Pattern Recognition: The Mark of Expertise
(Originally published May 2021)
Last month’s Sounding Board, “Critical Thinking in Child Welfare,” discussed decision making errors associated with overreliance on gut feelings and quick intuitive judgments based on thin information ( “What You See Is All There Is”). Much of this discussion was taken from Daniel Kahneman’s, Thinking Fast and Slow, (2011), one of the most important books in psychology published in recent decades. This commentary discusses a different perspective regarding intuitive judgments, seemingly conflicting with Kahneman’s distrust of intuition, based on Gary Klein’s naturalistic studies of decision making summarized in Sources of Power: How People Make Decisions (1998). In studying expert decision makers (as identified by their peers) in firefighting, nursing, submarine command and chess, Klein found that experts did not typically use logic models; rather they depended on holistic pattern recognition, especially in high stakes decisions made under extreme time pressures. Furthermore, the intuitive pattern recognition of experts led to skilled responses. Experts not only recognized patterns (which Klein calls prototypes) that their less skilled professional peers failed to recognize; they also had an immediate skilled response in their repertoire. Klein refers to the combination of holistic pattern recognition and skilled responses as “recognition primed decision making.”
Pattern recognition in child welfare
Child welfare caseworkers with at least 12-18 months experience should be able to quickly recognize the following:
Sue, a young single mother, has recently regained custody of her 2-year-old son, Justin, after almost a year of inpatient and outpatient drug/alcohol treatment. Sue is enrolled in a parenting skills program. The family caseworker is impressed with Sue’s commitment to her son and her desire to be a good parent. “I was a terrible parent when Justin was a baby, but I’ll never neglect my son again,” Sue has said to the caseworker. During the first 3 months following reunification, Justin attends a therapeutic child-care program regularly and appears to be thriving. Justin then begins to miss his childcare program about every other day. Sue is not at home on three successive visits. She calls the caseworker with excuses after each missed appointment and assures the caseworker that everything is ok.
Sarah’s two children, ages 2 and 4, have been in foster care for 15 months due to Sarah’s polydrug abuse and related neglect. Sarah’s children have adjusted well to their foster home, even though the 4-year-old is distraught at the end of visits. Sarah has been in inpatient treatment on two separate occasions during the past 15 months and has successfully completed parenting classes. The caseworker has planned to return the children on two occasions due to the mother’s progress in drug treatment. However, three days before the scheduled date of reunification, Sarah fails a UA and then accuses the caseworker and drug testing facility of falsifying the results. Sarah subsequently relapsed and reentered drug treatment. Sarah has now moved in with an old boyfriend, a convicted drug dealer with a DV history, two weeks before the children are scheduled to be returned for a trial home visit. The boyfriend has refused to engage in a substance abuse assessment or engage with the caseworker in any way. Sarah is furious with the caseworker whom she accuses of “finding any reason she can to keep my kids in foster care.” Sarah is demanding that a new caseworker be assigned to her case, someone who is “on my side.”
The first scenario is relapse; the second scenario is self-sabotage, both of which occur so frequently during reunification that they should be easily recognizable to experienced caseworkers, drug treatment staff and therapists. When relapse occurs post-reunification, a competent conscientious caseworker will seek out children immediately (without delay) to check on their safety, and then quickly arrange for a parent to meet with a substance abuse treatment specialist acquainted with the parent. A skilled caseworker will neither exaggerate or minimize the significance of relapse; and will make safety decisions based on her/his safety assessment and after consultation with the substance abuse counselor, not on something she/he may have heard in training regarding the meaning of relapse in recovery; or based on the assumption that relapse inevitably leads to safety threats.
Self-sabotage accompanied by adamant denial and blaming of others is something different than conscious ambivalence of someone on the verge of achieving a goal; and requires a different response. Ambivalence of a parent on the verge of reunification is as common as the reservations and uncertainty of engaged persons approaching their wedding date. Some parents are honest with themselves (at least) regarding ambivalence as they are about regain custody of their child; others are not, despite the loud and clear message embodied in their behavior. An experienced skilled caseworker will understand quickly whether she/he can have an honest candid conversation with a parent engaging in self sabotage regarding her anxieties and conflicted motivation; or be forced to gradually “chip way” at denial by pointing out the conflict between what the parent says they want and their behavior, without any aggression or disapproval in the caseworker’s tone of voice. Denial with deep emotional roots cannot be overcome merely by reasoning or argument, much less by ridicule or aggressive confrontation.
It is much more difficult to recognize patterns that occur rarely, or infrequently, because expert intuition is developed by repeated exposure to phenomena, not through brief trainings. Consider the following scenario:
Timmy, age 7, was adopted at age 5 by his foster mother, a paralegal, after almost 3 years in foster care and several prior placements. Timmy was both abused and neglected by his birth mother from the time he was a baby. The adoptive mother is frustrated with Timmy because he won’t follow rules, and because of conflicts over meals when Timmy sometimes refuses to eat his dinner. The adoptive mother admits that she has become so frustrated with Timmy’s refusal to eat certain foods she has sent him to bed without supper on a few occasions.
CPS intake receives a report from school that Timmy has been begging for food and juice, but his adoptive mother has insisted that school personnel refuse to “give in” to Timmy’s demands. She has also requested that someone accompany Timmy to the bathroom so that he cannot drink water out of the sink or toilet. The mother states that her pediatrician agrees that Timmy should receive no food or water except at mealtimes or snack time. Timmy’s teachers are upset with these directions as Timmy has been lethargic and ‘spacey’ in class. Timmy has also had unexplained bruises on his arms and face on several occasions.
This scenario is taken from a famous case in Washington State in which a 7-year-old boy died of dehydration after attempting to break out the screen of his bedroom window the evening before his death to eat snow! This child had been systematically denied adequate food and water for months during a power struggle gone wrong; one of the first cases of torture of a school age child to result in death in this state. There have been several similar cases in Washington and around the country in recent years in which CPS caseworkers failed to recognize cases of torture that included deliberate starvation of a child (sometimes by a foster parent or adoptive parent), as well as physical abuse and humiliation of the child. In some of these cases, caseworkers have been criminally prosecuted for failure to protect a child from horrible mistreatment and death despite numerous CPS reports and investigations. In cases of deliberate starvation of a child, it is common to read child death reports and news stories in which not a single child welfare practitioner, professional, law enforcement officer or reporter refers to torture, or to the dynamics of ugly power struggles in which a caregiver is determined to break a child’s will whatever the cost.
Caseworkers and other professionals encounter cases of torture infrequently, and so even experienced caseworkers and supervisors may miss the warning signs, i.e., bizarre punishments, combined with ongoing denial of food and water along with large doses of humiliation and scapegoating of a child. The development of expert intuition is developed through repeated exposure to cases that embody specific patterns. It is not enough to listen to descriptions of torture of a child, though watching the TV documentary, “The Trial of Gabriel Fernandez” may be instructive.
Pattern recognition in child protection
Standardized assessment tools that seek to routinize and dumb down assessments of child maltreatment will never be an adequate substitute for expert pattern recognition and recognition primed decision making in child protection. Risk and safety assessment tools have serious deficiencies, some of which are inherent in the tools themselves and others which reflect how they are used, or (more likely) not used:
Their predictive power is too weak (e.g., too many false positives) to be useful in specific cases. Classifying cases as “high risk”, “moderate risk” or “low risk” has a limited utility in decision making, except for the decision whether to open or keep open cases following investigation, or to close cases.
They are of little use in case planning because they do not provide an understanding of family dynamics or of parental needs.
Caseworkers often use the tools primarily at case closure to document assessments and decisions made without use of the tool.
Most of the tools do not connect the dots between and among risk factors; nor do they assist caseworkers in developing a plausible narrative of conditions in families that endanger or protect children.
Assessment tools based on algorithms are not helpful in predicting low base rate phenomena such as child maltreatment deaths, an inherent flaw that cannot be overcome (see Kahneman, p. 192) because of how the tools are developed. In addition, the misuse of these tools in policy and practice has led to a misunderstanding of expertise in child protection and of what is required to improve decision making.
CPS caseworkers and supervisors with a high degree of expertise can do the following:
Recognize the difference between harsh punishment and cruelty in a parent’s approach to discipline of a child and recognize instances of parental cruelty as a warning sign.
Recognize emotional abuse and neglect of a child and understand its impact on child development.
Recognize indicators of nurturance, or lack of nurturance, in parent-child interactions. A skilled CPS caseworker will be attuned to the emotional milieu in which child maltreatment occurs.
Recognize parents’ protective capacities, including their motivation and ability to protect their child from another caregiver.
Recognize the difference between situational neglect and chronic neglect; and focus attention on patterns in CPS report histories rather than the most recent alleged incident of neglect.
Recognize the difference between substance abusing parents motivated to engage in harm reduction behaviors and parents too damaged by their substance dependence or addiction to protect a child.
Recognize differences in the dynamics between different types of physical abuse such as excessive discipline, battered child syndrome, misguided attempts at education and torture of a child.
Recognize indicators of trauma for both children and parents, e.g., PTSD symptoms, emotion dysregulation and explosive behaviors, paralysis, shame, and self-harming behaviors. Experts may or may not be able to analyze and verbally articulate the basis of intuition. For this reason, some expert decision makers are not good coaches or trainers for less experienced staff.
Kahneman and Klein
Given Kahneman’s distrust of intuition and preference for algorithmic based tools and Klein’s view that expert decision makers depend on intuition (along with his low opinion of algorithmic decision making), it is not surprising that these two psychologists have engaged in a spirited discussion and debate, some of it published in journal articles. What is surprising (and admirable) is that Kahneman and Klein were able to maintain civility in this debate, persist in frequent exchange of conflicting views and reach a large amount of agreement regarding the conditions in which expert intuition is possible.
Kahneman asks: “how can we evaluate the probable validity of an intuitive judgment? When do judgments reflect true expertise? When do they display an illusion of validity? The answer comes from the two basic conditions for acquiring a skill:
an environment that is sufficiently regular to be predictable
an opportunity to learn these regularities through prolonged practice.”
Kahneman maintains that “the accurate intuitions Gary Klein has described are due to highly valid cues that the expert’s System 1 has learned to use even if System 2 has not learned to name them.” Kahneman points out that “Expertise is not a single skill, it is a collection of skills, and the same professional may be highly expert in some of the tasks in her domain while remaining a novice in others.”
Kahneman reminds readers that “Studies of chess masters have shown that at least 10,000 hours of dedicated practice (about 6 years of playing chess 5-6 hours a day) are required to attain the highest levels of performance.” The question for child welfare agencies is how to increase the exposure of inexperienced staff to case scenarios consisting of varied combinations of four main factors: poverty, substance abuse, chronic mental health conditions and domestic violence (and a few other less common factors) without overwhelming new staff. One idea is to use the staffing of cases as a learning opportunity for inexperienced staff. A supervisor or experienced caseworker would be required to point out features of cases and patterns that are obvious to them, but unfamiliar to an inexperienced caseworker, either during the staffing or immediately following it.
A model of expertise
Daniel Kahneman has demonstrated an exemplary ability to listen to critical feedback, and to change his mind about issues in which he is emotionally invested. Kahneman’s views regarding the use of intuition in decision making have altered to a significant degree, though he remains fully cognizant of the deficiencies of intuition, e.g., its over confidence, proclivity for stories with fictional elements and willingness to make statistically unlikely predictions.
Kahneman asserts that he has learned “not simply to trust intuitive judgment - your own or that of others - but do not dismiss it either.” Kahneman proposes using a knowledge of base rates, i.e., “facts about a population to which a case belongs,” to inform and anchor intuitions. One main difference between experts and non-experts in any profession is knowledge of base rates as discussed in last month’s Sounding Board. Examples of base rates in child welfare include:
Percentage of families investigated by CPS who are re-reported within 12 months, by risk category (high risk, moderate risk, low risk at case closure).
Percentage of children reunified with a parent who re-enter foster care within 12 months or 24 months of return to the parent’s home.
Inexperienced and veteran caseworkers should steadily work at increasing their knowledge of relevant base rates.
Kahneman’s suggestion is to use analytical abilities and statistical understanding to inform and even re-program expert intuition. The ability to quickly recognize key patterns in cases can be strengthened by statistical information and by research findings.
Kahneman, D., Thinking Fast and Slow (2011), Farrar, Straus & Giroux, New York City.
Klein, G., Sources of Power: How People Make Decisions (1998), The MIT Press, Cambridge, Mass.