Allegations of family violence and child abuse in family law children's proceedings
- 1. Family violence, child abuse and family law
- 2. International research on allegations of family violence and child abuse in family law
- The roads ahead - "walking the talk"
- 3. Australian research on allegations of family violence and child abuse in family law
- 4. Methodology
- 5. The prevalence and nature of allegations
- 6. Evidentiary material and responses to allegations
- 7. Parenting proposals and court outcomes
- 8. Discussion and implications
- Appendix A: Summary tables of research literature
- Appendix B: Coding frame for the three-category typology of family violence and child abuse176
- Appendix C: Case summary for each category in the typology
- Appendix D: Classification of allegations relating to physical abuse, sexual abuse and emotional/verbal abuse
- Appendix E: Classification of evidentiary material
- Appendix F: Type of family violence or child abuse: All cases
2. International research on allegations of family violence and child abuse in family law
This chapter provides a brief overview of some of the key international studies of allegations of family violence and child abuse in the context of post-separation parenting disputes. Australian studies are summarised in the next chapter.
These studies are arbitrarily grouped into three distinct lines of inquiry: (a) those that focus on allegations of family violence; (b) those that focus on allegations of child abuse; and (c) those that focus on allegations of family violence, child abuse and child neglect as potentially overlapping forms of parental and interpersonal transgressions. (Appendix A contains a summary of the key attributes of each study for each line of inquiry.)
Though the review is not exhaustive, it provides a useful empirical context in which to locate the present study's methodology and findings. Of course, the extent to which the findings from overseas studies readily transfer to the Australian context cannot be certain. Some readers might prefer to read only the summary at the end of the chapter rather than immerse themselves in the detail of divergent methodologies and findings.
2.1 Empirical studies of family violence in family law
Studies of family violence in the context of custody/access disputes have been conducted in the US, Canada and the United Kingdom.23 Some of the key studies from these countries are now briefly described.
2.1.1 Empirical studies of family violence in the United States of America
In the US context, Newmark, Harrell, and Salem (1995) explored the sorts of challenges facing courts and court-connected services when family violence is raised in the context of court-ordered mediation for child custody and access disputes. A total of 422 separated parents from 293 families were recruited from mediation services in Oregon and Minnesota in early 1993. Parents were surveyed about whether the other parent had ever engaged in different types of abusive behaviour against them (such as "intimidation", "physical abuse", "severe abuse" and/or "use of a weapon"), and, if so, how often this abuse had occurred in the 6 months (a) prior to attending the mediation service, and (b) prior to the survey. Perceptions of empowerment in dealing with the other parent and the court system and the risk of harm by the other parent were also assessed.
Newark et al. (1995) found that "a great deal of abuse was reported" (p. 57).24 Specifically, 80% of women and 72% of the men in the sample reported experiencing some form of spousal abuse. While "intimidation" was reported as the most frequent and ongoing type of abuse, 55% of men and 68% of women reported being physically abused (being slapped, grabbed, pushed, kicked and punched to physically hurt), and 38% of the women and 20% of the men reported "severe abuse" (being beaten or choked). Those who reported the experience of abuse (both women and men) were more likely than those who did not report abuse to perceive that their former partner controlled decision-making in their relationship and that their former partner might physically harm them in the following 6 months. Most physical abuse was reported by women and men as having occurred more than 6 months prior to entering the mediation or evaluation services.
Newark et al. (1995) also found that women reported being abused more often than men, and that women who reported abuse felt a diminished sense of empowerment than women who did not report experiencing abuse. No such difference emerged for men.
On the basis of these data, Newark et al. (1995) suggested that family court programs develop specialised responses for dealing with allegations of family violence, and made six recommendations in relation to the need for specialised training, screening protocols, information, resources, practical support and a (shuttle) mediation process. More recently, Logan, Walker, Jordan, and Horvath (2002) conducted an exploratory study of court records to compare cases with and without family violence, paying particular attention to custody evaluation reports and recommendations, and case outcomes. A total of 135 cases containing at least one custody evaluation or child representative report were sampled from a Kentucky Circuit Court. Using a standardised coding protocol, two psychology students analysed the content of a random sample of 82 cases. Logan et al. found that 56% of the cases involved family violence, while 8% involved reciprocal violence between males and females.25
Cases classified as involving family violence were more likely than other cases to mention child abuse and substance abuse in the court file; the files were also more likely to mention the involvement of younger children, marital debts and parents returning to court for child support matters. By contrast, only minor differences emerged between the violence and non-violence cases in relation to the custody evaluations and recommendations. On the basis of these findings, Logan et al. (2002) suggested that custody evaluators do not seem to look into the nature or extent of family violence or to explore it as a way of addressing the safety of children. They speculated that evaluators may be overly focused on "parental rights" in the custody evaluation process (p. 737).
Logan, Walker, Horvath, and Leukefeld (2003) sought to examine, among other things, the characteristics of post-separation parenting arrangements, paying particular attention to whether these arrangements differed for divorcing couples who raised allegations of spousal violence. They conducted a content analysis of a random sample of 20% of divorce cases settled in Kentucky in the 1997/1998 financial year.26 Logan et al. found that 20% of the cases within this sample (n = 258) and 33% of the cases with children mentioned spousal violence. More specifically, they found that these cases were no more likely to require a judicial determination than other cases, with most cases in the sample settled by agreement through lawyers or self-representation. However, they also found that cases involving allegations of spousal violence were more likely than cases without violence to note substance abuse and post-decree activity in relation to child support. The cases with violence allegations also typically involved shorter marriages, shorter periods from filing to divorce, and more structured parent-child contact (or did not stipulate contact). It is noteworthy that the authors did not formally define "spousal violence" or mention whether the alleged violence was substantiated or remained an allegation.27
2.1.2 Empirical studies of family violence in Canada
In Canada, Shaffer and Bala (2003) sought to examine the circumstances in which, and under what terms, abusive husbands were being allowed parent-child contact by the court. They also wanted to assess the extent to which judges believed claims by women of "wife abuse". They conducted a computerised search of children's matters reported in the Canadian family law database Quicklaw, which produced 42 cases in which wife abuse was likely to have been a consideration in a custody or access dispute.28 As noted by Shaffer and Bala, cases in this database cover legal decisions with written reasons by judges and typically represent contentious, complex or novel legal arguments. These cases nonetheless provide an opportunity to access judicial thinking and decision-making.
Shafer and Bala (2003) concluded that a broad array of abusive male behaviour (pre- and post-separation) was evident in the cases they examined; covering such behaviours as enduring severe physical and sexual violence, some physical abuse in tandem with other types of abusive and controlling behaviour, and primarily emotional and verbal abuse. The authors suggested that many of the men denied or minimised their violent behaviour, which the court rejected when "physical violence was extreme or well documented" (p. 259); less extreme violence, however, tended to be seen as mutual by the court.
Shafer and Bala (2003) found that:
the general picture in terms of custody in the cases we surveyed was that abusive men were not succeeding in getting either sole or joint custody, provided that the court accepted the allegations of abuse. ... [But] the situation is quite different when dealing with access. Most of the men found to have abused their wives were granted access rights to their children, usually on an unsupervised basis. (pp. 263-64)
They also noted that the court found women's allegations to be exaggerated or unsubstantiated in 11 of the 42 cases:
In some of these cases, the court gave no reasons for concluding that the women had fabricated or embellished their claims, making the validity of these judicial decisions impossible to assess. It is possible that the courts were correct and the claims had been fabricated or exaggerated. It is also possible, however, that judges failed to recognize abuse because it was not well documented or because the abuse took a predominantly emotional, rather than physical, form.
In the absence of medical records, police reports or witnesses to the abuse, judges may have difficulty finding that abuse occurred, and in fact, judges mention this as a problem in some of the cases. In part this may be a function of the law's requirement that the person making an allegation of abuse prove it true on the balance of probabilities. ... Often the only adult witnesses to spousal abuse are the spouses. While children often see or hear spousal abuse, there are a variety of evidentiary and ethical concerns about calling them as witnesses. Many women do not disclose their abuse, report it to doctors, or call the police; therefore, it may be difficult to prove abuse in many cases. Absent evidence corroborating the women's allegations, judges may be reluctant to find that abuse has been proven to the court's satisfaction. Judges may decide the allegations are "unfounded" simply because the woman cannot muster sufficient evidence that it has occurred.
Judges may also conclude that allegations are unfounded where women raise allegations of emotional or verbal abuse as involving significant physical violence. If the abuse is primarily emotional or verbal, courts may have difficulty conceptualizing the conduct as abusive, viewing it instead as mutual conflict or discord. (pp. 259-60)
Shaffer and Bala (2003)29 also reflected:
Some lawyers may advise at least some of their women clients who claim to have been victims of spousal abuse not to raise abuse concerns in their court documents. Some lawyers do this because they consider spousal abuse irrelevant to custody and access issues if there is no direct abuse of the children. In other cases, there is a concern that if the issue of spousal abuse is raised but not proven, that this will reflect poorly on the mother's claim for custody, suggesting that she is an "unfriendly parent" who will thwart the father's access to the children for selfish reasons. (p. 269)
Shaffer and Bala (2003) expressed concern at what they concluded to be an apparent lack of judicial sensitivity to abuse in some of the cases in their sample, and called for the development of more refined judicial responses for different configurations and levels of abuse so that children are not put at risk when in the care of a parent.
2.1.3 Empirical studies of family violence in the United Kingdom
In the UK, Humphreys and Thiara (2003) surveyed 161 women about the women's experiences of post-separation violence. Women were recruited by domestic violence project outreach workers with the Women's Aid Federation. Half of the women recruited had lived in a women's refuge at some point, while the other half of women had only made use of the outreach service.30 Humphreys and Thiara also conducted face-to-face in-depth interviews with 20 women who had not completed the survey questionnaire. Both samples were purposive.31
The researchers found that 76% of the women surveyed reported further abuse and harassment from their former partners after separation. They also noted that for many women, the violence and abuse stopped after 6-12 months; sometimes because women relocated, in other instances, because the abuse stopped. Nonetheless, 36% of women reported ongoing post-separation violence. All 20 of the women interviewed referred to an escalating pattern of violence and coercive control, characteristic of Johnson's category of "intimate terrorism" (Humphreys & Thiara, 2003).
While verbal and emotional abuse (76%), and serious physical threats (41%), were reported more frequently than physical and sexual assaults (23% and 6%), Humphreys and Thiara (2003) found that women often experienced threats, verbal and emotional abuse as "terrifying", and "some women found this harder to cope with than actual physical violence" (p. 201).
Humphreys and Thiara (2003) noted that:
69 per cent [of the women in the purposive sample] feared for their emotional wellbeing; 66 per cent said that the abuse was becoming worse; 60 per cent of women feared that they would be killed; 60 per cent feared for their mental health; 54 per cent left [the relationship] when they could see the abuse affecting their children; 25 per cent said they feared for their children's lives; 25 per cent said they feared they would kill their abuser. (p. 200)
Humphreys and Thiara (2003) also found that women reported diverse experiences of post-separation violence and found it difficult to achieve separation from an abusive former partner. They suggested that a lack of effective action by the legal system often put women and children at serious risk of harm, with father-child contact being a particular point of vulnerability for ongoing post-separation violence and abuse.
2.2 Empirical studies of allegations of child abuse in family law
We begin with an examination of some of the early North American clinical studies of "false allegations" of child sexual abuse.32 Larger empirical studies from North America and the United Kingdom are then examined.33
2.2.1 Early clinical studies: North America
Much of the early work into allegations of abuse in the context of parental separation was conducted in North America during the 1980s. These small-scale clinical studies - drawing on data from the private practices of psychiatrists who conducted clinical assessments for litigants, the court or child welfare agencies - focused primarily on allegations of child sexual abuse in custody disputes.34 These studies suggested that such allegations were on the rise, were largely false, and were made by mothers against fathers (Faller & DeVoe, 1995; McGraw & Smith, 1992; Thoennes & Tjaden, 1990).35
For instance, Kaplan and Kaplan (1981) reported on a single case from their clinical practice (with a passing reference to one other similar case). The case was referred to them by a family court judge who wanted to determine whether allegations of sexual abuse made by a child in a custody dispute were "true". Although no definitive assessment was offered, Kaplan and Kaplan implied that the allegations were unfounded, suggesting that the allegations stemmed from a folie à deux, in which several family members shared similar delusional beliefs.
In another frequently cited study, Benedek and Schetky (1985) examined 18 cases of alleged incest (14 of which were related to custody disputes after divorce)36 and concluded that 10 of these cases (55%) were based on "false" allegations. They suggested that although false allegations of sexual abuse by children and their parents are rare, such allegations were particularly likely to occur in custody disputes. They speculated that some parents might make false allegations "to obtain sole custody, to terminate visitation, to terminate parental rights, or to harass a non-custodial parent" (p. 156).
Green (1986) reported on 11 cases from his private practice in which the children, in the context of a custody or access dispute, claimed that they had been sexually assaulted by their non-resident father. Green concluded that four of these cases (36%) involved false allegations of abuse, which, he suggested, mirrored Benedek and Schetky's (1985) "strikingly high" rate of false allegations (p. 449).37 (Green also suggested that false denials by children were "common", whereas false allegations by children were rare (p. 451).
Moreover, Schuman (1986) described seven cases from his clinical practice in which child physical and sexual abuse were alleged. He concluded that "all of the claims of abuse were ultimately shown to be nonvalid" (p. 6), and suggested that "domestic relations cases are unfortunately fertile ground for nonvalid perceptions and/or allegations of misconduct of all forms" (p. 19).
However, as several researchers (such as Faller, 2003) have noted, small-scale clinical studies are likely to be limited in several ways. First, they typically have little generalisability because of their small selective samples.38 The examination of low base-rate conditions, such as child sexual abuse, requires large samples for statistical power; clinical case studies are not designed to estimate prevalence rates (Bala et al., 1998; McIntosh & Prinz, 1993; Thoennes & Tjaden, 1990). Second, case studies might reflect a clinician's own views about, and theoretical orientation towards, abuse, which can shape what has been described and concluded ("confirmation bias"). This is especially likely in the emotionally charged area of custody proceedings, in which there is pressure to reach a decision one way or the other.39 An apparent gender bias against mothers in this early work has not gone unnoticed (see Corwin, Berliner, Goodman, Goodwin, & White, 1987). Third, clinicians in these early studies tended to categorise cases idiosyncratically rather than rely on empirically validated and consensually agreed observations (Corwin et al., 1987; Paradise, Rostain, & Nathanson, 1988).
More problematic, perhaps, is that these early clinical studies employed a simple binary classification system to assess allegations ("true" or "false"), in which unsubstantiated allegations were treated as "false". This is understandable, given that a key aim of these studies was to try to identify clinical criteria that could distinguish between "true" and "false" reports (Thoennes & Tjaden, 1990). But, as several researchers have noted, allegations can remain unsubstantiated for many reasons, and there remains much confusion about definitional issues (Fahn, 1991; Myers, 1990; Penfold, 1995; Robin, 1991; Trocmé & Bala, 2005).40
For instance, Awad (1987) recommended that allegations be assessed as either "probably true", "probably not true", or indeterminate, because it may be impossible in many instances to determine the veracity of an allegation with any certainty (see also Faller & DeVoe, 1995). Likewise, Penfold (1995) suggested that allegations be classified into one of three types: substantiated reports (variously termed "true", "found", "proved" or "confirmed"), unsubstantiated (or "unfounded", "unproven" or "insufficient information"), and false (or "fictitious", "erroneous" or "manufactured").
More recently, Trocmé and Bala (2005) emphasised the importance of distinguishing unsubstantiated investigations from "intentionally false" reports:
Most unsubstantiated investigations are the result of well intentioned reports triggered by a suspicious injury or concerning behavior or a misunderstood story. Mandatory reporting laws require the reporting of reasonably suspected child abuse or neglect, and do not expect reporters to conduct their own investigations prior to reporting.
In contrast to unsubstantiated allegations, intentionally false allegations are intentional fabrications that are made in the hope of manipulating the legal system, or made to seek revenge against an estranged former partner, or may be the product of the emotional disturbance of the reporter. If there is a deliberate fabrication made, it is important to distinguish between cases in which a parent or other adult who is taking the lead in the fabricating from those where it is the child who is fabricating the allegation without adult influence.
It is also important to distinguish allegations that are clearly unsubstantiated or false, from those where abuse cannot be substantiated but remains suspected [italics in original]. (p. 1335)41
Accordingly, Trocmé and Bala (2005) recently adopted a four-category classification scheme in which allegations were classified as "substantiated", "suspected", "unsubstantiated, [in] good faith", or "intentionally false".42 They suggested that confusion often arises in the interpretation of allegation statistics because of a lack of conceptual or definitional clarity. For instance, Thoennes and Pearson (1988) defined "false" allegations as those "offered in good faith, but where, for a variety of reasons, the abuse was unlikely" (p. 23), whereas Trocmé and Bala would define these as "unsubstantiated, good faith".43 The issue of conceptual clarity is also germane when examining larger empirical studies of child abuse, to which we now turn.
2.2.2 Empirical studies of child abuse
From the late 1980s, a number of large-scale international empirical studies have examined allegations of child abuse in the context of "child custody" and "access disputes.44 Most of these studies focused on false allegations of child sexual abuse.
For clarity, these studies are grouped into three strands: (a) rigorous clinical studies in the US; (b) studies based on data from child protection services; and (c) studies using data from domestic relations courts (that is, courts with family law jurisdiction).
220.127.116.11 Clinical studies of child abuse allegations in the US
In one of the earliest, more rigorous clinical studies of child abuse allegations in the context of divorce, Jones and Seig (1988) reviewed the files of 20 cases of child abuse evaluated at the Kempe Center for the Prevention and Treatment of Child Abuse and Neglect, based in Denver, Colorado.45 All cases at the Center were screened for alleged sexual abuse and the presence of a custody/access dispute, and the first 20 cases that met these criteria - working backwards in time from 1985 to 1983 - were extracted from the Center's files and analysed.46 The allegations in 14 of the 20 cases (70%) were deemed "reliable" as evaluated by a clinical team, 4 cases (20%) appeared to be "fictitious", 1 (5%) was deemed "unsubstantiated suspicion", and the allegations in the other case (5%) remained uncertain. Jones and Seig noted that mothers alleged abuse in 14 of the 20 cases (70%), fathers made the report in 3 cases (15%), and children in 3 cases (15%). Three of the four cases believed to be fictitious were instigated by the mothers. On the basis of their findings and those of earlier studies (such as Benedek & Schetky, 1985; Green, 1986), Jones and Seig concluded that custody and access disputes seemed to increase the likelihood of clinicians finding fictitious allegations. It is important to note that a comparison group was not included in the study's design.
McGraw and Smith (1992) re-examined 18 cases of allegations of child sexual abuse made in the context of contested custody disputes that were investigated by the Boulder County Sexual Abuse Team in Colorado. These cases were selected from a larger pool of 290 cases investigated in 1987, most of which did not involve a post-divorce custody dispute. In the original 1987 analysis, only 1 of the 18 cases (6%) was assessed as being founded (94% were unfounded). But on re-assessment of the cases using a more comprehensive validation process, McGraw and Smith reported that the number of cases determined to be founded increased to 8 (44%); 10 cases (56%) remained unfounded, of which 5 cases (28%) were classified as "unsubstantiated suspicion", 2 (11%) were deemed to contain insufficient information, 2 (11%) were believed to be fictitious adult reports, and 1 (5.5%) a fictitious report by a child. Fictitious reports of child sexual abuse in the context of custody disputes thus totalled 17% in this small clinical sample.47 McGraw and Smith suggested that this estimate was slightly lower than the 20% estimate of "false" allegations reported by Jones and Seig (1988), and was much lower than the 36% figure reported by Green (1986), and the 55% figure of unsubstantiated reports found by Benedek and Schetky (1985). But all of these studies appear to mean "unsubstantiated" when they use the term "false".
Faller and DeVoe (1995) sought to estimate the prevalence of false allegations of child sexual abuse in the context of divorce. They examined a clinical sample comprising 215 divorce cases involving allegations of sexual abuse. These cases were drawn from 15 years of practice at a multi-disciplinary university-based clinic in the mid-west of USA. Most of the referrals to the clinic had come from child welfare services. This study remains one of the most thorough clinical studies conducted in the area of divorce and false allegations of abuse.
Faller and DeVoe (1995) found that most allegations of sexual abuse were made against biological fathers (69% of named offenders), followed by step-parents and biological mothers (9% and 8% respectively), followed by other relatives, non-relatives and others (6%, 4% and 3%). Based on the clinical judgment of the multi-disciplinary team, alleged abuse was substantiated in 73% of cases, 20% of cases were classified as "unlikely", and 7% as "uncertain". By contrast, the domestic relations court substantiated 35% of these cases, did not substantiate 44% of cases, made no mention of the alleged sexual abuse in 14% of cases, and was "unclear" in 7% (for example, because of a vague court order, no decision by the court, or a lack of information).
Faller and DeVoe (1995) suggested that there were four causal pathways between divorce and allegations of child sexual abuse: (a) divorce follows the discovery of abuse (14% of cases); (b) divorce occurs before the discovery of abuse that has been occurring during the marriage (25%); (c) abuse begins to occur during or after parental separation (27%); and (d) allegations of abuse are false (14%) or potentially false (7%), with 5% classified as "knowingly made false"48 This low rate of deliberately false allegations, as noted by Faller and DeVoe, accords with other studies (such as Thoennes & Tjaden, 1990, see below). It is noteworthy that Faller and DeVoe found that the substantiation rate by the court of allegations of child sexual abuse was around half the clinical substantiation rate.
In summary, these three clinical studies of alleged child sexual abuse produced mixed findings as to whether "fictitious" allegations disproportionately occur in child custody disputes: Jones and Seig (1988) suggested that custody disputes may increase the likelihood of fictitious allegations, McGraw and Smith (1992) found that allegations of child sexual abuse were "only slightly more likely" to be determined fictitious in the context of custody disputes, while Faller and DeVoe (1995) found only a small number of deliberately false allegations.
These studies point to the difficulty of establishing the veracity of allegations of child sexual abuse. It is also important to note that these samples may not be representative of the population of separating parents involved in parenting disputes, given that they were derived from referrals to specialist centres.
18.104.22.168 Studies of alleged child abuse based on child protection service data
Three studies of child protection investigations involving post-separation custody disputes also warrant brief mention: one study was conducted in England, the other two in Canada.
In England, Anthony and Watkeys (1991) examined all referrals of suspected child sexual abuse received by child health, social services and police between 1986 and 1989 in the South Wales area. Sexual abuse was proven in 197 of the 350 referrals (56%) that were fully investigated by a multidisciplinary team, 44% were deemed to be "unsubstantiated", 18% "false", while 9% of allegations were deemed to be "false and malicious".49
Twenty-four of the children referred were also the subject of a custody dispute by their parents. Only 5 of these cases (21%) were substantiated. Anthony and Watkeys (1991) concluded that the aim of the deliberately false allegations in the 19 unsubstantiated cases was to prevent parent-child contact from occurring, and to make life more difficult (mostly for the fathers). They noted that these allegations were often followed by counter-accusations by fathers against mothers' new cohabiting partners or associates.
In Canada, Hlady and Gunter (1990) sought to determine the frequency of custody and access disputes within the caseload of a child protection unit in the British Columbia Children's Hospital. The unit assessed 370 children in 1988, 41 (11%) of whom were involved a custody/access dispute. Sexual abuse was alleged in 34 of these cases (83%), for which there was physical evidence in 6 cases; physical abuse was alleged in seven cases (17%), for which there were corroborating findings in 5 cases.
Hlady and Gunter (1990) found that child sexual abuse was substantiated at roughly the same rate for the custody/access group and the group where custody and access was not an issue (18% and 15%). This was in contrast to allegations of physical abuse, where such allegations were more likely to be substantiated in the custody/access group than in the non-custody dispute group (71% vs 48%). The finding requires caution, however, because the ‘custody' sample involved only 7 allegations of physical abuse (of which 5 were substantiated). They noted that allegations were frequently vague, and that children's behavioural problems were typically the trigger for a report in the absence of physical findings. Hlady and Gunter concluded that, although false allegations of child sexual abuse were believed to be a common concern in divorce proceedings, their data did not support this concern.
More recently, two large specialised data collections that examine the incidence and characteristics of alleged child maltreatment have been developed in Canada: (a) the 1993 Ontario Incidence Study of Reported Child Abuse and Neglect (OIS-93) (Trocmé, McPhee, Tam, & Hay, 1994), which was based on surveys completed by child protection personnel in Ontario for a representative sample of 2447 children, and was replicated in 1998 and 2003; and (b) the 1998 Canadian Incidence Study of Reported Child Abuse and Neglect (CIS-98), which expanded the scope and coverage of the Ontario study.
Drawing on data from the OIS-93 (Trocmé et al., 1994), Bala et al. (2001) noted that 46% of the investigations involved children from families in which the parents had separated or divorced.50 Resident mothers made two-thirds of the allegations (with child sexual abuse being the most common allegation made by mothers - 13% of all investigated sexual abuse allegations), while non-resident fathers raised a third of the allegations (see Table 2 in Bala et al., 2001).51
Of the allegations made by resident mothers against non-resident fathers, 23% were deemed to be substantiated, 27% were suspected, 50% were unfounded, while only 1% of the allegations were considered to be malicious (that is, deliberately false). There was a police investigation in almost one third of these cases, with criminal charges being laid against the suspected non-resident father in 8% of the cases. By contrast, of the allegations raised by non-resident fathers against resident mothers, only 10% were deemed to be substantiated, 18% were suspected, 72% were unfounded, while 21% were considered to be malicious. There was a police investigation in 10% of these cases, with no criminal charges being laid against any of the resident mothers (again, see Table 2 in Bala et al., 2001).
Commenting on the OIS-93 data, Bala and Schuman (1999) noted that, while resident mothers made more than twice as many allegations of child sexual abuse against non-resident fathers than fathers did against mothers, non-resident fathers' allegations were less likely to have been substantiated and more likely to have been considered as deliberately false than the converse.
Drawing on national data from the CIS-98, Trocmé and Bala (2005) examined data from 7,672 child maltreatment investigations reported to a national random sample of Canadian child welfare services in late 1998. They sought to estimate the prevalence of deliberately false allegations of abuse and neglect in the context of post-separation custody and access disputes. Using a standard set of definitions, child protection workers in the sample reported the details of maltreatment incidents and the results of their investigations. Their assessment involved a clinical judgment as to whether the reports were substantiated, suspected, unsubstantiated (in good faith), or were made maliciously (that is, were deliberately false). Their attempt to identify "intentionally false" reports within a large national random sample is a unique feature of the study.
In line with other national studies of child maltreatment, data from the Canadian Incidence Study indicated that only 4% of all cases were considered to be deliberately false (42% were substantiated; 23% were suspected; and 31% were unsubstantiated but made in good faith). This contrasts with cases in which there had been a custody or access dispute: 12% of these cases were considered to have been intentionally fabricated (40% were substantiated; 14% were suspected; and 34% were unsubstantiated in good faith).
More specifically, Trocmé and Bala (2005) found, among other things, that: (a) deliberately false accusations of child neglect were far more common than false allegations of abuse; (b) non-resident parents (mostly fathers) were more likely than resident parents (mostly mothers) to make false reports, which runs counter to the view put forward by some that false allegations are widespread and made by mothers against fathers; and (c) children rarely made deliberately false allegations of sexual abuse.
Specifically, they concluded that:
noncustodial parents were responsible for 43% of intentionally false reports in cases involving custody or access disputes, with relatives, neighbors, or aquaintances accounting for another 19% of these cases. Custodial parents (14%) and children (2%) were responsible for relatively few intentionally false allegations in cases arising in a context in which there was an ongoing dispute over custody and access. (p. 1341)
Trocmé and Bala (2005) suggested that, although intentionally fabricated reports are an important issue (particularly in relation to custody disputes), of greater concern is that such allegations may overshadow unsubstantiated allegations raised in good faith and unresolved reports where suspicions of abuse and neglect remain, both of which were found to be far more prevalent.
22.214.171.124 Studies of child abuse based on family court data
At least three significant studies of allegations of child sexual (and physical) abuse in the context of family law proceedings have been conducted: two in the US, and one in Canada.
In what continues to be one of the largest focused empirical studies to date, Thoennes and Tjaden (1990) examined the incidence, type and apparent veracity of allegations of child sexual abuse in the context of custody and access disputes. They analysed data from 169 custody/access disputes that involved an allegation of child sexual abuse. These data were collected across a six-month period by mediators and custody evaluators based at domestic relations courts in jurisdictions in 12 US states.52 Staff completed a summary form each time they encountered a case in which an allegation of child sexual abuse was raised. These data were then augmented by a detailed review of family court files (including any custody evaluation and court orders), and information from child protective service agency files (for example, Was the case reported? Was there an investigation? What was found?). Cases for over 9,000 families in dispute about custody or access were screened for allegations of sexual abuse.
Thoennes and Tjaden (1990) found that just under 2% of custody/access disputes involved an allegation of child sexual abuse (ranging from 1-8% across different jurisdictions).53 Mothers accused fathers of the abuse in just under half (48%) of these cases, or accused step-fathers in 6% of cases. Fathers accused mothers or a mothers' new partner in 6% and 10% of cases respectively. Almost 20% of cases involved allegations by mothers (13%) or fathers (6%) against other family friends or relatives; a third party alleged abuse in another 11% of the cases. Four per cent of the cases involved a counter-accusation, in which the original allegation was denied by the alleged perpetrator, who in turn alleged the other parent perpetrated the abuse.
Based on the assessment of either a custody evaluator, child protection worker or both in the 129 cases where assessment data were available, Thoennes and Tjaden (1990) found that sexual abuse was believed to have occurred in half the cases, no abuse was thought to have occurred in one third of cases, and the remaining 17% of cases were indeterminate. More specifically, where mothers accused fathers of abuse, Thoennes and Tjaden found that half the cases were perceived by the professionals to be likely, one third were deemed unlikely, and 18% were indeterminate; in the less common situation where fathers accused mothers, 42% of cases were deemed "likely", 41% as "unlikely", and 17% could not be determined.54
Thoennes and Tjaden concluded that:
allegations of sexual abuse among families in dispute over custody and visitation are no more likely to be determined false than are allegations of child sexual abuse in the general population. Further, mothers are no more likely than fathers to make false allegations. (p. 161)
It is important to note that these researchers were equating allegations that were perceived to be unlikely with allegations that were perceived to be false - something that probably reflects the under-developed conceptual frames operating in the late 1980s.
Amid growing concern in the US that allegations of child sexual abuse were commonplace in child custody/access disputes, McIntosh and Prinz (1993) sought to determine the extent to which this was the case by analysing the entire one-year caseload of a county family court. Specifically, they reviewed the files of 603 applications in 1987 for divorce by parents (98% of the sample) or cases involving a post-divorce child custody/access dispute (2% of the sample); 14% of the overall sample involved a custody/access dispute.
McIntosh and Prinz (1993) found that allegations of child abuse were relatively rare. Allegations of child physical abuse were made in only 2% of all cases (6% of custody dispute cases), and sexual abuse allegations were raised in less than 1% of all cases (2% of the custody dispute cases). They noted that in the five cases where sexual abuse allegations had been made, not one of the final court judgments addressed the veracity of these allegations. It is noteworthy that this study included independent coding of a random sub-sample of 34 cases, as well as a review by the entire team of the relatively small number of cases where abuse was identified, as a reliability check of the coding.
In Canada, Bala and Schuman (1999) reviewed family law judges' written decisions in 196 judgments between 1990 and 1998, where allegations of child (physical or sexual) abuse had been in the context of parental separation. Of these cases, 23% were substantiated, 45% were unfounded and 35% contained evidence of abuse, but the judge did not make a finding that abuse had occurred. According to Bala and Schuman, the judge believed a deliberately false allegation had been made in 45 of the 150 cases (30%) where abuse was unsubstantiated. Most (71%) of the allegations raised in the 196 judgments examined were made by mothers, compared with only 17% by fathers and 2% by grandparents or foster parents. Children were found to instigate allegations in only 9% of the cases under review. Thus, fathers were most likely to be accused of abuse (74%), followed by mothers (13%), stepfathers or a mother's boyfriend (7%), then a grandparent (3%), followed by other kin (including siblings) (3%) (Bala & Schuman, 1999). The researchers were quick to point out, however, that their findings may not generalise to all cases where allegations are made in family law because perpetrators are unlikely to contest allegations of abuse where there is strong evidence against them. In addition, the Quicklaw sample of judgments is likely to be biased towards judgments involving contentious, complex and novel legal arguments.
2.3 Empirical studies of allegations of family violence, child abuse and neglect
At least three significant empirical studies of allegations of family violence, child abuse, child neglect and their potential overlap, have been conducted in North America and England.
The California Administrative Office of the Courts collects state-wide longitudinal data on a range of family law issues, including the extent to which concerns about abuse are raised in court-mandated mediation and custody investigations. These data relate to over 18,000 child custody cases and are analysed by the AOC's Center for Families, Children and the Courts. A baseline survey was conducted in 1991 by the Center, with subsequent waves of data collected in 1993, 1996 and 1999.
In its Snapshot Study (Center for Families, Children and the Courts, 1993), the Center found that concerns about family violence, child abuse and substance abuse were frequently raised by either or both parents in mediation sessions. Specifically, family violence was raised in 39% of all sessions, co-existing with other problems in 31% of sessions. Substance abuse problems were raised almost as frequently as family violence (38% of sessions), followed by child neglect (30%), child physical abuse (18%), child sexual abuse (8%) and child abduction (6%). Concerns about other criminal activities were raised in 8% of sessions, typically in tandem with other forms of alleged parental misconduct.
In 1999, based on reports from 2,500 separating mothers and fathers, family violence and substance abuse continued to remain the two most common issues raised in a case (29% and 25%, respectively) (California Administrative Office of the Courts, 2003). More broadly, in 55% of cases, at least one issue related to parental misconduct was raised.55 In just over three quarters (76%) of court-based custody mediation cases, at least one indicator of prior interparental violence was raised by a parent (California Administrative Office of the Courts, 2003, p. 3). The Center also noted that "even in cases with a history of relatively severe acts of domestic violence or restraining orders, the parents surveyed often did not raise issues of violence before or during mediation" (California Administrative Office of the Courts, 2003, p. 9).
In addition, mediators reported that mothers and fathers raised different types of issues about each other: family violence, substance abuse and harassment were raised as issues by mothers, whereas fathers tended to raise issues about child neglect and psychological disorders.
In Florida, Sorensen et al. (1995) examined the relationship between allegations of spousal violence, child abuse, child neglect and substance abuse on the one hand, and custody awards on the other.56 They examined data reviewed by child representatives in 60 contested custody cases requiring judicial determination between September 1988 and October 1990.
Sorensen et al. (1995) found that allegations of parental misconduct occurred frequently in contested child custody cases: 83% of the cases contained at least one allegation, 70% of cases contained multiple allegations and 45% of the cases involved mutual allegations. Allegations of spousal emotional abuse were most common, whereas allegations of child sexual abuse were least common. They also found that allegations against mothers and stepfathers were more likely to be about the "previous abuse" of children (18% of cases) and recent child neglect (35%), more so than similar allegations against fathers and stepmothers. Conversely, allegations against fathers and stepmothers most often related to recent spousal abuse (physical and emotional) and child sexual abuse (35%, 42% and 17% respectively).
According to Sorensen et al. (1995), 30% of the allegations raised were substantiated; allegations of prior child abuse, child physical abuse and child neglect were substantiated most frequently (50%-43%), compared with only 20% of allegations of spousal physical abuse, alcohol abuse and child emotional abuse.57 Allegations of child sexual abuse were substantiated the least (7%). Sorenson et al. noted how this estimate was markedly smaller than the 50% substantiation rate found by Thoennes and Tjaden (1990) in their study of custody cases and the study conducted by Jones and McGraw (1987) that examined reports to child welfare services. Sorenson et al. speculated that parents facing strong corroborating evidence of abuse or neglect might "fear negative responses from the court, and thus be more inclined to settle through mediation, or in family court, prior to dissolution of marriage" (p. 259). Sorenson et al. noted that they were unable to distinguish unsubstantiated allegations made in good faith from deliberately false allegations.
One of the key findings of the study was that judges seemed to be sensitive to allegations of abuse when making custody awards, despite the lack of corroborating evidence. Sorenson et al. (1995) suggested that:
judges do not require sufficient evidence for substantiation of the abuse/neglect allegations to consider such information relevant. The researchers believe this finding may indicate the presence of judicial sensitivity to the frequent absence of corroborative evidence of child abuse/neglect or spousal abuse. Another possible interpretation of this finding is that judges are more concerned with children's welfare than with parental rights. ... [Conversely] it is possible that judges are not willing to severely limit one parent's function vis-a-vis child care based on unsubstantiated allegations. (p. 258-259)
More recently, Smart, May, Wade, and Furniss (2003) investigated residence and contact disputes between litigating parents in three county courts in different parts of England. They obtained a random sample of 430 cases relating to residence and contact disputes under the Children Act 1989.58 These cases went to court in the year 2000. They also conducted an in-depth content analysis of 281 files in order to perform a more fine-grained analysis of the nature of post-separation parenting disputes, including allegations of family violence and child abuse.
Smart et al. (2003) found that 22% of cases contained an allegation of physical or emotional violence59 and 6% of cases contained allegations of child sexual abuse. They noted that the cases involving allegations of family violence could be classified into three broad types: (a) cases where mothers had applied for residence, typically because the mother had to leave the home without the children or because the father had taken (or had threatened to take) the children from the mother's care (one third of cases involving allegations of family violence); (b) cases where fathers were seeking in-person contact with children and perceived contact was being minimised or obstructed, but ongoing violence had been flagged as the reason for the cessation of contact (another third of cases); and (c) cases where both parents (or sometimes others) sought residence. Cases involving allegations of child sexual abuse by one of the parents unfolded in three main ways: (a) the case dissipated when the alleged perpetrator withdrew; (b) the court found that there was insufficient evidence to support the allegations; or (c) the allegations became the centre of prolonged investigations.
A key insight from the study was that where children had witnessed family violence or were the subject of abuse, this appeared to impact on judicial decision-making. Where there was little or inconclusive evidence, allegations of violence did not appear to influence residence or parent-child contact. Smart et al. (2003) suggested that it was rare for face-to-face parent-child contact to ever be denied by the courts. Even where the court thought that children might be at risk, indirect contact could still be ordered.
In the US, Johnston et al. (2005) investigated allegations of child abuse, neglect and family violence among 120 divorced families referred for child custody evaluations or custody counselling and the extent to which these allegations were substantiated.60 The families in the study, according to Johnston et al., were at the high end of the conflict spectrum, had been unable to resolve their custody disputes after mandatory mediation and had been referred by family courts in the San Francisco Bay Area between 1989 and 2002.
Following coding of the documentary records61 of these cases by experienced clinicians working independently, Johnston et al. (2005) found that at least one allegation of abuse was raised against mothers in 56% of families and against fathers in 77% of the families in the sample (and substantiated in 34% and 57% respectively). More specifically, allegations of child physical/verbal abuse were made against mothers in 15% of families and against fathers in 21% of families (and substantiated in 9% and 6% respectively); allegations of child sexual abuse were made against mothers in 6% of families and against fathers in 23% of families (and substantiated in 3% and 6% respectively); and allegations of family violence were made against mothers in 30% of families and against fathers in 55% of families (and substantiated in 15% and 41% respectively).
More broadly, Johnston et al. (2005) found that just over half of the allegations of abuse, neglect and family violence were substantiated in some way (52% against mothers, 51% against fathers); around one fourth (24%) of mutual allegations were also substantiated. Johnston and her colleagues interpreted these findings to suggest that mothers are "no more likely to allege unsubstantiated abuse against their children's other parent than are men" (p. 290); a finding that they believe offers little support for Gardner's (1999) "parental alienation syndrome" thesis that mothers were more likely than fathers to make false allegations. In line with larger national incidence studies in the US context, both mothers and fathers were also found to be equally responsible for child physical abuse and neglect.
Some gender differences did nonetheless emerge. Johnston et al. (2005) found that mothers were more likely to allege child sexual abuse and drug abuse against fathers than the reverse, although neither types of allegations were substantiated at a relatively high rate (unlike mothers' allegations against fathers of alcohol abuse and spousal violence). Johnston et al. suggested that the latter finding supports feminist approaches that point to the gendered nature of family violence.
It is important to note some of the methodological limitations of the study, many of which Johnston and her colleagues (2005) noted themselves. First, the findings are based on a moderately sized, non-random sample from a single US state, and thus may not generalise to other populations of divorced parents in dispute about child custody - either in the US or elsewhere. Second, despite care to define and standardise "substantiations" of abuse, other researchers might believe this coding to be under- or over-inclusive. Third, Johnston et al.'s definition of domestic violence included any act of physical aggression or coercive control such as the use of physical restraint, force, or threats of force by one parent to compel the other parent to do something against his or her will. However, the range and severity of violence (and other forms of abuse) were not rated. So, pushing and slapping, for instance, was not distinguished from murder or the use of a weapon. Fourth, "substantiations of abuse" were coded on the basis of any corroborating evidence - except where this had been dismissed as entirely unfounded by the court - and where the abuse was deemed by the researchers to have been "severe enough" to have affected parenting and to have warranted consideration in custody decision-making (p. 288). Fifth, no attempt was made to distinguish intentionally false allegations from other forms of unsubstantiated allegations, or to distinguish alleged parent perpetrators from other alleged perpetrators.
At the same time, this work is one of the most comprehensive studies of abuse conducted to date. It has extended much of the earlier work by including spousal violence, substance abuse and child neglect in its analytic purview, not just child abuse. It also has expansive coverage in relation to documentary evidence, in that it examined concerns raised in mediation, custody counselling and custody evaluations, not just allegations raised in formal complaints to the court.
It seems fair to say that the international research literature focusing on allegations of family violence and child abuse in the context of post-separation parenting disputes is somewhat of a methodological potpourri: key concepts and terms are often used loosely, differentially or not defined at all; foci, aims, sampling strategies (and representativeness) and methodologies vary considerably across studies; the findings and pattern of results are often complex; and there appears to be little (if any) replication or regard for building on the methodological approaches of prior work. Put simply, the international literature looks to comprise - at best - a disparate collection of partially overlapping investigations, with little convergence among the various lines of inquiry.
The complexity and disparate nature of this literature makes it difficult to distil or articulate key findings into a definitive list. Nonetheless, several findings can be used as foundation stones in the absence of solid convergence between studies.
First, allegations of child sexual abuse in the context of custody disputes appear to occur relatively infrequently; the frequency of such allegations ranges from 2% (McIntosh & Prinz, 1993; Thoennes & Tjaden, 1990) to 6% (Smart, May, Wade, & Furniss, 2005). However, higher rates of allegations have been cited in studies with smaller samples that include high-conflict parents. In Johnston et al.'s (2005) study, child sexual abuse was alleged against mothers in 6% of families, while similar allegations were made against fathers in 23% of the 120 high-conflict families in that study. In Sorenson et al.'s (1995) study of 60 contested custody cases, 15% of cases involved allegations of child sexual abuse.
Second, though small, the incidence of child sexual abuse appears to be greater in the cases where the custody of children is in dispute than the incidence of such abuse reported in the general population (Thoennes & Tjaden, 1990).62
Third, allegations of child sexual abuse considered to be "deliberately false" (that is, "malicious" or "fictitious") appear to be relatively uncommon in some studies, but elevated in other studies among certain subgroups - most notably, allegations made by non-resident fathers. Faller and DeVoe (1995), for example, classified 5% of cases in their clinical sample as "knowingly made false", and Trocmé et al. (1994) classified 1% of allegations made by resident mothers against non-resident fathers as "malicious", compared with 21% of allegations made by non-resident fathers against resident mothers. More recently, Trocmé and Bala (2005) found that only 4% of all child protection cases in Canada in 1998 were considered to be deliberately false. This contrasts with cases in which there had been a custody or access dispute, of which 12% of those cases were considered to have been intentionally fabricated. Within the latter cases, non-resident parents (mostly fathers) were judged as being more likely than resident parents (mostly mothers) to make deliberately false reports of child sexual abuse (43% compared with 14%). Commenting on the Canadian data reported by Trocmé et al. (1994), Bala and Schuman (1999) suggested that even though mothers were more likely than fathers to make allegations of child sexual abuse, mothers' allegations were more likely than fathers' allegations to be substantiated.
Fourth, allegations of family violence appear to be more widespread than allegations of child abuse, ranging from 22% to 83% of custody/access dispute cases (Smart et al., 2005; Sorenson et al., 1995), with other estimates falling between these upper and lower bounds (Johnston et al., 2005; Logan et al., 2002; Logan et al., 2003).63 That said, allegations of family violence and child abuse in the context of custody disputes can also co-occur. Logan et al. (2002), for example, found that cases involving family violence were more likely to have child abuse issues noted than cases in which family violence was not flagged.64 Johnston et al. noted that "the more intransigent conflict-ridden divorcing families are likely to be troubled by multiple indicators of domestic violence, child neglect, molestation and abuse, parental substance abuse, mental health problems, and child abduction" (p. 291). High-end conflict can sometimes be part and parcel of a complex amalgam of interlocking psychosocial problems.
Fifth, several studies have noted that mothers and fathers often make different allegations: mothers are more likely to make allegations of family violence, child sexual abuse and substance abuse against fathers than fathers against mothers, with higher rates of substantiation in relation to family violence alleged against fathers than alleged against mothers (Johnston et al., 2005). By contrast, fathers appear to be more likely to allege child neglect and psychological problems against mothers than other types of allegations (California Administrative Office of the Courts, 2003). Nonetheless, mothers and fathers have been found by Johnston et al. to be equally likely to be responsible for child physical abuse and neglect. It is also noteworthy that two-sided allegations were raised in around half (49%) of the 120 families in Johnston et al.'s sample, and in 45% of cases in Sorensen et al.'s (1995) sample of 60 contested custody cases.
Sixth, allegations of violence have been found to influence applications for the custody (residence) of children (such that abusive fathers often do not gain the full-time care of children), but it seems that family courts rarely disallow parent-child contact (Shaffer & Bala, 2003; Smart et al., 2003). A potent predictor of case outcome in disputed parenting matters - at least in the UK - seems to be whether children witnessed any violence between their parents or experienced any violence or abuse themselves (Smart et al., 2005).
One final observation: As might be expected, the veracity of allegations is far more likely to be assessed in clinical or child protection settings than in family court jurisdictions. This is likely to reflect the different mandates and principles of evidence operating within each setting. As noted in Chapter 1, the role of the Family Court of Australia:
is not to determine whether an alleged perpetrator ... is guilty of any offence, nor necessarily to make a finding as to its occurrence. Rather, the High Court has made it clear that the Court must, in such circumstances, protect the child from an unacceptable risk. (FCoA, 2003, p. 9)65
In the North American context, it is noteworthy that Faller and DeVoe (1995) found that the substantiation rate by court of allegations of child sexual abuse was around half the clinical substantiation rate. It is also worth noting that no attempt was made by Sorensen et al. (1995) or Johnston et al. (2005) to assess the veracity of allegations made in the context of family law proceedings. While assessing the veracity of allegations is difficult even where a plethora of clinical and collateral information is available, attempting to do so from the confines of formal court documents is to walk on shaky ground. While a judge may address the veracity of an allegation in setting out her or his written reasons (see, for example, Bala and Schuman, 1999), the international literature suggests that the court may sometimes look away from issues related to intimate partner violence or child abuse allegations (McIntosh & Prinz, 1993; Shaffer & Bala, 2003; Smart et al., 2003).
23 Small-scale qualitative studies have been excluded from this review, such as those that investigated the experiences of: battered women in refuges (e.g., Radford, Hester, Humphries, & Woodfield, 1997), mothers who shared the care of children with abusive partners (Shalansky, Eriksen, & Henderson, 1999), abused mothers whose children attended supervised contact centres (e.g., Shepard, 1992), mothers who had left abusive relationships but continued to support father-child contact while trying to protect the children (Varcoe & Irwin, 2004). This is not to suggest that these studies of women's experience of violence are unimportant.
24 The authors appeared to use domestic violence and abuse interchangeably. While no formal definition of either was offered, "abuse" appeared to be defined by the occurrence of any of the four types of abuse assessed. These four domains were adapted from the Conflict Tactics Scale (Straus, 1979).
25 In this study, the presence of a Domestic Violence Order (DVO) was used to classify the case as a "domestic violence case". In this jurisdiction, a DVO is issued "after a court hearing with both parties present and findings of fact and law that support issuance of the order" (p. 724).
26 Cases sealed by the court were excluded from this analysis.
27 The researchers write: "Spousal violence information in this study includes both allegations (e.g., in pleadings and petitions) and findings (e.g., when a protective order is in evidence)" (p. 271).
28 These cases spanned the period January 1997 through May 2000. The original search actually produced 45 cases, 3 of which were excluded because they involved allegations of abuse against a mother or a stepfather.
29 See also Brown and Blenkinsop's (1994) comment (cited in Brown et al. (1998), Chapter 3) on the difficulty some courts have in processing complex information.
30 One quarter of the sample were from black and minority ethnic backgrounds.
31 That is, participants were not selected randomly but were selected on the basis of being typical of the category of cases under investigation.
32 Although these studies are now somewhat anachronistic, they nonetheless continue to be cited by some advocacy groups as evidence of the high rate of "false" allegations of child abuse in family law matters.
33 See also the excellent summaries provided by Bala et al. (2001) and Faller (2003). There are also clinical studies with medium-sized samples, but which suffer from many of the problems of the smaller studies (such as Gunter et al., 2000; Paradise, Rostain & Nathanson, 1988; Wakefield & Underwager, 1990).
34 In the context of researching family violence, Gelles (1990) noted that clinical studies typically have provided a frequent source of data, since psychiatrists, psychologists and other mental health professionals have the most direct access to cases. The same appears to hold for allegations of child sexual abuse, which is perhaps why clinical studies appeared at the vanguard of the early "research" into false allegations (and false denials) of child abuse in the context of divorce.
35 Some of these studies are cited as conference presentations (for example, Brant & Sink, 1984; Guyer & Ash, 1986; Yates & Musty, 1987); such presentations are typically difficult to obtain (and therefore assess), and are generally not peer-reviewed. Accordingly, these studies are not reported here.
36 Faller and DeVoe (1995) noted this; the original paper is somewhat opaque on this breakdown. Faller and DeVoe thus suggested that the derived rate of false allegations is actually 71% (10 out of14 cases) rather than 55% (10 out of18 cases).
37 See Corwin, Berliner, Goodman, Goodwin, and White's (1987) critique of Green's research.
38 Only the more difficult cases or ambiguous cases are generally referred to specialist child psychiatrists. Faller (2003) has pointed out that the study conducted by Benedek and Schetky is probably a study of cases "predetermined to be likely false" (p. 243) since Benedek herself noted that she tended to decline briefs in which there was evidence that sexual abuse may have occurred.
39 McGraw and Smith (1992) found that in many cases caseworkers prejudged allegations as "false" because they were made in the context of divorce proceedings.
40 Oates et al. (1998) have also cautioned against the use of a simple binary "substantiated" vs "non-substantiated" division. In their view, the latter category suggests that child sexual abuse "might well have occurred but was not able to be substantiated on this occasion" (p. 154). They believe such a view should be discouraged and have attempted to address this issue in their own categorisation scheme.
Oates and his colleagues (1998) reviewed the case notes of all (n = 551) child sexual abuse reports to the Denver Department of Social Security that were opened and concluded over a 12-month period. Oates and his colleagues divided the results into: (a) substantiated - assessment that abuse had occurred made on one or more of the following grounds: information from child or family, medical evidence, court findings and/or perpetrator's confession; (b) no sexual abuse - judged to have definitely not occurred or occurrence highly unlikely, but not in the "erroneous concerns" category (see (d)); (c) inconclusive - judged to have insufficient evidence to put in the "substantiated" category; and (d) erroneous concerns by children - consisting of cases of collusion between parent and child, cases in which a child thought sexual abuse had occurred but a professional did not, and cases in which a child had knowingly fabricated a story of sexual abuse. Inter-rater reliability with respect to capacity to assign to the above four categories was 89.5% (based on every 20th case).
Substantiation occurred in 43% of the cases, while 21% of cases were rated as inconclusive. In 34% of cases, it was judged that no sexual abuse had occurred, while 3% were assessed as erroneous concerns.
The rate of erroneous concerns was in keeping with the range between 2% and 8% found by others (Everson & Boat, 1989; Faller, 1991; Goodwin, Sahd, & Rada, 1978; Jones & McGraw, 1987; Peters, 1976). Oates et al. (1998) point out that although considerably higher incidents have been reported (Benedek & Schetky, 1985; Green, 1986), these have comprised "small samples of selected cases referred for forensic evaluation where allegations about sexual abuse were complicating custody decisions" (p. 153).
41 The complexity of arguments around the truth or otherwise of such allegations is also well illustrated in a statement on allegations of child sexual abuse in Bala, Mitnick, Trocmé, and Houston (in press). Drawing on data taken from 10,756 useable cases in the 2003 Canadian Incidence Study (also reported on in Trocmé et al., 2005), Bala and his colleagues summarised the situation as follows: "A significant proportion of allegations of child abuse made in the context of parental separation are true, but this is a context with a relatively high rate of unfounded allegations. While some cases of untrue allegations are due to fabrication, more commonly unfounded allegations are made in good faith. Pre-existing distrust or hostility may result in misunderstandings and unfounded allegations, especially in cases where the children involved are young and the allegations are reported through a parent. Some cases of unfounded allegations may be the product of the emotional disturbance of the accusing parent".
42 The "intentionally false" category is actually subsumed under the "unsubstantiated" category.
43 Thoennes and Pearson (1988) wrote: "In the absence of a generally accepted set of phrases to describe these cases, we refer to true, false, indeterminate, and spiteful or fictitious allegations" (p. 23).
44 Even though Australia stopped using the language of "custody" and "access", and more recently, with the introduction of the Family Law Amendment (Shared Parental Responsibility) Act 2006, no longer uses their replacements ("residence" and "contact"), we have retained the language of "custody" and "access" (or "visitation") in this report to stay true to the use of these terms in the North American literature.
45 The Kempe National Center for the Prevention and Treatment of Child Abuse and Neglect was established in 1972 to provide a clinically based resource for helping to treat abused children.
46 The re-analysis was actually conducted by the second author, since the first author had been involved in the original clinical assessments at the Center.
47 The reported rate was actually 16.5%, but all figures reported in the text of this chapter have been rounded for readability.
48 Percentages do not total 100% due to rounding. Faller and DeVoe (1995) also noted two other sets of cases: 13% of cases in which the allegations of abuse were not related to custody proceedings because the alleged abuser was not a biological parent; and 4% of cases involving apparently false allegations made by children (of which only 1 of 9 instances appeared to be intentionally false). It should be noted that the 10 "deliberately false" allegations made by adults actually only involved 6 parents; 4 of the allegations were made by one father.
49 This last estimate is based on 6% malicious false reports by adults and 3% by children.
50 While data were not collected in the OIS about whether children were involved in a custody dispute, "cross-custody allegations", in which separated parents raised allegations against each other, could be identified as a rough proxy for disputes (Bala et al., 2001, p. 32).
51 Similar relative proportions emerged in the OIS 1998 (see Table 8-1a in Trocmé et al., 2002, p. 116).
52 This data collection followed an extensive round of preliminary interviews with, and surveys of, over 200 court personnel (judges, administrators, mediators and custody evaluators) across the US, complemented by a subsequent round of personal interviews with 70 court professionals in 5 states involved in cases where allegations of sexual abuse have been raised. Some of these preliminary data are discussed by Thoennes and Pearson (1988).
53 They noted that this estimate is similar to those reported by Duryee (1987), who found that 6% of allegations of sexual abuse were raised in 524 custody/access disputes during a three-month block in 1985-86; 5% of allegations of sexual abuse were raised in 861 disputes during a similar period in 1986-87 (cited in Thoennes & Tjaden, 1990).
54 Where a parent accused a third party, the estimates were 30% "likely", 45% "unlikely" and 25% "indeterminate".
55 The issues included: "domestic violence, substance abuse, maligning, harassing, emotional abuse of child, physical abuse of child, sexual abuse of child, child neglect, psychological disorder, stalking, and child abduction" (California Administrative Office of the Courts, 2003, p. 2).
56 They were especially interested in the way in which judges interpreted and applied the law relating to children's matters, but noted that existing research into judicial decision-making had generally relied on surveys of judges (e.g., Pearson & Ring, 1981, cited in Jones & McGraw, 1987; Settle & Lowery, 1982; Waller & Daniel, 2004). They noted that asking judges to rate the importance of certain criteria in their decisions (or asking lawyers to second-guess judges' thinking) is problematic at a number of methodological levels - not the least of which is the potential for social desirability bias. Moreover, it may be difficult at the best of times to articulate one's own decision-making processes. Sorenson et al. (1995) sought to overcome some of these methodological issues by modelling judicial decision-making from a sample of contested custody proceedings.
57Substantiation was based on "corroborating information such as provided by a medical examination, report of witnesses, police, or child protection personnel" (Sorenson et al., 1995, p. 254).
58 The project also conducted in-depth interviews with parents, and step-kin and other kin involved in the disputes.
59 Smart et al. (2003) defined "violence" and "abuse" as "physical, verbal and emotional violence and harassment of either a parent or child" (see note 18, p. 131).
60 While equal numbers of parents were in the custody evaluation and custody counselling groups, Johnston et al. (2005) noted that one-third of those in the custody counselling group had also received a prior custody evaluation.
61 According to Johnston and her colleagues (2005), the data had been collected, on average, about three to four years after parents had separated.
62 According to Thoennes and Tjaden (1990), there are several possible reasons for this disparity; similar to other pyschosocial stressors, child abuse may create stress and lead to separation, parental separation may itself create opportunities for abuse in certain family circumstances, and children may be more open to disclosing abuse by a parent because secrecy is less enforceable and the non-abusing parent may be more willing to believe the child in the face of inter-parental distrust and self-reliance (see also Corwin et al., 1987).
63 Johnston et al. (2005) found that allegations of domestic violence were raised against mothers in 30% of families and against fathers in 55% of families in their sample.
64 However, only 2% of cases in Smart et al.'s (2003) sample involved allegations of both domestic violence and child sexual abuse - though this estimate is likely to be tempered by the low frequency of allegations of child sexual abuse.
65 See in this regard an excellent review of the concept of unacceptable risk by Fogarty (2006).