Minimizing the Influence of Unconscious Bias in Evaluations: A Practical Guide

Origins of Emotion and Unconscious Motivation

The Expert's Past. Motivations may arise as a result of resonance between the present situation and the expert's past. For example, an expert hired by an attorney whose imperious demeanor resembles that of her father may experience an unrecognized motivation to please him. Alternately, an expert may be infuriated by an attorney who is well liked, but who reminds her of her contentious sister, and may unthinkingly opine in a manner unfavorable to the attorney's client. Similarly, an expert whose spouse died during a routine operation and who is now evaluating a surgeon accused of malpractice may be motivated to punish the surgeon with an opinion damaging to him.

The Expert's Personality. Certain types of motivations may be typical of the expert's personality, rather than unique to a given situation. For example, one expert's characteristic motivation to protect the weak may cause him to favor personal injury plaintiffs, whereas another expert's motivation to enforce personal responsibility may cause her to evaluate the same plaintiffs more skeptically. An expert characteristically motivated to avoid doing harm may selectively discount information unfavorable to her evaluee. An expert characteristically motivated by the wish for financial gain to the exclusion of other wishes may be more likely than others to allow this motivation to overwhelm the motivation to be objective. It should be noted that, because an individual's past experience often influences personality, motivations that stem from the expert's personality typically also stem from the expert's past.

Many authors have addressed bias that arises when the expert deviates “from the role of forensic scientist into the role of zealot or crusader, whose efforts constitute a personal or political agenda” (Ref. 1, p 221). Indeed, sociopolitical beliefs, when they are strong enough to influence the outcome of expert work, may be seen as sources of bias stemming from the expert's personality. For example, consider an expert strongly opposed to the death penalty. When asked to evaluate a defendant facing potential execution, she accepts the case rather than recognize that her beliefs might impair her objectivity. Her evaluation is ultimately biased to favor the defendant. This example parallels recent attention to the notion that ethics-related beliefs are likely to influence psychiatrists evaluating the competence to consent to assisted suicide.⁵ In addition to conscious aspects of her sociopolitical beliefs, the strength of the bias may reflect unconscious, personality-based motivations, such as the motivation to protect the defendant or to defy the authority of an unjust government.

Current Situation. Certain situations, by their nature, may give rise to biasing motivations. This truism is reflected in several areas of concern to forensic psychiatrists. Examples include universal tensions such as that between the expert's duty to objectivity and the attorney's advocacy role, as well as biases that arise from specific factors such as the expert's financial desperation or inexperience. An understanding of situational bias also underlies admonitions against evaluating one's patient (because one is motivated to help the patient) and against contingency fees (because one is motivated to help the retaining attorney for financial gain).

At times, situations merely enhance a motivation that already exists because of the expert's personality or past. For example, an expert who has a competitive personality may be particularly vulnerable to the situation-induced motivation to advocate for the party retaining her.

Provocation by Others. It is possible for the expert to experience motivations that stem primarily from provocation by another individual. The distinguishing feature of such situations is that the individual in question provokes similar responses in people other than the expert. For example, a particularly attractive evaluee might motivate many experts to testify on his behalf, a particularly infuriating or off-putting evaluee might provoke many experts to want to discredit him, and a highly controlling attorney might provoke many experts to want to defy him through opinions unfavorable to his client.

Information-Processing Style

The expert's information-processing style may impair objectivity. This topic has been explored by the heuristics and biases literature, which emphasizes that judgments are often biased because individuals interpret complex situations based on inaccurate but simplifying assumptions known as heuristics. One such heuristic is the “availability heuristic,” a tendency to base one's judgments more on prominent memories than on normative data. An expert operating under this heuristic might overvalue the violence risk of a psychotic evaluee based on her memory of one psychotic patient from her training. She would discount research showing that most psychotic individuals are not at increased risk for violence. Recently, attention has been focused on the biasing effect of heuristics on psychiatric disability evaluations.⁶ For an approach to minimizing the effects of heuristics and biases on forensic evaluations, the reader is referred to an excellent article on the subject (see Ref. 7).

Theoretical perspective also may inform the expert's information-processing style. For example, experts trained predominantly in the diagnosis of mental disorders with presumed biological origins, such as schizophrenia or bipolar disorder, may overestimate the likelihood of such disorders in evaluees. Conversely, experts trained in psychodynamic, cognitive-behavioral, or other psychological modes of understanding symptoms may downplay the role of biological etiologies in their forensic evaluations.

Fund of Knowledge

The expert's fund of knowledge that is relevant to an evaluation may affect objectivity. Differences in fund of knowledge may stem from the expert's professional background, including training, clinical experience, or awareness of the most current literature. For example, as Gold has noted, the forensic psychiatrist who does not know that women frequently fail to report unwanted sexual advances in the workplace may, in evaluating a plaintiff alleging sexual harassment, interpret an evaluee's lack of reporting as evidence that she was not distressed.⁸

Bias also may reflect the effects on fund of knowledge of personal factors, including the expert's race, sex, religion, culture, ethnicity, sexual orientation, or early-life exposure. For example, an expert evaluating a litigant from a different culture may mistakenly interpret a history of teenage stealing as evidence of incipient antisocial features, oblivious to the fact that such behavior is a typical rite of passage in the evaluee's culture.

Example

The following example synthesizes concepts discussed herein. It may be helpful to refer to Figure 1 while reading it: Ms. Z. is suing for emotional damages from the owner of a parking lot where she tripped and fell with minimal physical injury. She alleges that she has been very depressed since the incident. In fact, Ms. Z. is malingering her depression, which is no worse than her baseline dysthymia. The fact that she is malingering could be detected by noting inconsistencies in her account, by exploring these discrepancies in an interview, and by interviewing her coworkers.The plaintiff's expert, Dr. A., misses these indications of malingering and opines that Ms. Z. has depression triggered by her fall. This erroneous conclusion—which Dr. A. defends sincerely—is based on a series of decisions made under the influence of unconscious bias favoring Ms. Z. Usually diligent, Dr. A. inadvertently overlooks several inconsistencies in Ms. Z.'s account of her symptoms. During the interview, Dr. A. is impressed by Ms. Z.'s tears and accepts her statements that she “can't concentrate” and that she “is not interested in work” as signs of depression without carefully assessing her baseline levels of concentration or interest in work. Finally, Dr. A. asks the attorney for copies of depositions of Ms. Z.'s coworkers, but does not follow up when the attorney fails to respond to these requests.Dr. A.'s bias stems from at least two factors. First, her memory of a patient who became depressed around the time of a fall leads her to overvalue the link between falls and depression (an effect of information processing). Second, she feels affinity and compassion for Ms. Z. and is motivated to favor her. This motivation has to do with the fact that Ms. Z. has a likable demeanor and inspires many to want to help her (the influence of provocation), that Ms. Z. reminds Dr. A. of a favorite grandmother (the influence of the past), that Dr. A. is generally altruistic (the influence of personality), that Dr. A. has been anxious to do more civil forensic work and wishes to please the attorney (the influence of the situation), and that Dr. A. wishes to be helpful to counteract guilty feelings connected to her husband's recent criticism of forensic work as “heartless” (the influence of provocation).

This example illustrates three important points. First, bias may be the product of numerous underlying factors. Second, motivations are almost always directed at other individuals, some of whom may be connected to the case (e.g., the plaintiff and attorney) and some of whom may be unconnected individuals from the expert's past (e.g., Dr. A.'s grandmother) or present (e.g., Dr. A.'s husband). Third, experts may not recognize their biasing emotional motivations (which may be too difficult to acknowledge) or their nonemotional biases (which may be so ingrained that they are transparent).

Why Detect and Minimize Bias Proactively?

Consider an expert who, during a case, notices herself feeling negative and wonders if she may be biased. She begins to consider her motivations and decides to consult a colleague about potential sources of bias. Such a response to an inkling of bias is probably the primary mode of bias detection among experts, but it is problematic, because most experts will fail to experience such an inkling in response to unconscious biases and thus will fail to engage in the needed introspection.

Simon and Wettstein⁹ have advised that “the forensic psychiatrist should endeavor to be aware of his or her personal biases while, at the same time, making the appropriate correction for these biases.” I propose that a proactive approach—in which the expert actively attempts to discern potential sources of bias, rather than passively awaiting an inkling of them—would greatly improve bias detection and minimization. This suggestion is based on the hypothesis that deliberate cueing may alert the expert to biasing motivations or to nonemotional factors that otherwise might remain unrecognized. To express the idea another way, since a subjective sense of bias is an insensitive measure of unconscious bias, proactive bias detection is indicated to maintain forensic objectivity.

An Approach to Proactive Bias-Detection and Prevention

The following set of introspective tasks is offered as an approach to proactive detection and prevention of bias. The tasks are designed to cue the expert's recognition of motivations and nonemotional factors that previously have been transparent or difficult to acknowledge. Each task asks the expert to examine potential sources of bias from a slightly different perspective. The expert should complete each task deliberately and actively, probably in writing. The case of Dr. A. and Ms. Z. is used to illustrate each step.

1. Consider each question in Table 2. Are there clues that I may be biased?

Example: Consider how Dr. A. might complete this task. She might note that she has been thinking about the case inadvertently, that she likes Ms. Z., that she has felt uneasy with her husband's critical view of forensic work, that she has neglected to insist on review of depositions of Ms. Z.'s coworkers, and that she forgot to follow up on discrepancies that arose during Ms. Z.'s interview.

2. Who are the significant individuals connected to this case? Who are significant individuals in my personal life, past or present, who motivate me in some way?

Example: Dr. A. might list the attorney, the plaintiff and others connected with the case. She also would list people from her personal life, including her grandmother and husband.

3. Consider each individual on the list, asking: do I feel positive emotions (admiration, attraction, affinity, desire, love, pride, tenderness, compassion, or pity) or negative emotions (fear, anger, humiliation, tension, anxiety, apathy, ambivalence, shame, or guilt) in connection with this person?

Example: Dr. A. might note affinity and compassion for Ms. Z., anxiety about the attorney's hiring her in the future, and guilt about her husband's suggestion that forensic work is “heartless.”

4. Consider each individual on the list, asking: have my emotions toward this person given rise to motivations? Am I motivated to help, protect, humiliate, punish, compete with, gain materially from, be liked by, please, control, take advantage of, defy, or appear superior to this person? Am I motivated to avoid harming, being disliked by, being controlled by, being harmed by, or being humiliated by this person?

Example: This task might cue Dr. A. to recognize that affinity and guilt about being heartless motivate her to help Ms. Z. and that anxiety about being hired again motivates her to want to please the attorney.

5. For the emotions and motivations identified thus far, what are the relative influences of situation, of provocation, of my past, and of my personality?

Example: This task might cause Dr. A. to examine the sources of her affinity for and desire to help Ms. Z. The cue to consider the situation may alert her to her theretofore unconscious desire to please the attorney. In considering the role of provocation, she might imagine that Ms. Z. would be likable to many people. In considering the past, she might recognize similarities between Ms. Z. and her grandmother. Initially, she might be unable to recognize any personality factors relevant to her motivation to help Ms. Z.

6. Do emotions, motivations, or nonemotional factors that I have experienced in previous cases play a role in this case?

Example: This task might cause Dr. A. to note that, in the past, others have chided her about the fact that her altruism causes her to “see only the good” in her evaluees. After thinking about it, she realizes that this is an aspect of her personality that may have enhanced her motivation to help Ms. Z.

7. Is my professional background limited with regard to this case?

Example: Dr. A. initially might be unable to think of any limitations in her professional background that would affect the case.

8. Does my personal background limit my fund of knowledge with regard to this case?

Example: This question might cue Dr. A. to recognize that she is unfamiliar with Ms. Z.'s culture and how it would affect the presentation of depression.

9. Which of my decisions could have been approached in several acceptable ways? Included might be decisions regarding:

• The pursuit of collateral information;

• What to address during interviews;

• How to investigate discrepancies;

• Whether the evaluee's mental condition meets given DSM criteria or diagnoses;

• Cause and effect;

• Whether the evaluee's mental state meets specific legal criteria;

• Disputed psychiatric issues; or

• The degree of probability to which opinions are held.

10. Were any of these decisions influenced by the potential motivations or nonemotional factors identified in previous steps?

Example: This task may cue Dr. A. to recognize that her failure to confront discrepancies in Ms. Z.'s account of the fall, her uncritical acceptance of Ms. Z.'s reports of depressive symptoms, and her failure to insist on depositions from Ms. Z.'s coworkers may all have been influenced by her motivation to help Ms. Z.

11. How would an expert without my motivations or nonemotional biases—or with opposite motivations or nonemotional factors—have judged or reasoned differently? Was my work biased by ignoring such reasoning?

Example: Dr. A. might recognize that an expert who is not motivated to help Ms. Z. might have explored all discrepant data, pursued the missing depositions more assiduously, and focused more critically on malingering before diagnosing depression. She might then recognize her failure to take these measures as signs of bias.

12. If work was biased, revise judgments and opinions to incorporate the reasoning identified in step 10.

Example: If Dr. A. recognizes her bias before she has completed work on her report, she might revise her approach accordingly. For instance, she might obtain the missing depositions and learn that Ms. Z.'s symptoms do not reflect a new depression. She might re-interview Ms. Z. (if possible) and find that Ms. Z. responds to probing questions by becoming increasingly evasive and reporting more severe symptoms that conflict with earlier reports. Taking this additional information into account, Dr. A. might opine that Ms. Z.'s symptoms do not meet criteria for major depression, might question the causal link between her symptoms and her fall, and might raise the question of malingering in her report.

Pitfalls in Proactive Bias Detection

Proactive bias detection is challenging. It requires the vigilance to suspect bias when it is not overt, the discipline to question oneself without external compulsion, the strength to resist becoming defensive, and the courage to risk confronting painful motivations in the service of forensic objectivity.

Failure to consider each step fully may reflect the expert's belief that potential bias deserves attention only if overt, ignoring the fact that bias is most problematic when it is unconscious. More difficult may be the expert's belief that he is too intelligent or too sophisticated to overlook bias. In fact, a desire to dismiss any of the steps outlined most likely reflects the fact that it can be difficult to confront biasing aspects of motivation, information-processing style, or fund of knowledge.

The expert can overcome such difficulty by deliberate adherence to the attitude that objectivity is compromised unless proven otherwise. Diligence may be facilitated by imagining a discerning peer who continually suspects bias and demands that the expert prove him wrong. This approach resembles the diagnostic approach to a medical complaint: possible diagnoses are enumerated systematically and none is discounted without explicit consideration and reasoning. Equivalent diligence is appropriate to the expert's consideration of factors that may compromise objectivity.

The proactive attempts at bias detection discussed herein will inevitably fail to identify some aspects of unconscious emotions, motivations, information-processing styles, or fund of knowledge. These factors—which may be too difficult or too transparent to recognize, even after deliberate introspection—may be identified through supervision and peer-review, forums useful not only for pointing out potential biases, but also for reassuring the expert that bias is common, present in worthwhile individuals, and appropriate to explore. Thus, experts should have a low threshold for soliciting colleagues' input.

It should be noted that the approaches to detecting bias that have been described herein do not address the topic of bias prevention—a topic that has been discussed elsewhere and bears more thorough examination in the literature. Approaches to bias prevention may include further education in psychiatry, forensic psychiatry, and ethics;¹⁰ ongoing peer supervision and consultation; psychotherapy targeting the emotional factors that may impair objectivity;¹¹ training and supervision from more experienced forensic psychiatrists; testifying on both “sides” of significant issues; and sheer experience as a forensic psychiatrist.¹²