Allegations of family violence and child abuse in family law children's proceedings

A pre-reform exploratory study
Research Report No. 15 – May 2007

3. Australian research on allegations of family violence and child abuse in family law

A substantial Australian conceptual literature exists with respect to the impact of family violence on parenting and property applications in family law (see, for, example, Alexander, 2002, 2006; Behrens, 1993, 1995, 2006; Graycar, 1995; Kovacs, 2002; Monahan & Young, 2006; Parkinson, 1995; Victorian Law Reform Commission [VLRC], 2006). This chapter provides a brief overview of several key Australian empirical studies of family violence and child abuse, drawing on primary data in the context of post-separation parenting disputes.

As with the previous chapter, the Australian studies are arbitrarily grouped into three distinct lines of inquiry: (a) studies that focus on allegations of family violence; (b) studies that focus on allegations of child abuse; and (c) studies that focus on allegations of family violence and child abuse.66 Some readers may prefer to skip to the summary at the end of this chapter rather than immerse themselves in the detail of diverse methodologies and findings. (Appendix A contains a summary of the key attributes of each study to help readers traverse the detail of these Australian studies.)

3.1 Empirical studies of allegations of family violence in family law

At least three empirical studies of family violence in the context of divorce have been conducted in Australia over the past decade: one study explored allegations of violence in the context of mediation (Keys Young, 1996); another examined the reports of a random sample of divorced adults (Sheehan & Smyth, 2000); a third study was primarily concerned with changes associated with the operation of the Family Law Reform Act 1995 (Rhoades, Graycar, & Harrison, 2000).

3.1.1 Family violence in the context of mediation

Keys Young (1996) were commissioned by the Australian Government Attorney-General's Department to undertake a study into the issue of family violence and the practice of mediation. The study consulted with staff in 12 of the Australian Government-funded community-based family mediation services, at least one of which was in one of the states or territories. Exit surveys were conducted with mediation clients in all of these locations. Men and women (n = 27 and n = 101 respectively) were surveyed separately.67 The surveys were designed to gather information about the prevalence of violence in cases presenting to mediation. Qualitative data were also gathered from 50 women and men who were willing to discuss their experiences of mediation. Finally, in three locations, information and referral network staff from organisations such as women's domestic violence and legal aid services were also consulted.

The key findings with respect to violence can be summarised as follows. The agencies themselves had identified almost a third of the women surveyed as clients in cases involving violence. On the other hand, almost three quarters of the women surveyed reported experiencing what they described as some type of violence or abuse. Thus, of the 65 cases in which agencies did not identify violence, 41 of the women surveyed by the researchers reported at least one form of violence directed at them or their children. Of the 27 men who responded to the survey, only 5 reported some form of violence or abuse.

The report suggested that a broad range of violence was common in families presenting for mediation and that agencies tended not to identify a considerable number of these cases. On the other hand, it found that women, in many cases identified as involving violence, reported satisfaction with the mediation processes and with the outcomes.

As Cleak, Bickerdike, and Moloney (2006) noted, the Keys Young report helped to galvanise the mediation sector into adopting more systematic and more reliable instruments and protocols to assist in the identification of family violence and to make a judgment about victims' capacity to participate effectively in family mediation. Thus, for some years, specialist intake and screening has been routinely used and indeed is required by government as part of its accreditation of family dispute resolution programs. Screening aims not only to assess the suitability of the dispute and the disputants for mediation, but to make a referral to more appropriate services when this is not the case.

3.1.2 A national survey of divorced parents

Sheehan and Smyth (2000) reported on telephone interviews with a random sample of Australian men and women who had divorced after January 1988.68 Using questions derived from the Women's Safety Survey (Australian Bureau of Statistics [ABS], 1996), the study sought to ascertain the rate at which family violence had been experienced both during and after the marriage. In order to gain some idea of the emotional and physical consequences of reported violence, questions were also asked about experiences of fear during and after the marriage and the extent to which any physical injuries had required treatment by a doctor or nurse. In addition, the study sought information on post-separation property distribution, living standards and workforce participation.

Sheehan and Smyth (2000) found that the broader the definition of violence, the higher was its reported incidence and the less was the gender differentiation. Thus, when spousal violence was defined as all actions that would be considered as an offence under criminal law, 55% of the men (n = 152) and 65% of the women (n = 244) reported its existence. When violence was defined as "actual or threatened conduct that would cause fear about wellbeing or safety", the percentages decreased to 24% for men and 53% for women.69 With respect to reports of injury that required medical treatment, the relevant percentages reduced further to 3% and 14%.

A significant finding in this study was that:

despite the extent of the financial difficulties experienced, and their ongoing responsibility for the care of children, women who report spousal violence are more likely than women who report no spousal violence to have received a minority of the share of property. (p. 118)

Further, the more severe the reported violence, the greater was the likelihood that a minority of both the domestic assets and the total marital assets would be awarded to the female partner and the greater was the likelihood that the female partner would be in difficult financial circumstances following separation. In addition, the more severe the violence, the less likely it was that the female victims would be participating in the workforce at the time the interview was conducted.

Sheehan and Smyth (2000) concluded that legally defined violence appears to be common in the divorcing population, but that an adequate understanding of its consequences requires greater specificity with respect to definitions. Fear, combined with multiple domains of violence (physical/sexual, emotional and harassment) are associated with poorer post-separation financial outcomes for women. It was suggested that these more severe forms of violence are likely to be associated with difficulties in negotiating a fair property settlement, and a situation in which concerns for safety may give way to the right to a fair share.

3.1.3 Assessing the impact of the Family Law Reform Act 1995

Rhoades et al. (2000) were especially interested in what impact the Family Law Reform Act 1995 may have had on court-ordered parenting arrangements. They reported on a content analysis of 674 interim and final Family Court judgments, the vast majority of which (n = 635) were unreported judgments collected in three batches: 1995, 1996-97 and 1998-99. Of the last batch of cases, 78 out of the 229 were in the "final judgments" category, of which 52 (67%) involved allegations of violence.

In addition to an analysis of contents and outcomes of court cases, the report by Rhoades et al. (2000) made use of a wide range of other data sources, including interviews with court personnel, solicitors, workers in women's services and litigants. Among other things, these interviews pointed to a perception that court orders were being made for increased amounts of time with non-resident parents, and that resident parents were feeling greater pressure to agree to endorsing such parenting arrangements, even in circumstances where they had safety concerns for themselves or their children of both.

Interim hearings were identified as an especially problematic area for judges who were trying to balance the claims by each parent for a continued relationship with their children against claims that wished to reduce or terminate the involvement on the grounds of alleged violence or abuse. One judge described decision-making processes at the interim stage as more like "artful dodging" than a judicial exercise (Rhoades et al., 2000, p. 76).

Rhoades et al. (2000) concluded that the Reform Act had contributed to an increase in the expectations of fathers regarding post-separation parenting arrangements with their children, and that even in the face of allegations of violence against fathers, judges were now more reluctant than they had previously been to make orders that precluded fathers' involvement. The grounds upon which these conclusions were drawn have been questioned (Moloney, 2001; Parkinson, 2001).

3.2 Empirical studies of allegations of child abuse in family law

At least three empirical studies of child abuse in the context of family law proceedings have been conducted in Australia.70 Two were conducted in South Australia.71

3.2.1 Child sexual abuse and Family Court judgments

In one of the earliest Australian studies of child sexual abuse in the context of family law proceedings, Kiel (1988) examined seven transcripts of judges' decisions in which strong corroborative evidence of child sexual abuse was provided to the Family Court. Kiel sought to assess how the Court weighed the competing interests of "the preservation of family relations on one side and the protection of children on the other" (p. 3), where child sexual abuse allegations are raised.72 She did not describe how she selected the seven cases in her sample.

Kiel (1988) identified two key themes from her analysis: (a) judicial scepticism towards allegations of child sexual abuse; and (b) a "family and welfare ideology" (p. 9), in which the "best interest" principle appeared to favour the interests of parents over children's interests. Specifically, she noted that a judge had made a formal finding of abuse in only two of the seven cases. Kiel concluded:

The Family Court has been said to rarely deny access, even in cases where there is strong evidence of abuse. This was true of the seven cases examined here. Access was denied in only one case. Normal access was granted in one case, and supervised access in five. (p. 4)

While Kiel (1988) noted that her sample was small and not necessarily representative, she suggested that the purpose of her analysis was to encourage consideration of the issue of child protection issues in family law matters.

3.2.2 Assessing the impact of spousal violence on children

In what was effectively a clinical sample, Mertin (1995) studied changes in the emotional and behavioural functioning of 34 children whose mothers had sought refuge from violent partners in a women's shelter. Just over half the children were reported to have been hit by their fathers. Shortly after they had left their partners (Time 1), Mertin found that 62% of the children were rated by their mothers (n = 27) as having borderline to severe behaviour problems. Responses to standardised measures filled out by the children themselves indicated that 18% had significantly elevated anxiety scores, while 35% had significant levels of depression. All but three of the children (the three having reported no fear of their fathers) had no contact with their fathers over the next 10 months.

Mertin (1995) found that at re-test at the end of this 10-month interval (Time 2), many of the children who had had psychological difficulties at Time 1, were now scoring in the normal range. This prompted him to speculate that many of these situations are reactive and are thus likely to improve if the source of the difficulties is removed. Mertin suggested that his findings were "consistent with" those of Church (1984). In a self-published study, Church had found similar recovery rates among children who, after their mothers had left a violent relationship, had had no contact with their fathers except in the cases in which they were unafraid of them.

Mertin (1995) concluded by suggesting that cessation of father-child contact following separation after a violent relationship is likely to be of benefit to the child. However, he acknowledged two significant design limitations. The first was the absence of a control group of children who continued contact with their fathers. The second was the lack of any data on the mothers' psychological health at Time 1 and Time 2.

3.2.3 Allegations of child sexual abuse in family law proceedings

Hume (1996) investigated 50 cases in which child sexual assault allegations had been made in proceedings in the Family Court of Australia in Adelaide. The cases involved a total of 105 children (45 male, 60 female), 70 of whom (21 male, 49 female) were alleged to have been sexually abused. Of the 50 cases, 36 contained specific allegations of sexual abuse, 11 contained allegations that a child was at risk, and 3 expressed concerns about inappropriate behaviour that indicated the possibility of sexual abuse.

Across all cases, 42% of the allegations were confirmed, 16% proved inconclusive, and 8% were found to be false. In 34% of the cases, no investigation into the allegations was made. However, in the 36 cases in which the allegations of abuse were specific, 56% were confirmed, 14% were inconclusive, 11% were found to be false, and 19% were not investigated. Hume (1996) concluded that the more specific the allegation, the greater was the possibility of confirmation. The increase in possibility was only modest, however, and may have largely reflected the fact that specificity of allegation was more likely (though by no means certain) to trigger an investigation. At the same time, a slightly greater percentage of false allegations was also found in this group.

In the 32 cases in which the alleged abuser was the father, 59% were confirmed, 13% were inconclusive and 6% were found to be false. Again, however, the "no investigation" category (30%) was relatively large. In the 27 cases in which the alleged abuser was the father, and the allegations were specific, the confirmation rate climbed to 70%, while the inconclusive outcomes and false allegation findings were both 7%. Even in this group, however, 15% were not investigated.

Hume (1996) found that for the sample as a whole, the pattern of confirmations, inconclusive findings, and false allegations were roughly similar to those reported by the state's child protective services. It was difficult to be clear if the non-investigation rate in the court group was higher. This was because the state sample contained separate figures in the "non-investigation", "partial investigation" and "not located" categories. When added together, these categories suggested a profile of non-investigations that are probably similar to that of the court. But Hume's findings that confirmation rates were greater as allegations became more specific, and considerably greater when that specificity was directed at the father, suggests that in these circumstances Family Court confirmation rates may be significantly greater than confirmation rates by state authorities.

Of the 50 cases investigated, 10 withdrew their Family Court applications and 1 had not been finalised at the time of the completion of the study. Of the remaining cases, 17 resulted in the alleged abuser having no further contact with the children, while 22 proceeded to "normal" post-separation parenting arrangements. However, this number was reduced to 8 "normal" arrangements when the alleged abuser was the father. Only 9 cases proceeded to a fully defended hearing. In 7 of these, allegations of sexual abuse were confirmed by an investigative body, though in only one case did the Court itself make a finding (on the balance of probabilities) that sexual abuse had occurred.

Hume (1996) summarised her study by suggesting that the Court, "despite [usually making] no legal determination that abuse had occurred, tended to take a child protective stance in its judgments. In only two cases, did the judgment suggest that the sexual abuse allegation had been made maliciously" (p. 18). Hume believed that her study supported the hypothesis that false allegations of child sexual abuse are no more likely to occur in the context of a family law dispute than in other circumstances.

3.3 Empirical studies of allegations of family violence, child abuse and neglect

At least five sets of empirical studies of family violence and child abuse have been conducted in Australia.

3.3.1 Allegations of family violence and child abuse in fully defended cases

The first significant study of the outcome of child-related cases in the Family Court of Australia was especially interested in the extent to which the Court's decisions reflected gendered assumptions about parenting. The Court at the time was being criticised for reinforcing gendered stereotypes in its custody decisions, the majority of which were made in favour of mothers.

The study, by Horwill and Bordow (1983), examined judicial decisions in 100 fully defended cases heard in the Melbourne and Sydney registries in 1980, approximately 5 years after the Court had commenced operations. The report made no reference to allegations of violence and child abuse. Rather, it examined the extent to which decisions were informed by gender-neutral principles - the "tender years" presumption (which traditionally favoured mothers) - and arguments based on the status quo. "Other factors" were briefly explored, but again, violence and child abuse were not included in this exploration. Indeed, it was a sign of the times in which the report was written that the authors described the sort of cases that went on to defended hearings in the Family Court in the following terms:

It should be noted that defended cases in general can be regarded as atypical because of the "selection" process involved in getting to that stage - a process affected by such things as how people respond to the conciliation process, how the parents' lawyers see the case and its prospects and the advice they give, parents' ability to pay the costs involved, their ability to cope with a lengthy time of uncertainty and conflict, and their determination to push their view as far as possible. (p. 369-370)

Bordow (1994) set out to replicate this study, drawing this time on a sample of 294 child-related judgments supplied by judges throughout Australia during 1990. Again, the words "violence" and "abuse" were not mentioned in Bordow's reporting of the results. However, the summary of findings made the following observation:

Another atypical feature of defended custody cases is the presence of a significant number of allegations of physical and emotional misconduct by one or both of the parents. (p. 262)

The word "misconduct" sits uneasily in the context of later research findings that began to identify serious violence and abuse as a significant issue in a percentage of family separations and in a percentage of child-related applications made to the Family Court. Parkinson (1995), however, suggested that Bordow (personal communication, 1994) revealed that in this same sample of cases 22% contained affidavits that alleged violence towards the wife (presumably by the former husband or partner), while 2% of cases contained allegations of violence by the wife's new partner. In addition, 7% of the cases contained allegations of child sexual abuse by the father and another 3% contained allegations of abuse (type unspecified) towards children by others, such as stepfathers or family relatives.

3.3.2 Project Magellan

Brown, Frederico, Hewitt, and Sheehan's (1998) Violence in families: Report number one is perhaps the most frequently cited study of child abuse and neglect in the context of Australian family law. The report consists of a series of conference papers (with the addition of an initial and a concluding chapter), all of which speak to research conducted in the Family Court of Australia in Melbourne and Canberra. The first paper (Chapter 2) drew initial conclusions based on a sub-sample of 30 Family Court cases, representing approximately 3% of a larger sample, which, the authors noted, were drawn from "probably 50% of the total population of case files from the years 1992-1993" (p. 15). The authors found that family violence in this sample was "widespread".

Brown et al. (1998) also suggested that there was evidence from these cases that adult-to-adult violence masked child abuse. By this, they meant that a careful reading of the files indicated a link between adult violence and child abuse that was not immediately clear when one looked only at key data such as the orders being sought. Indeed, the authors flagged a more general concern in this regard - the fact that many of the issues raised in material, such as affidavits, are not pursued in the court process. They speculated that this might link to an observation of Brown and Blenkinsop (1994), whose research in the Victorian Children's Court led them to suggest that "the Court could not deal with numbers of perpetrators or numbers of victims in one family" [italics added] (cited in Brown et al., 1998, p.18).73

In Chapter 3 of their report, Brown and her colleagues spoke to a sample of "the first 40 families studied in the Family Court" (p. 23). This presumably included the 30 cases referred to in Chapter 2. In this sample, they found that, "seventy per cent of the child abuse cases studied had spouse violence occurring" (p. 23). In the same chapter, they noted that "in the 40 cases studied initially, all involved abuse of the children as well as the female partner" [italics added] (p. 24).

It would appear, therefore, that this is a sample of cases selected for the presence of partner violence or abuse or both. Thus, the statement that "the review of the initial 40 cases showed widespread and pervasive violence occurring in families which are in dispute" (p. 26) is not unexpected. What may be more significant is the authors' observation of how multi-layered the violence appeared to be. However, this observation must be tempered by two qualifications. First, the sample size (n = 40) is again small. Second, it is not clear how the sample was drawn; this may have been a group of cases with unusually violent and abusive profiles.

Brown et al. (1998) noted that the Court had begun "to base decision-making on a definition of child abuse that saw the witnessing of parental abuse acknowledged as a form of child abuse" (p. 24). On the other hand, the study repeated the suggestion referred to in Chapter 2 that violence tended to be minimised by the Court. The example given in support of this contention was the use of the word "alleged" with respect to a case in which it was claimed that a husband had fired a bullet into his house and into the house of a neighbour. Perhaps the incident had been proven or admitted to, but this was not made clear.

Significantly, it was found that one third of these initial 40 cases was settled by consent orders that subsequently broke down. Brown et al. (1998) concluded that consent orders can mask serious violence. This finding was linked to a more general observation that final resolution of these matters seemed difficult to achieve, with cases taking between 2 years and 10 years to reach completion. Indeed, some matters were said to have resolved themselves by default only because the child became too old to continue to be part of Family Court proceedings.

The relatively small samples referred to in Chapters 2 and 3 were drawn with the aim of assisting the researchers to get a feel for the Family Court population and for court processes. Chapter 4 drew on a larger sample and addressed the question: "Who are the families involved in custody and access disputes74 where child abuse allegations occur and what are the problems or what is the nature of the abuse allegations they bring to the Court?" (p. 34).

This question was addressed by:

  • an examination of case files from the Family Court's Melbourne and Canberra registries that had become active as either new or re-activated applications between January 1994 and June 1995, and which were then followed up until July 1996
  • interviews with relevant Court staff from both registries
  • observation of pre-hearing conferences
  • interviews with state child protection staff.

The study drew on information from files from both Melbourne and Canberra. The design of the study made use of the fact that in the Melbourne Registry, cases in which child abuse is alleged or in other ways appear to be an issue are tagged by the Court. In the period in question, 373 such cases were identified. This represented 3% of the total population of files during that time and 5% of the child-related matters. The final Melbourne sample consisted of 117 of these cases. In addition, 20 defended cases taken from a total of 143 cases that had no allegations of violence and abuse were included for comparison purposes. The authors also included 10 threatened violence cases and 3 others whose characteristics were unspecified.

The Canberra Registry did not tag child abuse cases. Thus, all 102 cases in which there were disputes over custody and access during the period noted above were reviewed. From this, 38 were identified as including child abuse allegations. These then became part of the study. Thus, the total number of cases analysed was 188.

Brown et al. (1998) summarised the results of their study as follows. Family Court cases that contained allegations or other evidence of child abuse had an unexpectedly high proportion of fathers with high rates of unemployment and mothers with high rates of non-employment. The men also had high rates of criminal convictions. The women's criminal conviction rates were less high than those of the men, but compared to the general population were high nonetheless. It was suggested that "violence is fundamentally the problem that these families presented to the Court. Not only were they families in which abuse allegations had emerged, they were families in which partner-to-partner violence and child abuse had played a large part in precipitating the family relationship breakdown" (p. 64).

The authors also suggested that child abuse in parenting disputes in the Family Court was underestimated "in that while almost every case of formally identified child abuse was found to be present in those cases, it was also found to have been reported in over half of the custody and access cases which were categorised by the Court as not involving child abuse" (p. 65). In light of the substantiation rates discussed by the authors in Chapter 5 (see below), the first part of this statement is difficult to interpret. With regard to the second part of the statement, several qualifications need to be made. First, the sample of cases not involving child abuse (either 20 or 24) was a small one. Second, it is not clear what method was used to select these cases from the population of 143 cases in this category. Third, it is also unclear how the presence of child abuse was detected or defined, or what, if any, reliability checks were made with respect to the identification of these cases. On the basis of the evidence presented in their paper, the finding that most defended hearings of custody and access disputes also involved child abuse should be treated with caution.

The focus of Chapter 5 was on the relationship between the Family Court and state-based child protective services in the State of Victoria. Much of this chapter was concerned with comparing data drawn from those cases identified as involving child abuse allegations (the 3% of all files and 5% of children's matters noted above) with 10,000 closed cases from Victoria's child protective services.75

With respect to the Family Court cases, Brown et al. (1998) found that, although the child abuse group began as 5% of children's matters, by the time they reach the stage of a pre-hearing conference (judged to be at about the half-way mark towards completion in a typical defended hearing at the time), they make up about 50% of this category of cases.

Interestingly, in light of the fact that these were identified cases of child abuse allegations, Brown et al. (1998) also found that child abuse was cited as the cause of the relationship breakdown in a relatively small percentage of cases, the most common cause being cited as arguments and next most common as spousal violence. The potential for child abuse to be a catalyst for separation has been noted by others (for example, Faller & DeVoe, 1995, see Chapter 2 of the present study).

In seeking to explain this finding, Brown et al. (1998) speculated that in view of the fact that their preliminary study of 30 cases (see Chapter 2) found a high correlation between child abuse and spousal abuse, it may be that the alleged spousal abuse and spousal arguments were more prominent in their minds. They also noted that when the research staff were asked to determine if child abuse was a factor in the separation via an examination of the files, they "thought it was" in 41% of the cases. It is not clear, however, what criteria the researchers used to arrive at this conclusion.

With regard to the parents themselves, it has already been noted that both sexes had higher than normal rates of unemployment or non-employment and higher than normal rates of criminal convictions. In addition, Brown et al. (1998) reported high levels of substance abuse and partner-to-partner violence. Comparing these parents to those involved in the 10,000 child protection cases, it was found that they were more likely to be violent, more likely to be involved in substance abuse and a little more likely to suffer from psychiatric disorders.

Of potential significance also was the finding that the abuse profile for the Family Court and the protective services families appeared quite different. The two most common forms of abuse recorded by state authorities - neglect and emotional abuse - were the two least common allegations in the Family Court. Brown et al.'s (1998) suggestion that this was a largely different group of families was reinforced by the finding that 76% of the Family Court group was unknown to the state services.

The question of substantiation of allegations is, of course, at the core of many of the debates in this difficult area. Here, it must be said that Brown et al. (1998) did not add clarity. Rather, for this factor only, they switched back to their first study, in which they assessed 9% of the allegations to be false. It is not clear from that study how the assessment of "false" allegations was made.

But there is a more critical problem. The casual reader might assume that in a chapter that is dealing with a respectable number of Family Court cases (at least 117 cases and possibly has high as 373), the 9% would be drawn from a numerically meaningful sample. In fact, the percentage is drawn from the first sample which, it will be recalled, had only 30 cases.

An analysis of the child protection reports that resulted from referral of the Family Court cases to this service revealed that abuse was substantiated in 22% of cases, not substantiated in 47% of cases and judged to be not known 28% of the time. In addition, a small number of cases were referred but not investigated. It remains difficult to know how to interpret these figures. Brown et al. (1998) noted that the two organisations had quite different cultures and were likely to be dealing with overlapping but largely different populations.

The authors also cited an interesting observation by Armytage (1997). Armatyge claimed that Family Court referrals were dealt with in the same way as all other referrals to child protective services and that 50% of the cases were discarded in the initial screening. Furthermore, this process of discarding cases from the outset is not evident in the reports subsequently sent to the Court. None of the discarded cases, it is assumed, would have reports substantiating abuse.

3.3.3 Revisiting the "truth" or "falsity" of allegations

In a more recent publication, Brown (2003) revisited the theme of the truth or falsity of allegations of child abuse in the context of family law. She began by citing a large British study's findings (Cawson, Wattam, Brooker, & Kelly, 2000) that 40% of fathers and 49% of mothers were responsible for the physical abuse of their children. In this category of abuse, there was no significant difference between mothers and fathers with respect to seriousness. The victims were equally male and female, with younger children being at greater risk.

In terms of child sexual abuse, however, Brown's (2003) reading of the research suggested that "men have been shown to be the most common perpetrators ... with proportions ranging from 80% to 95%" (p. 370). Brown cited Cawson et al.'s (2000) finding that female children were 1.5 to 3 times more likely than boys to be the victims. This study also found that while the victims were most likely to be known to the perpetrator, with girls more likely than boys to be abused by a family member, the most common perpetrators were not fathers or father equivalents. Thus, while about 4% to 5% of girls reported abuse by a father figure, the abuse rate by biological fathers was between 1% and 2.8%.

Brown (2003) contextualised the issue of abuse allegations in Australian post-separation parenting disputes by reiterating the finding (Brown et al., 1998) that they comprise about 5% of child-related matters. Of these, 22% were substantiated in the study previously reported upon, but 52% were substantiated in a subsequent study by Brown and her colleagues (2001). The primary purpose of the subsequent study was to monitor the efficacy of a more proactive relationship between the Family Court and the state protective services. To this end, the study sampled and therefore reported on only serious allegations of abuse. The sampling strategy was likely to account for at least some - and possibly most - of the increased substantiation rate.

In the 2001 study, Brown and her colleagues found that in their sample of family law disputants, in which there were serious allegations of child abuse, mothers were twice as likely to make allegations as fathers. But the allegations made by mothers were four times as likely to be substantiated as those made by fathers.76 Of a total of 11 allegations found to be false, 6 were made by fathers and 5 by mothers. Of the 52 cases of substantiated abuse, 32 involved the father as abuser and 4 involved the mother. This left 16 cases of substantiated abuse perpetrated by another family member. A little under half (n = 23) of the 52 substantiated abuse cases were in the category of sexual abuse, with all but one of the remainder of cases being equally divided between physical abuse and emotional abuse. There was one case of neglect.

Brown (2003) set out to explore the stereotypes portrayed on the one hand by those who assert the prevalence of the "parental alienation syndrome" (referred to in Chapter 1 of the present study), and on the other hand by those who see fathers as "perpetrators of patriarchal power and control through the use of family violence such as child abuse" (p. 378). The sample size of 100 cases reported in Brown et al. (2001) reduced to numbers that would again suggest caution in making predictive statements. But combined with her reading of related research, Brown (2003) concluded as follows:

It is clear that allegations are made against fathers more frequently than against any other family member and that the person making the allegations is most commonly the mother. Furthermore, many fathers are not found to be the perpetrators as alleged. This gives some support to the belief of fathers that their former partners pursue them with malicious allegations of child abuse. However, fathers are the ones found to be the most common perpetrator of the abuse alleged in this context; allegations against fathers were found to be substantiated in just over half the cases. False allegations are few. They are not the prerogative of mothers but are almost equally initiated by fathers and mothers. At the same time, the uncertain conclusions from so many of the child protection investigations that leave fathers as the most frequently unsubstantiated perpetrators encourage their ongoing anger and frustration. Thus the stereotype presented in the Parental Alienation Syndrome is not supported ... but the feelings and views of fathers as expressed by individual men's groups are partially supported by the research. (p. 378)

3.3.4 Court administrative data

As part of its submissions to the Inquiry into Joint Custody Arrangements in the Event of Family Separation, the Family Court of Australia (2003) examined allegations of family violence and child abuse in a random sample of cases from its case file management information system. It extracted a sample of cases for the period January through June 2003 from the Court's three largest registries: Sydney, Melbourne and Brisbane. The final sample comprised three sub-groups: 450 "consent applications" (parenting orders made by consent), 300 "settled applications" (applications for final parenting orders that were settled prior to a trial) and 91 "judicially determined matters" (parenting matters that required a judicial determination).

The Family Court of Australia (2003) noted that allegations of physical abuse (or the risk of physical abuse) were made in a "high percentage of cases, particularly those which went to judgment. In that sample [91], physical violence was an issue in more than two thirds (67%) of the judgments" (p. 17).77 In 59% of these cases, "the Court either made a finding upholding the allegation (51%) or found that there was an unacceptable risk of family violence (a further 8%)" (p. 18). No finding was made in 30% of the cases; the allegation was not upheld in another 10% of cases.

Allegations of child sexual abuse (or the risk of child sexual abuse) were made in around one quarter (26%) of the judgments sampled. There was found to be an unacceptable risk of child sexual abuse in 38% of these cases. A finding of no unacceptable risk was made in 29% of the cases, while no finding was made in another 29% of cases.78

3.3.5 Allegations of family violence in fully litigated cases

Kaspiew (2005a)79 examined a randomised sample of 40 Family Court files involving children's matters that were fully litigated in the Melbourne Registry of the Court in 1999 and 2000. The focus of her analysis was on the content of affidavit material, psychosocial assessments and judicial determinations.

Kaspiew (2005a) reported that "violence was a factor" (p. 51) in 23 of the 40 cases (57.5%). In most cases, it remained a "background" factor, but in 10 cases it was judged to be highly significant. In three of the cases, fathers had received criminal convictions for violence against former partners. And domestic violence orders were mentioned in "more than half" the files.

Allegations of physical and emotional abuse of children were made in 16 cases and of sexual abuse in three cases. Issues relating to mental illness were raised in a total of 18 cases. Substance abuse was highly significant in two cases and a "background" factor in a further 10.

A key claim resulting from this study was that the fathers in the sample were seeking bigger roles in their children's lives in the context of family histories that often involved violence, entrenched conflict, mental illnesses and substance addiction. Kaspiew (2005a) also suggested that although the circumstances of many of the cases prevented the fathers from realising these objectives, judges remained strongly committed to maintaining father-child relationships.

3.3.6 Separated mothers' experiences of an abusive former partner

Kaye, Stubbs, and Tolmie (2003) interviewed 40 mothers, recruited largely from women's refuges and women's health services, who were negotiating and facilitating parenting arrangements with an abusive ex-partner. They found that 80% had obtained a violence restraining order at some point, with only one finding it to be a fully effective measure. Kaye and her colleagues also interviewed 22 individuals who were professionally involved in family law, domestic violence and children's contact services.

The mothers in this study reported multiple forms of abuse, with 85% of the sample disclosing serious physical and/or sexual abuse. They reported that almost two-thirds of the children had witnessed physical violence against their mothers and that 33% of children had themselves experienced physical violence. All but one reported a continuation of violence after separation. Kaye et al. (2003) found the professionals were divided over whether contact should go ahead when the non-resident parent had abused the children - most assuming that violence against mothers was separable from consideration of the wellbeing of the child.

Even though most of the mothers did not endorse father-child contact because of safety concerns, all of the fathers had obtained contact orders, and only four mothers reported that contact was to be supervised. The most common means of making contact and residence arrangements were reported to be through consent orders (44%), court orders (37%) and private negotiation (19%). In 13 families where the children had been direct targets of physical violence by their father, six had unsupervised contact, fathers had residence in four cases, contact was supervised in two cases and one case involved phone contact only. Mothers who had denied contact when there were orders or an agreement in place referred to circumstances such as child illness, child safety fears, child distress at contact or child allegations of sexual abuse during contact by the parent or their associates.

3.3.7 Allegations of violence in contested contact disputes

Shea Hart (2004) reported on all judgments that related to disputed parenting cases in the Adelaide Registry of the Family Court of Australia between 1991 and 2001. The time interval divided itself equally into a period prior to and a period following the Family Law Reform Act 1995 (Cth), which was introduced in Australia in June 1996. The total number of contested child-related cases prior to the Reform Act was 338; the number after the Reform Act was 447, an increase of 32%. Of the total number in this 10-year period, 399 were contested contact disputes, 159 of which took place prior to and 240 of which occurred after the Reform Act. This represented an increase in contact disputes of 51%. In 183 of these cases (46%), domestic violence was referred to as "an issue" by the judge - 55 pre-reform and 128 post-reform. This represented a pre- to post-reform increase of 123% in cases in which domestic violence was seen as an issue.

Shea Hart (2004) then examined contested contact disputes in judgments that had been delivered and in which a finding of domestic violence had been made - 43 cases pre-reform and 66 post-reform. Interestingly, both represent 27% of the total applications pre- and post-reform; however, the absolute number of applications post-reform had increased by 53%. In 36 (84%) of the pre-reform and in 52 (79%) of the post-reform cases in which a finding had been made, the father was referred to by the judge as the perpetrator of at least one act of domestic violence. This represented 23% of the pre-reform applications and 22% of the post-reform applications for contact. Domestic violence restraining orders had been in place in just over half the cases, both pre- and post-reform, in which family violence had been found to be present. More than half of these were alleged to have been breached.

The study also found a significant concurrence of allegations of child abuse with findings of domestic violence - 26 (60%) pre-reform and 48 (73%) post-reform. This represented 16% of the total applications for contact pre-reform and 20% post-reform. Of the pre-reform allegations, 23 were against the person (usually the father) seeking contact; post-reform, the number was 39. A core finding was that the considerable increase in applications for contact that occurred post-reform was associated with roughly the same proportion of orders for contact being made, despite findings of "at least one instance" of domestic violence in the majority of cases. Thus, the absolute number of contact orders in cases in which there had been a finding of at least one instance of domestic violence had increased significantly following the reforms.

Similar to the findings of Brown and her colleagues cited above, this study pointed to the resource-intensive nature of these cases. Indeed, 17 (40%) of the pre-reform cases and 35 (53%) of the post-reform cases involved multiple (between 2 and 11) applications and multiple determinations. The length of these cases ranged from 2 to 8 years, with most using a considerable range of services, such as counselling, mediation, Family Reports and separate legal representation, and a range of clinical services for the children. There was a general (though not universal) tendency for more of these services to be employed post-reform and for the average time of litigation to increase.

Shea Hart (2004) discussed her findings in the context of a distinction between conflict and violence, which she saw as qualitatively different. She noted that:

From a feminist perspective, reconstructing violence as conflict is serious as it masks unique power relations and ongoing fear and risk factors inherent in cases where there is, or has been violence. It also implies that couples are mutually involved and equally responsible. (p. 182)

Shea Hart (2004) also endorsed Tolman's (1992) view that:

Domestic violence involves a power imbalance where the perpetrator maintains intimidation and fear in the adult and child victims and this differs significantly in nature and effect from interpersonal conflict where both parents engage in the dispute with a reasonably equal power relationship. (p. 183)

From this perspective, Shea Hart (2004) discusses the dangers of too readily assuming a "culture of cooperation and negotiation" following separation, and of paying insufficient attention to safety issues at the expense of promoting an ongoing relationship with both parents. She points to a range of studies that address the impact of family violence on children and protective factors, such as the findings by Mullender et al. (2002) of the efficacy of support from siblings and mothers. She also noted that there was a dearth of research on the nature of the relationship between a child and a father who has perpetrated violence against the child's mother.

Shea Hart (2004) called for more sophisticated analyses of allegations of violence in family law cases and an increase in resources to allow this to happen. She suggested that every case of alleged violence requires "thorough and early investigation to establish any possible adverse effects on the child ... before gambling with the child's wellbeing through making orders for contact prior to a thorough specialised assessment of the child's situation" (p. 189).

We return to this important issue in the final chapter. At the same time, we note that Shea Hart's (2004) own position - that couples cannot be "mutually involved and equally responsible" for violence - appears to negate a number of the research findings in this area discussed in Chapter 1.

3.3.8 Resident mothers' views of post-separation parenting arrangements

McInnes (2006) used focus groups to assist initially in the shaping of a survey of separated resident mothers' attitudes to child-father parenting arrangements (known in Australia at the time as "contact"). She drew her sample from networks of the National Council of Single Mothers and their Children and e-lists for women in South Australia. Acceptance into the focus groups was determined by an affirmative response to one or both the following questions:

  • Have you ever been involved in family court proceedings with respect to children's matters?
  • Have you ever feared for your own or your children's safety from your ex-partner? (p. 21)

McInnes noted that the above questions were:

aimed at identifying a past or current conflict between the parents. [Thus,] mothers who answered affirmatively to one or both of these questions were scheduled to attend a separate focus group from mothers who answered both questions negatively to ensure that low-conflict cases were not overwhelmed by cases involving conflict. (p. 21)

This strategy resulted in six intending participants being scheduled to attend what the author referred to as the "no conflict" group. At the scheduled time, however, only one of the six individuals attended. By contrast, 16 women responded by saying they had been involved in child-related family court proceedings and/or had at some time feared for their own or their children's safety from their ex-partner. They were assigned to a "conflict" focus group and, at the scheduled time, 14 showed up.

It is not clear how many of this group had made applications to the Family Court. Selected citations from the transcripts would suggest that most, if not all, were in this category. What is clear is that all of the 14 women in the focus group had at some time had cause to be fearful of their former partners on their own behalf or on behalf of their children. It would seem, therefore, that most, if not all, of the women sampled answered both the trigger questions in the affirmative.

The key themes emerging from this focus group were the difficulties mothers reported in being able to guarantee safe, quality child-father contact. In addition, police, child protection and family law services, such as mediation and counselling, were frequently seen as ineffective. Participants felt that mediation and counselling service staff were often uninformed or withdrew services due to the men's violence. Some were labelled as "crazy" by their abusers and found it was the abusers who ended up with greater credibility in the family law decision-making system.

Mothers were also concerned that children's needs and interests were not central and that quality of parenting care was typically not examined. Rather, these mothers believed that a willingness to support child-father contact was the focus of the family law intervention system.

McInnes (2006) made use of the focus group material to inform the development of a set of survey questions. The survey attracted 170 valid responses. Of these, 100 were analysed. The remainder of the data are to be analysed at a future date.

According to McInnes (2006), the resident mothers in the survey reported strong support for fathers maintaining contact with their children after separation, with mothers wanting more father-child contact outnumbering those who were seeking less contact. McInnes also reported, however, that "most respondents were unhappy with the contact pattern, but nevertheless enabled contact to take place" (p. 65). Such findings suggest a complex set of dynamics that would probably require further teasing out in individual interviews.

Experiencing fear at the time of separation, experiencing continuing fear of the father and having contact determined by the court system were the most common factors linked to mothers wanting children to have less time with their fathers. Yet mothers who reported being afraid of their ex-partner and afraid for their children's safety suggested that they were nevertheless compliant with court orders and agreements in most instances. When mothers did stop contact, children's safety and health were the main reasons and most stopped contact only once or twice. According to the mothers, fathers' desires to see their children and their availability for contact had a much greater impact on whether contact took place, with twice as many fathers as mothers cancelling or not attending planned contact.

McInnes (2006) summarised her findings by suggesting that they highlight "the exposure of children to long-term continuing harm through court-ordered contact with parents who abuse or terrorise them" (p. 65). An obvious limitation of the research in this regard is that it reported on perceptions of the mothers only.

3.4 Summary

The Australian empirical literature generally echoes that of overseas studies, both in its findings and in the many gaps and the many methodological problems that are evident. Clearly, researchers have been struggling to pin down issues that are of considerable complexity, that have generally not begun to come to the fore until some years after the Family Court began operations and that have contested explanations that, at their extremes, reflect polarised constructions of gender.

There is little doubt that various forms of violence exist in a considerable proportion of separating families and that this was generally not appreciated when the idea of "no fault divorce" was being promoted in the 1970s.

It is also probably true to suggest, as do Brown and her colleagues (1998), that violence and abuse increasingly become "core business" as cases move further and further into family court processes. Indeed, it is almost certain, as Shea Hart (2004) also found, that a hugely disproportionate percentage of court resources are taken up with litigation by couples - for whom violence is quite often accompanied by unemployment or underemployment, mental health issues, and drug and alcohol abuse, which have been part of their "normal" lives.

The Australian evidence around child abuse, especially sexual abuse, remains sketchy. Brown's (2003) summary suggests that we have a long way to go in unpicking questions about frequency and what is meant by sexual abuse in the context of family law; the fact that allegations may be made in good faith does not mean that they are necessarily "true". Parents and step-parents remain vulnerable to such claims, and both they and the children can suffer great anguish if claims are later shown to be unsubstantiated, or even if the verdict is one that signals continuing uncertainty. This is where the principle of safety for children can collide with benefits that can derive from an ongoing relationship between children and both parents.

For all its methodological difficulties, the Australian research reviewed above does tell a story of a gradually increasing recognition that family law litigation over children has been tapping into a hitherto largely unrecognised (or at least under-recognised) problem. Compare, for example, the reports of Horwill and Bordow in 1983, and Bordow ten years later, with the increasingly strongly worded comments from researchers from about 1995 onwards. Why did violence and abuse virtually go unmentioned in the 1983 study, and why in the 1994 study was it labelled "misconduct" and probably underestimated in terms of its prevalence? It is possible, though unlikely, that this was a sampling issue. The more likely answer is that we tend to see only that for which we are searching.

The evidence about family violence in family law disputes has been hard won. It is therefore understandable that in seeking to bring this to the attention of sceptical professionals and decision makers in the context of family law, some researchers have been more concerned with getting the stories out there than with careful sampling or careful definitions.

At the same time, what most of the studies fail to address adequately is the range of expressions of violence and abuse. The stories tend to be of coercive control by men, perhaps because some theorists and some research samples are mainly tapping the experiences of women and children in refuges and perhaps because even in more broadly based qualitative research, these are the stories that are difficult to ignore.

Having highlighted such stories, the discussion sections in these studies tend to move from citing relatively large percentages of family violence to making overt or covert presumptions that all of this violence is of this type. They tend not to clearly identify how family violence was defined in the first place, and they rarely concede the possibility that at least some of the violence may be situational, one-off, reciprocated, or even at times initiated by women.

In sum, there appears to be a strange disjuncture between the research and debates on definitions and causes of family violence in the general literature reviewed in Chapter 1, and the family law-oriented research reviewed in both this and the preceding chapter. Perhaps this is a developmental issue. Perhaps there is also something about the urgency of family law that places pressure on researchers, when in doubt, to err on the side of safety at all costs.

For there can be little doubt that for some Australian families, separation is a time of danger. There can also be little doubt that for some children, spending time alone with a parent who has a history of violence and abuse puts them at risk. In decision-making about post-separation parenting, there is indeed a sense in which safety has to be the core concern. In an unsafe atmosphere, relationships inevitably grow into very twisted shapes that are likely, in turn, to impact on parent-child relationships in the next generation.

Thus, the next real challenge in this very difficult area is not to continue to attempt to convince policy makers, decision makers, mediators and others that family violence and abuse of children is a reality. The next challenge is to understand the nature of this violence and abuse, to recognise that it is not all the same, and that it requires a range of responses - from total exclusion of the perpetrator, to placing protective structures around parents and children, to various forms of rehabilitation, right through to a recognition that the violence or abuse was a single instance that is most unlikely to happen again.

Shea Hart (2004) has appealed for better and more sophisticated forensic resources that can support the decision makers by clarifying the nature of the violence and abuse that occurs in separating families. This may indeed prove to be a necessary way to break the seeming impasse that has resulted from the largely gender-based criticisms of family courts both in Australia and elsewhere. In the meantime, questions addressed in the present study - such as: What precisely is being alleged by whom? What evidence is there to support the allegations? and What responses are made by the alleged perpetrators? - offer a step forward in becoming clearer in this difficult area.

Footnotes

66 Johnson's (2002) Western Australian study of post-separation familicide is also relevant and has been noted as part of the broader analysis of violence in Chapter 1.

67 The response rate was 47% for women but only 26% for men.

68 Western Australia was excluded from this survey.

69 This figure for women is very similar to Hunter et al.'s (2000) finding. After examining 176 Family Court files, Hunter and her colleagues found that 54% (n = 95) "contained evidence of domestic violence".

70 Bromfield (2005) also investigated the prevalence, course and predictors of chronic child maltreatment in a sample of 100 child protection cases in regional Victoria; 38 of the 100 cases involved allegations made in the context of a custody dispute. It is noteworthy that 8% of the 100 cases in the sample involved allegations that were classified by Bromfield as "malicious".

71 Relevant studies (e.g., Cashmore, Chisholm, & Waters, 1992; Young, 1998) drawing on secondary data from courts with family law jurisdiction are not included in this review. It is nonetheless worth noting that figures for the 1993 calendar year provided by the Family Court of Western Australia to Young suggested to her that the rate of child sexual abuse allegations seemed to be "somewhere in the range of 1.2% of all cases involving parenting disputes" (p. 103). Young noted that this estimate was consistent with the findings of Cashmore et al. (1992), who wrote that "there is as yet little Australian data on these questions, but what is available does not support the notion that child sexual assault allegations are common in the resolution of custody and access disputes: a study by the Family Court Counselling Service found that in a 3-month period only 1.6% of all new cases registered with the Court Counselling Service involved allegations of child sexual assaults" (p. 32).

72 Kiel (1988) also analysed the Family Court's submission to the 1985 New South Wales Child Sexual Assault Taskforce.

73 The suggestion that court processes may struggle to deal with multiple allegations and counter allegations, and that legal strategists may at times advise that only one or two principal allegations be focused on (a "less is more" approach), seems to be worth further investigation.

74 This was the language of the day.

75 It is not clear if the total number (n = 373) were compared, or if the child protection data were compared with the Family Court sub-sample (n = 117) described in the previous chapter.

76 The absolute numbers here are relatively small: 7 fathers and 32 mothers.

77 In its 2004-2005 Family Violence Strategy report (FCoA, 2005, p. 3), the Family Court of Australia suggested that this analysis "showed that violence was a factor in 68 out of the 91matters judicially determined during that period"; that is, 75% of the judgments contained an allegation. The reason for this small disparity between reports is unclear. This latter estimate almost mirrors that produced by the random sample of judgments for the entire year in 2003 in the present study (see Chapter 5).

78 This would appear to amount to 9 cases (10%) of unacceptable risk, 7 cases (8%) of no unacceptable risk and 7 cases (8%) in which there was no finding.

79 See Kaspiew (2005b) for a published summary of this research.