Coercive control and situational couple violence in families with child protection involvement: A case-file analysis

Introduction

Domestic and family violence (DFV) is a significant issue in families who are involved with statutory child protection services (Coulter & Mercado-Crespo, 2015; Henry, 2018; Holmes et al., 2019; Humphreys & Healey, 2017; Lawson, 2019). DFV can cause harm to children and young people; for example, if they witness it or are injured during violent incidents. It is also associated with increased risk of child abuse and parenting difficulties for both the perpetrators and victims of DFV (Ainsworth, 2020; Harwin & Barlow, 2022). Further, children are not merely bystanders or witnesses to DFV but are victim–survivors of it in their own right. Child victims of DFV may live in fear of the perpetrator, may feel terror that one of their parents will be harmed or killed by the other, and may continue to be harmed by DFV, even after parental separation (Katz, 2016). Further, DFV commonly occurs together with other forms of child maltreatment, meaning that many children who live with and/or survive DFV are also impacted by physical abuse, emotional abuse and/or neglect (Higgins et al., 2023). There has been growing recognition that not all DFV is the same, with some researchers differentiating between DFV that is characterised by coercive control and DFV that is situational in nature (Johnson, 2008; McKay et al., 2022; Myhill, 2017; Nancarrow et al., 2020; Ross, 2011; Stark, 2007). Coercive control does not always involve physical violence. It is a pattern of behaviour in which a perpetrator controls multiple aspects of the victim’s life, often by using threats, intimidation or violence. It results in victims having limited autonomy and often living in fear of the perpetrator (Stark, 2007).

Situational couple violence occurs in the context of conflict between a couple and does not involve an overarching pattern of controlling and dominating behaviour. Although situational couple violence can involve serious violence and may be unilateral (i.e. only one person using physical violence), the conflict is mutual in nature and victims of this violence type are likely to maintain autonomy and unlikely to fear the perpetrator at times other than during incidents of violence. Research on situational couple violence has found that both men and women use violence at similar rates; however, when severity and impact is considered, women are more likely to experience injury or severe violence by men than the other way around (Johnson et al., 2014). As such, situational couple violence does not occur independently of gendered power dynamics (Johnson et al., 2014), and responses to it should prioritise holding perpetrators accountable for their use of violence and managing the impacts this has on both adult and child victims.

The importance of recognising and understanding coercive control and the impact it has on victims and children has been highlighted in research on DFV in the context of child protection practice (Humphreys et al., 2020; Mandel & Wright, 2019). Researchers have found that child protection interventions with families impacted by DFV are often characterised by mother blame and a failure to recognise how a perpetrator’s use of coercive control impacts on both adult victims and child victims (Humphreys & Absler, 2011; Humphreys et al., 2011; Mandel & Wright, 2019). Some researchers have argued that child protection practice frameworks that define DFV as coercive control are vital in ensuring that child protection practitioners identify and address the controlling and manipulative behaviours used by many perpetrators of DFV (Humphreys & Healey, 2017; Humphreys et al., 2018; Mandel & Wright, 2019). Others, however, have suggested that approaching DFV as a homogeneous phenomenon, and always characterised by coercive control, may not meet the needs of all families with child protection involvement (Ferguson et al., 2020; Lawson, 2019). To date, there has been little research that considers the difference between coercive controlling violence and situational couple violence in the child protection context (Lawson, 2019).

Understanding whether DFV in families involved with child protection services includes both coercive control and situational couple violence is important because these differing types of DFV may have different impacts on adult and child victims (Johnson, 2006; Johnston, 2006; Katz, 2016). Further, coercive control and situational couple violence may have differing causal factors and require different responses (Armenti et al., 2016; Bernardi & Day, 2015; Cleary Bradley & Gottman, 2012; Haselschwerdt, 2014; Love et al., 2020; McCann, 2021; Schneider & Brimhall, 2014; Stith & McCollum, 2011). For example, researchers have suggested that situational couple violence may be caused by issues such as communication or conflict management difficulties, which can be addressed through joint couple counselling (Armenti et al., 2016; Karakurt et al., 2016; Stith & McCollum, 2011), and associated with life stressors such as poverty (Johnson & Ooms, 2016). Coercive control, on the other hand, may be grounded in the perpetrator’s beliefs about gendered roles in relationships and desire for power and control (Johnson, 2008; Stark, 2007) and researchers have argued that joint couple approaches, or other approaches focussed on addressing conflict and communication issues, are not appropriate or safe in this context (Karakurt et al., 2016).

Therefore, the aim of this study was to explore whether it was possible to identify both violence types, and to differentiate between them on the basis of information contained in a sample of intake reports provided by the South Australian Department for Child Protection (DCP), a government department responsible for investigating and responding to child abuse and neglect.

Positionality

The first author of this article is a social worker with experience in child protection practice and conducted this study as part of a doctoral thesis. The other authors are the first author’s doctoral supervisors. All of the authors are non-Indigenous, which limits our perspectives. As we have not collaborated with Aboriginal researchers for this study, we have been cautious in exploring the potential meaning and impacts of our research for First Nations families, and the privilege we hold as white Australians who live and work on colonised land has undoubtedly meant that we do not share the same insights and experiences Aboriginal researchers may have. Further, the first author’s child protection practice experience has been beneficial to many aspects of this research; for example, in developing a trusting relationship with DCP, which supported access to sensitive data. However, practicing in child protection in Australia also means participating in a system that still harms First Nations families and that, at times, continues to be informed by (and to reinforce) racist ideas and policies. While it is not our intent to reinforce such ideas in this paper, the fact that none of the authors are Aboriginal and our lack of engagement with First Nations researchers may have resulted in us missing opportunities to design and carry out our study in a culturally informed way that prioritises the needs of Aboriginal children and families. For example, as a non-Indigenous practitioner who has developed assessment and analysis skills through working in a system that is often not culturally informed or culturally safe, the first author may have analysed the data differently to how an Aboriginal researcher would have. Ideally, there would have been joint data analysis together with at least one First Nations researcher, but due to the nature of the research agreement and the extremely sensitive nature of the data and the need to protect the privacy of families whose data were shared, only the first researcher was able to access and analyse the data. We acknowledge that our limited perspectives have impacted on this study and hope that in the future there may be opportunity for collaboration with First Nations researchers, or opportunity for First Nations researchers to critique and/or build on our work with the aim of making it more relevant for Indigenous Australians.

Method

For this study, we used a case-file analysis process to analyse a data set of reports detailing child protection notifications and child protection history. The process of families receiving a child protection response from DCP involves the generation of an ‘intake report’. This is a document that is generated when a report is made by members of the public or professionals to the ‘Child Abuse Report Line’ – the screening arm of DCP that determines whether concerns about children meet the threshold for a response from DCP. A separate intake report is generated for each child subject to a notification; that is, the report pertains to a child rather than a family. This report records the current notification and includes a summary of all prior notifications made about that child.

Approval for this study was granted by DCP and the Australian Catholic University Human Research Ethics Committee, approval number 2020-133H. The approval terms ensured that the confidentiality of client families who were the subject of the intake reports was protected.

For this study, DCP data systems staff used a random number generation tool to select a sample of 100 intake reports that were ‘screened in’ during a 12-month period between 2021 and 2022, and that met criteria for the risk/harm type of ‘Domestic and Family Violence’ as the grounds for statutory intervention (Bromfield & Higgins, 2005). As intake reports include sensitive confidential information, including names, dates of birth and addresses of children and their family members, DCP provided the reports to the lead researcher via a secure one-time access link. The lead researcher then extracted information from the original intake reports and recorded it in de-identified form. A number was allocated to each family and, where multiple reports concerned the same family (i.e. the family had multiple children who had been included in the sample), these were noted as duplicates and only one report was used for analysis. Where a report did not include concerns of DFV (i.e. the worker undertaking the screening accidentally selected the wrong risk/harm type or ‘grounds for intervention’), they were identified as invalid and not included in the analysis. In total, 100 cases were provided. After eliminating duplicates and invalid cases, 77 remained.

The lead researcher conducted a qualitative analysis using the de-identified information to determine whether it contained indicators of coercive control or situational couple violence. Many studies on DFV primarily focus on physical violence but prior research on coercive control has demonstrated that identifying indicators and impacts of controlling behaviour is an effective way of conducting a case-file analysis with this focus (Myhill & Hohl, 2019). For this study, we used key literature about situational couple violence and coercive control (Johnson, 2008; Stark, 2007) to develop classification criteria based on behaviours and impacts of each violence type. For a detailed description of the criteria used to classify the cases, see Appendix I. Using these indicators, we found that the cases in the sample could be placed into one of five categories – ‘coercive control’ and ‘situational couple violence’, which were cases with clear indicators of each DFV type, ‘possible situational couple violence’ and ‘possible coercive control’, which were cases with one indicator of the DFV type but not enough to make them clearly identifiable as one or the other, and ‘unclear’, in which the information was not indicative of either DFV type.

Results

Of the 77 cases analysed, 20 were classified as ‘coercive control’, 27 as ‘situational couple violence’, 10 as ‘possible coercive control’, 6 as ‘possible situational couple violence’, and 14 as ‘unclear’.

Of the cases classed as ‘situational couple violence’, three involved conflict/violence between adult family members who were not a couple. These cases were included in the ‘situational couple violence’ category because the nature of violent conflict between these adult family members was such that the impact on children in the household would have been similar if they had been a couple (i.e. the adults were household members who shared caregiving responsibilities for the children).

Although all cases classified as ‘situational couple violence’ included mutual violence or aggression, use of violence was not equal in all cases (see Fig. 1). Overall, 12 of the 27 cases involved equal use of violence by both parents/caregivers, two involved the mother being the primary perpetrator and 10 involved the father being the primary perpetrator. Two of these cases involved verbal aggression only (no physical violence). One of these involved the mother using more severe verbal aggression, the other the father. The remaining three cases involved violence that was between parents and extended family members.

Figure 1. Characteristics of mutual violence in situational couple violence cases (n = 27)

In eight ‘situational couple violence’ cases, the violence was severe, resulting in either hospitalisation, significant injury or criminal charges against one or both parties. In three of the cases involving serious violence, both parents appeared to have used equally severe violence, either during the same incident, or there was a history of both parties causing significant injury to the other in separate incidents. In five, the father/stepfather had used significantly more serious violence, leading to only the mother having a significant injury.

The ‘possible situational couple violence’ category consisted primarily of cases where there was use of violence by only one parent/caregiver (the father/stepfather in all cases), but some reference to mutual aggression/conflict or indicators that the mother/victim was not scared of the perpetrator and/or that they did not have to modify their behaviour due to the violence. This category, as with the ‘situational couple violence’ category, included some cases in which there was significant physical violence. There were two cases in which the violence had resulted in injury, and two more that involved previous concerns that had resulted in intervention orders being issued against the father.

Of the 20 cases classified as ‘coercive control’, most had more than three behavioural indicators of coercive control; that is, the information in the intake report suggested the perpetrator was using a range of multiple controlling behaviours. Five cases had five or more indicators, and six had four indicators. Four cases had three indicators, and two had two indicators, the minimum number required for cases to be classed as ‘coercive control’. Cases with only one indicator of controlling behaviour were classed as ‘possible coercive control’.

In the majority of the ‘coercive control’ cases, the information in the intake report suggested the victim was scared of the perpetrator; however, there were only five cases in which coercive control was accompanied by significant physical violence. In the majority of the ‘coercive control’ cases, the perpetrator’s behaviour consisted of non-physical forms of abuse and intimidation, including threats, stalking, preventing the victim from leaving the relationship, sexual abuse or attending the victim’s home despite intervention orders being in place. A substantial proportion of the ‘coercive control’ cases (7 out of 20; 35%), involved separated couples. In all of these cases, mothers were seeking safety from abuse for themselves and their children but the perpetrator continued to use controlling, intimidating and threatening behaviour to cause fear. In comparison, only 3 out of 27 cases (11%) in the ‘situational couple violence’ group involved separated families and none of those involved one person being afraid of the other after separation, with both parties seeming to play a role in maintaining contact and conflict. Only one case in the ‘coercive control’ category involved a mother being the perpetrator.

Although mutual violence was not a contra-indicator of coercive control in our classification criteria (to account for potential violent resistance), only 3 of the 20 ‘coercive control’ cases included mutual violence. In the ‘possible coercive control’ category, 3 of the 10 cases involved mutual violence, making it twice as prevalent in this category as in the ‘coercive control’ category.

In total, 39 of the 77 intake reports (just over 50%) concerned Aboriginal families; that is, families in which the child/children were Aboriginal. In most of these cases, both parents were Aboriginal, but in some cases just one parent was, or this information was missing for one parent, or not all children in the family had the same father and mother. There were no Torres Strait Islander, South Sea Islander or Māori families.

Figure 2 shows how many Aboriginal and non-Aboriginal families from our sample were in each DFV category. When the DFV cases among Aboriginal (n = 39) and non-Aboriginal families (n = 38) were compared, there was a slightly larger proportion in the ‘situational couple violence’ category (41% or 16 out of 39 cases vs 37% or 14 out of 38 cases) and a substantially smaller proportion in the ‘coercive control’ category (15% or 6 out of 39 cases vs 29% or 11 out of 28 cases) in the Aboriginal group. The larger proportion of situational couple violence cases for Aboriginal families was also apparent when looking at the cases in the ‘possible’ categories: with a greater proportion of the ‘possible situational couple violence’ compared to ‘possible coercive control’ categories for Aboriginal families compared with non-Aboriginal families. Of the six ‘possible situational couple violence’ cases, five were Aboriginal families, which meant that this category made up 13% of DFV cases for Aboriginal families, but only 2.5% of cases for non-Aboriginal families. There were four Aboriginal families in the ‘possible coercive control’ category (10% of all Aboriginal family cases), compared with six non-Aboriginal family cases (15% of all non-Aboriginal family cases). A close-to-equal proportion of Aboriginal (7 out of 39) and non-Aboriginal family cases (7 out of 38) were in the ‘unclear’ category.

Discussion

This study identified that, among families impacted by DFV and involved with child protection services, there is likely to be a range of types or dynamics of DFV, as well as significant complexity. In particular, this study highlighted that although coercive control does characterise DFV in many families involved with child protection systems, it is likely that there is also a substantial proportion of families in which DFV is not characterised by coercive control and may involve mutual violence between parents/caregivers. This could mean that children do not have a non-violent parent/caregiver. It may also mean that there is a proportion of families with child protection involvement that may not benefit from the kinds of interventions commonly used to address DFV, which operate on the assumption that only one person is using violence, and that all DFV is characterised by coercive control (Ferguson et al., 2020; Lawson, 2019). These families may require interventions that work with both parents to address underlying causes of violence and/or mutual conflict. This could include addressing drug and alcohol use issues, the impacts of parents/caregivers’ own experiences of childhood abuse and trauma and/or mental health challenges, and recognising the role of life stressors, such as poverty and social disadvantage (Ferguson et al., 2020; Love et al., 2020; Stover et al., 2022).

It is not our intent to suggest that child protection practitioners should use a classification process such as the one adopted in this study to assign families into strict categories. The emergence of the ‘unclear’, ‘possible coercive control’ and ‘possible situational violence’ categories in our data suggested that some cases of DFV do not fit neatly into the category of either coercive control or situational couple violence. Rather, our hope is that this study encourages child protection services to move towards an approach that recognises the heterogeneity of DFV and the resultant complexity that is likely to be apparent in typical child protection caseloads.

To adequately respond to DFV-related child protection concerns, case-management and family support processes need to be responsive to complexity and heterogeneity rather than using a one-size-fits-all approach. It is important to note that coercive control may be missed if services working with families are not adequately skilled in recognising it, or if victims are too scared to disclose it (Humphreys & Healey, 2017). Disclosing DFV, including coercive control, may be more difficult for victims impacted by socio-economic disadvantage and Aboriginal women and children who may, for good reason, mistrust authorities such as child protection services due to the ongoing harms of unjust child removal policies (Fiolet et al., 2021).

Our findings were consistent with much prior research about coercive control and situational couple violence. Researchers have found that coercive control is likely to persist, or even increase, after separation (Johnson, 2006; Johnson et al., 2014; Katz et al., 2020). In our sample, many cases in the coercive control category involved post-separation abuse, whereas few of the cases in the situational couple violence category involved separated families. Researchers have found that men are more likely to perpetrate coercive control, whereas situational couple violence often involves violence by both men and women (Johnson, 2008; Stark, 2007). In our sample, only one case of coercive control involved a mother who was the identified perpetrator, whereas the situational violence group involved both mothers and fathers/stepfathers using violence in many cases. Importantly, however, our results supported the findings of researchers who have found that, even when both people in a heterosexual couple use violence, men are more likely than women to use severe violence (Johnson et al., 2014). This suggests that although identifying and addressing mutual violence in families with child protection involvement is important, child protection systems and practitioners should still consider gendered societal and relationship dynamics that increase the risk of harm DFV poses to women and children. Even when DFV is mutual and/or conflict based in nature, there may be power differentials, both within the couple/family and within broader societal systems, that make women and children vulnerable to harm. For example, women may be more vulnerable to financial insecurity in the context of traditional gendered expectations around work and child-care. Such gendered power dynamics can be inadvertently reinforced by child protection systems; for example, if practitioners assume that mothers should be held responsible for protecting children but do not equally hold fathers responsible (Humphreys et al., 2020). It is always important for practitioners to consider power imbalances and how power is used in relationships, and to examine how violence and abuse is used by, and/or impacts on, each family member, rather than assuming that mutual use of violence always equates to equal use of violence and/or equal risk of harm. Further, severity and patterns of violence are essential considerations. For example, practitioners should consider how often each person is using violence, whether the violence resulted in significant fear, distress or physical harm, and whether the person is able to acknowledge the impacts of their violence, take responsibility for the harm it has caused to others in the family (including children), and take steps to reduce the risk of violence and abuse continuing.

The large proportion of cases in our random extraction involving Aboriginal children was not unexpected, given the significant over-representation of Aboriginal and Torres Strait Islander children in child protection systems and the higher rates of family violence experienced by Aboriginal and Torres Strait Islander women (Higgins & Hunt, 2023; Moore et al., 2023; Morgan et al., 2022; SNAICC - National Voice for our Children, 2024). Both the over-representation of Aboriginal and Torres Strait Islander children in child protection systems, and the increased risk of harm from family violence faced by Aboriginal and Torres Strait Islander women, are linked to the ongoing harms of colonisation, including social and economic disadvantages, racism and the impacts of intergenerational trauma (Blagg et al., 2020; Day et al., 2012; Higgins & Hunt, 2023; Moore et al., 2023; Morgan et al., 2022). In our study, the finding that cases involving Aboriginal families were more likely to be classified in the situational violence category than in the coercive control category supports the work of other researchers who have argued that DFV definitions and responses that focus on coercive control may not be appropriate when working with and supporting Aboriginal and Torres Strait Islander families, and that holistic responses addressing underlying issues such as intergenerational trauma, disconnection from cultural identity and practices, social and economic disadvantage, and substance use are needed (Andrews et al., 2020; Blagg et al., 2020; Blagg et al., 2018; Day et al., 2012).

Limitations

It is a significant limitation of this paper that none of the authors is Indigenous. As noted earlier, this study was conducted as part of the first author’s doctoral research and it was not the intent of the study, nor the first author’s thesis as a whole, to focus on Aboriginal and Torres Strait Islander perspectives or experiences of DFV. Due to the nature of the project (as a doctoral research project this research did not have access to funding and was limited to the small team of the first author and two supervisors), the authors were not able to seek input from others who may have had valuable knowledge and views to contribute, such as Aboriginal researchers. However, although the large proportion of Aboriginal families in our data sample was incidental, it was not an unexpected result given the over-representation of Aboriginal and Torres Strait Islander children and families in Australian child protection systems. For the reasons noted in the previous section, including the voices and perspectives of First Nations researchers and/or participants in Australian child protection research is vital. Without the inclusion of these voices, the ability of this paper to explore issues relevant to Aboriginal and Torres Strait Islander children, women, families and communities is limited. Further, the broader body of research this study was a part of did not seek input from First Nations researchers and this has limited the overall perspective used in the research, the study design and the way data were analysed. For example, in an earlier stage of the research (Marwitz et al., 2024), we interviewed child protection practitioners and asked them to comment on a series of case vignettes. As non-Indigenous researchers, we chose not to use vignettes that specifically portrayed Indigenous families and none of the practitioner participants were Indigenous themselves. This limited the extent to which we were able to reflect on the relevance of our study for First Nations children and families. Aboriginal and Torres Strait Islander people are experts on their own lives and are best placed to understand and comment on the nature, causes and best interventions for Aboriginal children and families impacted by DFV and/or with child protection involvement. Further research by, or in collaboration with, Aboriginal and/or Torres Strait Islander researchers, centring First Nations perspectives, would be beneficial in understanding whether and how the issues raised in this study may be applicable to child protection work with Aboriginal and Torres Strait Islander children, families and communities.

Another significant limitation of this study was that intake reports, by their nature, provide only limited information and consist of allegations rather than substantiated facts. Further investigation of some of the cases coded as ‘situational couple violence’ could reveal hidden coercive control. Future research that analyses information obtained during child protection investigations would build on the findings of this study.

The confidentiality and ethics requirements of this study meant that the data were only able to be analysed by one person (the lead researcher). Conducting similar studies with other data sets that could be de-identified prior to release and therefore analysed by multiple researchers would enhance the validity of future research in this area.

Conclusion

The results of this study indicate that child protection systems and practitioners need to continue building and utilising awareness of coercive control and the impacts it has on adult victims and children. However, this should not come at the expense of acknowledging and meeting the experiences and needs of children and families impacted by situational couple violence. A substantial proportion DFV in families with child protection involvement may be situational in nature and may include mutual violence between parents/caregivers.

The key learning from the findings of this study is that child protection systems should encourage nuanced and curious assessment by practitioners working with families in which DFV is a risk factor, with a particular emphasis on recognising the heterogeneity of DFV. This should involve promoting an understanding of how coercive control and situational couple violence may impact differently on children and adult victims, and the different kinds of interventions and supports families may benefit from depending on the nature and dynamics of DFV. This learning presents opportunities for future development of practice frameworks and/or DFV-related training for child protection workers that highlights the heterogeneous nature of DFV and the differing experiences of children in families where one or both parents are using violence. At the same time, child protection systems and practitioners must maintain a focus on holding perpetrators of DFV responsible for their behaviour and managing the impacts it has on victims, including children.