Judges must consider domestic violence when determining child custody under state law. Many states guide the custody inquiry with statutory presumptions against awarding custody to abusers. With custody outcomes often hinging on allegations of domestic violence, judges increasingly turn to experts for answers. But expert assessments of domestic violence in the child custody context lack a uniform and reliable methodology. As this Note reveals, the formulation of state custody statutes and lax application of evidentiary rules nevertheless encourage admission of potentially faulty evidence. In response, this Note advocates statutory reform and reconsiders the role of experts in custody cases, endeavoring to balance scientific limitations with the dangerous prevalence of domestic violence.


“When warring parents head to court to fight over child cus­tody[,] . . . their lawyers often let them in on a little secret: The most powerful person in the process is not the judge. It is not the other parent, not one of the lawyers, not even a child.” 1 Leslie Eaton, For Arbiters in Custody Battles, Wide Power and Little Scrutiny, N.Y. Times (May 23, 2004), (on file with the Columbia Law Review). Ten years later, one is not hard-pressed to find pieces striking a similar tone. See, e.g., M.C. Moewe, The Most Powerful Person in a Child Custody Case Isn’t the Judge, Daily Kos (June 16, 2014, 9:06 am), [] (providing similar, cynical account of experts’ influence in custody cases).

An expert’s 2 Throughout this Note, the words “expert” and “evaluator” refer to persons providing expert testimony in custody proceedings. These experts are commonly mental health professionals, such as psychologists. For a more thorough discussion of the role of experts in custody proceedings, see infra section II.A. testimony often has a determinative impact on a child custody case. 3 See generally Milfred D. Dale & Jonathan W. Gould, Science, Mental Health Consultants, and Attorney–Expert Relationships in Child Custody, 48 Fam. L.Q. 1 (2014) (introducing role of experts in custody cases). A court finding of domestic violence 4 What constitutes “domestic violence” sufficient for a court finding differs across states and, as this Note argues, influences the sort of evidence offered in a custody proceeding. For in-depth discussion of these variations, see infra section I.C.1. is similarly influential to the outcome of a custody matter. 5 See infra section I.B.2 (describing impact of statutory provisions regarding domestic violence on judicial decisionmaking). “Domestic violence is best understood as one intimate partner’s attempt to control, dominate, and humiliate the other partner through a variety of means, including physical, sexual, psychological, financial, and spiritual abuse.” Peter G. Jaffe, Nancy K.D. Lemon & Samantha E. Poisson, Child Custody and Domestic Violence: A Call for Safety and Accountability 4 (2003) [hereinafter Jaffe et al., Child Custody and Domestic Violence]. This is but one definition of “domestic violence,” as there is no standardized definition. With the growing prevalence of ex­pert testimony in custody cases, 6 See infra section II.A (explaining role of experts in custody cases). the Department of Justice funded a 2010 study to examine expert treatment of domestic violence and court reliance on this expert testimony. 7 Michael S. Davis et al., N.Y. Legal Assistance Grp., Custody Evaluations When There Are Allegations of Domestic Violence: Practices, Beliefs and Recommendations of Professional Evaluators (2010), []. The results were startling:

The sum of the research findings suggest that the facts of the case have less influence on the final custody and visitation arrangements than the custody evaluator’s understanding of domestic violence. As a result, when a custody case ends up in court, the fate of parents and children most often lies in the hands of the evaluator. 8 Id. at viii.

The study went on to conclude that while evaluators’ understanding of domestic violence and relevant practice methods was sorely lacking, 85% of settlement agreements and 70% of court orders mirrored expert findings. 9 Id. at iv, 19–20, 65–66. Others have found that courts follow expert custody recom­mendations up to 90% of the time. Alix Spiegel, Evaluators in Child-Custody Cases Scrutinized, NPR (Nov. 21, 2007, 4:00 pm),
php?storyId=16523618 (on file with the Columbia Law Review).

State statutes governing child custody make domestic violence a key consideration. 10 See infra section I.B (describing how domestic violence factors into court’s cus­tody analysis). A wave of new child custody legislation followed a period of increased attention to the frequency and consequences of domestic violence in the 1990s, which led Congress to pass a concurrent resolution con­demning abuse and calling on states to act by creating statutory presump­tions against awarding custody to abusive parents (DV presumptions). 11 H.R. Con. Res. 172, 101st Cong. (1990); see Proclamation No. 6213, 104 Stat. 5432 (1990) (declaring October 1990 “Domestic Violence Awareness Month”); infra sec­tion I.B.1 (recounting history of adoption of DV presumptions). Intimate-partner violence affects a significant portion of the U.S. pop­ulation, 12 Michele C. Black et al., CDC, Nat’l Ctr. for Injury Prevention & Control, National Intimate Partner and Sexual Violence Survey: 2010 Summary Report 2 (2010), [
Z3WC-RWMP] (“More than 1 in 3 women (35.6%) and more than 1 in 4 men (28.5%) in the United States have experienced rape, physical violence, and/or stalking by an intimate partner in their lifetime.”).
with some studies placing the number of child cus­tody cases involving domestic violence near 75%. 13 Peter G. Jaffe, Claire V. Crooks & Samantha E. Poisson, Common Misconceptions in Addressing Domestic Violence in Child Custody Disputes, 54 Juv. & Fam. Ct. J., Fall 2003, at 57, 58 [hereinafter Jaffe et al., Common Misconceptions]. Many states have since responded to these facts and Congress’s call by implementing DV presump­tions, noting the dangers posed to children by domestic violence. 14 See, e.g., Cal. Fam. Code § 3020(a) (West 2004) (“The Legislature . . . finds and declares that the perpetration of child abuse or domestic violence in a household where a child resides is detrimental to the child.”). Current formulations of DV presumptions and related rules necessarily influence the evidence a party will offer, including expert testimony. 15 See infra sections I.C–D (detailing these statutes and rules).

Though court reliance on expert testimony in custody cases is not an entirely new phenomenon, social scientists and legal scholars have only recently scrutinized these experts. 16 For the seminal works in this area, see Robert E. Emery, Randy K. Otto & William T. O’Donohue, A Critical Assessment of Child Custody Evaluations: Limited Science and a Flawed System, 6 Psychol. Sci. Pub. Int. 1 (2005); Timothy M. Tippins & Jeffrey P. Wittmann, Empirical and Ethical Problems with Custody Recommendations: A Call for Clinical Humility and Judicial Vigilance, 43 Fam. Ct. Rev. 193 (2005). Many questions raised by this literature remain open today. See Elizabeth S. Scott & Robert E. Emery, Gender Politics and Child Custody: The Puzzling Persistence of the Best-Interests Standard, 77 Law & Contemp. Probs., no. 1, 2014, at 69, 71 (building on seminal studies questioning reliability of experts). Prior to ten years ago, partisan and discredited fathers’ advocacy organizations largely levied the criticism of evidence related to domestic violence. 17 For a survey of the efforts of these groups, see generally Kelly Alison Behre, Digging Beneath the Equality Language: The Influence of the Fathers’ Rights Movement on Intimate Partner Violence Public Policy Debates and Family Law Reform, 21 Wm. & Mary J. Women & L. 525 (2015) (discrediting efforts of “fathers’ rights” groups); see also, e.g., ACFC Mission Statement, Am. Coal. for Fathers & Children, [] (last visited Oct. 29, 2015) (exemplifying goals and efforts of such “fathers’ rights” group). Consequently, questioning the reliability of evidence related to claims of domestic violence has been politically dicey. 18 Cf. Scott & Emery, supra note 16, at 70 (describing strong politicization of substantive custody reform). The task also involves a host of competing interests. 19 Parental rights, children’s safety, safety of the parties, and governmental interest­s are all at stake. Without question, any credible work in this area must recognize the prevalence of domestic violence and the immediate danger it poses to victims and their children. 20 Claims by fathers’ advocates of rampant falsification of domestic violence claims are unsubstantiated. See infra note 138 (disputing falsification claims). Ever attentive to this reality, this Note builds upon the efforts of scholars and neutral observers in reaching its conclu­sion: Courts are admitting and relying on flawed evidence, and state law is encouraging it. Reliance on flawed expert evidence on domestic violence in the child custody arena creates the potential for erroneous decisions, 21 Choosing to rely on one piece of evidence or another logically has the potential to alter outcomes. This Note does not attempt to establish the rate of erroneous decisions nor the specific impact of these decisions on families. But see infra text accompanying note 209 (noting courts, in some cases, have awarded custody to abusers). Rather, it high­lights that courts may give potentially flawed testimony dispositive weight. See supra note 9 and accompanying text (explaining extent of court reliance on experts); infra text accompanying notes 153–155 (same). generates judicial inconsistency, 22 See infra section II.B (finding same testimony commonly admitted in custody cases is otherwise excluded in criminal cases). and compromises the judge’s legally prescribed role as fact-finder. 23 See infra notes 177, 181–182 and accompanying text (arguing fact-finding responsi­bility should remain with court because expert evaluations regarding domestic violence are unreliable).

This Note examines the scientific reliability, admissibility, and impact of expert testimony in custody cases involving domestic violence and how state statutes and evidentiary rules impact the use of such testimony. Part I introduces the relevant legal standards in custody cases and the implementation of DV presumptions as a state response to domestic violence, noting variations in these statutes and in state rules governing custody proceedings. Part II explains the current role of experts in custody cases involving domestic violence, details criticism, contrasts the use of similar experts in criminal cases, and explains how state law facil­itates the admission of problematic expert testimony. Finally, Part III offers an approach to reformulating legal standards with an eye toward scientific realities and raises questions regarding the appropriate role of experts.

I. Domestic Violence in the Child Custody Context

Increased attention to the prevalence and destructive effects of domestic violence 24 This increased attention includes the passage of the Violence Against Women Act of 1994, Pub. L. No. 103-322, tit. IV, 108 Stat. 1902 (codified as amended in scattered sec­tions of 8, 16, 18, 28, and 42 U.S.C.), and developments in social-science research, see Allen M. Bailey, Prioritizing Child Safety as the Prime Best-Interest Factor, 47 Fam. L.Q. 35, 39–40 (2013) (tracing developments in domestic violence research); see also supra note 11 and accompanying text (detailing related congressional action and presidential proclamation). has led state legislatures to mandate the consid­eration of domestic violence in civil child custody cases. 25 All states permit courts to consider evidence of domestic violence in child custody cases, and almost half of states have enacted DV presumptions. Bailey, supra note 24, at 35. As a result, allegations of domestic violence play a central and often decisive role in final custody determinations. 26 See, e.g., James N. Bow & Paul Boxer, Assessing Allegations of Domestic Violence in Child Custody Evaluations, 18 J. Interpersonal Violence 1394, 1395 (2003) (“A child custody evaluator’s opinion about alleged domestic violence can have a profound impact on the ultimate custody decision.”). The evolution of these state statutory schemes has been accompanied by increased court reliance on experts in making custody determinations. 27 See Janet M. Bowermaster, Legal Presumptions and the Role of Mental Health Professionals in Child Custody Proceedings, 40 Duq. L. Rev. 265, 265 (2002) (“Because the best interests standard is not susceptible to traditional legal analysis, judges turned to men­tal health professionals for expertise regarding children and child development.”). Section I.A provides an introduction to the murky best-interests-of-the-child standard upon which custody actions are cast. Section I.B explains state responses to the problem of domestic violence, including the adoption of DV presumptions. Respectively, sec­tions I.C and I.D note variations in the formulation of DV presumptions and the evidentiary rules governing custody proceedings.

A. The Best Interests of the Child

At the heart of every custody case is the enigmatic 28 See, e.g., id. (describing standard as “broad and formless”). best-interests-of-the-child standard. 29 See id. at 267 (characterizing standard as “governing substantive principle in custody adjudications”). This section proceeds by introducing the family court system, the legal standard governing custody proceedings, and the role of the court in custody cases. Recognizing that custody actions provide one avenue by which domestic violence victims may seek pro­tection for their children, this section flags special considerations for cases involving allegations of domestic violence. 30 Victims of abuse may also seek protection through a temporary restraining order, see, e.g., Cal. Fam. Code § 6300 (West Supp. 2015) (providing for issuance of restraining order “to prevent acts of domestic violence” and “provide for a separation of the persons involved”) or in some states, through special family-offense proceedings, see, e.g., N.Y. Fam. Ct. Act § 812 (McKinney 2015) (delineating family-offense proceedings). For exam­ple, New York’s family courts and criminal courts have concurrent jurisdiction over certain family offenses, victims of which may seek orders of protection through family or criminal courts. Id.; see also Idaho Code § 32-1408 (Supp. 2015) (establishing domestic violence courts). This Note refers to “restraining orders” and “protective orders” or “orders of protection” interchangeably.

1. Commencing an Action in Family Court. — Family law cases can dominate state trial court dockets. 31 See Barbara A. Babb, Reevaluating Where We Stand: A Comprehensive Survey of America’s Family Justice Systems, 46 Fam. Ct. Rev. 230, 230 (2008) (finding family law cases compose more than 40% of trial court dockets in New Jersey and Maryland and nearly 60% in Nebraska). For this reason, many states have decided to implement specialized courts with subject matter jurisdiction over family law issues. 32 Id. at 231–33. As of 2006, thirty-seven states had specialized courts to handle family law matters. Id. at 232. “Family law” encompasses, inter alia, divorce, child custody and visitation, alimony and child support, juvenile cases, and domestic violence cases. Id. at 235 n.2. Regarding the function of family courts, the commission responsible for formulating Arizona’s family court system observed that “‘[the family court] must be a person-oriented court, one that makes the law work for the people, rather than merely fitting family prob­lems into a preconceived legal framework.’” Gerald W. Hardcastle, Adversarialism and the Family Court: A Family Court Judge’s Perspective, 9 U.C. Davis J. Juv. L. & Pol’y 57, 75 (2005) (quoting William J. O’Neil & Barry C. Schneider, Recommendations of the Committee to Study Family Issues in the Arizona Superior Court: A Family Court System, 37 Fam. & Conciliation Cts. Rev. 179, 181 (1999)). For an overview of the shortcomings of the family court system, see generally David J. Lansner, Abolish the Family Court, 40 Colum. J.L. & Soc. Probs. 637, 637–38 (2007) (lamenting “inferior[ity]” of establishment, reputa­tion, and function of New York’s family court system). States have accomplished this either by creating family court systems authorized by statute to hear family law cases 33 See Babb, supra note 31, at 232–33 (listing states implementing such schemes); see also, e.g., New York City Family Court, N.Y. State Unified Court Sys., http:// [] (last updated Jan. 3, 2013) (describing establishment of New York family court system).
or by electing to hear family law cases in separate divisions of courts of general jurisdiction. 34 See, e.g., Family Courts, Phila. Cts.: First Jud. District of Pa., http://www.courts. [] (last visited Oct. 29, 2015) (“The Family Division . . . is one of the three major divisions of the Court of Common Pleas.”). Just thirteen states continue to hear family law cases in general civil trial courts without any form of specialized system. Babb, supra note 31, at 233. Advocates of specialized courts emphasize the need for consistent management of the family law docket, experienced judges, user-friendly procedures, and access to social-ser­vice resources. Id. at 231–32. But as this Note explains, expert evidence admitted by family courts differs from similar evidence offered in criminal courts. See infra section II.B (explaining contrast).

A parent may begin the process of modifying child custody by filing an application in family court, 35 See, e.g., Cal. Fam. Code § 3120 (West 2004) (authorizing actions for exclusive custody). either during a divorce or separation proceeding or in a separate custody action. 36 See, e.g., id. § 3021(a) (dissolution of marriage); id. § 3021(c) (legal separation); id. § 3021(d) (action for exclusive custody). After initiating such a proceeding, if the parties are unable to settle the dispute through arbitration or some other means, 37 Due to growing caseloads, family courts have increasingly encouraged various forms of alternative dispute resolution as alternatives to litigation. John Lande, The Revolution in Family Law Dispute Resolution, 24 J. Am. Acad. Matrim. L. 411, 414–16 (2012). The use of mediation has greatly expanded, especially for cases involving child custody. Id. Supporters of alternative dispute resolution for custody cases cite the harmful collateral effects of litigation’s adversarial nature on children. See Gregory Firestone & Janet Weinstein, In the Best Interests of Children, 42 Fam. Ct. Rev. 203, 203–07 (2004) (outlining faults of adversarial system and harmful effects on children). But cf. Jana B. Singer, Dispute Resolution and the Postdivorce Family: Implications of a Paradigm Shift, 47 Fam. Ct. Rev. 363, 367 (2009) (noting overburdened system means “courts may have difficulty meeting both their basic adjudicative functions and the broader, more ambitious goals of the new family conflict resolution paradigm”). the case proceeds to an adversarial custody hearing. 38 Though “adjudication is relatively rare . . . [l]itigation involving parents who are in deep conflict . . . can take substantial time and resources, both for the family and the court.” Ira Mark Ellman et al., Family Law: Cases, Text, Problems 358 (abr. 5th ed. 2014). Notably, where there are allegations of domestic violence, cases may be subject to specialized screening procedures, regardless of whether parties proceed through medi­ation or the adversarial process. See generally Loretta Frederick, Questions About Family Court Domestic Violence Screening and Assessment, 46 Fam. Ct. Rev. 523 (2008) (describing purposes and goals of domestic violence screening). Though mediation and other alternatives to the adversarial process are beyond the scope of this Note, the issues raised here no doubt influence these processes, including what evidence may become admissible should the case proceed to trial. Courts hear custody cases involving domestic violence in the traditional adversarial process more often than other custody cases, “in part because of an understandable view that these allegations should be adjudicated” to protect victims. 39 Scott & Emery, supra note 16, at 96; see also Linda C. Nielson, At Cliff’s Edge: Judicial Dispute Resolution in Domestic Violence Cases, 52 Fam. Ct. Rev. 529, 538–39 (2014) (explaining when alternate forms of dispute resolution may be unsafe for parties). The adjudication itself involves the resolution of two key issues: legal custody and physical custody—in other words, “which parents are responsible for making decisions about children and where the children spend their time.” 40 Lande, supra note 37, at 423. Custody cases are intensely fact-specific, requiring careful weighing of evidence by judges. 41 See, e.g., Dale & Gould, supra note 3, at 7 (“Complex custody disputes almost always involve not just one theory or one question, but multiple questions and competing theories about highly disputed facts.”).

2. The Best-Interests-of-the-Child Standard. — Resolution of child cus­tody disputes depends on a holistic evaluation of the “best interests of the child.” 42 See, e.g., N.Y. Dom. Rel. Law § 240(1)(a) (McKinney 2015) (“In any action or proceeding brought . . . to obtain . . . the custody of or right to visitation with any child of a marriage . . . justice requires, having regard to . . . the best interests of the child . . . .”). The best-interests standard was originally implemented through the lens of maternal preference until the gradual decline of gender-based preferences in the latter third of the twentieth century. 43 Am. Law Inst., Principles of the Law of Family Dissolution: Analysis and Recommendations, 8 Duke J. Gender L. & Pol’y 1, 2 (2001). For a history of this decline and a detailed account of the changes in custody criteria from the 1700s to the late 1900s, see generally Jacob Goldstein & C. Abraham Fenster, Anglo-American Criteria for Resolving Child Custody Disputes from the Eighteenth Century to the Present: Reflections on the Role of Socio-Cultural Change, 19 J. Fam. Hist. 35, 36–42 (1994). Today this assessment generally requires consideration of both retrospective factors—such as a parent’s past conduct—and prospective factors—such as the probability of a continuing, positive parent–child relationship. 44 See, e.g., Fla. Stat. Ann. § 61.13(3)(a) (West Supp. 2015) (requiring retrospective consideration of history of domestic violence and prospective capacity of parents to “facilitate and encourage a close and continuing parent–child relationship”). Some factors may invite a holistic evaluation of both past conduct and future probabilities. For example, California requires consideration of “[t]he health, safety, and welfare of the child,” so far as they pertain to the best interests of the child; presumably this involves an ex post assessment of parental conduct and parents’ ability to ensure a child’s safety. See Cal. Fam. Code § 3011(a) (West 2004). For an example of a best-interests statute listing factors for court consideration, see Minn. Stat. Ann. § 518.17, subdiv. 1 (West 2004). State policy interests motivate the substance of the statutes. 45 See infra sections I.B–C, II.A.2 (discussing DV presumptions and legislative guidance for judicial decisionmaking). But even where statutes require a court to consider certain factors in its analysis, a court may usually consider any factors it deems relevant to the child’s best interests, again highlighting the amorphous nature of the best-interests inquiry. 46 E.g., Cal Fam. Code § 3011 (providing court may consider “any other factors it finds relevant”); see Ellman et al., supra note 38, at 358 (explaining statutes provide courts “no direction about rank ordering or about the weight that should be attached to any particular factor”).

Scholars have thoroughly ridiculed the best-interests standard and blame it for increased reliance on expert testimony in custody pro­ceedings. 47 See Scott & Emery, supra note 16, at 69 & n.1, 100 (noting academic criticism and arguing for replacement of best-interests standard with approximation standard based on past caretaking); infra notes 239–242 and accompanying text (discussing criticism of standard with regard to reliance on experts). Criticism underlying the best-interests standard includes the difficulty of accurately weighing the numerous relevant factors and the inevitable speculation and indeterminacy accompanying this task. 48 See Jon Elster, Solomonic Judgments: Against the Best Interest of the Child, 54 U. Chi. L. Rev. 1, 7 (1987) (arguing best-interests standard “is indeterminate, unjust, self-defeating, and liable to be overridden by more general policy considerations”). Describing the standard as “uniformly disparaged,” the American Law Institute (ALI) in its 2002 recommendations noted that “the unpre­dictability of results [under the best-interests standard] encourages parents to engage in strategic behavior, take their chances in litigation, and hire expensive experts to highlight each other’s shortcomings rather than work together to make the best of the inevitable.” 49 Am. Law Inst., supra note 43, at 2. For these reasons, alternate standards have been proposed, including the approximation standard, which the ALI has adopted. 50 See Elizabeth S. Scott, Pluralism, Parental Preference, and Child Custody, 80 Calif. L. Rev. 615, 617 (1992) (“[T]he law’s goal should be to approximate . . . the predivorce role of each parent in the child’s life.”); see also Scott & Emery, supra note 16, at 69, 101–04 (“[A]pproximation mirrors the underlying policy goals of custody law at least as well as do any of the psychological and emotional factors that currently serve as proxies for best interests. Basing custody on past parental care promotes continuity and stability in the child’s environment and relationships . . . .”). For further explanation of the approximation rule as adopted by ALI, see Ellman et al., supra note 38, at 392–95. The approx­imation standard looks solely to past parental practices, departing from the best-interests standard, which, approximation advocates argue, “masks the importance of the parents’ role in caring for the child during the marriage” and leaves custody to the whims of judges. 51 Scott, supra note 50, at 616. Past parental practices considered by the approx­imation rule include “the amount of time spent with the child, the extent to which the parent engaged in tasks that contributed to the child’s basic care and development, and the parent’s participation in decisions relevant to the child.” Id. at 637–38. West Virginia is currently the only state to employ the approximation rule. Katharine T. Bartlett, Prioritizing Past Caretaking in Child-Custody Decisionmaking, 77 Law & Contemp. Probs., no. 1, 2014, at 29, 46; see W. Va. Code Ann. § 48-9-206(a) (LexisNexis 2014) (mandating custody reflect “proportion of time each parent spent performing caretaking functions for the child prior to the parents’ separation”). It is in this complicated and highly debated context that one must evaluate eviden­tiary issues and the proper role of the judiciary in making custody determinations.

3. The Role of the Judge. — The difficult task 52 As Professor Jon Elster so aptly put it, “[K]nowledge that the [custody] decision will have momentous importance for the parties directly involved and the recognition that it may not be possible to have a rational preference for one parent over the other[] conspire to create a psychological tension in decision makers that many will be unable to tolerate.” Elster, supra note 48, at 2. of deciding a life-changing custody dispute usually falls to one individual—the judge. 53 Outside of Texas, a custody litigant generally does not have the option to present her case before a jury. See Anthony Champagne et al., Are Court-Appointed Experts the Solution to the Problems of Expert Testimony?, 84 Judicature 178, 180 (2001) (“Unlike other jurisdictions, family law cases can be heard by a jury in Texas.”); cf. Melissa L. Breger, Introducing the Construct of the Jury into Family Violence Proceedings and Family Court Jurisprudence, 13 Mich. J. Gender & L. 1, 2 & n.3 (2006) (“[I]n family violence proceedings in Family Court, the majority of the fifty states do not permit juries.”). Approximately 10% of custody cases are litigated, with most cases resolved privately or in mediation. Allison M. Nichols, Note, Toward a Child-Centered Approach to Evaluating Claims of Alienation in High-Conflict Custody Disputes, 112 Mich. L. Rev. 663, 664 (2014). The judge must not only engage in the highly subjective process of fact-finding, 54 See Lande, supra note 37, at 438 (describing custody factors as “quite subjective”). but she must also consider a host of complicated factors. 55 See Dale & Gould, supra note 3, at 7 (noting complicated nature of factors). The role of courts has been described as “rehabilitative or problem-solving,” with a single judge managing a case from start to finish, “adjudicating legal rights” along the way. 56 Lande, supra note 37, at 431–32. Systems of unified family courts allow “a single judge and professional team [to] deal with all issues of a particular family.” Id. at 432. Given the volume and complexity of relevant factors, courts have increasingly opted to appoint experts to aid in the decisionmaking process. 57 See id. at 438 (discussing frequent court appointment of experts); see also infra note 124 (discussing role of experts in custody disputes involving domestic violence). While some have welcomed this development because they are skeptical of judges’ abilities to make accurate credibility assessments in custody cases, 58 See Dana Harrington Conner, Abuse and Discretion: Evaluating Judicial Discretion in Custody Cases Involving Violence Against Women, 17 Am. U. J. Gender Soc. Pol’y & L. 163, 171–74 (2009) (arguing judges are not especially qualified to evaluate credibility and noting skills of experts); see also Breger, supra note 53, at 2 (arguing for introducing juries into civil family-violence proceedings to foster legitimacy). others have lamented this trend as an abdication of the court’s fact-finding function. 59 See infra section II.A.3 (explaining this criticism). On the other hand, the case for wide judicial discretion points to custody judges’ experience in the field and superior ability to assess witness credibility. See Conner, supra note 58, at 170–76 (summarizing arguments in support of strong role for judges in custody decisions). Though this notion has been challenged, see, e.g., id. at 176–78 (noting gender bias of courts), trial court decisions continue to receive extremely high deference on appeal. See id. at 183 (“[U]nless the ‘contradictory evidence is beyond belief,’ the appellate court will defer to the findings of the trial judge.” (quoting Lampe v. Lampe, 28 S.W.2d 414, 415 (Mo. Ct. App. 1930))); see also, e.g., In re Marriage of Burgess, 913 P.2d 473, 478 (Cal. 1996) (applying “deferential abuse of discretion test,” which asks “whether the trial court could have reasonably concluded that the order in question advanced the ‘best interest’ of the child;” and mandating courts must uphold rulings if “correct on any basis, regardless of whether such basis was actually invoked”). Though courts generally have wide discretion in deciding custody matters, 60 See Conner, supra note 58, at 169 (“Due to the great deference afforded to the custody trial judge, these particular judges enjoy a position of authority over decision-making unseen in other types of cases.” (footnote omitted)). many states have constrained this discretion by statute.

B. The Rebuttable Presumption Against Awarding Custody to Abusers

Following increased public attention to domestic violence in the 1990s, 61 See supra notes 11, 24 and accompanying text (describing government response to domestic violence in 1990s). many states responded with statutory protections to discourage awarding custody to abusers. 62 See Jaffe et al., Common Misconceptions, supra note 13, at 63–65 (describing legislative response to domestic violence). One popular response was the imple­mentation of legal presumptions against awarding custody to parents found to have committed domestic violence offenses (DV presumptions), in an effort to direct judicial decisionmaking. 63 See infra section I.B.1 (discussing motivations behind adoption of DV presumptions).

1. The History of DV Presumptions. — States may advance their policy goals by guiding the judicial inquiry with legal presumptions. 64 See generally J. Herbie DiFonzo, From the Rule of One to Shared Parenting: Custody Presumptions in Law and Policy, 52 Fam. Ct. Rev. 213, 214–15 (2014) (tracing development of legal presumptions in child custody context). Recognizing the dangers domestic violence poses to children, twenty-four states and the District of Columbia have implemented some form of rebuttable presumption against awarding custody to a party found to have perpetrated domestic violence. 65 Comm’n on Domestic & Sexual Violence, ABA, Child Custody and Domestic Violence by State (2014),
/domestic_violence1/Charts/2014%20Custody%20DV%20Chart.authcheckdam.pdf []. But see, e.g., Neb. Rev. Stat. § 42-364(2) (Supp. 2014) (explicitly providing that no presumption exists).
States that have not enacted DV presumptions nearly universally mandate that courts consider domestic violence in custody determinations. 66 Charts, 45 Fam. L.Q. 492, 494 chart 2 (2012). For example, though the legislature urged courts to ensure evidence of domestic violence is a “weighty consideration” in the best-interests-of-the-child calculus, New York does not have a DV presumption. See 1996 N.Y. Laws ch. 85, § 2 (“[D]omestic violence should be a weighty consideration in custody and visitation cases.”); Jennifer Jack, Note, Child Custody and Domestic Violence Allegations: New York’s Approach to Custody Proceedings Involving Intimate Partner Abuse, 5 Alb. Gov’t L. Rev. 885, 897–900 (2012) (providing overview of domestic violence’s impact on New York custody proceedings). Instead, it includes domestic violence among several factors to be considered by the trial court in making custody determinations. N.Y. Dom. Rel. Law § 240(1) (McKinney 2015). Even in jurisdictions where a finding of domestic violence does not carry presumptive effect, a finding of domestic violence often has determinative impacts on custody proceedings. 67 See, e.g., Joan S. Meier, Domestic Violence, Child Custody, and Child Protection: Understanding Judicial Resistance and Imagining the Solutions, 11 Am. U. J. Gender Soc. Pol’y & L. 657, 662–63 (2003) (describing increasing influence of domestic violence on custody cases). Many have advocated adopting DV presumptions in states without them, arguing that these states fail to adequately account for the effects of violence on victims. 68 See, e.g., Lynne R. Kurtz, Protecting New York’s Children: An Argument for the Creation of a Rebuttable Presumption Against Awarding a Spouse Abuser Custody of a Child, 60 Alb. L. Rev. 1345, 1345–46 (1997) (arguing for adoption of DV presumption); Jack, supra note 66, at 887–89, 912–13 (same).

2. Function of the DV Presumption. — DV presumptions shape the best-interests inquiry by prioritizing domestic violence over other relevant factors. 69 See Bowermaster, supra note 27, at 273–74 (positing presumptions set aside certain cases “from the unguided discretion of the best interests standard and mark[] [them] for special treatment”). Once a party convinces the court that the opposing party has perpetrated domestic violence, DV presumptions then shift the burden to the perpetrator, who must show that it is within the child’s best interests that the perpetrator retain custody rights, notwithstanding the violence. 70 E.g., Cal. Fam. Code § 3044 (West 2004) (“Upon a finding . . . that a party seeking custody of a child has perpetrated domestic violence[,] . . . there is a rebuttable presumption that an award of . . . custody of a child to a person who has perpetrated domestic violence is detrimental to the best interest of the child . . . .”). Some states require a showing of several enumerated factors before the presumption may be rebutted. See, e.g., Ariz. Rev. Stat. Ann. § 25-403.03(E) (West Supp. 2015) (requiring courts to consider “all” of six different factors in determining whether parent has rebutted presumption). Several states have mandatory, unrebuttable presumptions in cases of extreme violence. See, e.g., Ky. Rev. Stat. Ann. § 403.325 (LexisNexis 2010) (removing visitation rights from parent found guilty of killing other parent). In other words, a party found to have committed abuse must successfully dispute the now-presumed finding that he is not a suitable custodian. 71 See Bowermaster, supra note 27, at 275–76 (“Presumptions that shift the burden of proof are treated the same as presumptions that shift only the production burden, until the opposing party introduces evidence disputing the presumed fact.”). The relative ease of obtaining an initial finding of domestic violence or rebutting a finding of domestic violence depends on the DV presumption statute itself, including its definition of domestic violence 72 See infra section I.C.1 (noting variations in definitions). and the respective burdens of proof for invoking and rebutting the presumption. 73 See infra section I.C.2 (explaining differences). Statutes may require a court to explicitly state whether it has found domestic violence, whether it has applied a DV presumption, and if relevant, whether the perpetrator of domestic violence has rebutted the presumption. 74 See, e.g., F.T. v. L.J., 123 Cal. Rptr. 3d 120, 141–42 (Ct. App. 2011) (remanding case to trial court for explicit findings regarding application of DV presumption); see also N.Y. Dom. Rel. Law § 240(1)(a) (McKinney 2015) (requiring court to “state on the record” how “findings, facts and circumstances” of domestic violence factor into custody decisions). The DV presumptions implemented across the nation are not uniform, however. And each statute’s unique structure naturally influences the solicitation and consideration of expert testimony. 75 See infra section II.C (arguing formulation of DV presumptions encourages reliance on flawed expert testimony).

C. Statutory Variation Between States

Commentators have noted the differing formulations of DV presumptions. 76 E.g., Conner, supra note 58, at 198–99 (discussing several categories of DV presumptions). How a statute defines domestic violence—and the extent to which it provides enumerated factors relevant to meeting that definition—may affect the type of evidence offered by the parties to demonstrate the elements necessary to invoke the presumption. The specified burden of proof also affects how easily a presumption may be invoked. 77 For example, convincing a judge that it is more likely than not that domestic violence has occurred is much less onerous than proving the same beyond a reasonable doubt. Compare, e.g., Cal. Fam. Code § 3044(a) (West 2004) (preponderance of evidence), with Nev. Rev. Stat. Ann. § 125C.230 (LexisNexis 2010) (clear and convincing evidence), and Baker v. Long, 981 A.2d 1152, 1157 (Del. 2009) (requiring showing of criminal conviction, effectively imposing burden of beyond a reasonable doubt). Finally, some statutes constrain the extent to which courts may rely on expert testimony. 78 See infra notes 91–92 and accompanying text (providing examples).

1. The Definition of “Domestic Violence.” — Defining domestic violence involves two main considerations: what actions constitute “domestic violence” and the degree of violence that must be proven to invoke the presumption. DV presumptions sometimes incorporate by reference a definition of domestic violence from state penal law or a body of domestic violence law; 79 E.g., Fla. Stat. Ann. § 61.13(2)(c)(2) (West Supp. 2015) (“Evidence that a parent has been convicted of a misdemeanor of the third degree or higher involving domestic violence, as defined in s. 741. 28 and chapter 775, or meets the criteria of s. 39.806(1)(d), creates a rebuttable presumption of detriment to the child.”). otherwise the statute may include its own definition of domestic violence for the purposes of the presumption that may differ from other iterations of the term in the state’s laws. 80 For example, California’s definition follows:
“[A] person has “perpetrated domestic violence” when he or she is found . . . to have intentionally or recklessly caused or attempted to cause bodily injury, or sexual assault, or to have placed a person in reasonable apprehension of imminent serious bodily injury to that person or to another, or to have engaged in any behavior involving, but not limited to, threatening, striking, harassing, destroying personal property or disturbing the peace of another . . . .”
Cal. Fam. Code § 3044(c); see Conner, supra note 58, at 198–99 (explaining differences in statutory formulation of DV presumptions). This and other statutory definitions of domestic violence are fairly characterized as “incident-based,” rather than a “clinical model [that] focuses on how domestic violence plays a role in the relationship as opposed to a snapshot of what just occurred.” Id. at 220.

How a state defines domestic violence affects the type of evidence relevant to prevail on a claim of abuse. 81 See infra section II.C (providing examples of statutory definitions). In some states, the presumption may be invoked upon a court’s finding of a single incident of domestic violence. 82 See, e.g., Ariz. Rev. Stat. Ann. § 25-403.03(D) (Supp. 2015) (requiring finding of “an act of domestic violence”); Nev. Rev. Stat. Ann. § 125C.230 (“one or more acts”). Other states require a pattern or history of abuse. 83 E.g., Ark. Code Ann. § 9-13-101(c)(2) (Supp. 2015) (“pattern of domestic abuse”); Idaho Code Ann. § 32-717B(5) (West 2006) (“habitual perpetrator”); Iowa Code Ann. § 598.41(2)(c) (West Supp. 2015) (“history of domestic abuse”); Tex. Fam. Code Ann. § 153.004(b) (West 2014) (“history or pattern” of neglect or abuse). Yet other states require either a history of abuse or a single “serious” incident of abuse. 84 See Alaska Stat. § 25.24.150(g)–(h) (2014) (requiring finding of pattern of abuse, also met by showing one incident resulting in serious physical injury); Mass. Gen. Laws Ann. ch. 209C, § 10(c) (LexisNexis 2011) (same); Miss. Code Ann. § 93-5-24(9)(a)(i) (2013) (same); N.D. Cent. Code § 14-09-06.2(j) (Supp. 2015) (same). Even if a single incident of abuse is well documented, it will be relatively harder to trigger a DV presumption if its terms require a showing of a “history” of domestic violence or instances of “serious” domestic violence. 85 E.g., Ark. Code Ann. § 9-13-101(c)(2) (invoking DV presumption upon finding “parent has engaged in a pattern of domestic abuse”); La. Stat. Ann. § 9:364(A) (Supp. 2015) (invoking DV presumption upon finding “history of perpetrating family violence,” which may be shown by “one incident of family violence . . . result[ing] in serious bodily injury” or “more than one incident of family violence”); see also Jack, supra note 66, at 911 (“More often than not, the existence of this presumption does nothing more than ‘provide the illusion of protection for women and children, but also build obstacles that few victims have any hope of overcoming.’” (quoting Conner, supra note 58, at 200)). These definitional variations are one of several differences among state DV presumption statutes.

2. Statutory Guidance for Decisionmaking. — Statutes also vary in the extent to which they guide parties and courts and in setting the burden of proof. Trial courts generally have wide discretion to decide how much weight to afford each piece of evidence. 86 See supra note 60 and accompanying text (noting extent of trial court’s discretion). But Arizona, among other states, requires courts to consider an enumerated list of factors when determining whether domestic violence has occurred, such as previous court findings, medical and police reports, and family-services and school records. 87 Ariz. Rev. Stat. Ann. § 25-403.03(C). The statute requires consideration of the following factors: “1. Findings from another court of competent jurisdiction. 2. Police reports. 3. Medical reports. 4. Records of the department of child safety. 5. Domestic violence shelter records. 6. School records. 7. Witness testimony.” Id. States may also specify factors to be considered in rebuttal of a DV presumption—such as completion of batterer’s treatment programs, substance abuse programs, and compliance with parole and restraining-order conditions. 88 Cal. Fam. Code § 3044 (West 2004). These statutes highlight states’ ability to control the type of evidence considered in custody cases, but the extensiveness of factors varies widely by state: Some DV presumption statutes offer no guidance at all beyond setting forth the “domestic violence” definition and the standard of proof courts should apply. 89 E.g., Ark. Code Ann. § 9-15-215(c) (LexisNexis 2009) (“There shall be a rebuttable presumption that it is not in the best interest of the child to be placed in the custody of an abusive parent in cases in which there is a finding by a preponderance of the evidence that a pattern of abuse has occurred.”). Variations in the burden of proof by which a party must either prove an allegation of domestic violence or by which a party must rebut a finding of domestic violence range from “preponderance of the evidence” to “credible evidence” to “clear and convincing evidence,” 90 Most typically, the burden to prove or overcome the presumption is a preponderance of the evidence. E.g., Cal. Fam. Code § 3044(a). But see Tex. Fam. Code Ann. § 153.004(b) (West 2014) (imposing DV presumption upon finding of “credible evidence” of domestic violence). Some statutes require plaintiffs to prove domestic violence by clear and convincing evidence, e.g., Nev. Rev. Stat. Ann. § 125C.230 (LexisNexis 2010), and others bolster DV presumptions by raising the burden of proof once it has shifted to the other party, see N.D. Cent. Code § 14-09-06.2(1)(j) (Supp. 2015) (requiring showing by preponderance of evidence to invoke DV presumption, which “may be overcome only by clear and convincing evidence”); see also Bowermaster, supra note 27, at 287 (explaining North Dakota’s adoption of increased burden of proof for rebutting party only). representing yet another set of statutory variations.

Regarding expert testimony, statutes constraining a court’s factual inquiry are especially rare. Notably, California limits the extent to which courts may rely on expert testimony in reaching a finding of domestic violence, thereby prohibiting courts from making findings of domestic violence on the sole basis of expert testimony. 91 Cal. Fam. Code § 3044(e) (mandating consideration of “any relevant, admissible evidence submitted by the parties” in addition to any expert testimony). A similar rule applies in cases for terminating parental rights under federal tribal law, providing courts may not rely solely on expert testimony in reaching their decisions. 92 See 25 U.S.C. § 1912(f) (2012) (requiring consideration of evidence “including testimony of qualified expert witnesses” (emphasis added)); In re Adoption of Hannah S., 48 Cal. Rptr. 3d 605, 613 (Ct. App. 2006) (finding court’s exclusive reliance on expert testimony was contrary to meaning of 25 U.S.C. § 1912 and legislative intent). Such rules demonstrate reluctance, or at least formalistic reluctance, to permit an expert’s conclusion alone to determine the outcome of a case. These rules constraining consideration of expert testimony provide but one more example of statutory mandates used to guide courts’ decisionmaking. Further variance exists in state evidentiary rules, as explained below. This Note marshals these idiosyncrasies to point out both potential flaws and constructive practices, taking into account the shortcomings of expert evidence. 93 See infra section III.A (calling for revisions to statutes and rules).

D. State Evidentiary Rules and Consideration of Previous Judgments

Some states have enacted special rules of evidence for certain family court proceedings, 94 See infra note 104 and accompanying text (describing special rules of evidence for family-offense proceedings in New York). while others have not. 95 See, e.g., Tex. Fam. Code Ann. § 104.001 (providing general state evidentiary rules also govern custody cases). The relevance of prior findings of domestic violence in a subsequent custody proceeding also differs. 96 See infra section II.B (explaining differences). Evidentiary rules and res judicata principles together pose the risk that courts may subject evidence of domestic violence to disparate scrutiny, depending on the timing and nature of the proceeding. 97 See Janet R. Johnston et al., Allegations and Substantiations of Abuse in Custody-Disputing Families, 43 Fam. Ct. Rev. 283, 291 (2005) (“Somewhat different judicial and administrative philosophies affect the processing of cases in these different systems and fact-finding is subject to different standards of proof.”).

A prior domestic violence restraining order may constitute a relevant or even dispositive factor in determining the application of a DV presumption. 98 California appears to be one such example where restraining orders have a res judicata-like effect. See, e.g., Christina L. v. Chauncey B., 177 Cal. Rptr. 3d 178, 182 (Ct. App. 2014) (recognizing issuance of restraining order under Domestic Violence Prevention Act automatically invokes DV presumption). But see Toolan-Miller v. Yates, No. D062129, 2013 WL 1808727, at *8–9 (Cal. Ct. App. Apr. 30, 2013) (finding reliance on previously issued restraining order to invoke DV presumption was matter of “discretion”).
Other states seem more hesitant to recognize such a res judicata effect for restraining orders. For example, the defendant in a case for civil damages arising from an assault successfully argued that a related restraining order should not have res judicata effect; the court noted the “summary nature” of restraining-order proceedings and differences in the requisite burden of proof. L.T. v. F.M., 102 A.3d 398, 404 (N.J. Super. Ct. App. Div. 2014). Though this case did not involve the use of a restraining order in support of invoking a subsequent finding of domestic violence in a custody matter, the concerns mentioned by the court are also relevant in the custody context.
In states where prior findings of domestic violence are a relevant but not dispositive consideration the existence of a prior finding of domestic violence does not automatically trigger the DV presumption. See, e.g., Tex. Fam. Code Ann. § 153.004(f) (West 2014) (“In determining . . . whether there is credible evidence of a history or pattern of . . . abuse[,] . . . , the court shall consider whether a protective order was rendered . . . against the parent during the two-year period preceding the filing of the suit or during the pendency of the suit.”). The Texas House of Representatives noted that before this provision was added in 2003, it was “unclear whether the issuance of a family violence protective order [could] constitute credible evidence of a history or pattern of domestic violence for child custody purposes.” H. Comm. on Juvenile Justice & Family Issues, Bill Analysis, H. 78-2099, Reg. Sess., at 1 (Tex. 2003).
Such a finding can have a res judicata-like effect on a subsequent custody proceeding. 99 See supra note 98 and accompanying text (noting example where prior restraining order automatically invokes DV presumption). But the rules of evidence are generally lax in proceedings for restraining orders. 100 Cf. Glenna Goldis, When Family Courts Shun Adversarialism, 18 U.C. Davis J. Juv. L. & Pol’y 195, 233–35 (2014) (explaining lax application of evidentiary rules during hearings for temporary guardianship orders, including admission of speculative and hearsay testimony). The standard sufficient to issue a restraining order may be lower than the standard to invoke a DV presumption. 101 Compare, e.g., Cal. Fam. Code § 6300 (West Supp. 2015) (permitting issuance of restraining order upon showing of “reasonable proof of a past act or acts of abuse”), with Cal. Fam. Code § 3044(a) (West 2004) (requiring “a finding . . . of domestic violence” by preponderance of evidence to trigger DV presumption). New York is an example of a state with special evidentiary rules for domestic violence civil cases. 102 Though New York does not have a DV presumption, domestic violence is an important factor in its best-interests analysis. For an overview of how domestic violence influences New York custody proceedings, see generally Jack, supra note 66. Instead of a DV presumption, New York includes domestic violence as one of several factors to be considered by the trial court in making custody determinations. See N.Y. Dom. Rel. Law § 240(1) (McKinney 2015) (mandating consideration of domestic violence “together with such other facts and circumstances as the court deems relevant”). Though the legislature made explicit findings condemning domestic violence and urged courts to ensure the existence of domestic violence has a “weighty consideration” in the best-interest-of-the-child calculus, it did not go as far as to instate a legal presumption. Kim Susser, Weighing the Domestic Violence Factor in Custody Cases: Tipping of the Scales in Favor of Protecting Victims and Their Children, 27 Fordham Urb. L.J. 875, 877–78 (2000) (quoting Act of May 21, 1996, ch. 85, 1996 N.Y. Laws 273, 273). New York’s family courts and criminal courts have concurrent jurisdiction over certain “family offenses,” victims of which may seek restraining orders through family or criminal courts. 103 See N.Y. Fam. Ct. Act § 812 (McKinney 2015) (establishing concurrent criminal and family court jurisdiction over certain criminal conduct between members of same family or household). This is one of several relatively recent developments in New York’s response to domestic violence offenses, including the 2014 creation of the crime of “aggravated family offense,” bumping up repeat misdemeanor domestic violence convictions to felonies. See N.Y. Penal Law § 240.75 (McKinney 2015) (“A person is guilty of aggravated family offense when he or she commits a [specified] misdemeanor . . . and he or she has been convicted of one or more specified offenses within the immediately preceding five years.”). In a family-offense action, the rules of evidence are relaxed, as exemplified by hearsay exceptions for statements made by children. 104 See, e.g., N.Y. Fam. Ct. Act § 1046(a)(vi) (“[P]revious statements made by the child relating to any allegations of abuse or neglect shall be admissible in evidence . . . .”); id. § 1046(a)(vii) (eliminating certain forms of privilege in family-offense hearings). Relaxed evidentiary standards and lower burdens of proof for restraining orders combined with the relative persuasiveness of restraining orders call into question the use of prior court findings to trigger DV presumptions. This is problematic where a court admits evidence of prior findings of domestic violence, but the underlying evidence in obtaining that prior finding would not be admissible in the current proceeding: A party may be wholly unable to challenge the underlying bases of the previous finding. 105 See infra section II.C.3 (explaining possibility of piggybacking expert evidence into custody proceedings).

The type and amount of evidence a party might put forth in support of a claim of domestic violence necessarily depend on the statutory elements that must be proven, the burden of production, the burden of proof, and relevant procedural rules. These elements also influence a party’s decision to offer expert testimony and the nature of the testimony solicited. The next Part examines the dynamics of this critical interplay.

II. State Law Effects on Admissibility of Expert Testimony in Custody Cases

The impact, admissibility, and reliability of expert opinions in custody cases vary depending on state formulations of DV presumptions and rules of evidence and procedure. While current scholarship posits the incompatibility of the best-interests standard 106 See generally supra section I.A.2 (providing background on best-interests standard). with child-custody expert evaluations, 107 See infra notes 239–241 and accompanying text (explaining this critique). this Part submits that the formulation of custody statutes and limited scrutiny of experts work together to cause judicial overreliance on questionable expert testimony. Section II.A introduces the role of experts in custody disputes involving domestic violence and the shortcomings of mental health experts in this arena. Section II.B shows how experts in custody cases are subjected to limited scrutiny by courts, contrasting the use of similar testimony in criminal cases. Finally, section II.C identifies evidentiary loopholes legislatures and courts have created, making it possible to forego a thorough inquiry into the foundation of expert testimony in custody cases.

A. The Role of Experts in Custody Cases Involving Domestic Violence

The introduction of the best-interests standard has resulted in increased reliance on mental health experts. 108 See Daniel W. Shuman, The Role of Mental Health Experts in Custody Decisions: Science, Psychological Tests, and Clinical Judgment, 36 Fam. L.Q. 135, 135 (2002) (“[T]he standard for judicial determination of child custody arrangements has precipitated a flood of expert mental health testimony . . . .”); see also id. at 136 (noting writings that heavily influenced adoption of best-interests standard gave experts “societal mandate for involvement” in custody cases (internal quotation marks omitted) (quoting Robert Henley Woody, Behavioral Science Criteria in Child Custody Determinations, 3 J. Fam. & Marriage Couns. 11, 11 (1977))). By bringing to light the complicated, destructive consequences of abuse, mental health and social science experts can be essential in proving domestic violence allegations, which are often difficult to substantiate. 109 See Jane H. Aiken & Jane C. Murphy, Evidence Issues in Domestic Violence, 34 Fam. L.Q. 43, 45–52 (2000) (noting importance of experts in explaining effects of battering). Many states have recognized the uniqueness of crimes involving domestic violence by modifying their rules of evidence to permit otherwise inadmissible testimony in these cases, relaxing standards for the admissibility of character evidence to allow prosecutors to demonstrate patterns of domestic violence, which some state legislatures have described as “cyclical,” e.g., Colo. Rev. Stat. Ann. § 18-6-801.5 (West 2013); this puts prior bad acts involving domestic violence in a separate category from other evidence that might be excluded as impermissible propensity evidence. See, e.g., Cal. Evid. Code § 1109 (West 2009) (creating exemption from rule prohibiting propensity evidence for prior acts of domestic violence); see also Erin R. Collins, The Evidentiary Rules of Engagement in the War Against Domestic Violence, 90 N.Y.U. L. Rev. 397, 412–22 (2015) (presenting exceptions to character evidence rules for domestic violence and arguing against this trend). Custody recom­mendations and opinions as to whether domestic violence has occurred include particularized evaluations of the parties involved. 110 See generally Am. Psychol. Ass’n, Guidelines for Custody Evaluations in Family Law Proceedings, 65 Am. Psychology 863, 863–67 (2010) (establishing practice guidelines for experts in child custody matters); Emery et al., supra note 16, at 6–12 (using hypothetical to explain common practices and offering criticism of these techniques). Although experts face unique difficulties in custody cases involving allegations of domestic violence, their opinions and recommendations often prove highly influential in a court’s final decision.

1. Methodological Inadequacies and the Scope of Expert Testimony. — Courts follow an expert’s custody recommendation up to 90% of the time. 111 Spiegel, supra note 9. Evaluations are usually court-ordered 112 See Bow & Boxer, supra note 26, at 1401 (finding more than 90% of child custody referrals among mental health professionals are court-ordered). and performed by a mental health professional, such as a psychologist. 113 The extent of training of mental health professionals varies. See, e.g., id.; Dale & Gould, supra note 3, at 9 (“Not every evaluator is qualified to offer services in every situation.”). A typical evaluation involves psychological testing and interviews of the children, both parents, and persons directly involved in the care of the children. 114 See Emery et al., supra note 16, at 6 (providing hypothetical custody evaluation). The expert conducts additional interviews of teachers and family members, a review of school and medical records, and an examination of documents related to the litigation. 115 Id.

An expert’s resulting reportand subsequent testimony where a case proceeds to trial 116 Even if a case does not proceed to trial, custody evaluations can prove highly influential in terms of settlement. Mary Johanna McCurley et al., Protecting Children From Incompetent Forensic Evaluations and Expert Testimony, 19 Forensic Evaluations 277, 277 (2005). includes a variety of information, ranging from the expert’s personal observations to custody-specific recommendations for court action. 117 See Tippins & Wittmann, supra note 16, at 194–95 (categorizing typical evaluator conclusions in “clinical inference hierarchy”). These authors suggest expert reports contain four “levels” of data, a sort of sliding scale of reliability: (1) clinical observations (“anything that the clinician observes with his/her senses”); (2) psychological conclusions (“inferences and higher-level abstractions . . . without reference to custody/best-interests constructs); (3) custody-specific conclusions (“an even higher level of abstraction about the case, making reference to custody-specific constructs”); and (4) custody-specific recommendations for proposed court action. Id. Timothy Tippins and Jeffrey Wittmann categorized these types of expert opinions on a spectrum of reliability, mapping the relationship between the testimony offered and the boundaries of psychological knowledge. 118 See id. at 195–206 (summarizing scientific and ethical limitations associated with different “levels” of expert opinion). As expert inferences deviate from scien­tifically grounded observations and conclusions and approach custody-specific recommendationsthe ultimate issuethey become less reliable. 119 See id. at 202 (arguing at custody-specific level, clinicians present “their logic and personal values” disguised as scientific truth without disclosing absence of “reliable clinical method to do this weighing”). While the “overwhelming majority of judges and attorneys believe that psychologists should directly address the ultimate issue before the court,” the constraints of scientific knowledge and lack of uniform standards undermine the helpfulness and accuracy of outcome-specific opinions on the ultimate issue. 120 Id. at 193, 205–08. See generally Daniel A. Krauss & Bruce D. Sales, Legal Standards, Expertise, and Experts in the Resolution of Contested Child Custody Cases, 6 Psychol. Pub. Pol’y & L. 843 (2000) (providing overview of problems associated with science and testing underlying expert testimony in this context); Shuman, supra note 108 (same). Nor is there evidence that “current interview protocols, traditional psychological tests, or custody-specific tests are in any way able to reliably predict child adjustment to different plans, yet 94% of evaluating psychologists still make such recommendations.” Tippins & Wittmann, supra note 16, at 204 (citation omitted); see also Emery et al., supra note 16, at 6–10 (explaining “most . . . measures [used by evaluators] are deeply flawed when used in the custody context” and lack scientific support, noting “[n]o one particular psychological test was used by a majority of the respondents when assessing children”).
Though the tension between judicial utility and scientific constraints is not unique to the custody context, the fallibility of expert testimony regarding domestic violence as currently offered in custody casesand judicial reliance upon itcounsel special care against its use. See infra section II.A.3 (describing unreliability of expert assessments in this arena).

2. Courts’ Reliance on Experts. — Despite these limitations and the Supreme Court’s clarification of federal standards for evaluating the admissibility of expert testimony in Daubert v. Merrell Dow Pharmaceuticals, Inc., 121 509 U.S. 579 (1993). Expert testimony may be offered where a witness’s “scientific, technical, or other specialized knowledge will help the trier of fact,” provided it is “based on sufficient facts or data” and is “the product of reliable principles and methods” that are “reliably applied” to the case at hand. Fed. R. Evid. 702. Daubert set forth four relevant but not dispositive factors upon which courts should evaluate expert testimony: (1) “whether [a theory or technique] can be (and has been) tested”; (2) “whether the theory or technique has been subjected to peer review and publication”; (3) the “known or potential rate of error”; and (4) “[w]idespread acceptance” in the field. Daubert, 509 U.S. at 592–94.
Though many states have adopted the Daubert standard and others have retained a standard similar to that articulated in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), superseded by rule as stated in Daubert, 509 U.S. at 586–87, state law ultimately dictates how to evaluate an expert’s reliability and qualifications. For a survey of these state standards, see ABA Trial Evidence Comm., 50 State Survey of Daubert/Frye Applicability, ABA, [] (last visited Oct. 28, 2015).
the bases of expert opinions in custody cases have rarely been challenged. 122 See Emery et al., supra note 16, at 24 (“[A] review of appellate cases also suggests that the opinions of mental health experts are rarely excluded on the grounds that the basis for the expert opinions offered does not meet required scientific standards.”). Emery et al. argue that applying Daubert or other state standards for admissibility of expert testimony would in itself increase the reliability of evidence received in custody cases. Id. at 23–24. Other scholars have speculated that unwillingness to challenge expert opinions in custody cases may be the product of “the vague best-interests principle and the impossible dilemma it creates for judges,” id. at 24, or the fact that family lawyers are repeat players—“[w]hat they might challenge today, they might use tomorrow,” Shuman, supra note 108, at 155. Infrequent evidentiary challenges mean virtual nonappli­cation of the rules of evidence, fueling criticism of courts’ reliance on expert testimony in custody cases. This concern is compounded by the questionable methods associated with custody evaluations. 123 See supra section II.A.1 (explaining methodological inadequacies).

What starts as a psychological evaluation of the parties and children can often devolve into a credibility assessment by the expert, who may verify or refute individual instances of domestic violence. 124 See Scott & Emery, supra note 16, at 96–97 (“Not surprisingly, courts often turn to psychological experts for assistance in evaluating these claims, and often the clinician’s role is to endorse or challenge the alleged victim’s credibility, on which basis custody can be decided.”); Shuman, supra note 108, at 160 (“[T]he role of mental health professionals in custody litigation is being transformed from expert as expert to expert as judge.”); Tippins & Wittmann, supra note 16, at 207 (explaining when court accepts experts’ recommendations without questioning underlying methods, “court may be basing its decision on personal value judgments of witnesses who happen to have professional credentials”). Given the deference courts often accord to mental health professionals, questions arise regarding the proper role of experts, and whether courts have improperly delegated their fact-finding responsibility. 125 Compare Tippins & Wittmann, supra note 16, at 207 (arguing mental health practitioners have no special ability to judge witness credibility), with Conner, supra note 58, at 171–76 (arguing courts are no better suited than experts for this task); see also Tippins & Wittmann, supra note 16, at 208 (“To the extent . . . personal opinions are admitted under the guise of science and are allowed to impact the outcome of the case, the process is tainted and judicial power is usurped.”). Such a dele­gation is especially problematic where the court itself appoints the expert and a perceived absence of bias may fortify the expert’s opinion in the eyes of the court. 126 See Shuman, supra note 108, at 161 (“The use of court-appointed experts whose opinions determine the outcome of many, if not most, contested custody cases effectively delegates judicial power without formal legislative approval.”); see also Stephanie Domitrovich et al., State Trial Judge Use of Court Appointed Experts: Survey Results and Comparisons, 50 Jurimetrics 371, 387–88 (2010) (finding “[f]amily court judges . . . made a large number of discretionary appointments” of experts). The effects of this testimony could be especially damaging where the expert provides an opinion on the ultimate issue without having to provide the bases for that opinion. 127 See Scott & Emery, supra note 16, at 92 n.121 (arguing against permitting expert testimony on ultimate issue in custody cases); see also infra section II.B.2 (criticizing expert opinions on ultimate issue in custody cases). But see Fed. R. Evid. 704 (permitting testimony on ultimate issue).
Experts in custody cases often base their opinions on interviews with witnesses and other hearsay testimony, raising additional concerns. See, e.g., Scott & Emery, supra note 16, at 93–95 (describing typical sources of expert data). Although statements made by interviewees during the evaluation itself are potentially admissible as the basis of an expert opinion, e.g., Fed. R. Evid. 703, allowing hearsay testimony could transform the expert into a mere conduit, see, e.g., Ronald L. Carlson, Experts as Hearsay Conduits: Confrontation Abuses in Opinion Testimony, 76 Minn. L. Rev. 859, 860–66 (1992) (describing this issue generally in context of Confrontation Clause). Even if hearsay statements are not themselves offered, critics have warned against the admission of credibility assessments dressed as expert opinions. Scott & Emery, supra note 16, at 85; cf. also Conner, supra note 58, at 175–76 (“Demeanor evidence and credibility evidence . . . should not be confused. The use of expert testimony to explain the characteristics of a class of individuals is quite different from the admissibility of such evidence to prove truth-telling on the part of a particular witness . . . .”).

3. Special Difficulties in Cases Involving Domestic Violence. — The short­comings of mental health experts in custody cases become only more apparent in the context of domestic violence. 128 Domestic violence is highly prevalent in disputed custody cases. See Bow & Boxer, supra note 26, at 1396, 1407 (citing studies finding 72–80% of “high-conflict and/or entrenched custody cases” involve allegations of domestic violence and 37% of custody referrals to evaluators involve domestic violence). The limited and potentially unreliable evidence upon which expert opinions may rely has drawn criticism of experts in custody cases involving allegations of domestic violence. 129 See Meier, supra note 67, at 708 (highlighting lack of specific domestic-violence-related knowledge of child custody evaluators). But see Bow & Boxer, supra note 26, at 1405–08 (rebutting doubts regarding competence of domestic violence experts). In the face of domestic violence allegations, the mental health professional often only has access to witness statements and evidence offered by the party alleging abuse. 130 See Scott & Emery, supra note 16, at 96 (“Allegations of physical or sexual abuse of children are often based largely on evidence provided by the accusing parent, who might already be distrusting and suspicious of the alleged abuser.”). Determining whether a history of abuse exists may be impossible: “In some cases, verification of a history of abuse is not possible, because there are often few clues pointing to its existence.” 131 Jaffe et al., Child Custody and Domestic Violence, supra note 5, at 35. “The utility of an assessment is heavily contingent on the nature of the information gathered during the assessment process.” Id. at 37; see also Bow & Boxer, supra note 26, at 1395–96 (discussing difficulties in verifying domestic violence claims). Nonetheless, the complexity of custody cases involving domestic violence increases court deference to experts. 132 See Davis et al., supra note 7, at 2 (“The presence of domestic violence and the frequently co-occurring allegations of child abuse may increase judges’ reliance on custody evaluators who can observe interactions between parents and children and interview children about what they have seen and heard.”).

Few mental health professionals receive advanced, specialized domestic violence training. 133 See Bow & Boxer, supra note 26, at 1400 (finding fewer than 35% of mental health professionals had taken graduate courses regarding domestic violence). In fact, California is the only state that requires training on domestic violence for court-appointed experts. 134 Davis et al., supra note 7, at 20. The methods utilized to evaluate the existence and effects of domestic violence vary greatly; these largely unstandardized methods include psychological testing and specialized questionnaires—some dubiously created by evaluators themselves. 135 See Bow & Boxer, supra note 26, at 1405–07 (noting while time spent during evaluations involving domestic violence is greater than those not involving domestic violence, variations exist in methods used); see also Davis et al., supra note 7, at 20 (collecting criticism of “use of psychological tests in custody evaluations on the basis of insufficient empirical validation and other deficiencies”). Moreover, comprehensive violence risk-assessment models incorporating multiple variables are uncommon and lack empirical verification. 136 See Bow & Boxer, supra note 26, at 1406–07 (“Only a few respondents reported using a comprehensive domestic violence model in the assessment process . . . .”); see also Scott & Emery, supra note 16, at 96–97 & nn.147–150 (compiling research revealing lack of sound methods to evaluate credibility of domestic violence claims in custody context). Finally, experts employ various and inconsistent psychological theories to domestic violence cases, some specific to domestic violence, and others not. 137 See Davis et al., supra note 7, at 14–18 (explaining relative appropriateness of various domestic violence typologies). The “dominant” typology among victim advocates is the “power and control” model and the derivative “coercive control” model, which “construct domestic violence as a pattern of behavior involving power and control” and “distinguish between true and dangerous intimate partner abuse and more transient and less serious incidents of violence.” Id. at 15–17. The scientific foundation of custody evaluations involving domestic violence is thus shaky at best.

Some argue that unsound scientific footing facilitates false or marginal claims of domestic violence. 138 See Bow & Boxer, supra note 26, at 1397 (noting possibility of false claims and citing studies finding increase in these claims in 1980s–90s). The claim of false allegations, often raised by fathers’ advocacy groups, is probably overstated. Cf. Scott & Emery, supra note 16, at 86 (“False claims likely are rare, but more common might be allegations based on suspicions . . . or exaggeration of the seriousness of violent incidents due to distorted recollections.”); cf. also Jaffe et al., Child Custody and Domestic Violence, supra note 5, at 58–59 (acknowledging possibility of false claims but reminding “the most common problem in this area is that domestic violence tends to be underreported rather than fabricated or exaggerated”). But expert testimony based on limited evidence can also harm victims of abuse; abusers use allegations of domestic violence or counterclaims of domestic violence against their victims as a method of control. 139 See Jaffe et al., Child Custody and Domestic Violence, supra note 5, at 17–18 (explaining how abusers may use legal system against their victims); Mary Przekop, Note, One More Battleground: Domestic Violence, Child Custody, and the Batterers’ Relentless Pursuit of Their Victims Through the Courts, 9 Seattle J. Soc. Just. 1053, 1055–56 (same). “[A]busive fathers were at least twice as likely to dispute custody as nonabusive men.” Jaffe et al., Child Custody and Domestic Violence, supra note 5, at 89. Further, in response to allegations of domestic violence, abusers may wield claims of the now-discredited “parental alienation syndrome” (PAS), asserting that the abused parent has turned the child against the abuser. 140 See Jaffe et al., Child Custody & Domestic Violence, supra note 5, at 95 (noting increasing claims of PAS in cases involving domestic violence). See generally Richard A. Gardner, The Parental Alienation Syndrome: A Guide for Mental Health and Legal Professionals (2d ed. 1998) (proposing PAS and guidelines for application). Recently scholars have panned this so-called syndrome as “unsubstantiated,” Jaffe et al., Child Custody & Domestic Violence, supra note 5, at 53, and “fail[ing] to meet minimal requirements universally recognized in the scientific community,” Scott & Emery, supra note 16, at 98. This criticism has sometimes fallen silent on courts, with some admitting PAS testimony and others excluding it. See Jaffe et al., Child Custody & Domestic Violence, supra note 5, at 94–96 (describing mixed reception of PAS by courts); see also Rita Berg, Note, Parental Alienation Analysis, Domestic Violence, and Gender Bias in Minnesota Courts, 29 Law & Ineq. 5, 5 (2011) (“Despite wide criticism of PAS . . . this theory has permeated the legal system, appearing primarily in custody judgments.”); infra section II.C.2 (discussing how formulation of statutory factors may encourage admission of PAS testimony in custody cases involving domestic violence). In all, the strong eviden­tiary force of expert testimony in custody cases involving domestic violence becomes problematic when paired with limited scrutiny of the opinions offered. Reliance on faulty evidence risks erroneous decisions and creates judicial inconsistency.

B. Comparing Civil Custody Cases to Criminal Cases

Parties commonly introduce expert testimony regarding domestic violence in both the civil and criminal contexts. This section contrasts the typical uses of expert testimony in both contexts, revealing more lenient evidentiary standards in civil custody cases, especially with respect to testimony regarding specific incidents of violence. It highlights the alarming discrepancy that courts in custody cases admit a class of evidence deemed so unreliable as to justify its exclusion in criminal cases.

1. Expert Testimony in Domestic Violence Criminal Cases. — In criminal cases, prosecutors increasingly use experts to explain abusive relation­ships and the psychological effects of intimate-partner battering. 141 See Paula Finley Mangum, Note, Reconceptualizing Battered Woman Syndrome Evidence: Prosecution Use of Expert Testimony on Battering, 19 B.C. Third World L.J. 593, 611–19 (1999) (discussing use of this evidence in prosecutions); see also Cal. Evid. Code § 1107(a) (West 2009) (providing evidence on partner battering is admissible when offered by either prosecution or defense); Jennifer G. Long, Nat’l Dist. Attorneys Ass’n, Introducing Expert Testimony to Explain Victim Behavior in Sexual and Domestic Violence Prosecutions 9 (2007),
testimony.pdf [] (providing guide for prosecutors on how to use expert testimony in domestic violence prosecutions).
In other words, the expert takes on the role of teacher. This evidence often relates to “battered woman syndrome,” originally offered in self-defense cases to explain a defendant’s actions, but now also used by prosecutors to “explain[] the puzzling ‘whys’ of some victims’ behavior” in domestic violence prosecutions. 142 Caitlin Valiulis, Domestic Violence, 15 Geo. J. Gender & L. 123, 134–36 (2014). “Over the last two decades, at least twenty states have begun to accept some form of this expert testimony from prosecutors in domestic violence cases.” Id. at 136. When called by the prosecution, an expert may testify to explain to the jury why a victim may recant or refuse to testify, for instance. 143 See generally Audrey Rogers, Prosecutorial Use of Expert Testimony in Domestic Violence Cases: From Recantation to Refusal to Testify, 8 Colum. J. Gender & L. 67 (1998) (explaining use of expert witnesses in domestic violence prosecutions).

While several states exempt this kind of testimony from foundational challenges, 144 See, e.g., Cal. Evid. Code § 1107(a) (“[E]xpert testimony is admissible . . . regarding intimate partner battering and its effects, including the nature and effect of physical, emotional, or mental abuse on the beliefs, perceptions, or behavior of victims of domestic violence . . . .”); Nev. Rev. Stat. Ann. § 48.061(1) (LexisNexis 2012) (“[E]vidence of domestic violence and expert testimony concerning the effect of domestic violence . . . is admissible in a criminal proceeding for any relevant purpose . . . .”). they often also prevent experts from testifying on the ultimate question of whether an act of violence occurred. 145 See, e.g., Cal. Evid. Code § 1107(a) (allowing such evidence “except when offered against a criminal defendant to prove the occurrence of the act or acts of abuse which form the basis of the criminal charge”); Nev. Rev. Stat. Ann. § 48.061(2) (“Expert testimony concerning the effect of domestic violence may not be offered against a defendant . . . to prove the occurrence of an act which forms the basis of a criminal charge against the defendant.”). Courts prevent experts in criminal cases from testifying to the occurrence of an act because such testimony’s limited probative value is substantially out­weighed by its prejudicial effect. 146 Rogers, supra note 143, at 82–83. This is consistent with the notion that the prosecutor’s use of expert testimony in this context is permissible “primarily because experts on [battered woman’s syndrome] have knowledge that is beyond the purview of the jury’s experience.” 147 Valiulis, supra note 142, at 138. Similarly, conclusions that a particular victim’s behavior is consistent with a recognized syndrome are considered “both ineffective and vulnerable to attacks that [they are] unreliable.” 148 Long, supra note 141, at 24.

2. Evidence Solicited in Custody Cases. — But when asked to make a custody recommendation, experts may have to make exactly the sort of assessment typically deemed unreliable in the criminal context, including whether a specific incident of violence occurred at a particular point in time. 149 See supra section II.A (describing holistic nature of custody evaluations and related evidence). Courts commonly ask mental health professionals responsible for preparing a custody evaluation to make specific recommendations of appropriate remedies. 150 See supra notes 120, 124 and accompanying text (noting it is commonplace for experts to make specific, custody-based recommendations). At the very least, an expert’s determination of whether a specific act of violence occurred could influence a custody recommendation; at most it could prove determinative. 151 See supra section I.B (discussing impact of domestic violence findings on custody decisions). Unless the expert explicitly discloses an underlying conclusion regarding the occurrence of a specific instance of violence, the court has no way of knowing whether the expert’s conclusion is based on scientific methodology or a personal credibility assessment. This possibility is magnified in cases of “dueling experts,” where experts for adverse parties may present competing credibility assessments in support of their respective clients. 152 See Thomas A. Powell & John C. Holt, Forensic Psychological Evaluations: The Methods in Our Madness, Vt. B.J., Winter 2005–2006, at 37, 37 (mentioning this possibility).

These risks have been realized. When asked, experts have admitted the importance of credibility-based assessments in formulating custody recommendations. In fact, a study commissioned by the Department of Justice found that most experts not only admit to making such assessments, but “felt it was incumbent on them to assess whether the allegations of domestic violence were true.” 153 Davis et al., supra note 7, at vi–vii. They did this by attempting to establish “consistency of accounts across sources” and determine the relative reliability of sources 154 Id. —tasks that sound alarmingly similar to those of the fact-finder. Yet another study found that 37% of judges agree that experts should make recommendations on the ultimate issue “because the expert was the most knowledgeable,” apparently acquiescing to a role for experts in fact-finding. 155 Id. at 20. Thus, in contrast to the criminal context, which often utilizes the expert as “teacher,” in the custody context, experts take on the role of “evaluator”—and apparently in some instances, final arbiter.

3. Contrasting Expert Testimony in Criminal Cases and Custody Cases. — The evidence requested of experts in custody cases often does not reflect the limits of scientific expertise. In the custody context, courts run the risk of admitting evidence regarding the occurrence of a specific act otherwise inadmissible in the criminal courtroom. 156 See supra notes 146–147 and accompanying text (explaining inadmissibility of this testimony in criminal prosecutions). Besides the general lack of challenges to scientific foundation in custody cases, 157 See supra section II.A.2 (noting lack of foundational challenges to expert testimony). a key difference between the scrutiny of domestic violence expert testimony in the criminal and civil contexts lies in the carefulness of the evaluation of whether the testimony “help[s] the trier of fact.” 158 Fed. R. Evid. 702(a). See generally Daniel A. Krauss & Bruce D. Sales, The Problem of “Helpfulness” in Applying Daubert to Expert Testimony: Child Custody Determinations in Family Law as an Exemplar, 5 Psychol. Pub. Pol’y & L. 78, 90–92 (1999) [hereinafter Krauss & Sales, Problem of Helpfulness] (explaining nature of “helpfulness” inquiry in custody cases). For instance, while credibility-based opinions in custody cases arguably offer little permissible “help,” 159 See supra section II.A.2 (explaining reliance on such testimony inappropriately substitutes expert for fact-finder). an appropriately “helpful” opinion lies in explaining behavior consistent with a pattern or history of violence. 160 See infra section III.B (proposing reconsideration of types of opinions experts should offer); see also supra notes 82–84 and accompanying text (noting various definitions of domestic violence). Courts, however, stretch the boundaries of “helpfulness” in custody cases by liberally applying state rules of evidence 161 The Supreme Court of Connecticut, in a case involving termination of parental rights, went so far as to recognize that court reliance on expert “testimony as to the ultimate issue . . . [is] consistent with . . . the Connecticut Code of Evidence and common practice in Juvenile Court proceedings.” In re Melody L., 962 A.2d 81, 100–01 (Conn. 2009), overruled on other grounds by State v. Elson, 91 A.3d 862 (Conn. 2014); see also Conn. Code Evid. § 7-3(a) (2015) (“[A]n expert witness may give an opinion that embraces an ultimate issue where the trier of fact needs expert assistance in deciding the issue.”). and soliciting specific custody recommendations from experts, which may derive from credi­bility assessments and conclusions on the ultimate issue. 162 See supra section II.A.2 (describing this tendency).

Surely, criminal and child custody cases are distinct. For one, a prosecutor must prove criminal charges beyond a reasonable doubt, 163 E.g., Colo. Rev. Stat. Ann. § 18-1-402 (West 2013) (establishing beyond a reasonable doubt burden for criminal cases). while the civil standard is a preponderance of the evidence unless otherwise specified. 164 E.g., Colo. Rev. Stat. Ann. § 13-25-127(1) (West 2014)(establishing prepon­derance burden for civil cases). Perhaps more importantly, that jury trials are exceedingly rare in custody cases 165 See supra note 53 (noting nearly all custody trials are bench trials). may assuage concerns over admitting problematic evidence that could be misapplied by jurors. After all, in a bench trial judges should know upon which evidence they may or may not rely. 166 See Andrew J. Wistrich et al., Can Judges Ignore Inadmissible Information? The Difficulty of Deliberately Disregarding, 153 U. Pa. L. Rev. 1251, 1255–56 (2005) (surveying arguments that judges are better than jurors at disregarding prejudicial evidence). But studies revealing that judges in fact request and rely on ultimate-issue expert testimony in custody cases 167 See supra notes 153–155 and accompanying text (providing statistics indicating judges’ reliance on these opinions). upend this typical justification.

Custody cases involve ex ante predictions about the future, while a criminal trial involves an ex post determination of guilt. 168 See Ellman et al., supra note 38, at 358 (explaining while “[m]ost adjudication ‘requires determination of past acts and facts,’” in custody determinations “‘[p]roof of what happened in the past is relevant only insofar as it enables the court to decide what is likely to happen in the future’” (quoting Robert H. Mnookin, Child-Custody Adjudication: Judicial Functions in the Face of Indeterminacy, Law & Contemp. Probs., Summer 1975, at 226, 251–52)). But where questions of domestic violence are concerned, this line is blurred: The inquiries posed to courts, and therefore to experts, look a lot like the ex post determinations typical of a criminal trial. 169 Incident-based definitions depart from the spirit of the “person-oriented” best-interests standard and demand instead an “act-oriented” inquiry. See Ellman et al., supra note 38, at 357–58 (explaining this distinction). Courts do not seem to have recognized this shift. Rather than tailoring opinions to best assist courts in making the required future-oriented best-interests determination, experts in custody cases venture into territory otherwise prohibited in similar legal contexts, creating a troubling inconsistency.

C. State Law and the Admissibility of Problematic Expert Testimony on Domestic Violence

Legislatures’ statutory commands regarding the definition of dom­estic violence and the factors courts must consider in determining the best interests of the child naturally influence which facts are relevant in custody proceedings. It follows that a court will solicit expert testimony relevant to these factors. Thus, the combination of current state law and limited scrutiny of expert testimony encourages the admission of scientifically unsound expert opinions.

1. Effects of Statutory Definitions on Admissibility of Expert Testimony. — As explained earlier, the evidence sufficient to invoke a DV presumption varies by state. 170 See supra section I.C.1 (discussing statutory variation). While some states require a showing of a single instance of domestic violence to invoke a presumption against awarding custody to an abuser, other states require a showing of a history or pattern of domestic violence. 171 See supra section I.C.1 (comparing state definitions of domestic violence in custody statutes). Because of the burden-shifting consequences of invoking a DV presumption, evidence of domestic violence becomes crucial. 172 See supra section I.B.2 (explaining function of DV presumptions). Inevitably, mental health professionals must also consider evidence of domestic violence in custody evaluations and recommendations. 173 If the court requests a specific custody recommendation from the expert, it implicitly asks the expert to consider the relevant statutory provisions. Scholars have questioned experts’ ability to properly account for DV presumptions in their recommendations. See Bowermaster, supra note 27, at 284–89 (arguing experts lack proper legal training to account for DV presumptions in their conclusions and recommendations).

Legal definitions fail to fully encapsulate what constitutes abuse between intimate partners. 174 Conner, supra note 58, at 219–20 (“[C]omparing clinical and legal definitions of domestic violence . . . suggests that our legal definition of domestic violence may fail to provide an accurate characterization of what constitutes abuse between intimate partners.”). The “incident-based” definitions of crim­inal law have carried over to custody law. 175 See id. (describing both civil and criminal responses to domestic violence as “incident-based”). And incident-based analyses are often not amenable to expert evaluations. 176 See supra sections II.A.3–B.1 (noting difficulties of expert evaluations of domestic violence cases and exclusion of incident-based testimony in criminal cases). Where legal definitions fail to accurately encapsulate the effects of abuse, experts cannot make scientifically reliable conclusions that directly respond to the statutory inquiry. 177 Indeed, some have argued that judges are no better suited to undertake an incident-based evaluation. See supra note 58 (citing such arguments). If this were the case, perhaps the comparative capability of experts to decide such matters would be more desirable than judicial determination of the ultimate issue. But with no evidence to support the assertion that existing scientific knowledge fosters comparatively better evaluation by experts, see supra section II.A (explaining shortcomings of experts), the question becomes one of allocation of responsibility rather than absolute reliability. Legislatures have delegated fact-finding responsibility to judges in the custody context, and judges commonly eschew it. See supra notes 59, 126 and accompanying text (summarizing improper-delegation argument).

In cases where allegations of domestic violence are not documented by convictions, police reports, or previous restraining orders, an expert’s evaluation may be central to a court’s analysis. This is problematic in several scenarios. Take first a state in which a single act of domestic violence is enough to invoke a DV presumption. 178 For two such statutes, see supra note 82. Assuming a party alleges a single instance of domestic violence, the evidence available to ensure the expert properly accounts for domestic violence is very limited. Without substantial corroborating evidence, the expert may be forced to base her opinion solely on the interviews of the parties. Reliable psychol­ogical tools for this undertaking are currently unavailable. 179 See supra section II.A.3 (explaining experts’ difficulty in assessing allegations of domestic violence in custody evaluations).

Even in states requiring a victim to demonstrate a pattern or history of abuse, DV presumptions require findings of specific acts of domestic violence. The statute may further specify timing and severity requirements for the violent acts, 180 See, e.g., Cal. Fam. Code § 3044(a) (West 2004) (requiring commission of violent act within last five years); see also supra section I.C.1 (discussing varying degrees of violence required to invoke DV presumptions). elements typically decided by the fact-finder, but which courts may implicitly delegate to experts. Courts delegate such decisions when they accept expert recommendations based on a determination of whether a past event actually occurred. While mental health professionals may be well suited to offer opinions on the psychological symptoms displayed by a victim of abuse or on the dynamics of the family relationship, 181 See Tippins & Wittmann, supra note 16, at 196 (recognizing strengths of mental health professionals in drawing these types of conclusions). it is within the court’s purview to decide whether specific acts of abuse have occurred. 182 This finding of fact is required in order to invoke the DV presumption. See supra section I.C.1 (describing when domestic violence triggers DV presumptions).

Because the best-interests analysis and the formulation of DV presumptions do not account well for the psychological effects of abuse 183 See infra notes 216–222 and accompanying text (discussing problem and calling for reformulation of statutes to better account for psychological harm); cf. Bailey, supra note 24, at 35 (“[D]omestic violence is almost always defined in terms of physical abuse, rather than psychological.”). and judges commonly ask for a holistic assessment of the parties, 184 See Tippins & Wittmann, supra note 16, at 203–04 (explaining consideration of many factors involved in making custody recommendations). the opinion the court seeks may be discordant with the scientific realities constraining experts. Incident-based definitions of domestic violence and those that fail to take into account the psychological effects of violence are out of line with prevailing typologies of domestic violence. 185 See infra notes 217–222 and accompanying text (describing weaknesses of “incident-based” definitions of violence and explaining accepted typologies focus on power dynamics and distinguish between types of violence). This critique finds support in the fact that a “custody evaluator’s knowledge of domestic violence” is more predictive of the safety of parenting plans than “the severity of the physical, emotional and social abuse in [a] couple’s history.” 186 See Davis et al., supra note 7, at 85 (finding startling inconsistencies in expert evaluations in cases involving domestic violence).

2. Effects of Enumerated Statutory Factors on Admissibility of Expert Testimony. — After a court makes a finding of domestic violence, some statutes specify how a party may rebut the resulting presumption against custody. 187 See supra note 70 (citing Arizona statute providing guidance for rebuttal). Oftentimes there is a catch-all provision within the enumerated lists of factors to be considered on rebuttal. Such a provision allows a party to show by a preponderance of the evidence that the parent found to have committed domestic violence should retain custody because it is within a child’s best interests. 188 E.g., Cal. Fam. Code § 3044(b)(1) (West 2004) (containing catch-all consideration of “[w]hether the perpetrator of domestic violence has demonstrated that giving sole or joint physical or legal custody of a child to the perpetrator is in the best interest of the child”). At that point, one must look to the relevant statute defining the best-interests standard, which often contains another litany of enumerated factors.

The inclusion of criticized “friendly parent” provisions 189 See DiFonzo, supra note 64, at 224–25 (warning that victims of abuse may be disadvantaged by friendly-parent claims of abuser); Peter Jaffe, A Presumption Against Shared Parenting for Family Court Litigants, 52 Fam. Ct. Rev. 187, 191 (2014) (same). Friendly-parent provisions are in force in thirty states. Bailey, supra note 24, at 58. among these factors provides one example of how statutory factors may encourage admission of scientifically questionable evidence. Friendly-parent provisions instruct courts to consider the ability of a parent to facilitate cooperation and a continuing relationship with the other parent. 190 E.g., Fla. Stat. Ann. § 61.13(3)(a) (West Supp. 2015) (requiring courts to consider “demonstrated capacity and disposition of each parent to facilitate and encourage a close and continuing parent-child relationship, to honor the time-sharing schedule, and to be reasonable when changes are required”); id. § 61.13(3)(l) (requiring courts to consider “demonstrated capacity of each parent to communicate with and keep the other parent informed of issues and activities regarding the minor child, and the willingness of each parent to adopt a unified front on all major issues when dealing with the child”). The language in these statutes is closely aligned with the much-maligned PAS. 191 See supra note 140 and accompanying text (noting criticism of PAS). Claims that children suffer from this “syndrome” have been brought against victims of abuse—who in response to violence may shield children from the offending parent—to rebut domestic violence claims. 192 See supra note 140 (explaining use of PAS “diagnoses” against abusers). Paired with limited court review of expert testimony in custody cases, state statutes such as these invite scientifically unreliable evidence. 193 Some states provide that friendly-parent provisions do not apply where the case involves a finding of domestic violence. See Lisa Bolotin, When Parents Fight: Alaska’s Presumption Against Awarding Custody to Perpetrators of Domestic Violence, 25 Alaska L. Rev. 263, 279–80 (2008) (recounting history of such provision in Alaska, found necessary after public hearings revealing awarding custody to abusers on basis of friendly-parent provision).

Though many courts will not explicitly admit PAS evidence, 194 Cf. Clare Dalton et al., Navigating Custody and Visitation Evaluations in Cases with Domestic Violence: A Judge’s Guide 24–25 (2006),
default/files/navigating_cust.pdf [] (“Any testimony that a party to a custody case suffers from [PAS] or ‘parental alienation’ should . . . be ruled inadmissible.”).
its influence can still be found. Take for instance a 2013 California case, In re Marriage of Crystal and Shawn H. 195 No. D061388, 2013 WL 2940952 (Cal. Ct. App. June 17, 2013). Despite a finding that the husband sexually abused his wife, substantiated by audio recording and resulting in conviction and a six-year prison sentence, 196 Id. at *1–3. a trial court still mandated joint custody based in part on evidence of parental alienation. 197 Id. at *1, *14–16. The court gave credence to the notion that the mother may have turned her children against their abusive father, suggesting that the mother had been “bad mouthing” the father. Id. at *5. The order was reversed on appeal, as the trial court improperly relied on ex parte discussions with an expert about parental alienation. 198 Id. at *14–16. Though the admission of this evidence happened to be nullified through the appeals process, it provides a stark reminder of the possible lingering influence of PAS evidence. Indeed in some cases, evidence of parental alienation has been admitted so long as not framed in clinical terms as a “syndrome.” 199 See, e.g., McRoberts v. Superior Court of L.A. Cty., No. B234877, 2012 WL 2317714, at *10–12 (Cal. Ct. App. June 19, 2012) (admitting testimony of “alienation” where expert “did not offer an opinion based on PAS or any other syndrome or diagnosis” and “instead simply discussed a dynamic that he frequently observed during his three decades of practice and offered his expert opinion regarding the matter”); see also Scott & Emery, supra note 16, at 99–100 (“Courts routinely consider expert testimony on PAS . . . despite the lack of any scientific foundation for this diagnosis.”). And this testimony is encouraged by the current form­ulation of best-interest factors and a lack of uniformity in legislative guidance regarding friendly-parent evidence in cases involving domestic violence. 200 See D.M.S. v. I.D.S., No. 2014-CA-0364, 2015 WL 926777, at *18–19 & n.17 (La. Ct. App. Mar. 4, 2015) (“[T]he trial judge was well within the ambit of her discretion in ruling that while PAS may be relevant to a child custody matter, the purported syndrome has no bearing in a protection case . . . .”); Palazzolo v. Mire, 10 So. 3d 748, 771–77 (La. Ct. App. 2009) (recognizing criticism of PAS but finding expert testimony regarding alienation relevant as related to “moral fitness” and other statutory factors); Margaret K. Dore, The “Friendly Parent” Concept: A Flawed Factor for Child Custody, 6 Loy. J. Pub. Int. L. 41, 51 (2004) (“Parental Alienation Syndrome contains . . . features of a friendly parent analysis.”).

3. “Piggybacking” of Otherwise Inadmissible Testimony. — Expert test­imony offered in previous judicial proceedings may also indirectly impact later custody proceedings—a sort of “piggybacking” of evidence created by the consideration of previous judgments based on unchecked expert testimony. This is made possible when a judge accepts a previous finding of domestic violence, such as a civil restraining order, as evidence supporting a claim of domestic violence in a later custody case. 201 See supra section I.D (explaining possibility).

Though courts have cautioned against treating domestic violence restraining orders as final custody determinations, 202 E.g., Keith R. v. Superior Court, 96 Cal. Rptr. 3d 298, 304 (Ct. App. 2009). The court reasoned as follows:
“[D]omestic violence orders should not be treated as the functional equivalent of final judicial custody determinations. Domestic violence orders often must issue quickly and in highly charged situations. The focus understandably is on protection and prevention, particularly where the evidence concerning prior domestic abuse centers on the relationship between current or former spouses. Treating domestic violence orders as de facto final custody determinations would unnecessarily escalate the issues at stake, ignore essential factors (such as the children’s best interest) and impose added costs and delays.”
these orders may take on such an effect if sufficient to trigger a DV presumption—at the very least, any evidence relied upon in the previous proceeding may not be subject to the same scrutiny as other evidence in the current pro­ceeding. Some states have adopted separate rules of evidence for restraining-order proceedings, 203 See supra notes 100–101 and accompanying text (explaining evidentiary standards in civil restraining-order proceedings). and courts disagree whether evidence received in a restraining-order proceeding is admissible in a concurrent or subsequent custody hearing. 204 Compare In re Nilda S., 754 N.Y.S.2d 281, 281 (App. Div. 2003) (admitting therapist report and colloquy under family-offense exception, otherwise inadmissible hearsay, because family-offense and custody proceedings were “inextricably interwoven”), with Khan-Soleil v. Rashad, 969 N.Y.S.2d 104, 106 (App. Div. 2013) (refusing to apply same exception because proceedings were not “inextricably interwoven”). These differing outcomes beg the question of why domestic violence family-offense proceedings and child custody proceedings would ever tend not to be so related as to be “inextricably interwoven.”

A prior restraining order is compelling evidence of domestic violence, however. Though proceedings for restraining orders do not involve extensive evidence, parties are usually free to call expert witnesses in support of their claims. 205 See Jane K. Stoever, Enjoining Abuse: The Case for Indefinite Domestic Violence Protection Orders, 67 Vand. L. Rev. 1015, 1078 (2014) (“Although protection order cases are heard on an accelerated schedule due to the underlying allegations of violence, either party in a protection order case may . . . call expert witnesses.”); cf. Furry v. Rickles, 68 So.3d 389, 390 (Fla. Dist. Ct. App. 2011) (reversing lower court ruling denying appellant opportunity to cross-examine opposing party’s expert witness); Brandt v. Brandt, 523 N.W.2d 264, 265 (N.D. 1994) (recounting expert’s testimony in restraining order case). If the court granting the previous restraining order fails to detail the bases of its findings and the scope of any expert testimony offered, unreliable evidence may piggyback its way into custody proceedings. Even though “[t]he summary nature of restraining-order hearings make calling an expert unlikely,” 206 Aiken & Murphy, supra note 109, at 52. the possibility of such testimony has led some to call for appointed counsel in restraining-order cases, recognizing the potential for unchallenged expert testimony in largely uncounseled proceedings. 207 See Beverly Balos, Domestic Violence Matters: The Case for Appointed Counsel in Protective Order Proceedings, 15 Temp. Pol. & Civ. Rts. L. Rev. 557, 600–01 & n.429 (2006) (noting possibility of expert testimony and arguing “in favor of the appointment of counsel in protective order proceedings”).

The cost of error in custody cases is especially high because every decision has a significant and immediate impact on the lives of children and their parents and familial instability accompanies indeterminacy. 208 See Ellman et al., supra note 38, at 355–57 (quoting Mavis Hetherington et al., Marital Transitions: A Child’s Perspective, 44 Am. Psychol. 303, 303–10 (1989))(noting harmful effects of marital discord). Lawyers and victims alike have even attributed grants of custody to an abuser to expert “ignorance of the dynamics of domestic violence.” 209 Davis et al., supra note 7, at 5. These concerns, along with a desire for judicial consistency in appli­cation of legislative mandates, compel legislative and judicial action.

III. Reconciling Scientific Limitations with the Prevalence of Domestic Violence

The current statutory framework and relevant procedures governing child custody cases facilitate the introduction of scientific evidence of questionable value, making it necessary to revisit these rules and under­lying policies. Merely pointing out that courts may be admitting problematic evidence in child custody proceedings does little in the face of the devastating effects of domestic violence 210 See Bailey, supra note 24, at 48–52 (explaining harmful effects of domestic violence on children, even where violence was not carried out directly against children). and the already difficult task of substantiating domestic violence allegations. 211 See supra notes 129–131 and accompanying text (explaining these difficulties). Section III.A suggests modifications to DV presumptions and related procedures to improve the quality of evidence considered by courts. Section III.B reconsiders the role of experts in child custody proceedings and recom­mends further examination of current practice standards. Section III.C considers alternate methods of facilitating presentation of evidence of domestic violence, recognizing the already arduous task of proving up these claims.

A. Reformulating Procedures and DV Presumptions with Science in Mind

Given the prevalence of expert testimony in custody cases 212 See supra section II.A (explaining role of experts in custody cases). and the fact that experts can help judges in difficult fact-finding, 213 See supra section II.A.1 (noting strengths and weaknesses of custody evaluators). legislatures should write DV presumptions with the strengths and shortcomings of scientific knowledge in mind. Further, states can consider closing evidentiary loopholes and eliminating statutory language encouraging introduction of faulty evidence. 214 See supra section II.C.2 (providing example).

1. Crafting Statutory Factors. — Legislatures can modify the factors in custody statutes relating to domestic violence to make them more amenable to a mental health professional’s expertise. 215 In other words, legislatures should craft statutory factors with science in mind. This means recognizing the scope of available, reliable evidence of domestic violence and inserting reliable indicators of domestic violence as factors in a court’s evaluation. It also means implementing rules excluding unreliable evidence. Such factors should emphasize the extent of the psychological effects of abuse on victims. 216 See Bailey, supra note 24, at 35 (“[D]omestic violence is almost always defined in terms of physical abuse, rather than psychological.”). By moving away from exclusively incident-based definitions, statutes could mandate consideration of the psychological effects of violence on victims and their children. So instead of only asking whether a history of violence or multiple instances of violence exist, 217 See, e.g., La. Stat. Ann. § 9:364(A) (Supp. 2015) (requiring “history of perpetrating family violence”). a statute could also require consideration of the violence’s effect on the family’s psychological well-being. 218 In a similar manner, Alaska requires consideration of how a parent’s substance abuse has affected the psychological health of the child. Alaska Stat. § 25.24.150(c)(8) (2014). This type of evaluation is better grounded in available psychological methods and testing, 219 See Nancy Ver Steegh, Differentiating Types of Domestic Violence: Implications for Child Custody, 65 La. L. Rev. 1379, 1420–21 (2005) (noting custody evaluators can provide valuable input in differentiating and explaining psychological effects of abuse). more so than specific credibility-driven conclusions about whether particular incidents of violence have occurred. 220 See supra section II.A.3 (discussing expert evaluations in custody context). Such a modification would allow courts to specifically cite psychological effects of violence in a DV presumption analysis rather than in the context of up-or-down findings of specific instances of conduct. 221 See supra section II.B.3 (contrasting questionable admission of expert opinions regarding specific conduct in custody setting with criminal setting, where such evidence is not admissible); see also Bailey, supra note 24, at 46 (noting court practices also fail to take into account damaging psychological effects of high-conflict custody battles on children). Taking into account psychological factors also more adequately reflects the reality that domestic violence, though always damaging and unacceptable, exists in varying degrees of severity. 222 See Ver Steegh, supra note 219, at 1379–80 (“[C]hild custody courts could more effectively protect children through identification and consideration of the type of domestic violence a family has experienced . . . . Current statutes are not drafted with sufficient precision to adequately protect children.”). Michael P. Johnson identified four types of domestic violence “based on the motivation of the aggressor and the overall pattern of the violence.” Id. at 1384 (citing Michael P. Johnson & Kathleen J. Ferraro, Research on Domestic Violence in the 1990s: Making Distinctions, 62 J. Marriage & Fam. 948 (2000)). These include “Intimate Terrorism, which involves an escalating pattern of coercive control, . . . Situational Couple Violence, which involves isolated conflict-based incidents,” and “Violent Resistance, which involves self-defense.” Id. Because statutory definitions of domestic violence do not differentiate among these different forms of domestic violence, they fail to adequately target abusive relationships involving coercive control. See id. at 1415–19, 1423–24 (“Even if the court weighs the domestic violence factor heavily, . . . the statutory definition of domestic violence is unlikely to address patterns of coercive control.”).

On the other hand, statutory factors that implicitly invite question­able evidence, like friendly-parent provisions, should be eliminated for cases involving domestic violence. 223 See id. at 1421 & n.260 (arguing “‘[f]riendly parent’ provisions should not be used in cases involving domestic violence” and noting Minnesota statute that explicitly precludes consideration of friendly-parent factor in custody cases involving domestic violence (citing Minn. Stat. Ann. § 518.17(1)(a)(13) (West 2006)). Alaska took similar steps following hearings revealing that courts were admitting evidence regarding parental alienation at higher rates where one spouse alleged domestic violence. 224 See Bolotin, supra note 193, at 279–80 (explaining, after hearings revealed five fatalities stemming from friendly-parent provision, “Alaska courts do not consider the friendly parent provision where one parent has a history of perpetrating domestic violence”). Factors such as the ability of one parent to maintain a cooperative and continuing relationship with the other parent and to make joint decisions with the other parent on major issues 225 See, e.g., supra note 190 (providing example of friendly-parent statute formulation). —assuming legal custody is not awarded to one parent alone—fail to take heed of the situation and mindset of a victim of abuse and only encourage introduction of parental-alienation evidence, which should not be considered by courts. 226 See, e.g., Dalton et al., supra note 194, at 24–25 (warning courts against admitting evidence relating to parental alienation).

Finally, legislatures can include specific guidance regarding the weighing of factors, requiring courts to make specific findings of whether domestic violence exists, note whether the court has applied a presumption, and detail the bases of those findings. 227 See, e.g., Fla. Stat. Ann. § 61.13(3)(m) (West Supp. 2015) (“If the court accepts evidence of prior or pending actions regarding domestic violence . . . the court must specifically acknowledge in writing that such evidence was considered when evaluating the best interests of the child.”). This would allow for greater transparency and easier evaluation of opinions on review. 228 See Davis et al., supra note 7, at 88 (praising New York legislation “requiring judges to state on the record how any findings regarding domestic violence influenced their decisions on custody and visitation”). States can also require judges to consider corroborating evidence of domestic violence outside of an expert’s conclusions, similar to the California provision requiring consideration of additional substantiating evidence in addition to expert conclusions. 229 Cal. Fam. Code § 3044(e) (West 2004) (mandating consideration of “any relevant, admissible evidence submitted by the parties” in addition to any expert testimony). Such a provision recognizes the proper role of the judge as fact-finder.

2. Closing Evidentiary Loopholes. — Two currently existing evidentiary loopholes include (1) consideration of problematic expert opinions on the ultimate issue of the existence of individual instances of domestic violence and (2) admission of previous findings of domestic violence where underlying evidence may not have been subject to scrutiny. 230 See supra section II.C (addressing state laws allowing admission of problematic expert testimony).

Prohibiting expert ultimate-issue opinions, including specific custody recommendations and testimony on specific instances of conduct, is an appropriate remedy and is similar to criminal evidentiary rules in domestic violence prosecutions. 231 See supra note 145 and accompanying text (explaining rules in criminal prosecutions). Both the scientific and legal communities support this recommendation. 232 See Davis et al., supra note 7, at 86–87 (recommending “courts make their own assessments of whether domestic violence has been committed by one family member against another” and concluding “it is more appropriate for custody evaluators to present conclusions without recommending a specific parenting plan”). Even if states do not explicitly prescribe such rules, application of existing evidentiary rules by courts would likely prevent admission of problematic testimony. 233 See Emery et al., supra note 16, at 24 (noting survey of appellate decisions reveals lack of review under Daubert). A review of available scientific techniques shows that ultimate-issue conclusions about domestic violence fail to withstand review under the Frye or Daubert standard 234 Cf. Tippins & Wittmann, supra note 16, at 210 (comparing evaluation of domestic violence expert testimony in custody proceedings under Frye and Daubert). and state evidence rules mirroring Federal Rule of Evidence 702, requiring that scientific or technical testimony “help the trier of fact.” 235 Fed. R. Evid. 702(a); see also Tippins & Wittmann, supra note 16, at 208–11 (putting forth proposed application of Daubert and suggesting current practice is flawed); cf. also Krauss & Sales, Problem of Helpfulness, supra note 158, at 96–97 (reminding judges that best-interests-based expert opinions involve many subfactors, underlying assumptions, and series of conclusions, all of which must withstand scientific scrutiny prior to admission).

States should also consider the evidentiary implications of the availability of multiple forums for civil domestic violence cases. Courts should be aware that differing burdens and varied levels of scrutiny might be applied to evidence in earlier proceedings. 236 See supra section II.C.3 (explaining potential for judgments based on otherwise inadmissible evidence to affect later custody proceedings); see also supra notes 202–204 and accompanying text (noting differing evidentiary standards and burdens of proof). Because some states give presumptive effect to previous findings of domestic violence, 237 Supra note 98 and accompanying text. it is important to reconsider the extent to which findings in earlier proceedings should have a res judicata effect on later custody cases. While a criminal conviction may inarguably constitute sufficient evidence to trigger a DV presumption, 238 Many states explicitly recognize the importance of considering prior criminal convictions. See, e.g., Fla. Stat. Ann. § 61.13(2)(c)(2) (West Supp. 2015) (providing conviction automatically triggers DV presumption). a finding of domestic violence in an expedited restraining-order proceeding with a limited record should not be afforded the same weight. That is not to say that the issuing of a restraining order is not relevant, but rather that courts should also consider the extent of the evidence supporting the order in addition to the issuance of the order itself.

B. Rethinking the Role of the Expert

Many scholars have pointed to the best-interests standard as the cause of courts’ overreliance on experts. 239 For discussion of these criticisms, see supra notes 47–51 and accompanying text. They argue that moving away from the best-interests inquiry and toward another standard would limit the improper use of experts; the broad, forward-looking best-interests standard encompassing numerous factors leads courts to depend on experts to a fault. 240 See, e.g., Scott & Emery, supra note 16, at 100 (“As long as the best-interests standard persists as the custody decision rule, judges are likely to urge mental-health experts to offer opinions on the ultimate issue of custody unless they are legally restricted from doing so by the evidentiary screen that applies to other legal proceedings.”). Experts’ questionable intertwining of multiple analyses mirrors the ambiguous best-interests standard. 241 See, e.g., id. at 24 (“[T]he low scientific standards for expert testimony again can be traced to the vague best-interests principle and the impossible dilemma it creates for judges.”); see also Bowermaster, supra note 27, at 310 (advocating approximation standard in consideration of DV presumptions). But because this standard has been entrenched in family law for decades, the mental health profession must articulate clearer standards for expert testimony in child custody cases. 242 Cf. Robert Geffner et al., Conducting Child Custody Evaluations in the Context of Family Violence Allegations: Practical Techniques and Suggestions for Ethical Practice, 6 J. Child Custody 189, 212–13 (2009) (advancing guidelines for custody evaluations involving allegations of domestic violence).

Rethinking the role of the expert in custody proceedings involves a careful balancing of the expert’s potential roles as “teacher” and “evaluator.” 243 See Tippins & Wittmann, supra note 16, at 210 (“The implications of [the] dwindling attachment to scientific method with respect to therapeutic settings aside, in the forensic context such unanchored opinions can determine whether or how often a specific child gets to see a particular parent.”). Until advances in psychological science result in reliable methods for evaluating specific claims of domestic violence, experts should limit their opinions accordingly. 244 The American Psychological Association (APA) recognizes internal disagreement among mental health professionals regarding the propriety of offering opinions on the ultimate issue. See Am. Psychology Ass’n, supra note 110, at 866 (noting lack of “consensus”). Though “psychologists seek to remain aware of the arguments on both sides of this issue and are able to articulate the logic of their positions on this issue,” the APA seems reluctant to oppose specific custody recommendations. Id. (citations omitted). Organizations capable of surveying the community of experts, such as the American Psychological Association, should first ask whether domestic violence should be considered separately by experts with specialized knowledge and apart from other relevant custody factors. 245 Cf. Dale & Gould, supra note 3, at 8–9 (advocating careful selection of experts because specialized knowledge of protocols in areas such as domestic violence is necessary for sound opinions regarding these issues). This important step would serve to recognize the difference between determining the occurrence of a specific event and the psychological and psychosocial elements of expert custody evaluations. 246 See Bowermaster, supra note 27, at 276–94 (arguing experts inappropriately apply DV presumptions in their evaluations, leading to distortion of relevant statutory factors where judge heavily relies on expert’s opinion). The scientific community must also carefully assess whether their methods are consistent with the field’s ethical standards, which arguably prohibit certain ultimate-issue conclusions. 247 See McCurley et al., supra note 116, at 317–18 (noting relevant ethical standards for mental health professionals to consider regarding reliability of methods and presentation of evidence in custody cases); cf. Bowermaster, supra note 27, at 301 (“The kinds of social, moral, financial and legal concerns involved in child custody deter­minations are simply outside the scientific expertise of mental health professionals.”).

The development of uniform standards for experts should include further empirical study of the existing methodology for evaluating domestic violence claims and adherence to an honest assessment of current scientific limitations. Experts may be most useful in the fact-gathering process, where “the simple and cautious recording of best-interests-related admissions on the part of parents, or reports about family life on the part of children, can provide a valuable glimpse into how the family functions.” 248 Tippins & Wittmann, supra note 16, at 195. Expert opinions can also provide “a more humanized understanding of the emotional nuance behind petitions” 249 Id. at 197. and “statements about potential risks [and] advantages, as long as they are grounded in case-specific facts and reliable empirical literature.” 250 Id. at 200. Opinions on the ultimate issue should be avoided. 251 See Scott & Emery, supra note 16, at 92 n.121 (arguing experts should limit testimony on ultimate issue); Tippins & Wittmann, supra note 16, at 203, 218 (noting dangers of expert testimony on ultimate issue and suggesting inclusion of warnings in expert reports); see also supra sections II.B.3–C.1 (noting how current practices encourage expert opinions on ultimate issue).

Looking to how prosecutors utilize experts in criminal cases highlights how the expert as “teacher” is useful to explain relevant concepts to the fact-finder. 252 See supra section II.B.1 (explaining use of expert testimony in domestic violence prosecutions). These opinions provide valuable insight into the dynamics of an abusive relationship and allow the court to make inferences about the existence of domestic violence without soliciting ultimate-issue opinions from experts. 253 See supra notes 141–143 and accompanying text (describing usefulness of expert testimony in criminal context). The scope of this testimony would necessarily differ in the child custody context, given the prospective nature of the best-interests standard as opposed to the ex post nature of criminal proceedings. Though family law judges pre­sumably do not need extensive primers in abusive relationships as a jury might, there is room for experts to provide valuable insight short of making ultimate-issue conclusions. For example, it may be appropriate for an expert to make specific diagnoses and draw other psychological conclusions, even though such evidence is presented in a circumspect manner in criminal cases, avoiding reference to the specific parties. 254 See, e.g., Long, supra note 141, at 20 (recommending prosecution experts seeking to explain victim behavior should not review case file or mention victim by name). Emphasizing the special abilities of experts to put testimony and family dynamics of individual cases into context may help to reduce improper reliance on experts in fact-finding.

C. Additional Tools for Victims: Alternate Forms of Evidence

Stringent evidentiary standards for expert testimony, even if necessary, may deprive victims of a strong form of substantiation for valid claims. For this reason, it is useful to consider alternate means of reliably supporting claims of abuse.

One possibility is better facilitating the ability of a victim to present evidence of abuse. Some states have allowed videotaped statements of children in cases of child abuse, for instance, which avoids subjecting victims to multiple cross-examinations over the course of a proceeding. 255 See, e.g., Conn. Gen. Stat. Ann. § 54-86g (West 2009) (permitting testimony of child victims be taken outside courtroom). California permits use of videotaped or stenographical victim testimony at preliminary hearings upon motion by the State in a criminal domestic violence prosecution. 256 Cal. Penal Code § 1346.1 (West Supp. 2015) So long as the evidence is admissible at the preliminary hearing, it may later be introduced at trial. 257 Id. § 1346.1(d). In the custody context, such a provision could allow victims of abuseor perhaps childrento introduce statements made in prior civil restraining-order hearings, family-offense hearings, or preliminary hearings. Some have also suggested exceptions to the hearsay rule for initial reports of domestic violence to police, purporting to solve the problem of re­cantation, common in domestic violence cases. 258 E.g., Douglas E. Beloof & Joel Shapiro, Let the Truth Be Told: Proposed Hearsay Exceptions to Admit Domestic Violence Victims’ Out of Court Statements as Substantive Evidence, 11 Colum. J. Gender & L. 1, 2 (2002) (advocating for exception).

It is worth further exploring these options, though they carry reliability concerns of their own. As another option, many have encouraged the use of alternative dispute resolution for custody cases, limiting trials and thus issues of problematic evidence. 259 See, e.g., Emery et al., supra note 16, at 20–22 (proposing further reliance on alternative dispute resolution); Scott & Emery, supra note 16, at 105–08 (supporting use of collaborative divorce and mediation).


A judge’s decision to admit or reject evidence of domestic violence in custody cases can have profound consequences for parents’ rights and their children’s lives. For this reason, courts should be especially careful when scrutinizing expert testimony regarding domestic violence in child custody cases. 260 See Dale & Gould, supra note 3, at 6 (“The best interests of children are ill-served when flawed reports become the basis upon which the trier of fact rests his or her judicial decision.”). The evidence offered in these cases, and encouraged by current state rules and statutes, may often fail to meet the threshold of scientific reliability. In order to prevent problematic expert testimony from unduly influencing custody cases, legislatures, experts, and courts must reconsider current laws and practices, while remaining vigilant of the dangerous prevalence of domestic violence. 261 The words of Justice Cardozo ring true for judges, lawmakers, and scientists alike: “‘The work of a judge is in one sense enduring and in another ephemeral . . . . In the endless process of testing and retesting, there is a constant rejection of the dross and a constant retention of whatever is pure and sound and fine.’” Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 597 n.13 (1993) (quoting Benjamin Cardozo, The Nature of the Judicial Process 178–79 (1921)).