Foundational Scrutiny of Forensic Custody Opinions

Empirical Deficiencies in Custody Assessment Process

While there is a wealth of information that forensic evaluators can offer courts as they deal with the tangled issues connected to custody disputes, the deficiencies in our knowledge listed below can be effectively utilized during cross examination to make clear to the court that the final custody “choice” by an evaluator, while sounding logical, may be less reliable than it sounds ' that it may actually rest on sand.

There Is No Agreed-Upon Definition of 'Psychological Best Interests'

Custody assessment, together with a choice to tell a court which particular access and decision-making plan is best, essentially involves gathering data in order to make a prediction about which arrangement is psychologically the best (or the least detrimental) option for a child. To make such a prediction, one should have a definition of “psychological best interests” that is clearly defined and empirically proven. Unfortunately there is no such agreed-upon definition in the mental health field.

This yields a child custody evaluation (CCE) process often riddled with the subjective values of an individual evaluator regarding what is good and bad for a child. Are the best interests of the child (BIC) defined by the level of happiness achieved by a child? By the level of self-control? By the display of good academic achievement? By an absence of symptoms? By a disciplinary approach that is firm or one that encourages resistance to authority? By sole placement with one primary parent rather than by shared placement? By the number of the child's close friends? By access to a role model for a good work ethic?

By what combination of these, and the innumerable other possibilities, is the best interest of the child defined? Without an empirically grounded definition of BIC for a road map in the evaluation process, how can the process as a whole be deemed sufficiently reliable?

There is No Agreed-Upon Method for Assessing Psychological Best Interests

There are certain things that virtually all evaluators do, including interviewing parents, interviewing children, and seeing the children together with their parents. However, a glance at texts guiding CCE practitioners quickly makes clear that there are as many models for assessment as there are texts. Bow & Quinell (2001, Professional Psychology: Research and Practice, 32, #3) surveyed CCE practitioners and found that the range of hours spent ranged from five to 90; that 61% of those surveyed did psychological testing with the children in a majority of their cases (meaning that 39% chose not to); that 31% percent held a conjoint session with the parents in most cases (so, 69% did not); that 47% gave intelligence tests to adults (and 53% did not); that 48% gave intelligence tests to children (while 52% did not); that 52% gave the Millon Clinical Multi-Axial Inventory (meaning that 48% did not); and 44% gave the Rorschach Inkblot Test (and 56% did not). To say that there is a lack of consensus about exactly how children and their parents should be assessed would appear to be a gross understatement.

This state of affairs means we have no way of knowing the degree to which the outcome of the CCE process ' a recommended custody plan that may profoundly affect a child ' is the product of the realities of the family before the court or of the unique variations in a particular practitioner's method. It is quite possible that, given no standard approach in the field, different methods may lead to different recommendations for families.

No Instruments Have Been Shown to Adequately Predict 'Best Plans' for Children

When an evaluator chooses a method ' interviews, tests, observations protocols ' and then decides to tell a court which custodial plan is best for a child, he or she is implying that the chosen method (or combination of methods) is sufficiently valid and reliable for that prediction to be made. However, for a variety of reasons (including the shear complexity of the research task), there is virtually no objective research that has confirmed that the methods and instruments we have available, when used in certain combinations, are actually valid and reliable for custody-plan selection. Those instrument-developers who claim they have found a valid and reliable method for doing so have seen their methodologies panned by psychometric reviewers. We have no idea if our processes for evaluating families are in fact valid and reliable.

There Are No Reliable Techniques for Weighing Things We Discover About Families

When custody evaluators offer opinions about the “best” custodial parent or decision-making plan they are, in effect implying they can do the following: 1) Determine the variables appropriate to assess for BIC; 2) Validly and reliably evaluate the presence and intensity of such variables; 3) Estimate a weight that should be given to each of these variables in the ultimate multivariate prediction; and 4) Combine a massively complex field of intra-psychic, behavioral, parental, family-level and extended-family-level variables; apply weights to them; and generate a prediction most likely to produce the best child outcomes.

The process of applying weights offers the clearest glimpse of why the process of selecting custody plans really involves the imposition of subjective values. Should superior empathy skills in one parent be viewed as more valuable for a child than the other parent's superior skills at authoritative limit-setting? Should a mother's occasional dysfunctional drinking be weighted more heavily when reasoning about custody than the father's emotionally harsh parenting style? The choice of how to apply relative weights to such variables is essentially a subjective values choice about which our discipline has little to say ' yet it is precisely that weighting that represents the guts of forensic reasoning in contested custody matters.

The construct “best interests of the child” and the implied weighting process is a socio-moral process (involving competing, subjective values), not a psychological one, and should reside squarely with the finder-of-fact. The devices for measuring or discerning it are not in the mental health practitioner's toolbox.

There Is No Sufficient Base-Rate Data for Most Parent-Child Relationship Variables

Custody evaluators make many judgments about whether certain aspects of parent-child relationships appear to cross a line into a zone of psychological dysfunction or risk. This author has reviewed scores of evaluations in which it is asserted that a parent's low empathy skills, high level of permissiveness, favoring of a busy extracurricular schedule for a child, or struggles with emotional communication are tagged for the reader as problems suggesting dysfunction. However, the psychological literature has minimal information about the rates at which such “deficits” occur in the general population.

Without knowing the rate at which a particular parenting trait occurs in the general population, how can we know if its intensity or frequency in a particular parent should be viewed as outside of normality and worthy of consideration? If we do not have data, for example, on whether the level of parental empathy displayed for a child is in fact low for the general population, should it be asserted ' based simply on an evaluators subjective “take” on a case ' that this empathy deficit ought to be viewed as a worrisome finding?

No Corrective Experience for Evaluators

It is not uncommon in custody matters for a moment to come in the court proceedings when, having been asked for the basis of a particular forensic opinion, the forensic evaluator utters the words, “Well, based on my clinical experience I concluded ' .” In addition, it is commonly assumed among legal professionals that one good way of supporting a forensic evaluator's assertions is to amplify for the court the number of years of clinical experience a particular expert has had, the number of evaluations they have completed, etc. However, the reality is that the clinical experience many experts are referring to has little bearing on the selection of custody plans.

One's “clinical experience” is only relevant if it is direct experience with the issue at hand in a manner that can teach the professional, over time, how to become more and more accurate; i.e., to know when one is “getting it wrong.” If an orthopedic physician suggests a patient do X rather than Y for knee pain, and then hears a week later that this approach made the pain worse, the professional is getting direct (albeit imperfect) corrective feedback so the he can change course in the treatment (and possibly do things differently in future cases). As such feedback comes in over the years, it helps the doctor develop internal rules about what appears, over time, to have worked well or to have made things worse.

Custody evaluators have virtually no similar opportunity to learn. Their reports are mailed in, and they almost never hear of the cases again. Thus, there is no chance to get corrective feedback about whether the recommendations were followed, whether they “worked,” or whether they led to a child's deterioration. Even in custody evaluators' “clinical” practices, where they may do treatment with children of divorce, they have virtually no systematic way to learn what works and doesn't work for children when choosing among custody plans in a controlled fashion. (By no fault of their own, collecting such data in a manner that would be meaningful is, at this point, almost impossible, because we cannot randomly assign children to custody schedules). In a sense there is almost no chance to get the information needed to know if one is getting it right and when one is missing the mark.

Techniques for Measuring Parenting Capacity and Predicting Parenting Behavior Are Still in Their Infancy

The field of psychology has available various methods for assessing parenting attitudes and behavior. However, many of these methods are not ready for prime time and few are designed for the specific application of child custody assessment, where litigants are oriented toward presenting positively in order to prevail.

No Data on Error-Rate Is Available

One important step in the development of diagnostic assessment instruments is research on how often the test accurately identifies the trait or syndrome in question (“hits”) and how often it fails to do so (“misses”). Without such information, we can coast along thinking an instrument is reliable and valid when, in fact, as used in the real world, it is often mistaken in the conclusions it suggests about patients. While this information exists for a reasonable number of instruments, it is non-existent for the CCE process. Custody evaluators have no information to offer courts about how often their particular constellation of interviews, tests, and observation sessions gets it wrong and/or gets it right with respect to the selection of custody plans. This gaping hole in our data on the CCE process is perfectly understandable, given the massive complexity and thorny problems with conducting the necessary research. However, this lack of information should still give the courts pause given that the experts opining on custody do not know how often their methods actually work.

Implications for Practice

Students of evidentiary law will quickly recognize that the empirical deficiencies summarized above offer rich opportunities to mount Frye or Daubert reliability challenges to the methods used for, and to the conclusions derived from, forensic custody evaluations. Strong cross-questions can be easily developed to amplify each of the issues noted above. Going beyond more granular challenges to the process used by an evaluator to a meta-challenge to the CCE process as a whole essentially involves educating the court about what the mental health professional, in fact, does not know despite his or her willingness to recommend a custody plan for a child ' a recommendation that could change the trajectory of the child's life.

It is also important to note the following: Forensic custody assessments are often critically important tools that can offer helpful guidance for the finder-of-fact. The psychological “library” of research on child development, parenting, post-divorce functioning, conflict-resolution, family violence, etc., is massive and extremely relevant to custody matters. However, the limitations noted above suggest a more circumscribed and nuanced role for custody evaluators that does in fact have reasonable empirical support, a role suggested by this author, together with Timothy Tippins, in 2005 (Empirical and Ethical Problems with Custody Recommendations: A Call for Clinical Humility and Judicial Vigilance. Family Court Review, 43).

This approach emphasizes evaluation of psychological risk factors and benefits associated with various custody options without any pretense that our discipline can reliably choose among plans. Such an approach emphasizes educating the court as thoroughly as possible about the advantages and disadvantages of various options, and grounds the conclusions in the best available research. It calls for evaluators to stop short of doing what is essentially the job of the court, a job for which our empirical library leaves us inadequately equipped.

Jeffrey P. Wittmann, Ph.D., a member of this newsletter's Board of Editors, is a licensed psychologist and trial consultant whose national practice concentrates on trial support for attorneys in custody and access matters and on forensic work-product reviews. He is author of Evaluating Evaluations: An Attorney's Handbook for Analyzing Child Custody Reports (MatLaw, 2013). Additional information can be found at www.childcustodyforensics.com.