Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Though estimates have been reported that 40% of marriages with children end in divorce in the USA, only 10% of divorces require the involvement of a child custody evaluation (Bow, 2006). The remaining 30% of families are able to come to agreements outside of court regarding placement of the child(ren) and visitation of the nonresidential parent (Bow, 2006). Though many attempts have been made to establish guidelines, little research guides what constitutes the best interest of the child (BIC) in the divorce. The first step in developing empirically supported child custody evaluation guidelines is to describe the history and controversies of child custody evaluations, and the problematic current state of science regarding how child custody evaluations are conducted.

Custody Arrangement Controversies

Over that last two centuries, there have been dramatic shifts in legal standards for determining custody. Most of these legal controversies continue. Pre-nineteenth century assumptions about custody were gender biased, with fathers initially being favored based on several assumptions including the belief that the father was better able to provide for the child financially, and the view that children were the property of the father (Kushner, 2006; Waller & Daniel, 2005).

The Tender Years Doctrine

This was followed by a complete reversal seen in the early nineteenth century—by a bias toward the mother largely due to the Tender Years Doctrine, which suggested that mothers should be the preferred guardian for children due to their natural role as caregiver and nurturer, especially in the infant and toddler years of the child’s life (Kushner, 2006; Waller & Daniel, 2005). For more than a century, essentially all custody arrangements were awarded to the mother, with fathers given the burden to prove mothers as unfit (Kushner, 2006).

The Tender Years Doctrine continues to influence court decisions, with studies reporting that older judges, especially, are still more likely to award custody to the mother (Hellman, 1988 as cited in Kushner, 2006). Custody decisions were also heavily influenced by the general societal morals of the time, which resulted in discrimination against homosexual and cohabiting, but unmarried, parents (Waller & Daniel, 2005).

The Approximation Rule or Primary Parent

The American Law Institute (ALI) proposed in its Principles of the Law of Family Dissolution (1996) that parents in contested custody disputes be awarded custody of their children that most closely approximates the allocation of time parents had with children before the dissolution of the marriage. If, for example, the mother was primary caretaker and spent 80% of the time with the children and the father 20% before the divorce, the mother and father would be awarded the same proportion of time, and therefore promote some form of consistency in the child(ren)’s life pre to post-divorce. While the Approximation Rule attempts to overcome vague language related to the BIC construct and also attempts to mitigate parents’ dispositions to viciously attack the integrity or parenting competence of their ex-spouse, it has received much criticism (Warshak, 2007). The Approximation Rule assumes that decisions made regarding primary caretaking roles in the home prior to the separation were permanent, and the primary caretaker, up until the time of the divorce, was the better parent, when this is frequently not the case (Warshak, 2007). An example would be when a couple agrees that a father will work outside the home and the mother will temporarily stay home and care for a young child, with the understanding that she is temporarily putting her career on hold. If the couple were to divorce before she returns to work, this should not freeze time allocations for visitation and prevent the father from having more time with the child, or the mother from re-entering the work force. Furthermore, it is often the case that a stay-at-home parent, due to changed financial circumstances, needs to reenter the workforce thus making time allocations prior to the divorce at least somewhat moot.

Shared Parenting

Shared parenting, joint custody or 50/50 shared custody, is another custody ­arrangement that describes a family arrangement in which parents have equal time, custody, and decision-making powers with their child(ren). The literature has generally supported joint custody or substantial contact with both parents as this arrangement has been associated with better child outcomes post-divorce (Emery, 1999).

While this seems like an important finding, it is interesting to note that joint custody is rarely recommended in cases that involve litigation, as it is uncommon in these situations that both parents agree to joint custody. That is, if one or both parents contest a shared parenting arrangement, it is unlikely that family courts will recommend this. It is estimated that only 5–10% of custody cases end in a joint custody arrangement (Emery, 1999). The definition of joint custody also appears to differ from state to state (i.e., primary physical residence with one parent but substantial time with another, versus completely equal time at both parents’ homes, etc.).

The Least Detrimental Alternative or Psychological Parent

Though not adopted as a statute in any state, the psychological parent is an “all-or-nothing” custody arrangement, which gives sole legal and physical custody to only one parent, and has been an influential theory in the field (Melton et al., 1997). According to this theory, custody is recommended for the parent that has the strongest bond with the child, and that this parent should also have control over the child’s amount of time available to the noncustodial parent. This theory has received much criticism regarding how much control it gives to one parent as well as to problems it raises for a custody evaluator to be able to accurately determine the one parent who is the “psychological parent,” as frequently children are bonded to both parents (Buehler & Gerard, 1995). In addition, no assessment strategy has yet been established to accurately measure this construct.

Same Sex Custody

Though not frequently recommended as a custody arrangement, giving custody of sons to fathers and daughters to mothers has become a recent recommendation (Powell & Downey, 1997). This arrangement contends that children adjust better in situations in which they are with their same sex parent as opposed to the opposite (Powell & Downey, 1997). This is not to say that both parents are not important to children’s adjustment, but that this model assumes that children should primarily reside with same gender parents. Research examining children’s adjustment in same sex custody arrangements indicates positive adjustment in school performance and in relationships (Guttman et al., 1999). However, there is still a lot of missing data regarding outcomes, and no longitudinal studies involving direct comparisons of various custody arrangements have been conducted. It is also important to note that in families with children of both genders, this model would also involve separating the children from one another and the effects of this separation of siblings would also need to be carefully studied.

Father’s Rights Movement

As a part of a backlash to the Tender Years Doctrine, which is alleged to still permeate family courts, father’s rights groups have cropped up in the USA over the past decade (Crowley, 2003). The majority of father’s rights groups contend that men are the victims in family law and are unfairly and unjustifiably discriminated against with regard to child custody and child support (Bertoia & Drakich, 1993). In the 1990s, groups such as the “Divorce Racquet Busters” were established with the purpose of protesting the courts’ discriminatory actions against fathers that they felt caused men’s financial and emotional devastation. The American Coalition for Fathers and Children (ACFC) was established in the 1990s to strive for a unified pro-fatherhood agenda. Local chapters such as Parents and Children for Equality (PACE) in Ohio and Fathers Are Parents, Too (FAPT) in Georgia were organized to attempt to influence state legislators and state courts to become more friendly to a pro-fatherhood agenda (Crowley, 2003).

BIC Doctrine

The doctrine that has received most attention in the past 30 years is the BIC Doctrine. The BIC Doctrine, conceptualized in the mid-1960s, was alleged to be a gender-neutral, child-centered model for custody decisions. The BIC Doctrine replaced parental preference, making room for a child to be awarded to a nonparent, as was seen in Painter v. Bannister, the first case in which custody of the child in question was awarded to the maternal grandparents. Though the first joint custody statute was passed in North Carolina in 1957, awarding joint custody was not a common practice until the 1980s (Waller & Daniel, 2005). Researchers have argued that the BIC Doctrine is problematic, suggesting that the “best interest” is a concrete solution that the courts must find, while the ambiguous nature of the definition often involves potential gray areas or disagreements as to what the “best interest” really is (Emery, Otto, & O’Donohue, 2005). The vagueness of this model has been noted for nearly a third of a century. Mnookin (1975) stated:

Deciding what is best for a child poses a question no less ultimate than the purposes and values of life itself. Should the judge be primarily concerned with the child’s happiness? Or with the child’s spiritual and religious training? Should the judge be concerned with the economic “productivity” of the child when he grows up? Are the primary values of life in warm, interpersonal relationships, or in discipline and self-sacrifice? Is stability and security for a child more desirable than intellectual stimulation? (pp. 260–261).

Controversies in State Law

In addition to the various models described above, states across the USA adopted what they consider to be important factors in deciding the best interests of the child. There is some homogeneity among states regarding important factors to consider; however, as is explained in more detail below, there is vast heterogeneity as well. The fact that states use different factors for conducting custody evaluations has huge implications with regard to the reliability and validity of child custody evaluation methods.

Though states include their own BIC guidelines, most stem from the Uniform Marriage and Divorce Act (1979) which includes: (1) the wishes of the child’s parent or parents involved; (2) the wishes of the children regarding their custody; (3) the interaction or relationship of the children and their parent or parents, siblings, or other individuals who may affect the BIC; (4) the children’s adjustment (to their home, community, school, etc.); and (5) the physical and mental well-being of the parties involved. Though many states adhere to these, there is vast heterogeneity with regard to state law regarding BIC.

In addition to Uniform Marriage and Divorce Act (1979) guidelines, 48 states provide other factors to be considered in determining the best interests of the child. The number of additional factors included in state guidelines differ both in number and in content with the average number of factors to consider being 7, the highest being 17 factors (for both Maine and the District of Colombia) and the lowest being one other factor to consider.

Heterogeneity of the State Laws

Some states include having both parents construct a cooperative parenting plan, though states differ in their guidelines regarding specificity of the plan. Specifically, some states provide areas that they want each parent to cover, while others do not. Often the suggested guidelines differ from state to state. For example, Massachusetts’ state law specifies, “If the issue of custody is contested and either party seeks shared legal or physical custody, the parties, jointly or individually, shall submit to the court at the trial a shared custody implementation plan setting forth the details of shared custody including, but not limited to, the child’s education; the child’s health care; procedures for resolving disputes between the parties with respect to ­child-raising decisions and duties; and the periods of time during which each party will have the child reside or visit with him, including holidays and vacations, or the procedure by which such periods of time shall be determined” (Massachusetts General Laws—Chapter 208—Sections: 28 and 31), whereas Washington state law specifies: “All custody cases must have a proposed parenting plan or agreement to be presented to the court for approval before the final order is put in place. The primary goals of the parenting plan are to: 1) Provide for the child’s physical care; 2) Maintain the child’s emotional stability; 3) Provide for the child’s changing needs as the child grows and matures, in a way that minimizes the need for future modifications to the permanent parenting plan; 4) Set forth the authority and responsibilities of each parent with respect to the child; 5) Minimize the child’s exposure to harmful parental conflict; 6) Encourage the parents to meet their responsibilities to their minor children through agreements in the permanent parenting plan, rather than by relying on judicial intervention; and 7) To otherwise protect the best interests of the child.” (Revised Code of Washington—Title 26—Chapters: 26.09.181, 26.09.220). Washington’s parenting plan guidelines go on to specify: “The plan shall allocate decision-making authority to one or both parties regarding the children’s education, health care, and religious upbringing. The parties may incorporate an agreement related to the care and growth of the child in these specified areas, or in other areas, into their plan. Regardless of the allocation of decision making in the parenting plan, either parent may make emergency decisions affecting the health or safety of the child. a) Each parent may make decisions regarding the day-to-day care and control of the child while the child is residing with that parent; b) When mutual decision making is designated but cannot be achieved, the parties shall make a good-faith effort to resolve the issue through the dispute resolution process.”

The following additional factors are included to demonstrate the vast heterogeneity of state guidelines regarding factors to consider in determining the BIC:

  • The moral fitness of the parties involved (Michigan Compiled Laws—Sections: 552.16 and 722.23)

  • The child’s cultural background (Minnesota)

  • Geographic distance between the parents’ residences (Wyoming Statutes—Title 20—Chapters: 20-2-104, 20-2-107, and 20-2-201)

  • The stability of the family unit (Tennessee Code—Volume 6A, Title 36, Sections 36-4-106)

  • The age of the child (Maine Revised Statutes—Title 19A—Sections: 1501 and 1653)

  • If the child is under 1 year of age, whether the child is being breast-fed (Maine Revised Statutes—Title 19A—Sections: 1501 and 1653)

  • The existence of a parent’s conviction for a sex offense or a sexually violent offense (Maine Revised Statutes—Title 19A—Sections: 1501 and 1653)

  • The demands of parental employment (District of Colombia)

  • The age and number of children (District of Colombia)

  • The sincerity of each parent’s request (District of Colombia)

  • The parent’s ability to financially support a joint custody arrangement (District of Colombia)

  • The consideration of whether the party satisfactorily completed participation in a parenting education program established pursuant to section 46b-69b (Connecticut General Statutes—Title 46b—Chapters 56 and 84)

  • The religious faith of the parents (Code of Laws for South Carolina— Chapter 3 ; Sections 20-3-160, 20-7-100, and 20-7-1520)

It is not difficult to see that the heterogeneity inherent in the differing state guidelines could be problematic. Is it actually the case that a child’s best interest is a function of their state of residency? Or is this unfortunate state of affairs due to the fact that lawmakers are trying their best to respond to the problem addressed in this book: i.e., how to accurately explicate what constructs are components of the best interests of the child? In the absence of clear tested models in behavioral science, lawmakers have done their best to explicate this based on their common sense—but again, this has resulted in tremendous heterogeneity across states. A custody evaluator working in one state would be assessing vastly different issues in his or her custody evaluation from another evaluator in a different state. In addition, if evaluators change location and practice in different states, they will need to adhere to entirely different guidelines. Though it is impossible to know how differently any given custody evaluation might turn out if it were conducted in different states, the possibility for a different outcome (based solely on location) is very concerning. A further examination of these differing state guidelines is provided in Appendix A.

It is probably fair to say that no state systematically developed BIC criteria in consultation with expert input from mental health professionals. This could account for the wide degree of variability across states in their specification of the BIC standard. It might be the case that if this construct was more clearly explicated and found to produce better custody evaluations, states could adopt this and the unwanted variability across states could decrease.

Homogeneity of State Laws

Despite the varying numbers of factors set forth in different states’ laws, there is some consensus across states regarding the most salient factors. However, even within the homogenous factors, there are still significant discrepancies.

Thirty-three states are in agreement that the wishes of the child should be taken into consideration, though differences are seen in determining when and under what conditions the child’s wishes should be considered. Some states’ statutes report that the child’s wishes should be considered “if he or she is sufficiently mature to express reasoned and independent preferences as to the parenting time schedule” (Colorado Statutes—Article 10—Sections: 14-20-123, 14-20-124, and 14-20-129), or “if the child is of sufficient age and capable of forming an intelligent preference” (Connecticut General Statutes—Title 46b—Chapters 56 and 84) or a “meaningful preference” (Maine Revised Statutes—Title 19A—Sections: 1501 and 1653), though no specific age is provided leaving this as a source of heterogeneity in actual custody evaluations. Some states leave this discretion to the judge, such as “if the court considers the child to be of sufficient age to express preference” (Michigan Compiled Laws—Sections: 552.16 and 722.23).

Other states add age provisions, such as “the wishes of the child, with more consideration given to the child’s wishes if the child is at least fourteen (14) years of age” (Indiana Code—Title 31—Article 15—Chapters: 17-2-8, 17-2-8.5, and 17-2-15), though these are inconsistent as well, as other states report, “the reasonable preference of a child over 12 years of age” should be considered (Tennessee Code—Volume 6A, Title 36, Section 36-4-106). Utah’s state Code specifies, “The desires of a child 16 years of age or older shall be given added weight, but is not the single controlling factor” (Sections: 30-2-10, 30-3-5, and 30-3-10). Still other states report that “the desires and wishes of the minor child if of an age of comprehension regardless of chronological age, when such desires and wishes are based on sound reasoning” should be considered (Nebraska Statutes—Chapter 42—Section: 364).

Twenty-eight states are in agreement that the quality of the relationship the child has with the parent is important to consider, though states differ in how they define this. For example, some states report it is important to consider “the love, affection, and other emotional ties between each party and the child” (Louisiana Code of Civil Procedure—Article: 131, 132, 133, and 134), while other states report “the intimacy of the relationship between each parent and the child” should be considered (Minnesota), and the majority cite the Uniform Marriage and Divorce Act (1979) consideration stating “the interaction or relationship of the children and their parent or parents, siblings or other individuals who may affect the best interest of the child” should be considered.

Twenty-two states’ laws include guidelines related to parents’ abilities to co-parent. The way in which states define this differ somewhat from “the past and present ability of the parents to cooperate with each other and make decisions jointly” (Alabama State Divorce Code, Chapter 3, Sections 30-3-150 and 30-3-152), to “the willingness and ability of each parent to facilitate and encourage a close and continuing relationship between the other parent and the child” (Alaska State Divorce Code), to “the ability of parents to cooperate in the rearing of their children; and methods for resolving disputes regarding any major decision concerning the life of the child, and the parents’ willingness to use those methods” (Minnesota State Divorce Code), and to “the ability and disposition of each parent to foster a positive relationship and frequent and continuing physical, written, and telephonic contact with the other parent, except where contact will result in harm to the child or to a parent. (a) The support of each parent for the child’s contact with the other parent as shown by allowing and promoting such contact. (b) The support of each parent for the child’s relationship with the other parent” (New Hampshire Statutes). Still other states’ statutes discuss co-parenting in terms of negative behavior one or both parents have engaged in that would assume poor ability to co-parent (e.g., “Whether the residential parent or one of the parents subject to a shared parenting decree has continuously and willfully denied the other parent’s right to parenting time in accordance with an order of the court”—Ohio Code).

Additional similarities that were found in the states’ statutes on determining the BIC include:

  • Twenty-six states report that the environmental stability of the home environment and the parent’s ability to meet the needs of the child are important to consider.

  • Twenty-five states report that domestic violence between spouses is an important factor to consider.

  • Twenty-four states make explicit statements about considerations of child abuse or potential for child abuse.

  • Twenty-one states statutes include the consideration of the parents’ mental health, including considerations about substance abuse or dependence.

A more complete review of the state laws regarding child custody evaluations is provided in Appendix A. Though it is perhaps a step in the right direction to offer concrete factors that may determine the BIC, it is unknown what formula various states have used to determine their guidelines. The guidelines do not offer evaluators guidance regarding the relative weight and importance that should be given to various factors or how to assess these factors (or even more radically if there are any assessment strategies to validly measure these) or how to combine these to produce actual custody recommendations.

Controversies Regarding Existing Guidelines

In addition to common custody doctrines and statutes that have been provided from specific states, several organizations have published standards and guidelines for conducting child custody evaluations. These organizations include the American Psychiatric Association Task Force on Clinical Assessment in Child Custody (1988), the Association of Family and Conciliation Courts (1994), and the American Psychological Association (APA) (1994). We provided the APA guidelines to illustrate several points:

  1. I.

    Orienting Guidelines

    1. 1.

      The primary purpose of the evaluation is to assess the best psychological interests of the child. (Note: this is interesting and potentially problematic because not all states have adopted this standard).

    2. 2.

      The child’s interest and well-being are paramount.

    3. 3.

      The focus of the evaluation is on parenting capacity, the psychological and developmental needs of the child and the resulting fit. (Note: again, the APA guidelines are mute on the variability across state laws regarding what these specify and how they should be measured).

  2. II.

    General Guidelines

    1. 4.

      The role of the psychologist is that of a professional expert who strives to maintain an objective, impartial stance.

    2. 5.

      The psychologist gains specialized competence. (Note: but how, especially given the dearth of knowledge in this area?).

    3. 6.

      The psychologist is aware of personal and societal biases and engages in nondiscriminatory practice. (Note: what exactly are these biases and are there any specific to custody evaluations, e.g., there are allegations that these are biased against both genders—which is true?).

    4. 7.

      The psychologist avoids multiple relationships.

  3. III.

    Procedural Guidelines

    1. 8.

      The scope of the evaluation is determined by the evaluator, based on the nature of the referral question. (But isn’t there a commonality here—what arrangements are best for the child?).

    2. 9.

      The psychologist obtains informed consent from all adult participants, and as appropriate informs child participants.

    3. 10.

      The psychologist informs participants about the limits of confidentiality and the disclosure of information.

    4. 11.

      The psychologist uses multiple methods of data gathering. (Note: again, this only rules out single source evaluations—which isn’t ruling out much).

    5. 12.

      The psychologist neither overinterprets nor inappropriately interprets clinical or assessment data. (Note: this is obvious; who would argue otherwise?).

    6. 13.

      The psychologist does not give any opinion regarding the psychological functioning of any individual who has not been personally evaluated.

    7. 14.

      Recommendations, if any, are based on the psychological best interests of the child. (Note: again, but what exactly does this mean exactly; what constructs are involved and how should these be measured and synthesized into custody recommendations?).

    8. 15.

      The psychologist clarifies financial arrangements.

    9. 16.

      The psychologist maintains written records.

Though face valid, the APA guidelines have been considered “largely truisms” (O’Donohue & Bradley, 1999, p. 317) and read more as a collection of general ethical standards rather than providing the custody evaluator with a systematic model and concrete tools to conduct high quality, standardized evaluations. In addition, these guidelines make some practical sense, though the broad domains covered in the standards (i.e., interviews, testing, direct observation, record reviews, collateral contacts, etc.) leave too much room for subjective bias, uncontrolled variance, and inappropriate, overreliance on clinical judgment. In addition, recommended assessment tools used in custody evaluations have not been standardized, leaving room for the use of inappropriate testing either because of general validity problems of the test e.g., the Rorschach because these measure irrelevant constructs.

The standard practice in conducting child custody evaluations often includes an interview with both parents (either alone or together), an interview with the child or children, direct observation of child–parent interactions, conducting a home visit to each parent’s home to assess the quality of the living environment, and an interview with collateral sources provided by both parents (i.e., teachers, primary care physicians, nannies, family members, etc.), a review of relevant records or materials provided by both parents, and psychological/intelligence testing of both parents and children (Mart, 2007).

In addition to organizations publishing guidelines for child custody evaluations, individual psychologists have suggested methods for the best way to conduct custody evaluations, with varying attention to the construct of the BIC. For example, Gould’s (1998) Conducting Scientifically Crafted Child Custody Evaluations provided information regarding child custody guidelines, and what other researchers (i.e., Grisso’s Objectives of Competency Evaluations) have purported to be useful in conducting custody evaluations. Gould discussed interviewing parents, children, and collateral sources; the limitations surrounding validity in evaluations and assessments [and mentions the ASPECT, Bricklin Perceptual Scales (BPS), Minnesota Mutliphasic Personality Inventory-2 (MMPI-2) and Millon Clinical Multiaxial Inventory-II (MCMI-II) as frequently used scales]; suggested variables to consider in child custody decision making; and finally discussed positive, as well as deficient, parenting behaviors to consider. Gould reported, “Now, the question becomes how to integrate the data; assign weights to the relative meaning of each factor; and make a responsible, fact-based decision about the lives of each person within this changing family system” (Grisso, 1998, p. 225). Gould reported that there are no rules to follow and that “any custodial determination is, after all, a judgment call” (p. 226). This unfortunately given the present state of affairs is correct. For example, interviews can focus on a number of constructs, e.g., job performance, substance use, sexual interests, presence of phobias, early childhood history, future plans, domestic violence, relationship with siblings, and so on. Which constructs ought to be covered in the interview and which ought not? There is insufficient specification of this key issue.

Additionally, Gould provided a list of 13 factors for the custody evaluator to consider. These factors include:

  • Do not rely on memory (i.e., don’t rely on recollection of interview data over other methods)

  • Document (i.e., videotape, audiotape, or take thorough notes)

  • Observe, Infer, and then Conclude (i.e., make a number of hypotheses regarding behavioral observations to minimize memory bias before assigning weight to any one hypothesis)

  • Consider simple versus complex decision making (i.e., be aware of the issue of incremental validity in gathering data and consider your decision-making strategy and how you weigh each variable in making recommendations)

  • Formulate specific psycholegal questions (i.e., ask the court what they want you to evaluate/if there are specific concerns to address)

  • Good questions lead to well-chosen measurement tools (i.e., if the court is concerned about psychopathology of the father, choosing measures of adjustment or psychopathology should be used)

  • Behavioral prevalence and base rates (i.e., having knowledge of the population and base rate behaviors in that population—e.g., behaviors related to depression can be commonly found in parents in contentious divorces)

  • Confirmatory bias (i.e., tendency to look to information that supports evaluator’s hypotheses rather than information that might be contradictory to expectations)

  • Covariation and illusory correlations (i.e., whether variables have a true relationship between them or they may only appear to but in actuality are not related)

  • Hindsight bias (i.e., a tendency to examine behavior retrospectively and draw conclusions from it rather than place it in context)

  • Overconfidence (i.e., similar to confirmatory bias, it is important for evaluators to not become too confident in their assessments as this might cause them to overlook opposing ideas)

  • Focus and use of unique data (i.e., using multiple sources to establish trends in a person’s behavior rather than a single event)

  • Correlated measures (i.e., be cautious when using measurements with high correlation as they may be redundant)

While these are useful factors to consider, these are also common heuristic errors that psychologists frequently make in all areas of practice (i.e., in individual therapy, case conceptualization, etc. These are important to be aware of when conducting child custody evaluations though, because of the generality of these there still may not be enough structure to provide adequate guidance that would lead to reliable and valid custody evaluations.

Rohrbaugh’s Comprehensive Guide to Child Custody Evaluations (2008) provided extensive information regarding conducting evaluations. She reported that evaluators should assess: (1) the relationship between the child and parent/caregiver, (2) the relationship between parents/caregivers, (3) parenting abilities of each parent/caregiver, (4) psychological health of each parent/caregiver, (5) psychological health of each child, and (6) family dynamics including domestic violence or child abuse. She also attempted to explicate the primary caretaker’s functions which she reported are:

  1. 1.

    Planning and preparing meals

  2. 2.

    Bathing, grooming, and dressing

  3. 3.

    Purchasing and care of clothes

  4. 4.

    Medical care, including nursing and trips to the doctor

  5. 5.

    Arranging and transporting children to afterschool activities and social engagements

  6. 6.

    Arranging alternative childcare such as babysitters and day care

  7. 7.

    Sleeping—bedtime, care during the night, waking in the morning

  8. 8.

    Disciplining, general manners, and toilet training

  9. 9.

    Education, including religious, cultural, and social information

  10. 10.

    Teaching elementary skills such as reading, writing, and arithmetic

Finally, Rohrbaugh identified what she termed “research-based criteria for ­identifying attachment figures” as (1) provision of physical care, (2) provision of emotional care, (3) quality of care provided, (4) time spent with the child, (5) continuity or consistency, and (6) emotional investment in the child. Taken together, Rohrbaugh reported that these make up what data should be collected in each ­custody evaluation. Rohrbaugh reported the use of testing, interviewing parents, interviewing children, interviewing collateral witnesses, conducting home visits, parent–child observations, and reviewing records as being important components of the evaluation (2008). She further provided structured questions to ask in the interviews, though she did not report specifically how these structured questions were selected (i.e., the use of incomplete sentences with child interviews such as “my favorite food is…”). Furthermore, it is not made clear as to how this information ultimately contributes to the formulation of custody recommendations. Rohrbaugh (2008) extended typical practices to attempt to increase the reliability of evaluations. She included numerous additional areas that increase the comprehensiveness of her approach (i.e., by including risk factors and their effects on children, by including nontraditional families and recommendations for them, and by including more specific, structured approaches to interviews).

Rohrbaugh’s book is useful as it provides more specificity than APA’s vague guidelines and more than other models. However, she still fails to present arguments that her model is directly attempting to explicate the construct of the best interests of the children, and, in fact, her model can be critiqued as not being well grounded in this legally relevant concept.

Stahl’s Conducting Child Custody Evaluations (1994) also advocated the use of parent interviews, child interviews, collateral interviews, psychological testing, parent–child observations, and home visits in evaluations. Stahl provided sample questions, sample reports, and psychological tests and games for parents and children. The sample questions included “what was your marriage like before the problems began?”; however, he did not specifically explicate the key constructs that ought to be assessed. He also frequently provided examples that suggested he was basing decisions on clinical judgment by using numerous constructs with no clear valid measurement operations (e.g., “…He has a tendency to be defensive, deny common human frailties, and externalize responsibility for many of the problems in his life. He is easily overwhelmed by his emotions and works hard to control them as much as possible…” p. 86—how are any of these to be validly measured?). Because his reasoning is not always clearly explicated, evaluators may find his model difficult to replicate, which can decrease inter-rater reliability. And his model, like all others, has not been shown to have predictive validity, i.e., to actually produce custody arrangements that are in the BIC.

Taken together, the informal models provided by these authors appear to be roughly derived from some understanding of key legal concepts as well as from clinical judgment. However, it is important to note that although the models generally agree on the assessment methods that ought to be used, they are either silent or disagree on the specific constructs that the custody evaluator ought to be assessing. The models are not explicit about key issues such as how to synthesize these ­measurements into actual custody recommendations, and have not been shown to actually increase inter-rater reliability. They also confuse two distinct issues: what constructs ought to be measured in determining the BIC and what measurement operations (i.e., interviews, parent–child observations, home visits, etc.) ought to be used to assess these constructs. Too many models seem to think that it is sufficient to specify an assessment strategy without specifying the aim of that strategy. This is like telling someone who wants to find out something, to make observations. The key is to have a focus, i.e., to observe what, and when, where, and how?

This gap is helpful in explaining why some of the current evaluations lack reliability. Because of the problem and the absence of such a model, which explicates the key constructs that are components of the best interest standard, clinical scientists have called for a moratorium on conducting custody evaluations (O’Donohue & Bradley, 1999). It is clear that despite the BIC Doctrine being the predominant standard over the last 30 years, no clear model for operationalizing the construct has yet been developed. And though it is not known what evaluators have used to base their assessment models and their recommendations on, researchers have studied custody evaluator’s common practices.

Custody Evaluator’s Practices

Many researchers have suggested reasons as to why mental health professionals have come to be relied upon in conducting child custody evaluations. One possible reason for this may be the emphasis placed upon the psychological well-being of the child in custody arrangements (Bala & Saunders, 2003; Bolocofsky, 1989; Mason & Quirk, 1997). Melton et al. (1997) have found that since the adoption of the BIC Doctrine, courts have been depending more and more on expert parenting assessments. Melton et al. (1997) proposed that mental health professionals who specialize in family and child development are desirable professionals to conduct these assessments. It has also been proposed that the ambiguity inherent in the BIC Doctrine combined with the ambiguous and unexplicated psychological constructs which are thought to be important in custody decisions have led to the reliance on mental health professionals in child custody evaluations (Bolocofsky, 1989). Mason and Quirk (1997) have suggested that complex psychological factors (e.g., allegations of substance use, domestic violence, physical or sexual abuse, mental illness, etc.) often involved in custody evaluations make mental health professionals a natural choice for conducting custody evaluations. In addition, many mental health professionals have turned to forensic work to broaden their practices, because of practice and financial restrictions placed upon them by managed care companies (Gould, 2006). Finally, it has been proposed that judges often find custody cases stressful and outside the realm of their profession, and therefore turn to professionals who appear to be best able to conduct custody evaluations with the BIC criteria in mind (Kushner, 2006).

Several studies have examined the practices of child custody evaluators. Gourley and Stolberg (2000) received surveys from 21 psychologists who had been judged to be “highly credible” by attorneys in the area of child custody evaluations. Of the 21 psychologists, 90% indicated they had private practices and 42% described their theoretical orientation as being cognitive-behavioral (with 14% describing themselves as eclectic or interpersonal and 10% as psychodynamic). Over 90% of the psychologists reported that, in their evaluations, they assessed parental mental health, child mental health, parent–child conflict, parental conflict, parenting skills, sibling relationships, and the role of other extended family members. Psychologists also ranked the five most influential factors in making a recommendation as follows: parent mental health, parenting skills, child mental health, parent–child conflict, and parent conflict. However, there was still significant variability in the rankings of these constructs.

Keilin and Bloom (1986) surveyed 82 custody evaluators and discovered that 69% observed parent–child interactions, 50% had observed the parents interacting, 30% went to the children’s schools, and finally approximately 30% contacted third parties, or collateral contacts, to gain more information about the parents and children. Keilin and Bloom asked psychologists to rank, in importance, factors that influenced their custody evaluation and found that the psychologists ranked preference of a 15-year old child as most important, followed by parental attempts to alienate the other parent, and the parent–child relationship (Keilin & Bloom, 1986). In 1997, Ackerman and Ackerman replicated this study with 201 psychologists who conducted child custody evaluations. When asked to rank factors in order of importance in custody arrangements, psychologists in this study ranked parent substance abuse as the number one most important factor, followed by parenting skills, and parents’ attempts to alienate the other parent (Ackerman & Ackerman, 1997). These surveys can be looked at as an attempt to explicate the informal models that custody evaluators are using to operationalize the BIC.

In a method that did not completely rely on self-report of the child custody evaluators, Bow and Quinnell (2002) reviewed 52 child custody reports of doctorate-level psychologists. Demographic information of the psychologists revealed that 62% of the sample were male, 78% Ph.D.s (as opposed to Ed.D.s or Psy.D.s) with average clinical experience of 21 years. Psychologists reported that the average time devoted to child custody cases was about 40% with an average of 22 custody evaluations conducted per year. Upon examining the reports of these evaluators, it was discovered that all evaluators interviewed parents individually, and with the exception of children under 5, almost all also conducted interviews with children (92%). Psychological testing of parents was conducted in almost all cases (90%), though only one-third of children were tested. The majority of cases included parent–child observations (83%) with approximately one-third taking place during a home visit. Significant others were interviewed 75% of the time, although they were tested only 33% of the time. Collateral contacts were therapists 78% of the time. When children were of school age, school personnel were contacted only 62% of the time. Sixty percent of the reports included detailed and specific information provided by collateral contacts, while 17% provided general information, 8% provided a list of collateral contacts but no information provided from them, and 15% did not report using collateral contacts. Less than half of the reports reviewed addressed the best interests of the child, with most instead addressing strengths and weaknesses of the parents, summary of findings, general and then explicit recommendations regarding custody. Physical custody recommendations were made 92% of the time, with legal custody and visitations being recommended less often (85% and 81%, respectively). Therapy was recommended for parents 64% of the time, and for children 40% of the time. Parenting classes were recommended only 11% of the time, and divorce groups for children and parents recommended only 2% and 5% of the time, respectively (Bow & Quinnell, 2002).

Horvath, Logar, and Walker (2002) also examined forensic child custody reports to establish how closely evaluators were adhering to child custody evaluation guidelines set forth by the APA (1994) and by psychologist Clark (1995). Nine guidelines that were taken into consideration in the study included: (1) the psychologist uses multiple methods of gathering data; (2) the focus of the evaluation is on parenting capacity, the psychological and developmental needs of the child and the resulting fit; (3) the same procedures are used for both parties; (4) the evaluator should conduct interviews with both parents, all children, any adults directly responsible for the care of the children, and any party living in the custodial or visiting home. If relevant, day care providers, medical, psychotherapy and school personnel should be interviewed; (5) interviews with parents should assess personal and legal history, drug and alcohol use, emotional problems, current living situation, health status, and employment status; (6) formal and psychological testing for adults is recommended; (7) parents and children should be observed interacting with each other; (8) children should be interviewed and assessed with psychological testing as is age appropriate; and (9) home or school visits may be conducted if further information is needed. Results indicated that of the 82 evaluations examined, 87% assessed parenting skills, 80% assessed psychological and developmental needs of the child, and 72% assessed parent’s ability to meet those needs. Eighty-nine percent of the evaluations assessed the mother, and 90% the father, though only 65% assessed the mother’s history and 71% the father’s history. Approximately 60–62% assessed child–parent interactions while in only 9% of the cases, physicians or day care workers were interviewed. Thirty percent interviewed other counselors, and only 24% interviewed teachers. Psychological testing of the adults was used in only 19% of the reviewed cases and testing of children only 11% of the time. Children were assessed 68% of the time, with only 44% assessing the children’s preference (Horvath et al., 2002).

It is clear from both this study and the Bow and Quinnell (2002) study that while child custody evaluators are using recommended guidelines in their evaluations, they are not using them consistently. This is problematic and speaks to the need for the implementation of more structured guidelines, which should be guided by empirical research on the negative impact of divorce on children. Some of this variability again may come from the fact that a conceptual model of the BIC standard has not been explicated. The research cited above emphasizes what assessment methods are being used, but it importantly misses the question of what constructs are being measured by these methods. Guidelines ought to follow such a model, rather than be developed independently of it. In addition, this is a self-report which raises two issues: are evaluators actually doing this (i.e., would their reports alone reflect that they followed guidelines) and second, the validity of their inferences are unknown and tell us nothing about the ultimate question—who gets the children and what guided their decision.

Controversies Concerning Assessments

It is difficult to justify the use of the MMPI-2, the Rorschach, the TAT or any intelligence tests, none of which were originally designed to be used in child custody evaluations, though many evaluators use them in evaluating parents (Emery et al., 2005; O’Donohue, Beitz, & Cummings, 2007). The lack of empirically established alternatives, or even agreed upon constructs for some of these factors (e.g., parent competence, quality of parent–child relationships) may cause custody evaluators to be reticent to stray from the conventional practice of psychological testing that is used in custody evaluations. Several alternative assessments have been developed for the purpose of establishing the BIC in custody evaluations; however, there are numerous problems with them (Emery et al., 2005; Grisso, 2003). Specifically, one of the most widely used tests is the BPS (Bricklin, 1990). The BPS is a projective measure that assesses children’s “unconscious preferences” toward their parents by using a stylus to assess both parents across 32 activities thought to capture parents’ competence, supportiveness, follow-up consistency, and possession of admirable traits. The problem with the BPS is that it has not been empirically validated, or normed, and there does not appear to be any evidence that it actually captures children’s “unconscious preferences” (Emery et al., 2005; Otto & Edens, 2003). While projective tests may be problematic for conducting child custody evaluations, it is possible that they are used because of the problem that parents are highly motivated to present themselves in a positive light. This can distort test findings if parents minimize symptoms (Carr, Moretti, & Cue, 2005). Parents presenting themselves in a positive light can be especially difficult when observing parent–child interactions. Some researchers are even skeptical about the utility of observing parent–child interactions, because the process of being observed changes the authenticity of the interaction, especially in the context of child custody evaluations (Bricklin, 1995).

Similar to best interest guidelines, guidelines for the observation of child–parent interactions are highly face valid. Oftentimes, guidelines include observing parents interacting with each child individually, then together (if there is more than one child in the home), in a structured and then less structured environment (allowing the parent and child to pick a task to engage in or giving them a problem solving or cooperative task). The evaluator is encouraged to look for signs of attachment, communication, and expectations of the parent with the child’s behavior (Ackerman, 1995; Gould, 1998; Schutz, Dixon, Lindenberger, & Ruther, 1989). The use of home visits is encouraged to observe the natural environment in which the majority of parent–child interactions take place so that more natural behaviors may be observed (Schutz et al., 1989). Some problems with these guidelines, which are similar to the problems of the BIC guidelines, are that they are not standardized and are too vague to be useful. They involve numerous threats to validity including reactivity, unrepresentative samples of behavior, unreliable coding systems, and no clear way to synthesize information gathered from them to make sense of how they relate to the BIC (Emery et al., 2005).

Custody Evaluators’ Testing Practices

In examining custody evaluators’ common practices regarding the use of psychological testing, virtually all psychologists reported the use of interviews to gather information, but again, without a formal model of constructs to guide the aims of these methods. They also report using psychological testing with regard to parent and child mental health. Gourley and Stolberg (2000) found that psychologists reported using the following psychological tests, in order of most frequently used: the MMPI, the Child Behaviors Checklist (CBCL), the Thematic Apperception Test (TAT), the Rorschach, intelligence tests, projective drawings and the Children’s Depression Inventory (Gourley & Stolberg, 2000). The use of projective devices is particularly troubling given their psychometric inadequacies and the complete lack of validity data regarding inferences directly relevant to custody evaluations. In addition, when studied, projective testing has been found to have negative incremental validity, i.e., adding projectives has been shown to decrease the accuracy of clinician inferences in areas like predicting violence and suicide (Sechrest).

Keilin and Bloom (1986) surveyed 82 custody evaluators and discovered that 76% used psychological tests with adults (most commonly used in order were the MMPI, the Rorschach, and the TAT), and 74% with children (most commonly the TAT, followed by the Children’s Apperception Test, projective drawings, and the Rorschach). Ackerman and Ackerman (1997) indicated that psychologists continued to use intelligence tests, the TAT and Rorschach when assessing children (92% of the sample reporting using psychological testing with children), and the MMPI, the Rorschach, and the TAT when assessing adults (98% reported using psychological testing with adults). In their study of custody reports, Bow and Quinnell (2002) found that with regard to psychological testing of the parent, the objective personality measures often used included the MMPI (93%) and the MCMI-III (44% of the time). Parenting inventories were used 45% of the time, ­followed by the Rorschach and other projective tests (40%) with child ratings scales and IQ tests being used most infrequently (30% and 22%, respectively). Only 17% of reports included actual test scores to substantiate conclusions made from them.

Based on the lack of standardization of practices or accountability for evaluators, it is not surprising that as many as 35% of a sample of 198 psychologists conducting custody evaluations have received at least one board or ethics complaint, and 10% have received a malpractice suit related to child custody work (Bow & Quinnell, 2002 as cited in Bow, 2006). Obviously the two major pathways for this kind of trouble can be actual problematic practice or disgruntled clients who did not like the outcome of sound practice. Clearly, there is a significant variability with regard to existing doctrines guiding custody arrangements, state statutes regarding important considerations, custody evaluation researchers’ guidelines, organizational guidelines, and custody evaluators’ practices. In addition, there are grounds to have deep concerns about the quality of these custody evaluations as no model has specified target constructs; there is evidence of problematic test use, and there is little to no information about how results from assessment are synthesized into actual custody recommendations.

Research examining children’s adjustment post-divorce is conducted after the evaluation has been made and thus provides little information regarding how the recommendations of the custody arrangement might have influenced this adjustment. Furthermore, some potentially important research questions are impossible to answer as they would involve ethically questionable or impossible methodologies. For example, it would not be ethically permissible to randomly assign children to different custody arrangements and then follow their adjustment. One potential resource to guide evaluations, however, would be to examine longitudinal research that examines important factors that predict children’s outcomes.