Gender Paradigm in Domestic Violence Research Vis The Information Website of the American Bar Association
Gender Paradigm in Domestic Violence Research Vis The Information Website of the American Bar AssociationDonald G. Dutton, Kenneth N. Corvo, John Hamel
Aggression and Violent Behavior, Volume 14, Issue 1, 2009, Pages 30-38
Reproduced with permission of the author
| EJF Home | Find Help | Help the EJF | Comments? | Get EJF newsletter | Newsletters |
| Domestic Violence Book | DV Site Map | Data tables | DV bibliography | DV index |
| Chapter 2 — Domestic Violence And The Rule Of Law |
| Next — License To Kill: Men And Women, Crime And Punishment by Cathy Young |
| Back — American Bar Association Goes Over To The Dark Side by Carey Roberts |
The web site of the American Bar Association (ABA) set out to correct ten purported myths about domestic or intimate partner violence (IPV). The critique of these myths appears to be empirically based. However, a close reading of the studies used to debunk these “ myths” shows that they are either:
(1) government publications with no empirical data, or
(2) empirical studies that do not refute the targeted myth.
The problems with the false conclusions on their website are varied, but three main ones are:
1) confusion of allegations of abuse with real incidence of abuse;
2) interpretations of unsubstantiated claims of child abuse that are based on varied sources for corroboration that use vague decision criteria in studies not designed to assess malingered claims; and
3) over simplification of the complex causality of psychological phenomena, such as Parental Alienation Syndrome.
In many of these studies, social science methodology may be poorly suited to answer questions best left to an unbiased weighting of facts in an individual case.
As Wilson (1983) points out in the above quote, criminal justice policies that mis-conceptualize the problem they seek to solve are doomed from their inception. This mis-conceptualization has been ubiquitous in North American criminal justice policy on domestic violence (DV), mis-directed by a gender paradigm that interferes with effective and empirically based interventions. In a series of reviews of numerous empirical studies (Dutton, 2005, Dutton and Corvo, 2006, Dutton and Corvo, 2007, Dutton and Nicholls, 2005, Hamel, 2007), we have argued that this predominant “ paradigm” in domestic violence is politically driven and not supported by the data, constituting a lens through which domestic, or intimate partner violence (IPV) is viewed and what policy options are implemented to resolve the problem.
In fact, the very belief that the problem is resolvable mainly by criminal justice action is predicated on this paradigm. The gender paradigm frames IPV as solely or primarily, male perpetrated, used primarily as an instrument of control (serving the patriarchal function of suppression of women, e.g., Dobash & Dobash, 1979), and presents female IPV as self-defensive (e.g., Dobash & Dobash, 2004)).
Contrary to this paradigm, contemporary North American society is among the most gender egalitarian of societies both geographically and historically (Archer, 2005). Studies attributing IPV in North America to patriarchal causes have repeatedly been dis-confirmed. For example, only 2% of North American males agree that it permissible “ hit your wife to keep her in line” (Simon et al., 2001), less than 10% of North American marriages are male dominant (Coleman & Straus, 1986), and most IPV is bilateral (Stets and Straus, 1989, Whittaker et al., 2007) even when matched for level of severity. Only 6% of couples who are violent demonstrate “ wife battering” pattern (of severe male violence and no female violence, Stets & Straus, 1989).
Conversely, tendencies to use IPV develop early in women and remain as an aggressive trait (Capaldi et al., 2004;Moffitt et al., 2001, Serbin et al., 2004), and are not, as the gender paradigm portrays, survival-based reactions to male violence.
Numerous studies have focused exclusively on male violence perpetration while women in the study clearly also initiated and were contributing to the IPV at high rates (e.g. even in samples selected for male violence, women initiated the violence in 40% of cases: Gondolf and Jones, 2001, Jacobson et al., 1994, Dutton and Corvo, 2007), creating the misimpression that male aggression, operating in a vacuum, is the sole causes of the IPV.
Without restating the entire argument, the notion that domestic violence is solely motivated by male domination of women has been rejected on several grounds, including huge and representative data sets showing female IPV to be more commonplace than male perpetrated IPV (Archer, 2000, Stets and Straus, 1989, Whittaker et al., 2007), to generate only moderately more injuries (Whittaker et al., 2007), and to be generated by the same motives (Fiebert, 2004, Follingstad et al., 1991).
Large sample data sets also show that bilateral violence is the most common form of IPV (Stets and Straus, 1989, Whittaker et al., 2007); and that developmental factors drive women and men to assortative (”birds of a feather”) mating (Capaldi et al., 2004;Serbin et al., 2004) by which violence-prone individuals seek out similarly disposed others. For these reasons, bilateral couple violence rates, when they are properly assessed, range from 42% (Stets & Straus, 1989) to 63% (Neidig, 1993).
Often, with the gender-political focus exclusively on “male violence,” proper assessments are not made, or the data showing female violence are buried or suppressed. Incidence rates of female violence, even towards their children, have been largely suppressed (Dutton, 2005, Dutton, 2006a, Dutton and Corvo, 2006). The result is a professional mindset that is based on false information and which leads to errors of judgment (Follingstad et al., 2004, Hamel et al., in press, Lidz et al., 1993).
In previous papers (e.g., Dutton & Corvo 2006, 2007), we have outlined the origins of this belief system in Marxist thought (Dobash and Dobash, 1978, MacKinnon, 1989) and showed how such a Manichean view oversimplifies the complexities of IPV. In this paper, we examine the application of the gender paradigm to the mis-education of practitioners, specifically, the legal profession and custody assessors.
Despite repeated empirical dis-confirmations, the gender paradigm is maintained by advocates and presented to professional groups as if supported by the empirical research, when, as we have argued, the best studies show the opposite.
It is noteworthy, then, to examine the conceptualization of IPV, offered by professional associations to their members. Again, to paraphrase JQ Wilson, less than a clear understanding of the nature of IPV dynamics will result in an improper design of public policies and practices to deal with the issue.
We begin with the depiction of domestic violence by the American Bar Association website. In July 2006, the ABA distributed a flyer in its Quarterly newsletter that listed “ 10 Myths about Custody and Domestic Violence and How to Counter them.” This flyer was reproduced on the ABA website in 2006 and remained there at the time of writing (American Bar Association Commission on Domestic Violence, 2006). The flyer and website were prepared by the ABA Commission on Domestic Violence, and the flyer states that the Commissions' purpose “ Mobilizing the legal profession to provide access to justice and safety for victims of domestic violence.” The American Bar Association Commission on Domestic Violence argues that “attorneys who represent victims of domestic violence in custody matters often encounter the following false claims. To assist with overcoming these myths, the ABA Commission on Domestic Violence provides these facts and statistics for use in litigation.” The website then offers up a list of ten “myths” that it purports to counter with empirical research.1
It should be noted that these “myths” are placed in the context of resource materials to which the reader is directed and which could, in themselves, promote a gender view that all IPV is male perpetrated. The problem is that, almost without exception, the resource materials offered to dispel these myths are flawed, biased, outdated, and/or inconclusive research studies (or other government reports that include no empirical data). Elsewhere on their site, the ABA states: “The ABA Commission on Domestic violence does not engage in research, and cannot vouch for the quality or accuracy of any of the data excerpted here. Users are advised to independently confirm data with source documents cited.”
However, this disclaimer is not applied to their list of “myths” and purported disconfirming evidence. Presentations of “myths” which are then summarily dispelled or disproved are a frequent rhetorical device. This process conveys to the reader a convenient sense of knowing with absolute certainty without rigorous investigation. Often the “myths” and their refutation involve common logical fallacies such as the straw man argument.
To investigate the scientific credibility of the “myths” and their dismissals, we will approach the evidence like any forensic analysis (Weissman & DeBow, 2003) starting with competing assumptions as to the truth. We will examine the scientific evidence for the dismissals in particular since many of the myths cited appear to have no basis in evidence. It is not our intent to confirm the “myths,” but rather to reveal the evidentiary weakness in their dismissal and the corresponding risks for distorted understanding of DV by users of the ABA website.2
The first “myth” stated is that domestic violence is rare amongst custody litigants. It is not clear who believed this “myth” and no reference is provided. Two references are cited to dispel this myth; however, the studies cited (Johnston, 1994, Keilitz, 1997) do not answer the question posed because both studies assess allegations of DV rather than actual DV. This issue is problematic throughout the ABA “refutations” because it is not known whether the allegations represent actual abuse rates or a legal ploy by the complainant to gain advantage in a custody case. While the ABA website accepts every allegation as a veridical truth, the data say something very different.
The first ABA citation is to a resource handbook (Keilitz, 1997) that presents no direct empirical evidence except a survey of state courts on DV incidence. Evidence for DV presented by these state courts “included civil protection orders, documents related to criminal charges in the custody case files, self-reports in questionnaires and interviews, allegations in the pleadings, and other evidence in the case record” (p. 5).
Obviously, from this sort of compilation of various evidence sources, it is impossible to differentiate allegations of DV from actual incidence. The reader is not told in what proportions these various sources, of uneven evidentiary weight, contribute to the final conclusion. The reader does not know whether DV incidence is based solely on self-reports or allegations, or what “other evidence in the case” consists of. The Keilitz study, however, treats these all sources as “evidence of domestic violence” (p. 5).
Such conflation of allegations with evidence runs through the bulk of the literature cited by the ABA website, transforming what are often merely complaints to police into corroborated incidents of domestic violence.
The study by Johnston1994) cites earlier studies, one by the same author (Johnston & Campbell, 1993) and one by Depner, Cannata, and Simon (1992) as finding “physical aggression had occurred between 75% and 70% of the (high conflict divorce) parents.” However, the Depner et al study again reported only allegations of abuse.
The Johnston and Campbell (1993) study, while sparse on methodological details, gave out Conflict Tactics Scales (CTS) (still an uncorroborated self-report measure of DV) to two small samples (n = 80 and 60) of divorcing couples in San Francisco. Any type of domestic violence was presumably counted, including “throwing or smashing objects.” The authors then developed a typology “based on clinical inference.” They identified five patterns of IPV in their clinical sample of high conflict custody litigants: ongoing male battering, female initiated violence, male-controlling violence, separation/divorce violence, and psychotic/paranoid reactions (op, cit., pp. 288–289). The “ongoing male battering” (which constitutes the stereotype of all IPV in the gender-paradigm) at “moderate or low levels of severity” was found in 8% of couples in sample one and 11% in sample two.
Little is known about the validity of the CTS in these circumstances of reporting, but the CTS is susceptible to social desirable reporting in court-mandated samples (Dutton & Hemphill, 1992). It is designed to measure incidence from anonymous survey respondents and cannot be interpreted as providing valid incidences in highly emotionally charged conditions. In this light, Johnston and Campbell reported the intra-couple reliability of the CTS scores “ranged from .2 to  .62. It is clear that the couples were not in agreement on the CTS. The authors did not report breakdowns of items by respondent, so the exact nature of the disagreement cannot be assessed.
A later and methodologically superior study, by the same author (Johnston, Lee, Olesen, & Walters, 2005), makes the importance of this distinction between allegation rates and actual incidence rates abundantly clear. In that study (reviewed below), allegations of sexual abuse of children were made against fathers in 23% of the cases studied but substantiated by a judges' decision in only 6% (op. cit., Table 3). Allegations of physical child abuse were made in 21% of cases and substantiated in only 6%. For the category “any child abuse,” allegations were made against fathers in 51% of the cases studies but substantiated in only 15%. Of all the incidence studies cited as evidence on the ABA website, not one used a measure of actual rates of abuse, yet the Johnston et al data underscore the difference between allegations and substantiated abuse. This substantiation rate, using a judges' decision as the criterion measure, was about 1/4 of the allegation rate.
The Straw Man of this myth is that effects of domestic violence on children are minimal or short term. Of course, they are not. The problem is that the studies cited (Felitti et al., 1998) abstracted cause and effect relationships from multiple-problem dysfunctional families. The Fellitti study assessed the long term impact of exposure to multiple sources of abuse and family dysfunction. The greatest long term health problems were found in people who had exposure to four or more adverse experiences as children, including family drug abuse, alcoholism, sexual abuse, and mental illness.
While the Fellitti et al study assessed “violence towards mother,” it did not assess violence towards fathers although surveys show these to be equally prevalent (Archer, 2000, Stets and Straus, 1989; Whittaker et al., 2007). Other studies cited in support of the website argument include those by Jaffe et al. (1990) and Bancroft and Silverman (2002). The former drew their sample from a battered woman's shelter, the latter from a treatment group for males convicted of domestic violence. Then both studies were generalized to a community population (like the one studied by Fellitti et al.) disregarding that both samples were highly selective and not representative. For a more thorough discussion of the shortcomings of these studies, see Dutton (2005).
The message to readers from the spin put on these studies by the ABA Website is that DV is committed only by fathers (e.g., Jaffe, Wolfe, & Wilson, 1990) against mothers and children (Edelson, 1999, Appel and Holden, 1998) with long term consequences for both (Fellitti et al.).
As Dutton stated, (2005) “there is a priming of assessors to look only at the male as the abuse perpetrator, and having done so, to suspect his denial of abuse” (p. 25). Such priming occurs in a more indirect fashion on the ABA website. The studies cited on the effects of DV on children are all on male-perpetrated violence. One is Jaffe et al.'s book drawn from a shelter sample. Another, by Morrill et al. (Morrill, Dai, Dunn, Sung, & Smith, 2005) is based on a cases where “the father has already perpetrated violence against the mother.” In fact, it was not a study of the effects of DV on children but a study of whether Family Court Judges had correctly (sic) assimilated the teachings on DV of the National Council of Juvenile and Family Court Judges in making their decisions (p. 1089). (The authors found the judges attitudes to be “deficient” (p. 1076)). The authors request better “quality of DV education” (p. 1076) by which they mean, promulgation of the gender paradigm.
The third study, by Edelson (1999), is from another researcher who studied the overlap of woman battering and child abuse based only on samples from men's court-mandated treatment groups. In sum, all studies cited as evidence against the “myths” are from biased and self-selected samples that cannot be generalized to the community. These are then combined with a community sample of multi-problem families to imply that “male perpetrated” abuse causes long term physical problems. While no one is trivializing child abuse, Kaufman and Zigler (1993) did find that “transmission rates” (abused children who abuse as adults) were overstated in the literature and are around 30%. Hence, over 2/3 of abused children function adequately as parents/spouses. Emotionally supportive experiences with other adults seem to mediate the transmission rate and stop inter-generational replications of abuse.
Furthermore, while the ABA website cites articles emphasizing exposure to male IPV as a risk to children, recent studies indicate that the greater risk (the greater risk of what? Becoming an abuser themselves?) is from abuse at the hands of mothers (Gaudioisi, 2006, Trocme, 2001). McDonald et al. (McDonald, Jouriles, Ramlisetty-Mikler, Caetano, & Green, 2006) specifically examined number of children in the US exposed to IPV and the gender of the perpetrator. In a sample of 1,615 dual-parent households, they found, through face-to-face interviews, that children in these families were exposed to any type of male to female violence in 13.7% of the families and exposed to severe male to female violence in 3.6%. Corresponding incidence figures for violence by an adult female violence (to a male) were 18.2% and 7.5%. Hence, the greater risk was for children exposed to violence by adult women.
Although the vast majority of research has focused on male-perpetrated IPV, effects on children who have witnessed their parents physically abuse one another, both directly and indirectly (e.g., emotional and conduct disturbance, deterioration in peer and family relations, and poor school performance), occur regardless of the perpetrator's gender (Davies and Sturge-Apple, 2007, English et al., 2003, Fergusson and Horwood, 1998, Johnston and Roseby, 1997, Mahoney et al., 2003). Furthermore, correlational studies indicate that child witnesses to interparental violence are at equal, or greater, risk for becoming depressed, engaging in substance abuse, and perpetrating intimate partner abuse themselves as adults when their mother was the abuser (Kaura and Allen, 2004, Langhinrichsen-Rohling et al., 1995, Margolin and Gordis, 2003, Sommer, 1994, Straus, 1992).
Myth 3 is that allegations of sexual abuse are invented by mothers to win custody cases. The evidence provided against this claim comes from a study by Tjaden and Thoennes (2000) of 169 cases that purportedly found that sexual abuse allegations were rare (6%) in custody cases and 2/3 were “substantiated” based on unsubstantiated judgments of child protection workers or custody evaluators.
In another study used to refute the myth that false allegations are manufactured in custody disputes (Brown et al., 2000), two Australian samples had allegation rates of 18.2% and 36.1% for physical abuse and 12.1% and 48.6% for sexual abuse. In fact, Brown et al reported that “substantiation rates were found to vary between (Australian) states.” Although children in the study had high levels of psychological distress, it was not known whether this stemmed from exposure to abuse or parental divorce and substantiation rates for the child abuse allegations were not published in this study.
The ABA web does not cite these concerns or the high allegation rates reported by Brown et al., choosing instead to cherry pick the lower rates (6%) reported by Thoennes and Tjaden. The Thoennes and Tjaden study was based on unsubstantiated judgments from workers trained to suspect abuse, even from “soft signs” of child distress, that may, in fact, have been due to abuse or marital breakdown.
The Bala and Schuman paper (1999) warns the reader that “there are legitimate concerns about the possibility that accusing parents or children may be lying (or more likely may be mistaken), those who have abused children usually deny or minimize their abuse” (p. 192). The error in logic here is to put the denial (of abuse) before the evidence, so the denial magically becomes “evidence” for what is denied.
This is exactly the same witch-hunt procedure that Jaffe et al. (2003) use to prime custody assessors about purportedly violent men (see Dutton (2005) for an expanded discussion). Bala and Schuman confuse an abuse allegation with actual abuse, a confusion repeated throughout the ABA website. At an early point in the custody assessment, it is not accurate to equate an accused person with a proven abuser. Also, Bala and Schuman put it “while in some cases of false allegations there may be deliberate effort to lie, more commonly the parent who brings forward the unfounded allegation of abuse following separation has an honest belief in the allegation” (pp. 193–194). In other words, accusing parents are more likely to be mistaken than to lie. This claim is made without any empirical evidence offered to support it.
In fact, in their sample, Bala and Schuman found that only 23% of sexual and physical abuse allegations made by mothers were substantiated by a judicial written decision on the basis of a “balance of probabilities” (the civil standard), the remainder were unproven. We use the notion of “innocent until proven guilty” here, unlike Bala and Schuman. There is, in fact, an unstated premise running through this literature that all allegations are true, some are just difficult, or impossible to prove. This is surely a most curious position for lawyers to take.
A later paper by Trocme and Bala (2005) examined 7,672 child maltreatment investigations, some of which were made as a result of a custody-separation allegation. In general, 1/3 of the maltreatment investigations were unsubstantiated. Of all cases, 4% were considered to have been intentionally fabricated. In the custody sub-set, this intentional fabrication group increased to 12%. While the authors data tables did not disaggregate cases by gender, the authors did report that “non-custodial parents (usually fathers) are more likely to make false allegations than are custodial parents (usually mothers).” How this is known is not reported. A more serious question, however, is not answered by Trocme and Bala's methodology. This is specifically; did child protection workers (who made these judgments) decide that these allegations were “intentionally false” as opposed to unfounded? Bala and Schuman themselves say there is usually no conclusive evidence to support a claim of an intentional false allegation (p. 192).
Absence of evidence allowing substantiation could be either unfounded or intentionally false. The proof of intention would be much more difficult to determine and one wonders whether the effort to make this assertion might occur in what was, in effect, a child maltreatment investigation. It seems peculiar that a peer reviewed journal would accept data sets based on judgments where the judgment heuristic is not elucidated, or not the focus of the investigation (which was to substantiate the child abuse claim) and is susceptible to bias.
This “evidence” for Myth 3, as cited by the ABA saying that fathers are more likely than mothers to make intentionally false claims against the other parent, comes from judgments made by “child protection workers” in a study done in Canada (cited in Bala & Schuman, p. 196 and Trocme & Bala, p. 1337). These judgments were never subjected to evidentiary tests in court.
The hysteria surrounding child abuse in Canada was so great at that time, that a pediatric pathologist was later revealed to have manufactured “evidence” in his conclusions in 20 cases of wrongful prosecution and 12 cases of wrongful conviction of “abusive parents.” Both this pathologist and the child protection workers were working with the same mindset, described this way “concerns about child abuse had reached almost hysterical proportions...which often produced tunnel vision amongst investigators, who would sometimes look for evidence to substantiate their suspicions, rather than conducting investigations with an open mind” (Vancouver Sun, January 31, 2008).
This mindset led to the belief that satanic cults committed infanticide (Victor, 1996) and that child abuse was rampant in daycares (such as the McMartin Case in Los Angeles) and the firm belief that fathers, accused of sexual abuse of children probably were guilty even if the evidence was difficult to adduce. Hence, unproven judgments by child protection workers primed to suspect abuse in an era of abuse-hysteria constitutes the “evidence” base for the ABA refutation of Myth #3.
It seems peculiar that a peer reviewed journal would accept data sets based on judgments where the judgment heuristic was not elucidated, not the focus of the (investigation (which was to substantiate a child abuse claim) and susceptible to bias (see Dutton, 2005, and Kahneman, Slovic, and Tversky, 1982). It also seems peculiar that a professional group who believes in due process and has standards (e.g., Frye, Daubert) for admitting scientific evidence to court would accept these subjective judgments as evidence.
When a higher standard of proof for abuse is required, the picture changes. As we mentioned above, Johnston et al. (2005) conducted a large sample study of allegations and substantiations of abuse in custody-disputing families in California. Substantiations in this study were defined as any corroborating evidence of abuse to back up the allegations, that, in their words “had not been dismissed as entirely unfounded.” In comparison to the ABA claim that child sexual abuse allegations in custody cases are rare (6%), Johnston et al. found them to be made against fathers in 23% of cases studied (and against mothers in another 6%). In these cases, only 6% of sex abuse allegations against fathers were substantiated (and 3% against mothers). The findings of Johnston et al. may cause one to reconsider whether Myth 3 is, in fact, a Myth since the child sexual abuse allegation rate was about four times higher than the ABA claims (23% vs. 6%) and the substantiation rate (when actual evidence instead of judgments is used) is 6%. Furthermore, 51% of custody cases involved allegations of “any child abuse” against fathers, with 15% substantiated. For allegations against mothers, the corresponding figures were 38% and 17% (op. cit., Table 3).
To provide some baseline general population incidence data on this issue, two large scale studies have found the following. A huge (135,573) and nationally representative study of child abuse allegations was made by the National Clearinghouse on family Violence in Canada (Trocme et al., 2001). Substantiation rates in general ran from 52% to 58%. Biological mothers were more likely to commit physical child abuse (47% vs. 42%), emotional maltreatment (61% vs. 55%), and neglect (86% vs. 33%). Compared to biological mothers, biological fathers were more likely to commit sexual abuse (15% vs. 5%). These data, drawn from a nationally representative sample rather than from a pre-selected sample of women from a shelter house or men from a treatment group, give a very different picture of risk to children than that presented by the sources cited on the ABA Website.
The gender of the substantiated child abuse perpetrator in the general population is more likely to be female and the allegation/substantiation ratio is higher (1/2) than that obtained in custody cases (about 1/4), suggesting that false allegations may more frequently be made in the latter. This “different view” of the gender of the perpetrator is echoed by the even larger sample of child abuse perpetration in the 2004 US Department of Health and Human Services study. In their national study of risks to children (n = 718,948), 57.8% of the perpetrators were women and 42.2% were men. Mothers were involved in 51% of child fatalities; fathers in 38.6%. Large sample studies without a gender-political agenda paint a very different picture than the small sample cherry-picked results available on the ABA website.
The Johnston et al. (2005) study shows that allegation/substantiation rates drop to about 1/4 in custody cases compared to 1/2 in non-custodial cases. This statistic is telling. It does not prove “false allegations” but suggests higher rates of unsubstantiated accusations when custody is at stake. In our opinion, social science research is misleading when it purports to prove false allegation rates using data from child abuse investigations that were not designed to investigate false allegations. In fact, we would argue that social science cannot prove false allegations since such absolute proof would require reports on intent from uncooperative witnesses. On the rare occasion when a false allegation is discovered by police, it involves someone recanting in the face of evidence contradicting their statement. This requires time and effort that would not occur in a large scale social science study.
A final observation should be made on “substantiation” through judgments of child protection workers. Follingstad, DeHart, & Green (2004) found that actions presented to clinical psychologists in experimental vignettes were more likely to be deemed “abusive” when described as being enacted by males. The same action, enacted by a female was not seen as abusive.
This result applied to both physical actions and verbal actions, such as asking someone their whereabouts. This finding suggests that professionals whose judgment is required for “substantiation” may not substantiate actions equally by gender and may be primed to see abuse when hard evidence is lacking. The Johnston et al. study used “hard evidence” (eyewitness reports, self admissions, medical records, etc.) as well as “child protective service reports,” and it is a better test than those cited by the ABA, given the Follingstad et al finding. The one pervasive problem with research in this area (including Johnston et al., 2005), however, is that it conflates various sources of “corroboration.” One would like to know how substantiation rates vary with the presence of hard evidence compared to rates based on judgments alone.
We are not sure who would actually hold the belief that DV and child abuse are completely unrelated. Be that as it may, the implication and interpretation of evidence provided by the ABA appear to suggest that the assault of children is primarily a product of partner (male to female) assault. As can be seen, this “myth” is situated between a previous myth of mothers who invent allegations of abuse by fathers and followed by another “myth” concerning abusive fathers who get custody anyway. The Appel and Holden (1998) study, which the ABA cites as revealing a “significant overlap between domestic violence and child abuse,” actually found a co-occurrence of 40% between wife and child abuse in a shelter sample but only 6% in a community sample illustrating the lack of generalizability from shelter samples.
The study by Ross (1996) shows a connection between spousal abuse and child abuse but in a different way than the Appel and Holden (1998) study. Rather than trying to show “overlaps,” which create a huge false positive rate (generated, according to Appel and Holden, the item “pushed or shoved”), Ross regresses the amount of spousal violence onto child abuse probabilities finding that in extremely abusive families (50 or more acts of marital violence), the probabilities of child abuse approach 100% for male perpetrators and 30% for female perpetrators. We doubt anyone would argue that someone this frequently abusive should have custody of a child. This is a different argument however, from the implication that if you find spousal abuse, you should suspect child abuse. Again, the mechanism of “myth-dispelling” obscures a more complex phenomenon for the purpose of maintaining a gender-biased perspective. For example, the US Administration for Children and Families (Gaudioisi, 2006) reports that in 2005, women were more than 1.3 times more likely to abuse children than were men. When acting alone, mothers are twice as likely to abuse their children as are fathers.
To the extent that there is a correlation between perpetration of spousal abuse and child abuse, that correlation exists for both genders (Margolin and Gordis, 2003, Straus and Smith, 1990). Overall, the impact on children having to witness interparental violence versus having been physically abused are comparable (Kitzmann, Gaylord, Holt, & Kenny, 2003); and several well-designed studies have found that verbal and emotional abuse directed by a parent against a child may cause the greatest damage, again regardless of that parent's gender (English et al., 2003, Moore and Pepler, 1998).
The ABA cites the Violence And The Family: Report Of The American Psychological Association Presidential Task Force On Violence And The Family (1996), in support of the statement: “Abusive parents are more likely to seek sole custody than nonviolent ones...” There are no data, nor even a statement, concerning this in the cited material.
Further, the ABA cites the American Judges Foundation3 indicating that abusive parents are successful in securing sole custody about 70% of the time. The actual quote from the American Judges Foundation is: “Studies show that batterers have been able to convince authorities that the victim is unfit or undeserving of sole custody in approximately 70% of challenged cases.” There is no citation in this report documenting any empirical study. It has been suggested that male batterers are able to project a non-abusive image in court (Bow & Boxer, 2003), but this is solely a hypothesis, nothing more.
Lastly, the ABA states: “Allegations of domestic violence have no demonstrated effect on the rate at which fathers are awarded custody of their children, nor do such allegations affect the rate at which fathers are ordered into supervised visitation (i.e., abusers win unsupervised custody and visitation at the same rate as non abusers).” The study cited in support of this statement (Kernic, Monary-Ernsdorff, Koepsell, & Holt, 2005) compared couples with substantiated DV with allegations of DV, and a “randomly selected” (sic) non-violent sample of parents petitioning for marriage dissolution with no history of IPV and no evidence of either allegations of IPV or substantiated IPV (pp. 1000–1001). Data were collected from couples with minor children petitioning for dissolution of marriage. Inter-group comparability was problematic. Although the authors had access to comprehensive criminal justice records, substantiation was defined as “allegations of IPV that were supported by formal documentation including police reports, court reports, professional agency reports or confessions by the abusive partner” (Kernic et al., 2005, p. 1000).
Police reports or ex parte restraining orders constitute a form of “formal documentation.” Nevertheless, they are nothing more than self-reports (allegations) of IPV recorded by the police or by the court; the accused has not had a chance to defend him/herself. There is no requirement for the police to supply independent evidence to generate a police report. Also, the authors do not report how many men were classified as IPV solely on the basis of an allegation or a police report. Their conclusion that the courts failed to identify DV in their case files may have simply reflected the unwillingness of court officials to count allegations that the researchers considered substantiated but that the officials did not.
The researchers refer constantly to the “substantiated history” of DV in their IPV group. The researchers find the alleged perpetrator guilty despite a lack of due process. Of interest is the ABA reference to this article as finding “allegations of domestic violence have no demonstrated effect on the rate at which fathers are awarded custody.” This is indented under a section called “Myth 5: Abusive Fathers don't get custody.” The ABA, in effect, recapitulates the error of equating an allegation with guilt.
There are serious questions about whether the Kernic et al. methodology really “substantiates” DV since it again conflates hard evidence for abuse with “police reports” (i.e., reports made by one partner that have not yet been investigated). As described above, the differential effect of substantiation by hard and soft evidence is not available in the methodology nor the data tables. The paradigm influence on the Kernic study is apparent when the authors state that due to disproportionate share of severe abuse experienced by female victims, they dropped all cases where the female was the abuse perpetrator (p. 1000). The source for this take is the “crime victim survey” (Tjaden & Thoennes, 2000) that we have critiqued for imposing selective filters on male reporting of DV (Dutton & Corvo 2006). The authors reassure the reader though that “barring any other compelling factors, comparison group fathers do not warrant any restriction of their parental rights” (p. 1000).
Although allegations of domestic violence may have “no demonstrated effect on the rate at which fathers are awarded custody of their children,” as Kernic et al. argued, their own study found allegations that were even partially substantiated did have such an effect. Fully 71% of fathers with a history of some “substantiated” IPV had restrictions on their visitation, in comparison to 17.5% of fathers with no history of IPV, and fathers were more likely to be denied visitation altogether; supervised visits were ordered for 25.6% of fathers with some substantiated IPV, versus 4.6% of fathers in the comparison group.
Other studies also lament the alleged unresponsiveness of the family court system to the needs of female victims. Mediators who participated in a study of 400 disputed child custody cases in San Diego, California apparently ignored the existence of IPV in 36 of 70 cases in which an IPV allegation was accompanied by the filing of a temporary restraining order (TRO) (Johnson, Saccuzon, & Koen, 2005). Although mediators were as likely to recommend joint custody in IPV as non-IPV cases, supervised visitation was ordered in 75.3% of cases in which IPV was alleged). Neither this study nor the one by Kernic et al. provides evidence that violence against women was minimized or ignored. The problem is not that IPV against women is not taken seriously, but rather that it is identified as the only problem worth investigating.
The current obsession with IPV, and specifically male-perpetrated IPV, prevents investigators from fully exploring the context in which IPV occurs in families, including the extent of mutual violence, and it diminishes the likelihood of identifying other forms of abuse and dysfunction and their impact on children. For example, among the 36 cases of alleged IPV cited by Johnson et al. (2005), 20 also involved allegations of substance abuse and 9 documented serious parental psychopathology.
The elaboration of this myth reads: “Mothers who are victims of DV are often depressed and suffering from post traumatic stress disorder, and as a result, can present poorly in court and to best-interest attorneys and/or custody evaluators.”
Golding's (1999) review article is the primary basis for the ABA “refutation” of this myth. Golding also focuses exclusively on the gender paradigm: “intimate partner violence is restricted to violence by men against women” because studies that find women to be just as violent are “methodologically flawed” (p. 102), “women suffer more severe injuries,” and females have greater perception of risk. Same-sex violence is also dismissed (p. 102). Golding finds mental health problems in battered women based on a meta-analysis.
We can only reiterate that other research (Pimlott-Kubiak & Cortina, 2003) has found that the amount of exposure to trauma rather than gender is the main predictor of mental health sequelae, and that repeated studies find mutual violence (matched for level of severity), to be 2.5 times as common as “wife-battering” (Stets and Straus, 1989, Whittaker et al., 2007). Stress and trauma present a problem with self presentation in court for everyone, but the myth does not read: Fit parents don't lose custody.
The other source for refutation of this myth is the Kernic study which cites Golding as its primary source but adds a citation to Meier (2003). Findings from the Meier study are based on a sample of two case studies. The findings from a small N of case studies can only be interpreted as exploratory or suggestive, not explanatory or confirmatory.
The promotion of this myth suggests that battered mothers appear less fit as a result of the abuse rather than any deficiencies on their part, and that they relent out of fear of more abuse or for financial reasons (Jaffe & Geffner, 1998), when on a case by case basis, any of these explanations may apply. Support for the “refutation” has depended on victim reports and makes sense only if it assumed that women never initiate abuse or engage in mutual violence.
In fact, symptoms such as anxiety and depression may be evidence of victimization, perpetration, or involvement in the court process for both genders (Anderson, 2002, Pimlott-Kubiak and Cortina, 2003, Stets and Straus, 1992). The “myth” also ignores research finding both mothers and fathers involved in child custody disputes experience fear of one another. When asked if they were afraid to disagree with their partner because that partner might hurt them or their children, 28% of mothers and 27% of fathers said “often” (Newmark, Hartnell, & Salem, 1995).
The elaboration of this myth reads: “The American Psychological Association has noted lack of data to support so-called 'parental alienation syndrome,' and raised concern about the term's use.” What is not cited, however, is the concluding statement from APA: “However, we have no official position on the purported syndrome.”
It seems to us, somewhat disingenuous of the ABA, to invoke the concept of “scientifically sound” after their failure to provide methodologically sound studies for their own dismissals of the above “myths.”
Gardner (1987) described the “Parental Alienation Syndrome” as occurring when one parent attempts to alienate the child from the other parent. According to Gardner, signs of the syndrome include:
(1) the child is obsessed with 'hatred' of a parent;
(2) these children speak of the hated parent with every vilification and profanity in their vocabulary — without embarrassment or guilt;
(3) not only is there a rehearsed quality to the speech about the hated parent, but one also often hears phraseology that is not usually used by the child (many expressions are identical to those used by the 'loved' parent);
(4) when these children are asked to give compelling reasons for the hatred, they are unable to provide them.
According to Gardner, there is a campaign of denigration by the alienating parent against the targeted parent along with frivolous rationalizations for this denigration. The alienated child denigrates the targeted parent's entire family, regardless of quality of past relationships, expresses no ambivalence in hostile feelings towards the rejected parent, and insists that his/her negative views are completely their own (the “independent thinker” phenomenon). There is absence of guilt from the child about maltreatment of the targeted parent, and the child is unwilling to be impartial or hear other points of view.
Finally, the child's accusations do not appear genuine but are rather made up from the alienating parent's exact words or phrases. It seems that the Gardener work is a plausible hypothesis for use in evaluation in a custody case. Any forensic evaluation would assess this possibility and a competing possibility that the “alienation” was based on actual abusive or neglectful behavior. This is a case by case issue, not a general hypothesis that can or cannot be refuted generically. Although parental alienation is not a “syndrome” in the diagnostic sense of specific symptomatology leading to DSM inclusion, it has adequate psychosocial and psychological validity to be found in over 250 referenced works in the PsycINFO database, including almost 200 peer-reviewed academic journals.
In the first ever study on the long term consequences of parental alienation, Baker (2007) conducted a series of in-depth interviews with a selected sample of 40 adults, 15 male and 25 female, who had been alienated from a parent as children. All subjects had been pre-screened to determine the alienation did in fact occur, as opposed to estrangement. In 36 of the cases, the alienating parent was the mother. Baker found evidence among her subjects of parental alienation, as manifested in the specific ways previously described by Gardner. Baker identified three broad familial patterns of PAS:
(1) narcissistic mothers in divorced families who alienated the children from the father;
(2) narcissistic mothers in intact families who alienated the children from the father; and
(3) cold, rejecting or abusive parents of either gender, in divorced or intact families.
Baker's subjects reported that their alienating parents behaved like cult leaders requiring excessive devotion from them which cultivated dependency by insisting that the targeted parent does not love them thus erasing all traces of the targeted parent from their lives and withdrawing love and affection when the child showed any positive feelings for the targeted parent. Baker argued that such behaviors were a pervasive and serious form of child abuse. As adults, 70% of Baker's subjects suffered from depression and 35% experienced drug or alcohol problems.
The PAS, as formulated by Gardner, has been criticized for its exclusive focus on the behavior of the alienating parent as the sole cause of a child's rejection of the targeted parent, and it is thought to be used in court by fathers to gain a tactical advantage. A child's rejection of a parent, it is argued, may reflect realistic estrangement due to that parent's abuse or neglect or other factors rather than anything the other parent might be doing (Johnston, 2001).
To test this view, Johnston and her colleagues (Johnston, 2003) analyzed 215 disputed child custody cases litigated in the San Francisco Bay Area from 1989–2002. Among the factors found to be significantly correlated with rejection of a mother were separation anxiety around the father and child abuse by the mother, but the highest correlations were for alienating behaviors by the father. Significantly correlated with a child's rejection of the father were the father's lack of warmth and parental involvement with the child, the father's abuse of the child, and the mother's positive warmth and involvement separation anxiety around mother, and the mother's alienating behavior. Johnston's results produced a much more complex model (called a Structural Equation) of “child rejection of father” comprised of actions by both parents and at odds with both the notion that all disparagement of a parent is induced by the other parent through alienation or by the simplistic “black and white” presentation of the problem by the ABA.
It might be a more reasonable position to say that, while evidence for parental alienation syndrome is weak, alienating behaviors can certainly occur and can contribute to a child's reaction to the rejected parent.
One reference basis for this dismissal of this myth is a book by Bancroft and Silverman (2002) that is based on males drawn from court mandated treatment groups and then generalized to custody assessments of males who are not domestically violent or for whom DV is alleged but not proven. The Bancroft et al. book is not a homicide study.
The other is a study by Langford, Isaac, & Kabat (1999) that makes no comparative analysis of risk of homicide to children by whether or not their parents are married or divorced. It used a sample of 34 over a 5 year period, but 12 of the cases were actually teens killed by their dating partners. With a five year sample of 24, and no actual comparison of risk, the Langford study is irrelevant to the myth. As we pointed out above, under Myth 2 and Myth 4, the greatest risk for either child abuse or child homicide is from the mother (Gaudioisi, 2006).
It seems the intent of dispelling this myth is to uncouple DV perpetration from mental health problems. Some of the citations provided to disconfirm this myth are 20 years old; the most recent is 9 years old.
In contrast to the “refutation” of this myth, all studies of males in court-mandated treatment groups for spouse assault found elevations in incidence of Axis 2 disorders (personality disorders: Hamberger and Hastings, 1986, Hamberger and Hastings, 1988, Holtzworth-Munroe and Stuart, 1994, Holtzworth-Munroe et al., 1997, Saunders, 1992).
This is also true for females convicted of spousal assault (Carney and Buttell, 2004, Henning and Feder, 2004).
The only exception to this finding (Gondolph, 1999) was reported in a sample where reporting was so guarded (and corresponding response scale scores so out of range) that reports of personality disorder were suppressed. While it is possible for someone to commit repeat intimate partner violence without a personality defect, it is exceptional.
In addition to elevated rates of personality disorders, DV perpetrators also demonstrate higher rates of anxiety, depression, and other disorders (Dutton, 2006b) and neurological deficits (Corvo, Halpern, & Ferraro, 2006).
While there is debate about whether an Axis 2 personality disorder constitutes “mental illness,” they do constitute unusual and enduring patterns of cognition, affect, and behavior that are deviant within the host culture and cause significant distress (DSM-111-R, p. 689).
This is a sort of a “damned if you do, damned if you don't” type of myth, which seems to suggest that even if a child seems comfortable and secure with a parent, one should still suspect abuse since the child may have traumatically bonded with that parent.
The reference for this dismissal of the myth is the book by Bancroft and Silverman (2002) that is based on males drawn from court mandated treatment groups and then generalized to custody assessments of males who are not domestically violent or for whom DV is alleged but not proven. Traumatic bonding, based on work by Dutton & Painter (1981), describes the formation of powerful bonds in relationships of intermittent abuse.
It is not clear however, as the ABA claims, that these bonds would not be accompanied by “fear and aversion.” Fear and anxiety may be present independently of a traumatic bond. Traumatic bonding may present as an admixture of fear and attachment.
Traumatic bonding requires exposure to intermittent abuse followed by positive contact. It should not be presumed in the absence of evidence for such abuse. The awarding of unsupervised contact would still depend on a careful assessment of evidence for an abuse history against claims that the other parent has malingered this history.
What we see in the ABA's listing of “10 Myths About Custody and Domestic Violence and How to Counter Them” is the sort of superficial treatment of DV that is excused when it represents a perspective that is ideologically congruent with a gender-biased paradigm. This view that patriarchy is the sole cause of DV is accepted by advocates and some practitioners but has much less support in the research community. 12/30/25
We do not expect attorneys to operate from the same standards of scientific rigor.
Note: The ABA removed the 10 myths page from their web site after this review was published.
For links to references cited in this review go here.
| EJF Home | Find Help | Help the EJF | Comments? | Get EJF newsletter | Newsletters |
| Domestic Violence Book | DV Site Map | Data tables | DV bibliography | DV index |
| Chapter 2 — Domestic Violence And The Rule Of Law |
| Next — License To Kill: Men And Women, Crime And Punishment by Cathy Young |
| Back — American Bar Association Goes Over To The Dark Side by Carey Roberts |
This site is supported and maintained by the Equal Justice Foundation.