This study sought to better understand individual differences in child mental health during the COVID-19 pandemic and pinpoint parent-level risk and resilience factors underlying differences to target in clinical interventions with families. Based partly on the conceptual model set forward by Prime and colleagues (2020), this study tested two moderated mediation models in which parent depression and anxiety indirectly impacted child internalizing (Model 1) and externalizing (Model 2) problems during the pandemic through negative effects on multiple aspects of parenting, with these relationships moderated by families’ exposure to COVID-19-stressors. These relationships were tested in a large national U.S. sample of parents of 5–16-year-old children with income, race/ethnicity, and regional distribution matching the U.S. population.
Parent Mental Health Effects on Parenting During the COVID-19 Pandemic
In moderated mediation models, controlling for parent and child gender and age and household income, greater levels of parental depression/anxiety were directly associated with greater inconsistent discipline and less parent supportiveness. Parent depression/anxiety also interacted with COVID-19 stressors; this interaction was characterized by stronger, positive associations between parents’ depression/anxiety symptoms and both parent hostility and poor supervision when greater levels of COVID-19 stressors were reported. On the other hand, depression/anxiety symptoms were more strongly, negatively associated with positive reinforcement, daily routines, and parenting self-efficacy when fewer COVID-19 stressors were reported. This pattern of findings seems to suggest that the combination of high COVID-19 stressors and greater depression/anxiety for parents may have influenced greater frequency of negative parenting (poor supervision, hostility), whereas for parents who reported high COVID-19 stressors, depression/anxiety symptoms had less of an influence on positive aspects of parenting (routines, self-efficacy, positive reinforcement). This may have occurred because greater level of COVID-19 stressors resulted in families spending more time together or needing to provide more support for each other, such that positive parenting behaviors were maintained regardless of parents’ level of depression/anxiety, or that at high levels of COVID-19 stressors, the disruption/stress due to these COVID-19 events was all that influenced these particular parenting variables. Notably, one exception to this pattern was parent supportiveness, which was not associated with COVID-19 stressors but was negatively associated with parent depression/anxiety. This may suggest that parents did not alter their support of children’s thoughts and feelings based on COVID-19 disruptions and stressors. Inconsistent discipline was independently predicted by both parent depression/anxiety and COVID-19 stressors such that higher levels of either were associated with more inconsistent discipline. This makes sense given that disruptions due to pandemic stressors likely impacted parents’ discipline practices and consistency, and concerns over COVID-19 impacts on children may have also made parents feel guilty about following through with punishments. At the same time, parents who were experiencing more depression/anxiety may also have had difficulty following through with discipline or may have threatened punishments that they did not mean.
These findings add further evidence in support of recent research that has indicated negative impacts of parent mental health on parenting during the pandemic (Fosco et al., 2021; Roos et al., 2021; Russell et al., 2020), but adds to this prior research by showing that detrimental effects of parent mental health on parenting can vary as a function of families’ differing exposures to COVID-19 stressors. In particular, these findings confirm that negative features of parenting (e.g., hostility) are particularly sensitive to the combination of COVID-19 stressors and parent depression/anxiety, and should be monitored, especially for families who have experienced more pandemic stressors. Conversely, interventions aimed at improving positive parenting features (e.g., positive reinforcement) should attend to parent depression/anxiety and also consider how COVID-19 stressors may have already brought the family together or increased family support in ways that could be capitalized on. Implications of this bifurcation of COVID-19 stressors’ impact on parenting are important, given evidence for some differential associations between negative and positive parenting and important child outcomes (e.g. Dallaire et al., 2006).
Parenting Effects on Child Mental Health during the COVID-19 Pandemic
Direct links between parenting variables and child mental health were observed in moderated mediation models, regardless of level of COVID-19 stressors, parent mental health symptoms, and covariates. Specifically, more inconsistent discipline, poorer supervision, greater hostility, fewer routines, and less parent supportiveness were associated with worse child internalizing problems. More inconsistent discipline, greater hostility, fewer routines, less parental supportiveness, and lower levels of parenting self-efficacy were associated with worse child externalizing problems. It is interesting to note that with the exception of poor supervision for child internalizing, and parent self-efficacy for child externalizing, there was overlap of the parenting variables predicting internalizing and externalizing (though this was less consistent when models were examined by child age group, see SM). This may have occurred because these factors (consistent discipline, low hostility, routines, supportiveness) generally support well-being for children and may therefore represent transdiagnostic parenting factors. Critically, Prime and colleagues (2020) noted harsh and coercive parenting (i.e., hostility) and routines as especial areas of concern for child mental health during the pandemic, and others have also highlighted the importance of family routines, the regulatory role of parents (i.e., consistent discipline), and emotional support—as a feature of the attachment system—during times of stress (Masten & Cicchetti, 2016). Therefore, the crosscutting importance of these parenting factors are consistent with prior family resilience theories. Shared findings across internalizing and externalizing may also have occurred because of the high correlation between these symptom domains in our sample (r = 0.69), which we discuss further in the limitations section.
In contrast, poor supervision was uniquely associated with child internalizing and parent self-efficacy uniquely with child externalizing. We note that the effects of supervision/monitoring on child internalizing appear to be driven by the youngest (ages 5–8) and oldest (ages 13–16) age groups (see SM). For younger children, low supervision may represent less time spent together with their parent, perhaps leading to increased withdrawn behavior or feelings of sadness or anxiety for the child. For adolescents, low supervision may result in adolescents spending more time alone at home, withdrawing from others, or increasing phone or social media use, which may result in greater internalizing symptoms. This also may be in part bidirectional, wherein adolescents’ withdrawal/internalizing symptoms occurred first and led to lower supervision/monitoring. The unique effects of parent self-efficacy on child externalizing appear to be driven by the 5–8 year old and 9–12 year old groups (see SM). This is consistent with previous parenting literature: low parenting self-efficacy, as a feature of parents’ sense of competence, has often been associated with more behavior problems in children, although it is thought that this operates in a bidirectional fashion (Colalillo & Johnston, 2016). This may have been driven by the younger age groups because parents of older children have more experience and therefore experience greater self-efficacy in the parenting role.
It is also notable that, in preliminary analyses testing moderation of the M1-7 to Y paths, there was no significant moderation of any path by number of COVID-19 stressors experienced, suggesting that the well-documented pathways between parenting and child internalizing and externalizing behaviors were maintained regardless of level of exposures to COVID-19 stressors. This information is critical in terms of underlining the utility of existing evidence-based parenting interventions during the pandemic that emphasize changes in these parenting factors. This provides reassuring evidence insofar as clinical practice may lean on the existing evidence-based intervention armamentarium rather than redirect crucial resources to the development and implementation of new COVID-19 informed interventions and protocols.
Parent Mental Health Effects on Child Mental Health during the COVID-19 Pandemic
That parenting mediators did not fully explain the association between parent depression/anxiety and internalizing and externalizing problems in children (i.e., direct links remained) is an important finding as it suggests that parent and child mental health are linked through additional mechanisms other than parenting behavior. Considering this finding in the context of Prime and colleagues’ (2020) model, the present study considered only the parent–child “subsystem.” It could be that additional consideration of indirect paths through other family subsystems (marital, sibling, whole family) would more fully explain the link between parent and child mental health. Nevertheless, this finding also strongly suggests that interventions would benefit from addressing the individual parent’s mental health alongside their parenting.
It is also important to note the particular pathways through which parent depression/anxiety exerted influence on child internalizing and externalizing problems. Parent depression/anxiety was associated with worse child internalizing problems by means of lower parent supportiveness, poorer supervision, fewer daily routines, and increased parent hostility. However, indirect paths through supervision, daily routines, and hostility were moderated such that for families with higher exposure to COVID-19 stressors, there were stronger effects through hostility and poor supervision and weaker effects through daily routines. Lower parent supportiveness, more inconsistent discipline, fewer routines, increased parent hostility, and lower parenting self-efficacy mediated the association between parent depression/anxiety and child externalizing; however, again families with greater exposure to COVID-19 stressors had weaker indirect effects through routines and self-efficacy and stronger indirect effects through hostility. These models identify parent hostility, parent supportiveness, and daily routines as important avenues for intervention across child mental health domains. Additionally, they pinpoint supervision/monitoring as a specific target for child internalizing problems, and inconsistent discipline and parenting self-efficacy as specific targets for child externalizing problems.
Critically, reducing parent hostility appears particularly important for child internalizing and externalizing difficulties in families with greater exposure to COVID-19 stressors. This finding aligns with Prime and colleagues’ (2020) hypothesis that increases in coercive and harsh parenting were likely to be observed when parents’ well-being is detrimentally impacted by the pandemic. In addition, for families with greater exposure to COVID-19 stressors, supervision/monitoring appears to be important to understand and address for child internalizing difficulties. This observation may reflect families who had to continue working in person while children were at home, or families in which children were withdrawing from parental monitoring more often due to increased time at home, contributing to depression symptoms.
Effects of COVID-19 Stressors
The level of COVID-19 stressors that families experienced did not directly contribute to child mental health problems; rather, parent depression/anxiety showed significant effects. This emphasizes the importance of parents as a buffer for child mental health in times of adversity, consistent with decades of research (Masten & Narayan, 2012). However, it also highlights that COVID-19 stress is likely impacting children indirectly through parents, and therefore that parents’ stress and mental health is important to continue to investigate and address (Peris & Ehrenreich-May, 2021). It is notable that COVID-19 stressors also interacted with parent mental health in predicting a variety of parenting variables, or were directly associated with parenting behaviors, indicating that family risk and resilience pathways do differ based on COVID-19 events experienced. We observed either direct or interactive effects of COVID-19 stressors on every parenting variable except parent supportiveness, revealing pandemic-related effects across behavioral, affective, and cognitive features of parenting. Future research should evaluate to what extent this is due to the effects of specific stressors (e.g., death or severe illness in the family, reduction in income, parent essential worker) or cumulative stressors as measured here, particularly given that certain events—such as a death in the family—are likely to affect families more heavily or in different ways than other pandemic events, such as a stay-at-home order.
Findings help speak to the “how” of child mental health during the pandemic by suggesting particular cognitive (self-efficacy), affective (hostility, parent supportiveness), and behavioral (routines, consistent discipline, monitoring) aspects of parenting, along with parent depression and anxiety, as important areas for intervention to support children’s emotional and behavioral health. Findings also speak to “for whom” by suggesting that, for families who have experienced more COVID-19-related stressors, there may be more particular need to address parent hostility and supervision/monitoring, and by demonstrating specificity in terms of particular parenting mediators of child internalizing (supervision/monitoring) vs. child externalizing (parenting self-efficacy, inconsistent discipline) problems. Behavioral parenting interventions such as Parent Management Training or PMT (e.g. Barkley, 2013; McMahon & Forehand, 2003) and Parent–Child Interaction Therapy (Eyberg, 1988) that are widely disseminated have documented evidence for addressing parenting behaviors such as consistent discipline, routines, supervision, and supportiveness; notably, PMTs have also demonstrated positive effects for parenting cognitions (Colalillo & Johnston, 2016). Parents may also benefit from treatments to address their own symptoms and improve emotion regulation (in order to reduce hostility), such as Dialectical Behavioral Therapy (Linehan, 1993). Mentalization-based parent interventions may also be important for increasing parents’ supportiveness, identification and management of emotions, and understanding of children’s emotions and behavior (Arnsten et al., 2021). The findings reported here support a recent comprehensive review of evidence-based interventions, including family- and parent-based interventions, recommended for addressing child mental health problems from the pandemic (Arnsten et al., 2021). Yet, in leveraging or adapting existing evidence-based interventions to address pandemic mental health, the current findings clearly demonstrate it is important for clinicians to assess families’ level of exposure to COVID-19 stressors given its impact on family pathways to child mental health.
Limitations, Alternative Models, and Directions for Future Research
Data were collected at one time point and therefore mediational findings cannot speak to effects over time, only to explanation of variance at the time when data was collected. In addition, though our study focuses only on the paths from parent to child mental health in keeping with Prime and colleagues’ model, there are likely bidirectional effects between parent and child mental health (and between child mental health and parenting variables). It is possible that current findings were partially a result of child to parent effects, and future research on this topic would benefit from using longitudinal data to evaluate competing models and reciprocal parent–child influences. Another alternative is that effects could be driven by shared biological factors between parents and children (genes, temperament, etc.): for example, shared genetic risks could increase vulnerability to psychopathology for both parent and child during COVID-19, representing possible gene by environment interactions. In addition, other shared traits such as emotion regulation behaviors, which children often learn through parents’ modeling (Hajal & Paley, 2020), may result in increased parent–child psychopathology during times of stress. Collectively, it is important to consider both alternative causal directions of influence and potential shared underlying mechanisms that may influence parent and child mental health.
Another limitation is that data presented here are only a snapshot of one period during the pandemic (February-April 2021). One benefit of the study’s time period, however, is that the COVID-19 stressors measure included a year of pandemic-related events, increasing variability and supporting the variable as a predictor. Data were collected only by parent report, meaning that shared method variance likely effects the magnitude of associations. A related limitation is that the child internalizing and externalizing variables were strongly associated in our sample (r = 0.69), which may have led to a similar pattern of findings across the two models. Correlations of r = 0.45-0.54 have been reported between internalizing and externalizing problems in a large review of studies (Achenbach et al., 2016). It is possible that parents were reporting problems more consistently across domains due to observing children’s behavior more during increased time at home, or perceiving that in general their child’s behavior was more problematic during the pandemic, leading to a large correlation in our sample. Future studies would benefit from multi-method and multi-rater assessment to reduce shared method variance.
A crucial need for future research is also to understand social and cultural influences within Prime’s model by testing differences based on family racial/ethnic background. This is particularly important given past research showing racial/ethnic differences in parenting styles (Smetana, 2017) and their association with child mental health (e.g. Greening et al., 2010). Such research could help advise culturally informed adaptations of evidence-based mental health interventions.