Comment on Child Custody Evaluations

The American Psychological Association has asked for comment on the assessment practice of child custody evaluations.  This is my comment:

---

---

I am a clinical child and family psychologist.  My professional background includes work with a full range of childhood pathologies, with a specialty in ADHD, early childhood mental health, attachment trauma, oppositional-defiant pathology and family conflict. 

In the early 2000s, I was on medical staff at Children’s Hospital of Orange County working on a collaborative intervention project with the University of California, Irvine Child Development Center (Dr. Swanson) on the identification and treatment of ADHD in preschool-age children.  I then became the Clinical Director for an early childhood assessment and treatment center dealing primarily with children in the foster care system.

In 2008, I left the Clinical Director position to enter private practice as I began to wind down my career.  It was when I entered private practice that I first became acquainted with court-involved family conflict and attachment-related family pathology surrounding divorce.  As I began to unravel the pathology, and the mental health response to attachment-related family pathology surrounding court-involved high-conflict divorce, I have increasingly been called on to provide expert consultantion and expert testimony for family law attorneys regarding the assessment, diagnosis, and treatment of family pathology.

In this capacity as an expert consultant and witness regarding clinical child and family psychology, I have often been asked to review as a clinical psychologist the information contained in a child custody evaluation that has been conducted for the family.  As a clinical psychologist with extensive experience with a range of child and family pathology, who has worked throughout my career with top level institutions and projects, I am deeply concerned regarding the apparent profound professional ignorance and deficient standards of professional practice that I have encountered with regard to the practice of child custody evaluations.

Clinical psychologists are trained to specialty practice in assessment. As a clinical psychologist, the practice of child custody evaluations is particularly disturbing to me because child custody evaluations violate every standard of professional practice for the construction of psychological assessment procedures.  In my professional opinion as a clinical psychologist, the practice of child custody evaluation is substantially below the professional standards of practice expected in clinical psychology.

Issues of Prominent Concern:

1.)  No Inter-Rater Reliability

If an assessment procedure is not reliable, it cannot, by definition, be valid.  If an IQ test gives a score for a client of 120 this week (above average) and when this same test is administered the following week it produces a score of 70 (below average) for this same client, this IQ test is not reliable (test-retest reliability).  If the IQ test is not reliable, the results and conclusions of the test cannot possibly be valid (true) because the test results are not stable (one week the test results say the client is a genius, and the next week the test results say the client is cognitively challenged).  It is axiomatic in professional assessment that an assessment procedure MUST be reliable in order to be valid.  Reliability does not ensure that the test results are valid, but reliability is required for validity.

There are four types of reliability in professional assessment, 1) test-retest reliability, 2) inter-rater reliability, 3) alternate forms reliability, and 4) split-half internal consistency reliability.  For the assessment procedure of child custody evaluations that rely on the opinions of the evaluator for interpreting the meaning of the data, the appropriate form of reliability would be inter-rater reliability (that two evaluators would reach the same conclusions based on the same data).  

There is no inter-rater reliability established for child custody evaluations.  Zero.  None.

Without established inter-rater reliability, two different child custody evaluators can reach entirely different conclusions and recommendations based on exactly the same data.  This means that the conclusions and recommendations reached by child custody evaluations represent the lone opinion of a single individual evaluator, and are not necessarily based on any underlying constructs or principles of professional psychology.

If an assessment procedure is not reliable – in the case of child custody evaluations; inter-rater reliability – then the assessment procedure cannot, by definition, be valid.  This is axiomatic in assessment.  This means that the conclusions and recommendations reached by child custody evaluations cannot, by definition, be valid because they are not stable across evaluators. 

The absence of scientifically established inter-rater reliability and hence validity to the conclusions and recommendations reached by child custody evaluations, by itself, represents sufficient reason for discontinuing the assessment practice of child custody evaluations, since the conclusions and recommendations reached by child custody evaluations (and therefore, the assessment procedure itself) are not valid. 

2.)  No Established Validity

The assessment procedure’s reliability is just the ground foundation of establishing the validity of an assessment procedure.  Once reliability is established, the next psychometric procedure is to establish the validity of the assessment.  There are a variety of methods used to establish the validity of an assessment procedure, including face validity, construct validity, content validity, predictive validity, concurrent validity, and discriminant validity. 

No study has ever even tried to establish the validity of the conclusions and recommendations of child custody evaluations.  The conclusions and recommendations of child custody evaluations have no established validity.  The conclusions and recommendations of child custody evaluations are merely the opinions of one person, and these opinions may or may not be accurate.

The absence of established validity (face validity, construct validity, content validity, predictive validity, concurrent validity, discriminant validity) for the conclusions and recommendations of child custody evaluations represents sufficient reason for discontinuing the practice of child custody evaluations because it is not a valid assessment procedure.

3.)  No Operational Definitions

Foundational to the construction of an assessment procedure is to begin by operationally defining the key constructs of the assessment.  With regard to child custody decision-making surrounding the custody evaluation, the key construct is the child’s “best interests.”  However, there is no operational definition for the construct of the child’s “best interests” for the purposes of the child custody assessment.

In their review of forensic practice in child custody evaluation, two leading figures in forensic psychology, Stahl and Simon, describe the problematic definition of “best interests of the child” (note who published this work, The Family Law Section of the American Bar Association):

From Stahl & Simon:  “A critical subject facing those working in the field of family law, whether they’re legal professionals or psychological professionals, is the concept of the best interests of the children. Even recognized experts in this concept differ with regard to what it means, how it should be determined, and what factors should be considered in determining what is in the best interest of a child. Thus, this ubiquitous term escapes consensus and remains fundamentally vague.” (Stahl & Simon, 2013, p. 10-11)

From Stahl & Simon:  “It is defined differently from state to state; and even in Arizona, where there are nine statutory factors associated with the best interest of the child, the meaning behind many of the factors is obscure.  Additionally, when psychologists refer to the best interests of children, they are referring to a hierarchical set of factors that may have different meanings to different children with different families and that may be understood differently by psychologists with different backgrounds and different training.” (Stahl & Simon, 2013, p. 11)

Stahl, P.M. and Simon, R.A. (2013). Forensic Psychology Consultation in Child Custody Litigation: A Handbook for Work Product Review, Case Preparation, and Expert Testimony, Chicago, IL: Section of Family Law of the American Bar Association.

There is no operational definition in the assessment procedure for the key construct of the child’s “best interests.”  Substantial information is collected as part of the child custody evaluation, but what constructs and principles from professional psychology are then applied to the information to reach a conclusion about the “best interests” of the child remains arbitrary and undefined.

Each individual custody evaluator is allowed to apply, not apply, or misapply, any, some, or none of the established constructs and principles of professional psychology.  Child custody evaluators are even allowed to make up idiosyncratic new forms of pathology that are absent a professional-level definition in clinical psychology.

The absence of an operational definition for the construct of the child’s “best interests” would be analogous to conducing an assessment of intelligence without first providing a definition for the construct of intelligence.  In the history of developing an assessment protocol for intelligence, there was a vigorous professional discussion regarding the meaning of the construct “intelligence.” This vigorous professional debate is a good thing because it helps elaborate the nature of the construct being assessed, with differing professional definitions of the construct producing differing assessment protocols that are based on the definition.

For a construct as important as the “best interests” of the child surrounding the child’s post-divorce custody and visitation schedule, a decision that can have profound and life-long consequences for the child, a similarly robust and vigorous professional debate is needed regarding the definition of the construct, the “best interests” of the child.  However, as noted by Stahl and Simon, the construct of the child’s “best interests” has never been defined.  Instead, each individual custody evaluator is allowed to make up their own idiosyncratic definition for the meaning of this term in each individual case, and this definition is not necessarily based on any established constructs or principles of professional psychology.

In the development of an assessment procedure, the first step is to define the construct being assessed; in the case of custody evaluations, it is the “best interests” of the child construct.  The second step is to identify the procedures used to assess the construct definition; the operational definition for the construct.

However, the first step of a vigorous professional debate regarding how the construct of the child’s “best interests” is to be defined has not occurred, and the practice of child custody evaluation skips the second step of developing an operational definition of the construct for assessment purposes.  Instead, child custody evaluations skip directly to collecting the data to assess a non-defined construct.  In skipping the first steps in the professional standards of practice for developing an assessment protocol of defining the construct to be assessed and then creating an operational definition of the construct for assessment purposes, the practice in forensic psychology of conducting child custody evaluations has preempted professional debate regarding the meaning of the construct, “best interests” of the child.

Furthermore, as a clinical psychologist familiar with the scientific literature on parenting and child development, I would submit that the definition of the child’s “best interests” is an undefinable construct in professional psychology because it involves too many variables and unknown parameters.  

There is no information from professional psychology that can provide supported criteria to differentiate the possible outcomes for the child from a 60-40%, 70-30%, 80-20%, or 90-10% custody visitation schedule in any given situation.  No criteria are available from professional psychology for these differential opinions regarding the “best interests” construct.  Furthermore, rendering an opinion on the “best interests” of the child requires prognosticating an outcome for the future development of the child based on these fine-grained current custody visitation schedules.  Predicting the future is beyond the capacity of professional psychology.

The only scientifically and theoretically supported recommendation from professional psychology for post-divorce child custody is for shared 50-50% custody visitation in all cases except diagnosed child abuse (see cultural considerations below).

In family systems therapy (Bowen, Minuchin, Haley), the family is transitioning from an intact family structure that was united by the marriage, to a new separated family structure that is united by the child through the child’s shared bonds of affection with both parents.   It is always in the child’s best interests for the family to make a successful transition to a healthy and cooperative separated family structure. This is the only definition of the child’s best interests supported by the professional literature.

Beyond that broad outcome definition for the child’s best interests (that the family makes a successful transition to a healthy separated family structure following divorce), children benefit from a complex relationship with both parents.  There are four types of primary relationship, each central to the child’s emotional and psychological development:

Mother-son,
Mother-daughter,
Father-son,
Father-daughter. 

Each of these relationship types is unique, and each is of profound emotional and psychological value for the child.  None of these relationships is expendable to the child’s healthy development.  The loss of any of these primary parent-child bonds during childhood will be damaging and traumatic to the child’s development (the death of a parent, the loss of a parent, is a traumatic childhood experience), and there is no supported foundation in the scientific or professional literature that would allow for a professional opinion on the relative costs and benefits to the child’s emotional and psychological development from a 60-40%, 70-30%, 80-20%, or 90-10% visitation time-share in any given situation.

Example: Two Scenarios

Before any assessment procedure can be developed to determine the “best interests” of the child, both the broader definition of the child’s “best interests” must be specified, and an operational definition of this construct for assessment purposes must be provided.  Child custody evaluations should be discontinued as an assessment procedure until an operational definition for the key assessment construct of the child’s “best interests” is identified.

4.) Violation of Principle D of the APA Ethics Code

The forensic psychology practice of child custody evaluation is in violation of Principle D: Justice of the APA’s ethics code.

Principle D: Justice
Psychologists recognize that fairness and justice entitle all persons to access to and benefit from the contributions of psychology and to equal quality in the processes, procedures, and services being conducted by psychologists.

Child custody evaluations are prohibitively expensive for a large number of families, typically costing between $20,000 to $40,000 to complete, and yet child custody evaluations are required by established standards of practice in forensic psychology before a mental health professional can render an opinion on child custody visitation schedules for the family.  The excessive financial cost of child custody evaluations effectively denies lower-income families “access to and benefit from the contributions of psychology” regarding the court’s child custody decision-making with their families.

Child custody evaluations also deny “equal quality in the processes, procedures, and services being conducted by psychologists” by having no established inter-rater reliability for the assessment procedure.  Without any established inter-rater reliability for the assessment procedure, different custody evaluations can reach entirely different conclusions and recommendations based on the exactly same data.  The arbitrary, idiosyncratic, and potentially differing conclusions and recommendations reached by child custody evaluators – who are free to arbitrarily apply, misapply, or not apply, any, some, or none of the established constructs and principles of professional psychology – denies “equal quality in the processes, procedures, and services being conducted by psychologists.”

Child custody evaluations are in violation of two separate components of Principle D: Justice of the APA’s ethics code.  Each violation individually would warrant the discontinuation of the assessment procedure. 

5.) Over-broad & Unanswerable Referral Question

Axiomatic in professional assessment is that the referral question organizes the assessment procedures.  The psychometrics underlying assessment are based on probability and statistics.  An over-broad and unfocused referral question (such as “What’s wrong with Johnny?”) leads to the collection of large amounts of information in a wide range of domains that statistically raise the probability of spurious findings based on statistical probability alone (associations in the data that are just chance associations).  A more focused referral question, on the other hand (such as, “Does Johnny have autism?”), limits the scope of data collection to the information necessary to answer the referral question, thereby limiting the probability of obtaining spurious associations based on chance.

Child custody evaluations seek to answer the referral question, “What should the child’s custody visitation schedule be?”  This is an overbroad referral question that results in the unfocused collection of family history information.  Spurious associations will occur in the information simply as a result of the over-extended collection of data and the psychometric context created by an over-broad referral question. 

In clinical psychology, when an over-broad referral question is initially offered by the client, clinical psychologists (who are knowledgeable in assessment) work with the client to develop a more appropriate limited-scope referral question.  In court-involved family conflict surrounding attachment-related family pathology following divorce, I would propose that a more professionally responsible referral question for assessment is:

Limited Scope Referral Question: “Which parent is the source of pathogenic parenting creating the child’s attachment-related family pathology surrounding divorce, and what are the treatment implications?”

This more limited scope referral question avoids having to operationally define the “best interests” of the child and provides the structure necessary for a structured and limited-scope assessment of pathology (that substantially reduces the financial costs of the assessment, making the benefits from professional psychology accessible to all families, including lower-income families).

A Limited Scope Assessment Protocol

6.) Cultural and Personal Bias

The practice of child custody evaluations is highly vulnerable to cultural and personal bias.  Personal bias would enter the assessment through the evaluator’s own unresolved family of origin issues (counter-transference).  Cultural bias would similarly enter the assessment process through the evaluator’s own cultural context for interpreting the data.

While some may argue that personal bias can be eliminated by self-awareness (not a position I would take), cultural bias is absolutely present in the evaluator’s assessment.  Everything we do as psychologists is embedded within a historical and cultural context.  This is established foundational knowledge of cultural psychology.  The child custody evaluator will be influenced by this evaluator’s historical and cultural context. That is a fact established by the field of cultural psychology.

The issue is not whether the historical and cultural context is influencing the custody evaluator, the issue is limiting the bias inherent to the assessment process against cultural value systems that differ from the White Protestant Northern-European values of the surrounding culture in the United States, and from the culturally embedded personal beliefs and values of the individual custody evaluator.

The more latitude the evaluator is allowed regarding the interpretation of the evaluation data, the greater the potential for unconscious personal and cultural bias to enter the conclusions and recommendations reached by the evaluator.  Both personal and cultural bias in assessment can be substantially limited the more structured the assessment protocol becomes. 

What specific steps do child custody evaluations take to limit the personal and cultural bias of the evaluator?  None.  Guidelines can encourage evaluators to refrain from bias due to personal or cultural beliefs, but these Guidelines have no practical impact on the actual conduct of any individual custody evaluation.  Custody evaluations are conducted by one person and the reports are typically sealed by the court.  Custody evaluations never receive review by any other psychologist regarding the accuracy of the interpretations contained in the evaluation, nor for the potential of cultural bias in the conclusions and recommendations offered.  Because custody evaluations are sealed by the court and are not reviewed for accuracy and bias, an evaluator who introduces  personal bias and the evaluator’s own personal cultural bias into the interpretation of the data, the conclusions reached, and the recommendations made, is never revealed.

There exist NO structural safeguards with the child custody evaluation protocol to limit the impact of personal and cultural bias on the interpretation of family data, on the conclusions reached, and on the recommendations made in child custody evaluations, and to the degree that there are no guidelines specific to how designated principles and constructs of professional psychology are to be applied to interpret the child custody data (identification of relevance and weighting), child custody evaluators are allowed to apply, not apply, or misapply, any, some, or none of the constructs and principles of professional psychology.  This is exactly the type of arbitrary assessment procedure that becomes extremely vulnerable to the introduction of personal and cultural bias into the interpretation of data.

7.) No Professional Oversight or Review

Child custody evaluations are typically sealed by the court to protect the privacy of the child, which is a laudable goal.  However, this creates a professional problem in that the interpretations, conclusions, and recommendations of child custody evaluations are never subject to professional review for accuracy. 

When child custody reports are reviewed, they are typically reviewed by another forensic child custody evaluator as to whether the proper procedures were followed, NOT as to whether the interpretations, conclusions, and recommendations made are accurate.

The financial cost and extensive data collection procedures associated with child custody evaluations essentially prevent parents from seeking second opinions regarding the custody decisions being addressed by the court.  A more limited scope treatment focused clinical assessment protocol can be conducted in a relatively brief time frame (six to eight weeks) at a substantially reduced cost ($2,500) compared to financial cost of a child custody evaluation ($20,000 – $40,000).  This reduced financial cost and time frame for a limited scope treatment focused assessment allows parents to seek a second opinion regarding the symptoms evident and the interpretation of these symptoms.

8.) The Custody Prize

Child custody evaluations support the family pathology of the child’s triangulation into the spousal conflict through the formation of a cross-generational coalition with one parent against the other parent (Minuchin, Haley, Bowen) by making the child a prize to be won by the supposedly “better parent” – which is a symbolic substitution for the “better spouse” within the spousal conflict surrounding the divorce.  Professional psychology has been seduced by the premise that the child represents a prize to be awarded to the “better parent,” as determined by the custody evaluator.

It is an inappropriate role for professional psychology to become the arbiter of who is the “better parent” who should be awarded the “custody prize” of the child following divorce.  Based on the four types of parent-child relationship (father-son, father-daughter, mother-son, mother-daughter) and the foundational principle that children benefit from a complex relationship with both parents, the only recommendation from professional psychology for child custody that is supported by the scientific and professional literature is a 50-50% custody time-share in all cases except child abuse.

If there are family conflicts, this is a treatment issue, not a custody issue.  If the child is evidencing parent-child attachment-related pathology surrounding the divorce, the differential diagnosis becomes identifying which parent is the source of pathogenic parenting creating the child’s attachment-related pathology following divorce

Furthermore, when the custody evaluation procedure seeks and values the child’s expressed preferences for parent, this supports the family pathology of triangulating the child into the spousal conflict by making the child’s beliefs and expressed wishes a prize to be won by the parent, with each parent seeking to convince the child to choose them as the child’s “preferred” parent.  This creates a destructive family environment where each parent seeks to convince, manipulate, and coerce the child into choosing them as the “preferred parent” (creating a loyalty conflict that can rip the child apart psychologically). 

If either parent has empathy for the child and does not put the child in the middle of having to choose a parent, this parent will lose custody to the parent who does put the child in the middle, and who does seek to convince, manipulate, and psychologically coerce the child to select them as the “preferred parent.”  Children should be neutral in the spousal conflict surrounding divorce, and children should never be placed in a position of choosing between parents following divorce. 

To the extent that child custody evaluations seek, value, and report on the child’s “preferences” for parents, the child custody evaluation supports the pathology in the family of a cross-generational coalition by turning the child into a custody prize to be won by whichever parent convinces, manipulates, and psychologically coerces the child to choose them as the child’s supposedly “preferred parent” following the divorce.

Professional psychology should not be in the role of determining the “better parent” who should be awarded the “custody prize” of the child.

Conclusion

Each of these eight issues of prominent concern would individually warrant the discontinuation of the practice of child custody evaluation.  Taken together, they represent a robust and compelling argument for the discontinuation of child custody evaluations.

My recommendation as a clinical psychologist is for professional psychology to get out of the business of identifying the “better parent” who is to be awarded the “custody prize” of the child following divorce.  This can be achieved by the professional recommendation for shared 50-50% custody in all cases except child abuse.

Craig Childress, Psy.D.
Clinical Psychologist, PSY 18857