Multilevel Factors Associated with Unsupportive Emotion Socialization: An Examination of Child Maltreatment and its Sequelae

Abstract

Purpose: Adults who have been maltreated as children are at risk for a variety of adverse sequalae that can have a negative impact on parents’ emotion-related socialization behaviours (ERSBs) and contribute to the intergenerational transmission of emotion regulation difficulties. However, various supports may reduce unsupportive behaviours. Informed by Belsky’s (1984) determinants of parenting model, the goal of the present study was to examine multi-level stressors and supports that may contribute to, or discourage, parents’ use of unsupportive ERSBs. Methods: Mothers and young adults (aged 18–25) from across Canada participated in an online study (N = 185 dyads). Mothers responded to questionnaires assessing multi-level stressors and supports, while young adults reported on their mothers’ ERSBs during adolescence. Results: A hierarchical regression analysis revealed that mothers who reported more impulse control difficulties, who experienced greater revictimization, and who had more severe dissociative symptoms were rated as higher in their use of unsupportive contingencies. However, when contextual supports were added into the model, only revictimization remained associated with unsupportive contingencies. Furthermore, mothers’ positive perceptions of the parent-child relationship were negatively associated with unsupportive contingencies. Conclusions: Preventing revictimization amongst survivors of child maltreatment may be an effective way to prevent the intergenerational continuity of unsupportive emotion socialization and emotion regulation difficulties. Interventions that improve parent-child attachment relationships may also be beneficial to survivors who want to break the cycle of maladaptive parenting.

Parents’ emotion-related socialization behaviours (ERSBs) describe the ways in which parents discuss and react to their children’s emotions (Eisenberg et al., 1998). Parents’ reactions to children’s emotions can be classified into supportive (e.g., problem-solving) or unsupportive (e.g., punitive) contingencies (Eisenberg, 2020), and there is evidence to suggest that unsupportive contingencies can contribute to the intergenerational transmission of emotion regulation difficulties (Cabecinha-Alati et al., 2022). It is important to study determinants of ERSBs so that we can identify targets for prevention and intervention, particularly among survivors of child maltreatment (CM) who may be more likely to experience parenting problems (Plant et al., 2018). Thus, the present study will examine CM and its sequelae as social determinants of parents’ unsupportive contingencies.

The Determinants of Parenting

Over three decades ago, Belsky (1984) proposed a process model aimed at elucidating the social determinants of parenting behaviours. The model posits that parental functioning is multiply determined and outlines three factors that influence parenting including: (a) parents’ individual characteristics and psychological resources; (b) individual characteristics of the child (including the effect of the child on the child-caregiver microsystem); and (c) the broader exosystem in which the parent-child relationship is embedded including contextual sources of stress and support (Belsky, 1984; Belsky & Jaffee, 2015). The model also assumes that parents’ psychological wellbeing and parental functioning may be traced back to parents’ developmental histories. Consequently, research on the determinants of parents’ ERSBs has garnered increased attention in recent years (Eisenberg, 2020) and evidence suggests that these behaviours are strongly influenced by parents’ own capacity for emotion regulation (Hajal & Paley, 2020).

Mothers’ Developmental History

Child Maltreatment

Individuals who have experienced CM are less likely to be exposed to parents who model supportive contingencies and more likely to be subjected to emotionally invalidating behaviours (Rea & Shaffer, 2016). A recent study by Milan and colleagues (2021) demonstrated that mothers with a history of CM recalled being subjected to more unsupportive contingencies during their own childhoods, which in turn predicted their own use of unsupportive contingencies when their children expressed negative emotions. Given that different forms of CM tend to co-occur (Burczycka & Conroy, 2017) and contribute to relational difficulties in a dose-dependent fashion (Steine et al., 2017), future studies must examine the influence of cumulative CM on unsupportive contingencies.

Insecure Attachment

CM has also been associated with an increased risk for developing insecure attachment styles that can persist into adulthood (Oshri et al., 2015), and cumulative experiences of CM may contribute to higher levels of attachment insecurity (Godbout et al., 2017). As such, mothers who have insecure attachment styles (i.e., attachment anxiety and avoidance) may be more likely to rely on unsupportive contingencies (Jones et al., 2014). However, no studies to our knowledge have examined the relative influence of cumulative CM and adult attachment on mothers’ ERSBs.

Revictimization

Another adverse effect of cumulative CM is a greater risk of experiencing multiple types of interpersonal victimization after childhood (Edalati et al., 2016). Cumulative CM is associated with intimate partner violence (IPV) victimization over and above the effects of any single type of CM (Brassard et al., 2020). In turn, IPV is associated with destructive interparental conflict and unsupportive contingencies (Lee & Brophy-Herb, 2018). However, relative to mothers who are victimized exclusively during childhood or adulthood, mothers who have been revictimized present a greater risk for negative parenting behaviours (Stephenson & Renk, 2019). Although this effect appears to be best explained by maternal emotion regulation difficulties, a limitation of this research is that revictimization was represented by a dichotomous variable, which may have masked the impact of cumulative experiences of revictimization.

Mothers’ Individual Characteristics

As a result of developmental stressors including CM, the development of insecure attachment styles, and revictimization, survivors of CM are also prone to emotion regulation difficulties (see Gruhn & Compas, 2020 for review).

Emotion Regulation Difficulties – Impulse Control

Parenting is an emotionally demanding task where parents need to regulate their own emotions while simultaneously engaging in co-regulation with their distressed child (Rutherford et al., 2015). According to Gratz and Roemer (2004), emotional dysregulation is characterized by impairments in emotional awareness, clarity, understanding, and acceptance, as well as the ability to engage in goal-directed behaviours, inhibit impulses in the face of negative emotions, and modulate emotional responses with appropriate regulatory strategies. Impulse control specifically is a key aspect of emotion regulation that can be impaired after CM (Sigrist et al., 2021; Williams et al., 2015). A previous study conducted by our research team found that a maternal history of CM was positively associated with mothers’ impulse control difficulties, which in turn, predicted young adults’ recollections of unsupportive contingencies during adolescence (Cabecinha-Alati et al., 2022). Moreover, difficulty with impulse control was the dimension of emotion regulation that was most strongly correlated with unsupportive contingencies (Cabecinha-Alati et al., 2021). However, this study did not account for the potential influence of revictimization. As such, more research is needed to determine whether revictimization and impulse control difficulties exert independent influences on unsupportive contingencies.

Post-Traumatic Stress Symptoms

Concomitant with revictimization and impulse control difficulties, mothers who have experienced CM are at an increased risk for post-traumatic stress symptoms (PTSS), which can interfere with parenting (Samuelson et al., 2017; Williams et al., 2021). In addition, symptoms of depression, substance use, and personality disorders, which are known to be present in parents with a history of CM, may put such parents at a greater risk for unsupportive contingencies (e.g., Breaux et al., 2016; Gurtovenko & Katz, 2020). Maternal anxiety symptoms (which included PTSS) have also been associated with the use of unsupportive contingencies (Breaux et al., 2016). However, no studies to our knowledge have examined the unique influence of trauma-related psychopathology (i.e., PTSS) on unsupportive contingencies that are common among survivors of CM (Vang et al., 2018).

Dissociation

Congruent with the pattern observed in PTSS, mothers who experienced interpersonal trauma in childhood and revictimization in adulthood, reported higher levels of dissociation compared to mothers who were not revictimized (Hulette et al., 2011). The tendency to dissociate may be conceptualized as a disengagement strategy that protects the individual from overwhelming hyperarousal (Lewis et al., 2020). Dissociative symptoms can generalize to affect the quality of the parent-child relationship and impair maternal sensitivity (Moser et al., 2013). Individuals prone to dissociation also appear to have lower levels of cognitive empathy (Chiu et al., 2016) and this association was partially explained by deficits in emotional awareness, a pre-requisite for effective emotion socialization (Gottman et al., 1996). Taken together, parents who struggle with dissociation may be more likely to rely on unsupportive contingencies, but the influence of dissociation on unsupportive contingencies is equivocal.

Maternal Age at Childbirth

Studies have identified a link between a history of CM and teenage motherhood (Madigan et al., 2014). Evidence suggests that younger mothers who have experienced severe CM may be more likely to engage in emotionally unsupportive parenting behaviours (i.e., psychological control, hostility, and unavailability) compared to older mothers with similar abuse histories, regardless their level of emotion regulation difficulties (McCullough et al., 2015). Consequently, motherhood at a young age may also negatively affect  ERSBs that are independent from emotion regulation abilities.

Familial / Contextual Sources of Stress and Support

Family Economic Wellbeing

Becoming a parent at a young age has been associated with single parenthood, poverty, lower levels of education, and job instability, which can exacerbate parenting stress (Steele et al., 2016). However, many individuals who are survivors of CM do not experience these difficulties because of contextual sources of support, such as higher levels of education or living in a low-problem neighbourhood (Meng et al., 2018). Thus, parents with higher education and family incomes may be less likely to use unsupportive contingencies (Lugo-Candelas et al., 2016).

Neighbourhood Economic Disadvantage

In contrast, contextual stressors such as neighbourhood economic disadvantage may have a negative effect on parenting (Belsky & Jaffee, 2015). Unemployment and poverty were found to be associated with increased social isolation and parenting stress (Belsky & Jaffee, 2015), both of which are known to increase the demand on parents’ self-regulatory capacities. Thus, contextual sources of stress and support may contribute to, or discourage, unsupportive contingencies.

Characteristics of the Child – The Child’s Influence on the Child-Caregiver Microsystem

Mother’s Perceptions of the Child and Parent-Child Relationship

Finally, Belsky (1984) argues that individual characteristics of the child (or the effect of the child on the child-caregiver microsystem) may also influence parenting behaviours, albeit less strongly than parents’ individual characteristics or contextual sources of stress and support. Compared to mothers without a history of CM, mothers who have experienced abuse or neglect may be more likely to perceive their child’s temperament as “difficult” (Casanueva et al., 2010) and have unrealistic expectations that lead them to make negative attributions about their child’s behaviour (Malone et al., 2010). The parent-child relationship may become even more challenging during adolescence as teens strive for greater autonomy and individuation from their parents, resulting in elevated levels of conflict, negative emotionality, and lower levels of maternal supportiveness (De Goede et al., 2009). However, more positive parent-child relationships characterized by perceptions of openness, warmth, and mutual responsiveness have been positively associated with parents’ reliance on supportive contingencies (Criss et al., 2016). Consequently, mothers who perceive their child and the parent-child relationship more positively may be less likely to use unsupportive contingencies.

The Present Study

In sum, several studies have identified determinants of parental ERSBs, however a major shortcoming is that ERSBs are often self-reported (Eisenberg, 2020). Additionally, many studies have focused on child or parent characteristics, while fewer have examined interpersonal and contextual factors (e.g., Godleski et al., 2020; Shaffer et al., 2012). As such, the present dyadic study aimed to examine the relative influence of multilevel stressors and supports on parents’ unsupportive contingencies with the goal of elucidating the relative impact of a parental history of CM and its sequalae (see Fig. 1).

Fig. 1.

Model of the present study for the determinants of parenting

Given that the literature presented in this review focused predominantly on mothers, the present study focused on maternal ERSBs. Additionally, we chose to use young adults’ ratings of mother’s ERSBs since ratings from the child’s point of view may be less vulnerable to self-report biases (Briscoe et al., 2019). This method also reduces the potential inflation issues that can arise when the predictors and outcome are reported by the same informant. Research has only just begun to acknowledge how ERSBs from earlier stages of development continue to impact individuals in young adulthood (McKee et al., 2021). Only two studies have examined recollected ERSBs from the period of adolescence (Cabecinha-Alati et al., 2022; Lugo-Candelas et al., 2016). This is surprising given that adolescence is a critical period with new psychosocial challenges that necessitate increasingly sophisticated emotion regulation abilities (Zimmermann & Iwanski, 2014). As such, it is crucial to better understand the stressors and supports that contribute to, or discourage, parents’ use of these potentially harmful unsupportive behaviours.

In line with Belsky’s model, it was hypothesized that parents’ individual characteristics (i.e., maternal impulse control difficulties, PTSS, dissociation, age at childbirth) would be more strongly associated with unsupportive contingencies than stressors originating from their developmental history (i.e., cumulative CM, insecure attachment, interpersonal revictimization), as well as any contextual stressors (i.e., neighbourhood economic disadvantage). We also hypothesized that contextual sources of support (i.e., mothers’ education, family economic well-being) and mothers’ positive perceptions of their child / the parent-child relationship would be negatively associated with unsupportive contingencies when controlling for the above multi-level stressors, but that these effects would be smaller than the effects of parents’ individual characteristics. In the context of the present study, mothers’ perceptions of the parent-child attachment relationship, as opposed to the young adults’ perceptions, were used to represent the effects of the child on the child-caregiver microsystem.

Methods

Participants

Mothers and their young adult children (18–25 years old) were recruited from across Canada using social media (e.g., Facebook ads), online research platforms (e.g., Research Stream, Honeybee), and ads that were distributed through university departmental listservs, as part of a larger study on CM. The final sample included 252 mothers, 578 young adults, and 185 complete dyads. Complete dyads were included in the present study. See Table 1 for demographic characteristics of the mothers (100% female, M = 51.19 years old, SD = 5.81) and young adults (89.7% female, M = 20.88 years old, SD = 2.18).

Table 1.

Demographic variables and mothers’ victimization characteristics

Variables		Mother n (%)	Young adult n (%)
Ethnicity (n = 365)	White	130(70.3)	127(68.6)
	Asian	35(18.9)	35(18.9)
	Black	4(2.2)	4(2.2)
	Arab / Middle Eastern	4(2.2)	4(2.2)
	Hispanic	3(1.6)	3(1.6)
	Indigenous / Native American	3(1.6)	2(1.1)
	Mixed race	1(0.5)	10(5.4)
Educational attainment (n = 368)	Elementary school or less	3(1.6)	0(0)
	High school	31(16.8)	63(34.1)
	CEGEP or professional school	42(22.7)	25(13.5)
	Undergraduate degree or certificate	74(40)	83(44.9)
	Graduate degree	33(17.8)	14(7.6)
Household income (n = 314)	Less than $20, 000/yr	4(2.2)	37(20)
	$20, 000 - $59, 999/yr	35(18.9)	43(23.3)
	$60, 000 - $99, 999/yr	48 (26.0)	25 (13.5)
	$100, 000 or more/yr	76(41.0)	46(24.8)
Physical Abuse (n = 185)	At least one form	77(41.6)	--
Emotional Maltreatment (n = 185)	At least one form	74(40.0)	--
Sexual Abuse (n = 185)	At least one form	53(28.6)	--
Neglect (n = 184)	At least one form	52(28.1)	--
Witness to IPV (n = 185)	At least one form	54(29.2)	--
Revictimization (n = 182)	At least one form	87(47.0)	--
Physical Assault IPV (n = 86)	At least one form	28(15.1)	--
Psych. Aggression IPV (n = 86)	At least one form	82(44.3)	--
Sexual Violence (n = 87)	At least one form	28(15.1)	--

Note: CEGEP = Collège d’enseignement général et professionnel is a precollege university program unique to Quebec. Psych. = Psychological. IPV = Intimate partner violence. Sexual Violence refers to sexual assault experiences in adulthood that occurred inside or outside the context.

Measures

Demographic Information

All participants (i.e., mothers and young adults) were asked to provide demographic information including their gender, ethnicity, educational attainment, annual household income, and age at first childbirth (if the participant was a parent).

Maternal History of Cumulative Child Maltreatment

Three subscales from the Early Trauma Inventory Self-Report – Short Form (ETISR-SF; Bremner et al., 2007) were used to assess mothers’ experiences of physical abuse (five items), sexual abuse (six items), and emotional maltreatment (five items). The short form of the ETI has good psychometric properties and can distinguish patients with known trauma histories from comparison subjects (Bremner et al., 2007). Additionally, five items from the IPSCAN Child Abuse Screening Tool – Retrospective Version (ICAST-R; Dunne et al., 2009) were used to assess mothers’ experiences of childhood physical neglect. The development of the ICAST-R was based on consensus from international experts in child protection and it has adequate measurement properties as a retrospective survey tool (Dunne et al., 2009). Lastly, three questions were adapted from the Revised Conflict Tactics Scale (CTS2; Straus et al., 1996) to assess whether mothers had witnessed IPV prior to the age of 18. The CTS2 is one of the most widely used instruments for assessing IPV and has good psychometric properties (Chapman & Gillespie, 2019). In the present study, mothers were asked to indicate whether they experienced or witnessed specific acts prior to the age of 18 using a dichotomous (Yes/No) response format, for example, “Were you ever slapped in the face with an open hand?” (Physical Abuse). Questions were specific to perpetration by a parent or caregiver except for sexual abuse (intra and extrafamilial abuse was included). Internal consistencies for the five subscales ranged from α = .71 to .87. Like previous studies (e.g., Berthelot et al., 2014; Brassard et al., 2020), the present study used a composite score to quantify mothers’ cumulative CM. First, a dichotomous score (i.e., 0 or 1) was created for each type of CM that was endorsed. Participants who endorsed an item on any of the five subscales were classified as having experienced that type of CM. Second, a sum was computed to reflect the total number of maltreatment types that mothers reported ranging from 0 (no CM) to 5 (experienced all five types of CM).

Mothers’ Insecure Attachment

The Experiences in Close Relationships Questionnaire – Short Form (ECR-12; Lafontaine et al., 2015) is a 12-item abbreviated self-report measure, of which 6 items that assess mothers’ attachment avoidance (e.g., “I don’t feel comfortable opening up to romantic partners) and another 6 items used to measure anxiety (e.g., “I worry a fair amount about losing my partner”). Responses were recorded on a Likert-type scale ranging from 1 (Disagree strongly) to 7 (Agree strongly). Like the original ECR, the ECR-12 has good psychometric properties (Lafontaine et al., 2015). A mean score was used for attachment avoidance (α = .88) and anxiety (α = .85).

Mothers’ Experiences of Interpersonal Revictimization

Ten items from the Revised Conflict Tactics Scale – Short Form (CTS2-SF; Straus & Douglas, 2004) were used to assess mothers’ experiences of IPV victimization. Mothers were presented with a series of statements (e.g., “My partner pushed, shoved, or slapped me”) and were asked to indicate how often this more specific event occurred within the last year (e.g., 1 = Once in the past year; 7 = Not in the past year, but it did happen before). The CTS2-SF is like the original, with internal consistencies ranging from .79 to .95 (Straus et al., 1996). The present study focused on subscales measuring the lifetime prevalence of Physical Assault IPV (α = .84), Psychological Aggression (α = .55) IPV, and Sexual Coercion IPV (α = .81). Additionally, one item from the Life Events Checklist for the DSM-5 (LEC-5; Weathers et al., 2013) was used to assess whether mothers experienced sexual assault (i.e., rape or attempted rape) after the age of 18. This item was included to account for sexual victimization that may have occurred outside the context of a couple relationship. A dichotomous score (i.e., 0 or 1) was computed for each type of interpersonal victimization after childhood (Physical Assault, Psychological Aggression, Sexual Violence). A revictimization score was computed indicating whether mothers also reported a history of CM. Mothers who did not report a history of CM were assigned a score of 0. Mothers who endorsed CM were given a cumulative score (Edalati et al., 2016) from 0 (no revictimization) to 3 (three types of revictimization).

Mothers’ Other Adverse Events in Adulthood

The LEC-5 (Weathers et al., 2013) was used to account for non-interpersonal adverse events in adulthood (e.g., transportation accident, life-threatening illness). Mothers were assigned a count score ranging from 0 to 9 representing the number of events they endorsed.

Maternal Emotion Regulation – Impulse Control Difficulties

The Difficulties in Emotion Regulation Scale – Brief Version (DERS-18; Victor & Klonsky, 2016) is an 18-item self-report measure of emotion regulation difficulties with good psychometric properties. The Impulse subscale (i.e., difficulties with controlling impulses in response to negative emotions) was used (e.g., “When I’m upset, I have difficulty controlling my behaviours”). Responses were recorded on a Likert-type scale ranging from 1 (Almost never) to 5 (Almost always). The Impulse subscale displayed good internal consistency (α = .93).

Mothers’ Trauma-Related Psychopathology

The PTSD Checklist for the DSM-5 (PCL-5; Weathers et al., 2013). The PCL-5 is a psychometrically sound 20-item self-report measure that assessed mothers’ PTSS (Blevins et al., 2015). After completing the LEC-5 and other measures related to interpersonal victimization, mothers were instructed to keep in mind the worst experience they ever had while responding to the questionnaire. For each item, mothers were asked to indicate how much they were troubled by their symptoms in the past month (e.g., “feeling jumpy or easily startled”). Responses were recorded on a Likert-type scale ranging from 0 (Not at all) to 4 (Extremely). A sum score was computed for mothers’ PTSS severity (α = .96).

The Dissociative Experiences Scale Taxon (DES-T; Waller & Ross, 1997). The DES-T is an eight-item subscale of the DES that was used to assess mothers’ dissociative symptoms (e.g., “Some people are told that they sometimes do not recognize friends or family members”). The original DES has good test-retest reliability (r = .93) and high internal consistency (α = .96) (Dubester & Braun, 1995). The DES-T is used to detect more extreme, pathological forms of dissociation (Waller et al., 1996). Mothers were asked to indicate the percentage of the time that a particular dissociative symptom was experienced (0–100%). A mean score was used to assess the severity of mothers’ dissociative symptoms (α = .95).

Mothers’ Household Economic Well-Being

The Index of Material Deprivation (Azeredo & Payeur, 2015) is a 17-item self-report measure adapted from the Canadian Survey of Economic Wellbeing (Statistics Canada, 2013). Mothers were asked to respond to questions about their level of financial comfort. For example, “Can you afford to pay your bills on time?”. Responses are recorded using a dichotomous (Yes/No) response format and a sum score (Yes = 1, No = 0) was computed for household economic wellbeing (α = .87), with higher scores representing greater financial comfort.

Neighbourhood Disadvantage

Mothers provided the first three characters of their postal code. The first three characters of a postal code, referred to as forward sortation areas, represent a specific postal delivery area within a major geographical region (Statistics Canada, 2017). The first three digits of postal codes were entered into Postal Code OM Conversion File Plus (PCCF+) software, which was used to generate a neighbourhood income quintile variable (ranging from 1 to 5). A lower income quintile represents greater levels of neighbourhood disadvantage (Canadian Institute for Health Information, 2018).

Mothers’ Perceptions of their Child and the Parent-Child Relationship

The Revised Inventory of Parent Attachment (R-IPA; Johnson et al., 2003) is a 30-item self-report measure that was used to assess mothers’ perceptions of their child and the parent-child attachment relationship. The R-IPA includes two dimensions: Trust / Avoidance (23 items; e.g., “I trust my child”, “I feel my child is good”, “my child cares about my point of view”) and Communication (7 items; e.g., “I like to get my child’s point of view on things I am concerned about”, “my child can tell when I’m upset about something”) that can be combined to create a sum score. Responses are recorded on a Likert-type scale ranging from 1 (Almost never or never true) to 5 (Almost always or always true). The measure had acceptable psychometric properties in the development sample (Johnson et al., 2003) and in the current study (α = .94). In the context of the present study mothers’ perceptions, as opposed to the young adults’ perceptions, were used so that this variable would capture the impact of the child on the child-caregiver microsystem. In contrast, the young adult version of the measure represents the young adult’s perception of their mother, and the impact of the mother on the child-caregiver microsystem. Higher scores on this measure indicate that mothers have more positive perceptions of their child and of the parent-child relationship.

Young Adult Ratings of Mothers’ Emotion-Related Socialization Behaviours

The Coping with Children’s Negative Emotions Scale – Adolescents’ Perceptions Version (CCNES-AP; Fabes & Eisenberg, 1998) was used to assess young adults’ perceptions of their mother’s responses to their negative affect during adolescence. Like Lugo-Candelas and colleagues (2016), the existing scale was modified by switching the nine vignettes into past tense and asking participants to recall their teenage years. For example, “When my mother saw me becoming anxious about something at school, she would usually…”. Six possible reactions were presented for each scenario and three subscales that assessed unsupportive contingencies were used: Distress Reactions (e.g., become nervous and uneasy in dealing with my anxiety), Punitive Reactions (e.g., get angry at me for not dealing with things better), and Minimizing Reactions (e.g., tell me that I was making too big a deal out of it). Responses were recorded on a Likert-type scale ranging from 1 (Very unlikely) to 7 (Very likely). Internal consistencies for these subscales were α = .91, .92, and .91, respectively, and a mean score was computed for mothers’ unsupportive contingencies (e.g., Mezulis et al., 2015; Nelson et al., 2009).

Procedure

Once the study received ethical approval from the primary researcher’s institution, (Research Ethics Board at McGill University; REB # 278–1118) measures were administered via an online Qualtrics survey (Qualtrics, Provo, UT). Prior to completing questionnaires, participants selected a language (English or French) and provided informed consent. The first participant filled out contact information for themselves as well as for the second member of their dyad. Each participant who completed the survey in mother-young adult pairs was compensated with a $5 e-gift card. Every participant was entered into a draw to win an iPad, regardless of whether their other family member participated.

Data Analysis

Data analysis was conducted using SPSS Version 24. Participants were excluded if they were duplicate responses (n = 68); they did not provide a valid identification number or used the same email address for both members of the dyad (n = 39); they completed less than 75% of the survey (n = 36); or they were outside of the age range specified in our inclusion criteria (n = 69). Three additional measures were implemented to screen out careless responders. First, cases (n = 38) were excluded if the participant answered “no” to a question that asked them to indicate whether they felt that their data was valid. Second, participants (n = 121) were excluded if they failed to follow the instructions to more than 3 of the 5 directed questions that were randomly distributed throughout the survey. Lastly, participants (n = 17) were excluded if they completed the survey in less than half of the modal completion time. The final sample included 185 complete dyads. Following preliminary analyses, a three-step hierarchical regression analysis was conducted to predict young adults’ ratings of mothers’ unsupportive contingencies. In line with Belsky’s (1984) theoretical model, mothers with a history of cumulative CM and associated developmental stressors (i.e., revictimization and insecure attachment) were entered into the first block of the model. Parents’ individual characteristics (i.e., maternal impulse control difficulties, PTSS, dissociative symptoms, and age at first childbirth) were entered into the second block. Finally, contextual sources of support (i.e., economic well-being) and mothers’ perceptions of the parent-child relationship were entered into the third block.

Results

Descriptive Analysis – Mother’s Victimization Characteristics

Our results showed that 64.3% of mothers in the present study reported experiencing at least one type of CM. For those with cumulative CM experiences, 16.2% reported experiencing two types of CM, 11.4% reported three types, 13.5% reported four types, and 5.9% of mothers reported experiencing all five types of CM. Additionally, 47% of mothers who endorsed a history of CM reported that they also experienced interpersonal revictimization after childhood. For those who were revictimized, 27.0% reported one type of interpersonal revictimization, 12.4% reported two types, and 7.6% reported experiencing all three types. Rates for each type of CM and revictimization are displayed in Table 1.

Preliminary Analyses

Missing data on the variables of interest was minimal (less than 5%), except for neighbourhood disadvantage (8.1% missing). Little’s MCAR test, χ2(92) = 91.92, p = .48, revealed that data was likely missing completely at random. As such, listwise deletion was used (Tabachnik & Fidell, 2007). Four variables, including mothers’ attachment avoidance, other adverse life events in adulthood, educational attainment, and neighbourhood disadvantage, were excluded from analyses since they were not correlated with unsupportive contingencies (see Table 2). Two outliers with large, standardized residuals (> 3SD) were also removed. The means, standard deviations, and correlations are highlighted in Table 2. Bivariate correlations revealed that unsupportive contingencies were correlated with all study variables except for neighbourhood disadvantage, mother’s experiences of other adverse life events in adulthood, mother’s attachment avoidance, and mother’s education.

Table 2.

Means, standard deviations, and correlations (N = 154)

	1) UC	2) ND	3) CCM	4) RV	5) OAE	6) A-AV	7) A-AN	8) ICD	9) PTSS	10) DS	11) ACB	12) EDU	13) HEW	14) M-C A
1	2.47 (1.24)	− 0.07	0.18*	0.24**	0.06	0.08	0.26**	0.35**	0.20**	0.30**	− 0.17*	− 0.11	− 0.25**	− 0.51**
2	-	3.23 (1.41)	− 0.00	− 0.09	− 0.04	− 0.11	− 0.10	− 0.14	− 0.09	− 0.20*	0.08	0.10	0.15	0.12
3	-	-	1.58 (1.60)	0.69**	0.17*	0.20*	0.34**	0.37**	0.49**	0.33**	− 0.10	− 0.03	− 0.38**	− 0.32**
4	-	-	-	0.71 (0.94)	0.33**	0.22**	0.34**	0.40**	0.62**	0.40**	− 0.16	− 0.08	− 0.45**	− 0.31**
5	-	-	-	-	1.20 (1.29)	0.07	0.12	0.12	0.22**	0.11	0.03	0.04	− 0.25**	− 0.14
6	-	-	-	-	-	2.52 (1.27)	0.23**	0.30**	0.44**	0.30**	− 0.11	− 0.02	− 0.28**	− 0.32**
7	-	-	-	-	-	-	3.11 (1.51)	0.37**	0.45**	0.32**	− 0.06	− 0.10	− 0.35**	− 0.44**
8	-	-	-	-	-	-	-	4.69 (2.43)	0.49**	0.50**	− 0.08	− 0.06	− 0.30**	− 0.46**
8	-	-	-	-	-	-	-	-	14.29 (15.87)	0.53**	− 0.20*	− 0.11	− 0.50**	− 0.39**
9	-	-	-	-	-	-	-	-	-	6.81 (14.76)	− 0.23**	− 0.13	− 0.54**	− 0.43**
10	-	-	-	-	-	-	-	-	-	-	25.75 (5.09)	0.24**	0.24**	0.09
11	-	-	-	-	-	-	-	-	-	-	-	3.48 (1.01)	0.24**	0.07
12	-	-	-	-	-	-	-	-	-	-	-	-	16.03 (2.22)	0.40**
14	-	-	-	-	-	-	-	-	-	-	-	-	-	119.68 (17.18)

Note: *Significant at p < .05 level. ** Significant at p < .01 level. The means and standard deviations are displayed on the diagonal. UC = Unsupportive Contingencies. ND = Neighbourhood Disadvantage. CCM = Cumulative Child Maltreatment. RV = Revictimization. OAE = Other Adverse Events. A-AV = Attachment Avoidance. A-AN = Attachment Anxiety. ICD = Impulse Control Difficulties. PTSS = Post-traumatic Stress Symptoms. DS = Dissociative Symptoms. ACB = Age at Childbirth. EDU = Education. HEW = Household Economic Wellbeing. M-C A = Mother-Child Attachment Relationship

Hierarchical Regression Analysis Predicting Mothers’ Unsupportive Contingencies

Regression coefficients and effect sizes for each predictor are displayed in Table 3. Effect sizes are measured using Cohen’s f² (.02 = small, .15 = medium .35 = large; f² = R²_inc/1-R²_inc) (Geert van den Berg, 2020). Developmental stressors were significantly associated with unsupportive contingencies, F(3, 169) = 8.25, p < .001, accounting for 12.8% of the variance. Revictimization and attachment anxiety were both positively associated with unsupportive contingencies, but cumulative CM was not.

Table 3.

Regression coefficients (N = 173)

	Predictors	Model 1				Model 2				Model 3
		B	SE	rpart	f 2	B	SE	rpart	f 2	B	SE	rpart	f 2
Step 1	Cumulative CM	− 0.08	0.07	− 0.08	0.01	− 0.09	0.07	− 0.09	0.01	− 0.11	0.07	− 0.11	0.01
	Revictimization	0.42***	0.12	0.25	0.06	0.31*	0.13	0.16	0.03	0.32**	0.12	0.17	0.03
	Attachment Anxiety	0.14*	0.06	0.16	0.03	0.09	0.06	0.10	0.01	0.00	0.06	0.00	0.00
Step 2	IC Difficulties	-	-	-	-	0.12**	0.04	0.20	0.04	0.07	0.04	0.11	0.01
	PTSS	-	-	-	-	− 0.01	0.01	− 0.01	0.01	− 0.01	0.01	− 0.12	0.01
	Dissociative Symptoms	-	-	-	-	0.02*	0.01	0.15	0.02	0.01	0.01	0.08	0.01
	Age at childbirth	-	-	-	-	− 0.02	0.02	− 0.09	0.01	− 0.02	0.02	− 0.09	0.01
Step 3	Economic Wellbeing	-	-	-	-	-	-	-	-	− 0.01	0.04	− 0.02	0.00
	M-C Relationship	-	-	-	-	-	-	-	-	− 0.03***	0.01	− 0.33	0.13

Note: Regression coefficients are presented in unstandardized form. CM = Child Maltreatment. IC = Impulse Control. PTSS = Posttraumatic stress symptoms. M-C = Mother-Child. * Significant at p < .05 level. ** Significant at p < .01 level. *** Significant at p < .001 level. r_part is the semipartial correlation for each predictor variable. Cohen’s f² = R²_inc/1-R²_inc, where R²_inc is equivalent to the squared semipartial correlation for each predictor. f² = 0.02 = small effect, 0.15 = medium effect, 0.35 = large effect

When individual level characteristics were added in the second block, the overall model was significant, F(7,165) = 7.18, p < .001, and increased in its predictive power (𝛥F(4, 165) = 5.70, p < .001), accounting for an additional 10.6% of the variance in unsupportive contingencies. Revictimization remained associated with unsupportive contingencies, but not attachment anxiety. Maternal impulse control difficulties and dissociative symptoms were also positively associated with unsupportive contingencies; however, mothers’ PTSS and age at first childbirth were not. The strongest predictor was maternal impulse control difficulties followed by revictimization, and dissociative symptoms, which all had a small effect on unsupportive contingencies.

Finally, entering contextual supports and mothers’ perceptions of the parent-child relationship into the model explained an additional 11.6% of the variance in unsupportive contingencies, and this change in R² was also significant, 𝛥F(2, 163) = 14.41, p < .001. Revictimization remained associated with unsupportive contingencies while mothers’ positive perceptions of mother-child relationship were negatively associated with unsupportive contingencies. Economic wellbeing was not associated with unsupportive contingencies and maternal dissociative symptoms as well as impulse control difficulties were no longer significant. The overall model was significant, F(9, 163) = 9.72, p < .001, and together, the stressors and supports factors accounted for 34.9% of the variance in unsupportive contingencies. The strongest predictor in the last step was mother-child attachment quality, which had a small-medium effect on unsupportive contingencies.

Discussion

Associations between a history of CM with parenting problems have been documented (Hughes & Cossar, 2016), yet only a handful of studies have examined how CM and its sequalae may influence parental ERSBs (Cabecinha-Alati et al., 2021, 2022; Martin et al., 2018; Milan et al., 2021). To expand on previous research on ERSBs that has predominantly focused on parents’ individual characteristics (Eisenberg, 2020), the present study used Belsky’s (1984) determinants of parenting model to examine stressors and supports occurring across multiple levels.

Mothers’ Developmental History and Individual Characteristics

The hypothesis that mothers’ individual characteristics would be more strongly associated with unsupportive contingencies than stressors at other levels was supported, since maternal impulse control difficulties had the strongest effect on unsupportive contingencies when controlling for developmental stressors. With all other stressors statistically controlled for, mothers who reported more impulse control difficulties, experienced greater revictimization, and had more severe dissociative symptoms were rated as higher in their use of unsupportive contingencies. These findings are consistent with previous research, which showed that revictimization (rather than CM alone) contributed to mothers’ negative emotional responsivity when adolescents expressed negative emotions (Martin et al., 2018; Molina et al., 2020), as well as research that suggests that parental difficulties with impulse control may contribute to unsupportive contingencies (Cabecinha-Alati et al., 2022; Hajal & Paley, 2020).

Moreover, consistent with Belsky’s (1984) model, CM (as an individual predictor) was not associated with ERSBs, since it is argued that mothers’ developmental history serves as a starting point that then influences mother’s individual characteristics, which seemed to be the most influential variable. It is possible that the impact of CM on ERSBs is contingent on the level of emotional dysregulation that mothers experience. For instance, Choi and Kangas (2020) found that mothers with a history of high betrayal trauma (including CM) were more likely to experience parenting stress when maternal emotion regulation difficulties were moderate or high in contrast to milder. Similarly, McCullough and colleagues (2015) found that mothers who had experienced high levels of CM were more likely to engage in negative parenting behaviours (i.e., hostility, psychological control, and unavailability) in the context of high levels of emotional dysregulation but were not at risk for these behaviours when emotional dysregulation was low.

The finding that anxious attachment was no longer significant when impulse control difficulties were entered into the model aligns with a previous study (Jones et al., 2014), which demonstrated that the influence of maternal attachment anxiety on unsupportive contingencies occurred through maternal emotion regulation difficulties. Similarly, although maternal age at childbirth and PTSS were associated with unsupportive contingencies in bivariate correlations, these associations did not persist once entered in the regression model. Previous literature suggests that these variables may interact with emotion regulation to influence parenting behaviours (e.g., Gurtovenko & Katz, 2020; McCullough et al., 2015) and as such, future studies should examine whether this is also the case for parental ERSBs.

Lastly, although this is the first study, to our knowledge, that has examined the relationship between dissociation and parental ERSBs, our findings coincide with studies that have linked dissociation with punitive parenting behaviours (Collin-Vézina et al., 2005) and parenting stress (Williams et al., 2021), which can negatively impact supportive ERSBs (Wu et al., 2019). Thus, dissociation may represent a maladaptive strategy that traumatized parents use to disengage from the stress elicited by their child’s negative affect.

Mothers’ Perceptions of the Parent-Child Relationship and Contextual Supports

In addition to examining the relative influence of different stressors, the present study aimed to identify factors that might discourage parents from using unsupportive contingencies. When contextual supports (i.e., household economic wellbeing) and mothers’ perceptions of the parent-child relationship were entered into the last step of the model, the associations between parents’ individual characteristics (i.e., impulse control and dissociation) and unsupportive contingencies were no longer significant.

The perceived quality of the mother-child relationship was the only factor that was negatively associated with unsupportive contingencies and contrary to our second hypothesis, this was the strongest predictor in the final step. It is possible that mothers who can cultivate secure parent-child relationships despite their histories of victimization are more resilient and have been able to, develop strategies (e.g., cognitive reappraisal; Bao & Kato, 2020) or skills (e.g., mentalizing; Milan et al., 2021) that have enabled them to manage their emotional and psychological difficulties so that these problems do not spill over into the parental role and negatively impact ERSBs.

In contrast, mothers who have been unable to resolve their abuse experiences may be more likely to engage in unsupportive contingencies that contribute to the transmission of disorganized attachment (DeOliveira et al., 2004). Our finding that mothers’ positive perceptions of the mother-child relationship related to lower ratings of unsupportive contingencies align with a recent longitudinal study (Wang et al., 2021). However, given our cross-sectional design, we cannot rule out the possibility that lower unsupportive contingencies in adolescence may be contributing to more positive parent-child relationships at present.

The Influence of Revictimization

After contextual supports and the parent-child relationship were entered into the model, the only stressor that remained associated with unsupportive contingencies was revictimization. According to Banyard and colleagues (2003), a history of CM may put mothers at risk for more proximal stressors, such as IPV, that can negatively affect parenting. IPV has been associated with increased parenting stress (Pinto et al., 2019), which in turn, can contribute to emotionally unsupportive parenting behaviours (Loucks & Shaffer, 2014). Previous studies have also identified a link between destructive partner conflict and parents’ use of unsupportive contingencies (e.g., Lee & Brophy-Herb, 2018), and these effects may be specific to mothers who have been victimized in both childhood and adulthood (Martin et al., 2018).

Unlike previous research, which found that the association between revictimization and mothers’ negative emotional responsivity towards adolescents was non-significant after maternal emotion regulation difficulties (Martin et al., 2018), the present study found that revictimization and difficulties with impulse control exerted independent influences on mothers’ unsupportiveness contingencies (see Molina et al., 2020 for similar findings).

Consequently, it may be important to provide interventions for all IPV-exposed mothers, rather than targeting mothers with emotion regulation difficulties more specifically (Molina et al., 2020). Interventions that combine emotion regulation skills training with training in emotion-coaching appear promising in that they have been shown to increase supportive contingencies (e.g., validation) and reduce unsupportive contingencies (e.g., scolding) among IPV-exposed mothers (Katz et al., 2020). Our results suggest that these interventions may also be important for mothers who have experienced multiple forms of revictimization.

Limitations and Future Directions

Although the present study helps shed light on determinants of maternal ERSBs, our findings are not without limitations. First, our reliance on a retrospective self-report measure of CM is a limitation given that using a single method to identify CM may overlook certain cases (Shaffer et al., 2008). Agreement on prospective and retrospective report measures of CM are typically poor as they tend to capture distinct groups of individuals (Baldwin et al., 2019). As such, future studies would benefit from combining multiple sources of information (Newbury et al., 2018) to assess for CM. Secondly, most mothers in our sample were university educated, endorsed high economic well-being, and did not reside in disadvantaged neighbourhoods, which may have prevented us from detecting relationships between these factors and maternal ERSBs.

Our findings also highlight that CM and ERSBs were not associated with one another, which might be indicative of a potential moderation effect that was not included in our analyses. Past research has shown that the effect of CM on parenting behaviours or ERSBs was dependent on the level of emotion regulation difficulties (Dix, 1991; Rutherford et al., 2015). Future studies should explore this interaction effect between CM and ERSBs. Additionally, our definition of trauma-related psychopathology was limited to dissociation and PTSS rather than including more general psychological difficulties (e.g., depressive symptoms) that have been associated with a history of CM and ERSBs (Choi & Kang, 2021). We also did not include other child-related variables (e.g., problem behaviours during adolescence) that may have influenced bidirectional parent-child interactions and mothers’ ERSBs (Wang et al., 2021).

Being that Belsky’s (1984) model and Fig. 1 are bidirectional and that attachment is also a dyadic construct, research should ideally include measures that reflect reports from mothers and their children. In the present study, mothers’ perceptions of their child and the parent-child relationship were used to represent the influence of the child on the parent-caregiver microsystem. However, the child’s perceptions of their mother and the parent-child relationship also impacts the child-caregiver microsystem, which may in turn, influence their mother’s relationship perceptions and behaviours. As such, future studies should explore the bidirectional nature of this relationship in more depth.

Furthermore, this research used young adults’ retrospective reports of emotion socialization during adolescence within the context of a cross-sectional design. As such, we cannot rule out the possibility that current levels of emotion regulation difficulties, parent-child attachment quality, psychological distress, or depression had an effect on young adults’ recollections of their mothers’ parenting behaviours (Ehrlich et al., 2014; Garside & Klimes-Dougan, 2002).

Lastly, our cross-sectional design and retrospective measure of ERSBs preclude us from establishing directionality. The transition from adolescence to young adulthood is characterized by realignment (i.e., warmth and closeness) in parent-child relationships (Lanz, 2017) and recollections of parental ERSBs may have been influenced by the current quality of the parent-child attachment relationship (e.g., Ehrlich et al., 2014). However, given the mean age of the young adult sample (20.88 years old) and the fact that 52% still lived at home, adolescence is relatively recent for most participants and may even be considered a protracted developmental stage due to delays in the assumption of adult roles in Western society (Riediger & Klipker, 2014). There is also some evidence to suggest that parent-child closeness remains relatively stable between the ages of 17 to 22 (Fang et al., 2021). Consequently, retrospective ratings of ERSBs may still be useful, especially since the parent-child attachment relationship appears to be a more stable predictor of ERSBs than vice versa (Wang et al., 2021).

Strengths and Implications

Notwithstanding these limitations, a notable strength of the present study is its use of data from both members of the parent-child dyad. Previous studies examining predictors of maternal ERSBs have used self-report data from the same informant (typically the mother), which inflate the associations between the predictors and the outcome variable. Additionally, by assessing different types of interpersonal victimization, the present study was able to tease apart the impacts of cumulative CM and revictimization, which builds off extant literature on cumulative risk (Cohen et al., 2008), as well as research that has attempted to ascertain the relative influence of CM and adult victimization on parental ERSBs (e.g., Martin et al., 2018).

Taken together, findings from the present study support the notion that a history of CM is not necessarily deterministic of poor parenting. Rather, there are several pathways through which CM may heighten one’s vulnerability for parenting difficulties and conversely, many pathways through which the intergenerational transmission of risk may be disrupted (Morelen et al., 2018). Our results do, however, suggest that preventing revictimization amongst survivors of CM may be an effective way also circumvent the continuity of unsupportive ERSBs (Leerkes et al., 2020) and emotion regulation difficulties (Cabecinha-Alati et al., 2022). Interventions that promote parent-child attachment relationships (e.g., Mom Power; Rosenblum et al., 2017), particularly amongst parents and their adolescent children (e.g., Tuning Relationships with Music; Colegrove et al., 2019), may help to decrease unsupportive contingencies amongst survivors, and in turn, improve adolescents’ capacity for emotion regulation (Herd et al., 2021) so that these difficulties do not persist in future generations.

Data availability

Data not available – participant consent: The participants of this study did not give written consent for their data to be shared publicly, so due to the sensitive nature of the research, supporting data is not available.

Declarations

Conflict of interest

Authors have no conflicts of interest to disclose.