which tracked changes in parent–child aggression risk among first-time mothers and their male partners in the Southeast U.S. At Time 1, 203 primiparous women in their third trimester of pregnancy were recruited for a three-wave study, along with 151 of their male partners. Families were reevaluated when their children were 6 months (±2 weeks; Time 2) and 18 months old (±3 weeks; Time 3). By Time 2, two families were ineligible to continue due to the death of their infant; of the remainder, 186 mothers and 146 male partners were retained. The present study focuses on Time 3 (T3) and Time 4 (T4). By T3, one family lost custody of their child; 180 mothers with 136 male partners were included in this study. A fourth wave (T4) was later added when children were 4–4½ years old. By this time, an additional family had lost custody, and 119 mothers and 93 male partners participated. Based on Optimal Design 2.01 software (Spybrook et al., 2011), sample size was estimated using a medium effect size and an ICC = .30, indicating power exceeds .80 at sample sizes in excess of 79.

At T3, M_age was 27.6 years (SD = 5.76) for mothers and 30.7 years (SD = 6.28) for fathers. Mothers reported their racial/ethnic identity as follows: 48.9% White, 48.3% African American, 1.1% Asian, and 1.7% Native American; 3.3% of women also identified as Hispanic/Latina and 7.8% as biracial. Fathers reported their racial/ethnic identity as follows: 59.6% White and 40.4% African American; 4.4% of men also identified as Hispanic/Latino and 7.4% as biracial. Mothers reported their educational attainment as: 25.6% high school or less; 25.0% some college or vocational training; 21.1% college degree; and 28.3% > college degree. Fathers reported their education as: 28.7% high school or less; 23.5% some college or vocational training; 27.9% college degree; and 19.8% > college degree. At T3, 48% of mothers reported an annual household income ≤$40,000 and 39% of mothers reported receipt of public assistance. At T3, 87.8% of mothers reported being in an intimate partner relationship (91.6% at T4).

Measures

Identical measures were administered at T3 and T4 except for comparable, developmentally appropriate measures of child behavior problems and the new measures of perceived family functioning at T4, as described below.

Parent Psychological Distress

The Brief Symptom Inventory-18 (BSI; Derogatis & Melisaratos, 1983) is an 18-item measure of current psychological distress, in which respondents report on the frequency of depression and anxiety symptoms experienced in the past week. Items are rated on a 5-point scale from 0 (Not at all) to 4 (Extremely) and summed to create a total score. Higher total scores signify greater distress. The BSI demonstrates good convergent and factorial validity (Prinz et al., 2013). In the current sample, internal consistency was good for mothers (α = .91–.92) and fathers (α = .81–.93) across time.

Child Behavior Problems

At Time 3, the Brief Infant Toddler Social Emotional Assessment (BITSEA; Briggs-Gowan et al., 2004) was used to assess children’s socioemotional adjustment. Parents responded to 42 items about their children’s feelings and behaviors on a 3-point scale, and two items about their worries about children on a 4-point scale. The Total score comprises all 44 items, with higher scores indicating greater internalizing and externalizing problems and emotion dysregulation. The BITSEA demonstrates good test–retest reliability, interrater reliability, and 1-year stability (Briggs-Gowan et al., 2004). Reliability was good in the current sample (α = .81 for mothers, .87 for fathers).

At Time 4, the Child Behavior Checklist, preschool version (CBCL/1.5-5; Rescorla, 2005) was used to evaluate children’s internalizing and externalizing problems. On this 99-item checklist, parents rate child behaviors from the last 2 months on a 3-point scale. Items are summed, and higher total scores reflect greater problem behaviors. The preschool and original versions of the CBCL both have high individual item intraclass coefficients (>.90) and demonstrate convergent validity with other measures of child behavior problems (Achenbach & Rescorla, 2001; Rescorla, 2005). The CBCL exhibited excellent internal consistency in the current sample (α = .94 for mothers, .96 for fathers).

Parent Resources

The Coping Self-Efficacy Scale (CSES; Chesney et al., 2006) evaluates participants’ perceived ability to use problem-focused coping strategies effectively. The CSES uses 13 items rated on an 11-point scale from 0 (Cannot do at all) to 10 (Certain I can do). Items are summed and higher total scores indicate greater sense of effective coping abilities. This measure has evidenced convergent validity with other problem-focused coping measures (Chesney et al., 2006). The current sample demonstrated strong internal consistency for mothers (α = .93–.94) and fathers (α = .92–.95) across time.

The Negative Mood Regulation Scale (NMRS; Catanzaro & Mearns, 1990) is a 30-item measure of perceived negative emotion regulation abilities. Participants rate items relating to different emotion regulation strategies on a 5-point scale from 1 (Strongly Agree) to 5 (Strongly Disagree). Items are summed, and higher total scores suggest poorer emotion regulation abilities. The NMRS demonstrates good stability, concurrent, and predictive validity with negative affect (Catanzaro & Mearns, 1990), and convergent validity with other emotion regulation measures (Bardeen et al., 2016). In this sample, internal consistency was strong for women (α = .92 both time points) and men (α = .91–.92) across time.

The Interpersonal Reactivity Index (IRI; Davis, 1983) is a 14-item measure of empathic ability. The IRI uses two 7-item subscales: Empathic Concern, or affective sympathy, and Perspective Taking, or the ability to adopt others’ viewpoints. Items are rated on a 5-point scale from 1 (Describes me very well) to 5 (Does not describe me well) and summed across both subscales; higher total scores indicate greater empathy. The current sample exhibited acceptable reliability for mothers (α = .82–.86) and fathers (α = .77–.85) across time.

Couple/Family Functioning

The Revised Conflict Tactics Scale-Short Form (CTS-2S; Straus & Douglas, 2004) requires participants to estimate how frequently IPV victimization and perpetration occurred in the past year. Of the 20 items on this measure, the eight items on physical or psychological IPV victimization were used in this study. Scores are weighted as a frequency count and summed, with higher totals indicating more frequent IPV. The CTS-2S demonstrates concurrent validity with the longer, frequently utilized CTS-2 (Straus & Douglas, 2004).

The Couple Satisfaction Index (CSI; Funk & Rogge, 2007) is a 10-item measure of partner relationship satisfaction. Items are summed, and higher total scores reflect greater partner satisfaction. The CSI exhibits discriminant validity and convergent validity with other measures of dyadic, marital, or relationship satisfaction (Funk & Rogge, 2007). This sample had strong internal consistency for women (α = .97–.98) and men (α = .96–.97) across time.

In addition to the CTS-2S and CSI, two measures of family functioning were included at T4. The General Functioning (GF) subscale of the McMaster Family Assessment Device (FAD; Epstein et al., 1983) was used to assess overall family functioning quality at T4. Based on the McMaster Model of Family Functioning, the FAD is a 60-item inventory that is widely used to evaluate family functioning along six dimensions: Problem solving, communication, roles, affective responsiveness, affective involvement, and behavior control (Epstein et al., 1978). Evidence supports use of the 12-item GF subscale as a reliable and valid proxy for the lengthier FAD (Byles et al., 1988; Mansfield et al., 2015). On this measure, parents rate items on a 4-point scale from 1 (Strongly Agree) to 4 (Strongly Disagree), with higher total scores reflecting more family dysfunction. The GF subscale demonstrates high test–retest stability (Byles et al., 1988) and intraclass correlation coefficients indicate moderate agreement between husbands and wives (.67; Mansfield et al., 2015). In the current sample, the GF subscale exhibited good internal consistency (α = .90 for mothers, .86 for fathers).

The Conflict Subscale of the Family Environment Scale-Revised (FES-R; Moos & Moos, 1986) was used as a second indicator of poor family functioning at T4. The Conflict Subscale is often used in research and clinical settings to assess the frequency and intensity of overall family conflict (Kline et al., 2003). Given inconsistent use of the Conflict Subscale in prior literature, administration in this study followed recommendations to rate the nine items on a 6-point scale from 1 (Definitely false) to 6 (Definitely True) (with higher scores indicating greater conflict) rather than the original True/False format, allowing for a wider range in responses and increased sensitivity (Kline et al., 2003). Measured this way, the Conflict Subscale correlates highly with the original FES conflict subscale and exhibits similarly high test–retest reliability (Kline et al., 2003). Internal consistency was modest in this study (α = .72 for mothers, .67 for fathers).

Demographic Covariates

At each wave, parents reported their educational attainment and annual household income. Due to high correlations between income and education level (r = .60–.75, p < .001 across time), these were standardized and combined to create a composite socioeconomic status (SES) score at each time point for mothers and fathers separately.

Procedure

At study enrollment, first-time mothers were recruited with flyers at local community health centers and hospital obstetrics/gynecology clinics. Mothers in T3 and T4 were scheduled for 2- to 3-hr sessions with their male partner, if available. All measures were delivered electronically. Parents provided informed consent independently and completed the protocol in separate private rooms on laptops. The study was not preregistered. The university’s Institutional Review Board approved all study procedures.

Data Analytic Plan

Missing Data

Missing data within timepoints were minimal: Less than 1% at T3 and less than 2% at T4. Due to participant loss to follow-up, differential attrition analyses were performed to determine whether participants who did not return at T4 differed from those who participated at both T3 and T4. Chi-square tests and independent sample t tests were performed for mothers and fathers separately. Analyses indicated that participants not retained did not significantly differ on any study variables at T3 (i.e., BSI, BITSEA, CSES, NMRS, IRI, CSI, and CTS Victimization) nor on T3 sociodemographic variables such as household income, education level, age, minority status, and relationship status with the other parent.

Analyses

All data are based on measures reported in this study with no data exclusions or manipulations. Descriptive statistics and correlations were performed in SPSS 25. Hypotheses were tested with autoregressive path models in Mplus 8.1 with Full Information Maximum Likelihood (FIML) estimation, using all available data to preserve the full sample. Path models were estimated for mothers and fathers separately (see Figure 1 model) and then dyadically, using family level data with parents nested as a couple incorporating covariances between mothers and fathers’ variables to account for their nonindependence. To identify gender differences in the dyadic model, Wald statistics were used to constrain selected paths as equal for mothers and fathers. Only paths with significant effects for mothers versus fathers were tested; paths of interest were constrained one at a time.

Model fit was evaluated with root mean square error of approximation (RMSEA), standardized root-mean-square residual (SRMR), and comparative fit index (CFI). Good model fit is indicated by RMSEA and SRMR values <.08 and CFI values >.95 (Kline, 2011). All reported path coefficients are standardized.

Data Reduction

Because multiple measures were used to assess parents’ intrapersonal resources (CSES, NMRS, and IRI) and couple/family functioning (CTS-2S Victimization and CSI, with FAD and FES added at T4), data reduction was achieved by creating composite scores for use in the path models. Variables were standardized and summed to create composite scores for each. For parent resources, higher composite scores are oriented to indicate greater resources, and for couple/family functioning, higher scores indicate greater dysfunction. Confirmatory factor analyses (CFAs) performed in Mplus affirmed that all variables significantly loaded onto their respective latent factors (e.g., for mothers across time, parent resource loadings from .61 to .96, couple/family functioning loadings from .38 to .83, all p < .001; for fathers, parent resource loadings from .55 to .89, couple/family functioning loadings from .30 to .85, all p < .01). CFA model fit was good for mothers (e.g., at T3: RMSEA = .000; SRMR = .017; CFI = 1.00) and adequate for fathers (e.g., at T3: RMSEA = .091; SRMR = .041; CFI = .963). Requests for analytic code and output are available upon request.

Results

Preliminary Analyses

Sample means, standard deviations, and correlations for all variables appear in Table 1, by time point and by parent gender. With regard to the level of child behavior problems reported on the CBCL, 3.8% of parents reported problems in the borderline range, and 6.7% of parents reported problems that met or exceeded the clinical cutoff. SES and child sex were examined as possible covariates. For mothers, SES was significantly positively related to self-reported empathy (r = .22–.28, p ≤ .003) and couple satisfaction (r = .26–.38, p ≤ .005) across time, and negatively associated with perceived family dysfunction at T4 (r = −.31 to −.35, p ≤ .001). For fathers, SES was significantly positively associated with self-reported coping skills across time (r = .21–.29, p ≤ .02) and negatively associated with perceived family dysfunction at T4 (r = −.22 to −.34, p ≤ .04). Thus, SES was included as a time-varying demographic covariate in all path models. Neither mothers nor fathers reported significant differences in behavior problems between girls and boys on the BITSEA at T3 or the CBCL at T4. Therefore, child sex was not included as a covariate in the path models.

Table 1.

Means, Standard Deviations, and Correlations Between Measures

Time/Measure	Time 3							Time 4
	Parent	Child	Parent resources			Couple		Parent	Child	Parent resources			Family/couple functioning
	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.	14.	15.	16.
1. T3 BSI		.21**	−.52***	.30***	−.11	.10	−.40***	.25*	−.16	−.20	.00	−.03	.07	−.25*	.28*	.13
2. T3 BITSEA	.39***		−.18*	.22*	−.13	.30***	−.19*	.37**	.36**	−.27*	.10	−.09	.20	−.15	.18	.21
3. T3 CSES	−.44***	−.27***		−.56***	.26**	.00	.27 **	−.24*	−.24*	.46***	−.35**	.12	.02	.24*	−.28*	−.25*
4. T3 NMRS	.35***	.35***	−.59***		−.58***	.21*	−.28**	.16	−.05	−.47***	.58***	−.53***	−.02	−.28*	.34**	.31**
5. T3 IRI	−.13	−.31***	.43***	−.53***		−.11	.24**	−.02	.05	.31**	−.38***	.65***	−.05	.20	−.31**	−.37**
6. T3 CTS Vict	.31***	.30***	−.16*	.27***	−.13		−.16	.36**	.01	−.34**	−.07	−.19	.24*	−.42***	.24*	.37**
7. T3 CSI	−.27***	−.21**	.18*	−.29***	.23**	−.26***		−.14	.03	.16	−.06	.03	−.13	.70***	−.41***	−.28*
8. T4 BSI	.69***	.25**	−.32***	.29**	−.08	.08	−.11		.23*	−.51***	.29**	−.21*	.29**	−.35***	.30**	.28**
9. T4 CBCL	.45***	.29 **	−.40***	.19*	−.14	.08	−.17	.44***		−.23*	.14	−.09	.33**	.00	.21*	.25*
10. T4 CSES	−.30**	−.16	.53***	−.42***	.23*	−.18	.17	−.34***	−.25**		−.63***	.48***	−.18	.44***	−.53***	−.49***
11. T4 NMRS	.18	.09	−.56***	.51***	−.39***	.19	−.13	.28**	.27**	−.68***		−.42***	.08	−.26*	.36***	.26*
12. T4 IRI	−.02	−.14	.44***	−.48***	.64***	−.17	.19*	.04	−.09	.40***	−.65***		−.22*	.22*	−.42***	−.43***
13. T4 CTS Vict	.06	.01	−.09	−.08	.04	.30**	−.29**	.03	.28**	−.13	.02	−.05		−.17	.25*	.25*
14. T4 CSI	−.27**	−.09	.18	−.24*	.19*	−.29**	.63***	−.19*	−.18	.26**	−.18	.24**	−.29**		−.58***	−.46***
15. T4 FAD	.25**	.20*	−.45***	.42***	−.42***	.12	−.43***	.24**	.26**	−.40***	.33***	−.40***	.29**	−.58***		.61***
16. T4 FES	.22*	.14	−.34***	.29**	−.17	.27**	−.33***	.16	.31***	−.31***	.35***	−.36***	.34***	−.48***	.57***
Mothers M	5.63	13.73	96.66	65.44	55.47	6.97	47.46	6.69	26.64	96.51	61.84	57.39	5.02	47.20	20.26	22.82
Mothers SD	6.75	7.80	25.67	18.43	8.35	13.85	13.15	9.57	18.25	24.03	18.92	9.04	8.32	12.72	6.48	7.73
Fathers M	4.07	11.79	103.17	66.29	53.00	5.21	49.80	4.14	25.09	100.08	62.27	53.65	5.02	47.30	20.47	21.95
Fathers SD	6.17	7.47	22.26	16.82	7.80	8.82	10.77	5.11	19.53	22.10	18.46	8.99	7.85	11.99	5.60	6.94

Note. Mothers’ correlations below the diagonal & fathers’ correlations above the diagonal. Parent = parent challenges; Child = child challenges; Couple = couple functioning. Time 3:1 = T3 Brief Symptom Inventory-18 (BSI), 2 = T3 Brief Infant–Toddler Social and Emotional Assessment (BITSEA), 3 = T3 Coping Self-Efficacy Scale (CSES), 4 = T3 Negative Mood Regulation Scale (NMRS), 5 = T3 Interpersonal Reactivity Index (IRI), 6 = T3 Revised Conflict Tactics Scale-Short Form (CTS-2S) IPV victimization, 7 = T3 Couple Satisfaction Index (CSI). Time 4: 8 = T4 Brief Symptom Inventory-18 (BSI), 9 = T4 Child Behavior Checklist (CBCL), 10 = T4 Coping Self-Efficacy Scale (CSES), 11 = T4 Negative Mood Regulation Scale (NMRS), 12 = T4 Interpersonal Reactivity Index (IRI), 13 = T4 Revised Conflict Tactics Scale-Short Form (CTS-2S) IPV victimization, 14 = T4 Couple Satisfaction Index (CSI), 15 = T4 Family Assessment Device-General Functioning Scale (FAD), 16 = Family Environment Scale-Conflict Scale (FES).

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Path Analyses

Mothers

Findings appear in Table 2. Model fit for mothers was strong: χ²(8) = 4.78, p = .78; RMSEA = .000; SRMR = .008; CFI = 1.000. Mothers’ higher psychological distress at T3 significantly predicted their reports of more perceived child behavior problems at T4 and greater perceived family dysfunction at T4. Mothers’ greater intrapersonal resources at T3 significantly predicted their reports of less perceived family dysfunction at T4. No significant paths were identified between T3 perceived child behavior problems and subsequent functioning, or between T3 couple functioning and subsequent functioning in the fully controlled path model.

Table 2.

Standardized Coefficients for Individual and Dyadic Path Models

Parameter estimate	Individual				Dyadic
	Mothers		Fathers		Mothers		Fathers
	β	p	β	p	β	p	β	p
T3 BSI → T4 BSI	.74	.000	.20	.073	.77	.000	.23	.036
T3 BSI → T4 CBCL	.43	.000	−.24	.025	.48	.000	−.23	.039
T3 BSI → T4 resources	−.10	.291	.16	.044	−.04	.670	.14	.080
T3 BSI → T4 family	.28	.005	.08	.434	.25	.019	.13	.160
T3 BITSEA → T4 BSI	−.05	.548	.34	.006	−.06	.432	.35	.003
T3 BITSEA → T4 CBCL	.09	.387	.54	.000	.07	.534	.51	.000
T3 BITSEA → T4 resources	.14	.143	−.15	.096	.12	.219	−.17	.068
T3 BITSEA → T4 family	−.14	.210	.29	.005	−.14	.196	.25	.011
T3 resources → T4 BSI	−.08	.228	−.12	.345	−.08	.226	−.08	.531
T3 resources → T4 CBCL	−.15	.096	.01	.953	−.13	.140	.08	.551
T3 resources → T4 resources	.68	.000	.75	.000	.67	.000	.73	.000
T3 resources → T4 family	−.23	.010	−.31	.003	−.18	.037	−.22	.032
T3 couple → T4 BSI	−.06	.314	−.11	.356	−.05	.413	−.11	.328
T3 couple → T4 CBCL	.02	.801	.19	.076	.02	.846	.22	.040
T3 couple → T4 resources	.13	.084	.01	.886	.15	.043	.02	.849
T3 couple → T4 family	.07	.419	.06	.550	.01	.890	.11	.206

Note. T3 = Time 3 (18 months); T4 = Time 4 (4 years); BSI = Brief Symptom Inventory; BITSEA = Brief Infant–Toddler Social and Emotional Assessment; CBCL = Child Behavior Checklist; SES = socioeconomic status; Resources = parents’ intrapersonal resources; Couple = couple dysfunction; Family = family dysfunction. Bolded values indicate statistical significance (p < .05). All pathways include SES as a time-varying covariate. Covariances are available upon request.

Fathers

Path analyses for fathers are also in Table 2. Model fit was good: χ²(8) = 10.56, p = .23; RMSEA = .046; SRMR = .012; CFI = .993. Fathers’ perceptions of greater perceived child behavior problems at T3 significantly predicted their reports of higher psychological distress at T4 and greater perceived family dysfunction at T4. Fathers’ greater intrapersonal resources at T3 significantly predicted their reports of less perceived family dysfunction at T4. Yet fathers’ higher psychological distress at T3 also significantly predicted their reports of fewer perceived child behavior problems at T4 and greater internal resources at T4. No significant paths were identified between fathers’ T3 couple functioning and subsequent functioning in this model.

Dyadic

Results for the dyadic model also appear in Table 2. Model fit was good: χ²(95) = 161.14, p < .001; RMSEA = .061; SRMR = .084; CFI = .939. Dyadic analyses indicated similar effects to the individual models for mothers and fathers, with two exceptions. Fathers’ greater couple dysfunction at T3 significantly predicted their reports of more perceived child behavior problems at T4 in the dyadic model. On the other hand, fathers’ psychological distress at T3 became nonsignificant in predicting their internal resources at T4.

Of the seven paths tested for possible gender differences using Wald statistics, three significant differences were identified. Higher psychological distress at T3 predicted more perceived child behavior problems at T4 for mothers but fewer perceived behavior problems at T4 for fathers: Wald(1) = 19.712, p < .001. In addition, two paths were significant for fathers only: fathers’ perceptions of greater child behavior problems at T3 predicted their reports of higher psychological distress at T4, Wald(1) = 5.216, p = .02, as well as greater perceived family dysfunction at T4, Wald(1) = 7.396, p = .007. These two paths were nonsignificant for mothers.

Discussion

As guided by family systems theory and the family stress model, the present study assessed mothers’ and fathers’ reports of their psychological distress, intrapersonal resources, child behavior problems, and perceived couple/family functioning from toddlerhood into the preschool years. Results partially support the hypotheses, with evidence of significant risk and protective factors for perceived family functioning that varied for mothers versus fathers. Several gender differences were identified in the dyadic model, including stronger child evocative effects for fathers than for mothers. Overall, findings are suggestive of mutual influence in the family and highlight that perspectives and processes are distinct between mothers versus fathers.

For mothers but not fathers, greater psychological distress predicted poorer perceived family functioning, as hypothesized. Maternal findings are consistent with prior research on the deleterious effects of parent distress and depression on the family (Hughes & Gullone, 2008), whereas the absence of significant findings for fathers are inconsistent with this literature. Instead, for fathers only, perceptions of greater child behavior problems predicted their subsequent reports of greater family dysfunction and higher psychological distress, indicating child evocative effects. These father findings support the spillover hypothesis (Combs-Ronto et al., 2009), in that child challenges “spilled over” or transferred over to their fathers, resulting in further perceived difficulties in the family. Yet the absence of child evocative effects for mothers differs from prior longitudinal evidence of child-driven effects on mothers’ harsh parenting and couple functioning and stronger spillover effects for mothers than fathers (Pu & Rodriguez, 2020, 2021).

The inconsistencies of the present findings compared with previous literature can perhaps be explained by the current model’ s highly controlled design and simultaneous inclusion of protective factors (i.e., parents’ intrapersonal resources) along with risk factors. Protective or resilience factors as evident in parent strengths are less frequently assessed in models of family functioning than factors related to stress or vulnerability (Newland, 2015). Notably, the present study found that parents’ intrapersonal resources (i.e., coping, emotion regulation, and empathy) may serve a protective function against later perceived family dysfunction for both mothers and fathers. In existing literature, the ameliorating effects of parents’ strong coping, emotion regulation, and empathic abilities on child developmental outcomes are well-documented (Altiere & von Kluge, 2009; Geiger et al., 2016). However, far less research has considered the impact of parents’ internal resources on overall family functioning. Given current findings, continued investigation into these and other individual-level strengths will help broaden our understanding of family well-being and clarify clinically relevant mechanisms that can be promoted to build child and family resilience.

Perhaps the most unexpected finding of this study was that parental greater psychological distress predicted later higher perceived child behavior problems for mothers but lower perceived behavior problems for fathers. Likewise, this study also found that fathers’ psychological distress predicted subsequent reports of strong interpersonal resources while controlling for concurrent interpersonal resources and psychological distress. Results for mothers align with abundant prior literature linking parental depression with adverse child functioning (Breaux et al., 2014; Letourneau et al., 2010; Rominov et al., 2016), but the current findings for fathers contradict such patterns. Although these paternal findings initially seem counterintuitive, distressed fathers may conceivably fail to notice children’s problematic behaviors, exacerbating fathers’ tendency to underreport child difficulties in comparison to mothers (Schroeder et al., 2010), perhaps due to less involvement and direct exposure to problem behaviors. Alternatively, distressed fathers may inaccurately perceive and report on their own functioning. These surprising results highlight the need for continued inquiry on how fathers and fathering shape children’s development, including the onset and maintenance of child internalizing and externalizing disorders (Barker et al., 2017), as well as how they contribute to overall family functioning.

Limitations

Although the present study has strengths such as its use of a rigorously controlled statistical design with nested mother–father data across two timepoints, some methodological limitations should be noted. First, given our focus on parents’ perceptions of family functioning, the study relied on parents’ self-reported data and therefore may be subject to self-report or source biases related to attention, selective recall, and social desirability, particularly related to sensitive or stigmatizing information such as participants’ IPV experiences (Möricke et al., 2016). Future studies could include direct laboratory observations of child, couple, or family functioning to supplement self-reports or consider other multimethod innovations to enhance family assessment. Future work can also consider how our findings on parents’ perceived family functioning may differ when utilizing combined reports from multiple family members about their family functioning. In addition, the FAD and FES were administered at Time 4 only, as the original 3-wave study focused on child abuse risk and did not seek to track family functioning specifically. Although couple functioning measures (i.e., CSI and CTS-2S Victimization) were used to assess family conflict as a theoretical proxy for family functioning at Time 3 (Hindman et al., 2013), the measurement variance reduces the stability of the estimates and diminishes the cross-lagged aspect of this design. A similar issue arises from the different yet developmentally appropriate measures used to assess parental perceptions of child functioning at 18 months versus 4 years. To rectify this, more frequent multimethod assessment over time, with larger samples that continue to be evaluated after child age 4, would provide greater stability in model estimates and could reveal further changes in perceived family functioning as children age. Of course, future models could include any number of other potential risk or protective factors such as parental substance use or community involvement, and studies could integrate teacher informant data as children enter the school system. Lastly, despite good racial and socioeconomic heterogeneity in the sample, parents identifying as Latinx were somewhat underrepresented, and all families were recruited from the community. Replication in samples that are more culturally diverse, cross-cultural, clinical or treatment-seeking, or substantiated for IPV could help determine the generalizability of these findings.

Clinical Implications.

Overall, results suggest that both risk and protective factors at the individual-level can shape and predict perceived family functioning over time. Early family prevention and intervention services are vital, with current findings highlighting the importance of detecting and treating parents’ depression, as well as bolstering parents’ existing internal strengths and skills. Unexpected findings relating to distressed fathers’ functioning suggest that more research is needed to understand fathers’ perceptions and unique contributions to the family, particularly when fathers are impacted by stressors such as psychological disorders. Investigating paternal postpartum depression, which is currently under-screened, under-diagnosed, and under-treated (Eddy et al., 2019), will be especially important not only to support new fathers’ own well-being, but also to understand distressed men’s perceptions of and behaviors toward their children and families. Given that fathers with paternal postpartum depression tend to distance themselves from their child due to their symptomatology (Eddy et al., 2019), rapid identification and effective treatment of new fathers’ depressive symptoms could be beneficial for the whole family system, in addition to benefiting fathers themselves. Specifically, fathers may benefit from increased screening of their mental health after the birth of a child and/or when seeking treatment for a child presenting with behavioral difficulties; adaptive impacts for their children and families are likely to follow. Thoughtful consideration of these bidirectional and reciprocal influences within the family can thereby enrich future family research and promote positive well-being and long-term resilience for children and families.