Examining Children’s Problem Behaviors and Mothers’ Dating for Mothers’ Depressive Symptoms Following Divorce

Abstract

Although divorce is typically stressful for mothers, the formation of post-divorce dating relationships can help to ease this stress. Unfortunately, research has yet to empirically consider children’s post-divorce adjustment for mothers’ wellbeing leading up to and during mothers’ post-divorce dating. This study addresses the following questions: 1) How do children’s problem behaviors predict mothers’ depressive symptoms following divorce? 2) How do children’s problem behaviors predict the quality of mothers’ dating relationships and the rapport between children and mothers’ dating partners? 3) How do children’s problem behaviors, the quality of mothers’ dating relationships, child-dating partner rapport, and length of mothers’ dating simultaneously impact mothers’ depressive symptoms? Data for this study comes from a longitudinal investigation of recently divorced mothers and their children (N = 232). Hierarchical linear models revealed that mothers experienced more depressive symptoms when their children exhibited more internalizing behaviors. Children’s internalizing behaviors were negatively associated with the quality of mothers’ dating relationships. When examining these variables simultaneously, increases in children’s internalizing behaviors and decreases in relationship quality predicted increases in mothers’ depressive symptoms. Promoting family-level adjustment appears best for mothers’ wellbeing following divorce. Other implications for post-divorce adjustment are discussed.

Mothers’ Mental Health Following Divorce: Examining the Individual and Joint Contributions of Children’s Problem Behaviors and Mothers’ Post-divorce Relationship Quality

Divorce is an extremely salient and emotionally-taxing experience (Amato, 2010; Langlais, Anderson, & Greene, 2016a; Symoens et al., 2014; Shafer et al., 2017), and some divorcing individuals seek new romantic relationships to cope with that stress (Amato, 2010; Anderson & Greene, 2005; Garneau et al., 2015). While the benefits of cohabiting relationships and remarriages are fairly well-documented, few researchers have examined the dating relationships that precede cohabitation and remarriage (i.e., post-divorce dating relationships). Of that limited research, however, is evidence that initiating high-quality dating relationships is positively associated with residential mothers’ wellbeing (Langlais, Anderson, & Greene, 2016b; Symoens et al., 2014). These results, as well as the divorce-stress adjustment perspective, indicate that entering a high-quality dating relationship is (at least initially) beneficial for mothers’ post-divorce adjustment.

While the identified benefits of high-quality dating for mothers’ post-divorce wellbeing are optimistic, we must remember that mothers’ post-divorce relationships are not self-contained between the mother and new partner. Many divorced women have and care for children, and divorce is notably also stressful for children (Ehrlich, 2014; Finley & Schwartz, 2010; Morrison et al., 2017). As children’s post-divorce adjustment difficulties can hinder parents’ repartnering efforts and success (Schnor et al., 2017; Shafer et al., 2016), it is possible that children’s post-divorce adjustment challenges serve as independent stressors on mothers’ post-divorce adjustment. Moreover, it is possible that children’s post-divorce adjustment challenges hinder the quality of their mothers’ new romantic relationships, the rapport mothers’ new romantic partners have with the children, as well as any benefits these relationships may provide for mothers’ adjustment. As such, the question remains: Do children’s behavior problems outweigh the benefits of repartnering for mothers’ post-divorce adjustment?

The goal of this study is threefold. This study aims to examine 1) how children’s problem behaviors are associated with mothers’ depressive symptoms following divorce; 2) how children’s problem behaviors are associated with the rapport between children and dating partners and quality of mothers’ post-divorce dating relationships; and 3) the combined influence of children’s behavior problems, rapport between children and dating partners, and quality of relationships between mothers and dating partners for mothers’ depressive symptoms following divorce. This study examines these relationships using a sample of divorced mothers with residential custody, as divorced mothers tend to face increased challenges to post-divorce adjustment and repartnering compared to other groups (Di Niallo, 2019; Koster et al., 2021). For instance, divorced mothers often retain residential custody of their children and, therefore, face heightened parental responsibilities, as well as reduced family income and social networks (Symoens et al., 2014; Tach & Eads, 2015). Divorced residential mothers must therefore navigate these challenges while simultaneously caring for their (adjusting) children and attempting to reenter the dating market (Schnor et al., 2017; Vanassche et al., 2015). Additionally, we focus on mothers’ dating relationships, as these relationships are mothers’ first attempts at repartnering following divorce (Langlais, Anderson, & Greene, 2016a).

Children’s Post-Divorce Adjustment and Mothers’ Depressive Symptoms

Although divorce is typically stressful for the separating individuals, many divorcing individuals have children; research has consistently noted that children also experience adjustment challenges both leading up to and following their parents’ separation (Ehrlich, 2014; Finley & Schwartz, 2010; Morrison et al., 2017; Shafer et al., 2016). Children’s post-divorce adjustment challenges are generally linked to parents’ pre- and post-divorce wellbeing (i.e., parents’ mental health, interparental conflict, ineffective parenting) and disruptive life events, such as relocating and changing schools (e.g., Amato, 2010; Nielsen, 2017; Weaver & Schofield, 2015). As a result, children with divorced parents are noted to exhibit more behavior problems, social difficulties, psychological distress, and academic challenges when compared to children with married parents (e.g., Ehrlich, 2014).

Although research largely associates children’s post-divorce adjustment with parents’ pre- and post-divorce wellbeing (Demir-Dagdas, 2021; Nielsen, 2017; Weaver & Schofield, 2015), it is possible that some children’s post-divorce adjustment precedes parents’ post-divorce adjustment challenges. These “child effects,” (Ambert, 2001; Montgomery et al., 1992) although theorized, have received limited empirical attention. Research on children’s temperament demonstrate that parents are, indeed, affected by their children’s adjustment difficulties. Some studies have demonstrated that when infants have more difficult temperaments, mothers experience more depressive symptoms (Murray et al., 1996; Whiffen & Gotlib, 1989) and inconsistent parenting (Lengua & Kovacs, 2005). Another study, which assesses the implications of adult children’s divorce on parents’ wellbeing, indicates that when adult children experience divorce, their parents experience significantly increased depressive symptoms; these effects are even greater when the societal acceptance of divorce is low (Tosi & Albertini, 2018). Taken together, these findings illustrate that parents are deeply invested in their children’s wellbeing and that the implications of child adjustment challenges on parents’ adjustment can be seen across the lifespan.

Family systems theory provides theoretical support for this type of bidirectional relationship between mother and children’s adjustment and behaviors. According to this theory (Bowen, 1991; Broderick, 1993), families are comprised of dyads, such as the “mother-child” dyad and the “mother-father” dyad. All of these dyads combine to create the family system. When one (or more) of the dyads experience stress or challenges, the rest of the family system is in flux. As a result, one family member’s stressors often lead to challenges for the remaining family member(s). Previous empirical research shows that if there is difficulty between two members of a dyad, such as a mother and their spouse, members outside of that specific dyad can exhibit adjustment challenges (Langlais, DeAnda, Anderson, & Greene, 2018). This rationale exhibits why, not only could parents’ divorce (i.e., dissolution of mother-father dyad) have independent implications for children’s adjustment, but also how children’s adjustment could impact mothers’ adjustment, even if the mother was not previously experiencing post-divorce adjustment challenges. In other words, the difficulty of parents’ divorce may not only hinder children’s adjustment (Amato, 2010; Coleman et al., 2018), but children’s post-divorce adjustment may hinder their mothers’ adjustment. From the theoretical and empirical evidence, it makes sense that children’s problem behaviors may independently elevate mothers’ depressive symptoms following divorce. We hypothesize that children’s problem behaviors after divorce will be positively associated with changes in mothers’ depressive symptoms within the first two years after divorce-filing.

Children’s Post-divorce Adjustment and Mothers’ Repartnering

As noted through the divorce-stress adjustment perspective (Amato, 2000) and prior research (Cartwright, 2010; Langlais, Anderson, & Greene, 2016b), forming high-quality romantic relationships after divorce is typically beneficial for mothers’ mental health. However, we must remember that mothers’ post-divorce romantic relationships are not isolated between the mother and her new partner (Jensen & Ganong, 2020). According to the family systems theory, the addition of a new dyad (i.e., mother-new dating partner dyad) may predict adjustment challenges for another family member (i.e., child) or dyad (i.e., mother-child dyad) (Hetherington & Clingempeel, 1992; Jensen & Ganong, 2020; Jensen & Lippold, 2018; Jensen et al., 2018; Shafer et al., 2016; Wang & Amato, 2000).

This association has been empirically demonstrated in a number of ways. First, studies have shown that children can inhibit their parents’ dating after divorce (Di Niallo, 2019; Koster et al., 2021). It is common for children to exhibit increased problem behaviors after divorce, as children yearn for a stable family structure (often, the former marital relationship; Hetherington & Kelly, 2002). Children exhibit more problem behaviors when they perceive threats to their original family structure, such as when mothers begin to introduce new dating partners to the family (Ganong & Coleman, 2016). In the presence of these increased behavior problems, mothers may be less able to focus on their romantic relationships, which could hinder relationship formation. Also, if children are not adjusting well to divorce, mothers’ post-divorce adjustment may also be impeded, further inhibiting the formation of mothers’ repartnering relationships. If mothers exhibit signs of separation distress or depression, they may be less likely to fulfill their role as a romantic partner (e.g., DeAnda, Langlais, Greene, & Anderson, 2020), which is also likely to hinder romantic relationship quality.

Moreover, mothers’ introduction and inclusion of new dating partners can also impact the rapport that children have with dating partners. Family systems theory demonstrates that the establishment of the child-dating partner dyad is likely to be stressful for children, as children attempt to resolve family disequilibrium. Children who exhibit problem behaviors following divorce may be averse to accepting the new partner (i.e., mother-partner dyad) in the family unit. Following divorce, many children yearn for their parents to reconcile (Christian, 2005) and could seemingly become resistant to the inclusion of a new romantic partner. When children experience post-divorce adjustment challenges and, hypothetically, are unprepared for their father to be “replaced” in the family unit, children may resist forming a relationship with their mother’s new dating partner.

If children exhibit further adjustment difficulties upon mothers’ repartnering (in addition to other post-divorce adjustment challenges), partners may experience difficulties forming high-quality relationships with the children (Langlais, DeAnda, Anderson, & Greene, 2018; Montgomery et al., 1992). For example, children who are withdrawn or act out may be difficult to form relationships with in the first place. In parents’ new romantic relationships, partners’ parental boundaries or roles may be undefined or ambiguous. Research has identified that a salient challenge in repartnering relationships is parents’ and partner’s navigation of whether and when the dating partner should play a parental or disciplinarian role (Amato, 2010; Hetherington & Kelly, 2002). If children exhibit problem behaviors following divorce and have reservations about including the partner in the family unit, any disciplinary action by the partner is likely to weaken the relationship between children and dating partner.

Finally, research has shown that having a mutually satisfying relationship between the partner and child(ren) serves as the strongest predictor of emotional integration, harmony and stability throughout the repartnering process (Ganong et al., 2019; Jensen & Ganong, 2020). It is possible that, when these relationships are not formed successfully, mothers may experience increased stress and acknowledge the challenges in family integration. These perceptions commonly lead to self-fulfilling prophecies. In other words, if cohesive relationships are not formed between children and mothers’ new romantic partners, mothers may recognize these relationship mismatches as stressors that are challenging to overcome; mothers may then communicate this relationship doubt and, in turn further facilitate low-quality relationships with their dating partners. Given the dyadic relationship between mother and children, we hypothesize that children’s problem behaviors will be negatively associated with the rapport between children and dating partners, as well as the quality of mothers’ dating relationships within the first two years following divorce-filing.

Not only is it possible for children’s adjustment challenges to hinder the quality of mothers’ post-divorce dating relationships, but it may also be possible that children’s adjustment challenges outweigh the benefits offered by high-quality repartnering relationships. Thus, it remains unclear the capacity to which children’s post-divorce adjustment and repartnering relationships simultaneously affect mothers’ post-divorce adjustment. As previously mentioned, children’s behavior is positively associated to parents’ future adjustment (Lengua & Kovacs, 2005; Whiffen & Gotlieb, 1989), such that when children exhibit problem behaviors, their parents experience greater depressive symptoms. Parents’ investment in their children continue throughout the lifespan (Tosi & Albertini, 2018) demonstrating just how critical the parent-child tie is for parents’ wellbeing. Because the mother-child bond preexists the formation of a post-divorce relationship, it is possible that, regardless of any benefits high-quality relationships may provide, mothers will still experience significant challenges when their children experience adjustment challenges following divorce. This may also be true even if mothers maintain a romantic relationship over a long period of time. Although the length of dating relationships usually signals increased commitment and provides more time for a child to form a relationship with a dating partner, this relationship can still be strained if children are exhibiting problematic behaviors. Essentially, children’s problem behaviors may thwart the quality of mothers’ romantic relationships regardless of the length of these relationships, which could hinder maternal wellbeing.

On the other hand, it could be argued that children’s behaviors may not impact mothers’ depressive symptoms as greatly when mothers’ post-divorce dating relationships are high-quality or stable over time. High-quality partners likely assist family development, by providing a variety of family-level supports, such as emotional support for the parent and child, financial support to the family, and/or parenting assistance with children. To best promote family-level adjustment to divorce, it remains important to examine the simultaneous effects of children’s post-divorce adjustment and mothers’ post-divorce dating for mothers’ post-divorce depressive symptoms. Given the limited empirical evidence, we propose the following research question: What are the simultaneous effects of children’s problem behaviors, the rapport between children and dating partners, the quality of mothers’ dating relationships, and the length of these relationships for mothers’ depressive symptoms within the first two years following divorce-filing?

Method

Participants

Participants were obtained through divorce court records from a metropolitan area in the south-central United States between spring 2004 and 2006. Eligible families were those who had filed for divorce within the past 120 days and had at least one elementary-school aged (i.e., kindergarten through 5^th grade) child who resided with the mother at least 50% of each week (N = 319 mother-child dyads). The demographics of the complete sample reflect the area and time in which participants were recruited (United States Census Bureau, 2000).

To adequately address our research aims, the current study is limited to mothers who had not entered “serious” (i.e., serious, cohabiting, remarried) relationships prior to the onset of the study (N = 232). Among these mothers were those who self-identified as having “serious” dating relationships (n = 107), only having “casual” relationships (n = 49) and those who never dated (n = 76) throughout the study. While all mothers in this sample were included in the model investigating the implications of children’s adjustment for mothers’ depressive symptoms (Research Question 1), only mothers who identified as being in serious relationships were included in the models investigating the implications of children’s adjustment for mothers’ repartnering (Research Questions 2a, 2b, and 3). Please refer to Figure 1 for a summary of our subsamples in this study.

Figure 1.

Participants included in the current study.

Note. We limited our sample only to those boxes that appear in black with solid lines.

The average length of marriage (between the mother and her former spouse) was 10.87 years (SD = 5.39) and the average length of separation from the former spouse was 0.94 years (SD = 1.57). The average mother’s age was 37.75 years (SD = 6.68). The majority of mothers identified as non-Hispanic white (65.1%), and the remaining mothers identified as Hispanic (23.7%) or African American (11.3%). Mothers’ education level varied from less than high school (6.0%) to doctoral degree (1.3%), with the median level being “some college”. All children who participated in the study were the biological or adoptive children of the parents who were ending the current marriage. The children were almost evenly split by gender (49% female) and the mean number of children living in the household was 2.00 (SD = 0.83). The average target child was 8.36 years (SD = 2.01) at the baseline assessment.

Procedures

Using contact information provided in the divorce court records, project staff mailed recruitment brochures (which provided information about the study) and made telephone calls to verify families’ eligibility. Eligible families were invited to participate in a “get-acquainted visit” (which occurred in the families’ homes) to answer questions about participation (N = 363). Eighty-eight percent of families who agreed to the get-acquainted visit accepted participation in the study (N = 319). Prior to any data collection, mothers provided written consent for their own, as well as their children’s, participation; children also provided verbal assent for their own participation.

Participating families were first interviewed within 120 days of divorce-filing (i.e., the “baseline” assessment). Standard assessments took place in the families’ homes at the baseline assessment, as well as in 12-month and 24-month follow-up assessments. Two-person interview teams conducted structured interviews with the target child and the mother in separate areas of the home. During breaks in the interview, mothers completed self-report questionnaires by themselves. Additionally, mothers completed assessments at 6- and 18-months of the study, either by mail or phone. If mothers experienced a transition in their dating histories (e.g., began cohabitating or broke up with a partner), the mother and target child completed in-home supplemental assessments, which included specific questions about the romantic partner and relationship (i.e., mothers’ relationship quality, child-partner rapport). Mothers received $50 for completing the baseline assessment, $75 for the 12-month assessment, $100 for the 24-month assessment, and $30 for the 6- and 18-month assessments.

Given the inclusion of the repartnering-based “supplemental” assessments. data collection schedule followed an individually tailored assessment schedule superimposed upon a standard longitudinal panel design. The assessment schedule was time-unstructured (the data collection schedule varied across families), and unbalanced (the number of assessments varied across families). All aspects of this study were approved by the appropriate institutional review board.

Measures

Children’s Internalizing Behaviors

Children’s externalizing symptoms were measured at the baseline, 6-month, 12-month, 18-month, 24-month, and supplemental assessments using the Zill Behavior Problem Index (Peterson & Zill, 1986). This scale was composed of the sum of 12 items and assessed externalizing behaviors over the last six months. Mothers responded to statements such as, “(Child) is unhappy, sad, or depressed” and “(Child) is too fearful or anxious.” Response options ranged from 0 (Not true) to 2 (Often true). Mothers’ responses were multiplied by 10, resulting in responses ranging between 0 and 20. Across all mothers and waves of data, children’s average internalizing score was 3.15 (SD = 2.99), indicating that the average child exhibited very few internalizing symptoms. Cronbach alphas averaged across all five waves of data demonstrated good internal consistency (α = 0.81). Descriptive statistics (mean, minimum, maximum, range) for this and all other measures, as well as the bivariate correlations between these measures, are provided in Table 1.

Table 1.

Correlations Among the Primary Constructs

Study Variable	Mean (SD)	Min	Max	Range	N	1	2	3	4	5	6	7	8
1. Children’s internalizing behaviors	3.15 (2.99)	0.00	17.50	0–20	232	--	.65***	.38***	−.34***	−.27***	−.18**	−.14*	−.22**
2. Children’s externalizing behaviors	4.67 (3.39)	0.00	16.88	0–20	232		--	.31***	−.26***	.07	.06	−.02	−.01
3. Mothers’ depressive symptoms	13.58 (10.41)	0.00	53.68	0–60	232			--	−.21**	−.36***	−.33***	−.40***	−.41***
4. Child-dating partner rapport	2.86 (.79)	0.00	4.00	0–4	104				--	.41***	.36***	.28***	.39***
5. Mother-dating partner rapport	3.60 (.47)	2.00	4.00	0–5	106					--	.69***	.68***	.89***
6. Mothers’ relationship commitment	7.14 (1.23)	0.00	8.00	0–8	106						--	.67***	.89***
7. Mothers’ relationship satisfaction	5.10 (1.02)	0.00	6.00	0–6	106							--	.88***
8. Mothers’ (mean) relationship quality	0.00 (.89)	−4.16	0.83	--	106								--

Note. Descriptive statistics and correlations were calculated using the mean values for all mothers across the study. Min = Minimum. Max = Maximum.

^***p < .001.

^**p < .01.

^*p < .05.

Children’s Externalizing Behaviors

Children’s externalizing symptoms were measured at the baseline, 6-month, 12-month, 18-month, 24-month, and supplemental assessments using the Zill Behavior Problem Index (Peterson & Zill, 1986). This scale was composed of the sum of 16 items and assessed externalizing behaviors over the last six months. Mothers responded to statements such as “(Child) has a very strong temper and loses it easily” and “(Child) is withdrawn, doesn’t get involved with others.” Mothers’ responses were multiplied by 10, resulting in responses ranging between 0 and 20. Across all mothers and waves of data, children’s average externalizing score was 4.67 (SD = 3.39), indicating that the average child exhibited very few externalizing symptoms. Cronbach alphas averaged across all five waves of data demonstrated good internal consistency (α = 0.87).

Mothers’ Depressive Symptoms

Mothers depressive symptoms were measured at the baseline, 6-month, 12-month, 18-month, 24-month, and supplemental assessments using the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977). This scale was composed of the sum of 20 items and assessed depressive symptoms over the last week. Example items were, “I felt lonely” and “I felt sad,” and response choices ranged from 0 (Not at all) to 3 (Most or all of the time). Across all mothers and waves of data, children’s average internalizing score was 13.58 (SD = 10.41), indicating that the average mother exhibited depressive symptoms that nearly met the threshold for clinical depression (clinical cutoff = 14.00). Cronbach alphas averaged across all waves of data displayed good internal consistency (α = 0.92).

Mothers’ Relationship Status and Length

At the baseline, 6-month, 12-month, 18-month, 24-month and during supplemental assessments, mothers reported whether they were 1) not dating, 2) in a “casual” relationship, or 3) in a “serious “relationship (e.g., serious dating relationship, engaged). In the present study, we limited the sample to mothers who had not entered “serious” (i.e., seriously dating, cohabiting, remarried) relationships prior to the onset of the study (N = 232). Among these mothers were those who self-identified as ever having “serious” dating relationships (n = 107), only having “casual” relationships (n = 49) and those who never dated (n = 76) throughout the study. If mothers reported that they were in “serious” relationships, they identified the exact date this relationship began. “Length of relationship” with the new partner (years) was calculated from the mothers’ reports of the when their relationship began. Because mothers only reported on child-partner rapport and relationship quality if they were in serious relationships, we further limit the sample (Research Questions 2a, 2b, 3) to n = 107 mothers. For a summary of how we limited the sample, please see Figure 1. Independent samples t-tests were calculated to determine if there were differences in means with demographic, independent, and dependent variables by dating status. Only one significant difference was found, where mothers who did not date were significantly older than mothers who seriously dated (M = 39.14, 36.37 respectively; F = 4.45; p < .05). Mothers’ age is included as a between-person (i.e., level-2) covariate in all models.

Child-Partner Rapport

Mothers reported on their children’s rapport with their dating partners during assessments that the mother indicated she was in a serious relationship. This measure was composed using the mean of a 5-item scale of rapport (Hetherington & Clingempeel, 1992), with sample items including, “How well does (child) get along with (dating partner)?” and “How much does (child) like spending time with (dating partner)?” Responses ranged from 0 (Not at all) to 4 (Extremely), such that higher scores indicate greater rapport between the child and dating partner. The mean of this variable was 2.86 (SD = .79), indicating that mothers reported fairly high rapport between their children and dating partners. The scale demonstrated good internal consistency (α = 0.93).

Mother-Partner Rapport

Mothers who self-reported that they were in serious repartnered relationships reported their rapport with the partner. This scale was offered at the baseline, 6-month, 12-month, 18-month, and 24-month assessments and during in-home supplemental assessments. Mothers reported their rapport with their partners using the mean of a seven-item scale (Hetherington & Clingempeel, 1992). Sample items included, “How well do you get along with (partner)?” and “How much do you enjoy spending time with (partner)?” Responses ranged from 0 (not at all) to 5 (extremely), such that higher scores indicated greater rapport. Across all waves of data, mothers’ average rapport = 3.60 (SD = 0.47). This scale demonstrated good internal consistency across all waves (α range = .83 - .89).

Mothers’ Commitment to the Relationship

Mothers who self-reported that they were in serious repartnered relationships reported their commitment to the partner. This scale was offered at the baseline, 6-month, 12-month, 18-month, and 24-month assessments and during in-home supplemental assessments. Mothers’ commitment is derived from the mean of six items from Rusbult’s (1980) commitment scale. A sample item includes, “Tell me how committed you feel to maintaining the relationship with (partner)” with responses ranging from 0 (not at all) to 8 (completely/extremely). Items that indicate less commitment are reverse scored so that larger scores represent higher commitment. Across all waves of data, mothers’ average commitment = 7.14 (SD = 1.23) and the scale shows good internal consistency (α range = .82 - .88).

Mothers’ Relationship Satisfaction

Mothers who self-reported that they were in serious repartnered relationships reported their relationship satisfaction. This scale was offered at the baseline, 6-month, 12-month, 18-month, and 24-month assessments and during in-home supplemental assessments. Mothers’ relationship satisfaction is assessed using the mean of 11 items from the Marital Opinion Questionnaire (Huston & Vangelisti, 1991). Mothers were told, “I am going to read two words or descriptions and I would like you to tell me where you think your relationship fits. The first item is ‘hopeful versus discouraging.’ Where does your relationship fall between these two descriptions on the scale?” Response choices range from 0 to 6, with higher scores indicating the second choice presented. When negative items are the second choice presented (e.g., hopeful versus discouraging), items are reverse scored so that higher scores indicate more satisfaction. Across all waves of data, mothers’ average satisfaction = 5.10 (SD = 1.02). This scale demonstrated good internal consistency across all waves (α range = .88 - .92).

Mothers’ Relationship Quality

Mothers’ relationship quality was calculated using the mean of the three relationship subscales (i.e., rapport, commitment, and satisfaction). Mothers’ relationship rapport, commitment, and satisfaction subscales were highly correlated (between r = .67 and .69), suggesting that computing an average value across subscales was appropriate. Given the differences in response option ranges across the three subscales, we transformed each individual measure to its z-score prior to computing mothers’ average relationship quality at each wave.

Covariates

Adjusted models also include several variables that have been consistently associated with mothers’ post-divorce adjustment: mothers’ age, income, length of separation from the ex-spouse, and the number of children living in the household at least 50% of the time (e.g., Raley & Sweeney, 2020). These variables are controlled for at the between-person level (i.e., level-2) and data was collected at the baseline assessment.

Analytic Approach

In the present study, we limited the sample to mothers who had not entered serious relationships by the baseline assessment (N = 232). Limiting the sample in this way allowed us to understand the independent associations between children’s and mothers’ adjustment and repartnering by eliminating mothers who had entered serious relationships before legal separation. To address goals of the study, we conducted multi-level models using hierarchical linear modeling techniques in HLM (Raudenbush & Bryk, 2002). This approach captures within-person (i.e., level-1) and between-person (i.e., level-2) effects of children’s problem behaviors, mothers’ depressive symptoms, and mothers’ repartnering. In addition, HLM accounts for missing data using Full Maximum Likelihood. HLM deletes cases for which data is missing at the between-person level.

For each of the models, we engaged in iterative model building and compared the unadjusted means, unadjusted growth, conditional growth without covariates, and conditional growth with covariates. In each of the models, “unconditional means” model included only the intercept of the outcome. “Unconditional growth” models included the intercept of the outcome and time. When examining mothers’ depressive symptoms independent of their repartnering (Research Question 1), “time” was represented by time in the study (years). Time 0.00 represented the baseline assessment. In analyses that explored mothers’ repartnering (Research Questions 2a, 2b, and 3), time was represented by the length of the relationship between the mother and her partner. Here, “time” 0.00 represented the start of the relationship. Conditional growth models (both without and with covariates) included the intercept, time, time-varying predictors, and controls for the baseline levels of the predictors on the outcomes’ intercepts. Conditional growth models with covariates also included mothers’ ages, income categories, years since separation from the ex-spouse, and numbers of children in the household as level-2 predictors on the intercept (Singer & Willett, 2003). For parsimony, only the adjusted conditional growth models are discussed.

Results

The first goal of the study was to examine the impact of children’s externalizing and internalizing behaviors for changes in mothers’ depressive symptoms following divorce. The results are presented Table 2. The intercept for mothers’ depressive symptoms was significant (p < .001), indicating that mothers’ depressive symptoms at the beginning of the study were significantly different from zero. Children’s baseline internalizing and externalizing symptoms were not associated with mothers’ baseline depressive symptoms. The slope (indicated by time in the study) was also significant (p < .001), indicating that mothers’ depressive symptoms significantly decreased across the first two years following divorce-filing. Changes in children’s internalizing behaviors were positively associated with changes in mothers’ depressive symptoms (p < .001), though changes in children’s externalizing symptoms were only marginally associated with changes in mothers’ depressive symptoms.

Table 2.

The Consequences of Children’s Problem Behaviors for Mothers’ Depressive Symptoms within the First Two Years Following Divorce-Filing (n = 232).

	Unconditional means	Unconditional growth	Conditional growth (Unadjusted)	Conditional growth (Adjusted)
Intercept	13.80 (.57)***	15.92 (.62)***	11.45 (1.04)***	11.68 (1.12)***
Children’s baseline internalizing behaviors	--	--	.24 (.25)	.22 (.24)
Children’s baseline externalizing behaviors	--	--	−.07 (.23)	−.07 (.23)
Mothers’ length of separation from the ex-spouse	--	--	--	−.13 (.36)
Mothers’ age	--	--	--	.16 (.08)~
Mothers’ income	--	--	--	−.36 (.12)**
Number of children living in the household	--	--	--	−.07 (.61)
Time (Time in study)	--	−2.29 (.30)***	−1.98 (.29)***	−1.99 (.29)***
Children’s time-varying internalizing behaviors	--	--	.78 (.15)***	.76 (.15)***
Children’s time-varying externalizing behaviors	--	--	.26 (.15)~	.26 (.15)~
Variance components
Level-1	47.52	44.14	42.03	42.07
Level-2	61.74	61.87	49.59	47.06
Model fit statistics
Deviance	7029.07	6971.09	6754.69	6745.78
AIC	7035.07	6979.09	6770.69	6769.78
BIC	7045.41	6992.88	6798.26	6811.14

Note. Statistics are standardized beta coefficients and are presented as B(SE).

^***p < .001.

^**p < .01.

^~p < .10.

Our second goal was to examine the effects of children’s post-divorce problem behaviors on the rapport between children and dating partners and mothers’ relationship quality in serious relationships. We predicted that children’s problem behaviors would be negatively associated with children’s rapport with dating partners, as well as the quality of mothers’ dating relationships. Results are presented in Tables 3 and 4, respectively. The intercept for child-partner rapport was significant (p < .001), indicating that when mothers first reported on the rapport between their child and dating partner, child-partner rapport was significantly different from zero. Children’s internalizing and externalizing behaviors at the baseline assessment were not associated with initial levels of child-partner rapport, indicating that children’s problem behaviors at the baseline assessment were not associated with the rapport children and partners had at the beginning of mothers’ relationships. The slope was positive and significant (p < .01), indicating that child-partner rapport significantly increased as mothers spent more time in those relationsihps. Neither children’s internalizing nor externalizing behaviors were associated with changes in child-partner rapport.

Table 3.

The Consequences of Children’s Problem Behaviors for Child-Partner Rapport within the First Two Years Following Divorce-Filing (n = 104).

	Unconditional means	Unconditional growth	Conditional growth (Unadjusted)	Conditional growth (Adjusted)
Intercept	2.84 (.07)***	2.72 (.08)***	3.10 (.13)***	3.15 (.14)***
Children’s baseline internalizing behaviors	--	--	−.04 (.03)	−.03 (.03)
Children’s baseline externalizing behaviors	--	--	.01 (.03)	−.00 (.03)
Mothers’ length of separation from the ex-spouse	--	--	--	−.07 (.06)
Mothers’ age	--	--	--	−.03 (.02)*
Mothers’ income	--	--	--	.01 (.02)
Number of children living in the household	--	--	--	.05 (.07)
Time (Length of mothers’ relationships)	--	.23 (.06)**	.20 (.06)**	.21 (.07)**
Children’s time-varying internalizing behaviors	--	--	−.04 (.02)	−.03 (.02)
Children’s time-varying externalizing behaviors	--	--	−.03 (.02)	−.04 (.02)
Variance components
Level-1	.28	.26	.26	.27
Level-2	.37	.37	.29	.26
Model fit statistics
Deviance	571.96	558.07	491.18	483.39
AIC	577.96	571.35	511.56	507.14
BIC	585.89	581.93	532.72	538.87

Note. Statistics are standardized beta coefficients and are presented as B(SE).

^***p < .001.

^**p < .01.

^*p < .05.

Table 4.

The Consequences of Children’s Problem Behaviors for the Quality of Mothers’ Repartnered Relationships within the First Two Years Following Divorce-Filing (n = 106).

	Unconditional means	Unconditional growth	Conditional growth (Unadjusted)	Conditional growth (Adjusted)
Intercept	−.03 (.08)	.08 (.09)	.09 (.16)	.05 (.17)
Children’s baseline internalizing behaviors	--	--	−.04 (.03)	−.04 (.03)
Children’s baseline externalizing behaviors	--	--	.05 (.03)	.03 (.03)
Mothers’ length of separation from the ex-spouse	--	--	--	−.13 (.07)~
Mothers’ age	--	--	--	−.02 (.01)
Mothers’ income	--	--	--	.01 (.02)
Number of children living in the household	--	--	--	.15 (.09)~
Time (Length of mothers’ relationships)	--	.10 (.09)	.09 (.09)	.10 (.09)
Children’s time-varying internalizing behaviors	--	--	−.05 (.03)~	−.05 (.03)
Children’s time-varying externalizing behaviors	--	--	.01 (.03)	.01 (.03)
Variance components
Level-1	.39	.38	.33	.33
Level-2	.40	.41	.40	.36
Model fit statistics
Deviance	601.29	599.48	511.41	503.06
AIC	607.29	607.48	527.41	527.06
BIC	615.28	618.13	548.72	559.02

Note. Statistics are standardized beta coefficients and are presented as B(SE).

^~p < .10.

When examining the associations between children’s problem behaviors and mothers’ relationship quality, the intercept for mothers’ relationship quality was not significant, indicating that, when mothers first entered serious relationships, their mean relationship quality was not significantly different from zero. Children’s internalizing and externalizing behaviors at the baseline assessment were not associated with initial levels of mothers’ relationship quality, indicating that children’s problem behaviors at the baseline assessment were not associated with initial quality of mothers’ serious relationships. The slope was also not significant, indicating that, after accounting for the other variables, relationship quality did not significantly increase or decrease as mothers spent more time in their relationships. Neither children’s internalizing nor externalizing behaviors were associated with changes in mothers’ relationship quality over time.

Our final goal examined the concurrent associations between children’s post-divorce problem behaviors and mothers’ post-divorce dating and mothers’ depressive symptoms. The findings are presented in Table 5. The intercept for mothers’ depressive symptoms was positive and significant, indicating that mothers’ depressive symptoms at the beginning of their relationships were still significantly different from zero. In contrast to when relationship quality was not accounted for (Table 2), children’s baseline internalizing symptoms were positively significantly associated with mothers’ baseline depressive symptoms. Children’s baseline externalizing symptoms were only marginally associated with mothers’ baseline depressive symptoms, indicating that children’s externalizing behaviors at the beginning of the study were not significantly associated with mothers’ depressive symptoms at that time. The slope was only marginally significant, indicating that, after accounting for all other variables in the model, mothers’ depressive symptoms did not significantly decrease across the two years following divorce. After accounting for all other variables in the model, changes in children’s internalizing behaviors were significantly positively associated with changes in mothers’ depressive symptoms, while changes in mothers’ relationship quality were significantly negatively associated with changes in mothers’ depressive symptoms. In other words, when children exhibited more internalizing behaviors, mothers’ depressive symptoms typically increased, but simultaneously, increases in the quality of mothers’ dating relationships lessened these depressive symptoms. Changes in children’s externalizing behaviors and child-partner rapport were not associated with changes in mothers’ depressive symptoms over time.

Table 5.

The Joint Influence of Children’s Problem Behaviors, Child-Partner Rapport, and Mothers’ Relationship Quality on Mothers’ Depressive Symptoms within the First Two Years Following Divorce-Filing (n = 104)

	Unconditional means	Unconditional growth	Conditional growth (Unadjusted)	Conditional growth (Adjusted)
Intercept	13.27 (.78)***	13.55 (.81)***	6.00 (2.98)*	6.18 (3.11)~
Children’s internalizing behaviors	--	--	.62 (.27)*	.57 (.29)*
Children’s externalizing behaviors	--	--	−.55 (.29)~	−.52 (.30)~
Child-dating (first reported) partner rapport	--	--	−.67 (.98)	−.49 (1.00)
Mothers’ (first reported) relationship quality	--	--	.04 (.87)	0.00 (.87)
Mothers’ length of separation from the ex-spouse	--	--	--	−.28 (.60)
Mothers’ age	--	--	--	.04 (.11)
Mothers’ income	--	--	--	−.13 (.16)
Number of children living in the household	--	--	--	.12 (.76)
Time (Length of mothers’ relationships)	--	−2.33 (.42)***	.40 (.78)	.39 (.78)
Children’s time-varying internalizing behaviors	--	--	.54 (.24)*	.55 (.24)*
Children’s time-varying externalizing behaviors	--	--	.34 (.25)	.32 (.25)
Child-dating partner time-varying rapport	--	--	1.20 (.90)	1.10 (.91)
Mother-dating partner time-varying relationship quality	--	--	−2.71 (.76)***	−2.71 (.77)***
Variance components
Level-1	52.50	47.88	26.52	26.42
Level-2	51.89	56.61	21.69	21.58
Model fit statistics
Deviance	3412.44	3384.00	1392.83	1391.97
AIC	3418.44	3392.00	1416.83	1423.97
BIC	3428.78	3405.79	1458.19	1479.12

Note. Statistics are standardized beta coefficients and presented as B(SE).

^***p < .001.

^*p < .05.

^~p < .10.

Discussion

This paper sought to answer three questions: 1) How do children’s problem behaviors impact mothers’ depressive symptoms following divorce? 2) How do children’s problem behaviors impact the rapport between children and dating partners, as well as the quality of mothers’ post-divorce dating relationships? 3) How do children’s problem behaviors and mothers’ dating relationships simultaneously impact mothers’ depressive symptoms following divorce? To answer these questions, we specifically examined two sets of children’s problem behaviors – externalizing and internalizing behaviors – with the rapport between children and mothers’ partners and the quality of mothers’ dating relationships after divorce. Results from this study demonstrated that children’s internalizing behaviors were associated with increases in mothers’ depressive symptoms. When assessing the implications of children’s behavior for mothers’ relationships, neither children’s internalizing nor externalizing behaviors were associated with the rapport between the child and partner or quality of mothers’ post-divorce dating relationships. When assessing the effect of children’s behavior problems and measures of mothers’ repartnering (i.e., child-partner rapport, mothers’ relationship quality) on mothers’ depressive symptoms simultaneously, results were consistent with previous findings; children’s internalizing behaviors were positively associated with mothers’ depressive symptoms, while the quality of mothers’ dating relationships was negatively associated with mothers’ depressive symptoms. Taken together, children’s post-divorce adjustment critically impacts mothers’ relationship quality and depressive symptoms within the first two years following divorce-filing.

Results demonstrated a direct link between children’s problem behaviors and mothers’ depressive symptoms, such that increased internalizing behaviors were associated with increases in mothers’ depressive symptoms within the first two years following divorce-filing. These findings are supported by family systems theory (Bowen, 1991; Broderick, 1993), which illustrates that the mother-child dyad is symbiotic; just as mothers’ wellbeing influences children’s adjustment, results demonstrated that children’s problem behaviors impact mothers’ wellbeing. Although research had previously found links between children’s temperament and life stressors and parents’ later wellbeing (Ehrlich, 2014; Tosi & Albertini, 2018), empirical evidence on the child-driven effects of children’s post-divorce adjustment for parents’ post-divorce adjustment had been limited; the current study demonstrates that children’s post-divorce problem behaviors can also influence mothers’ adjustment within the most sensitive period (i.e., within the first two years) following divorce.

One possible explanation for this finding is that mothers may become depressed when they are unable to ease their children’s internalizing behaviors. Mothers may be indirectly or directly impacted by a child who appears withdrawn or depressed. Given the innate bond between parents and children, parents are deeply invested in their children’s lives. If their children exhibit withdrawn and depressive behaviors, mothers may mirror these same behaviors. Moreover, when children display problem behaviors, mothers may view their parenting as ineffective or poor, which can be damaging to self-esteem. Parents tend to view their offspring as extensions of themselves, even after their children are grown (Levitzki, 2009). Thus, the strong relationship between children and mothers can explain how children’s post-divorce adjustment independently impacts changes in mothers’ depressive symptoms following divorce.

Children’s problem behaviors were neither associated with the rapport between children and mothers’ new partners nor the quality of mothers’ relationships. These findings may stem from the fact that mothers and children only report on “serious” relationships. It is possible that by the time mothers’ relationships become “serious,” they have already weeded out the partners who did not get along with their children or the relationships that did not facilitate the post-divorce adjustment process for their families. In other words, by the time mothers’ relationships become serious, mothers are already aware of the extent to which partners respond to their children’s behavior. It is possible that, if a potential partner responded negatively to the child’s adjustment difficulties, the relationship could not have become serious to begin with.

Moreover, according to the post-divorce-stress adjustment perspective, repartnering relationships can either ease or exacerbate family adjustment challenges (Amato, 2000; Langlais, DeAnda, Anderson & Greene, 2018). It is possible that mothers or partners would terminate a relationship that did not help to facilitate family adjustment. If residential mothers must invest their resources into accommodating their children’s post-divorce lives and adjustment, adding a low-quality relationship to their plates would seemingly add (versus reduce) stress. This explanation is further supported by family systems theory, which demonstrates that children and partners need to have and maintain rapport with one another to form a family “system” (DeAnda, Langlais, Anderson, & Greene, 2021). If mothers and new partners cannot form this system, mothers are more likely to and benefit from terminating those relationships (Langlais, DeAnda, Anderson, & Greene, 2018).

Upon examining the joint influences of children’s problem behaviors, rapport between the child and dating partner, length of time in a relationship, and mothers’ relationship quality for mothers’ depressive symptoms, results demonstrated that, children’s internalizing behaviors and mothers’ relationship quality differentially influence to mothers’ depressive symptoms. These findings are noteworthy, as it appears that the roles of parent and romantic partner both independently influence mothers’ wellbeing. Even after accounting all other variables in the model, children’s internalizing behaviors were still significantly associated with mothers’ depressive symptoms. These findings provide some evidence that mothers’ and children’s post-divorce adjustment are interrelated, even if mothers form new post-divorce dyads (i.e., dating relationships). Mothers may experience role conflict after divorce by trying to fulfill the role of dating partner and mother, which is likely to heighten stress (Anderson & Greene, 2011). In the context of divorce, children’s post-divorce adjustment may have more implications for mothers’ post-divorce adjustment than the relationships between children and dating partners. Mothers may be more concerned about their children’s mental state than the relationship between their child and their dating partner. This cognition may be explained by perceived control (Thuen & Rise, 2006), as mothers may view themselves as responsible for their children’s wellbeing, whereas the rapport between children and dating partners may be more of a task for dating partners than mothers.

On the other hand, results demonstrated that, even after controlling for all other variables, forming high-quality relationships was associated with reductions in mothers’ depressive symptoms. These findings are consistent with the divorce-stress adjustment perspective (Amato, 2000) with previous studies (Langlais, Anderson, & Greene, 2016b; Symoens et al., 2014). High-quality partners can provide varying levels of support, such as financial, emotional, and/or parental support that can mitigate some of the consequences of children’s internalizing behaviors. Low-quality relationship partners will, likely, not show as much investment or commitment to the relationships and may simply act as another stressor in mothers’ post-divorce lives.

It is also possible that having high-quality relationships affirms mothers’ self-perceived mate values, possibly boosting their confidence as romantic partners and parents (Hetherington & Kelly, 2002; Langlais, Anderson, & Greene, 2016a). As mentioned, residential mothers can face extreme challenges adjusting to divorce, parenting, and repartnering simultaneously. Overcoming these challenges and experiencing high-quality relationships is likely both affirming and rewarding. If mothers enter low-quality (or even average quality) relationships, they could view these relationships as direct representations of their mate values, and subsequently experience increases in depressive symptoms. Again, when mothers are in low-quality relationships, their depressive symptoms are likely to increase as mothers may view these relationships as another stressor (in addition to their children’s post-divorce adjustment) that they have to deal with.

Although it is beneficial for mothers to assist children’s adjustment and form high-quality dating relationships following divorce, it may be difficult to perform both of these tasks simultaneously. The ability of mothers to do so may be dependent on their focus on their own and their children’s post-divorce adjustment, which is likely to be difficult. Based on research from Anderson and Greene (2011), some mothers are child-focused (i.e., mothers focus primarily on their children’s adjustment after divorce), while others are adult-focused (i.e., mothers focus mostly on their own adjustment after divorce, often by forming new romantic relationships). Mothers who are child-focused are more aware of their children’s concerns and are less likely to concentrate on forming new romantic relationships until they are more certain of their children’s adjustment. Mothers who are adult-focused, on the other hand, typically focus on children’s concerns as means to help children form bonds with new dating partners (Anderson & Greene, 2011). Based on this study, these approaches are distinctively different. The findings in the current study reveal the importance of being child- and adult-focused for mothers’ depressive symptoms. Future studies should examine ways that divorced parents can incorporate both of these focuses in order to assist their own post-divorce adjustment.

Limitations and Conclusions

Although this study is one of the first to examine child-driven effects of children’s post-divorce behavior problems, mothers’ dating experiences, and mothers’ subsequent wellbeing, it is not without limitations. First, the current study’s sample may have implications for the generalizability of the findings. The sample consists of mothers with primary residential custody. Because many post-divorce families live with and/or rely heavily on extended family members for support (Monostori & Murinkó), it is possible that co-residential caregivers play a role in children and mothers’ post-divorce adjustment and repartnering. The data for this study does not include information on multigenerational co-residence and, therefore, precludes us from examining the extent to which co-residential caregivers impact mother-child or partner-child relationships. In this same vein, it should be noted that it is no longer the “norm” for mothers to have primary residential custody of children following divorce (Meyer et al., 2017). Shared parenting arrangements (i.e., both parents have equal residential custody) are increasingly common and certainly have implications for child and mother adjustment (DeAnda, Langlais, Greene, & Anderson, 2020), as well as mothers’ repartnering (Di Niallo, 2019 Koster et al., 2021). Samples with parents who have different parenting arrangements should attempt to replicate the current findings for generalizability. Finally, the study’s findings are heteronormative. The exclusion of same-sex couples is a serious limitation but was not intentional; at the time of data collection (spring 2004), same-sex marriages were not yet legalized. Without testing this limitation directly, we cannot make any claims to generalizability of our findings past heterosexual residential mothers, but we predict that the results would be the same for same-sex couples with children as has been identified previously (Anderssen et al., 2002).

A second set of limitations relate to the study’s methodology. Mothers reported on their children’s internalizing and externalizing behaviors. Although mothers in this study were likely to spend significant amounts of time with their children (given the custody arrangements required for eligibility), some mothers may be biased in favor of their children. This bias may stem with a fear that their responses may be a reflection of their parenting. Future studies should incorporate methods to measure children’s behaviors more precisely. Third, there are many other variables that are likely to contribute to mothers’ mental health. For instance, co-parenting stress may have significant implications for mothers’ wellbeing (Amato, 2010). Although the current study adds to the literature by examining two different relational processes for mothers’ mental health, examining other relationship processes, such as those that involve the other parent, could further clarify the influence of post-divorce variables for mothers’ wellbeing.

In sum, this study examines an understudied topic with implications for family adjustment after divorce. Results illustrate that children’s behavior problems and mothers’ relationship quality both (perhaps in opposition of each other) influence mothers’ long-term wellbeing. These results also reveal that post-divorce dating experiences not only impact mothers, but the entire family system, and in return, the family system influences mothers’ dating. It is important to consider how each member of the family (and each dyad) contributes to relationship processes that could impact post-divorce adjustment. Evidence from this study emphasizes the importance of promoting children’s adjustment after divorce and forming high-quality dating relationships in order to assist mothers’ post-divorce adjustment.

Highlights:

1. Children’s internalizing behavior predicted mothers’ depressive symptoms.

2. Children’s internalizing behaviors predicted declines in mothers’ relationships.

3. Relationship quality and children’s internalizing behaviors predicted mothers’ depressive symptoms.