Bias in Custody Evaluations

Detecting Bias

There is no blood test for evaluator bias. It is something that happens between the ears of the mental health professional (MHP) and has to be inferred from the way the MHP related to the parties, the data s/he chose to gather, and how the data was ultimately interpreted. It is a fundamental reality that affects evaluators, attorneys, and judges alike simply by virtue of membership in the species homo sapiens. The best an attorney can do when faced with an adverse forensic opinion by an evaluator is to be alert to the red flags or hints of its existence in the evaluator's process or work product.

The list of bias-types that follows represents a partial catalogue of some of the predispositions that:

|

1. the general psychological literature suggests affect judgment ( e.g., Garb, H.N. (1998). Studying the clinician: Judgment research and psychological assessment. Washington, DC: American Psychological Assn., Turk, D. & Salovey, P. (Eds., 1988). Resoning, inference, and judgment in clinical Psychology. New York: The Free Press; Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press); and that
2. were detected by this author during the process of reviewing scores of forensic work-products by others; and
3. have been outlined by previous commentators on this forensic issue ( e.g., Martindale, D.A. (2005a). Confirmatory bias and confirmatory distortion. J Child Custody, 2(1/2), 31-48; Martindale, D.A. (2005b). Bias and prejudice in custody evaluations. The Matrimonial Strategist, 23:9, 3-8; Martindale, D.A. (2010). Top Ten Biases Often Overlooked by Child Custody Evaluators. Association of Family and Conciliation Courts eNews, 5:8; Tippins, T.M. (2007). Custody Evaluations ' Part XXIV: Confirmatory bias. New York Law Journal, 5/01/07, 3; Wittmann, J.J.P. (2002). Evaluating the evaluator: Assessing the quality of forensic custody assessments. The Matrimonial Strategist, December 2002, XX (11), 1-4).

Relational Biases

Affiliative Bias: The tendency to arrive at more positive judgments about litigants whom one finds attractive, to have similar worldviews or beliefs, or to be generally likable, etc. One example would be the politically liberal MHP who favors a father who is a former war protester over a mother who is a conservative Republican.

Disaffiliative Bias: The tendency to arrive at more negative judgments about litigants one finds interpersonally unattractive, distasteful, or difficult, like a father who calls the MHP's office many times between sessions to report, at great length, the latest in his ex-wife's upsetting behavior.

Group Membership Bias: The tendency to have one's judgments skewed by gender, cultural or racial stereotypes, as when an evaluator with privately held, borderline-mysogenistic views comes to be known as the person to go to if the desired outcome is father-custody.

Credibility Bias: Referred to by Martindale (2010) as “Jiminy Cricket Bias,” this is the belief in one's power to discern truth-telling, and the effect of this assumption on judgments about two parents, despite little evidence of the MHP's actual ability to accurately detect lying.

Retention Bias: The tendency to have one's judgments of litigants skewed by a disproportionately involved, positive, or lucrative relationship with one side's legal team.

Data-Gathering Biases

Confirmatory Bias: The tendency to seek information that supports one's hypotheses about a family while neglecting the equally important step of seeking disconfirmatory information. One illustration of this bias is when an evaluator suspects a mother is depressed, asks collaterals if they have ever seen her depressed, yet administers no tests that might suggest otherwise and fails to attend to information suggesting the depression issue is resolved.

Data-Selection Bias: The tendency to favor the gathering of a class of information that is irrelevant or marginally related to parenting capacity, leading the evaluator to miss information that is much more probative about this issue. For instance, a psychiatrist might favor the highly detailed gathering of medical history, but fail to ask about attachment/separation issues; or an MHP might emphasize a detailed analysis of why a marriage failed rather than a detailed analysis of parenting skills.

Instrument Bias: The tendency to routinely administer instruments one is familiar with or has developed despite a lack of data establishing the usefulness of those instruments for reliably assessing parenting-related questions or despite data suggesting such instruments may tend toward worrisome inaccuracy (e.g., when an evaluator regularly gives parents the Thematic Apperception Test despite the absence of data indicating the TAT reliably measures parenting constructs).

Inferential Biases

Pathology Bias: A tendency on the part of evaluators, due to their
mental health backgrounds, to perceive, seek evidence for, or emphasize in their reports any indication of psychological symptoms despite the infrequency of such symptoms or their frequent irrelevance to parenting or to the case at hand (e.g., an evaluator with a therapy-focused practice writes pages in her report about the internal psychological conflicts of the parents, possible diagnoses, etc., despite an absence of mental health allegations in the case).

Internal-Factors Bias: The tendency to attribute psychological symptoms or undesirable parent behavior to factors internal to the parent while ignoring or under-weighting situational/environmental explanations. For example, if a father presents as anxious and disorganized, the evaluator would speak about these symptoms as abiding traits, despite the alternative possibility that the evaluation setting and the prospect of losing access to his children are time-limited stressors that might more accurately explain his emotional state.

Primacy/Recency Bias: The tendency to weight the earliest and/or latest information received during an assessment process most heavily, leading to a disproportionate influence by such sequence factors on the overall evaluation. This may be shown when an evaluator receives and reviews a hefty stack of accusatory documents from the father against the mother ' before seeing the mother ' and his notes suggest he may have concluded early on that such accusations were true.

Intervention Bias: This factor, suggested by Martindale (2010), likely flows from the fact that many evaluators started their careers emphasizing therapy. It involves the tendency to provide counseling guidance or mediation-like interventions in cases where their professional role is defined as evaluative (e.g., an evaluator counsels both parents to start interacting with their child in certain new ways, and tries to facilitate a reconciliation between father and son).

Shock Bias: Referred to in the judgment literature as the “availability bias,” this refers to a tendency to weight most heavily information that is most easily recalled, and ease of recall is determined by such factors as the salience, severity, or repetition of the information. For instance, an evaluator might weight most heavily the information about a momentary lapse in judgment by a mother 12 years ago that led to her child's brief coma and concussion, despite the fact that, for all of the subsequent years, the father was drunk most evenings and the mother acted competently.

Conceptual Bias: A tendency to weight assessment factors according to subjective theoretical preferences rather than empirically confirmed principles. To illustrate: A psychodynamically oriented evaluator heavily weights (in favor of the mother) a slight difference in attachment security between a child and her parents, despite the fact that the father is far superior in his capacity to set firm, clear limits on the child's oppositional behavior, and despite a lack of empirical confirmation that such a weighting makes sense.

Parenting Values Bias: A tendency to view parenting behavior that is more congruent with one's own values as superior to that which is not, despite a lack of empirical confirmation that such preferences are associated with better child outcomes (e.g., where an evaluator who works hard to minimize her own child's extracurricular demands writes critically in her report about a mother whose straight-A student takes a violin lesson before school and plays on two basketball teams).

Implications for Practice

It is because the biases and judgment-shortcuts summarized above are so ubiquitous that effective legal advocacy demands careful attention to any hints that an evaluator may have had his or her reasoning distorted in a manner that was unfair to a litigant. This usually requires a careful review of the forensic case-file, including notes, marginal comments, the way a report is structured and the language used to describe the parent-subjects of the assessment. In addition, the particular evaluator's prior statements in other cases and written works may help determine if there is reason to believe he or she has a historical leaning that manifests in biased treatment of the case at issue. With this information in hand, the family law practitioner can better combat the biases that negatively impact a client's case.

The hallmarks of effective and trustworthy forensic work are: 1) even-handed and fair-minded interaction with litigants; 2) collection of data; and 3) interpretation of what has been learned about a family. “Bias,” in the broadest sense, refers to an emotional or cognitive inclination that interferes with an unprejudiced consideration of the data that has been gathered.

Bias can come in many forms. An evaluator may unconsciously favor fathers, a prejudice that quickly becomes evident when he disdains some of the mother's assertions without reason. Another evaluator may set herself up for a preference for the mother's position in a case simply by interviewing the mother first ' and repeatedly ' in advance of ever seeing the father. Still another evaluator may enter an assessment with firmly held pre-conceptions about what is good and bad for children (e.g., the family bed is bad; young children should be with their mothers; etc.) despite the fact that the empirical research in his or her discipline fails to support such ideas.

These cognitive sets and assumptions, however formed, create a kind of lens through which data that is gathered on a family is processed and interpreted. And these biases create the very real potential for errors to be made at the stage where the court is being given an evaluator's “bottom line” about a particular child's needs or a certain parent's skills and capacities.

Detecting Bias

There is no blood test for evaluator bias. It is something that happens between the ears of the mental health professional (MHP) and has to be inferred from the way the MHP related to the parties, the data s/he chose to gather, and how the data was ultimately interpreted. It is a fundamental reality that affects evaluators, attorneys, and judges alike simply by virtue of membership in the species homo sapiens. The best an attorney can do when faced with an adverse forensic opinion by an evaluator is to be alert to the red flags or hints of its existence in the evaluator's process or work product.

The list of bias-types that follows represents a partial catalogue of some of the predispositions that:

|

1. the general psychological literature suggests affect judgment ( e.g., Garb, H.N. (1998). Studying the clinician: Judgment research and psychological assessment. Washington, DC: American Psychological Assn., Turk, D. & Salovey, P. (Eds., 1988). Resoning, inference, and judgment in clinical Psychology. New York: The Free Press; Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press); and that
2. were detected by this author during the process of reviewing scores of forensic work-products by others; and
3. have been outlined by previous commentators on this forensic issue ( e.g., Martindale, D.A. (2005a). Confirmatory bias and confirmatory distortion. J Child Custody, 2(1/2), 31-48; Martindale, D.A. (2005b). Bias and prejudice in custody evaluations. The Matrimonial Strategist, 23:9, 3-8; Martindale, D.A. (2010). Top Ten Biases Often Overlooked by Child Custody Evaluators. Association of Family and Conciliation Courts eNews, 5:8; Tippins, T.M. (2007). Custody Evaluations ' Part XXIV: Confirmatory bias. New York Law Journal, 5/01/07, 3; Wittmann, J.J.P. (2002). Evaluating the evaluator: Assessing the quality of forensic custody assessments. The Matrimonial Strategist, December 2002, XX (11), 1-4).

Relational Biases

Affiliative Bias: The tendency to arrive at more positive judgments about litigants whom one finds attractive, to have similar worldviews or beliefs, or to be generally likable, etc. One example would be the politically liberal MHP who favors a father who is a former war protester over a mother who is a conservative Republican.

Disaffiliative Bias: The tendency to arrive at more negative judgments about litigants one finds interpersonally unattractive, distasteful, or difficult, like a father who calls the MHP's office many times between sessions to report, at great length, the latest in his ex-wife's upsetting behavior.

Group Membership Bias: The tendency to have one's judgments skewed by gender, cultural or racial stereotypes, as when an evaluator with privately held, borderline-mysogenistic views comes to be known as the person to go to if the desired outcome is father-custody.

Credibility Bias: Referred to by Martindale (2010) as “Jiminy Cricket Bias,” this is the belief in one's power to discern truth-telling, and the effect of this assumption on judgments about two parents, despite little evidence of the MHP's actual ability to accurately detect lying.

Retention Bias: The tendency to have one's judgments of litigants skewed by a disproportionately involved, positive, or lucrative relationship with one side's legal team.

Data-Gathering Biases

Confirmatory Bias: The tendency to seek information that supports one's hypotheses about a family while neglecting the equally important step of seeking disconfirmatory information. One illustration of this bias is when an evaluator suspects a mother is depressed, asks collaterals if they have ever seen her depressed, yet administers no tests that might suggest otherwise and fails to attend to information suggesting the depression issue is resolved.

Data-Selection Bias: The tendency to favor the gathering of a class of information that is irrelevant or marginally related to parenting capacity, leading the evaluator to miss information that is much more probative about this issue. For instance, a psychiatrist might favor the highly detailed gathering of medical history, but fail to ask about attachment/separation issues; or an MHP might emphasize a detailed analysis of why a marriage failed rather than a detailed analysis of parenting skills.

Instrument Bias: The tendency to routinely administer instruments one is familiar with or has developed despite a lack of data establishing the usefulness of those instruments for reliably assessing parenting-related questions or despite data suggesting such instruments may tend toward worrisome inaccuracy (e.g., when an evaluator regularly gives parents the Thematic Apperception Test despite the absence of data indicating the TAT reliably measures parenting constructs).

Inferential Biases

Pathology Bias: A tendency on the part of evaluators, due to their
mental health backgrounds, to perceive, seek evidence for, or emphasize in their reports any indication of psychological symptoms despite the infrequency of such symptoms or their frequent irrelevance to parenting or to the case at hand (e.g., an evaluator with a therapy-focused practice writes pages in her report about the internal psychological conflicts of the parents, possible diagnoses, etc., despite an absence of mental health allegations in the case).

Internal-Factors Bias: The tendency to attribute psychological symptoms or undesirable parent behavior to factors internal to the parent while ignoring or under-weighting situational/environmental explanations. For example, if a father presents as anxious and disorganized, the evaluator would speak about these symptoms as abiding traits, despite the alternative possibility that the evaluation setting and the prospect of losing access to his children are time-limited stressors that might more accurately explain his emotional state.

Primacy/Recency Bias: The tendency to weight the earliest and/or latest information received during an assessment process most heavily, leading to a disproportionate influence by such sequence factors on the overall evaluation. This may be shown when an evaluator receives and reviews a hefty stack of accusatory documents from the father against the mother ' before seeing the mother ' and his notes suggest he may have concluded early on that such accusations were true.

Intervention Bias: This factor, suggested by Martindale (2010), likely flows from the fact that many evaluators started their careers emphasizing therapy. It involves the tendency to provide counseling guidance or mediation-like interventions in cases where their professional role is defined as evaluative (e.g., an evaluator counsels both parents to start interacting with their child in certain new ways, and tries to facilitate a reconciliation between father and son).

Shock Bias: Referred to in the judgment literature as the “availability bias,” this refers to a tendency to weight most heavily information that is most easily recalled, and ease of recall is determined by such factors as the salience, severity, or repetition of the information. For instance, an evaluator might weight most heavily the information about a momentary lapse in judgment by a mother 12 years ago that led to her child's brief coma and concussion, despite the fact that, for all of the subsequent years, the father was drunk most evenings and the mother acted competently.

Conceptual Bias: A tendency to weight assessment factors according to subjective theoretical preferences rather than empirically confirmed principles. To illustrate: A psychodynamically oriented evaluator heavily weights (in favor of the mother) a slight difference in attachment security between a child and her parents, despite the fact that the father is far superior in his capacity to set firm, clear limits on the child's oppositional behavior, and despite a lack of empirical confirmation that such a weighting makes sense.

Parenting Values Bias: A tendency to view parenting behavior that is more congruent with one's own values as superior to that which is not, despite a lack of empirical confirmation that such preferences are associated with better child outcomes (e.g., where an evaluator who works hard to minimize her own child's extracurricular demands writes critically in her report about a mother whose straight-A student takes a violin lesson before school and plays on two basketball teams).

Implications for Practice

It is because the biases and judgment-shortcuts summarized above are so ubiquitous that effective legal advocacy demands careful attention to any hints that an evaluator may have had his or her reasoning distorted in a manner that was unfair to a litigant. This usually requires a careful review of the forensic case-file, including notes, marginal comments, the way a report is structured and the language used to describe the parent-subjects of the assessment. In addition, the particular evaluator's prior statements in other cases and written works may help determine if there is reason to believe he or she has a historical leaning that manifests in biased treatment of the case at issue. With this information in hand, the family law practitioner can better combat the biases that negatively impact a client's case.