Parental Problem Drinking, Parent Emotion Socialization and Child Emotion Regulation

Abstract

The current study examines mother and father PPD as a risk factor for child emotion regulation difficulties. A model was tested in which parent depression and parent emotion socialization of children were serial intervening variables. Parent emotion socialization was assessed via parent self-reported supportive and nonsupportive reactions to child negative emotions. Differences in associations based on parent and child sex were also tested. Data were drawn from an online survey of 337 parents (58.2% mothers) reporting on a single child aged 5 to 12 years (M = 7.10 years of age; 45.1% female). The majority of parents identified as White (72.7%); whereas 10.1% identified as Black, 6.4% identified as Hispanic, 8.4% identified as Asian or Pacific Islander, and 2.4% identified as multiracial/biracial or another racial/ethnic identity not listed. PPD was associated with higher parental depression, which in turn was associated with greater nonsupportive and lower supportive parent reactions to child negative emotions; PPD was indirectly associated with poorer child emotion regulation through lower supportive parent reactions. Associations were not moderated by parent or child sex. Findings support parent emotion socialization practices as an important mechanism through which PPD conveys risk for child psychological problems and is a possible target for intervention.

Parental problem drinking (PPD) includes parent heavy drinking, symptoms of alcohol use disorder (e.g., lack of control, experiences of withdrawal), and drinking in hazardous ways (e.g., driving intoxicated, taking medications with alcohol contraindications; Keller et al., 2005). PPD is a major social concern due to the large number of children exposed (McKetta & Keyes, 2019) and the numerous adverse consequences for child development (Rossow et al., 2016a). According to the most recent National Survey on Drug Use and Health, 84.2% of US adults have a positive lifetime history for alcohol use, more than half consume alcohol on a regular basis and 23.5% engage in binge drinking; nearly 30 million Americans meet criteria for alcohol use disorder (https://www.samhsa.gov/data).

Adverse consequences of PPD for children include anxiety, depression, behavioral issues, lower academic achievement, difficulties in peer relationships, and substance use problems (Lund et al., 2023). What many of these negative consequences have in common is that they involve difficulties with child emotion regulation (Cludius et al., 2020). Emotion regulation is increasingly understood to be important for mental health, as it overlaps with three of the research domain criteria proposed by the National Institute of Mental Health (regulation, positive valence and negative valence) to improve diagnosis and treatment of mental illness. However, there is little research on associations between PPD and child emotion regulation. This has left an important gap in understanding the mechanisms through which PPD conveys risk for maladaptive developmental outcomes; effective prevention and treatment in this population requires a thorough understanding of these mechanisms. The current study helps to address this need. Using a non-clinical sample, we test a model of associations between PPD and child emotion regulation and lability. In this model we propose that PPD is related to higher parental depression, that parental depression will be associated with poor parent socialization of emotion, and that poor parent socialization of emotion will be related to child emotion regulation difficulties.

Significance of Parental Problem Drinking for Child Development

The gap in research on PPD and parent emotion socialization is important to address because PPD may affect a large number of children. Alcohol use disorder (AUD) is one of the most prevalent mental health problems in the US: In 2022, 29.5 million Americans were estimated to meet criteria for AUD (https://www.samhsa.gov/data). As noted above, alcohol use and binge drinking is even more common than AUD in the United States. Parental problem drinking falls along a spectrum, and there is important variability within the diagnostic category of AUD and across subclinical levels of problem drinking (Merrill & Aston, 2020). An important implication is that comparisons of families with and without parental AUD may not fully capture the full consequences of PPD for child development. Rather, the use of non-clinical samples may more easily permit estimation of the magnitude of effects within a population than do clinical samples, and such studies have provided valuable insights.

These valuable insights include multiple ways in which non-clinical parental problem drinking may affect parenting. For example, van der Zwaluw et al. (2008) found that PPD was associated with lower parental support, more permissive rules for adolescent drinking, and greater overall permissiveness with adolescents over one year. Ohannessian (2012) found that poor parent-child communication mediated the longitudinal association between PPD and adolescent mental health for girls but not for boys. Finan et al. (2018) found that mother problem drinking was associated with lower mother support of adolescents, but this was not the case for father problem drinking and father support. In all of these studies, the percentage of the sample suffering from AUD was low, yet significant associations between PPD, family dysfunction, and adverse child outcomes were observed.

Theoretical Models

Theoretical models of child emotion regulation provide guidance for understanding these associations. Morris et al (2007) proposed the Tripartite Model of Family Impact on Children’s Emotion Regulation and Adjustment (henceforth referred to as the Tripartite Model), a model in which parent mental health plays an important role. According to the Tripartite Model, there are three primary mechanisms through which the family contributes to the development of child emotion regulation: family emotional climate, parent modeling of emotion regulation, and parenting practices that socialize emotion. An example of parent emotion socialization practices is how parents react to children’s negative emotions. Parent characteristics such as mental health are proposed to influence each of these mechanisms. Morris et al. (2007) provide evidence that parental depression models poor emotion regulation and undermines parents’ abilities to adaptively respond to children’s negative emotions. However, they note that there is very little research on dimensions of parent mental health beyond parent depression and call for additional research to better understand how parent mental health influences the key mechanisms identified by the Tripartite Model. The current study offers a test of PPD as an important parent mental health variable that may undermne parent emotion socialization, focusing on parent reactions to child negative emotions as the parental emotion socialization mechanism, which in turn may contribute to child emotion regulation. Parental depression is also included in the tested model as a mechanism through which PPD may influence parent emotion socialization.

The current study also draws on the Integrated Model of the Intergenerational Transmission of Self-Regulation (henceforth referred to as the Integrated Model; Bridgett et al., 2015). Parent mental health is an important contributor to child emotion regulation in the Integrated Model, but only parent depression and ADHD are reviewed by Bridgett et al. (2015). The Integrated Model emphasizes a distinction between top-down and bottom-up dimensions of self-regulation. The current study employs this distinction by including both the top-down and bottom-up dimensions as separate variables. Top-down emotion regulation originates in cortical structures, is effortful, and encompasses cognitive strategies such as reappraisal, rumination, and suppression of emotions. The child emotion regulation subscale of the Emotion Regulation Checklist (Shields & Cicchetti, 1997) is a good assessment of top-down emotion regulation, as it includes items reflecting effective modulation of emotion in service of goals (ability to be positive in social situations, empathic towards others, can say when s/he is feeling sad, angry or mad, fearful or afraid, displays negative affect when appropriate) that leads to perceptions of the child as emotionally healthy (child is cheerful, not sullen or sad, does not have flat affect). Bottom-up emotion regulation originates in subcortical structures such as the amygdala, hippocampus, and stria terminalis, is automatic and reactive to the environment. The child emotion lability subscale of the Emotion Regulation Checklist appears to assess the bottom-up form of child emotion regulation, as it includes items that reflect experiences of intense and unpredictable affect (exhibits wide mood swings, overly excited, disruptive exuberance), impulsivity (cannot delay gratification, impulsive), and high levels of negative reactivity (prone to angry outbursts, reacts negatively to peers, temper tantrums, easily frustrated, intense reactions to changing activities).

Parental Problem Drinking and Parent Depressive Symptoms

As noted above, and consistent with Tripartite Model and Integrated Model proposals that parent depression can effect parent emotion socialization, the current study includes parent depression as an intervening variable in the association between PPD and parent emotion socialization. There is a strong rationale for PPD being associated with parent depression. Acute effects of alcohol and chronic heavy drinking can diminish parent emotion regulation (Becker, 2012; Nutt et al., 2021). This is readily observed in the associations between problem drinking and depression (e.g., Nunes, 2023). Longitudinal studies of associations between problem drinking and depression support problem drinking as the predictor of depression rather than depression as the predictor of problem drinking (Fergusson et al., 2009). Treatment studies also show reductions in depression following alcohol detoxification (Davidson, 1995). In addition, parental depression mediates associations between PPD and child internalizing symptoms (El-Sheikh & Flanagan, 2001). These findings also lend support to the role of parent depression in the proposed model.

Parent Reactions to Child Negative Emotions

Theory and empirical research also support the proposed role of parent reactions to child negative emotions. Morris and colleagues’ Tripartite Model (2007) includes these reactions as an important form of parent emotion socialization, one of the three processes through which the family influences the development of child emotion regulation. Parents can react in supportive ways, such as by helping children solve the problem causing negative emotion (problem-focused reactions), helping children to feel better (emotion-focused reactions), and teaching children that negative emotions are a normal part of life (encouragement of expression; Fabes et al., 2002). Nonsupportive reactions include punishing negative emotions, minimizing or invalidating children’s feelings, or becoming distressed (Fabes et al., 2002). Supportive reactions teach children about the acceptability of negative emotions and how to regulate them, but nonsupportive reactions teach children to avoid and suppress negative emotions (Eisenberg et al., 1999).

Several research studies demonstrate the importance of parent reactions to child negative emotions. For example, mother distress reactions are associated with lower emotion control among typically-developing preschoolers (Woods et al., 2017). Parents’ reactions to kindergarten children’s negative emotions predicted child effortful control in the first grade (Swanson et al., 204). Differential associations with child emotion regulation versus child emotion lability are found in non-clinical samples: mother supportive reactions are associated with higher child emotion regulation and mother nonsupportive reactions are associated with higher child emotion lability (Morelen et al., 2016). Mother supportive reactions at age 5 years predict increases in child emotion regulation at age 7 years whereas mother nonsupportive reactions predict decreases in child emotion regulation in typically developing children (Blair et al., 2014). Despite a frequent focus on mother reactions, there is evidence that father reactions to negative emotions are also important (Baker et al., 2011). Based on these findings, the current study examines both mother and father reactions to child negative emotions.

Parental Problem Drinking and Parent Reactions to Child Negative Emotions

There is a strong rationale for expecting higher PPD to be associated with less supportive and more nonsupportive reactions to child negative emotions both directly and indirectly through parent depression. Adaptive emotion socialization requires parents to have healthy emotion regulation abilities themselves (Hajal & Paley, 2020) and Morris and colleagues’ (2007) Tripartite Model proposes that parent mental health problems can adversely affect parent emotion regulation. Consistent with this proposal, adult emotion regulation difficulties such as nonacceptance of emotions, struggles engaging in goal directed behavior, and limited repertoire of adult emotion regulation strategies are associated with problem drinking across diverse populations (Paulus et al., 2016; 2017). The implication is that problem drinking will make it difficult for parents to accept the emotions of their children, cope with child negative emotions, and instruct children in healthy emotion regulation strategies.

Despite the strong rationale for PPD as a risk factor for poor child emotion regulation and for parent emotion socialization as a mechanism of this risk, there are few studies of child emotion regulation in the context of PPD and even fewer studies of PPD and parents’ emotion socialization practices. One study found that adults exposed to PPD report greater suppression of their emotions in the presence of their parents (Jones & Houts, 1992). Keller et al. (2022) have shown that adult offspring’s retrospective reports of parent reactions to child negative emotions are intervening variables in the association between PPD exposure and offspring current emotion regulation; PPD is related to greater nonsupportive and lower supportive reactions to child negative emotions, which in turn are associated with greater anger and depression rumination. Another study found that severity of mother substance use disorder (primarily opioid use disorder) is associated with poorer emotion regulation in children aged three to eight years; nonsupportive reactions to child negative emotions are a more important intervening variable than general negative parenting (Shadur & Hussong, 2020). These studies lend initial support to the potential role of parent reactions to child negative emotions in associations between PPD and child emotion dysregulation.

The Current Study

The current study builds on this initial support to conduct the first study of associations between PPD and current (rather than retrospective) reports of parent emotion socialization, addressing a gap in understanding the psychological processes that explain why PPD confers risk for mental health problems in children. A thorough understanding of these processes is needed in order to ensure the best possible care is available to exposed children. The purpose of the current study is to address the need for basic research that can support the development of interventions for families characterized by PPD. We propose a model that integrates state-of-the-art empirical and theoretical work in the substance use and addiction fields with family and developmental psychology. Substance use research has established that alcohol consumption compromises a person’s emotion regulation and increases risk for depression; family and developmental psychology has developed models emphasizing the role of parent mental health and emotion socialization for child emotion regulation. In the proposed model, PPD is indirectly associated with child emotion regulation and lability. Specifically, we hypothesize that PPD will be related to greater parent depressive symptoms, which in turn will be associated with less supportive and more nonsupportive reactions to child negative emotions; these reactions will be related to poorer child emotion regulation and greater child emotion lability.

Because the current study includes a diverse sample with respect to parent and child sex, we are able to investigate both parent and child sex as moderators of associations. Alcohol consumption is overall higher among men than women, but women are more likely to use alcohol to cope with stress than men (Peltier et al., 2019). Women and girls are also higher in negative reactivity (Gardener et al., 2013) and report more often engaging in most emotion regulation strategies than men and boys (Nolen-Hoeksema, 2012). These findings suggest that associations between PPD, parent depression, and parent reactions to child negative emotions will be stronger for mothers than fathers, and that PPD and parenting will be more strongly associated with child emotion regulation and lability for girls than for boys.

Possible Confounding Variables

These associations are expected even after controlling for multiple possible confounding variables, including parent race, family socioeconomic status, marital conflict, and whether the parent is biologically related to the child. Including these variables as covariates in analyses helps ensure scientific rigor by ruling out some of the alternative explanations that may exist for the hypothesized associations. The selected variables are widely known to be associated with family and child variables (e.g., race, age, SES), making them important for inclusion. For example, developmental changes in emotion regulation occur across the period of 5 to 12 years (Carr, 2017; Woltering & Lewis, 2009). There is also evidence that White persons engage in more problem drinking than people with other racial identities, although people from racial minority groups may experience more negative consequences of their problem drinking than White people (Zapolski et al., 2014). Reactions to child negative emotions may differ between biological parents and step-parents because stepmothers feel higher parenting stress, experience greater depressive symptoms, and feel less positively toward children than biological mothers (Shapiro & Stewart, 2011). Higher SES is related to greater alcohol consumption but lower negative health consequences of drinking (Jones et al., 2015) and better emotion regulation (Côté et al., 2010). Finally, marital conflict is included as a covariate because the quality of the marital relationship is considered to be an important component of the emotional climate of the family, which is one of the three processes in the Tripartite Model through which the family influences the development of emotion regulation in children (Morris et al., 2007).

Method

Participants and Procedures

Participants were drawn from an online survey of 457 parents conducted in the summer of 2018. Only one of the parents participated from each family, with 173 (58.20%) being mothers. If participating parents had more than one child, they were asked to report on the child with whom they spent the most time or to report on the youngest child should they spend equal time with all their children. Most participants were biologically related to the child on whom they were reporting (91.6%) and had been living with the child for at least three years (96.6%). Among parents, the majority identified as White (72.7%); whereas 10.1% identified as Black, 6.4% identified as Hispanic, 8.4% identified as Asian or Pacific Islander, and 2.4% identified as multiracial/biracial or another racial/ethnic identity not listed. Parents also indicated the financial status of their family: 7.7% indicated that they were poor or very poor, 33.3% indicated that they were making just enough money to get by, 46.5% indicated that they were middle class and making more than enough to get by, and 12.4% indicated that they were upper class and able to afford more things than most people.

This study and its procedures were conducted with approval from the appropriate university Institutional Review Board. Participants were required to be currently rearing a child between five and 18 years old, reside in the US, and correctly answer all three attention-check questions in the survey. The three attention and validity check questions were included at random locations in the survey and asked participants to select a particular response. Persons who failed any of the attention check questions were removed from the data. Participants were recruited via Amazon Mechanical Turk. All participating parents gave informed consent prior to completing any part of the study and all ethical principles regarding research with human participants were followed. The self-paced online survey took approximately 40 minutes to complete, and participants were paid $1.

Because the measure of parent reactions to child negative emotions is only valid through 12 years of age, parents of children older than 12 years were excluded (n = 49). In addition, parents who had been living with the child for less than one year were excluded (n = 7) to help ensure sufficient exposure to the parent for associations to be observed. Due to the teratogenic effects of alcohol, parents of children for whom prenatal alcohol exposure was possible were excluded (n = 70). After excluding participants who failed one or more of the inclusion criteria, the resulting sample submitted to preliminary analyses was 337.

Measures

Parental Problem Drinking

PPD was measured with the Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993), a widely used and well-established measure (auditscreen.org). This measure contains 10 items assessing certain aspects of problem drinking behaviors, such as frequency of binge drinking, frequency of any consumption, inability to quit drinking or lower drinking habits, etc. Example items include “How often do you have six or more drinks on one occasion” and “How often during the last year have you had a feeling of guilt or remorse after drinking.” Following the standard scoring approach, eight of the items are scored on a five-point Likert-style scale where responses range from never (0) to daily or almost daily (4). Two of the items are scored on a three-point Likert-style scale where responses are no (0), yes, but not in the past year (2), or yes, during the past year (4). These two items use an alternative response scale because they assess symptoms that are likely to be rarer (injury, others noticing a drinking problem). Scores were computed by summing across all completed items and higher scores indicate higher problem drinking levels. Although the variable was treated as continuous, scores from one to seven are considered to be low-risk consumption, scores from eight to 14 may suggest hazardous or harmful consumption, and scores above 15 may suggest alcohol dependence (Saunders et al., 1993). Cronbach’s α was acceptable, α = .94.

Parent Depression

Parent depression was measured with the Center for Epidemiologic Studies Depression Scale (CESD; Radloff, 1977). The CESD is a 20-item measure that assess how frequently during the last week a person has experienced signs or symptoms of depression (e.g., “I thought my life had been a failure” or “I felt depressed”). Answers range from rarely (1) to most of the time (4). A total score was computed by summing across all of the completed items, including the four reverse-coded items (Radloff, 1977). Internal reliability for this measure was acceptable, α = .94.

Parent Reactions to Child Negative Emotions

The Coping with Children’s Negative Emotions Scale (CCNES; Fabes et al., 2002) measures how parents may tend to react to their children’s experience of negative emotions in 12 different hypothetical scenarios. An example scenario is that the child loses a prized possession and starts crying. For each scenario, parents are given six different ways in which they may respond and are asked to rate how likely they are to respond to the scenario in each specific manner on a 7-item Likert-style scale with responses ranging from very unlikely (1) to very likely (7). The six responses correspond to six subscales. The three supportive responses are problem-focused responses (e.g., “help my child think of places he/she hasn’t looked yet”), emotion-focused responses (e.g., “distract my child by talking about happy things”), and expressive encouragement (e.g., “tell him/her it’s OK to cry when you feel unhappy”). The three nonsupportive responses are distress (e.g., “get upset with him/her for being so careless and then crying about it”), minimizing child feelings (e.g., “tell my child that he/she is over-reacting”), and punitive responses (e.g., “tell him/her that’s what happens when you’re not careful”). Scores are computed by averaging responses across all available items, including reverse-coded questions, and higher scores indicate higher usage of that parent reaction. Cronbach’s α for the six subscales ranged from .78 (distress) to .91 (punitive), with an average across the six subscales of .89.

Child Emotion Regulation/Lability

Parents completed the Emotion Regulation Checklist (ERC; Shields & Cicchetti, 1997) for the child on whom they were reporting. The ERC contains two subscales with subscale scores computed by summing across all completed items. Higher scores on the emotion regulation subscale indicate greater child ability to regulate expression of emotions, capacity for empathy, and emotional self-awareness. An example item is “Is able to delay gratification.” Higher scores on the emotion lability subscale indicate emotional inflexibility, anger dysregulation, and mood lability (Shields & Cicchetti, 1997). An example item is “Is easily frustrated.” The 24 items are rated on a four-point Likert style scale with responses ranging from rarely or never (1) to almost always (4). Internal reliabilities were α = .71 for emotion regulation and α = .80 for emotion lability.

Covariates

All covariates were assessed via parent-report. Parent and child sex were dichotomous variables in which a value of one was assigned to females and zero was assigned to males. One-way ANOVAs did not indicate statistically significant differences among Black, Hispanic, Asian, or other minority races on study variables. Therefore, race was included as a dichotomous variable in which White participants were assigned a value of one and non-White participants were assigned a value of zero. Type of parent relationship was a dichotomous variable scored as biologically related to the child (1) or not biologically related to the child (0). Child age was measured in years. Subjective ratings of family socioeconomic status were assessed by a single item in which participants selected the response option that best describes their status on a scale ranging from very poor (0) to very rich or upper middle class (4); subjective ratings are more strongly associated with functioning than objective measures (Tan et al., 2020).

To assess marital conflict, the frequency and severity subscale of the Conflicts and Problem-Solving Scales (Kerig, 1996) was used. This two-item subscale is scored by adding the response to the item assessing frequency of minor conflicts to twice the rating of the frequency of major conflicts (to give greater weight to major conflict). Thus, higher scores indicate greater and more severe marital conflict.

Data Analysis Plan

Analyses were conducted in R version 4.0.4 (R Core Team, 2021) and proceeded in three steps: preliminary analyses, primary analyses, and exploratory analyses. Preliminary analyses focused on outliers, missing data, skewness in variables, and description of levels of PPD in the sample. To evaluate the potential for outliers, the Mahalanobis distance for each case was assessed and values meeting the criterion of p < .001 were considered outliers and removed (De Maesschalck et al., 2000). This conservative criterion was used to due to the large number of tests (a test was conducted for each participant, for a total of 337 tests). All other analyses used a p value of <.05 to determine statistical significance. Next, missing data were assessed. An a priori decision was made to employ casewise deletion if missing data were less than 5% and to use imputation if missing data were greater 5%. Skew was evaluated by dividing the skew statistic by its standard error for each variable, with values exceeding |2| considered indicative of significant skew. An a priori decision was made to use robust standard errors (Huber, 1967; Savalei & Rosseel, 2022) if significant skew was observed in the variables. Mean levels of PPD were examined, in addition to the percentages of parents in different categories (hazardous drinking, moderate to severe AUD), and rates of different symptoms of problem drinking.

Primary analyses utilized structural equation modeling (SEM) with the lavaan package (Rosseel, 2012). Two models were fit to the data. Models would be considered an acceptable fit to the data if: (1) the model had a nonsignificant chi-square test (Barrett, 2007), or (2) the RMSEA and SRMR values were less than or equal to .08 and the CFI value was greater than .95 (Hu & Bentler, 1999; MacCallum et al., 1996). The first model was the measurement model and assessed the factor loadings for two latent variables: parent supportive and nonsupportive reactions. Parent use of supportive reactions was modeled as a latent variable indicated by problem-focused, emotion-focused, and expressive encouragement reactions. Parent use of nonsupportive reactions was modeled as a latent variable indicated by minimization, distress, and punitive reactions. For each latent variable, one of the unstandardized factor loadings was fixed to 1. Appropriateness of the measurement model was determined by acceptable model fit.

The next model included all study variables and was the full structural model. PPD was an observed variable predicting of parent depression (observed variable), parent use of supportive and nonsupportive reactions (latent variables), and child emotion regulation and emotion lability (separate observed variables). Parent depression and parent use of supportive and nonsupportive reactions were included in the model as intervening variables in relations between PPD and child outcomes. Specifically, parent depression was modeled as a predictor of parent supportive and nonsupportive reactions, as well as child emotion regulation and lability. In addition, parent supportive and nonsupportive reactions were modeled as predictors of child emotion regulation and lability. Direct associations of PPD with parent reactions and child variables were also estimated. The residual variances for child emotion regulation and child lability were correlated. The residual variances for supportive and nonsupportive reactions were also correlated. The model controlled for parent sex, parent racial/ethnic identity, biological relation to child, family SES, child age, child sex, and marital aggression. Thus, the structured model was saturated. Testing of indirect associations was done with Sobel (1982) tests because bootstrapping is not possible when using robust standard errors (Rosseel, 2012).

Lastly, exploratory analyses focused on assessing parent and child sex as moderators of associations. A multigroup analysis was conducted with groups based on parent sex and a second one with groups based on child sex. For each multigroup analysis, measurement invariance across the two groups was first tested following established guidelines (Putnick & Bornstein, 2016). After model measurement invariance was established, equality constraints were added to all paths of interest across the two groups in order to test for an overall interaction (i.e., whether there are any significant differences in path coefficients between groups). The fit of the constrained model would be compared to the model without equality constraints utilizing a Δχ² test where a significant test would demonstrate differences in path coefficients based on parent or child sex.

Results

Preliminary Analyses

Two possible outliers were identified and removed from the data analysis sample. Only six cases out of the remaining 335 had missing data (1.8%), and therefore casewise deletion was used to handle missing data. Child emotion lability and several of the subscales for parent reactions to child negative emotions were highly skewed. This was addressed in two ways. For concerns about skew in bivariate relations, only the non-parametric Spearman’s rho correlation was used. For concerns about skew in multivariate relations (e.g., structural equation analyses), maximum likelihood estimation with robust standard errors was used.

Spearman’s rho (ρ) correlations among study variables are included in Table 1 and descriptives of study variables are included in Table 2. All potential covariates were associated with either parenting variables (e.g., different supportive and nonsupportive reactions to child negative emotions) or child variables. Data are available from the corresponding author upon request.

Table 1.

Correlations Among Study Variables

Variables	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.	14.	15.	16.
1. Child Age	-
2. Child Female	−.01	-
3. Biological Relation	−.03	.08	-
4. Family SES	−.05	.04	−.01	-
5. Parent Female	.09	.21**	.06	−.14*	-
6. Parent White	.01	.02	−.07	−.12*	.08	-
7. Marital Aggression	−.06	.05	−.04	−.01	−.08	−.12*	-
8. PPD	−.01	−.11*	−.06	.12*	−.25**	−.09	.34**	-
9. Parent Depression	−.04	−.08	−.07	−.12*	−.10	−.09	.52**	.42**	-
10. Problem-Focused	−.01	.13*	−.10*	−.20**	.29**	.03	−.14*	−19**	−.30**	-
11. Emotion-Focused	−.12*	.14*	−.07	−.19**	.26**	−.04	−.09	−.16*	−.24**	.88**	-
12. Ex. Encouragement	−.04	.13*	−.06	−.08	.28**	−.01	−.09	−.16*	−.18*	72**	.73**	-
13. Minimization	−.05	−.07	.01	.15*	−.24**	−.20**	.33**	.36**	.41**	−.34**	−.30**	−.31**	-
14. Punitive	−.08	−.11*	−.01	.17*	−.20**	−.15*	.33**	.37**	.45**	−.41**	−.36**	−.32**	.87**	-
15. Distress	−.05	−.08	−.06	.07	−.13*	−.07	.39**	.35**	.56**	−.47**	−.42**	−.35**	.67**	.77**	-
16. Child Emo. Reg.	−.05	.09	.05	−.05	.11*	−.05	−.08	−.07	−.20**	.42**	.41**	.33*	−.18**	−.19**	−.22**	-
17. Child Lability	.01	.03	.08	−.12*	.13*	.08	−.33**	−.34**	−.46**	.22**	.20**	.08	−.39**	−.44**	−.49**	.02

Note. SES = Socioeconomic Status, PPD = Parent Problem Drinking, Ex. = Expressive, Emo. Reg. = Emotion Regulation.

^*p < .05

^**p < .001.

Table 2.

Study Variable Descriptives

Dichotomous Variables	Response Categories	n	Percent	% Missing
Child Sex
	Female	152	45.4	-
	Male	183	54.6	-
Biological Relation to Parents
	Biologically Related	307	91.6	-
	Not Biologically Related	28	8.4	-
Parent Sex
	Female	202	60.3	-
	Male	133	39.7	-
Parent Ethic/Racial Identity
	Identified as ‘White’	242	72.2	-
	Identified as ‘Other than White’	93	27.8	-
Continuous Variables	M	SD	Range	% Missing
Child Age	7.63	2.30	5–12	0.9
Parent Age	34.75	6.55	20–60	0.3
Family SES	3.63	0.82	1–6	-
Marital Aggression	7.74	3.66	3–17	0.9
Problem Drinking	5.73	7.62	0–34	-
Parent Depression	35.66	12.83	20–68	-
Problem-Focused	5.37	1.12	1.67–7	-
Emotion-Focused	5.31	1.12	2–7	-
Expressive Encouragement	4.97	1.17	1.75–7	-
Minimization	3.06	1.25	1–6.42	-
Punitive	2.92	1.28	1–6.42	-
Distress	3.02	0.97	1–5.42	-
Emotion Regulation	22.32	2.74	14–29	-
Lability	43.00	5.85	21–55	-

Levels of Parental Problem Drinking in Sample

To aid in the interpretation of primary results, the extent of problem drinking by parents in the sample was evaluated. The mean score was 5.73 (SD = 7.62), which is in the range of low-risk alcohol consumption for the Alcohol Use Disorders Identification Test (1–7). Only 22.0% of parents indicated that they were nondrinkers. The percent of parents with scores indicating hazardous drinking (8–14) was 7.7% and the percent of parents with scores indicating potential moderate to severe alcohol use disorder was 16.6%. In addition, 37.7% of parents indicated that they typically drank 3 or more drinks per drinking occasion, 38.7% indicated that they had a binge drinking episode at least on rare occasions (less than monthly), and 17.9% indicated that they had a binge drinking episode at least monthly. The frequency of other alcohol use disorder symptoms ranged from 15.2% (memory loss from drinking) to 31.5% (feeling guilt or remorse after drinking) reporting they had experienced the symptom at least once.

Structural Equation Analyses

For the sake of brevity, all results reported are controlling for all other variables in the model.

The initial measurement model was an excellent fit to the data: robust χ² (6) = 5.75, p = .45, robust RMSEA = .00, robust RMSEA 90% CI [.00, .07], robust CFI = 1, robust TLI = 1, and robust SRMR = .02. Robust model fit statistics were used over typical model fit statistics in accordance with the use of robust standard errors (Rosseel, 2012). Within the model, all six subscales for parent reactions to child negative emotions loaded significantly onto their respective latent variable with all standardized loadings greater than .64. The covariance between the two latent variables was also significant, estimate = −.42, p < .001.

The full structural model was an acceptable fit to the data: robust χ² (49) = 135.30, p < .001, robust RMSEA = .07, robust RMSEA 90% CI [.06, .08], robust CFI = .96, robust TLI = .92, and robust SRMR = .03. Figure 1 depicts all significant paths. Table 3 provides all results of the model. Higher PPD was significantly associated with higher parent depression. In turn, higher parent depression was associated with lower use of parent supportive reactions. Lower use of parent supportive reactions was associated with lower child emotion regulation, but was unrelated to child emotion lability. In line with the data analysis plan, two indirect associations were tested. The indirect association between PPD and use of parent supportive reactions through parent depression was significant, ab = −.02, p < .001. The indirect association between parent depression and child emotion regulation through use of supportive reactions was also significant, ab = −.03, p < .001.

Figure 1. Depiction of Results From Structural Equation Model.

Note. Only significant associations are depicted, and coefficients are standardized.

Bio. = Biological, SES = Socioeconomic Status.

* = p <.05, ** = p <.01, *** = p <.001.

Table 3.

Results from the Full Structural Equation Model

Path Results	B	SE	Beta	P	R2
Parent Depression					.43
Child Age	−.10	.24	−.02	.69
Child Sex	−.99	1.14	−.04	.39
Bio. Relation	−2.15	2.25	−.05	.34
Family SES	−2.52	.77	−.16	.001
Parent Sex	.79	1.21	.03	.51
Parent Racial Identity	−.40	1.29	−.01	.76
Marital Aggression	1.41	.18	.40	<.001
Problem Drinking	.62	.10	.37	<.001
Supportive Reactions					.24
Child Age	−.04	.02	−.09	.10
Child Sex	.20	.11	.09	.07
Bio. Relation	−.57	.18	−.15	.002
Family SES	−.28	.07	−.22	<.001
Parent Sex	.49	.12	.23	<.001
Parent Racial Identity	−.20	.13	−.09	.11
Marital Aggression	.02	.02	.08	.19
Problem Drinking	−.01	.01	−.03	.65
Parent Depression	−.03	.01	−.35	<.001
Nonsupportive Reactions					.42
Child Age	−.03	.02	−.06	.21
Child Sex	−.09	.11	−.04	.43
Bio. Relation	−.01	.18	−.01	.94
Family SES	.17	.08	.13	.16
Parent Sex	−.05	.12	−.02	.65
Parent Racial Identity	−.03	.13	−.01	.84
Marital Aggression	.05	.02	.18	.005
Problem Drinking	.02	.01	.16	.01
Parent Depression	.03	.01	.40	<.001
Emotion Regulation					.20
Child Age	−.01	.06	−.01	.83
Child Sex	.07	.28	.01	.82
Bio. Relation	.74	.55	.07	.18
Family SES	.09	.20	.03	.65
Parent Sex	−.16	.31	−.03	.62
Parent Racial Identity	−.21	.33	−.03	.54
Marital Aggression	−.02	.06	−.03	.72
Problem Drinking	.01	.03	.04	.60
Parent Depression	−.01	.02	−.05	.51
Supportive Reactions	1.16	.17	.45	<.001
Nonsupportive Reactions	.04	.24	.02	.86
Emotion Lability					.1 6
Child Age	−.06	.10	−.02	.55
Child Sex	−.33	.56	−.03	.55
Bio. Relation	.52	.95	.02	.59
Family SES	.15	.09	.11	.24
Parent Sex	.78	.61	.07	.20
Parent Racial Identity	−.14	.64	−.01	.82
Marital Aggression	−.10	.10	−.06	.29
Problem Drinking	−.08	.05	−.10	.14
Parent Depression	−.01	.02	−.04	.49
Supportive Reactions	−.28	.42	−.04	.48
Nonsupportive Reactions	.07	.48	.03	.81

Note. Bio. = Biological, SES = Socioeconomic Status.

Beta is the standardized coefficient from the std.all column.

Higher parent depression was also statistically significantly associated with more use of nonsupportive reactions. However, more use of parent nonsupportive reactions was not significantly associated with child emotion regulation nor was it associated with child emotion lability. Although the indirect association between PPD and use of parent nonsupportive reactions through parent depression was significant, ab = −.02, p < .001, the lack of an association between use of parent nonsupportive reactions and child variables means that the second indirect association was not significant. The direct associations of PPD with child emotion regulation and emotion lability were not statistically significant.

Several covariates included in the model were significantly associated with endogenous variables. Higher marital aggression was associated with higher parent depression and more use of nonsupportive reactions. Higher family SES was associated with lower parent depression, lower use of supportive reactions, and higher use of nonsupportive reactions. Fathers used less supportive reactions than did mothers, and parents who were biologically related to the child on whom they were reporting used less supportive reactions. Child sex, child age, and parent racial/ethnic identity were unrelated to parent depression, the use of parent supportive or non-supportive reactions, child emotion regulation, and child emotion lability.

Exploratory Analyses

For child sex, an multi-group model was conducted comparing model results between boys and girls. The model without equality constraints was an acceptable fit to the data: robust χ² (90) = 196.43, p < .001, robust RMSEA = .08, robust RMSEA 90% CI [.07, .10], robust CFI = .96, robust TLI = .91, and robust SRMR = .04. The model displayed strict (i.e., residual) measurement invariance. Application of equality constraints on structural paths did not result in a significant reduction in model fit, χ²(13) = 16.56, p = .22, indicating child sex was not a moderator of associations.

The parent sex multi-group analysis compared model results between mothers and fathers. The unconstrained model was an acceptable fit to the data: robust χ² (90) = 177.38, p < .001, robust RMSEA = .08, robust RMSEA 90% CI [.06, .09], robust CFI = .96, robust TLI = .92, and robust SRMR = .04. The model displayed both metric and scalar invariance, but not strict variance. When constraints of the residual variance terms were added to the model, the model exhibited a significant decline in fit; interestingly, releasing the constraints for the intercepts of distress reactions and the punitive reactions addressed the issue. This modified model was used to test parent sex as a moderator. Application of equality constraints on structural paths did not result in a significant reduction in model fit, χ²(13) = 19.74, p = .10, indicating that parent sex did not moderate associations.

Discussion

Despite the high prevalence of PPD and its close ties to emotion dysregulation in drinkers, there is a gap in knowledge in the growing field of parent emotion socialization research regarding potential consequences of PPD for the development of child emotion regulation. The current study innovatively extends theories regarding the determinants of parent emotion socialization by addressing PPD as an important aspect of parent mental health that can impair parent emotion socialization. PPD and other parent substance use problems may be unique from other parent mental health problems because of the acute and chronic effects of substances on the neurological structures underlying parent emotion regulation (Nutt et al., 2021). Simultaneously, in the substance use literature there is a significant gap in understanding the importance of child emotion regulation in the intergenerational transmission of problem drinking, leading to a failure to target child emotion regulation in programs for at-risk families. Without evidence of the importance of parent emotion socialization, this is unlikely to be remedied.

Summary of Findings

The current study examined indirect associations between PPD and child emotion regulation through parent depression and parent reactions to child negative emotions. PPD was associated with greater parent depressive symptoms, which in turn was related to nonsupportive reactions to child negative emotions (i.e., minimization reactions, punitive reactions, and distress reactions) and lower supportive reactions to child negative emotions (i.e., problem focused reactions, emotion focused reactions, and encouragement of expression). However, only lower supportive reactions to child negative emotions were associated with child emotion regulation difficulties. These associations controlled for child sex, child age, parent sex, parent race, subjective SES, and relationship to child. No direct associations between PPD and child emotion regulation or lability were observed. Furthermore, associations did not differ based on parent or child sex.

Study Findings in Relation to Theory and Prior Research

Observed associations offer support for the Tripartite Model (Morris et al., 2007), which proposes that parent mental health will be indirectly associated with child emotion regulation via parent emotion socialization, parent modeling of emotion regulation, and family emotional climate. The current study specifically tests an important component of parent emotion socialization (i.e., parent reactions to child negative emotions) as an intervening variable, and extends the model to include PPD as an important dimension of parent mental health. Findings are also consistent with the limited prior research on parent substance use, parent emotion socialization, and child emotion regulation. For example, Keller et al. (2022) found that adult retrospective reports of PPD are indirectly associated with rumination through parent reactions to child negative emotions. In addition, mother drug use problems have been indirectly associated with child emotion regulation via parent nonsupportive reactions to child negative emotions (Shadur & Hussong, 2020). However, the current study advanced prior research by obtaining parent reports of their current behavior and focusing on PPD in a non-clinical sample. Important next steps for research include examination of the other mechanisms identified by the Tripartite Model (parent modeling of emotion regulation and family emotional climate) in the context of PPD.

Parents who engage in problem drinking may exhibit less supportive and more nonsupportive reactions to children’s negative emotions because these parents lack the emotion regulation skills needed to cope with child expressions of distress, fear, or anger. Indeed, the current study demonstrates that depressive symptoms serve as an intervening variable in associations between PPD and parent reactions to child negative emotions. Parent depressive symptoms reflect parent difficulties with emotion regulation, including rumination, catastrophizing, and inability to positively reframe events (Garnefski & Kraaij, 2006). Although the current study was cross-sectional and causality cannot be inferred, prior longitudinal research suggests that problem drinking is a predictor of depressive symptoms rather than the reverse direction of association (Furgusson et al., 2009). These parent emotion regulation deficits may be a negative consequence of the use of alcohol to cope with negative emotions (Stevenson et al., 2019), and may reflect neurological impairments resulting from chronic drinking (Carbia et al., 2021). For example, research indicates that problem drinking is associated with a limited adult emotion regulation strategy repertoire and frustration with negative emotions (Paulus et al., 2016; 2017). An important direction for future research is to examine specific parent emotion regulation strategies in associations between PPD and parent reactions to child negative emotions.

Although parent supportive reactions to child negative emotions were associated with better child emotion regulation as expected, the expected association between parent nonsupportive reactions to child negative emotions and poorer child emotion regulation was not observed. Yet, nonsupportive reactions to child negative emotions demonstrated significant bivariate correlations with child emotion regulation and child emotion lability. Null findings are difficult to interpret and it is possible that there was too much overlap in the predictive ability of supportive and nonsupportive reactions to observe both associations together in the same model. Shadur and Hussong (2020) examined parent reactions as a single construct, but Keller et al. (2022) found supportive and nonsupportive reactions to serve as simultaneous intervening variables. In the broader literature on parent reactions to child negative emotions, there are some studies that only examine parent supportive reactions (McElwain et al., 2007; Taylor et al., 2013). Studies that have included both supportive and non-supportive reactions have sometimes found that only nonsupportive reactions are associated with child emotion lability (Morelen et al., 2016; see also Shaffer et al., 2012), although other studies found that only supportive reactions were associated with child adjustment over time (Thompson et al., 2020). For children with anxiety disorders, supportive parent reactions are associated with lower child sadness inhibition and anger dysregulation but nonsupportive parent reactions are associated with lower sadness regulation (Hurrell et al., 2015). Therefore, additional research is needed to clarify associations with parent nonsupportive reactions. There is some evidence that profiles of coping with children’s negative emotions exist and the study of different combinations of supportive and nonsupportive reactions may be beneficial; parent emotion regulation and mindfulness are important correlates of profile membership cross-sectionally, suggesting both of these variables might mediate associations between PPD and parent reactions to child negative emotions (King et al., 2023; McKee et al., 2022; see also Acosta et al., 2021).

Non-clinical Context of Findings

It is important to note that levels of parental problem drinking in the current sample were relatively low in comparison to alcohol use disorder. The implication is that even sub-clinical symptoms of parental problem drinking may be associated with parent dysfunction and have adverse child outcomes. Associations between subclinical levels of parental problem drinking and marital conflict, parent-child conflict, child internalizing, and child externalizing symptoms are often observed (e.g., Keller et al., 2005; Lund et al., 2020). It is possible that parents under-report problem drinking, resulting in the minor levels of self-reported parent problem drinking actually reflecting more significant parental problem drinking in the sample. Alternatively, failure to meet criteria for alcohol use disorder does not necessarily indicate an absence of problems. For example, measures of parental drinking that do not include symptoms of alcohol use disorder (i.e., measures of drinking quantity and frequency) are associated with high stress levels in children (Kim et al., 2020), childhood psychiatric disorders (Suchithra et al., 2022), adolescent drinking to intoxication (Berglund et al., 2022), and sleep and anxiety problems in adulthood (Lund et al., 2023). In summary, PPD may be best thought of as a continuous variable and research with non-clinical samples that are predominated by low levels of self-reported PPD may be successful in explicating more widespread associations between PPD and child development than research with clinical samples.

Distinction Between Child Emotion Regulation and Emotion Lability

It is interesting that associations were observed with child emotion regulation but not child emotion lability. This finding may be due to the dispositional nature of emotion lability. Emotion lability is a construct similar to emotional reactivity, a component of biologically based temperament that is observable from infancy (Rothbart et al., 1994) and that is proposed to serve as a moderator of associations between parent mental health, family influences on child emotion regulation, and emotion regulation in children in the Tripartite Model (Morris et al., 2007). Furthermore, emotional reactivity has been distinguished from emotion regulation as being reflexive, automatic, and rapid rather than reflective, conscious, and planned in the IM and other models (Bridgett et al., 2015; Wills et al., 2013). As such, emotion lability may be less susceptible to the influence of the environment, including parenting influences. Additional research examining developmental trajectories of emotion lability is needed to investigate this possibility, and future research may benefit from examining emotional lability or other measures of emotional reactivity as moderators of associations. An alternative explanation is that the current study was not sufficiently powered to detect associations with emotional lability. There may also be alternative intervening variables for emotional lability. For example, the forms of parent emotion socialization examined in the current study may be less closely associated with emotion lability (i.e., problem focused reactions, emotion-focused reactions) than other forms of parent emotion socialization (e.g., modeling of emotions, emotion coaching, family emotional climate).

Implications for Intervention

A recent Cochran review evaluated the effectiveness of interventions for parents suffering from substance use problems, but defined effectiveness only in terms of reduced parent substance use and did not examine child outcomes (McGovern et al., 2022). The results indicated that the most effective interventions were those that combined substance use and parenting components. These programs were 24 sessions or longer, which may be burdensome and expensive. This may create barriers for patients (Farhoudian et al., 2022). A stronger focus on prevention efforts is therefore likely to be fruitful, but they require basic research identifying the specific psychological processes that are most likely to produce beneficial change in sub-clinical populations.

Findings from the current study contribute to this basic research need. Because associations between PPD, parent reactions to child negative emotions, and child emotion regulation difficulties were observed in a non-clinical sample, findings suggest that targeting of parent emotion socialization may be appropriate for population-level prevention programs. These observed associations also support parent emotion regulation difficulties as a target for intervention in alcohol problems, especially the emotion regulation difficulties that are characteristic of depression. Emotion regulation predicts alcohol use during and after cognitive behavior therapy for alcohol use disorder (Berking et al., 2011). Examination of emotion regulation enhanced therapy among parent problem drinkers is an important next step for prevention science. Mindfulness interventions are also beneficial for enhancing parent emotion regulation, reducing craving, and preventing relapse in problem drinkers (Cavicchioli et al., 2018). Better mindfulness is also related to more supportive reactions to child negative emotions (McKee et al., 2022). Thus, mindfulness training may be especially helpful for parents who consume alcohol. Third, interventions for parent emotion socialization exist that are effective at improving emotion coaching and reducing emotion dismissing among parents of young children (Havighurst et al., 2010) and adolescents (Kehoe et al., 2014). Findings of the current study also support inclusion of emotion socialization components in prevention efforts. Results indicated that promotion of supportive emotion socialization practices may be more valuable than discouragement of unsupportive practices, as supportive practices but not unsupportive practices were associated with child emotion regulation.

Study Limitations

Findings should be interpreted considering study limitations. First, parents were the sole reporters of variables and only a single parent per family participated. Possible consequences include inflated associations due to shared method variance, but also underestimated associations due to social desirability in reporting PPD, nonsupportive reactions, and child emotional difficulties. Future research should collect information from multiple sources (e.g., teachers or children themselves). The Integrated Model (Bridgett et al., 2015) presents compelling evidence that there are neurological and physiological substrates that are important for understanding child emotion regulation and lability, and therefore reliance on questionnaire measures of child emotion regulation and lability are a weakness of the study. Observations of parenting and child emotion regulation and collateral reports of PPD would also be valuable in future research. Second, the study was cross-sectional and no causal inferences can be made. It is likely that there are bidirectional or transactional effects among the study variables. For example, emotionally reactive children may provoke nonsupportive reactions to their emotions and may promote parental drinking as a form of coping (Pelham et al., 1998). Thus, associations may be inflated because they estimate both directions of causality and longitudinal research is needed to address this issue. Finally, the sample was homogeneous in terms of racial/ethnic identity and biological relation to the child. Caution should be taken when generalizing results to other populations and future research including more diverse samples is needed.

Conclusion

The current study advances understanding of associations between PPD and child psychosocial development through examination of child emotion regulation and lability. There have been relatively few studies of child emotion regulation in the context of PPD, despite emotion regulation difficulties underlying many of the child psychosocial problems associated with PPD. Strengths of the study include the examination of parent reactions to child negative emotions as an intervening variable and assessment of both mother and father reactions to child negative emotions (Bridgett et al., 2015). Findings indicate that low parent supportive reactions to children are a potential intervening variable in associations between PPD and child emotion regulation difficulties, and that parental depressive symptoms may help explain why PPD may result in low supportive reactions. These findings support emotion-focused interventions for parents who engage in problem drinking.

Highlights.

• Parental Problem Drinking (PPD) may influence child emotion regulation (ER).

• Relations between PPD, parent emotion socialization, and child ER were tested.

• PPD related to poorer emotion socialization indirectly through parent depression.

• PPD indirectly related to child ER through lower supportive reactions.

• Associations did not differ based on child or parent sex.